Advances in Experimental Medicine and Biology 1184

Akihiko Takashima
Benjamin Wolozin
Luc Buee Editors

Tau Biology
Advances in Experimental Medicine
and Biology

Volume 1184

Series Editors
Wim E. Crusio, CNRS and University of Bordeaux UMR 5287, Institut de
Neurosciences Cognitives et Intégratives d’Aquitaine, Pessac Cedex, France
John D. Lambris, University of Pennsylvania, Philadelphia, PA, USA
Nima Rezaei, Children’s Medical Center Hospital, Tehran University of
Medical Sciences, Tehran, Iran
More information about this series at http://www.springer.com/series/5584
Akihiko Takashima • Benjamin Wolozin
Luc Buee
Editors

Tau Biology
Editors
Akihiko Takashima Benjamin Wolozin
Department of Life Sciences Department of Pharmacology
Gakushuin University and Experimental Therapeutics
Tokyo, Japan Boston University School of Medicine
Boston, MA, USA
Luc Buee
Department of Neurology
University of Lille, INSERM,
Boston University School of Medicine
CHU-Lille, Alzheimer & Tauopathies,
Boston, MA, USA
LabEx DISTALZ
Lille, France Program in Neuroscience
Boston University School of Medicine
Boston, MA, USA

ISSN 0065-2598     ISSN 2214-8019 (electronic)

Advances in Experimental Medicine and Biology
ISBN 978-981-32-9357-1    ISBN 978-981-32-9358-8 (eBook)
https://doi.org/10.1007/978-981-32-9358-8

© Springer Nature Singapore Pte Ltd. 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, expressed or implied, with respect to the material
contained herein or for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore
189721, Singapore
Preface

The isolation of tau protein as a heat-stable protein from the microtubule frac-
tion of porcine brain was first reported in PNAS in 1975. The protein was shown
to be essential for microtubule assembly, and most early work on tau were done
in the context of microtubule assembly and the microtubule network.
In 1986, several groups reported that a major component of neurofibrillary
tangles, one of the major pathological hallmarks of Alzheimer’s disease
(AD), was hyperphosphorylated tau. Since then, studies of this protein
expanded to examining mechanisms of tau phosphorylation, aggregation, and
degradation. It was in this era that lovers of tau, so-called tauists, were born.
Mutations in the tau gene were found in frontotemporal dementia with par-
kinsonism-­17 (FTDP-17) in 1998, highlighting the pathophysiological func-
tion of tau, as well the neurotoxic effects of tau aggregation. While therapeutic
approaches targeting β-amyloid failed to show efficacy in slowing the pro-
gression of dementia, the last 20 years has witnessed an increase in attention
to the possibility that targeting tau might be a better approach to tackle AD –
this possibility provides a major stimulus for present-day research on tau.
In this book, we portray pioneering work alongside cutting-edge research
to illustrate the promise of tau as a target for the treatment of AD. Tau is well-­
known as a microtubule-associated protein that is normally localized in
axons. In the diseased brain, neuronal loss is prominently associated with
fibrillary deposits of hyperphosphorylated tau in the remaining neurons.
Important questions remain regarding the pathways and mechanisms respon-
sible for the observed neuropathology and whether tau aggregation represents
a toxic step that triggers neuronal death. New technologies are helping
researchers to address these long-standing questions. This book is intended to
provide the reader with basic knowledge on the biology of tau in order to
fertilize new ideas that may facilitate the development of new treatments for
AD, as well as improve our understanding on the role of tau in brain aging,
since age is the greatest risk factor for AD.
In Part I “Structure and Role of the Tau Molecule,” the authors discuss the
structure, posttranslational modification, regulation of splicing, and degrada-
tion of tau. Since tau is a disordered protein in aqueous solution, even after
heat treatment, studies on its structure have proved challenging. Tau bio-
chemistry is also still poorly understood, especially with respect to its metab-
olism and axonal sorting. This section also considers new knowledge about
tau hyperphosphorylation, how it becomes insoluble, and aggregates to give
rise to neurofibrillary tangles (NFT), a pathological hallmark of the

v
vi Preface

AD-affected brain. The section concludes with a description of the posttrans-

lational modifications of tau that appear to promote its aggregation.
“Tau localization and function” is the subject of Part II, where the physio-
logical role of tau in neurons and the other neural cell types is described. Under
physiological conditions, tau expression is largely restricted to axons where it
regulates axonal transport. The recent demonstration of tau in the postsynaptic
compartment is important because it suggests tau plays a physiological func-
tion in both the axon and the synapse. The precise function of synaptic tau still
awaits further definition; it will be interesting to understand how tau fulfils such
a role, whether by binding to other signaling, trafficking, or structural proteins.
Even more surprisingly, tau is reportedly secreted from dendritic sites to affect
proximal neurons and has recently been shown to contribute to ribosomal DNA
stabilization, opening new possibilities for onco-therapy.
Part III is dedicated to “tau and disease-related proteins,” such as ApoE
and amyloid β (Aβ), which have roles in AD.  The chapters in this section
focus on the relationship between tau and AD-related proteins, which pro-
vides insights into mechanisms that may underlie the clinicopathological
development of sporadic AD.  In addition, studies on dementia with Lewy
body (DLB), the second leading cause of cognitive impairment in the elderly,
are covered since tau deposition is a common accompaniment of DLB, prob-
ably sharing cause-effect properties with synuclein.
“Tauopathies, pathology, drivers, and markers” are considered in Part IV
since pathological observations are informative regarding the causes and/or driv-
ers of disease. The chapters in this section include descriptions of the diverse,
clinicopathological presentations of tauopathies. Since noninvasive imaging
modalities will continue to help investigators better understand the role of sec-
ondary risk factors, such as stress, trauma, diabetes, and inflammation, as causes
or drivers of tauopathy, space is given to description of developments in the
arena of tau PET and fMRI that are likely to aid longitudinal development of tau
pathology in parallel with the measures of impaired brain function.
The last section of this book addresses “tau aggregation and therapy.” In
its native form, tau is an unfolded protein, suggesting that it must be highly
soluble, but in tauopathies, it appears as β-sheeted aggregates. Also interest-
ing is that under physiological conditions, tau can phase-separate, thus
increasing the local concentration of tau and facilitating the formation of tau
“droplets” (membraneless organelles, e.g., stress granules) which may inad-
vertently stimulate the disease-linked aggregation pathway. This section also
discusses how the tau aggregation process may be linked to neurodegenera-
tion, as well as the propagation of the tau aggregation process from the ento-
rhinal cortex to the hippocampus and neocortex, as is thought to be the case
in clinical cases of AD; the work described forms a promising basis for devel-
oping disease-modifying treatments.
The editors are glad to be able to share their enthusiasm for tau pathology
with the reader and look forward to seeing new developments in ­understanding
the biology of this multifaceted protein which may lead to effective therapeu-
tic approaches for AD and other tauopathies.
Preface vii

This book would not be possible without the interest and dedication of its
individual authors – the editors thank them heartily for their time in preparing
timely reviews of their specific interests in tau.

Tokyo, Japan Akihiko Takashima

Boston, MA, USA Benjamin Wolozin
Lille, France Luc Buee
Contents

Part I Structure and Role of the Tau Molecule

1 Ordered Assembly of Tau Protein and Neurodegeneration ������    3

Michel Goedert and Maria Grazia Spillantini
2 Structure of NFT: Biochemical Approach ����������������������������������   23
Masato Hasegawa
3 Nuclear Magnetic Resonance Spectroscopy Insights
into Tau Structure in Solution:
Impact of Post-translational Modifications ��������������������������������   35
Clément Danis, Elian Dupré, Xavier Hanoulle,
Isabelle Landrieu, Alessia Lasorsa, João Filipe Neves,
Robert Schneider, and Caroline Smet-Nocca
4 Regulation of Tau Homeostasis and Toxicity
by Acetylation ��������������������������������������������������������������������������������   47
Tara Tracy, Kathryn C. Claiborn, and Li Gan
5 Tau Clearance Mechanisms����������������������������������������������������������   57
Maoping Tang, Jarreau Harrison, Carol A. Deaton,
and Gail V. W. Johnson
6 Mechanisms of Axonal Sorting of Tau and Influence
of the Axon Initial Segment on Tau Cell Polarity������������������������   69
Hans Zempel and Eckhard Mandelkow

Part II Tau Localization and Function

7 Tau and Axonal Transport Misregulation in Tauopathies���������   81

Benjamin Combs, Rebecca L. Mueller, Gerardo Morfini,
Scott T. Brady, and Nicholas M. Kanaan
8 Presynaptic Pathophysiology Encoded in Different
Domains of Tau – Hyper-Versus Hypoexcitability?��������������������   97
Jochen Martin Decker and Eva-Maria Mandelkow
9 Synaptic Localisation of Tau ��������������������������������������������������������  105
Diane P. Hanger, Despoina Goniotaki, and Wendy Noble

ix
x Contents

10 The Role of Tau in the Post-synapse��������������������������������������������  113

Philip Regan and Kwangwook Cho
11 Tau Secretion����������������������������������������������������������������������������������  123
Zhi Ruan and Tsuneya Ikezu
12 Emerging Connections Between Tau
and Nucleic Acids ��������������������������������������������������������������������������  135
Marie-Christine Galas, Eliette Bonnefoy, Luc Buee,
and Bruno Lefebvre
13 Tau Interacting Proteins: Gaining Insight
into the Roles of Tau in Health and Disease��������������������������������  145
Ilie-Cosmin Stancu, Mattia Ferraiolo, Dick Terwel,
and Ilse Dewachter

Part III Tau and Disease-Related Proteins

14 Relationship Between Tau, β Amyloid

and α-Synuclein Pathologies ��������������������������������������������������������  169
Lauren Walker and Johannes Attems
15 Associations Between APOE Variants,
Tau and α-Synuclein����������������������������������������������������������������������  177
Elena Rodriguez-Vieitez and Henrietta M. Nielsen
16 Amyloid-β and Tau at the Crossroads
of Alzheimer’s Disease ������������������������������������������������������������������  187
Gilbert Gallardo and David M. Holtzman

Part IV Tauopathies; Pathology, Drivers, and Marker

17 Myotonic Dystrophy: an RNA Toxic Gain

of Function Tauopathy?����������������������������������������������������������������  207
Francisco Fernandez-Gomez, Helene Tran,
Claire-Marie Dhaenens, Marie-Laure Caillet-Boudin,
Susanna Schraen-Maschke, David Blum,
Bernard Sablonnière, Valérie Buée-Scherrer,
Luc Buee, and Nicolas Sergeant
18 Tau PET Imaging ��������������������������������������������������������������������������  217
Makoto Higuchi
19 Tau Accumulation and Network Breakdown
in Alzheimer’s Disease ������������������������������������������������������������������  231
Hirohisa Watanabe, Epifanio Bagarinao, Takamasa Yokoi,
Hiroshi Yamaguchi, Shinsuke Ishigaki, Michihito Mausuda,
Masahisa Katsuno, and Gen Sobue
Contents xi

20 Stress and the Etiopathogenesis of Alzheimer’s

Disease and Depression������������������������������������������������������������������  241
Ioannis Sotiropoulos, Joana M. Silva, Patricia Gomes,
Nuno Sousa, and Osborne F. X. Almeida
21 Tau, Diabetes and Insulin��������������������������������������������������������������  259
Maud Gratuze, Aurélie Joly-Amado, Luc Buee,
Didier Vieau, and David Blum

Part V Tau Aggregation, and Propagation

22 Top-Down Projections Direct the Gradual Progression

of Alzheimer-Related Tau Pathology
Throughout the Neocortex������������������������������������������������������������  291
Heiko Braak and Kelly Del Tredici
23 Tau Prion-Like Propagation: State of the Art
and Current Challenges����������������������������������������������������������������  305
Simon Dujardin and Bradley T. Hyman
24 Tau Condensates����������������������������������������������������������������������������  327
Kenneth S. Kosik and Songi Han
25 Liquid-Liquid Phase Separation of Tau Protein
in Neurobiology and Pathology����������������������������������������������������  341
Susanne Wegmann
26 The Pathophysiology of Tau and Stress
Granules in Disease������������������������������������������������������������������������  359
Anna Cruz, Mamta Verma, and Benjamin Wolozin
27 Tau Oligomers��������������������������������������������������������������������������������  373
Sumihiro Maeda and Akihiko Takashima
28 Experimental Models of Tauopathy – From
Mechanisms to Therapies��������������������������������������������������������������  381
Julika J. Götz and Jürgen Götz
29 Cerebrospinal Fluid and Plasma Tau
as a Biomarker for Brain Tauopathy ������������������������������������������  393
Mikio Shoji
30 Dementia Therapy Targeting Tau������������������������������������������������  407
Luc Buee
 Part I
Structure and Role of the Tau Molecule
 Ordered Assembly of Tau Protein
and Neurodegeneration 1
Michel Goedert and Maria Grazia Spillantini

Introduction Tau Protein

Ordered assembly of a small number of proteins Tau monomers belong to the family of intrinsi-
into filaments characterises the majority of cases cally disordered proteins that, upon ordered
of age-related neurodegenerative diseases, assembly, form structured amyloid filaments [14,
including Alzheimer’s and Parkinson’s. Most 37, 85]. Their expression is largely confined to
cases are sporadic, but a small number is inher- central and peripheral nerve cells, where they are
ited in a dominant manner. Huntington’s disease highly enriched in axons [5]. However, tau
is always inherited. Work carried out over the assemblies are observed in both nerve and glial
past 35  years established a causal role for fila- cells in a number of neurodegenerative diseases.
ment formation in inherited forms of disease. By Since assembly is concentration-dependent, it
extrapolation, it appears likely that ordered remains to be established if tau assemblies can
assembly into filaments is also central for neuro- form de novo in glial cells or if glial tau pathol-
degeneration in sporadic cases of disease. ogy requires the uptake of seeds from neurons.
Tauopathies, which are characterised by the Tau protein can be divided into an N-terminal
assembly of microtubule-associated protein tau, region, a proline-rich domain consisting of two
are the most common proteinopathies of the separate parts, the repeat domain and a C-terminal
human nervous system (Table 1.1). They include region. The N-terminal region projects away
Alzheimer’s disease (AD), Pick’s disease (PiD), from the microtubule surface and is believed to
chronic traumatic encephalopathy (CTE), tangle-­ interact with components of the neuronal plasma
only dementia, progressive supranuclear palsy membrane. An interaction between exon 1 and
(PSP), corticobasal degeneration (CBD), argyro- annexins may help to explain the axonal localisa-
philic grain disease (AGD) and several rarer tion of tau [28], which may also be mediated, at
diseases. least in part, by the axon initial segment [57].
Exon 1 of human tau contains a primate-specific
sequence, which has been proposed to mediate
M. Goedert (*)
MRC Laboratory of Molecular Biology, interactions with neuronal proteins [78]. The
Cambridge, UK PXXP motifs in the proline-rich region are recog-
e-mail: mg@mrc-lmb.cam.ac.uk nised by SH3 domain-containing proteins of the
M. G. Spillantini Src family of non-receptor kinases, such as Fyn
Clifford Allbutt Building, Department of Clinical [56]. The repeat domain and some adjacent
Neurosciences, University of Cambridge, sequences mediate interactions between tau and
Cambridge, UK

© Springer Nature Singapore Pte Ltd. 2019 3

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_1
4 M. Goedert and M. G. Spillantini

Table 1.1  Neurodegenerative diseases They differ by the presence or absence of inserts
Alzheimer’s disease of 29 or 58 amino acids (encoded by exons 2 and
Parkinson’s disease 3, with exon 3 being only transcribed in conjunc-
Dementia with Lewy bodies tion with exon 2) in the N-terminal half and
Frontotemporal dementias (including Pick’s inclusion, or not, of the 31 amino acid
disease)
microtubule-­binding repeat, encoded by exon 10,
Progressive supranuclear palsy
Corticobasal degeneration in the C-terminal half. Inclusion of exon 10
Chronic traumatic encephalopathy results in the production of three isoforms with
Argyrophilic grain disease four repeats (4R) and its exclusion in a further
Tangle-only dementia three isoforms with three repeats (3R). The
Multiple system atrophy repeats comprise residues 244–368, in the num-
Huntington’s disease bering of the 441 amino acid isoform. In adult
Motor neuron diseases human brain, similar levels of 3R and 4R tau are
Prion diseases expressed [32]. The finding that a correct 3R and
4R tau isoform ratio is essential for preventing
microtubules. Electron cryo-microscopy (cryo- neurodegeneration came as a surprise. 2  N iso-
­EM) has shown that each tau repeat binds to the forms are underrepresented in comparison with
outer microtubule surface and adopts an extended isoforms that include exon 2 or exclude exons 2
structure along protofilaments, interacting with and 3; 2 N, 1 N and 0 N isoforms make up 9%,
alpha- and beta-tubulin [1, 50]. Single-molecule 54% and 37%, respectively. Big tau, which car-
tracking revealed a kiss-and-hop mechanism, ries an additional large exon in the N-terminal
with a dwell time of tau on individual microtu- half, is only expressed in the peripheral nervous
bules of about 40 ms [46, 65]. Despite these rapid system [20, 34].
dynamics, tau promotes microtubule assembly. The expression of isoforms is not conserved.
Microtubules have stable and labile domains. Tau Thus, in adult mouse brain, 4R tau is exlusively
is most abundant in the labile domain, which has present, whereas adult chicken brain expresses
led to the suggestion that it may not stabilise 3R, 4R and 5R isoforms [89]. One hyperphos-
microtubules, but enable them to have long labile phorylated 3R tau isoform lacking N-terminal
domains [6, 71]. Less is known about the func- repeats is characteristic of developing verte-
tion of the C-terminal region. brates. In mice, the switch from 3R to 4R tau
Although it lacks a typical low-complexity occurs between postnatal days 9 and 18, with tau
domain, full-length tau has been reported to phosphorylation decreasing over time [81].
undergo liquid-liquid phase separation, which However, isoform switching and phosphorylation
has been suggested to initiate aggregation and are regulated differently. Adult 4R isoforms are
neurodegeneration [84, 92]. RNA-binding pro- better at promoting microtubule assembly than
teins may influence these processes [2]. Tau is the 3R isoform expressed during development,
subject to a number of post-translational modifi- which is also more phosphorylated than both 3R
cations, including phosphorylation, acetylation, and 4R tau in adult brain [32]. This is consistent
methylation, glycation, isomerisation, O-Glc-­ with the need for a more dynamic cytoskeleton
NAcylation, nitration, sumoylation, ubiquitina- during the development of nerve cells.
tion and truncation [37].

Tau Assemblies
Tau Isoforms
Brain tau can assemble into filamentous inclu-
Six tau isoforms ranging from 352 to 441 amino sions [7, 37]. The repeats and some adjoining
acids in length are expressed in adult human sequences form the filament core, with the
brain from a single MAPT gene (Fig. 1.1a) [31]. N-terminal half and the C-terminus giving rise to
1  Ordered Assembly of Tau Protein and Neurodegeneration 5

A
Exons: 1 2 3 4 4a 5 6 7 8 9 10 11 12 13

1 441
R1 R2 R3 R4
1 412
R1 R2 R3 R4 4R
1 R1 R2 R3 R4
383

1 410
R1 R3 R4
1
381
R1 R3 R4 3R
1
352
R1 R3 R4

B +3 +4
+11 +12
-15 -10 +13 +14
+15 +16

1 441

E1 E9 E10 E11 E12 E13

R5H, R5L

∆K280

P364S

E372G
S352L
K257T

P301L, P301S, P301T

S356T
I260V
L266V
G272V
G273R
N279K
L284L, L284R
C291R
∆N296, N296N, N296H
K298E
G303V
G304S
S305I, S305N, S305S
L315R
K317M, K317N
S320F
P332S
G335S, G335V
Q336H, Q336R
V337M
E342V
D348G

V363I
G366R
K369I

G389R
R406W
N410H
T427M

Fig. 1.1  Human brain tau isoforms and disease-causing eral nervous system. The repeats (R1-R4) are shown, with
MAPT mutations three isoforms having four repeats (4R) and three isoforms
(a) MAPT and the six tau isoforms expressed in adult human with three repeats (3R). The core sequences of tau filaments
brain. MAPT consists of 14 exons (E). Alternative mRNA from Alzheimer’s disease (G273/305-E380) determined by
splicing of E2 (red), E3 (green) and E10 (yellow) gives rise cryo-EM are underlined (in orange); the core sequences of
to six tau isoforms (352–441 amino acids). The constitu- tau filaments from Pick’s disease (K254-F378 of 3R tau) are
tively spliced exons (E1, E4, E5, E7, E9, E11, E12 and E13) also underlined (in grey). (b) Mutations in MAPT in FTDP-
are shown in blue. E6 and E8 (violet) are not transcribed in 17T. Fifty coding region mutations and ten intronic muta-
human brain. E4a (orange) is only expressed in the periph- tions flanking E10 are shown
6 M. Goedert and M. G. Spillantini

the fuzzy coat [30, 87, 88]. Tau filaments from Table 1.2  Neurodegenerative diseases with abundant
brain, and those assembled in vitro from Tau inclusions
expressed protein, have a cross-β structure char- 3 + 4R Tauopathies
acteristic of amyloids [4]. Since the region that Alzheimer’s disease
Amyotrophic lateral sclerosis/parkinsonism-dementia
binds to microtubules also forms the core of tau complex (Guam and Kii peninsula)
filaments, physiological function and pathologi- Anti-IgLON5-related tauopathy
cal assembly may be mutually exclusive. Chronic traumatic encephalopathy
Phosphorylation negatively regulates the abil- Diffuse neurofibrillary tangles with calcification
ity of tau to interact with microtubules and fila- Down’s syndrome
mentous tau is abnormally hyperphosphorylated Familial British dementia
[44]. However, it remains to be proved that phos- Familial Danish dementia
phorylation is the trigger for tau assembly in Gerstmann-Sträussler-Scheinker disease
human diseases. Alternatively, a change in con- Niemann-Pick disease, type C
Nodding syndrome
formation as part of assembly may lead to hyper-
Non-Guamanian motor neuron disease with
phosphorylation. Since tau is hydrophilic, it is neurofibrillary tanges
not surprising that unmodified and full-length Postencephalitic parkinsonism
protein requires cofactors, such as sulphated gly- Primary age-related tauopathy
cosaminoglycans, nucleic acids or fatty acids, to Progressive ataxia and palatal tremor
assemble into filaments [35, 48, 68, 86]. Cofactors Tangle-only dementia
other than heparin and/or post-translational mod- Familial frontotemporal dementia and parkinsonism
(some MAPT mutations, such as V337 M and R406W)
ifications may cause the assembly of tau in
3R Tauopathies
human brain [25, 26].
Pick’s disease
Besides phosphorylation, other post-­Familial frontotemporal dementia and parkinsonism
translational modifications may also play a role. (some MAPT mutations, such as G272 V and Q336R)
Early studies on tau acetylation reported that it 4R Tauopathies
can promote both phosphorylation and assembly Age-related astrogliopathy
[15, 59]. However, subsequent work suggested an Argyrophilic grain disease
inverse correlation between acetylation and phos- Corticobasal degeneration
phorylation, with acetylation inhibiting tau Guadeloupean parkinsonism
Globular glial tauopathy
assembly [11, 17]. These discrepancies may have
Hippocampal tauopathy
resulted from the use of enzymes that acetylated Huntington’s disease
different residues. Site-specific acetylation of Progressive supranuclear palsy
K280 has been shown to enhance heparin-­ SLC9a-related parkinsonism
induced tau aggregation in vitro, while reducing Familial frontotemporal dementia and parkinsonism
microtubule assembly [40]. Unlike phosphoryla- (some MAPT mutations, such as P301L and P301S, all
tion, acetylation occurs on lysine residues. known intronic mutations, and many coding region
mutations in exon 10)
In AD, CTE, tangle-only dementia and other
tauopathies, all six tau isoforms are present in
disease filaments (Table  1.2). Pick bodies are
made of only 3R tau. In PSP, CBD, AGD and Genetics of MAPT
other diseases, isoforms with 4R tau are found in
the filaments. The morphologies of tau filaments The relevance of tau inclusion formation for neu-
vary in different diseases, even when they are rodegeneration became clear in June 1998, when
mainly made of the same isoforms. dominantly inherited mutations in MAPT were
1  Ordered Assembly of Tau Protein and Neurodegeneration 7

shown to cause a form of frontotemporal demen- electron microscopy, paired helical filaments
tia that can be associated with parkinsonism (PHFs) and straight filaments (SFs) are present.
(FTDP-17T, also known as familial FTLD-tau) The brains of many individuals with missense
[43, 70, 76]. Abundant filamentous tau inclusions MAPT mutations in exons 9–13 (K257T, L266V,
are present in either nerve cells or in both nerve S305N, G272V, L315R, S320F, S320Y, P332S,
and glial cells. Aβ deposits, a defining feature of Q336H, Q336R, K369I, E372G and G389R) are
AD, are not characteristic of FTDP-17T.  This characterised by abundant Pick bodies made pre-
work established that a pathological pathway dominantly of 3R tau. As in sporadic PiD, insol-
leading from monomeric to assembled tau is suf- uble tau shows strong bands of 60 and 64 kDa.
ficient for causing neurodegeneration and However, variable amounts of 68- and 72  kDa
dementia. bands are also present. A third pattern is charac-
Sixty mutations in MAPT have been identified teristic of MAPT mutations that affect the alterna-
in FTDP-17T (Fig.  1.1b). Filaments are com- tive mRNA splicing of exon 10, resulting in the
posed of either 3R or 4R tau, or of both 3R and relative overproduction of 4R tau (intronic muta-
4R tau. MAPT mutations account for approxi- tions and exonic mutations N279K, L284L,
mately 5% of cases of FTLD and are concen- L284R, ΔN296, N296D, N296H, N296N,
trated in exons 9–12 (encoding R1-R4) and the S305L, S205N and S305S). Insoluble tau runs as
introns flanking exon 10. They can be divided two strong bands of 64 and 68 kDa, and a weaker
into those with a primary effect at the protein band of 72  kDa; following dephosphorylation,
level and those affecting the alternative splicing three bands are present that align with recombi-
of tau pre-mRNA.  There is no obvious correla- nant 4R tau (isoforms of 383, 412 and 441 amino
tion between known mutations and post-­ acids). A similar pattern of pathological tau bands
translational modifications of tau. is observed for mutations in exon 10, such as
Mutations that act at the protein level change P301L and P301S, which have their primary
or delete single amino acids, reducing the ability effects at the protein level. Assembly of 4R tau
of tau to interact with microtubules [41]. Some has also been described for mutations I260V in
mutations also promote the assembly of tau into exon 9, K317N in exon 11, E342V in exon 12
filaments [36, 63]. Mutations with a primary and N410H in exon 13, showing that it is possible
effect at the RNA level are intronic or exonic and to alter 3R and 4R tau mRNAs through mutations
increase the alternative mRNA splicing of exon located outside exon 10.
10. This affects the ratio of 3R to 4R isoforms, The effects of MAPT mutations can vary.
resulting in the relative overproduction of 4R tau Neighbouring mutations in exon 12 (G335S,
and its assembly into filaments [43, 76]. One G335V, Q336H, Q336R and V337M) give rise to
mutation (ΔK280) has been reported to cause the structurally distinct assemblies and exert different
relative overexpression of 3R tau and its assem- functional effects. Mutation G335S is character-
bly into filaments [82]. ized by abundant filamentous tau inclusions in
Assembled tau shows different isoform pat- nerve cells and glial cells, in the absence of Pick
terns and filament morphologies, depending on bodies [77]. Mutations Q336H and Q336R give
the mutations in MAPT [29]. Mutations V337 M rise to what is essentially a familial form of PiD,
in exon 12 and R406W in exon 13 give rise to with abundant Pick bodies in nerve cells [69, 80],
insoluble tau bands of 60, 64 and 68 kDa and a whereas mutation V337M produces a neuronal
weaker band of 72 kDa. Following dephosphory- filamentous tau pathology like that of AD [70,
lation, six bands are present that align with 75]. These findings on MAPT mutations in three
recombinant tau, like what is seen in AD [33]. By adjacent codons reinforce the view that the mech-
8 M. Goedert and M. G. Spillantini

anisms underlying the formation of neurofibril- Heterozygous microdeletions of chromosome

lary lesions and Pick bodies are closely related. 17q21.31 give rise to a multisystem disorder with
Recombinant tau with the G335S, G335V [64], or intellectual disability, hypotonia and distinct
V337 M mutation shows a greatly reduced ability facial features (17q21.31 microdeletion syn-
to promote microtubule assembly. By contrast, drome or Koolen-de Vries syndrome) [51, 73,
mutations Q336H and Q336R increase the ability 74]. In addition to MAPT, three protein-coding
of tau to promote microtubule assembly. genes (CRHR1, SPPL2C and KANSL1) and two
Mutations G335V and V337M fail to increase putative genes (MGC57346 and CRHR-ITI) are
heparin-induced assembly into filaments signifi- found in this region. Deletions arise on the H2
cantly, whereas mutations Q336H and Q336R haplotype through low-copy, repeat-mediated,
increase the assembly of 3R, but not 4R, tau. nonallelic homologous recombination.
The architecture of MAPT on chromosome The 17q21.31 microdeletion syndrome is
17q21.31 is characterized by two haplotypes as caused by haploinsufficiency of KANSL1, which
the result of a 900  kb inversion (H1) or non-­ encodes a chromatin modifier that influences
inversion (H2) polymorphism [79]. Inheritance gene expression through the acetylation of lysine
of the H1 haplotype of MAPT is a risk factor for 16 of histone H4 [52, 95]. A 50% reduction in tau
PSP, CBD, Parkinson’s disease (PD) and amyo- levels does therefore not appear to have a detri-
trophic lateral sclerosis (ALS), but not for PiD [3, mental effect on development of the human brain.
16, 22, 61, 66, 90]. The association with PD and
ALS is particularly surprising, since they are not
characterized by tau inclusions. Neurodegeneration
Based on genome-wide association studies for and Propagation
PSP and CBD, it has been shown that association
with an allele at the MOBP/SLC25A38 locus Disease-causing mutations in MAPT have made
results in elevated levels of appoptosin, a protein it possible to produce transgenic rodent lines that
that activates caspase-3, which can cleave tau form abundant tau filaments and exhibit neurode-
[93]. This may cause aggregation of 4R tau. generation (Figs. 1.2 and 1.3) [37]. Tau assembly
Additional loci were unique to PSP or correlates with neurodegeneration. Reducing
CBD. Association of the H1 haplotype with PSP assembly and increasing aggregate degradation
had a higher odds ratio than that between apoli- are therefore therapeutic objectives. Since assem-
poprotein E epsilon 4 (APOEε4) and AD [42]. bly is concentration-dependent, decreasing the
APOEε4 is the major risk factor allele for late-­ level of soluble, monomeric tau is likely to result
onset AD [19]. H1 expresses more exon in reduced assembly [21, 55]. However, the
10-­containing mRNA than H2, especially in sub- molecular species of assembled tau that are
cortical regions [9, 54]. Moreover, H2 is associ- responsible for neurodegeneration remain to be
ated with increased expression of exon 3 of identified [58]. At a cellular level, it has been
MAPT in grey matter, suggesting that inclusion reported that the removal of senescent brain cells
of exon 3 may protect against PSP, CBD, PD and leads to a reduction in both tau assembly and
ALS [10]. In experimental studies, exon neurodegeneration in transgenic mice [8]. In
3-­
containing tau isoforms (those with both tauopathies, as in most neurodegenerative dis-
N-terminal inserts) have been found to assemble eases, inclusion formation manifests many years
less than those lacking this exon [94]. Even before clinical symptoms. In future, it will there-
though all six tau isoforms give rise to PHFs and fore be important to identify individuals at risk of
SFs, known mutations in MAPT do not cause disease. Early diagnostics,in particular imaging
AD.  They give rise to FTLD-tau. Tau with an of tau inclusions, is therefore likely to play an
A152T substitution has been reported to be a risk important role [38].
factor for AD [18], as well as for PSP, CBD and Transgenic mouse lines were also essential for
unusual tauopathies [18, 49, 53, but see also 67]. the identification of the prion-like properties of
1  Ordered Assembly of Tau Protein and Neurodegeneration 9

Fig. 1.2 Electron microscopy and immunoelectron cytoplasmic regions from (a, c, e). The electron micro-
microscopy of nerve cells in brain and spinal cord from graphs in (c, d) show immunogold labelling of filaments
mice transgenic for human mutant P301S tau using the phosphorylation-­ dependent anti-tau antibody
(a, b), Cerebral cortex; (c, d), Brainstem; (e, f), Spinal AT8. Scale bars in (C) and (E) must be in micrometers:
cord. (b, d, f), Higher magnifications of parts of the C, 1.5 m; E, 5.5 m (for a, e); F, 300 nm (for b, d, f)

assembled tau in vivo (Fig. 1.4) [62]. Aggregation aggregated 2N4R tau and, to a lesser extent, fol-
of hyperphosphorylated tau was induced follow- lowing intracerebral injection into wild-type
ing intracerebral injection of tau seeds from mice mice [12]. Tauopathy then spread to connected
transgenic for human mutant 0N4R P301S tau brain regions, indicative of seed endocytosis,
into transgenic mice expressing wild-type non-­ seeded aggregation, intracellular transport and
10 M. Goedert and M. G. Spillantini

release of tau seeds. This work was ­complemented

A by studies, which showed that short tau filaments
have the greatest seeding activity (Figs. 1.5 and
1.6) [45]. Seeded aggregation of tau was depen-
dent on the ability of expressed, monomeric tau
to aggregate [23].
Distinct conformers of assembled tau appear
to exist, reminiscent of prion strains. They may
explain the variety of human tauopathies.
Inclusions formed and spread of pathology
B occurred after intracerebral injection of brain
homogenates from cases of AD, tangle-only
dementia, PSP, CBD and AGD into a mouse line
transgenic for wild-type 4R tau [13]. PiD, the fila-
mentous inclusions of which consist of only 3R
tau, was an exception. Inclusions formed at the
injection sites, but spreading was not observed.
However, PiD is only rarely a pure 3R tauopathy,
since it can be associated with AD-type tau pathol-
ogy. We therefore cannot exclude that the activity
C in the PiD homogenate, which induced aggrega-
12
tion at the injection site, may have been due to the
presence of a small amount of aggregated 4R tau.
10
Sequence requirements for seeded tau aggre-
Number of neurons/mm2

8
gation in vivo remain to be defined. Tau assem-
blies reminiscent of those in the corresponding
6 human diseases were observed following the
injection of brain homogenates from patients
4 with PSP, CBD and AGD, which are 4R tauopa-
thies [13]. Although these findings are consistent
2 with the existence of distinct tau aggregate con-
formers, the definition of such conformers must
0 be structural.
Control P301S

Fig. 1.3  Nerve cell loss in spinal cord of mice transgenic

for human mutant P301S tau  igh-Resolution Structures of Tau
H
(a, b), Haematoxylin and eosin-stained sections of the
ventral grey matter of the lumbar spinal cord (L2-L3) Filaments from Alzheimer’s Disease
from 6-month-old non-transgenic (a) and transgenic (b)
mice. Swollen, abnormal material-containing motor neu- By cryo-EM, high-resolution structures of Tau
rons are indicated by arrows. Arrowheads point to atro- filaments were obtained from the frontal cortex
phic motor neurons, with dashed arrows pointing to
pyknotic cells that are surrounded by glial cells, sugges- of four individuals with AD, three sporadic and
tive of neuronophagia. (c) Graph showing the density of one inherited (Fig. 1.7) [25, 27]. The cores of tau
motor neurons (expressed as number of neurons per mil- filaments are made of two protofilaments consist-
limetre square) in the anterior horn of the lumbar spinal ing of residues G273/304-E380, which adopt a
cord of 6-month-­old non-transgenic (Control) and trans-
genic (P301S) mice. Nerve cell numbers were determined combined cross-β/β-helix structure (Fig.  1.8).
in five consecutive sections from each animal, with the Murine and human tau are identical in sequence
density of motor neurons from each section being repre- in this region. The N-terminal part of the cross-β
sented by a circle. The results are expressed as the means structure includes hexapeptide 306VQIVYK311,
± S.E.M. (n = 3). ∗p < 0.0001. Scale bar in (B) must be in
micrometers: 60 m (for a, b)
1  Ordered Assembly of Tau Protein and Neurodegeneration 11

Fig. 1.4  Induction of filamentous tau pathology in mice Gallyas-Braak silver. Non-injected (left), 15 months after
transgenic for 2N4R wild-type human tau (line ALZ17) injection with brain extract from non-transgenic control
following injection with brain extract from symptomatic mice (middle) and 15  months after injection with brain
mice transgenic for 0N4R P301S tau extract from 6-month-old mice transgenic for human
Staining of the hippocampal CA3 region of 18-month-old P301S tau (right). The sections were counterstained with
ALZ17 mice with anti-tau antibodies AT8 and AT100 and haematoxylin. The Scale bar must be given in micrometers

which is essential for the oligomerisation of PHFs and SFs are made of identical protofila-
recombinant tau and its assembly into filaments ments, but differ in inter-protofilament packing,
[72, 83]. It packs against 373THKLTF378, in agree- showing that they are ultrastructural polymorphs.
ment with the predicted heterozipper interaction PHF protofilaments are arranged base-to-base
between 306VQIVYK311 and 375KLTFR379 [60]. and SF protofilaments back-to-base. In PHFs,
Constructs K18 and K19 end at E372 [39]; they protofilaments are stabilised by backbone hydro-
can therefore not give rise to the human brain tau gen bonds between their 332PGGGQ336 sequences.
folds determined thus far. Moreover, the side-chains of K331 from one pro-
Each protofilament is made of eight β-strands, tofilament project towards the side-chains of
five of which give rise to two regions of antiparal- Q336 and E338 of the other protofilament, sug-
lel β-sheets, with the other three forming a β-helix gesting additional interactions that stabilise the
(Fig. 1.9). The C-terminal residues of R1 and R2 protofilament interface. Furthermore, in the pro-
form part of the first β-strand. R3 contributes tofilament interface of the PHF, extra densities
three and R4 four β-strands, with the final between the side-chains of K331 of one protofila-
β-strand being formed by 12 amino acids after R4 ment and the backbone of V337 of the other have
(residues K369-E380). Strands β1 and β2 pack been observed. They may correspond to a solvent
against β8, β3 packs against β7, with β4, β5 and molecule or a post-translational modification of
β6 giving rise to the C-shaped β-helix. K331, such as mono-methylation [25].
12 M. Goedert and M. G. Spillantini

Positive control (normalised)

4.4 weeks 24.4 weeks

Seeded with 10%

Seeded with 20%
Seeded with 30%
Seeded with 40%
Seeded with 50%

Seeded with 10%

Seeded with 20%
Seeded with 30%
Seeded with 40%
Seeded with 50%
Seeded with Top

Seeded with Top

Positive control
Unseeded

68kDa (HMW)
59kDa (LMW)

68kDa (HMW)

Insoluble cell preparation post-seeding

Fig. 1.5  Seeding of tau assembly with sucrose gradient HMW) and non-filamentous tau at 59 kDa (low molecular
fractions from the brains of mice transgenic for human weight, LMW). As a positive control, sarkosyl-­insoluble
mutant 0N4R P301S tau in a cell-based assay tau extracted from unfractionated brains of symptomatic
The mice were aged 4.4 weeks (no symptoms, no tau fila- transgenic P301S tau mice was used for seeding; the nor-
ments) or 24.4 weeks (symptoms, abundant tau filaments). malized positive control consisted in seeding with
Sucrose gradient fractions were used to seed assembly of sarkosyl-­insoluble tau that had been normalised for total
tau in HEK cells expressing 1N4R P301S tau. The pellet tau levels relative to those of the sucrose gradient frac-
from a 1000,000 g spin of the seeded cells was analysed tions. Seeding correlated with the presence of the 64 kDa
by immunoblotting for total tau (antibody DA9) and tau band in 24.4-week-old mice (20–50% sucrose gradient
phosphorylated at S202/T205 (antibody AT8). fractions). No seeding was observed upon addition of
Filamentous tau runs at 68 kDa (high molecular weight, sucrose gradient fractions from 4.4-week-old mice

In SFs, the protofilaments pack asymmetri-

cally. Their backbones are nearest each other  igh-Resolution Structures of Tau
H
between residues 321KCGS324 of the first and Filaments from Pick’s Disease
313
VDLSK317 of the second protofilament. The
inter-protofilament packing appears to be stabi- By cryo-EM, high-resolution structures of Tau
lised through the region of additional density that filaments were determined from the frontotempo-
interacts with the side-chains of K317, T319 and ral cortex of an individual with PiD [24]. Two
K321 of both protofilaments. This density may types of filament could be distinguished: a major-
correspond to residues 7EFE9, which constitute ity of narrow Pick filaments (NPFs) and a minor-
the N-terminal region of the discontinuous epit- ity of wide Pick filaments (WPFs) (Fig.  1.10).
ope of conformational anti-tau antibodies ALZ-­ The core of NPFs is made of a single protofila-
50 and MC-1 (the C-terminal epitope is ment that consists of residues K254-F378 of 3R
313
VDLSKVTSKC322) [47]. A similar density tau, which adopt an elongated cross-β structure.
also interacts with K317, T319 and K321  in Murine and human tau are identical in sequence
PHFs, where it does not contribute to the proto- in this region, with the exception of residue 257
filament interface. (K in human, R in mouse tau). WPFs are formed
1  Ordered Assembly of Tau Protein and Neurodegeneration 13

Fig. 1.6  Tau species in the 40% sucrose gradient frac- sucrose gradient fraction (10% fraction) were indistin-
tions from the brains of symptomatic mice transgenic for guishable from those injected with brainstem lysate from
human mutant P301S tau (24.4  weeks) mediate seeding non-transgenic mice. Injection of brainstem lysate [posi-
and spreading in brain tive control (normalised)] that had been normalised for
(a) Unilateral injection of brainstem lysate (positive con- total tau did not show robust tau aggregate propagation.
trol) from symptomatic transgenic mice into the hippo- (b) In the ipsilateral hippocampus, PG-5 immunoreac-
campus of asymptomatic transgenic mice led to the tivity was highest for the positive control, followed by
accumulation and spread of pathology (detected by anti- the 40% sucrose fraction. No significant increase in sig-
body PG-5, which recognises tau phosphorylated at nal was observed for the normalised positive control or
S409) in the injected (∨) and contralateral (∧) hippocam- the 10% fraction (relative to the negative control). (c) In
pus. Mice injected with brainstem lysate of non-trans- the contralateral hippocampus, a similar, but overall
genic mice (negative control) had minimal pathology. milder, pattern of tau pathology was observed compared
Injection of the 40% sucrose gradient fraction also to the injecteded side. The Scale bar must be given in
showed substantial accumulation and spread of pathol- micrometers. ∗p  <  0.05, ∗∗p  <  0.01, ♯p  <  0.01 for all
ogy (40% fraction), whereas mice injected with the 10% groupwise comparisons
14 M. Goedert and M. G. Spillantini

Fig. 1.7  Cryo-EM structures of paired helical filaments V717F in the amyloid precursor protein gene). All cases
(PHFs) and straight filaments (SFs) from the frontal cor- had a majority of PHFs and a minority of SFs. Yellow
tex of four AD cases. All structures show identical pairs arrows point to the extra densities, which were present in
of C-shaped protofilaments and the same inter-protofila- PHFs and SFs from all four cases, bordering the solvent-
ment packing in PHFs and SFs. Cases 1, 2 and 3 had exposed side-chains of R349 and K375, and of H362 and
sporadic AD, whereas case 4 had inherited AD (mutation K369

Fig. 1.8  Cryo-EM densities and atomic models of paired (PHFs) and green (SFs). Unsharpened 4.5 Å low-pass fil-
helical filaments (PHFs) and straight filaments (SFs) from tered densities are shown in grey. The models comprise
the frontal cortex of one AD case G273-E380 of 3R tau and G304-E380 of 4R tau. The pro-
PHFs and SFs were resolved to 3.2 Å and 3.3 Å, respec- tofilaments of PHFs and SFs are identical, but differ in
tively. Sharpened, high-resolution maps are shown in blue inter-protofilament packing (ultrastructural polymorphs)

by the association of two NPF protofilaments at each. These stacks pack together in a hairpin-like
their distal tips, where they form tight contacts fashion: β1 against β8, β2 against β7, β3 against
through van der Waals interactions. Each proto- β6 and β4 against β5. The final strand, β9, is
filament comprises nine β-strands, which are formed from the ten amino acids after R4 and
arranged into four cross-β packing stacks and are packs against the opposite side of β8.
connected by turns and arcs (Fig. 1.11). R1 pro- Three regions of less well-resolved density
vides two β-strands and R3 and R4 three β-strands bordering the solvent-exposed faces of β4, β5 and
1  Ordered Assembly of Tau Protein and Neurodegeneration 15

Fig. 1.9  Schematic view of the tau protofilament core of AD. The observed eight β-strand regions (β1-β8) are shown
as arrows. Each filament consists of two protofilaments

β9 are apparent in both NPFs and WPFs. They dementia, CTE, PiD, PSP, CBD and AGD. They
may represent less ordered, heterogeneous and/or may be different clinical diseases, because of the
transiently occupied structures. The density bor- existence of distinct conformers of assembled
dering β4 is similarly located, but more extended, tau. Assembly leads to propagation of pathology
than that found to interact with the side-chains of and neurodegeneration, which are characteristic
K317, T319 and K321 in AD filaments. of all tauopathies. It remains to be seen if the
It was not previously known why only 3R tau same or different molecular species of assembled
is present in Pick body filaments and why S262 is tau account for propagation and neurodegenera-
not phosphorylated. Our results suggest that tion. Small filaments are the major species of
despite sequence homology, the structure formed assembled tau responsible for the propagation of
by K254-K274 of R1 is inaccessible to the cor- pathology.
responding residues from R2 (S275-S305). Cryo-EM of filaments from human brain has
Moreover, because of steric constraints, the fila- established that distinct conformers of aggre-
ment structure precludes phosphorylation of gated tau are characteristic of AD and PiD [91].
S262. Even though both types of filament share resi-
dues G273-F378 of 3R tau, their structures are
very different (Fig. 1.12). Whereas PHFs and SFs
Conclusion of AD are made of two identical C-shaped proto-
filaments that each comprises eight β-strands and
Biochemistry, molecular biology and human a combined β-sheet/β-helix structure, NPFs of
genetics have shown that the ordered assembly of PiD are made of a single elongated protofilament
tau is at the heart of a large number of neurode- comprising nine β-strands and stacks of β-sheets.
generative diseases, including AD, tangle-only WPFs consist of two NPFs joined through their
16 M. Goedert and M. G. Spillantini

Fig. 1.10 (a) Sharpened, high-resolution cryo-EM of PiD.  Unsharpened cryo-EM densities of a narrow
map of narrow Pick filaments with the atomic model of (b) and wide (c) Pick filament. Wide Pick filaments
the Pick fold overlaid (3.2  Å resolution). Narrow comprise two narrow filaments that are joined at their
(93%) and wide (7%) Pick filaments are characteristic distal tips

Fig. 1.11  Schematic view of the tau protofilament core of PiD. The observed nine β-strands (β1-β9) are shown as
arrows. Narrow Pick filaments are made of one and wide Pick filaments of two protofilaments
1  Ordered Assembly of Tau Protein and Neurodegeneration 17

Fig. 1.12  Comparison of Alzheimer and Pick tau fila- consist of a single J-shaped protofilament. Wide Pick fila-
ment folds depicted as single rungs ments consist of two narrow filaments packed against
Paired helical and straight tau filaments of AD consist of each other. C322 (yellow) and D348 (red) are highlighted.
two identical C-shaped protofilaments that differ in inter- C322 is on the inside in the Alzheimer fold and the outside
protofilament packing (ultrastructural polymorphs). More in the Pick fold, whereas D348 is on the outside in the
than 90% of tau filaments of PiD are narrow filaments that Alzheimer fold and the inside in the Pick fold

4. Berriman J, Serpell LC, Oberg KA, Fink AL, Goedert

distal tips. Cryo-EM studies of tau filaments from M, Crowther RA.  Tau filaments from human brain
additional sporadic and inherited cases of AD, as and from in  vitro assembly of recombinant protein
well as negative-stain immunoelectron micros- show cross-beta structure. Proc Natl Acad Sci U S A.
2003;100:9034–8.
copy of tau filaments from multiple cases of AD 5. Binder LI, Frankfurter A, Rebhun LI.  The distribu-
and PiD, indicated that the cores of tau filaments tion of tau in the mammalian central nervous system.
from each disease case contain the same J Cell Biol. 1985;101:1371–8.
sequences. It therefore seems that the structures 6. Black MM, Slaughter T, Moshiach S, Obrocka M,
Fischer I.  Tau is enriched on dynamic microtubules
of tau filaments are distinct between diseases, but in the distal region of growing axons. J  Neurosci.
identical in different individuals with AD or 1996;16:3601–19.
PiD.  It appears likely that additional folds of 7. Brion JP, Passareiro H, Nunez J, Flament-Durand
assembled tau remain to be discovered in other J.  Mise en évidence immunologique de la protéine
tau au niveau des lésions de dégénérescence neu-
tauopathies. rofibrillaire de la maladie d’Alzheimer. Arch Biol.
1985;95:229–35.
8. Bussian TJ, Aziz A, Meyer CF, Swenson BL, Van
References Deursen JM, Baker DJ. Clearance of senescent glial
cells prevents tau-dependent pathology and cognitive
decline. Nature. 2018;562:578–82.
1. Al-Bassam J, Ozer RS, Safer D, Halpain DS, Milligan 9. Caffrey TM, Joachim C, Paracchini S, Esiri MM,
RA.  MAP2 and tau bind longitudinally along the Wade-Martins R.  Haplotype-specific expression of
outer ridges of microtubule protofilaments. J  Cell exon 10 at the human MAPT locus. Hum Mol Genet.
Biol. 2002;157:1187–96. 2006;15:3529–37.
2. Apicco DJ, Ash PEA, Maziuk B, LeBlang C, Medalla 10. Caffrey TM, Joachim C, Wade-Martins R. Haplotype-­
M, Al Abdullatif A, Ferragud A, Botelho E, Balance specific expression of the N-terminal exons 2 and
HI, Dhawan U, et al. Reducing the RNA binding pro- 3 at the human MAPT locus. Neurobiol Aging.
tein TIA1 protects against tau-mediated neurodegen- 2008;29:1923–9.
eration in vivo. Nat Neurosci. 2018;21:72–80. 11. Carlomagno Y, Chung D, Yue M, Castanedes-Casey
3. Baker M, Litvan I, Houlden H, Adamson J, Dickson M, Madden BJ, Dunmore J, Tong J, DeTure M,
D, Perez-Tur J, Hardy J, Lynch T, Bigio E, Hutton Dickson DW, Petrucelli L, et  al. An acetylation-­
M.  Association of an extended haplotype in the tau phosphorylation switch that regulates tau aggre-
gene with progressive supranuclear palsy. Hum Mol gation propensity and function. J  Biol Chem.
Genet. 1999;8:711–5. 2017;292:15277–86.
18 M. Goedert and M. G. Spillantini

12. Clavaguera F, Bolmont T, Crowther RA, Abramowski 24. Falcon B, Zhang W, Murzin AG, Murshudov G,

D, Frank S, Probst A, Fraser G, Stalder AK, Beibel Garringer HJ, Vidal R, Crowther RA, Ghetti B,
M, Staufenbiel M, et al. Transmission and spreading Scheres SHW, Goedert M.  Structures of filaments
of tauopathy in transgenic mouse brain. Nat Cell Biol. from Pick’s disease reveal a novel tau protein fold.
2009;11:909–13. Nature. 2018;561:137–40.
13. Clavaguera F, Akatsu H, Fraser G, Crowther RA,
25. Falcon B, Zhang W, Schweighauser M, Murzin AG,
Frank S, Hench J, Probst A, Winkler DT, Reichwald J, Vidal R, Garringer HJ, Ghetti B, Scheres SHW,
Staufenbiel M, et al. Brain homogenates from human Goedert M. Tau filaments from multiple cases of spo-
tauopathies induce tau inclusions in mouse brain. radic and inherited Alzheimer’s disease adopt a com-
Proc Natl Acad Sci U S A. 2013;110:9535–40. mon fold. Acta Neuropathol. 2018;136:699–708.
14. Cleveland DW, Hwo SY, Kirschner MW.  Physical 26. Fichou Y, Lin Y, Rauch JN, Vigers M, Zeng Z,

and chemical properties of purified tau factor and Srivasta M, Keller TJ, Freed JH, Kosik KS, Han
the role of tau in microtubule assembly. J Mol Biol. S.  Cofactors are essential constituents of stable and
1977;116:227–47. seeding-active tau fibrils. Proc Natl Acad Sci U S A.
15. Cohen TJ, Guo JL, Hurtado DE, Kwong LK, Mills 2018;115:13234–2239.
IP, Trojanowski JQ, Lee VMY. The acetylation of tau 27. Fitzpatrick AWP, Falcon B, He S, Murzin AG,

inhibits its function and promotes pathological tau Murshudov G, Garringer HJ, Crowther RA, Ghetti
aggregation. Nat Commun. 2011;2:52. B, Goedert M, Scheres SHW.  Cryo-EM structures
16. Conrad C, Andreadis A, Trojanowski JQ, Dickson of tau filaments from Alzheimer’s disease. Nature.
DW, Kang D, Chen X, Wiederholt W, Hansen L, 2017;547:185–90.
Masliah E, Thal LJ, et  al. Genetic evidence for the 28. Gauthier-Kemper A, Suárez-Alonso M, Sündermann
involvement of tau in progressive supranuclear palsy. F, Niewidok B, Fernandez MP, Bakota L, Heinisch
Ann Neurol. 1997;41:277–81. JJ, Brandt R.  Annexins A2 and A6 interact with
17. Cook C, Carlomagno Y, Gendron TF, Dunmore J, the extreme N-terminus of tau and thereby con-
Scheffel K, Stetler C, Davis M, Dickson D, Jarpe M, tribute to tau’s axonal localization. J  Biol Chem.
DeTure M, et  al. Acetylation of the KXGS motifs 2018;293:8065–76.
in tau is a critical determinant ion modulation of 29. Ghetti B, Oblak LA, Boeve BF, Johnson KA,

tau aggregation and clearance. Hum Mol Genet. Dickerson BC, Goedert M. Frontotemporal dementia
2014;23:104–16. caused by microtubule-associated protein tau gene
18. Coppola G, Chinnathambi S, Lee JJ, Dombrowski (MAPT) mutations: a chameleon for neuropathol-
BA, Baker MC, Soto-Ortolaza AI, Lee SE, Klein E, ogy and neuroimaging. Neuropathol Appl Neurobiol.
Huang AY, Sears R, et al. Evidence for a role of the 2015;41:24–46.
rare p.A152T variant in MAPT in increasing the risk 30. Goedert M, Wischik CM, Crowther RA, Walker JE,
for FTD-spectrum and Alzheimer’s diseases. Hum Klug A. Cloning and sequencing of the cDNA encod-
Mol Genet. 2012;21:3500–12. ing a core protein of the paired helical filament of
19. Corder EH, Saunders AM, Strittmatter WJ, Schmechel Alzheimer disease: identification as the microtubule-­
DE, Gaskell PC, Small GW, Roses AD, Haines JL, associated protein tau. Proc Natl Acad Sci U S A.
Pericak-Vance MA.  Gene dose of apolipoprotein E 1988;85:4051–5.
type 4 allele and the risk of Alzheimer’s disease in 31. Goedert M, Spillantini MG, Jakes R, Rutherford

late-onset families. Science. 1993;261:921–3. D, Crowther RA.  Multiple isoforms of human
20. Couchie D, Mavilia C, Georgieff IS, Liem RK,
microtubule-­ associated protein tau: sequences and
Shelanski ML, Nunez J.  Primary structure of high localization in neurofibrillary tangles of Alzheimer’s
molecular weight tau present in the peripheral nervous disease. Neuron. 1989;3:519–26.
system. Proc Natl Acad Sci U S A. 1992;89:4378–81. 32. Goedert M, Jakes R. Expression of separate isoforms
21. De Vos SL, Miller RL, Schoch KM, Holmes BB, of human tau protein: Correlation with the tau pattern
Kebodeaux CS, Wegener AJ, Chen G, et al. Tau reduc- in brain and effects on tubulin polymerization. EMBO
tion prevents neuronal loss and reverses pathological J. 1990;9:4225–30.
tau deposition and seeding in mice with tauopathy. Sci 33. Goedert M, Spillantini MG, Cairns NJ, Crowther

Transl Med. 2017;9:eaag0481. RA.  Tau proteins of Alzheimer paired helical fila-
22. Di Maria E, Tabaton M, Vigo T, Abbruzzese G,
ments: abnormal phosphorylation of all six brain iso-
Bellone E, Donati C, Frasson E, Marchese R, Montag forms. Neuron. 1992;8:159–68.
P, Munoz DG, et al. Corticobasal degeneration shares 34. Goedert M, Spillantini MG, Crowther RA. Cloning of
a common genetic background with progressive a big tau microtubule-associated protein characteristic
supranuclear palsy. Ann Neurol. 2000;47:374–7. of the peripheral nervous system. Proc Natl Acad Sci
23. Falcon B, Cavallini A, Angers R, Glover S, Murray U S A. 1992;89:1183–9.
TK, Barnham L, Jackson S, O’Neill MJ, Isaacs AM, 35. Goedert M, Jakes R, Spillantini MG, Hasegawa M,
Hutton ML, et al. Conformation determines the seed- Smith MJ, Crowther RA.  Assembly of microtubule-­
ing potencies of native and recombinant tau aggre- associated protein tau into Alzheimer-like filaments
gates. J Biol Chem. 2015;290:1049–65. induced by sulphated glycosaminoglycans. Nature.
1996;383:550–3.
1  Ordered Assembly of Tau Protein and Neurodegeneration 19

36. Goedert M, Jakes R, Crowther RA.  Effects of fron- cal clinical and neuropathological features. Neurobiol
totemporal dementia FTDP-17 mutations on heparin-­ Aging. 2012;33(2231):e7–e14.
induced assembly of tau filaments. FEBS Lett. 50. Kellogg EH, Hejab NMA, Poepsel S, Downing

1999;450:306–11. KH, DiMaio F, Nogales E.  Near-atomic model
37. Goedert M, Eisenberg DS, Crowther RA. Propagation of microtubule-­tau interactions. Science.
of tau aggregates and neurodegeneration. Annu Rev 2018;360:1242–6.
Neurosci. 2017;40:189–210. 51. Koolen DA, Viussers LE, Pfundt R, De Leeuw N,
38. Goedert M, Yamaguchi Y, Mishra SK, Higuchi M, Knight SJ, Regan R, Kooy RF, Reyniers E, Romano C,
Sahara N.  Tau filaments and the development of Fichera M, et al. A new chromosome 17q21.31 micro-
positron emission tomography tracers. Front Neurol. deletion syndrome associated with a common inver-
2018;9:70. sion polymorphism. Nat Genet. 2006;38:999–1001.
39. Gustke N, Trinczek B, Biernat J, Mandelkow EM, 52. Koolen DA, Kramer JM, Neveling K, Nilldsen WM,
Mandelkow E. Domains of tau protein and interactions Moore-Barton HL, Elmslie EV, Toutain A, Amiel J,
with microtubules. Biochemistry. 1994;33:9511–22. Malan V, Tsai AC, et al. Mutations in the chromatin
40. Haj-Yahya M, Lashuel HA.  Protein semisynthe-
modifier gene KANSL1 cause the 17q21.31 microde-
sis provides access to tau disease-associated post-­ letion syndrome. Nat Genet. 2012;44:639–41.
translational modifications (PTMs) and paves the way 53. Kovacs GG, Wöhrer A, Ströbel T, Botond G, Attems J,
to deciphering the tau PTM code in health and dis- Budka H. Unclassifiable tauopathy associated with an
eased states. J Am Chem Soc. 2018;140:6611–21. A152T variation in MAPT exon 7. Clin Neuropathol.
41. Hasegawa M, Smith MJ, Goedert M. Tau proteins with 2011;30:3–10.
FTDP-17 mutations have a reduced ability to promote 54. Kwok JBJ, Teber ET, Loy C, Hallupp M, Nicholson
microtubule assembly. FEBS Lett. 1998;437:207–10. G, Mellick GD, Buchanan DD, Silburn PA, Schofield
42. Höglinger GU, Melhem NM, Dickson DW, Sleiman PR.  Tau haplotypes regulate transcription and are
PMA, Wang LS, Klei L, Rademakers R, de Silva R, associated with Parkinson’s disease. Ann Neurol.
Litvan I, Riley DE, et  al. Identification of common 2004;55:329–34.
variants influencing risk of the tauopathy progressive 55. Lasagna-Reeves CA, de Haro M, Hao S, Park J,

supranuclear palsy. Nat Genet. 2011;43:699–705. Rousseaux MWC, Al-Ramahi I, Jafar-Nejad P,
43. Hutton M, Lendon CL, Rizzu P, Baker M, Froelich Vilanova-Velez L, See L, De Maio A, et al. Reduction
S, Houlden H, Pickering-Brown S, Chakraverty S, of Nuak1 decreases tau and reverses phenotypes in a
Isaacs A, Grover A, et  al. Association of missense tauopathy mouse model. Neuron. 2016;92:407–18.
and 5′-splice site mutations in tau with the inherited 56. Lee G, Newman ST, Gard DL, Band H, Panchamoorthy
dementia FTDP-17. Nature. 1998;393:702–5. G. Tau interacts with src-family non-receptor tyrosine
44. Iqbal K, Liu F, Gong CX.  Tau and neurodegen-
kinases. J Cell Sci. 1998;111:3167–77.
erative disease: the story so far. Nat Rev Neurol. 57. Li X, Kumar Y, Zempel H, Mandelkow EM, Biernat
2016;12:15–27. J, Mandelkow E.  Novel diffusion barrier for axonal
45. Jackson SJ, Kerridge C, Cooper J, Cavallini A,
retention of tau in neurons and its failure in neurode-
Falcon B, Cella CV, Landi A, Szekeres PG, Murray generation. EMBO J. 2011;30:4825–37.
TK, Ahmed Z, et  al. Short fibrils constitute the 58. Maeda S, Sahara N, Saito Y, Murayama S, Ikai A,
major species of seed-competent tau in the brains of Takashima A. Increased levels of granular tau oligo-
mice transgenic for human P301S tau. J  Neurosci. mers: an early sign of brain aging and Alzheimer’s
2016;36:762–72. disease. Neurosci Res. 2006;54:197–201.
46. Janning D, Igaev M, Sündermann F, Brühmann J, 59. Min SW, Cho SH, Zhou Y, Schroeder S, Haroutunian
Beutel O, Heinisch JJ, Bakota L, Piehler J, Junge W, V, Seeley WW, Huang EJ, Shen Y, Masliah E,
Brandt R. Single-molecule tracking of tau reveals fast Mukherjee C, et  al. Acetylation of tau inhibits its
kiss-and-hop interaction with microtubules in living degradation and contributes to tauopathy. Neuron.
neurons. Mol Biol Cell. 2014;25:3541–51. 2010;67:953–66.
47. Jicha GA, Bowser R, Kazam IG, Davies P. Alz-50 and 60. Moore CL, Huang MH, Robbenolt SA, Voss KR,

MC-1, a new monoclonal antibody raised to paired Combs B, Gamblin TC, Goux WJ. Secondary nucle-
helical filaments, recognize conformational epitopes ating sequences affect kinetics and thermodynamics
on recombinant tau. J Neurosci Res. 1997;48:128–32. of tau aggregation. Biochemistry. 2011;50:10876–86.
48. Kampers T, Friedhoff P, Biernat J, Mandelkow
61. Morris HR, Baker M, Yasojima K, Houlden H, Khan
EM, Mandelkow E.  RNA stimulates aggrega- MN, Wood NW, Hardy J, Grossman M, Trojanowski
tion of microtubule-­ associated protein tau into JQ, Revesz T, et  al. Analysis of tau haplotypes in
Alzheimer-like paired helical filaments. FEBS Lett. Pick’s disease. Neurology. 2002;59:443–5.
1996;399:344–9. 62. Mudher A, Colin M, Dujardin S, Medina M,

49. Kara E, Ling H, Pittman AM, Shaw K, de Silva
Dewachter I, Alavi Naini SM, Mandelkow EM,
R, Simone R, Holton JL, Warren JD, Rohrer JD, Mandelkow E, Buée L, Goedert M, et al. What is the
Xiromerisiou G, et al. The MAPT p.A152T variant is evidence that tau pathology spreads through prion-like
a risk factor associated with tauopathies with atypi- propagation? Acta Neuropathol Commun. 2017;5:99.
20 M. Goedert and M. G. Spillantini

63. Nacharaju P, Lewis J, Easson C, Yen S, Hackett J, with developmental delay and learning disability. Nat
Hutton M, Yen SH.  Accelerated filament formation Genet. 2006;38:1032–7.
from tau protein with specific FTDP-17 missense 75.
Spillantini MG, Crowther RA, Goedert
mutations. FEBS Lett. 1999;447:195–9. M.  Comparison of the neurofibrillary pathology in
64. Neumann M, Diekmann S, Bertsch U, Vanmassenhove Alzheimer’s disease and familial presenile dementia
B, Bogerts B, Kretzschmar HA. Novel G335V muta- with tangles. Acta Neuropathol. 1996;92:42–8.
tion in the tau gene associated with early onset 76. Spillantini MG, Murrell JR, Goedert M, Farlow MR,
familial frontotemporal dementia. Neurogenetics. Klug A, Ghetti B. Mutation in the tau gene in familial
2005;6:91–5. multiple system tauopathy with presenile dementia.
65. Niewidok B, Igaev M, Sündermann F, Janning DF, Proc Natl Acad Sci U S A. 1998;95:7737–41.
Bakota L, Brandt R. Presence of a carboxy-terminal 77. Spina S, Murrell JR, Yoshida H, Ghetti B, Bermingham
pseudorepeat and disease-like pseudohyperphos- N, Sweeney B, Dlouhy SR, Crowther RA, Goedert M,
phorylation critically influence tau’s interaction with Keohane C. The novel Tau mutation G335S: clinical,
microtubules in axon-like processes. Mol Biol Cell. neuropathological and molecular characterization.
2016;27:3537–49. Acta Neuropathol. 2007;113:461–70.
66. Pastor P, Ezquerra M, Munoz E, Marti MJ, Blersa R, 78. Stefanoska K, Volkerling A, Bertz J, Poljak A, Ke YD,
Tolosa E, Olivas R.  Significant association between Ittner LM.  An N-terminal motif unique to primate
the tau gene A0/A0 genotype and Parkinson’s disease. tau enables differential protein-protein interactions.
Ann Neurol. 2000;47:242–5. J Biol Chem. 2018;293:3710–9.
67. Pastor P, Moreno F, Clarimón J, Ruiz A, Combarros 79.
Stefansson H, Helgason A, Thorleifsson G,
O, Calero M, López de Munain A, Bullido MJ, de Steinthorsdottir V, Masson G, Barnard J, Baker A,
Pancorbo MM, Carro E, et  al. MAPT H1 haplotype Jonasdottir A, Ingason A, Gudnadottir VG, et  al. A
is associated with late-onset Alzheimer’s disease risk common inversion under selection in Europeans. Nat
in APOEɛ4 noncarriers: Results from the demen- Genet. 2005;37:129–37.
tia genetics Spanish consortium. J  Alzheimers Dis. 80. Tacik P, DeTure M, Hinkle KM, Lin WL, Sanchez-­
2016;49:343–52. Contreras M, Carlomagno Y, Pedraza O, Rademakers
68. Pérez M, Valpuesta JM, Medina M, Montejo de
R, Ross OA, Wszolek ZK, et al. A novel tau mutation
Garcini E, Avila J. Polymerization of tau into filaments in exon 12, p.Q336H, causes hereditary Pick disease.
in the presence of heparin: the minimal sequence J Neuropathol Exp Neurol. 2015;74:1042–52.
required for tau-tau interaction. J  Neurochem. 81. Tuerde D, Kimura T, Miyasaka T, Furusawa K,

1996;67:1183–90. Shimozawa A, Hasegawa M, Ando K, Hisanaga
69. Pickering-Brown SM, Baker M, Nonaka T, Ikeda K, SI. Isoform-independent and –dependent phosphory-
Sharma S, Mackenzie J, Simpson SA, Moore JW, lation of microtubule-associated protein tau in mouse
Snowden JS, de Silva R, et al. Frontotemporal demen- brain during postnatal development. J  Biol Chem.
tia with Pick-type histology associated with Q336R 2018;293:1781–93.
mutation in the tau gene. Brain. 2004;127:1415–26. 82. Van Swieten JC, Bronner IF, Azmani A, Severijnen
70. Poorkaj P, Bird TD, Wijsman E, Nemens E, Garruto LA, Kamphorst W, Ravid R, Rizzu P, Willemsen R,
RM, Anderson L, Andreadis A, Wiederholt WC, Heutink P.  The ΔK280 mutation in MAP tau favors
Raskind M, Schellenberg GD.  Tau is a candidate exon 10 skipping in vivo. J Neuropathol Exp Neurol.
for chromosome 17 frontotemporal dementia. Ann 2007;66:17–25.
Neurol. 1998;43:815–25. 83. Von Bergen M, Friedhoff P, Biernat J, Heberle J,
71. Qiang L, Sun X, Austin TO, Muralidharan H, Jean Mandelkow EM, Mandelkow E.  Assembly of tau
DC, Liu M, Yu W, Baas PW.  Tau does not stabilize protein into Alzheimer paired helical filaments
axonal microtubules but rather enables them to have depends on a local sequence motif (306VQIVYK311)
long labile domains. Curr Biol. 2018;28:2181–9. forming β-structure. Proc Natl Acad Sci U S A.
72. Sawaya MR, Nelson R, Ivanova MI, Sievers SA,
2000;97:5129–34.
Apostol MI, Thompson MJ, Balbirnie M, Wiltzius JJ, 84. Wegmann S, Eftakharzadeh B, Tepper K, Zoltowaka
McFarlane HT, Madsen AO, et al. Atomic structures KM, Bennett RE, Dujardin S, Laskowski PR,
of amyloid cross-beta spines reveal varied steric zip- MacKenzie D, Kamath T, Commins C, et  al. Tau
pers. Nature. 2007;447:453–7. protein liquid-liquid phase separation can initiate tau
73. Sharp AJ, Hansen S, Selzer RR, Cheng Z, Regan R, aggregation. EMBO J. 2018;37:e98049.
Hurst JA, Stewart H, Price SM, Blair E, Hennekam 85. Weingarten MD, Lockwood AH, Hwo SY, Kirschner
RC, et al. Discovery of previously identified genomic MW. A protein factor essential for microtubule assem-
disorders from the duplication architecture of the bly. Proc Natl Acad Sci U S A. 1975;72:1858–62.
human genome. Nat Genet. 2006;38:1038–42. 86. Wilson DM, Binder LI. Free fatty acids stimulate the
74. Shaw-Smith C, Pittman AM, Willatt L, Martin H, polymerization of tau and amyloid beta peptides. Am
Rickman L, Gribble S, Curley R, Cumming S, Dunn J Pathol. 1997;150:2181–95.
C, Kalaitzopoulos D, et al. Microdeletion encompass- 87. Wischik CM, Novak M, Thøgersen HC, Edwards PC,
ing MAPT at chromosome 17q21.31 is associated Runswick MJ, Jakes R, Walker JE, Milstein C, Roth
1  Ordered Assembly of Tau Protein and Neurodegeneration 21

M, Klug A. Isolation of a fragment of tau derived from Alzheimer’s and Pick’s diseases. eLife. 2019;8:43584.
the core of the paired helical filament of Alzheimer Biorxiv
disease. Proc Natl Acad Sci U S A. 1988;85:4506–10. 92. Zhang X, Lin Y, Eschmann NA, Zhou H, Rauch JN,
88. Wischik CM, Novak M, Edwards PC, Klug A,
Hernandez I, Guzman E, Kosik KS, Han S.  RNA
Tichelaar W, Crowther RA.  Structural character- stores tau reversibly in complex coacervates. PLoS
ization of the core of the paired helical filament Biol. 2017;15:e2002183.
of Alzheimer disease. Proc Natl Acad Sci U S A. 93. Zhao Y, Tseng IC, Heyser CH, Rockenstein E, Mante
1988;85:4884–8. M, Adame A, Zheng Q, Huang T, Wang X, Arslan PE,
89. Yoshida H, Goedert M. Molecular cloning and func- et  al. Appoptosin-mediated caspase cleavage of tau
tional characterization of chicken brain tau: Isoforms contributes to progressive supranuclear palsy patho-
with up to five tandem repeats. Biochemistry. genesis. Neuron. 2015;87:963–75.
2002;41:15203–11. 94. Zhong Q, Congdon EE, Nagaraja HN, Kuret J.  Tau
90. Zhang CC, Zhu JX, Wan Y, Tan L, Wang HF, Yu JT, isoform composition influences rate and extent of fila-
Tan L.  Meta-analysis of the association between ment formation. J Biol Chem. 2012;287:20711–9.
variants in MAPT and neurodegenerative diseases. 95. Zollino M, Orteschi D, Murdolo M, Lattante S,

Oncotarget. 2017;8:4494–45007. Battaglia D, Stefanini C, Mercuri E, Chiurazzi P,
91. Zhang W, Falcon B, Murzin AG, Fan J, Crowther Neri G, Marangi G. Mutations in KANSL1 cause the
RA, Goedert M, Scheres SHW. Heparin-induced tau 17q21.31 microdeletion syndrome phenotype. Nat
filaments are polymorphic and differ from those in Genet. 2012;44:636–8.
 Structure of NFT: Biochemical
Approach 2
Masato Hasegawa

Introduction have been found to closely correlate with disease

symptoms and progression of AD.  In addition,
Neurofibrillary tangles (NFTs) and senile plaques recent findings on prion-like spreading of intra-
(SPs) are the neuropathological hallmarks of cellular abnormal proteins support the idea that
Alzheimer’s disease (AD). SPs are extracellular the formation and propagation of abnormal tau
aggregates composed of amyloid β (Αβ) pep- protein pathologies are the central events leading
tides, which are cleaved from amyloid β protein to neurodegeneration in AD.  This chapter
precursor (APP). NFTs are intraneuronal patho- describes biochemical and immunochemical
logical structures, which are mainly composed of approaches used to study pathological tau in AD
microtubule-associated protein tau. Before these and other related disorders.
constituent proteins were found, there were many
discussions about which of them is related to the
pathogenesis and progression of AD, how they Tau Protein
are related each other, and so on. But, after the
discovery of mutations in the APP gene in famil- Tau is a heat-stable microtubule-associated pro-
ial forms of AD in 1990, the “amyloid hypothe- tein that promotes microtubule assembly and sta-
sis” was proposed and many researchers focused bility [1]. It is highly expressed in neurons and is
on the mechanisms of Aβ production and aggre- normally located in axons. In adult human brain,
gation, and their regulation. However, clinical tri- six tau isoforms (352–441 amino acid residues)
als targeting Αβ with Αβ vaccines have so far had are generated by mRNA alternative splicing from
no effect on tau pathologies or little effect on the MAPT gene on chromosome 17 [2].
clinical symptoms, even though Αβ deposits were Expression of tau isoforms is regulated in the
significantly reduced. These results suggest that developmental stages. Namely, the shortest iso-
aggregation of Αβ is not an appropriate target for form is expressed in fetal brain, and multiple iso-
AD therapy, and also that the Aβ hypothesis forms are gradually expressed during
should be reconsidered. Conversely, tau patholo- development by insertion of exons 2, 3 and 10;
gies in NFTs and their spread in brains of patients these isoforms have apparent molecular weights
of 48–67  kDa [3]. Among them, three-repeat
M. Hasegawa (*) (3R) tau and four-repeat (4R) tau are produced by
Department of Dementia and Higher Brain Function, the exclusion and inclusion of exon 10, respec-
Tokyo Metropolitan Institute of Medical Science, tively. The expression ratio of 3R tau and 4R tau
Tokyo, Japan is almost 1: 1 in adult human brain. This ratio is
e-mail: hasegawa-ms@igakuken.or.jp

© Springer Nature Singapore Pte Ltd. 2019 23

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_2
24 M. Hasegawa

important, because mutations that influence the peptides generated by digestion with lysyl-­
expression ratio of 3R tau and 4R tau induce endopeptidase, showing that they are composed
tauopathies exhibiting tau inclusions composed of the carboxyl-third of tau and ubiquitin [11,
of the increased tau isoforms [4]. Tau expression 12]. Meanwhile, the presence of SDS-soluble tau
in rodents is slightly different from that in human, in AD brains had been reported, after tau was
in that 3R tau is completely replaced by 4R tau in found to be a component of NFTs (SDS-soluble
the adult stage. The difference in the splicing of tau is easy to analyze by SDS-PAGE and immu-
exon10 between the species that determines the noblotting). Greenberg and Davis reported a
3R tau and 4R tau ratios appears to be due to dif- preparation method of PHFs from AD brains
ferences in the splicing enhancer and suppressor using Sarkosyl [13]. Lee’s group also purified
mRNA sequences and stem loop structures in SDS-soluble, Sarkosyl-insoluble PHFs by means
exon 10 and intron 10 [5]. Tau is a phosphopro- of a similar protocol and demonstrated that
tein and phosphorylation decreases its hyperphosphorylated tau is the major component
microtubule-­polymerizing ability. Therefore, it is of PHFs [14]. Since then, many tau researchers
suggested that phosphorylation may regulate have been working on SDS-soluble, Sarkosyl-­
microtubule dynamics. insoluble tau, because the SDS-soluble PHFs
may represent an early stage of PHFs (NFTs). It
is also advantageous that they are more treatable
Tau in AD Brain and easy to analyze. Thus, AD brains may con-
tain both SDS-soluble, relatively young PHFs
Tau is structurally less-ordered, and is classified and SDS-insoluble old PHFs (NFTs) that are
as a natively unfolded protein. Normal tau pro- mainly composed of full-length phosphorylated
tein is highly soluble and is mostly recovered in tau and ubiquitinated carboxyl-third of tau,
buffer-soluble fractions after ultracentrifugation respectively, because tau accumulation continues
of brain homogenates. It can be relatively easily for over 10 years from AD onset to death [15].
purified by boiling treatment and subsequent ion-­
exchange column chromatography [6] or by per-
chloric acid (PCA) extraction [7]. However, in Solubility of NFTs and PHFs
AD, the abnormal proteins are accumulated as
NFTs in neuronal cell bodies and neuropil threads The solubility of NFTs can be analyzed on slide
(NTs) in their processes. Ultrastructurally, they glasses by conventional Congo red staining or
are present as bundles of unique filamentous immunohistochemistry of smears of AD brain
structures, paired helical filaments (PHFs) and homogenates with or without various agents. As
straight filaments (SFs) in NFTs and NTs [8]. In shown in Fig. 2.1, numerous NFTs and NTs can
order to purify NFTs, Selkoe et  al. investigated be seen in AD brain homogenate in Tris-saline
their solubility in various buffers, solvents and (TS), as visualized with anti-tau antibodies such
denaturants, and reported that NFTs are highly as pS396 or AT8. The numbers of NFTs and NTs
insoluble in protein denaturants such as SDS, are not decreased by addition of 1% Triton-X100
urea, reducing agent and guanidine [9]. They (TX: a noionic detergent), but most of the NFTs
applied this unusual property to prepare NFTs disappear upon treatment with ionic detergents
from AD brains; they partially purified the NFTs such as 1% Sarkosyl or SDS (Fig.  2.1).
as SDS-insoluble tangles, and raised antibodies Immunoblot analysis of the pellets after 15 K rpm
to the tangles [10]. Ihara, who had worked with centrifugation of the homogenates confirmed that
Selkoe’s group at Harvard University, further pS396-positive tau in the 15  K-ppt decreased
investigated the SDS-insoluble tangles together after treatment with Sarkosyl or SDS (Fig. 2.1b)
with Kondo after his return to Tokyo. He solubi- (AT8 failed to detect tau because of the low con-
lized the tangles in formic acid and analyzing the centration). However, Sarkosyl-treated tau in the
2  Structure of NFT: Biochemical Approach 25

Fig. 2.1  Immunohistochemistry of AD brain homogenates smeared on slide glasses in extraction buffer with or without
Triton-X100 (TX), Sarkosyl (Sar) or SDS

supernatant was precipitated after ultracentrifu- Biochemical Approaches to PHF-Tau

gation at 50 K rpm and was detected with both in AD Brains
AT8 and pS396, whereas SDS-treated tau was
mostly recovered in the supernatant (Fig.  2.2). In order to clarify the molecular mechanisms of
These results suggest that NFTs are dissociated tau aggregation and NFTs formation in AD, we
to filamentous components by Sarkosyl, but still first took an immunochemial approach in Ihara’s
have structure, whereas most NFTs are solubi- lab at Tokyo Metropolitan Institute of Gerontology.
lized by SDS and recovered in the 50 K-sup. The At that time, there was no powerful method to ana-
presence of PHFs and SFs can be confirmed by lyze abnormal tau directly, and it was difficult to
immuno-EM observation of the Sarkosyl-­ purify tau from AD brains. So, we prepared
insoluble fractions of AD. Thus, the abnormal tau Sarkosyl-insoluble fractions from AD brains and
forming these filaments can be enriched from used them to immunize BALB/c mice, then
Sarkosyl-treated brain homogenates by means of screened clones for production of antibodies that
differential centrifugation [16]. This is the reason stained NFTs. Fortunately, several monoclonals
why many researchers use Sarkosyl extraction that strongly stained NFTs were obtained. All
protocols for preparation of pathological tau pro- these antibodies recognized tau in fetal or juvenile
teins from brains of patients and cellular and ani- brains, but hardly detected the tau in adult brains.
mal models. The purification, characterization We then analyzed the epitopes of two typical
and analysis of abnormal tau prepared from AD clones (M4 and C5) in fetal tau and PHF-tau.
brains have established that abnormally phos- When fetal tau was treated with E. coli alkaline
phorylated tau is the major component of PHFs phosphatase at 37 °C, a mobility shift of the tau
and SFs in NFTs. band was observed and immunoreactivity to C5
26 M. Hasegawa

Fig. 2.2  Immunoblot analyses of tau in AD brain. Brain fuged at 50 K-rpm for 20 min. The pellets and supernatant
homogenates were sonicated briefly and centrifuged at were solubilized in SDS sample buffer, applied to gel, and
15 K-rpm for 10 min. The supernatants were ultracentri- blotted with AT8 and pS396

was completely abolished, whereas the M4 immu- 406 and 198–250, respectively [17]. During these
noreactivity was unchanged, and was only lost studies, I learned a lot about protein-chemical
after alkaline phosphatase treatment at methods, and also thought it might be faster and
67 °C. These results suggested that the monoclo- more comprehensive if I could directly analyze the
nals recognize distinct phosphorylated epitopes of phosphorylation sites and other abnormalities of
tau [17]. Then, we tried to map the epitopes in fetal PHF-tau compared with normal tau. So, I started a
tau and PHF-tau. However, this proved difficult, direct approach to PHF-tau, in which all the
and the immunoreactivities often disappeared after digested peptides of purified PHF-tau and normal
protease digestion and/or separation by reverse- tau with or without dephosphorylation were ana-
phase HPLC. Therefore, we tested various chemi- lyzed using a protein sequencer and mass spec-
cal cleavage methods such as CNBr treatment and trometry. This was about 1990, when mass
enzymatic digestions with lysylendopeptidase, spectrometry was just being introduced for analy-
trypsin and other enzymes, together with peptide sis of peptides and proteins, and I spent about a
sequencing, and finally identified the recognition year in Dr. Takio’s lab at RIKEN to study the tech-
sites of the C5 and M4 antibodies as residues 386– niques and analyze the PHF-tau.
2  Structure of NFT: Biochemical Approach 27

Protein-Chemical Approach tion at these sites changes during development

to PHF-Tau [20]. In collaboration with Dr. Morishima, we
also investigated PHF-tau by protein sequencing
We succeeded in purifying the AD abnormal tau and mass spectrometric analyses of peptides
(PHF-tau) from Sarkosyl-insoluble fraction by digested with immobilized trypsin and some
solubilization with 6 M guanidine-HCl, gel filtra- other proteases, and determined ~20 abnormal
tion and reverse-phase HPLC. The purified insol- phosphorylation sites. Furthermore, we gener-
uble and soluble tau from AD brains with or ated several antibodies to these novel phosphory-
without alkaline phosphatase treatment were lation sites and confirmed that the sites are really
digested with lysyl endopeptidase, the peptide phosphorylated in tau in NFTs of AD brains [21].
maps were compared, and all the fractions were Half of the phosphorylation sites identified in the
analyzed by protein sequencing and mass spec- abnormal tau in AD brains are proline-directed,
trometric analysis. In the initial analysis with ion-­ but half are non-proline-directed, suggesting that
spray mass spectrometry, we identified three multiple kinases must be involved in the phos-
phosphorylation sites (Thr231, Ser235 and phorylation. During the course of analysis of
Ser262) and two extensively phosphorylated PHF-tau, we also noticed that ubiquitinated
regions located in the amino- and carboxyl-­ smeared tau can be separated from non-­
terminal portions of the microtubule-binding ubiquitinated tau. Then, we digested the samples,
domain [18]. Thr231 and Ser235 were fully separated the peptides by HPLC, and compared
phosphorylated in PHF-tau, but Ser262 was only the peptide maps to identify the targets of ubiqui-
partially phosphorylated, because unphosphory- tin. As expected, we identified several ubiquiti-
lated Ser262 peptide was detected in a larger nated tau peptides and a Lys-48-ubiquitinated
amount than the phosphorylated peptide. We also multi-ubiquitin peptide, demonstrating that the
found that two asparagine residues (Asn167 and ubiquitin target protein in NFTs is tau [15]. Since
Asn279) are deamidated. In addition, both 3R tau only a small proportion of the ubiquitin formed
and 4R tau specific peptides were identified in multi-ubiquitin chains, it seemed that most of the
PHF-tau at a similar ratio to that seen in normal ubiquitin in the tau occurred as a monoubiquiti-
tau [17]. These results suggested that both 3R and nated form. In addition, there was ubiquitin-­
4R tau isoforms are accumulated in AD, but in negative tau and smears, and the tau was much
abnormally phosphorylated and deamidated less processed, strongly suggesting that tau accu-
states. At the same time, Goedert et al. reported mulation and processing precede ubiquitination.
that all six brain tau isoforms are deposited in AD
brain in hyper-phosphorylated forms [19].
As described above, mass spectrometry is a Attempts to Recapitulate
very powerful technique to identify peptides with the Abnormal Phosphorylation
molecular mass of 500~3000, but is less effective of Tau
for heavily phosphorylated peptides, because it is
difficult to ionize these highly negatively charged These immunochemical and protein-chemical
peptides. Consequently, we struggled to deter- approaches to tau in NFTs demonstrated that the
mine the phosphorylation sites in these heavily major biochemical difference between PHF-tau
phosphorylated regions. So, we introduced a and normal tau was the phosphorylation.
method to determine the phosphorylation sites by Therefore, we attempted to recapitulate the phos-
beta elimination of phosphopeptides followed by phorylation of tau in vitro. Recombinant tau was
protein sequencing. We first analyzed tau in fetal phosphorylated with various kinases, such as
and adult rat brains, which are less highly phos- Cdk5, MAPK, GSK3, CaMK and PKA, in both
phorylated than PHF-tau, and determined ~10 single and sequential multiple passes, and the
phosphorylation sites in fetal and adult normal outcome was monitored with various
rat brains; we also established how phosphoryla- phosphorylation-­ dependent anti-tau antibodies.
28 M. Hasegawa

However, no heavily phosphorylated state similar positively charged regions such as the
to that in PHF-tau was reproduced in the in vitro microtubule-­binding region. It induces confor-
experiments with recombinant tau (unpublished mational changes of tau, as a result of which
data). We also tested tau phosphorylation with rat enzymes such as kinases can access the substrate
brain extracts in the presence of phosphatases more easily. Like heparin, tubulin, a physiologi-
inhibitors such as okadaic acid, NaF, etc. The cal partner of tau, can also bind to tau and pro-
brain extracts actually phosphorylated tau very mote phosphorylation. When I started work in
well and a big mobility shift of the tau band was Goedert’s lab in MRC in 1995, Goedert and Jakes
observed in the presence of inhibitors; however, were investigating the effect of heparin on in vitro
the phosphorylation state was not the same as tau phosphorylation. We noticed a high-­
that of PHF-tau, and tau aggregation or fibril for- molecular-­weight band on gel in the presence of
mation was never detected. Most tau researchers heparin, and Crowther looked at the sample by
were also trying to recapitulate the abnormal tau electron microscopy. Surprisingly, PHF-like tau
phosphorylation and find the responsible kinases filaments were seen in the sample, and after vari-
or phosphatases. At that time, Matsuo et  al. ous experiments it was found that heparin and
reported that normal soluble tau in biopsy brain negatively charged glycosaminoglycans induce
samples is highly phosphorylated, but is rapidly tau fibril formation [25, 26]. These results sug-
dephosphorylated in the postmortem period [22]. gest that tau can assemble into PHF-like fila-
This raised the question of why PHF-tau remains ments without phosphorylation.
hyperphosphorylated after a long (~10  h) post-
mortem delay. To address this, we investigated
dephosphorylation of fetal and PHF-tau  athological Tau Proteins in Various
P
(Sarkosyl-insoluble pellets prepared from AD Neurological Diseases
brain) by several protein phosphatases in  vitro.
Protein phosphatases 1 and 2A and calcineurin Tau pathologies are seen not only in AD, but also
could dephosphorylate fetal-tau, whereas intact in various other dementing neurological diseases,
PHF-tau without guanidine-HCl denaturation such as Pick’s disease (PiD), progressive supra-
was hardly dephosphorylated by these protein nuclear palsy (PSP), and corticobasal degenera-
phosphatases. However, after denaturing treat- tion (CBD) (Fig.  2.3). For this reason, it was
ments with guanidine-HCl or formic acid, the thought that tau lesions might be a final common
PHF-tau was easily dephosphorylated by these pathway of neurodegeneration, rather than the
protein phosphatases [23]. The results strongly cause. However, mutations in the tau gene (MAPT)
indicate that PHF-tau deposited in NFTs and NTs were discovered in familial forms of dementias
in AD brain is resistant to dephosphorylation, with tau pathologies in 1998 [27, 28, 29], demon-
because the tau is present in filamentous struc- strating that abnormality of tau is sufficient to
tures as PHFs in NFTs. Matsuo’s paper thus cause disease. Although the clinicopathological
revealed the limitations of the analysis of soluble features of dementia caused by MAPT mutations
proteins in postmortem brain tissues, and showed are diverse [30], there are strong correlations
that caution is needed in interpreting data between the positions (or effects) of mutations
obtained from postmortem samples. However, and tau pathologies, suggesting that neurodegen-
phosphorylation was still seen as central in the eration occurs via tau accumulation. Mutations
formation of pathological tau by AD researchers, outside exon 10 have been shown to affect tau
because PHF-tau is much more highly phosphor- conformation, cause tau dysfunction, and pro-
ylated than biopsy tau or fetal tau. Then, heparin mote tau aggregation, resulting in both 3R and 4R
was reported as a factor that accelerates tau phos- tau pathology. The intronic mutations in intron 10
phorylation [24]. Heparin is a negatively charged and many of the mutations in exon 10 affect tau
glycosaminoglycan, and binds to tau through splicing, changing the ratio of 3R tau and 4R tau,
2  Structure of NFT: Biochemical Approach 29

Fig. 2.3  Characteristic tau inclusions in AD, PiD, CBD, tion. All six tau isoforms are accumulated in AD, whereas
PSP and FTDP-17 with intron 10 + 16 mutation, and full-­ only 3R tau isoforms are deposited in PiD, and only 4R
length tau bands (isoforms) in these tauopathies detected tau isoforms are accumulated in PSP, CBD and FTDP-17
by immunoblot analysis after complete dephosphoryla- with intron 10 mutation

and resulting in accumulation of increased or C-terminal fragments (CTFs) of the insoluble tau
affected tau isoforms. These fi­ ndings clearly dem- were shown to be different between PSP and CBD
onstrated that tau abnormalities cause accumula- [33], suggesting that the difference in the CTFs
tion of tau and neurodegeneration. However, most banding pattern may reflect the structural differ-
of these tauopathies are sporadic, and no signifi- ences of tau filaments in PSP and CBD. In fact,
cant changes in the ratio of 3R/4R tau isoforms ultrastructure of the pathological tau filaments
were detected [31]. Furthermore, the morpholo- was shown to be morphologically different
gies of tau inclusions, their distributions, and their between the diseases [34]. To address this issue,
biochemical features are different among the dis- we investigated the pathological tau and the prote-
eases, suggesting that the molecular mechanisms ase-resistant fragments in these various tauopa-
may be different among the diseases and also dif- thies including AD, PiD, PSP, CBD and FTDP-17
ferent from the MAPT mutation cases. So, why with intronic mutations. We found that the tryp-
tau pathologies are so diverse and different sin-resistant core units of the tau filaments were
between the diseases? One reason is that tau iso- composed of different tau repeat regions located
forms accumulated in these tauopathy brains are between residues 243–406, indicating that the
different in the different diseases. In AD, all six conformations of the cores are disease-specific
brain tau isoforms (3R tau and 4R tau isoforms) [16] (Fig. 2.4). These patterns are reminiscent of
are deposited in neuronal cells as NFTs and NTs. the proteinase-­K-­resistant bands in human prion
In PiD, only 3R tau isoforms are deposited as disorders, and suggest that tauopathies may be
Pick’s bodies and related inclusions, while in PSP caused by accumulation of toxic tau ‘prions’ in
and CBD only 4R tau isoforms are accumulated the brain.
in neurons and glial cells, with some characteris- Thus, the phenotypical differences in the tau
tic structures [16, 32] (Fig.  2.3). However, both pathologies can be classified biochemically
CBD and PSP are 4R tauopathies, but they are according to the banding patterns of C-terminal
neuropathologically distinguished, suggesting fragments and protease-resistant tau, which rep-
that the other difference causes these tau pheno- resent the abnormal structures of tau aggregates
types. Interestingly, the banding patterns of [16, 33] (Fig. 2.4).
30 M. Hasegawa

Fig. 2.4  Summary of trypsin-resistant regions of patho- in terms of both the isoforms and regions. The cores are
logical tau from AD, PiD, PSP, CBD and FTDP-17 with localized to start from the middle of the 1st repeat of 3R
intron 10 mutation. The tau fragments identified as tau and the 2nd repeat of 4R tau, which are different
trypsin-­resistant cores in these tauopathies are indicated from those of the other tauopathies
by solid lines. AD-tau cores are distinct from the others

cannot be accessed by kinases [16]. In other

Significance of Tau Phosphorylation tauopathies including AD, the Ser262 residue is
located outside the core regions of the filaments,
The significance of tau phosphorylation for the so it can be partially phosphorylated. Furthermore,
aggregation or assembly of tau into filaments still Ser356, which is phosphorylated by the same
remains controversial. However, biochemical kinases that phosphorylate Ser262, has never
analyses of various kinds of pathological tau in been detected as a phosphorylated residue in the
tauopathy brains, including PiD, CBD, PSP, analyses of the pathological tau in theses tauopa-
FTDP-17, suggest that phosphorylation of tau thies [16]. This is in good agreement with the fact
may occur after aggregation or filament forma- that Ser356 is located in the trypsin-resistant
tion, because most of the phosphorylation sites cores of all these tau filaments. Recent Cryo-EM
are located outside the core regions of the fila- studies of tau filaments prepared from brains of
ments. It should be emphasized that, exception- patients confirmed that Ser356 is located in the
ally, Ser262 is not phosphorylated in tau in PiD cross-beta structures of tau filaments from AD
[35], which is a 3R tauopathy. In PiD, tau resi- and PiD [36, 37], and that Ser262 is located in a
dues 243–387 form the filament cores, suggest- position that cannot be accessed by kinases in
ing that Ser262 is buried in the core regions and PiD (Fig. 2.4).
2  Structure of NFT: Biochemical Approach 31

 olecular Mechanisms of Tau

M aggregates cannot be degraded by the attacks of
Aggregation and NFT Formation kinases, chaperones or proteases; further, phos-
phatases cannot dephosphorylate the aggregates.
As described above and shown in Fig. 2.5, both The resulting protein aggregates may have
3R and 4R tau proteins accumulate in AD chronic toxic effects on proteasome, lysosome
patient’s brain as amyloid-like filamentous struc- and autophagy systems in the cells (Fig.  2.6).
tures that show physicochemical properties of Furthermore, the protein seeds may propagate
β-sheet structure, such as Congo red or thioflavin from cell to cell and spread throughout the brains
staining. In PiD, only 3R tau isoforms assemble through both known and unknown mechanisms.
into filaments. In PSP and CBD, only 4R tau iso- It should be emphasized that considerable
forms assemble into filaments but the core amounts of pathological tau have already accu-
regions of these aggregates are different, forming mulated by the time symptoms appeared.
different tau filaments. As a mechanism of amy- Therefore, it is important to block the formation
loid fibril formation, seed-dependent aggregation of abnormal tau and the spreading or cell-to-cell
has been proposed and is now widely accepted. A propagation of abnormal tau protein for therapy
small aggregation nucleus is formed first and of AD and tauopathies.
then normal proteins bind to this seed nucleus
and assemble into filaments by converting the
normal protein to an abnormal form on the seed Conclusions
as a template. As illustrated in Figs. 2.5 and 2.6,
PHFs and SFs can be formed by coupling of 3R Immunochemical and protein-chemical
tau and 4R tau in neurons expressing both 3R and approaches have revealed that hyperphosphory-
4R tau isoforms, for reasons, such as aging, lated and partially ubiquitinated tau is the major
infection, injury, inflammation, or some other component of NFTs, one of the neuropathologi-
factors or events. The abnormal tau proteins cal hallmarks of AD and related neurological dis-
expand by prion-like conversion. The filamen- orders. The most important feature of the
tous tau is very stable and the abnormal protein pathological tau is its abnormal conformation,

Fig. 2.5  Schematic illustration of possible tau assemblies in filament core domains composed of 3-repeat tau isoforms
AD, PiD, CBD and PSP. The blue and red round shapes illus- and 4-repeat tau isoforms, respectively. These filament core
trate the normal 3R tau and 4R tau isoforms, respectively. domains may act as the seeds for the seed-­dependent prion-
The blue triangular and red tetragonal shapes illustrate the like conversion and abnormal tau filament formation
32 M. Hasegawa

Fig. 2.6  Schematic illustration of tau aggregation in a may be toxic to various intracellular functions, such as the
neuronal cell in an AD brain. Both 3R and 4R tau assem- proteasome, lysosome and autophagy systems, resulting
ble into filamentous structures by interacting with the in induction of cell death. Phosphatases do not dephos-
seed. Kinases, chaperones and proteases are unable to phorylate residues in these filaments because of the struc-
degrade or restore the proteins. The aggregated filaments tural stability of the filaments

which enables it to accumulate as amyloid-like pathogenesis and progression of AD and other

filamentous structures. The abnormal tau has neurodegenerative diseases.
prion-like properties, i.e., it converts normal tau Method: take 2  μL of each sample of brain
into filamentous form by working as a template homogenate, smear on slide glass, and stain with
seed, and can propagate from cell to cell. pS396 (or AT8)
Importantly, accumulated tau isoforms and the
abnormal structures of tau proteins are different
among different tauopathies. Clear structural dif- References
ferences can be detected in the protease-resistant
core regions of the tau filaments. Most of the 1. Weingarten MD, Lockwood AH, Hwo SY, Kirschner
MW. A protein factor essential for microtubule assem-
post-translational modifications found in the bly. Proc Natl Acad Sci U S A. 1975;72(5):1858–62.
pathological tau in brains of patients may be the 2. Goedert M, Crowther RA, Garner CC.  Molecular
results of responses by intracellular defense sys- characterization of microtubule-associated proteins
tems to the abnormal proteins, but may not be tau and MAP2. Trends Neurosci. 1991;14(5):193–9.
3. Goedert M, Jakes R. Expression of separate isoforms
causative for aggregation. Proteases and phos- of human tau protein: correlation with the tau pattern
phatases may be unable to degrade or dephos- in brain and effects on tubulin polymerization. EMBO
phorylate the filamentous tau, which explains J. 1990;9(13):4225–30.
why hyperphosphorylated tau can be detected in 4. Goedert M, Jakes R, Crowther RA.  Effects of fron-
totemporal dementia FTDP-17 mutations on heparin-­
the brains of patients even after long postmortem induced assembly of tau filaments. FEBS Lett.
intervals. Recent progress in Cryo-EM and its 1999;450(3):306–11.
application to amyloid-like pathological protein 5. D'Souza I, Schellenberg GD. Tau exon 10 expression
filaments should provide new insights into the involves a bipartite intron 10 regulatory sequence
2  Structure of NFT: Biochemical Approach 33

and weak 5′ and 3′ splice sites. J  Biol Chem. 20. Watanabe A, Hasegawa M, Suzuki M, Takio K,

2002;277(29):26587–99. Morishima-Kawashima M, Titani K, et  al. In vivo
6. Lindwall G, Cole RD. The purification of tau protein phosphorylation sites in fetal and adult rat tau. J Biol
and the occurrence of two phosphorylation states of Chem. 1993;268(34):25712–7.
tau in brain. J Biol Chem. 1984;259(19):12241–5. 21. Morishima-Kawashima M, Hasegawa M, Takio K,
7. Grundke-Iqbal I, Iqbal K, Quinlan M, Tung YC, Zaidi Suzuki M, Yoshida H, Titani K, et al. Proline-directed
MS, Wisniewski HM. Microtubule-associated protein and non-proline-directed phosphorylation of PHF-­
tau. A component of Alzheimer paired helical fila- tau. J Biol Chem. 1995;270(2):823–9.
ments. J Biol Chem. 1986;261(13):6084–9. 22. Matsuo ES, Shin RW, Billingsley ML, Van deVoorde
8. Kidd M.  Paired helical filaments in electron A, O'Connor M, Trojanowski JQ, et  al. Biopsy-­
microscopy of Alzheimer’s disease. Nature. derived adult human brain tau is phosphorylated at
1963;197:192–3. many of the same sites as Alzheimer's disease paired
9. Selkoe DJ, Ihara Y, Salazar FJ. Alzheimer’s disease: helical filament tau. Neuron. 1994;13(4):989–1002.
insolubility of partially purified paired helical fila- 23. Yamamoto H, Hasegawa M, Ono T, Tashima K,

ments in sodium dodecyl sulfate and urea. Science. Ihara Y, Miyamoto E.  Dephosphorylation of fetal-­
1982;215(4537):1243–5. tau and paired helical filaments-tau by protein
10. Ihara Y, Abraham C, Selkoe DJ.  Antibodies to
phosphatases 1 and 2A and calcineurin. J  Biochem.
paired helical filaments in Alzheimer's disease 1995;118(6):1224–31.
do not recognize normal brain proteins. Nature. 24. Yang SD, Yu JS, Shiah SG, Huang JJ. Protein kinase
1983;304(5928):727–30. FA/glycogen synthase kinase-3 alpha after hepa-
11. Mori H, Kondo J, Ihara Y. Ubiquitin is a component rin potentiation phosphorylates tau on sites abnor-
of paired helical filaments in Alzheimer’s disease. mally phosphorylated in Alzheimer's disease brain.
Science. 1987;235(4796):1641–4. J Neurochem. 1994;63(4):1416–25.
12. Kondo J, Honda T, Mori H, Hamada Y, Miura R, 25. Goedert M, Jakes R, Spillantini MG, Hasegawa M,
Ogawara M, et  al. The carboxyl third of tau is Smith MJ, Crowther RA.  Assembly of microtubule-­
tightly bound to paired helical filaments. Neuron. associated protein tau into Alzheimer-like filaments
1988;1(9):827–34. induced by sulphated glycosaminoglycans. Nature.
13. Greenberg SG, Davies P. A preparation of Alzheimer 1996;383(6600):550–3.
paired helical filaments that displays distinct tau pro- 26. Hasegawa M, Crowther RA, Jakes R, Goedert

teins by polyacrylamide gel electrophoresis. Proc Natl M. Alzheimer-like changes in microtubule-associated
Acad Sci U S A. 1990;87(15):5827–31. protein tau induced by sulfated glycosaminogly-
14. Lee VM, Balin BJ, Otvos L Jr, Trojanowski
cans. Inhibition of microtubule binding, stimula-
JQ.  A68: a major subunit of paired helical fila- tion of phosphorylation, and filament assembly
ments and derivatized forms of normal tau. Science. depend on the degree of sulfation. J  Biol Chem.
1991;251(4994):675–8. 1997;272(52):33118–24.
15. Morishima-Kawashima M, Hasegawa M, Takio K, 27. Hutton M, Lendon CL, Rizzu P, Baker M, Froelich
Suzuki M, Titani K, Ihara Y. Ubiquitin is conjugated S, Houlden H, et  al. Association of missense and
with amino-terminally processed tau in paired helical 5′-splice-site mutations in tau with the inherited
filaments. Neuron. 1993;10(6):1151–60. dementia FTDP-17. Nature. 1998;393(6686):702–5.
16. Taniguchi-Watanabe S, Arai T, Kametani F, Nonaka T, 28. Spillantini MG, Murrell JR, Goedert M, Farlow MR,
Masuda-Suzukake M, Tarutani A, et al. Biochemical Klug A, Ghetti B. Mutation in the tau gene in familial
classification of tauopathies by immunoblot, pro- multiple system tauopathy with presenile dementia.
tein sequence and mass spectrometric analyses of Proc Natl Acad Sci U S A. 1998;95(13):7737–41.
sarkosyl-insoluble and trypsin-resistant tau. Acta 29. Poorkaj P, Muma NA, Zhukareva V, Cochran EJ,

Neuropathol. 2016;131(2):267–80. Shannon KM, Hurtig H, et al. An R5L tau mutation in
17. Hasegawa M, Watanabe A, Takio K, Suzuki M, Arai a subject with a progressive supranuclear palsy phe-
T, Titani K, et  al. Characterization of two distinct notype. Ann Neurol. 2002;52(4):511–6.
monoclonal antibodies to paired helical filaments: 30. Ghetti B, Oblak AL, Boeve BF, Johnson KA,

further evidence for fetal-type phosphorylation of Dickerson BC, Goedert M.  Invited review: fronto-
the tau in paired helical filaments. J  Neurochem. temporal dementia caused by microtubule-associated
1993;60(6):2068–77. protein tau gene (MAPT) mutations: a chameleon for
18. Hasegawa M, Morishima-Kawashima M, Takio K, neuropathology and neuroimaging. Neuropathol Appl
Suzuki M, Titani K, Ihara Y.  Protein sequence and Neurobiol. 2015;41(1):24–46.
mass spectrometric analyses of tau in the Alzheimer's 31. Umeda Y, Taniguchi S, Arima K, Piao YS, Takahashi
disease brain. J Biol Chem. 1992;267(24):17047–54. H, Iwatsubo T, et  al. Alterations in human tau tran-
19. Goedert M, Spillantini MG, Cairns NJ, Crowther
scripts correlate with those of neurofilament in spo-
RA.  Tau proteins of Alzheimer paired helical fila- radic tauopathies. Neurosci Lett. 2004;359(3):151–4.
ments: abnormal phosphorylation of all six brain iso- 32. Goedert M.  Tau filaments in neurodegenerative dis-
forms. Neuron. 1992;8(1):159–68. eases. FEBS Lett. 2018;592:2383–91.
34 M. Hasegawa

33. Arai T, Ikeda K, Akiyama H, Nonaka T, Hasegawa M, basis in brains from patients with Alzheimer’s dis-
Ishiguro K, et  al. Identification of amino-terminally ease, Pick's disease, progressive supranuclear palsy
cleaved tau fragments that distinguish progressive and corticobasal degeneration. Acta Neuropathol.
supranuclear palsy from corticobasal degeneration. 2003;105(5):489–98.
Ann Neurol. 2004;55(1):72–9. 36. Fitzpatrick AWP, Falcon B, He S, Murzin AG,

34. Arima K.  Ultrastructural characteristics of tau fila- Murshudov G, Garringer HJ, et  al. Cryo-EM struc-
ments in tauopathies: immuno-electron microscopic tures of tau filaments from Alzheimer’s disease.
demonstration of tau filaments in tauopathies. Nature. 2017;547(7662):185–90.
Neuropathology. 2006;26(5):475–83. 37. Falcon B, Zhang W, Murzin AG, Murshudov G,

35. Arai T, Ikeda K, Akiyama H, Tsuchiya K, Iritani S, Garringer HJ, Vidal R, et  al. Structures of filaments
Ishiguro K, et al. Different immunoreactivities of the from Pick's disease reveal a novel tau protein fold.
microtubule-binding region of tau and its molecular Nature. 2018;561(7721):137–40.
 Nuclear Magnetic Resonance
Spectroscopy Insights into Tau 3
Structure in Solution: Impact
of Post-translational Modifications

Clément Danis, Elian Dupré, Xavier Hanoulle,

Isabelle Landrieu, Alessia Lasorsa,
João Filipe Neves, Robert Schneider,
and Caroline Smet-Nocca

Introduction bly domain (Ser198-Leu441) consists of the second

part of the proline-rich region (annotated PRR),
The neuronal Tau protein belongs to the three (Tau3R) or four (Tau4R) microtubule-­
microtubule-­associated protein 2 (MAP 2)/Tau binding repeats (MTBR, numbered R1 to R4) [9,
subfamily and is mainly described as a cytosolic 18], and a C-terminal extension. MTBR repeats
protein, localized in axons [5, 6] and involved in R1 to R4 (31 or 32 residues) have similar
the regulation of tubulin polymerization and sequences. Most NMR studies are performed
microtubule stability [47]. This simplified view either using the longest Tau isoform, splicing
has been enriched by multiple studies demon- variant Tau2N4R with 441 amino-acid residues,
strating roles of Tau in many additional biologi- or the shortest isoform, Tau2N3R with 352
cal processes in various localizations including amino-acid residues.
nucleus, dendrites and extracellular medium [20, As a disordered protein Tau has no stable two-
25, 44, 46]. Six Tau isoforms are present in or three-dimensional structures and behaves as an
human brain (ranging in length from 352 to 441 ensemble of dynamic conformers. Yet, Tau is
amino acid residues), resulting from alternative described as undergoing conformational changes
splicing [17, 19]. Tau is composed of two main that lead to pathological forms and ultimately to
domains (Fig.  3.1). In the projection domain its aggregation. However, the nature of the initial
(Met1-Tyr197) (residue numbering as in the lon- conformational change(s) has remained elusive.
gest human isoform of Tau) an N-terminal exten- Since NMR spectroscopy is well-suited to study
sion (with 0, 1, or 2 inserts abbreviated as 0 N, highly dynamic systems, such as intrinsically
1 N, 2 N) and the first part of a proline-rich region disordered proteins in general, its use has
(residues 165–197) are distinguished. The assem- increased in the field over the last 15 years. This
has come along with improved resolution and
sensitivity offered by higher magnetic fields and
C. Danis · E. Dupré · X. Hanoulle · I. Landrieu (*) cryogenically cooled NMR probeheads. Yet,
A. Lasorsa · J. F. Neves · R. Schneider
C. Smet-Nocca NMR has limitations, the major one being that
Université de Lille, CNRS, UMR 8576 – UGSF – the protein under study needs to be isotopically-­
Unité de Glycobiologie Structurale et Fonctionnelle, labelled (with 15N and 13C stable isotopes) since
Lille, France not all atomic nuclei are magnetically active and
e-mail: isabelle.landrieu@univ-lille.fr

© Springer Nature Singapore Pte Ltd. 2019 35

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_3
36 C. Danis et al.

Fig. 3.1  Domains of Tau protein. N1 and N2 corre- of Tau aggregation, the PHF6∗ and PHF6 hexapeptides,
spond to 2 regions included, or not, depending on the are indicated. Some elements of secondary structure are
splicing variants. PRR: proline-rich region and MTBR: represented, as red arrows for β-strands, blue cylinders for
microtubule-binding region. The MTBR consists of four α-helices and green cylinder for polyproline II helix
partially repeated sequences named R1 to R4. The nuclei

detectable by NMR.  An important consequence same environment and conformational sampling

is that NMR research mainly concerns recombi- as in the full-length protein, underlining the
nant proteins, and cannot address native proteins absence of stable secondary or tertiary structure
extracted from tissues. The other major limitation in Tau [40]. Finally, the 13C chemical shift values
is the low intrinsic sensitivity of NMR, which of alpha- and beta-carbons (CA, CB) of a selected
reinforces the notion that NMR is best suited for number of resonances match the random coil
studying proteins produced in recombinant chemical shift values found in databases, con-
systems. firming the disordered nature.
Later work has allowed extensive assignment
of Tau resonances, and with it came more detailed
 au: A Disordered Protein
T information on Tau conformation in solution
with Local Secondary Structure [34]. These data confirm that no rigid secondary
structure elements are present in Tau. However, a
For analysis of protein NMR spectra, resonance few transient elements of secondary structure,
assignment has to be performed. That is, the restricted to short sequence stretches, were found.
amino acid residue in the protein that gives rise to Small β-strand tendency was observed in the
a resonance in the spectrum has to be identified, repeat region of Tau (R1 to R4), with an occu-
such that information about its chemical environ- pancy estimated around 10–25%, the highest
ment and secondary structure can be extracted value being found for the PHF6* and PHF6,
(Fig.  3.2a). In early work, performed on a VQIINK and VQIVYK hexapeptides located in
medium-field NMR spectrometer, a limited num- the R2 and R3 repeat, respectively. These pep-
ber of resonances of a two-dimensional (2D) 1H, tides thus adopt a β-strand conformation in about
15
N spectrum of Tau were assigned (Fig. 3.3) [30, 10–25% of the conformers. Alternatively, an
40]. This was not an easy task for the Tau protein, individual conformer can be considered to adopt
since it is a large protein for NMR with consider- a β-strand conformation 10–25% of the time.
able amino-acid sequence degeneracy – with Gly, This result is of interest since it shows that this
Ser, Pro and Lys residues each constituting 10% structural tendency can be linked to functional
of its amino-acid sequence. The 1H, 15N 2D spec- consequences, as the PHF peptides are known to
trum of Tau is typical of a disordered protein, be nuclei of aggregation and can self-assemble
with narrow dispersion of resonances in the into β-sheet aggregates [45]. Short stretches of
amide proton dimension, confirming the absence helical tendency were found in the N-terminal
of a global fold. In addition, the resonances in (114–123) and C-terminal (428–437) regions
spectra of peptides corresponding to short (Fig.  3.1). This local structuration is confirmed
stretches of Tau sequence match the correspond- based on a reported ensemble of Tau conformers
ing resonances in the full-length Tau protein. An calculated from a large number of various NMR
amino acid in the peptide thus experiences the parameters combined with small-angle X-ray
3  Nuclear Magnetic Resonance Spectroscopy Insights into Tau Structure in Solution: Impact… 37

Fig. 3.2  Key NMR techniques used to study the struc- probe (up to 25–40 Å) have a decreased intensity in the
ture of Tau. (a) The 1H-15N HSQC spectrum provides paramagnetic sample compared to the diamagnetic con-
residue-specific information, since each resonance in this trol experiment. (c) NOESY spectra are recorded to obtain
spectrum corresponds to an N-H bond of an amino-acid NOEs, which provide short-range spatial information (up
residue of the protein. (b) Paramagnetic relaxation to 5–6 Å). When an NOE is observed between two atoms,
enhancement (PRE) is used to define the global structure it indicates that the two atoms involved are less than
of Tau. A paramagnetic probe is covalently bond to the 5–6  Å away from each other. (d) Residual Dipolar
side-chain of a Cys residue of the protein and induces Couplings (RDCs) can be measured when an alignment
transverse relaxation of atomic nuclei spatially close to medium is added to the NMR tube, which will induce a
the probe. An HSQC is recorded on a paramagnetic and a partial preferential alignment of the protein. This method
diamagnetic sample and the intensities of each corre- provides information on orientation of individual bonds
sponding signals are compared. The resonances corre- and inter-nuclear vectors in the protein
sponding to atomic nuclei that are close in space to the

scattering data [37]. Molecular ensemble large number of pre-generated conformers, which
approaches consist in in-silico selection of an fits the experimental data better than the purely
ensemble of molecular structures, among a very random statistical coil model (Fig.  3.4). In
38 C. Danis et al.

Fig. 3.3  Heteronuclear Single Quantum Correlation 8.32  ppm, 15N 119.3  ppm), the corresponding spectrum
Spectrum or HSQC of Tau. The 1H-15N HSQC (in red) (overlayed in blue) will be modified specifically for the
provides a fingerprint of Tau in a given chemical environ- resonances corresponding to the affected amino-acid resi-
ment. Each resonance corresponds to an N-H bond and dues (see N-H phospho-Ser404 at 1H 8.60  ppm, 15N
can be linked to a specific amino acid residue in Tau 120.2 ppm). The chemical shift values (in 1H and/or 15N
sequence (see labelled resonances for Asn286, Gln336, dimension) of these resonances will be modified (CSP,
Ser404). Upon phosphorylation of a Ser/Thr residue of chemical shift perturbation, compare Ser404 and pSer404
the protein (for example N-H Ser404 resonance at 1H chemical shift values) and/or their intensities

a­ddition this approach detects several stretches mated to reach a population of as much as 60%
located in the PRR and MTBR of Tau that have a [37]. In some cases, polyproline II helices can
tendency to adopt a polyproline II helical confor- represent binding sites for a number of protein
mation, an extended structure typical of IDPs partners. Peptide 216–223 for example, adopting
[37]. Although mainly associated with proline-­ the polyproline II helical conformation (Fig. 3.1),
rich sequences, this local structure can be adopted was later found to be within the interaction site of
by other amino-acid residues, as observed in Tau the SH3 domain of the BIN1 protein [43]. Here
for a number of short sequences. In proline-rich again, a conformational property can translate to
regions, such as the sequence of residues 212– a functional aspect, although the link is not
230, the population of polyproline II helical con- always obvious. Turn conformations were
formation exceeds that found in the statistical described within the repeats, based on NMR
coil model by up to 20%, and in some regions of residual dipolar couplings (RDCs, Fig.  3.2d),
the PRR, that particular conformation is esti- which provide orientation of inter-nuclear ­vectors
3  Nuclear Magnetic Resonance Spectroscopy Insights into Tau Structure in Solution: Impact… 39

(80%) in the fibril adopts a β-strand conforma-

tion, forming an overall C-shaped fold. The
β-strands are interrupted by β-breaking prolines
Pro312, Pro332, Pro364, β-turn glycines Gly323,
Gly355 or β -arc residues Glu342 and Asp348,
the latter being part of the DKFD motif of R4.
The motif DLSK in R3 is however part of β-strand
β2. These turn tendencies observed for the MTBR
of Tau thus partially reflect the conformation of
these sequences in the fibril. However, it was pro-
posed that alternatively, their presence could pre-
vent the aggregation, as β-strand breakers [33].

 Disordered Protein: Tau’s Global

A
Fold

Another property of the disordered Tau protein

that was characterized is its global fold. Indeed,
Tau does not behave as a rod but as a flexible
polymer that adopts a preferential fold, although
devoid of a hydrophobic core and thus mainly
driven by electrostatic interactions [34]. Thus,
that fold is dynamic and conformational exchange
occurs. To characterize the global fold of a disor-
dered protein, paramagnetic probes are used
which are attached to cysteines, with Cys291 and
Cys322 being the two native cysteines of Tau.
Fig. 3.4  Principle of ensemble calculation for intrinsi- Alternatively, these native cysteines have to be
cally disordered proteins. A large number of conformers mutated into alanines and a single Cys residue
are first generated, a sub-ensemble is next selected which introduced at a point of interest in the sequence.
fits part of the experimental data better than the initial
pool of conformers. The ensemble of conformations is The paramagnetic probes will enhance relaxation
lastly validated using another subset of experimental data, of atomic nuclei in their spatial proximity, up to a
independent from the one used in the selection step distance of 25–40  Å, which allows to define
regions close in space to the paramagnetic centre
in the protein, and molecular dynamics simula- (Fig.  3.2b). Applied to Tau, the paramagnetic
tions. Four turns were found, one in each repeat, relaxation enhancement (PRE) experiments show
with the amino-acid sequences DLKN, DLSN, several contacts between regions distant in the
DLSK, and DKFD in R1 to R4 [33]. Comparison sequence. A transient contact between the
with the recent electron microscopy (EM) struc- N-terminal region and the central region (PRR
ture of aggregates of Tau, corresponding to paired and R1 to R3 repeats) is supported by PRE data
helical filaments (PHFs) or straight filaments [34]. This long-range interaction was confirmed
(SF) extracted from Alzheimer’s disease (AD) based on a reported ensemble of Tau conformers
brain tissue [14], offers a view of these short calculated from a large number of various NMR
sequences within the native protomer. R1 and R2 parameters (including PREs) combined with
repeats are not embedded in the core of the pro- small-angle X-ray scattering data [37]. The
tomer in the EM structure, composed mainly of C-terminal domain transiently contacts the
the R3 and R4 repeats. Most of the Tau sequence N-terminal domain and the repeat domain R3-R4
40 C. Danis et al.

of Tau [34]. These results converge to the reported Phosphorylation of pThr212/pThr217 has no
global fold based on FRET (fluorescence reso- such effect, showing that the stabilization effect
nance energy transfer) distances, described as a is sequence specific. The conformations of phos-
paperclip shape [21]. phorylated Tau(225–246) peptides were addi-
tionally evaluated by molecular ensemble
calculations [38] based on (i) distance restraints
I mpact of Phosphorylation on Tau obtained from NOEs (nuclear Overhauser
Structure enhancements), NMR signals arising from the
close proximity of two protons in space (Fig.
The (average) conformation of Tau, already dif- 3.2c) and (ii) orientational restraints calculated
ficult to characterize for this large disordered pro- from RDCs for N-H bonds (Fig.  3.2d). The
tein, becomes even more complex to describe due resulting ensembles confirm the transient helix
to the multiple post-translational modifications presence, which does not dependent on phos-
(PTMs) that are possible in this protein. In par- phorylation of Thr231 but rather of Ser235, and
ticular, phosphorylation of Ser and Thr residues is further stabilized by additional phosphoryla-
has attracted much interest, since AD progression tion of Ser237 and Ser238 [38]. The enhanced
is characterized by an increased level of phos- stabilization effect could be due to the formation
phorylation and, PHFs are composed of hyper- of salt-bridges between the side-chains of
phosphorylated Tau. As phosphate groups pSer237/Lys234 and pSer238/Arg242. In addi-
introduce charges and can engage in hydrogen tion, based on these models of the phospho-­
bonds, they have the potential to impact the local Tau(225–246) peptide, the distances between the
and global structure of Tau. A first step in the phosphates pThr231/pSer235 and the nitrogens
characterization of potential changes of Tau con- in the basic groups of Arg230/Lys234, respec-
formation due to phosphorylation was the ana- tively, are shorter than 4.5  Å. This distance is
lytical capacity of NMR spectroscopy to identify compatible with formation of salt-bridges
phosphorylated residues in a given sample of Tau between the side-chains of the phosphorylated
(Fig.  3.3). Initial studies were conducted with residues and the side-chains of the directly pre-
Tau protein phosphorylated in vitro by recombi- ceding basic residues Arg230 or Lys234.
nant PKA kinase, followed by identification of Interestingly, phosphorylation by CDK2/
all modified serine residues in the sample [28]. CycA3 has a functional effect, since the Tau(208–
This point is of importance since NMR is able to 324) fragment loses its capacity to polymerize
link specific phosphorylations in Tau sequence tubulin into microtubules (MTs) once phosphor-
with the corresponding structural and/or func- ylated. When full-length Tau is phosphorylated
tional output. A first attempt to link specific phos- by the CDK2/CycA3 kinase in vitro, phosphory-
phorylations with local structural impact was lation at Ser202/Ser205 and Thr231/Ser235 sites
made on a Tau fragment phosphorylated in vitro are identified by NMR (and weak phosphoryla-
with recombinant CDK2/CycA3 kinase [39]. tion at Thr212/Thr217). These phosphorylations
This proline-directed kinase was found to modify do not significantly affect binding to MTs [2].
the Tau(208–324) fragment, containing the PRR Nevertheless, when at least three phosphates are
and part of the MTBR (R1-R2 and part of R3), on present in these four positions, Tau loses its
several sites within the PRR, corresponding to capacity to assemble tubulin into MTs. Additional
residues T231/S235 (the epitope of the AT180 experiments, using the shortest Tau isoform (Tau
antibody) as well as Thr212 and Thr217. The 0N3R) with residues Thr231 and Ser235 mutated
CDK-phosphorylated fragment remains globally to glutamate residues as phosphorylation mimic,
disordered, but a local structuration can be confirm that Glu231/Glu235 by themselves do
detected, corresponding to a helical tendency for not abolish the interaction of Tau with MTs [38].
about 10 amino-acid residues at the C-terminus However, NMR signals corresponding to resi-
of the pThr231/pSer235 phosphorylation sites. dues in the PRR were less attenuated upon
3  Nuclear Magnetic Resonance Spectroscopy Insights into Tau Structure in Solution: Impact… 41

a­ ddition of MTs to the mutated Tau 0N3R than molecular detail of the interaction of AT8 anti-
for the wild-type [38]. This might indicate that body with its epitope, revealed by a crystallo-
pseudophosphorylated Tau 0N3R was locally graphic study, shows that AT8 best accommodates
less tightly bound to the MTs. The salt-bridge a triply-phosphorylated sequence [32]. The link
between pThr231 and Arg230 side-chains is pro- between the dynamic turn conformation and the
posed to compete with salt-bridge formation with aggregation propensity of Tau is further sup-
MTs, participating in the effect of phospho- ported by a mutated Tau protein, with a Gly207
Thr231 [38]. A direct link between the conforma- residue replaced by valine. This Tau variant
tional modification and its functional exhibits no turn conformation around the
consequences is difficult to establish. However, pSer202/pThr205 sites and displays higher
this study proved that the effect of Tau phosphor- aggregation propensity in in vitro aggregation
ylation is not limited to the introduction of nega- assays (without heparin inducer). In this particu-
tive charges that may influence its interactions, lar case, the link between a small dynamic struc-
but that it additionally influences the local struc- tural motif, and a (dys)function of Tau can be
ture of Tau, with potential functional made. Numerous questions remain as to the
significance. mechanism of aggregation protection by the
Along the same lines, phosphorylation of resi- pSer202/pThr205 centred-turn, or whether there
dues Ser202/Thr205, associated with the AT8 might be other phosphorylation patterns with
epitope, was shown to induce formation of a local such properties.
dynamic turn, based on a combination of NMR Other studies used glutamic acid mutations to
data and molecular dynamics simulations [15]. mimic phosphorylation. This method of pseudo-­
This combination of phosphorylations is of major phosphorylation may not be ideal to study the
interest since the AT8 monoclonal antibody is impact of phosphorylation on Tau structure,
described as targeting the pathological AD-like since, for example, the hydrogen-bonding pattern
state of Tau [4] and immunocytochemistry of of a phosphate group is not reproduced perfectly.
brain slices using the AT8 allows to evaluate the However, a key advantage of glutamic acid muta-
stages of the disease [8]. The NMR data support- tions is that they result in a homogeneous sample,
ing the formation of a turn conformation consist as opposed to in vitro phosphorylation. Pseudo-­
in NOE contacts, for example between the amide phosphorylation might thus turn out to be a nec-
proton of Gly207 and the Hα, Hβ and Hɣ of essary evil in some cases. For example, glutamic
pThr205. The phosphate group of pThr205 con- acid mutation of the serine residues in the KxGS
tributes to stabilization of the dynamic turn con- motifs found in repeats R1 to R4 of Tau showed a
formation by engaging in a hydrogen bond with selective conformational effect in repeats R1-R2,
the amide proton of Gly207. In the 2D 1H, 15N based on N-H residual dipolar coupling experi-
spectrum of phosphorylated Tau, with pSer202/ ments (Fig. 3.2d). The sign of the RDCs changes
pThr205 detected, it results in an easily detect- for residues 265
NLK267 upon pseudo-­
able downfield shift of Gly207 resonance in the phosphorylation of Ser262, suggesting formation
1
H dimension. Salt-bridges between the pThr205 of a turn-like structure [13]. Since phosphoryla-
phosphate and the side-chain of Arg209, as well tion of Ser262 is described as protective against
as between the pSer202 phosphate and the side-­ aggregation [36], it is proposed that this turn-­
chain of Arg211 contribute to an additional stabi- conformation favours interaction of the
lization of the turn. An additional phosphorylation 259
KIGpS262 motif (in R1) with the end of R1 and
on Ser208, a non-proline directed phosphoryla- the beginning of R2, impacting the capacity of
tion site, disrupts turn formation. Of interest is the hexapeptide 275VQIINK280 in R2 for intermo-
that the Tau protein is more sensitive to aggrega- lecular interaction. Pseudophosphorylations at
tion without this turn conformation [11]. This the AT8 (in this study, mutations of Ser199,
might thus be the form recognized by the AT8 Ser202 and Thr205 into glutamic acid), AT100
antibody in immunocytochemistry, since the (Thr212Glu and Ser214Glu), and PHF1
42 C. Danis et al.

(Ser396Glu and Ser404Glu) epitopes were used

to evaluate the impact of phosphorylation on
global Tau conformation based on PREs com-
bined with an ensemble approach. The data show
a reduction of long-range electrostatic interac-
tions between the N-terminal region and the PRR
of Tau [3]. However, based on distance measure-
ments from FRET pairs, a Tau protein presenting
the same pseudo-phosphorylation mutations was
proposed to be more compact than the wild-type,
with increased contacts between the N- and
C-terminal regions, thus better reproducing the
conformation recognized by the conformational
antibody MC1 that targets AD-Tau [22]. This
apparent contradiction from these very thorough
studies serves to emphasize the challenge to cap-
ture the elusive conformational changes in a pro-
tein as flexible as Tau.
Other Tau PTMs have been studied by NMR
[23], although their impact on Tau structure
Fig. 3.5  cis/trans conformations of pSer-Pro bond
remains to date to be explored. These PTMs play Stick representation of a pSer-Pro dipeptide in trans and
roles in cross-talk with phosphorylation [7, 42]. cis conformations. Spontaneous conformational exchange
is slow. Detail of the 1H, 15N HSQC of CDK2/CycA3
phosphorylated Tau PRR (in gray), with phosphorylated
Ser202 detected, showing the resonance of N-H Gly204.
Tau prolyl cis/trans isomerization The overlayed EXSY (exchange spectroscopy) spectrum
in presence of Pin1 (in red, ratio PRR:Pin1 1:0.1) allows
Tau prolyl cis/trans isomerization is mediated by to detect exchange between cis and trans conformations
prolyl cis/trans isomerases (PPIases) that accel- of the pSer202-Pro203 bond. This exchange, accelerated
by the presence of Pin1, can be detected on the NMR time
erate the cis/trans conversion of the peptidic scale as an exchange peak (labelled c to t and t to c)
bond preceding a proline residue (Fig.  3.5). between the two resonances of Gly204, used as reporters
Prolyl cis/trans isomerases such as Pin1 and for the cis and trans pSer202-Pro203 bond
FKBP52 have been pointed out as being involved conformations
in the Tau pathway in AD [10, 16, 29, 31]. In par-
ticular, a cis-phospho Tau form, detected by a cis-­ phorylation sites [31, 41]. PPIase activity and
specific antibody, was proposed to represent an cis/trans conformation of proline residues in the
early pathogenic form of Tau in AD [26, 35], with substrate can both be characterized by NMR: the
the pThr231-Pro232 prolyl bond in the cis con- isoforms are identified by distinct sets of reso-
formation in AD neurons. This cis-phospho Tau nances for Xxx-Pro residue pairs in cis or trans
cannot promote MTs assembly, is more resistant conformations and, the activity is detected as a
to dephosphorylation and degradation, and more chemical exchange process during the mixing
prone to aggregation [26, 35]. These results have period of the NMR experiment, once the enzyme
been difficult to reconcile with molecular studies is added to accelerate the process (Fig.  3.5).
on Tau structure. FKBP52 has been shown to act PPiase activity of Pin1 is detected for a number
on Tau by interaction with its MTBR, rather than of phosphorylated Tau sites (Fig. 3.5), but not for
by its PPIase activity [24]. Pin1, on the other pThr231-Pro232 [12, 41]. Additionally, when
hand, is a particularly interesting PPIase, as it investigating the conformation of each Xxx-Pro
specifically recognizes phosphorylated Tau pro- bond in the fragment Tau(208–324), only a few
tein, and potentially its 17 pSer/pThr-Pro phos- were found with a detectable cis form, with each
3  Nuclear Magnetic Resonance Spectroscopy Insights into Tau Structure in Solution: Impact… 43

Xxx-Pro bond for over 90% in the trans confor- at AT8, AT100, and PHF1 epitopes on 441-residue
tau. J Am Chem Soc. 2011;133:15842–5. https://doi.
mation [1]. Phosphorylation by CDK2/CycA3, org/10.1021/ja205836j.
which phosphorylates four Ser/Thr-Pro motifs in 4. Biernat J, Mandelkow EM, Schroter C, et  al. The
Tau(208–324), does not change this conclusion, switch of tau protein to an Alzheimer-like state
with no preferential cis conformation for the includes the phosphorylation of two serine-proline
motifs upstream of the microtubule binding region.
pThr231-Pro232 bond. From a functional point EMBO J. 1992;11:1593–7.
of view, Pin1 was proposed to restore the capac- 5. Binder LI, Frankfurter A, Rebhun LI.  The distribu-
ity of phosphorylated Tau to bind to MTs and tion of tau in the mammalian central nervous system.
restore MTs assembly [31]. This model was how- J Cell Biol. 1985;101:1371–8.
6. Black MM, Slaughter T, Moshiach S, et  al. Tau is
ever also recently challenged, since Pin1 does not enriched on dynamic microtubules in the distal region
promote in vitro formation of MTs by phosphor- of growing axons. J Neurosci. 1996;16:3601–19.
ylated Tau [27]. Although, an involvement of 7. Bourré G, Cantrelle FX, Kamah A, et  al. Direct
Pin1 in AD is not ruled out, the molecular mecha- Crosstalk Between O-GlcNAcylation and
Phosphorylation of Tau Protein Investigated by
nisms sustaining its role, and in particular its NMR Spectroscopy. Front Endocrinol. 2018; 9:595.
PPiase activity, remain currently elusive. https://doi.org/10.3389/fendo.2018.00595.
8. Braak H, Alafuzoff I, Arzberger T, et  al. Staging of
Alzheimer disease-associated neurofibrillary pathol-
ogy using paraffin sections and immunocytochemis-
Conclusions try. Acta Neuropathol. 2006;112:389–404.
9. Butner KA, Kirschner MW.  Tau protein binds to
Although Tau is an intrinsically disordered pro- microtubules through a flexible array of distributed
tein, best characterized by an ensemble of weak sites. J Cell Biol. 1991;115:717–30.
10. Chambraud B, Sardin E, Giustiniani J, et al. A role for
dynamic conformations, some elements of struc- FKBP52 in Tau protein function. Proc Natl Acad Sci
turation can be defined. Given the dynamic char- U S A. 2010;107:2658–63. https://doi.org/10.1073/
acter of disordered proteins, biophysical pnas.0914957107.
techniques combined with computational meth- 11. Despres C, Byrne C, Qi H, et al. Identification of the
Tau phosphorylation pattern that drives its aggrega-
ods have to be used to describe their subtle local tion. Proc Natl Acad Sci U S A. 2017;114:9080–5.
and global preferential conformations. These https://doi.org/10.1073/pnas.1708448114.
characteristic elements of structuration can be 12. Eichner T, Kutter S, Labeikovsky W, et al. Molecular
influenced by post-translational modifications. Mechanism of Pin1-Tau Recognition and Catalysis.
J  Mol Biol. 2016;428:1760–75. https://doi.
The current challenge remains to link this struc- org/10.1016/j.jmb.2016.03.009.
turation, or its perturbation, to functional conse- 13.
Fischer D, Mukrasch MD, Biernat J, et  al.
quences. Much can be gained from such a Conformational changes specific for pseudophosphor-
characterization that may provide a definition for ylation at serine 262 selectively impair binding of tau
to microtubules. Biochemistry. 2009;48:10047–55.
a misfolded Tau protein, which seems at first 14. Fitzpatrick AWP, Falcon B, He S, et  al. Cryo-EM
glance antinomic to its disordered nature. structures of tau filaments from Alzheimer’s disease.
Nature. 2017;547:185–90. https://doi.org/10.1038/
nature23002.
15.
Gandhi NS, Landrieu I, Byrne C, et  al. A
References phosphorylation-­induced turn defines the Alzheimer’s
disease AT8 antibody epitope on the Tau protein.
1. Ahuja P, Cantrelle F-X, Huvent I, et  al. Proline Angew Chem Int Ed Engl. 2015;54:6819–23. https://
Conformation in a Functional Tau Fragment. J  Mol doi.org/10.1002/anie.201501898.
Biol. 2016;428:79–91. https://doi.org/10.1016/j. 16. Giustiniani J, Guillemeau K, Dounane O, et  al. The
jmb.2015.11.023. FK506-binding protein FKBP52 in  vitro induces
2. Amniai L, Barbier P, Sillen A, et  al. Alzheimer dis- aggregation of truncated Tau forms with prion-­
ease specific phosphoepitopes of Tau interfere with like behavior. FASEB J  Off Publ Fed Am Soc Exp
assembly of tubulin but not binding to microtubules. Biol. 2015;29:3171–81. https://doi.org/10.1096/
FASEB J. 2009;23:1146–52. https://doi.org/10.1096/ fj.14-268243.
fj.08-121590. 17. Goedert M, Spillantini MG, Jakes R, et  al. Multiple
3. Bibow S, Ozenne V, Biernat J, et  al. Structural isoforms of human microtubule-associated protein
impact of proline-directed pseudophosphorylation tau: sequences and localization in neurofibrillary
44 C. Danis et al.

t­angles of Alzheimer’s disease. Neuron. 1989;3:519– 31. Lu PJ, Wulf G, Zhou XZ, et al. The prolyl isomerase
26.. doi: 0896-6273(89)90210-9 [pii] Pin1 restores the function of Alzheimer-associated
18. Goode BL, Feinstein SC.  Identification of a novel phosphorylated tau protein. Nature. 1999;399:784–8.
microtubule binding and assembly domain in the 32. Malia TJ, Teplyakov A, Ernst R, et al. Epitope map-
developmentally regulated inter-repeat region of tau. ping and structural basis for the recognition of phos-
J Cell Biol. 1994;124:769–82. phorylated tau by the anti-tau antibody AT8. Proteins.
19. Himmler A, Drechsel D, Kirschner MW, Martin DW 2016;84:427–34. https://doi.org/10.1002/prot.24988.
Jr. Tau consists of a set of proteins with repeated 33. Mukrasch MD, Markwick P, Biernat J, et al. Highly
C-terminal microtubule-binding domains and variable populated turn conformations in natively unfolded
N-terminal domains. Mol Cell Biol. 1989;9:1381–8. tau protein identified from residual dipolar cou-
20. Hoover BR, Reed MN, Su J, et  al. Tau mislocal- plings and molecular simulation. J  Am Chem Soc.
ization to dendritic spines mediates synaptic dys- 2007;129:5235–43.
function independently of neurodegeneration. 34. Mukrasch MD, Bibow S, Korukottu J, et al. Structural
Neuron. 2010;68:1067–81. https://doi.org/10.1016/j. polymorphism of 441-residue tau at single resi-
neuron.2010.11.030. due resolution. PLoS Biol. 2009;7:e34. https://doi.
21. Jeganathan S, von Bergen M, Brutlach H, et al. Global org/10.1371/journal.pbio.1000034.
hairpin folding of tau in solution. Biochemistry. 35. Nakamura K, Greenwood A, Binder L, et al. Proline
2006;45:2283–93. https://doi.org/10.1021/ isomer-specific antibodies reveal the early patho-
bi0521543. genic tau conformation in Alzheimer’s disease.
22. Jeganathan S, Hascher A, Chinnathambi S, et  al.
Cell. 2012;149:232–44. https://doi.org/10.1016/j.
Proline-directed pseudo-phosphorylation at AT8 and cell.2012.02.016.
PHF1 epitopes induces a compaction of the paperclip 36. Schneider A, Biernat J, von Bergen M, et  al.

folding of Tau and generates a pathological (MC-1) Phosphorylation that detaches tau protein from
conformation. J  Biol Chem. 2008;283:32066–76. microtubules (Ser262, Ser214) also protects it against
https://doi.org/10.1074/jbc.M805300200. aggregation into Alzheimer paired helical filaments.
23. Kamah A, Huvent I, Cantrelle F-X, et  al. Nuclear Biochemistry. 1999;38:3549–58.
magnetic resonance analysis of the acetylation pat- 37. Schwalbe M, Ozenne V, Bibow S, et  al. Predictive
tern of the neuronal Tau protein. Biochemistry. atomic resolution descriptions of intrinsically dis-
2014;53:3020–32. https://doi.org/10.1021/ ordered hTau40 and alpha-synuclein in solution
bi500006v. from NMR and small angle scattering. Structure.
24. Kamah A, Cantrelle FX, Huvent I, et al. Isomerization 2014;22:238–49. https://doi.org/10.1016/j.
and Oligomerization of Truncated and Mutated Tau str.2013.10.020.
Forms by FKBP52 are Independent Processes. J Mol 38. Schwalbe M, Kadavath H, Biernat J, et al. Structural
Biol. 2016;428:1080–90. https://doi.org/10.1016/j. Impact of Tau Phosphorylation at Threonine
jmb.2016.02.015. 231. Structure. 2015;23:1448–58. https://doi.
25.
Katsinelos T, Zeitler M, Dimou E, et  al. org/10.1016/j.str.2015.06.002.
Unconventional Secretion Mediates the Trans-cellular 39. Sibille N, Huvent I, Fauquant C, et al. Structural char-
Spreading of Tau. Cell Rep. 2018;23:2039–55. https:// acterization by nuclear magnetic resonance of the
doi.org/10.1016/j.celrep.2018.04.056. impact of phosphorylation in the proline-rich region
26.
Kondo A, Shahpasand K, Mannix R, et  al. of the disordered Tau protein. Proteins. 2012;80:454–
Antibody against early driver of neurodegenera- 62. https://doi.org/10.1002/prot.23210.
tion cis P-tau blocks brain injury and tauopathy. 40. Smet C, Leroy A, Sillen A, et al. Accepting its random
Nature. 2015;523:431–6. https://doi.org/10.1038/ coil nature allows a partial NMR assignment of the
nature14658. neuronal Tau protein. Chembiochem. 2004a;5:1639–
27.
Kutter S, Eichner T, Deaconescu AM, Kern 46. https://doi.org/10.1002/cbic.200400145.
D.  Regulation of Microtubule Assembly by Tau and 41. Smet C, Sambo AV, Wieruszeski JM, et al. The pep-
not by Pin1. J Mol Biol. 2016;428:1742–59. https:// tidyl prolyl cis/trans-isomerase Pin1 recognizes the
doi.org/10.1016/j.jmb.2016.03.010. phospho-Thr212-Pro213 site on Tau. Biochemistry.
28. Landrieu I, Lacosse L, Leroy A, et  al. NMR analy- 2004b;43:2032–40. https://doi.org/10.1021/
sis of a Tau phosphorylation pattern. J  Am Chem bi035479x.
Soc. 2006;128:3575–83. https://doi.org/10.1021/ 42. Smet-Nocca C, Broncel M, Wieruszeski JM, et  al.
ja054656+. Identification of O-GlcNAc sites within peptides of
29. Liou YC, Sun A, Ryo A, et al. Role of the prolyl isom- the Tau protein and their impact on phosphorylation.
erase Pin1 in protecting against age-dependent neuro- Mol BioSyst. 2011;7:1420–9. https://doi.org/10.1039/
degeneration. Nature. 2003;424:556–61. c0mb00337a.
30. Lippens G, Wieruszeski JM, Leroy A, et al. Proline-­ 43. Sottejeau Y, Bretteville A, Cantrelle F-X, et  al. Tau
directed random-coil chemical shift values as a tool phosphorylation regulates the interaction between
for the NMR assignment of the tau phosphoryla- BIN1’s SH3 domain and Tau’s proline-rich domain.
tion sites. Chembiochem. 2004;5:73–8. https://doi. Acta Neuropathol Commun. 2015;3:58. https://doi.
org/10.1002/cbic.200300763. org/10.1186/s40478-015-0237-8.
3  Nuclear Magnetic Resonance Spectroscopy Insights into Tau Structure in Solution: Impact… 45

44. Sultan A, Nesslany F, Violet M, et al. Nuclear tau, a 46. Wang Y, Balaji V, Kaniyappan S, et al. The release and
key player in neuronal DNA protection. J Biol Chem. trans-synaptic transmission of Tau via exosomes. Mol
2011;286:4566–75. https://doi.org/10.1074/jbc. Neurodegener. 2017;12:5. https://doi.org/10.1186/
M110.199976. s13024-016-0143-y.
45. von Bergen M, Friedhoff P, Biernat J, et al. Assembly 47.
Weingarten MD, Lockwood AH, Hwo SY,
of tau protein into Alzheimer paired helical fila- Kirschner MW.  A protein factor essential for
ments depends on a local sequence motif ((306) microtubule assembly. Proc Natl Acad Sci U S A.
VQIVYK(311)) forming beta structure. Proc Natl Acad 1975;72:1858–62.
Sci U S A. 2000;97:5129–34.. doi: 97/10/5129 [pii]
 Regulation of Tau Homeostasis
and Toxicity by Acetylation 4
Tara Tracy, Kathryn C. Claiborn, and Li Gan

Introduction death; studies of tauopathy mouse models indi-

cate that toxic tau affects postsynaptic function
Tauopathies, including Alzheimer’s disease (AD), [35, 39, 41, 76, 80, 88] and triggers neuronal dys-
are neurodegenerative diseases characterized by function underlying cognitive impairments with-
aggregation of the microtubule-binding protein out widespread neuron loss [20, 28, 82, 87].
tau in the brain, which coincides with synapse However, the molecular alterations to the tau pro-
and neuronal loss [25, 38, 51]. While the majority tein that trigger mis-localization and the forma-
of tauopathy cases are sporadic, there are over 50 tion of aggregates are incompletely understood.
familial tau mutations that cause frontotemporal Here, we describe recent research highlighting
lobar degeneration with tau inclusions (FTLD- the role that post-translational modifications
tau), and the field has been shaped by studies of (PTMs), particularly acetylation, may play in
transgenic mice that express these human tau establishing the toxicity of tau, and how this may
mutations [17, 35, 53, 68, 78, 84, 87, 89]. shape future therapeutic efforts.
Tau is normally localized to the axons, and the
current prevailing view is that the mis-­localization
or secretion of aberrant forms is toxic to neurons Posttranslational Modifications
[64]. Indeed, in Alzheimer’s disease (AD), tau of Tau
accumulates in dendrites, a process associated
with accumulation of amyloid β [42]. Importantly, The nascent tau peptide is subject to extensive
tau accumulation may not always result in cell modification, including proteolytic cleavage,
phosphorylation [55], ubiquitination [14, 62],
glycosylation, methylation [23, 79], and acetyla-
tion [10, 60, 63, 80]. How these differently modi-
fied forms contribute to the pathogenesis of
T. Tracy · K. C. Claiborn tauopathies is an area of intense investigation.
Gladstone Institutes, San Francisco, CA, USA Phosphorylated tau was first described in tau iso-
L. Gan (*) lated from the brains of AD patients [32]; it has
Gladstone Institutes, San Francisco, CA, USA subsequently been shown that neurofibrillary
Helen and Robert Alzheimer’s Disease Research tangles contain highly phosphorylated tau, and
Institute, Feil Family Brain and Mind Research that these form in parallel with the manifestation
Institute, Weill Cornell Medicine, of cognitive impairments in AD [2, 29]. In sup-
New York, NY, USA port of a pathogenic role of hyperphosphoryla-
e-mail: li.gan@gladstone.ucsf.edu

© Springer Nature Singapore Pte Ltd. 2019 47

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_4
48 T. Tracy et al.

tion of the protein, expression of a tau mutant that tein [60], inhibits tau microtubule binding [10,
mimics the phosphorylation of 14 residues in 74], and promotes aggregation [10, 59]. To inves-
cultured hippocampal neurons resulted mis-­
­ tigate the role of specific acetylated residues,
localization of the protein to dendritic spines, and lysine-  >  glutamine (KQ) substitutions can be
reduced the number of AMPA-type glutamate made to mimic the structure and charge of acety-
receptors (AMPAR) at synapses [35]. Cleaved lated lysine [34, 49, 70, 83]. This approach has
forms of tau may also contribute to synaptoxicity, helped reveal the role of acetylated tau in neuro-
as blocking caspase-2-mediated cleavage of the degeneration [59], cytoskeletal dysregulation
protein restores synaptic transmission and res- [74], and synaptic dysfunction [80] related to AD
cues memory deficits in mice expressing the pathogenesis.
P301L tau mutation [90].
Acetylation of the N-terminus of tau was
described in early studies of the protein, but as Tau Acetylation
this modification was observed on protein iso- in Neurodegenerative Disease
lated from both diseased and healthy brains, it
was initially not considered a pathogenic feature More than 20 lysines of tau are subject to acetyla-
of tauopathies [32]. More than a decade later, it tion [60], of which several appear to be of par-
was discovered that tau is acetylated on various ticular pathological significance. Mass
lysine residues, and that tau acetylation was ele- spectrometry identified acetylation of tau K174
vated in the brains of patients with tauopathy as an early change in AD brains [59], and acetyla-
[60], a finding that has since been independently tion at two additional residues, K274 and K281,
validated in multiple studies of human cohorts as is found in the brains of patients who exhibit
well as mouse models of neurodegeneration [10, memory loss [80] (Fig. 4.1). Acetylation at K274
27, 59, 74, 80]. As described in more detail below, and K281 is also elevated in the brains of the
mechanistic investigations have revealed that hAPP-J20 mice [80], a model of AD that
acetylation of tau blocks degradation of the pro- expresses high levels of hAPP and amyloid beta

Fig. 4.1  Acetylated tau is enriched in tauopathy plasmatic neuronal inclusion (FTLP-17 – case 11). h glob-
brains. (a) High magnification view of acetylated-tau ular glial inclusions (White matter tauopathy with globular
inclusions located in inferior temporal cortex (a–c and e– glial inclusion – case 22). i glial inclusion (atypical tauop-
i) and midbrain (d), after immunostaining with an anti-ac-­ athy  – case 7). Scale bar represents 10  μm. Originally
tau antibody. a neuritic plaque (Alzheimer’s disease – case described in Grinberg et al. Acta Neuropathol. 2013 (35).
3). b neurofibrillary tangle (Alzheimer’s disease – case 3). (b). Levels of ac-K174 were significantly higher at early
c astrocytic plaque (corticobasal degeneration – case 9). d and late Braak stages than at Braak stage 0; n = 7 (Braak
globous tangles in oculormotor nucleus (progressive 0), n = 13 (Braak 1–2), n = 8 (Braak 3–5). ∗p < 0.05, one-­
supranuclear palsy – case 21). e Pick’s body (Pick’s dis- way ANOVA, Tukey-Kramer post hoc analyses. Originally
ease  – case 18). f intracytoplasmatic neuronal inclusion described in Min et al. Nat. Med. 2015 (34)
(chronic traumatic encephalopathy – case 8). g intracyto-
4  Regulation of Tau Homeostasis and Toxicity by Acetylation 49

(Aβ). A link between Aβ and tau acetylation in K274 occurs later in disease progression.
neurodegeneration is also supported by the Alternative splicing of the tau transcript also
­finding that treating cultured neurons with Aβ affects the acetylation status; while the full-­
oligomers increases tau acetylation [60]. A more length protein has four microtubule-binding
recent study showed that site-specific acetylation repeats (4R), the alternatively spliced version has
at K280, which was found only in AD brains only three (3R), and lacks the K281 site.
[10], significantly enhances the aggregation rate Neurofibrillary tangles in AD brains contain both
of tau and impairs microtubule assembly [30]. 3R and 4R tau [18], indicating the potential that
Hyperacetylation of tau in disease may be acetylation at K274 and K281 contribute to
related to modulation of the acetyltransferase pathology independently.
p300, which is upregulated in the brains of AD
patients [3, 85] and in neuronal models of AD
[50]. Tau also has intrinsic auto-acetylation capa-  he Effect of Acetylation on Toxic
T
bility [9], which has been linked to subsequent Tau Species
proteolytic cleavage and the production of tau
fragments [8]. Tau deacetylation is believed to be The mechanism by which tau acetylation induces
mediated largely by sirtuin 1 [60, 61]. There is pathogenesis may involve regulation of toxic
also evidence that HDAC6 can deacetylate tau, forms of the protein, such as hyperphosphory-
although the target lysines may be different from lated protein. However, the relationship between
those regulated by sirtuin 1 [7, 12]. The levels of tau acetylation and phosphorylation may be both
sirtuin 1 decrease in the brain during AD progres- residue- and context-specific. In transgenic mice
sion [40, 52], and treating cultured neurons with expressing an acetylation mimetic version of
Aβ reduces the expression of sirtuin 1 [46], per- human tau K274 and K281 (tauKQ), the levels of
petuating accumulation of acetylated tau. Taken phosphorylation at serine 202 are reduced, while
together, these findings indicate that enhanced phosphorylation at other residues associated with
acetylation and reduced deacetylation likely both tau pathology are unchanged [74, 80]. Expression
contribute to the pathological elevation of acety- of K174Q in mice has no effect on tau phosphor-
lated tau. ylation [59], while acetylation at amino acids
The regulation and functional consequence of 259/353 appears to block phosphorylation [12].
each acetylation site is a major point of interest in Interestingly, in a Drosophila model, expression
the field. Notably, K274 and K281 both occur in of the acetyl-mimetic K280Q resulted in
the microtubule-binding domain of tau, where increased levels of tau phosphorylation [26].
many of the familial FTLD-tau mutations are Another potential mechanism by which tau
located, and acetylation of K274 and K281 can acetylation could modulate tau toxicity is by
reduce the interaction between tau and microtu- altering the formation of caspase-cleaved tau
bules [10, 74]. Acetylation at K274 is present in fragments, which have been implicated in patho-
neurofibrillary tangles and neurotic plaques in logical tau accumulation in the brain [24].
AD, as well as many other tauopathies [27, 74]. However, tauKQ mice, which do display features
Acetylation at K274 is also enhanced in late stage of tauopathy, lack tau fragments [80], indicating
AD (Braak stages 5–6), compared to early onset that acetylation at K274 and K281 is sufficient to
disease (Braak stages 0–2) [74]. In contrast, drive synaptic and cognitive deficits without cre-
although acetylation of K281 is enhanced in the ating tau fragments.
brains of patients with mild dementia compared Tau acetylation may also affect the formation
to non-demented cases [80], modification of this of tau oligomers and aggregates, which are
residue not appear to correlate with Braak stage thought to play a role in the pathogenesis of AD
[74]. One interpretation of these findings is that [45, 54] and which are sufficient to impair
acetylation of K281 plays a role in the early memory and cause synaptic defects in mice
stages of cognitive decline, while acetylation of [44]. Acetylation of tau lysines blocks those
50 T. Tracy et al.

residues from being targeted for ubiquitination, and this secretion is elevated in the setting of AD
slowing the rate of protein turnover and leading [75]. It is possible that acetylated tau is preferen-
to ­accumulation [60]. Pathological tau, recog- tially secreted, which would contribute to the
nized by the conformation-specific MC1 anti- propagation of tau at synapses.
body, is detected in the brains of tauKQ mice Tau interacts directly with filamentous actin
[80], supporting a role for acetylation in the for- (F-actin), and can induce the organization of
mation of this species. MC1 positive tau has actin filaments [15, 22]. Synaptic activity can
been implicated in the spread of tau between promote microtubule polymerization from the
neurons in aged transgenic mice [48], and its dendrite into F-actin rich spines [58, 73], which
localization in both presynaptic and postsynap- affect the strength of glutamatergic transmission
tic compartments may indicate that tau oligo- [1]. This reorganization is required for the recruit-
mers can be spread across synapses [77]. ment of AMPARs and the maintenance of long-­
Whether acetylated tau propagates from cell to term potentiation [21, 43]. However, acetylated
cell in the brain is unknown, and understanding tau seems to have an antagonistic effect on the
if this property underlies its toxicity is an area cytoskeleton. Acetylation of K281 impairs micro-
of active research. tubule stabilization [81], and it has been shown
that acetylation at an additional site, K321, also
inhibits filament formation, at least in part by pre-
 he Effect of Acetylation on Tau
T venting phosphorylation of the neighboring resi-
Localization and Synaptic Plasticity due S324 [7]. In tauKQ mice, the interaction
between tau and F-actin is weakened, and
Although tau is normally localized to axons, activity-­dependent polymerization of actin is
acetylated tau appears in the somatodendritic impaired [80]. However, in this model, neither
compartment [74], raising the question of where basal actin polymerization nor glutamatergic
the acetylation takes place and how acetylated transmission are perturbed [80]. Furthermore,
tau translocates. The acetyltransferase p300 is other proteins involved in the regulation of
primarily active in the nucleus, and tau has also AMPARs during plasticity [36] may be affected
been detected in that compartment [47] and thus by interaction with acetylated tau. One of these is
may be acetylated there, although p300 has also the postsynaptic protein KIBRA, which in
been shown to act in the cytoplasm [71, 72]. The humans has been linked to memory performance
cytoskeletal network within the axon initial seg- [65] and to the risk for late-onset AD [6, 13, 67].
ment (AIS) is thought to play a role in partition- In the tauKQ model, the level of KIBRA in spines
ing the cell and retaining tau; interestingly, tau was reduced, but overexpression of KIBRA in
acetylation at K274 and K281 results in destabi- tauKQ hippocampal neurons restored the recruit-
lization of the AIS [74], which likely contributes ment of AMPARs to synapses [80]. Taken
to the translocation of the protein from the axons. together, these studies provide strong support for
In extracts from the brains of AD patients, tau a model in which acetylated tau disrupts F-actin
acetylated at K274 and K281 is found in the post- organization to specifically disrupt synaptic
synaptic fraction [80], raising the possibility that plasticity.
acetylated tau is transmitted from presynaptic to Notably, the effects of tau phosphorylation on
postsynaptic sites. Evidence from transgenic mod- actin filaments are quite distinct from the effects
els seems to confirm that tau can be propagated of acetylation. In a Drosophila model, phospho-­
from entorhinal cortical neurons into the hippo- mimetic mutations in tau increase F-actin assem-
campus [16, 31, 48], recapitulating the spread of bly [22]. This may be related in part to differential
neurofibrillary tangles during human AD progres- localization: while acetylated tau is present in
sion [5]. In addition, neurons secrete tau in dendritic spines [80], phosphorylated tau inter-
response to enhanced network activity [66, 86], acts with actin bundles in the soma [22].
4  Regulation of Tau Homeostasis and Toxicity by Acetylation 51

 cetylated Tau Promotes Memory

A acetyltransferase activity, blocks tau acetylation
Loss and restores memory deficits in FTLD-tau mice
[59]. Remarkably, salsalate did not protect
Patients in the early stages of AD display impair- against neurodegeneration in mice expressing
ments in episodic memory [4, 11]. A crucial K174Q [59], strong evidence that the effect of the
aspect of forming episodic memories is the abil- drug in that model is mediated by blocking acety-
ity to distinguish between similar experiences, an lation of tau. In a separate study, treatment with
event that is thought to occur in the granule cells the related drug sodium salicylate restored
of the dentate gyrus, a part of the hippocampus Aβ-induced synaptic and memory deficits in rats
[57]. In tauKQ mice, long-term potentiation is [19]. Furthermore, salsalate has anti-­inflammatory
inhibited in granule cells, and the mice exhibit effects [37], activates adenosine monophosphate-­
difficulty in differentiating between experiences activated protein kinase (AMPK), and increases
[80], indicating that tau acetylation may block autophagy [33], and thus may have multiple ben-
episodic memory formation. Tau acetylation has eficial effects in neurodegenerative disease.
also been implicated in disrupting the formation Further exploring the potential of targeting this
of spatial memories. For example, tauKQ mice mechanism to treat tauopathies is an exciting area
are able to learn the location of a hidden platform for future research.
in the Morris water maze, but then cannot retain
that memory [80]. A similar deficit is observed in
mice expressing K174Q tau [59]. Taken together, References
these findings support a model in which the tau
acetylation at multiple residues impairs the 1. Allison DW, Gelfand VI, Spector I, Craig AM. Role
of actin in anchoring postsynaptic receptors in cul-
encoding of memories in the hippocampus. tured hippocampal neurons: differential attachment of
NMDA versus AMPA receptors. J Neurosci Off J Soc
Neurosci. 1998;18:2423–36.
Targeting Tau Acetylation 2. Arriagada PV, Growdon JH, Hedley-Whyte ET,
Hyman BT.  Neurofibrillary tangles but not senile
as Potential Therapeutic Strategy plaques parallel duration and severity of Alzheimer’s
disease. Neurology. 1992;42:631–9.
Given the abundant evidence that acetylation of 3. Aubry S, Shin W, Crary JF, Lefort R, Qureshi YH,
tau is a critical element of the pathogenesis of Lefebvre C, Califano A, Shelanski ML. Assembly and
interrogation of Alzheimer’s disease genetic networks
tauopathies, strategies to reduce tau acetylation reveal novel regulators of progression. PLoS One.
represent an attractive approach to combat these 2015;10:e0120352.
devastating diseases. Among the potential targets 4. Bondi MW, Jak AJ, Delano-Wood L, Jacobson MW,
is sirtuin 1, which deacetylates tau in vivo [60]. Delis DC, Salmon DP. Neuropsychological contribu-
tions to the early identification of Alzheimer’s dis-
Indeed, overexpression of sirtuin 1 reduces the ease. Neuropsychol Rev. 2008;18:73–90.
spread of pathogenic tau in a mouse model of 5. Braak H, Braak E.  Neuropathological stageing of
tauopathy [61]. Sirtuin 1 activators, therefore, Alzheimer-related changes. Acta Neuropathol.
could be a promising method to reduce tau acety- 1991;82:239–59.
6. Burgess JD, Pedraza O, Graff-Radford NR, Hirpa M,
lation and prevent the progression of neurode- Zou F, Miles R, Nguyen T, Li M, Lucas JA, Ivnik RJ,
generation. In addition, the sirtuin 1 activator et  al. Association of common KIBRA variants with
resveratrol is already being explored as a thera- episodic memory and AD risk. Neurobiol Aging.
peutic for AD given its anti-amyloidogenic prop- 2011;32:557.e551-559.
7. Carlomagno Y, Chung DC, Yue M, Castanedes-­
erties [56, 69]. Casey M, Madden BJ, Dunmore J, Tong J, DeTure
An alternative would be to inhibit the activity M, Dickson DW, Petrucelli L, et  al. An acetylation-­
of p300, the acetyltransferase that targets tau. In phosphorylation switch that regulates tau aggre-
support of this approach, it has been shown that gation propensity and function. J  Biol Chem.
2017;292:15277–86.
treatment with salsalate, which inhibits p300
52 T. Tracy et al.

8. Cohen TJ, Constance BH, Hwang AW, James M, Alzheimer’s disease: A dysregulated symphony.
Yuan CX.  Intrinsic Tau acetylation is coupled to Trends Neurosci. 2017;40:347–57.
auto-proteolytic Tau fragmentation. PLoS One. 21. Fukazawa Y, Saitoh Y, Ozawa F, Ohta Y, Mizuno K,
2016;11:e0158470. Inokuchi K.  Hippocampal LTP is accompanied by
9. Cohen TJ, Friedmann D, Hwang AW, Marmorstein enhanced F-actin content within the dendritic spine
R, Lee VM.  The microtubule-associated tau protein that is essential for late LTP maintenance in  vivo.
has intrinsic acetyltransferase activity. Nat Struct Mol Neuron. 2003;38:447–60.
Biol. 2013;20:756–62. 22. Fulga TA, Elson-Schwab I, Khurana V, Steinhilb

10. Cohen TJ, Guo JL, Hurtado DE, Kwong LK, Mills ML, Spires TL, Hyman BT, Feany MB.  Abnormal
IP, Trojanowski JQ, Lee VM. The acetylation of tau bundling and accumulation of F-actin mediates tau-­
inhibits its function and promotes pathological tau induced neuronal degeneration in vivo. Nat Cell Biol.
aggregation. Nat Commun. 2011;2:252. 2007;9:139–48.
11. Collie A, Maruff P. The neuropsychology of preclini- 23. Funk KE, Thomas SN, Schafer KN, Cooper GL, Liao
cal Alzheimer’s disease and mild cognitive impair- Z, Clark DJ, Yang AJ, Kuret J. Lysine methylation is
ment. Neurosci Biobehav Rev. 2000;24:365–74. an endogenous post-translational modification of tau
12. Cook C, Carlomagno Y, Gendron TF, Dunmore J, protein in human brain and a modulator of aggrega-
Scheffel K, Stetler C, Davis M, Dickson D, Jarpe tion propensity. Biochem J. 2014;462:77–88.
M, DeTure M, et  al. Acetylation of the KXGS 24. Gamblin TC, Chen F, Zambrano A, Abraha A,

motifs in tau is a critical determinant in modulation Lagalwar S, Guillozet AL, Lu M, Fu Y, Garcia-Sierra
of tau aggregation and clearance. Hum Mol Genet. F, LaPointe N, et al. Caspase cleavage of tau: linking
2014;23:104–16. amyloid and neurofibrillary tangles in Alzheimer’s
13. Corneveaux JJ, Liang WS, Reiman EM, Webster JA, disease. Proc Natl Acad Sci U S A. 2003;100:10032–7.
Myers AJ, Zismann VL, Joshipura KD, Pearson JV, 25. Giannakopoulos P, Herrmann FR, Bussiere T, Bouras
Hu-Lince D, Craig DW, et al. Evidence for an asso- C, Kovari E, Perl DP, Morrison JH, Gold G, Hof
ciation between KIBRA and late-onset Alzheimer’s PR.  Tangle and neuron numbers, but not amyloid
disease. Neurobiol Aging. 2010;31:901–9. load, predict cognitive status in Alzheimer’s disease.
14. Cripps D, Thomas SN, Jeng Y, Yang F, Davies P, Neurology. 2003;60:1495–500.
Yang AJ.  Alzheimer disease-specific conformation 26. Gorsky MK, Burnouf S, Dols J, Mandelkow E,

of hyperphosphorylated paired helical filament-­ Partridge L.  Acetylation mimic of lysine 280 exac-
Tau is polyubiquitinated through Lys-48, Lys-11, erbates human Tau neurotoxicity in  vivo. Sci Rep.
and Lys-6 ubiquitin conjugation. J  Biol Chem. 2016;6:22685.
2006;281:10825–38. 27. Grinberg LT, Wang X, Wang C, Sohn PD, Theofilas
15. Cunningham CC, Leclerc N, Flanagan LA, Lu M, P, Sidhu M, Arevalo JB, Heinsen H, Huang EJ, Rosen
Janmey PA, Kosik KS.  Microtubule-associated pro- H, et  al. Argyrophilic grain disease differs from
tein 2c reorganizes both microtubules and micro- other tauopathies by lacking tau acetylation. Acta
filaments into distinct cytological structures in an Neuropathol. 2013;125:581–93.
actin-binding protein-280-deficient melanoma cell 28. Guerrero-Munoz MJ, Gerson J, Castillo-Carranza

line. J Cell Biol. 1997;136:845–57. DL.  Tau oligomers: the toxic player at synapses in
16. de Calignon A, Polydoro M, Suarez-Calvet M,
Alzheimer’s disease. Front Cell Neurosci. 2015;9:464.
William C, Adamowicz DH, Kopeikina KJ, Pitstick R, 29. Guillozet AL, Weintraub S, Mash DC, Mesulam

Sahara N, Ashe KH, Carlson GA, et al. Propagation of MM.  Neurofibrillary tangles, amyloid, and memory
tau pathology in a model of early Alzheimer’s disease. in aging and mild cognitive impairment. Arch Neurol.
Neuron. 2012;73:685–97. 2003;60:729–36.
17. Decker JM, Kruger L, Sydow A, Dennissen FJ,
30. Haj-Yahya M, Lashuel HA.  Protein semisynthe-

Siskova Z, Mandelkow E, Mandelkow EM. The Tau/ sis provides access to Tau disease-associated post-­
A152T mutation, a risk factor for frontotemporal-­ translational modifications (PTMs) and paves the way
spectrum disorders, leads to NR2B receptor-mediated to deciphering the Tau PTM code in health and dis-
excitotoxicity. EMBO Rep. 2016;17:552–69. eased states. J Am Chem Soc. 2018;140:6611–21.
18.
Dickson DW, Kouri N, Murray ME, Josephs 31. Harris JA, Koyama A, Maeda S, Ho K, Devidze N,
KA.  Neuropathology of frontotemporal lobar Dubal DB, Yu GQ, Masliah E, Mucke L.  Human
degeneration-tau (FTLD-tau). J  Mol Neurosci. P301L-mutant tau expression in mouse entorhinal-­
2011;45:384–9. hippocampal network causes tau aggregation and
19.
Doost Mohammadpour J, Hosseinmardi N, presynaptic pathology but no cognitive deficits. PLoS
Janahmadi M, Fathollahi Y, Motamedi F, Rohampour One. 2012;7:e45881.
K. Non-­selective NSAIDs improve the amyloid-beta-­ 32. Hasegawa M, Morishima-Kawashima M, Takio K,
mediated suppression of memory and synaptic plas- Suzuki M, Titani K, Ihara Y.  Protein sequence and
ticity. Pharmacol Biochem Behav. 2015;132:33–41. mass spectrometric analyses of tau in the Alzheimer’s
20. Forner S, Baglietto-Vargas D, Martini AC, Trujillo-­ disease brain. J Biol Chem. 1992;267:17047–54.
Estrada L, LaFerla FM.  Synaptic impairment in 33. Hawley SA, Fullerton MD, Ross FA, Schertzer JD,
Chevtzoff C, Walker KJ, Peggie MW, Zibrova D,
4  Regulation of Tau Homeostasis and Toxicity by Acetylation 53

Green KA, Mustard KJ, et al. The ancient drug salicy- neurotrophic factor in SH-SY5Y human neuroblas-
late directly activates AMP-activated protein kinase. toma cells. Exp Toxicol Pathol. 2016;68:271–6.
Science. 2012;336:918–22. 47. Liu C, Gotz J. Profiling murine tau with 0N, 1N and
34. Hecht A, Laroche T, Strahl-Bolsinger S, Gasser SM, 2N isoform-specific antibodies in brain and peripheral
Grunstein M. Histone H3 and H4 N-termini interact organs reveals distinct subcellular localization, with
with SIR3 and SIR4 proteins: a molecular model the 1N isoform being enriched in the nucleus. PLoS
for the formation of heterochromatin in yeast. Cell. One. 2013;8:e84849.
1995;80:583–92. 48. Liu L, Drouet V, Wu JW, Witter MP, Small SA,

35. Hoover BR, Reed MN, Su J, Penrod RD, Kotilinek Clelland C, Duff K.  Trans-synaptic spread of tau
LA, Grant MK, Pitstick R, Carlson GA, Lanier LM, pathology in vivo. PLoS One. 2012a;7:e31302.
Yuan LL, et al. Tau mislocalization to dendritic spines 49. Liu Y, Peng L, Seto E, Huang S, Qiu Y. Modulation
mediates synaptic dysfunction independently of neu- of histone deacetylase 6 (HDAC6) nuclear import
rodegeneration. Neuron. 2010;68:1067–81. and tubulin deacetylase activity through acetylation.
36. Huganir RL, Nicoll RA. AMPARs and synaptic plas- J Biol Chem. 2012b;287:29168–74.
ticity: the last 25 years. Neuron. 2013;80:704–17. 50. Lu X, Deng Y, Yu D, Cao H, Wang L, Liu L, Yu C,
37. in ‘t Veld BA, Launer LJ, Breteler MM, Hofman A, Zhang Y, Guo X, Yu G.  Histone acetyltransfer-
Stricker BH.  Pharmacologic agents associated with ase p300 mediates histone acetylation of PS1 and
a preventive effect on Alzheimer’s disease: a review BACE1  in a cellular model of Alzheimer’s disease.
of the epidemiologic evidence. Epidemiol Rev. PLoS One. 2014;9:e103067.
2002;24:248–68. 51.
Lue LF, Brachova L, Civin WH, Rogers
38. Ingelsson M, Fukumoto H, Newell KL, Growdon JH, J.  Inflammation, A beta deposition, and neurofibril-
Hedley-Whyte ET, Frosch MP, Albert MS, Hyman lary tangle formation as correlates of Alzheimer’s dis-
BT, Irizarry MC. Early Abeta accumulation and pro- ease neurodegeneration. J  Neuropathol Exp Neurol.
gressive synaptic loss, gliosis, and tangle formation in 1996;55:1083–8.
AD brain. Neurology. 2004;62:925–31. 52. Lutz MI, Milenkovic I, Regelsberger G, Kovacs

39. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A, van GG.  Distinct patterns of sirtuin expression during
Eersel J, Wolfing H, Chieng BC, Christie MJ, Napier progression of Alzheimer’s disease. NeuroMolecular
IA, et al. Dendritic function of tau mediates amyloid-­ Med. 2014;16:405–14.
beta toxicity in Alzheimer’s disease mouse models. 53. Maeda S, Djukic B, Taneja P, Yu GQ, Lo I, Davis A,
Cell. 2010;142:387–97. Craft R, Guo W, Wang X, Kim D, et  al. Expression
40. Julien C, Tremblay C, Emond V, Lebbadi M, Salem of A152T human tau causes age-dependent neuronal
N Jr, Bennett DA, Calon F.  Sirtuin 1 reduction par- dysfunction and loss in transgenic mice. EMBO Rep.
allels the accumulation of tau in Alzheimer disease. 2016;17:530–51.
J Neuropathol Exp Neurol. 2009;68:48–58. 54. Maeda S, Sahara N, Saito Y, Murayama S, Ikai A,
41. Kopeikina KJ, Polydoro M, Tai HC, Yaeger E, Carlson Takashima A. Increased levels of granular tau oligo-
GA, Pitstick R, Hyman BT, Spires-Jones TL. Synaptic mers: an early sign of brain aging and Alzheimer’s
alterations in the rTg4510 mouse model of tauopathy. disease. Neurosci Res. 2006;54:197–201.
J Comp Neurol. 2013;521:1334–53. 55. Mair W, Muntel J, Tepper K, Tang S, Biernat J,

42. Kowall NW, Kosik KS. Axonal disruption and aber- Seeley WW, Kosik KS, Mandelkow E, Steen H, Steen
rant localization of tau protein characterize the neu- JA. FLEXITau: quantifying post-translational modifi-
ropil pathology of Alzheimer’s disease. Ann Neurol. cations of Tau protein in vitro and in human disease.
1987;22:639–43. Anal Chem. 2016;88:3704–14.
43. Krucker T, Siggins GR, Halpain S.  Dynamic actin 56. Marambaud P, Zhao H, Davies P.  Resveratrol pro-
filaments are required for stable long-term potentia- motes clearance of Alzheimer’s disease amyloid-beta
tion (LTP) in area CA1 of the hippocampus. Proc Natl peptides. J Biol Chem. 2005;280:37377–82.
Acad Sci U S A. 2000;97:6856–61. 57. McHugh TJ, Jones MW, Quinn JJ, Balthasar N,

44. Lasagna-Reeves CA, Castillo-Carranza DL, Sengupta Coppari R, Elmquist JK, Lowell BB, Fanselow MS,
U, Clos AL, Jackson GR, Kayed R.  Tau oligomers Wilson MA, Tonegawa S.  Dentate gyrus NMDA
impair memory and induce synaptic and mitochondrial receptors mediate rapid pattern separation in the hip-
dysfunction in wild-type mice. Mol Neurodegener. pocampal network. Science. 2007;317:94–9.
2011;6:39. 58. Merriam EB, Millette M, Lumbard DC, Saengsawang
45. Lasagna-Reeves CA, Castillo-Carranza DL, Sengupta W, Fothergill T, Hu X, Ferhat L, Dent EW. Synaptic
U, Sarmiento J, Troncoso J, Jackson GR, Kayed regulation of microtubule dynamics in dendritic
R.  Identification of oligomers at early stages of spines by calcium, F-actin, and drebrin. J  Neurosci
tau aggregation in Alzheimer’s disease. FASEB Off J Soc Neurosci. 2013;33:16471–82.
J. 2012;26:1946–59. 59. Min SW, Chen X, Tracy TE, Li Y, Zhou Y, Wang
46. Lattanzio F, Carboni L, Carretta D, Candeletti S,
C, Shirakawa K, Minami SS, Defensor E, Mok SA,
Romualdi P.  Treatment with the neurotoxic Abeta et  al. Critical role of acetylation in tau-mediated
(25-35) peptide modulates the expression of neuro- neurodegeneration and cognitive deficits. Nat Med.
protective factors Pin1, Sirtuin 1, and brain-derived 2015;21:1154–62.
54 T. Tracy et al.

60. Min SW, Cho SH, Zhou Y, Schroeder S, Haroutunian 73. Shirao T, Gonzalez-Billault C.  Actin filaments and
V, Seeley WW, Huang EJ, Shen Y, Masliah E, microtubules in dendritic spines. J  Neurochem.
Mukherjee C, et  al. Acetylation of tau inhibits its 2013;126:155–64.
degradation and contributes to tauopathy. Neuron. 74. Sohn PD, Tracy TE, Son HI, Zhou Y, Leite RE, Miller
2010;67:953–66. BL, Seeley WW, Grinberg LT, Gan L. Acetylated tau
61. Min SW, Sohn PD, Li Y, Devidze N, Johnson JR, destabilizes the cytoskeleton in the axon initial seg-
Krogan NJ, Masliah E, Mok SA, Gestwicki JE, Gan ment and is mislocalized to the somatodendritic com-
L.  SIRT1 deacetylates Tau and reduces pathogenic partment. Mol Neurodegener. 2016;11:47.
Tau spread in a mouse model of tauopathy. J Neurosci 75. Sokolow S, Henkins KM, Bilousova T, Gonzalez

Off J Soc Neurosci. 2018;38:3680–8. B, Vinters HV, Miller CA, Cornwell L, Poon WW,
62. Morishima-Kawashima M, Hasegawa M, Takio K, Gylys KH.  Pre-synaptic C-terminal truncated tau is
Suzuki M, Titani K, Ihara Y. Ubiquitin is conjugated released from cortical synapses in Alzheimer’s dis-
with amino-terminally processed tau in paired helical ease. J Neurochem. 2015;133(3):368–79.
filaments. Neuron. 1993;10:1151–60. 76. Tai HC, Serrano-Pozo A, Hashimoto T, Frosch MP,
63. Morris M, Knudsen GM, Maeda S, Trinidad JC,
Spires-Jones TL, Hyman BT.  The synaptic accu-
Ioanoviciu A, Burlingame AL, Mucke L.  Tau post-­ mulation of hyperphosphorylated tau oligomers in
translational modifications in wild-type and human Alzheimer disease is associated with dysfunction
amyloid precursor protein transgenic mice. Nat of the ubiquitin-proteasome system. Am J  Pathol.
Neurosci. 2015;18:1183–9. 2012;181:1426–35.
64. Morris M, Maeda S, Vossel K, Mucke L. The many 77. Tai HC, Wang BY, Serrano-Pozo A, Frosch MP,

faces of tau. Neuron. 2011;70:410–26. Spires-Jones TL, Hyman BT. Frequent and symmetric
65. Papassotiropoulos A, Stephan DA, Huentelman MJ, deposition of misfolded tau oligomers within presyn-
Hoerndli FJ, Craig DW, Pearson JV, Huynh KD, aptic and postsynaptic terminals in Alzheimer’s dis-
Brunner F, Corneveaux J, Osborne D, et al. Common ease. Acta Neuropathol Commun. 2014;2:146.
Kibra alleles are associated with human memory per- 78. Tatebayashi Y, Miyasaka T, Chui DH, Akagi T,

formance. Science. 2006;314:475–8. Mishima K, Iwasaki K, Fujiwara M, Tanemura K,
66. Pooler AM, Phillips EC, Lau DH, Noble W,
Murayama M, Ishiguro K, et  al. Tau filament for-
Hanger DP.  Physiological release of endogenous mation and associative memory deficit in aged mice
tau is stimulated by neuronal activity. EMBO Rep. expressing mutant (R406W) human tau. Proc Natl
2013;14:389–94. Acad Sci U S A. 2002;99:13896–901.
67. Rodriguez-Rodriguez E, Infante J, Llorca J, Mateo I, 79. Thomas SN, Funk KE, Wan Y, Liao Z, Davies P,
Sanchez-Quintana C, Garcia-Gorostiaga I, Sanchez-­ Kuret J, Yang AJ.  Dual modification of Alzheimer’s
Juan P, Berciano J, Combarros O. Age-dependent asso- disease PHF-tau protein by lysine methylation and
ciation of KIBRA genetic variation and Alzheimer’s ubiquitylation: a mass spectrometry approach. Acta
disease risk. Neurobiol Aging. 2009;30:322–4. Neuropathol. 2012;123:105–17.
68. Santacruz K, Lewis J, Spires T, Paulson J, Kotilinek 80. Tracy TE, Sohn PD, Minami SS, Wang C, Min SW, Li
L, Ingelsson M, Guimaraes A, DeTure M, Ramsden Y, Zhou Y, Le D, Lo I, Ponnusamy R, et al. Acetylated
M, McGowan E, et al. Tau suppression in a neurode- Tau obstructs KIBRA-mediated signaling in synaptic
generative mouse model improves memory function. plasticity and promotes tauopathy-related memory
Science. 2005;309:476–81. loss. Neuron. 2016;90:245–60.
69. Sawda C, Moussa C, Turner RS.  Resveratrol
81. Trzeciakiewicz H, Tseng JH, Wander CM, Madden V,
for Alzheimer’s disease. Ann N Y Acad Sci. Tripathy A, Yuan CX, Cohen TJ.  A dual pathogenic
2017;1403:142–9. mechanism links Tau acetylation to sporadic tauopa-
70. Scroggins BT, Robzyk K, Wang D, Marcu MG,
thy. Sci Rep. 2017;7:44102.
Tsutsumi S, Beebe K, Cotter RJ, Felts S, Toft D, 82. Van der Jeugd A, Hochgrafe K, Ahmed T, Decker JM,
Karnitz L, et  al. An acetylation site in the middle Sydow A, Hofmann A, Wu D, Messing L, Balschun D,
domain of Hsp90 regulates chaperone function. Mol D’Hooge R, et al. Cognitive defects are reversible in
Cell. 2007;25:151–9. inducible mice expressing pro-aggregant full-length
71. Sebti S, Prebois C, Perez-Gracia E, Bauvy C,
human Tau. Acta Neuropathol. 2012;123:787–805.
Desmots F, Pirot N, Gongora C, Bach AS, Hubberstey 83. Wang YH, Tsay YG, Tan BC, Lo WY, Lee

AV, Palissot V, et al. BAT3 modulates p300-dependent SC.  Identification and characterization of a novel
acetylation of p53 and autophagy-related protein 7 p300-mediated p53 acetylation site, lysine 305. J Biol
(ATG7) during autophagy. Proc Natl Acad Sci U S A. Chem. 2003;278:25568–76.
2014;111:4115–20. 84. Warmus BA, Sekar DR, McCutchen E, Schellenberg
72. Shi D, Pop MS, Kulikov R, Love IM, Kung AL, GD, Roberts RC, McMahon LL, Roberson ED. Tau-­
Grossman SR.  CBP and p300 are cytoplasmic E4 mediated NMDA receptor impairment underlies
polyubiquitin ligases for p53. Proc Natl Acad Sci U S dysfunction of a selectively vulnerable network in a
A. 2009;106:16275–80. mouse model of frontotemporal dementia. J Neurosci
Off J Soc Neurosci. 2014;34:16482–95.
4  Regulation of Tau Homeostasis and Toxicity by Acetylation 55

85. Wong HK, Veremeyko T, Patel N, Lemere CA,

88. Zempel H, Thies E, Mandelkow E, Mandelkow

Walsh DM, Esau C, Vanderburg C, Krichevsky EM.  Abeta oligomers cause localized Ca(2+) eleva-
AM.  De-repression of FOXO3a death axis by tion, missorting of endogenous Tau into dendrites,
microRNA-132 and -212 causes neuronal apop- Tau phosphorylation, and destruction of microtu-
tosis in Alzheimer’s disease. Hum Mol Genet. bules and spines. J  Neurosci Off J  Soc Neurosci.
2013;22:3077–92. 2010;30:11938–50.
86. Yamada K, Holth JK, Liao F, Stewart FR, Mahan 89. Zhang B, Higuchi M, Yoshiyama Y, Ishihara T, Forman
TE, Jiang H, Cirrito JR, Patel TK, Hochgrafe K, MS, Martinez D, Joyce S, Trojanowski JQ, Lee
Mandelkow EM, et  al. Neuronal activity regulates VM. Retarded axonal transport of R406W mutant tau
extracellular tau in vivo. J Exp Med. 2014;211:387–93. in transgenic mice with a neurodegenerative tauopa-
87. Yoshiyama Y, Higuchi M, Zhang B, Huang SM, Iwata thy. J Neurosci Off J Soc Neurosci. 2004;24:4657–67.
N, Saido TC, Maeda J, Suhara T, Trojanowski JQ, Lee 90. Zhao X, Kotilinek LA, Smith B, Hlynialuk C, Zahs
VM.  Synapse loss and microglial activation precede K, Ramsden M, Cleary J, Ashe KH. Caspase-2 cleav-
tangles in a P301S tauopathy mouse model. Neuron. age of tau reversibly impairs memory. Nat Med.
2007;53:337–51. 2016;22:1268–76.
 Tau Clearance Mechanisms
5
Maoping Tang, Jarreau Harrison, Carol A. Deaton,
and Gail V. W. Johnson

The Proteasome dered proteins (IDPs) through an ATP- and

ubiquitin-­independent process. Further, there is
The catalytic core particle (CP) of the protea- increasing evidence that the 20S proteasome rep-
some is a barrel-shaped 700 kD multimeric cel- resents a prominent protein degradation route in
lular structure that has three proteolytic activities; mammalian cells. The majority of proteasomes in
chymotryptic-like, tryptic-like, and caspase-like mammalian cells are uncapped 20S proteasomes
[1, 2]. The CP, also known as the 20S protea- and there are multiple mechanisms by which the
some, has two outer rings made of multiple α proteolytic capacity of the 20S proteasome is
subunits and two inner rings made of β subunits. regulated [4]. For example, it has been suggested
The inner β subunits (β1, β2 and β5) are catalyti- that 20S proteasomes are predominantly in a
cally active, while the outer α subunits are pro- “closed gate” conformation, but they open inter-
teolytically inactive but play a role in defining the mittently to accept substrates, and the mere pres-
size of the pore which allows client proteins ence of substrates induces a shift towards the
inside the structure [3]. The 20S proteasome can open conformation allowing substrate entrance
be uncapped, singly  capped or doubly capped into the proteolytic core [5, 6]. Another level of
with the 19S regulatory particle (RP) (or in some regulation may involve “nanny” proteins that can
cases the 11S RP) to form the 26S proteasome. interact with IDPs and protect them from degra-
The RPs facilitate the recognition and deubiqui- dation by the 20S proteasomes [7]. However,
tylation of polyubiquitylated client proteins, further research needs to be done to fully delin-
and mediate the unfolding and targeting of the eate the functioning of 20S proteasomes in cells
proteins into the core of the 20S proteasome [2]. and how specific substrates are targeted to them
The 20S proteasome is able to directly degrade for degradation.
natively unfolded proteins or intrinsically disor- The 26S proteasome primarily degrades pro-
teins that have been polyubiquitylated with a
ubiquitin chain of at least four molecules. The
Authors Jarreau Harrison and Carol A. Deaton have been process of ubiquitylation is ATP dependent and
contributed equally to this chapter.
requires the action of three enzymes. The E1
enzyme forms a thio-ester bond with ubiquitin at
M. Tang · J. Harrison · C. A. Deaton
G. V. W. Johnson (*) the expense of ATP, then the activated ubiquitin is
University of Rochester, Rochester, NY, USA transferred from E1 to E2. Finally, the E3 enzyme
e-mail: angletangmp@gmail.com; jarreau_harrison@ catalyzes the transfer of the ubiquitin from the E2
urmc.rochester.edu; carol_deaton@urmc.rochester. enzyme, forming an isopeptide bond between the
edu; gail_johnsonvoll@urmc.rochester.edu

© Springer Nature Singapore Pte Ltd. 2019 57

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_5
58 M. Tang et al.

ubiquitin carboxyl terminus and a primary amino several different cell models have substantiated
group of a client protein (mono-ubiquitylation/ this initial finding that tau can be degraded by the
ubiquitin chain initiation) or another ubiquitin 20S proteasome in a ubiquitin independent man-
molecule (ubiquitin chain elongation) [8]. In ver- ner [14, 15]. Indeed, using a variety of different
tebrates, there are only two E1 enzymes, Uba1 approaches (ATP depletion, ubiquitylation-­
and Uba6, with Uba1 being the most abundant. In deficient cells, knockdown of a 19S proteasomal
contrast, there are approximately 50 E2 enzymes regulator subunit, as well as in  vitro ubiquity-
in humans, and more than 600 putative E3 lation studies) it was demonstrated that ubiquity-
enzymes have been identified in the human lation was not required for tau degradation by the
genome [8, 9]. proteasome [15]. Further, recent studies have
Immunoproteasomes are another class of pro- shown that small molecules which enhance the
teasomes and are, in essence, the 20S proteasome activity or levels of the 20S proteasome resulted
with inducible catalytic subunits: LMP2, MECL-­ in increased clearance of tau in various clonal
1, and LMP7, instead of β1, β2, and β5. These cell models [5, 16]. Given the fact that the 20S
inducible catalytic subunits are usually expressed proteasome efficiently degrades IDPs and the
in response to pro-inflammatory signals [10, 11]. fact that tau is an IDP, it is likely that in a physi-
Immunoproteasomes play a key role in the immune ological setting tau is degraded by the proteasome
response, and have well-defined functions in anti- in a ubiquitin-independent manner (Fig.  5.1).
gen-presenting cells [10]. However, there is a However, when tau undergoes pathological
growing awareness that immunoproteasomes can modifications such as phosphorylation and
modulate neurodegenerative processes. For exam- truncation, which occurs in Alzheimer’s disease
ple, when LMP7 knockout mice were crossed with (AD) and other neurodegenerative disorders,  it
APP/PS1 mice (which express human transgenes may become less disordered and take on a more
for APP with the Swedish mutation (K595N/ structured and perhaps “preaggregation” confor-
M596L) and PSEN1 containing an L166P muta- mation [17–20]. In this altered conformation/
tion, both under the control of the Thy1 promoter) structure tau, may not be as efficiently degraded
there was an altered cytokine response in microg- by the 20S proteasome; hence, other degradative
lia and an improvement in the cognitive deficits in mechanisms may need to be engaged to facilitate
the offspring compared to the APP/PS1 mice with tau’s clearance.
LMP7 [11]. However, given that the immunopro-
teasome appears to predominantly affect the func-
tioning of immune cells, its role in directly  au Degradation by the 26S
T
mediating tau clearance is not clear. Proteasome

Data indicate that tau is degraded by the pro-

 au Degradation by the 20S
T teasome; however the involvement of the 26S
Proteasome proteasome in degrading tau in neurons has not
been fully elucidated. Numerous studies have
Tau is one of the largest IDPs [12], and there are shown that in clonal cells, treatment with protea-
data suggesting that tau, like other IDPs, is some inhibitors (e.g., MG132, lactacystin, or
degraded by the 20S proteasome. An early study epoxomicin) increase tau levels or slow tau deg-
showed that treatment of human neuroblastoma radation [13–15, 21]. Since these inhibitors tar-
SH-SY5Y cells with the proteasome inhibitor get proteasome activity in general, this approach
lactacystin, significantly attenuated tau degrada- does not allow a determination of whether the
tion. Additionally, in an in vitro system tau was 20S or 26S proteasome or both are involved in
degraded by the 20S proteasome in the absence clearing tau. In vitro tau can be ubiquitylated
of ubiquitylation [13]. Ensuing studies using [14, 21, 22], and early work suggested that tau
5  Tau Clearance Mechanisms 59

Fig. 5.1  Degradation of tau by the 20S proteasome. (a) accept substrates, including tau, and the presence of tau
Tau is a intrinsically disordered protein (IDP) and there- (or other substrates) facilitates the open conformation and
fore is likely degraded by the 20S proteasome in an ATP- thus entrance of tau into the proteolytic core [5, 6]. (b)
and ubiquitin-independent manner [4]. It has been The degradation of tau by the 20S proteasome may also be
suggested that 20S proteasomes are predominantly in a regulated by “nanny” proteins that interact with IDPs and
“closed gate” conformation but open intermittently to protect them from degradation [7]

could be ­ubiquitylated in situ using a HEK cell hyperphosphorylated, is also polyubiquitylated

model [21]. However, in an immortalized mouse with K48, K11 and K6 linkages. In this study, it
cortical cell line this did not appear to be the case was suggested that modification at K6 inhibits
[14]. Nonetheless, mass spectrometric analyses ubiquitin-dependent degradation by 26S protea-
of wild type mouse tau revealed numerous sites some and thus could contribute to the formation
were ubiquitylated by inference from the detec- of neurofibrillary tangles (NFTs) [27]. In addi-
tion of GlyGly-modified lysine (K) residues due tion, the ubiquitin proteasome system (UPS),
to trypsin cleavage of ubiquitin [23]. In addition, including the activities of the proteasomal prote-
it is difficult to define the contribution of the pro- ases, has been reported to be impaired in AD
teasome, 20S or 26S, to the turnover of tau in brain [28]. This could be due in part to the pres-
neurons. This is due to the fact that treatment of ence of abnormally modified and/or oligomer-
cultured primary neurons with proteasome inhib- ized tau species (Fig.  5.2). For example,
itors results in compensatory activation of the impairment of the 26S proteasome has been
autophagy pathway [24] with a subsequent reported in the rTg4510 mouse model where
increase in tau turnover [25, 26]. Thus, the con- P301L tau (a mutant tau species which causes
tribution of ubiquitin modifications to tau turn- frontotemporal lobar degeneration (FTLD)) is
over by the 26S proteasome in a physiological robustly overexpressed [29]. In another study, it
setting has not been unequivocally delineated. As was shown that the deubiquitinase Otub1
indicated above, tau that has been pathologically removes K48-linked ubiquitin from P301S tau
modified in AD or other tauopathies may take on (which is another FTLD mutant). Furthermore,
an altered conformation and thus no longer be overexpression of Otub1 results in increased lev-
efficiently degraded by the 20S proteasome. In els of tau phosphorylated at S202/T205 (AT8
this case, tau may be ubiquitylated and turned antibody epitope), as well as increased tau oligo-
over by the 26S proteasome. Paired helical fila- merization and seeding both in primary neurons
ment (PHF) tau isolated from AD brain, which is from tau P301S mice (PS19 line) and in  vivo
60 M. Tang et al.

Fig. 5.2  Degradation of tau by the 26S proteasome. (a) teasome [30]. (b) Aggregated tau is ubiquitylated and is
Studies indicate that FTLD mutant tau species, such as likely not effectively degraded by the 26S proteasome but
P301S tau, are ubiquitylated and degraded by the 26S pro- instead may impair its function [28, 29]

[30]. Interestingly, the P301S mutation in tau are intriguing, but further studies are needed to
results in conformational changes which impact determine if PROTACs can facilitate tau clear-
structure and function [31], and these changes ance in vivo.
may make it differentially susceptible to modifi-
cation by certain E3 ligases and deubiquitinases
compared to wild type tau. Overall, these data Autophagy
indicate that the 26S proteasome plays a role
directly or indirectly in the turnover of P301S tau There are three defined types of autophagy, a pro-
(Fig. 5.2). cess that involves degradation of substrates by
PROTACs (PROteolysis TArgeting Chimeras) the lysosome. Microautophagy involves direct
are heterobiofunctional peptides that bind an E3 lysosomal or vacuolar engulfment of cytoplasmic
ligase and a target protein, and enable ubiquity- cargo. In chaperone-mediated autophagy (CMA),
lation and degradation by the 26S proteasome substrate proteins are directly delivered into the
[32]. Therefore it is possible that PROTACs may lumen of lysosomes through a Hsc70/Lamp2a-­
be useful in facilitating tau clearance. In a recent mediated process. Macroautophagy (which will
study, a PROTAC peptide that binds tau and be referred to simply as autophagy in the sections
Keap1, and thus recruits the Keap1-Cul3 ubiqui- below) is the most studied  type of autophagy,
tin ligase complex to tau, was shown to enter particularly in the context of AD and other neuro-
cells and induce tau degradation in models where degenerative diseases, and involves the delivery
tau is overexpressed  [33]. These initial findings of cargo to lysosomes via a vacuole-based
5  Tau Clearance Mechanisms 61

process. For more complete reviews of autoph- lation of autophagy-mediated tau clearance must
agy, as well as its role in neurodegenerative dis- be interpreted with caution and should be subse-
eases, see [34–38]. Given that the majority of quently validated using primary neuron or in vivo
studies that have examined autophagic tau clear- models.  As an example of this consideration,
ance have focused on macroautophagy/autoph- nutrient deprivation (starvation) and mTOR inhi-
agy, a brief review of this pathway will be bition (treatment with rapamycin or torin 1) reli-
presented here. ably and reproducibly upregulate autophagy in
Autophagy involves the formation and elon- many non-neuronal models; however, data indi-
gation of a double membrane structure called a cate that this is not always the case in neurons
phagophore which then closes around cargos to [50–53]. Interestingly, autophagy appears to be
form the autophagosome  which subsequently more efficient in younger neurons, as the expres-
fuses with lysosomes. This fusion of an autopha- sion levels of proteins that induce autophagy,
gosome and lysosome results in the formation of such as Beclin-1, ATG5 and ATG7, decline with
an autolysosome. Alternatively, the autophago- age [41, 54] and this attenuation of autophagy
some can fuse with an endosome to form an may potentially be a contributing risk factor for
amphisome which also then fuses with the lyso- the onset of age-related neurodegenerative dis-
some [39]. This process can be non-selective eases including AD.
and it is sometimes referred to as bulk autophagy,
when portions of the cytosol are engulfed in the
process of autophagosome formation.  The Tau Degradation by Autophagy
engulfed contents are then degraded to provide
the cell with recycled basic components for Numerous studies have provided compelling evi-
energy usually in response to nutrient depriva- dence that autophagy is a significant contributor
tion. On the other hand, selective autophagy in the degradation of tau (Fig. 5.3). A functioning
involves the recognition of specific clients by lysosome is essential for completion of autoph-
autophagy receptor complexes and targeting to agy and early reports demonstrated that tau was a
the developing autophagosomes [35, 40]. substrate of the lysosomal aspartyl protease,
However, the general scheme of how autolyso- cathepsin D.  In vitro, at a low pH, cathepsin D
somes form and degrade substrates is very simi- cleaved tau, and adding cathepsin D to rat brain
lar, if not the same, for both non-selective and homogenates also resulted in tau degradation
selective autophagy [37]. [55]. Treatment of hippocampal slices with chlo-
Neuronal autophagy is constitutively active roquine, which raises lysosomal pH and is often
and highly efficient [41]. Since neurons are used to inhibit autophagic flux, increased tau lev-
unique in being post-mitotic and highly asym- els, and in particular tau phosphorylated at
metric cells, autophagy is not a homogenous Ser396/404 (PHF1 epitope) [56]. Lastly, in a fly
event but rather is differentially regulated in the tauopathy model, ablation of cathepsin D potenti-
various compartments of the cell (soma, axon, ated tau-induced neurotoxicity [57].
dendrites, pre- and post-synapse) [42–45]. These Investigations carried out in non-neuronal
regulatory differences are key when considering models have also provided supporting evidence
autophagy-mediated tau turnover because tau, that tau is degraded by autophagy. In human neu-
though primarily axonal, also localizes to den- roblastoma SH-SY5Y cells overexpressing
drites, as well as the pre- and post-synaptic com- P301L tau, stimulation of autophagy by serum
partments [46–48], and the autophagy-mediated starvation or rapamycin treatment reduced tau
clearance of tau likely differs in these various cel- inclusions. This  effect was blocked when the
lular compartments [45, 49, 50].  With this in autophagy protein ATG5 was deleted [58]. When
mind, data from studies in which non-neuronal or N2A cells engineered to overexpress a tau con-
clonal cell models are used to examine the regu- struct of only the microtubule binding repeat
62 M. Tang et al.

Fig. 5.3  Degradation of tau by autophagy. Tau is likely tau, can also be incorporated into endosomes [87, 88]
degraded through different autophagic routes. Autophagy which then can go on to fuse directly with lysosomes
adaptor protein complexes interact with tau targeting it to forming endolysosomes, or the endosome can fuse with
the autophagosome which fuses with the lysosome to an autophagosome to form an amphisome prior to fusing
form autolysosomes. Cytosolic, as well as extracellular with lysosomes [39]

region of tau (referred to as “K18”) with the gesting that autophagy is a mediator of tau
FTLD ΔK280 mutation (which results in clearance.
increased aggregation) were treated with treha- In vivo and primary neuron models have been
lose, an mTOR-independent autophagy activa- used to provide strong support for a role of
tor,  there was a  significant reduction in autophagy in tau clearance. Interestingly, evi-
aggregated tau, as well as the  levels of the tau dence for the role of autophagy in clearing tau in
construct in both soluble and insoluble fractions neurons came from studies that were attempting
[25]. In a human neuroblastoma cell line that to elucidate the role of the proteasome in tau
inducibly expresses tau, treatment with the degradation [25]. In one study, treatment of rat
autophagy inhibitors chloroquine or 3-methylad- primary neurons with proteasomal inhibitors
enine, resulted in tau accumulation and increases unexpectedly reduced tau levels. This effect was
in tau in sarkosyl-insoluble fractions [59]. Lastly, likely due to a compensatory upregulation of
studies in an immortalized cortical cell model in autophagy, as evidenced by increased levels of
which either full-length tau or tau truncated at LC3-II, which is both an indicator of increased
D421 (to mimic caspase 3 cleavage which occurs autophagy induction, and increased numbers of
in pathological conditions such as AD) were autophagosomes [25, 26]. In hippocampal slices
inducibly expressed, revealed D421 truncated tau from JNPL3 mice, which express P301L tau
was preferentially degraded through the autoph- under the control of the mouse prion promoter,
agy pathway [14]. Although all these studies treatment with methylene blue to induce autoph-
were in clonal cell models, they provide data sug- agy resulted in a decrease of phosphorylated tau
5  Tau Clearance Mechanisms 63

and insoluble tau [60]. In vivo treatment of the Furthermore, in a clonal cell model, p62 played a
JNPL3 mice with methylene blue, as well as role in targeting tau aggregates to autophagy
treatment of another FTLD mouse model that [70]. Interestingly, in the same study, it was found
expresses P301S tau with trehalose, reduced the that while p62 bound tau aggregates, it did not
levels of insoluble tau [60–62]. In the P301S tau interact with the initial “tau seeds”. In contrast,
model trehalose treatment also significantly these seeds were recognized by NDP52 [70].
reduced tau phosphorylated at T212/S214 Indeed, an earlier study demonstrated that the
(AT100) (however, no other phosphorylation autophagy receptor NDP52 directly binds tau
sites were assessed) [61, 62]. Mice with a dele- through its SKICH domain and that, in primary
tion of autophagy gene ATG7 in their forebrain neurons, knocking down NDP52 results in
neurons develop age-dependent neurodegenera- increased levels of phosphorylated tau [67]. It
tion with accumulation of phosphorylated tau was also shown  that NDP52 co-localized with
inclusions [63]. Taken together these and other phospho-tau in the brains of APPsw/PS1dE9
studies provide compelling evidence that transgenic mice (which express mutant APP with
autophagy plays a key role in mediating tau the Swedish mutation and PSEN1ΔE9) [71].
clearance. NBR1 is another autophagy receptor, and in spo-
A critically important factor in clearance of radic inclusion body myositis muscle fiber inclu-
tau by selective autophagy is the involvement of sions containing phosphorylated tau also co-label
autophagy receptors (also referred to as autoph- with an antibody to NBR1 and it was postulated
agy cargo adaptors), chaperones, and co-­ that abnormalities in NBR1 may contribute to the
chaperones that work in concert to target tau to accumulation of phosphorylated tau species [72].
the autophagy pathway. Autophagy receptors are However, the role of NBR1  in mediating tau
key components of the selective autophagy path- clearance in neuronal models has not been sub-
way as they contain domains that allow them to stantially investigated. The role of optineurin in
engage specific cargos. Receptors also have an tau clearance has also not been extensively stud-
LC3 interacting region (LIR) for binding LC3 ied. However, a report in HeLa cells suggested
and targeting clients to the developing autopha- that optineurin may play a role in clearing soluble
gosome for degradation [40, 64, 65]. Autophagy tau species [69].
receptors have domains that allow them to inter- Recent studies have provided evidence that
act with ubiquitylated substrates; however, in autophagic degradation of tau is mediated in
some cases, they are also able to interact with part by Nrf2 induction of NDP52 [67].
non-ubiquitylated clients. For example, the Intriguingly, Nrf2 also appears to regulate the
autophagy receptor optineurin mediates the expression of BAG3 (BCL2-associated athano-
clearance of protein aggregates that were not gene 3) which is induced by proteotoxic stress,
ubiquitylated [66], and the binding of the autoph- is increased during aging [73], and plays a criti-
agy receptor, NDP52, to tau also occurs indepen- cal role in regulating soluble tau levels [26, 49,
dent of ubiquitylation [67]. 74]. BAG3 is a multidomain protein that plays a
Several different autophagy receptor proteins key role in regulating autophagy [49, 75, 76], as
have been implicated in facilitating the clearance well as functioning as part of a complex that tar-
of tau. Deletion of the autophagy receptor p62, in gets specific clients, including tau, to this degra-
a mouse model resulted in accumulation of dative pathway [26, 49, 77]. BAG3 binds heat
hyperphosphorylated tau, and thus, it was sug- shock proteins and it has been shown that a
gested that p62 facilitates tau clearance [68]. p62 BAG3-HspB8-Hsp70 complex can bind mis-
and LC3 co-stained tau positive inclusions in the folded protein aggregates and target them for
brains of mice that overexpress P301S tau (PS19 degradation via autophagy [78]. In young pri-
line) in an ubiquitin-independent manner [55, mary neurons, BAG3 overexpression induced
62]. In this same mouse model, overexpression of autophagy  which facilitated the clearance of
p62 resulted in a decrease in tau pathology [69]. endogenous tau [26]. Interestingly, in mature
64 M. Tang et al.

neurons BAG3 appears to regulate autophagy endosomal-microautophagy pathway [86]. It was

differentially depending on the cellular com- also recently demonstrated that tau can be deliv-
partment, preferentially contributing to autopha- ered to lysosome via early endosomes and multi-
gosome-lysosome fusion and ­ phosphorylated vesicular bodies (MVBs), and that this is
tau clearance in dendrites and at the post-­ dependent on Rab35 and the ESCRT system [87].
synapse. In a recent study it was demon- Further, the uptake of monomeric tau by human
strated that expression of BAG3 is significantly neurons can occur through endocytosis and
higher in inhibitory neurons compared to excit- micropinocytosis with the resulting vacuoles
atory neurons in control human brain. These likely routing tau to the lysosome [88]. However,
findings are significant  as in AD,  pathological the mechanisms involved and the exact fate of tau
tau species accumulate in excitatory neurons to once it is internalized through these routes have
a greater extent than inhibitory neurons not been fully elucidated.
[74]. Overall, these studies indicate that BAG3
is a significant contributor to the autophagic
processes that mediate tau clearance. Summary
PICALM/CALM (phosphatidylinositol-­
binding clathrin assembly protein) has been Overall, it is clear that tau is degraded by multi-
shown to associate with AD based on genome-­ ple mechanisms and that many factors determine
wide association studies [79]. In addition to regu- which route is taken. In addition, pathological
lating endocytosis, PICALM/CALM is able to conditions such as AD, likely have altered degra-
mediate different steps of autophagy and facili- dative pathways and, hence, altered tau clear-
tates tau clearance. Since PICALM/CALM plays ance. Decreased efficiency of any of these
a role in regulating autophagy and tau clearance clearance pathways is likely to have detrimental
it can be suggested that this may be why it is a effects on tau turnover, potentially enhancing tau
genetic risk factor for AD [80]. accumulation and pathology. Further studies to
elucidate tau clearance mechanisms are likely to
provide key insights into the underlying pro-
Tau Clearance by Chaperone-­ cesses that contribute to tau accumulation and its
Mediated Autophagy (CMA) subsequent role in the  pathogenesis in AD and
and Microautophagy other tauopathies.

Although the majority of studies have focused on

clearance of tau by macroautophagy/autophagy, References
CMA has also been examined. In CMA, substrate
proteins are first recognized by the cytosolic 1. DeMartino GN, Slaughter CA.  The proteasome, a
novel protease regulated by multiple mechanisms.
chaperone, hsc70 and directly targeted to the lys- J Biol Chem. 1999;274(32):22123–6. PubMed PMID:
osome by binding LAMP2a [81]. Most proteins 10428771
targeted by hsc70 contain a KFERQ targeting 2. Tanaka K.  The proteasome: from basic mechanisms
motif [82, 83]. Tau has  two imperfect KFERQ to emerging roles. Keio J  Med. 2013;62(1):1–12.
PubMed PMID: 23563787
motifs in the fourth repeat domain (336QVEVK340 3. Budenholzer L, Cheng CL, Li Y, Hochstrasser
and 347KDRVQ351), which are recognized by M.  Proteasome structure and assembly. J  Mol Biol.
Hsc70 [84, 85]. Using the mouse N2A neuroblas- 2017;429(22):3500–24. PubMed PMID: 28583440.
toma cell line with stable expression of full length Pubmed Central PMCID: 5675778
4. Ben-Nissan G, Sharon M.  Regulating the 20S pro-
wild type tau, A152T tau, or P301L tau, it was teasome ubiquitin-independent degradation pathway.
shown that CMA contributed to the degradation Biomol Ther. 2014;4(3):862–84. PubMed PMID:
of wild type and A152T tau, but not P301L tau. In 25250704. Pubmed Central PMCID: 4192676
this study it was also demonstrated that wild type 5. Njomen E, Osmulski PA, Jones CL, Gaczynska M,
Tepe JJ.  Small molecule modulation of proteasome
and A152T tau could be degraded through an
5  Tau Clearance Mechanisms 65

assembly. Biochemistry. 2018;57(28):4214–24. tion. J Cell Biol. 1989;109(4 Pt 1):1643–51. PubMed

PubMed PMID: 29897236 PMID: 2507554. Pubmed Central PMCID: 2115827
6. Osmulski PA, Hochstrasser M, Gaczynska M. A tet- 18. Kovacech B, Skrabana R, Novak M. Transition of tau
rahedral transition state at the active sites of the 20S protein from disordered to misordered in Alzheimer’s
proteasome is coupled to opening of the alpha-ring disease. Neurodegener Dis. 2010;7(1–3):24–7.
channel. Structure. 2009;17(8):1137–47. PubMed PubMed PMID: 20160453
PMID: 19679091. Pubmed Central PMCID: 2746003 19.
Sevcik J, Skrabana R, Kontsekova E, Novak
7. Tsvetkov P, Reuven N, Shaul Y. The nanny model for M.  Structure solution of misfolded conformations
IDPs. Nat Chem Biol. 2009;5(11):778–81. PubMed adopted by intrinsically disordered Alzheimer’s tau
PMID: 19841623 protein. Protein Pept Lett. 2009;16(1):61–4. PubMed
8. McDowell GS, Philpott A. New insights into the role PMID: 19149675
of ubiquitylation of proteins. Int Rev Cell Mol Biol. 20. Zhu S, Shala A, Bezginov A, Sljoka A, Audette G,
2016;325:35–88. PubMed PMID: 27241218 Wilson DJ.  Hyperphosphorylation of intrinsically
9. Li W, Bengtson MH, Ulbrich A, Matsuda A, Reddy disordered tau protein induces an amyloidogenic
VA, Orth A, et  al. Genome-wide and functional shift in its conformational ensemble. PLoS One.
annotation of human E3 ubiquitin ligases identi- 2015;10(3):e0120416. PubMed PMID: 25767879.
fies MULAN, a mitochondrial E3 that regulates Pubmed Central PMCID: 4359001
the organelle’s dynamics and signaling. PLoS One. 21. Petrucelli L, Dickson D, Kehoe K, Taylor J, Snyder H,
2008;3(1):e1487. PubMed PMID: 18213395. Pubmed Grover A, et al. CHIP and Hsp70 regulate tau ubiquiti-
Central PMCID: 2198940 nation, degradation and aggregation. Hum Mol Genet.
10. Ferrington DA, Gregerson DS. Immunoproteasomes: 2004;13(7):703–14. PubMed PMID: 14962978
structure, function, and antigen presentation. Prog 22. Shimura H, Schwartz D, Gygi SP, Kosik KS. CHIP-­
Mol Biol Transl Sci. 2012;109:75–112. PubMed Hsc70 complex ubiquitinates phosphorylated
PMID: 22727420. Pubmed Central PMCID: 4405001 tau and enhances cell survival. J  Biol Chem.
11. Wagner LK, Gilling KE, Schormann E, Kloetzel PM, 2004;279(6):4869–76. PubMed PMID: 14612456
Heppner FL, Kruger E, et  al. Immunoproteasome 23. Morris M, Knudsen GM, Maeda S, Trinidad JC,

deficiency alters microglial cytokine response and Ioanoviciu A, Burlingame AL, et  al. Tau post-­
improves cognitive deficits in Alzheimer’s disease-­ translational modifications in wild-type and human
like APPPS1 mice. Acta Neuropathol Commun. amyloid precursor protein transgenic mice. Nat
2017;5(1):52. PubMed PMID: 28646899. Pubmed Neurosci. 2015;18(8):1183–9. PubMed PMID:
Central PMCID: 5483273 26192747
12. Jebarupa B, Muralidharan M, Arun A, Mandal AK, 24. Lee MJ, Lee JH, Rubinsztein DC. Tau degradation: the
Mitra G.  Conformational heterogeneity of tau: ubiquitin-proteasome system versus the autophagy-­
implication on intrinsic disorder, acid stability and lysosome system. Prog Neurobiol. 2013;105:49–59.
fibrillation in Alzheimer’s disease. Biophys Chem. PubMed PMID: 23528736
2018;241:27–37. PubMed PMID: 30081240 25.
Kruger U, Wang Y, Kumar S, Mandelkow
13. David DC, Layfield R, Serpell L, Narain Y, Goedert EM.  Autophagic degradation of tau in primary neu-
M, Spillantini MG.  Proteasomal degradation of tau rons and its enhancement by trehalose. Neurobiol
protein. J  Neurochem. 2002;83(1):176–85. PubMed Aging. 2012;33(10):2291–305. PubMed PMID:
PMID: 12358741 22169203
14. Dolan PJ, Johnson GV.  A caspase cleaved form of 26. Lei Z, Brizzee C, Johnson GV. BAG3 facilitates the
tau is preferentially degraded through the autophagy clearance of endogenous tau in primary neurons.
pathway. J  Biol Chem. 2010;285(29):21978–87. Neurobiol Aging. 2015;36(1):241–8. PubMed PMID:
PubMed PMID: 20466727. Pubmed Central PMCID: 25212465. Pubmed Central PMCID: 4268232
2903354 27. Cripps D, Thomas SN, Jeng Y, Yang F, Davies P,
15. Grune T, Botzen D, Engels M, Voss P, Kaiser B, Jung Yang AJ.  Alzheimer disease-specific conformation
T, et  al. Tau protein degradation is catalyzed by the of hyperphosphorylated paired helical filament-­
ATP/ubiquitin-independent 20S proteasome under Tau is polyubiquitinated through Lys-48, Lys-11,
normal cell conditions. Arch Biochem Biophys. and Lys-6 ubiquitin conjugation. J  Biol Chem.
2010;500(2):181–8. PubMed PMID: 20478262. 2006;281(16):10825–38. PubMed PMID: 16443603
Pubmed Central PMCID: 2904402 28. McKinnon C, Tabrizi SJ. The ubiquitin-proteasome
16. Jones CL, Njomen E, Sjogren B, Dexheimer TS, Tepe system in neurodegeneration. Antioxid Redox
JJ.  Small molecule enhancement of 20S proteasome Signal. 2014;21(17):2302–21. PubMed PMID:
activity targets intrinsically disordered proteins. ACS 24437518
Chem Biol. 2017;12(9):2240–7. PubMed PMID: 29. Myeku N, Clelland CL, Emrani S, Kukushkin NV, Yu
28719185. Pubmed Central PMCID: 5720355 WH, Goldberg AL, et al. Tau-driven 26S proteasome
17. Hagestedt T, Lichtenberg B, Wille H, Mandelkow
impairment and cognitive dysfunction can be pre-
EM, Mandelkow E.  Tau protein becomes long and vented early in disease by activating cAMP-PKA sig-
stiff upon phosphorylation: correlation between naling. Nat Med. 2016;22(1):46–53. PubMed PMID:
paracrystalline structure and degree of phosphoryla- 26692334. Pubmed Central PMCID: 4787271
66 M. Tang et al.

30. Wang P, Joberty G, Buist A, Vanoosthuyse A, Stancu 43. Stavoe AKH, Holzbaur ELF.  Axonal autophagy:

IC, Vasconcelos B, et  al. Tau interactome map- mini-review for autophagy in the CNS. Neurosci Lett.
ping based identification of Otub1 as Tau deubiq- 2018 Mar 13. PubMed PMID: 29548988. Pubmed
uitinase involved in accumulation of pathological Central PMCID: 6136980.
Tau forms in  vitro and in  vivo. Acta Neuropathol. 44. Vijayan V, Verstreken P. Autophagy in the presynap-
2017;133(5):731–49. PubMed PMID: 28083634. tic compartment in health and disease. J  Cell Biol.
Pubmed Central PMCID: 5390007 2017;216(7):1895–906. PubMed PMID: 28515275.
31. Lathuiliere A, Valdes P, Papin S, Cacquevel M,
Pubmed Central PMCID: 5496617
Maclachlan C, Knott GW, et  al. Motifs in the tau 45.
Kulkarni VV, Maday S.  Compartment-specific
protein that control binding to microtubules and dynamics and functions of autophagy in neurons. Dev
aggregation determine pathological effects. Sci Neurobiol. 2018;78(3):298–310. PubMed PMID:
Rep. 2017;7(1):13556. PubMed PMID: 29051562. 29197160. Pubmed Central PMCID: 5816696
Pubmed Central PMCID: 5648870 46. Ittner A, Ittner LM.  Dendritic Tau in Alzheimer’s

32. Neklesa TK, Winkler JD, Crews CM.  Targeted pro- disease. Neuron. 2018;99(1):13–27. PubMed PMID:
tein degradation by PROTACs. Pharmacol Ther. 30001506
2017;174:138–44. PubMed PMID: 28223226 47.
Mondragon-Rodriguez S, Trillaud-Doppia E,
33. Lu M, Liu T, Jiao Q, Ji J, Tao M, Liu Y, et al. Discovery Dudilot A, Bourgeois C, Lauzon M, Leclerc N, et al.
of a Keap1-dependent peptide PROTAC to knock- Interaction of endogenous tau protein with synap-
down Tau by ubiquitination-proteasome degradation tic proteins is regulated by N-methyl-D-aspartate
pathway. Eur J Med Chem. 2018;146:251–9. PubMed receptor-­dependent tau phosphorylation. J Biol Chem.
PMID: 29407955 2012;287(38):32040–53. PubMed PMID: 22833681.
34.
Colacurcio DJ, Pensalfini A, Jiang Y, Nixon Pubmed Central PMCID: 3442535
RA.  Dysfunction of autophagy and endosomal-­ 48. Regan P, Whitcomb DJ, Cho K.  Physiological and
lysosomal pathways: roles in pathogenesis of down pathophysiological implications of synaptic Tau.
syndrome and Alzheimer’s disease. Free Radic Biol Neuroscientist. 2017;23(2):137–51. PubMed PMID:
Med. 2018;114:40–51. PubMed PMID: 28988799. 26908138
Pubmed Central PMCID: 5748263 49. Ji C, Tang M, Zeidler C, Höhfeld J, Johnson GVW.
35. Gatica D, Lahiri V, Klionsky DJ.  Cargo recognition BAG3 and synaptopodin facilitate phospho-MAPT/
and degradation by selective autophagy. Nat Cell Tau degradation via autophagy in neuronal processes.
Biol. 2018;20(3):233–42. PubMed PMID: 29476151. Autophagy. 2019;15(7):119–1213. PubMed PMID:
Pubmed Central PMCID: 6028034 307445
36. Nixon RA. The role of autophagy in neurodegenera- 50. Maday S, Holzbaur EL. Compartment-specific regu-
tive disease. Nat Med. 2013;19(8):983–97. PubMed lation of autophagy in primary neurons. J  Neurosci
PMID: 23921753 Off J  Soc Neurosci. 2016;36(22):5933–45. PubMed
37. Parzych KR, Klionsky DJ. An overview of autophagy: PMID: 27251616. Pubmed Central PMCID: 4887563
morphology, mechanism, and regulation. Antioxid 51. Rubinsztein DC, Nixon RA.  Rapamycin induces

Redox Signal. 2014;20(3):460–73. PubMed PMID: autophagic flux in neurons. Proc Natl Acad Sci U S A.
23725295. Pubmed Central PMCID: 3894687 2010;107(49):E181. author reply E2. PubMed PMID:
38. Ravikumar B, Sarkar S, Davies JE, Futter M, Garcia-­ 21115811. Pubmed Central PMCID: 3000262
Arencibia M, Green-Thompson ZW, et al. Regulation 52. Tsvetkov AS, Miller J, Arrasate M, Wong JS, Pleiss
of mammalian autophagy in physiology and patho- MA, Finkbeiner S. A small-molecule scaffold induces
physiology. Physiol Rev. 2010;90(4):1383–435. autophagy in primary neurons and protects against
PubMed PMID: 20959619 toxicity in a Huntington disease model. Proc Natl
39. Corona AK, Jackson WT. Finding the middle ground Acad Sci U S A. 2010;107(39):16982–7. PubMed
for autophagic fusion requirements. Trends Cell Biol. PMID: 20833817. Pubmed Central PMCID: 2947884
2018;28(11):869–81. PubMed PMID: 30115558. 53. Tsvetkov AS, Steven FS.  Reply to Rubinsztein and
Pubmed Central PMCID: 6197918 Nixon: potent and reliable autophagy induction in neu-
40. Deng Z, Purtell K, Lachance V, Wold MS, Chen S, rons. Proc Natl Acad Sci U S A. 2010;107(49):E182.
Yue Z.  Autophagy receptors and neurodegenerative 54. Lipinski MM, Zheng B, Lu T, Yan Z, Py BF, Ng A,
diseases. Trends Cell Biol. 2017;27(7):491–504. et  al. Genome-wide analysis reveals mechanisms
PubMed PMID: 28169082 modulating autophagy in normal brain aging and
41. Boland B, Kumar A, Lee S, Platt FM, Wegiel J, Yu in Alzheimer’s disease. Proc Natl Acad Sci U S A.
WH, et  al. Autophagy induction and autophago- 2010;107(32):14164–9. PubMed PMID: 20660724.
some clearance in neurons: relationship to autopha- Pubmed Central PMCID: 2922576
gic pathology in Alzheimer’s disease. J  Neurosci 55. Bednarski E, Lynch G. Cytosolic proteolysis of tau by
Off J  Soc Neurosci. 2008;28(27):6926–37. PubMed cathepsin D in hippocampus following suppression of
PMID: 18596167. Pubmed Central PMCID: 2676733 cathepsins B and L. J Neurochem. 1996;67(5):1846–
42. Nikoletopoulou V, Tavernarakis N.  Regulation and 55. PubMed PMID: 8863489
roles of autophagy at synapses. Trends Cell Biol. 56. Bendiske J, Bahr BA. Lysosomal activation is a com-
2018;28(8):646–61. PubMed PMID: 29731196 pensatory response against protein accumulation
5  Tau Clearance Mechanisms 67

and associated synaptopathogenesis  – an approach tivation of p62 leads to accumulation of hyperphos-

for slowing Alzheimer disease? J  Neuropathol phorylated tau and neurodegeneration. J Neurochem.
Exp Neurol. 2003;62(5):451–63. PubMed PMID: 2008;106(1):107–20. PubMed PMID: 18346206
12769185 69. Xu Y, Zhang S, Zheng H.  The cargo receptor

57. Khurana V, Elson-Schwab I, Fulga TA, Sharp KA, SQSTM1 ameliorates neurofibrillary tangle pathol-
Loewen CA, Mulkearns E, et  al. Lysosomal dys- ogy and spreading through selective targeting of
function promotes cleavage and neurotoxicity of tau pathological MAPT (microtubule associated pro-
in  vivo. PLoS Genet. 2010;6(7):e1001026. PubMed tein tau). Autophagy. 2018;5:1–16. PubMed PMID:
PMID: 20664788. Pubmed Central PMCID: 2904797 30290707
58. Wong ES, Tan JM, Soong WE, Hussein K, Nukina 70. Falcon B, Noad J, McMahon H, Randow F, Goedert
N, Dawson VL, et al. Autophagy-mediated clearance M.  Galectin-8-mediated selective autophagy pro-
of aggresomes is not a universal phenomenon. Hum tects against seeded tau aggregation. J  Biol Chem.
Mol Genet. 2008;17(16):2570–82. PubMed PMID: 2018;293(7):2438–51. PubMed PMID: 29282296.
18502787. Pubmed Central PMCID: 2722889 Pubmed Central PMCID: 5818177
59. Hamano T, Gendron TF, Causevic E, Yen SH, Lin 71. Kim S, Lee D, Song JC, Cho SJ, Yun SM, Koh YH,
WL, Isidoro C, et  al. Autophagic-lysosomal pertur- et  al. NDP52 associates with phosphorylated tau
bation enhances tau aggregation in transfectants with in brains of an Alzheimer disease mouse model.
induced wild-type tau expression. Eur J  Neurosci. Biochem Biophys Res Commun. 2014;454(1):196–
2008;27(5):1119–30. PubMed PMID: 18294209 201. PubMed PMID: 25450380
60. Congdon EE, Wu JW, Myeku N, Figueroa YH,
72. D’Agostino C, Nogalska A, Cacciottolo M, Engel
Herman M, Marinec PS, et  al. Methylthioninium WK, Askanas V.  Abnormalities of NBR1, a novel
chloride (methylene blue) induces autophagy and autophagy-associated protein, in muscle fibers of
attenuates tauopathy in vitro and in vivo. Autophagy. sporadic inclusion-body myositis. Acta Neuropathol.
2012;8(4):609–22. PubMed PMID: 22361619. 2011;122(5):627–36. PubMed PMID: 21935636
Pubmed Central PMCID: 3405840 73. Tang M, Ji C, Pallo S, Rahman I, Johnson GVW. Nrf2
61. Schaeffer V, Goedert M. Stimulation of autophagy is mediates the expression of BAG3 and autophagy
neuroprotective in a mouse model of human tauopa- cargo adaptor proteins and tau clearance in an age-­
thy. Autophagy. 2012;8(11):1686–7. PubMed PMID: dependent manner. Neurobiol Aging. 2018;63:128–
22874558. Pubmed Central PMCID: 3494601 39. PubMed PMID: 29304346. Pubmed Central
62. Schaeffer V, Lavenir I, Ozcelik S, Tolnay M, Winkler PMCID: 5801049
DT, Goedert M.  Stimulation of autophagy reduces 74. Fu H, Possenti A, Freer R, Nakano Y, Hernandez
neurodegeneration in a mouse model of human Villegas NC, Tang M, et al. A tau homeostasis
tauopathy. Brain J  Neurol. 2012;135(Pt 7):2169–77. signature is linked with the cellular and regional
PubMed PMID: 22689910. Pubmed Central PMCID: vulnerability of excitatory neurons to tau pathology.
3381726 Nat Neurosci. 2019;22(1):47–56. PubMed PMID:
63. Inoue K, Rispoli J, Kaphzan H, Klann E, Chen EI, 30559469
Kim J, et  al. Macroautophagy deficiency mediates 75. Klimek C, Kathage B, Wordehoff J, Hohfeld

age-dependent neurodegeneration through a phospho-­ J. BAG3-mediated proteostasis at a glance. J Cell Sci.
tau pathway. Mol Neurodegener. 2012;7:48. PubMed 2017;130(17):2781–8. PubMed PMID: 28808089
PMID: 22998728. Pubmed Central PMCID: 3544596 76. Ulbricht A, Arndt V, Hohfeld J.  Chaperone-assisted
64. Lippai M, Low P. The role of the selective adaptor p62 proteostasis is essential for mechanotransduc-
and ubiquitin-like proteins in autophagy. Biomed Res tion in mammalian cells. Commun Integr Biol.
Int. 2014;2014:832704. PubMed PMID: 25013806. 2013;6(4):e24925. PubMed PMID: 23986815.
Pubmed Central PMCID: 4075091 Pubmed Central PMCID: 3737759
65. Stolz A, Ernst A, Dikic I.  Cargo recognition and
77. Sturner E, Behl C.  The role of the multifunctional
trafficking in selective autophagy. Nat Cell Biol. BAG3 protein in cellular protein quality control and in
2014;16(6):495–501. PubMed PMID: 24875736 disease. Front Mol Neurosci. 2017;10:177. PubMed
66. Korac J, Schaeffer V, Kovacevic I, Clement AM,
PMID: 28680391. Pubmed Central PMCID: 5478690
Jungblut B, Behl C, et  al. Ubiquitin-independent 78. Behl C.  Breaking BAG: the co-chaperone BAG3  in
function of optineurin in autophagic clearance of pro- health and disease. Trends Pharmacol Sci.
tein aggregates. J  Cell Sci. 2013;126(Pt 2):580–92. 2016;37(8):672–88. PubMed PMID: 27162137
PubMed PMID: 23178947. Pubmed Central PMCID: 79. Harold D, Abraham R, Hollingworth P, Sims R,

3654196 Gerrish A, Hamshere ML, et al. Genome-wide asso-
67. Jo C, Gundemir S, Pritchard S, Jin YN, Rahman I, ciation study identifies variants at CLU and PICALM
Johnson GV.  Nrf2 reduces levels of phosphorylated associated with Alzheimer’s disease. Nat Genet.
tau protein by inducing autophagy adaptor protein 2009;41(10):1088–93. PubMed PMID: 19734902.
NDP52. Nat Commun. 2014;5:3496. PubMed PMID: Pubmed Central PMCID: 2845877
24667209. Pubmed Central PMCID: 3990284 80. Moreau K, Fleming A, Imarisio S, Lopez Ramirez A,
68. Ramesh Babu J, Lamar Seibenhener M, Peng J,
Mercer JL, Jimenez-Sanchez M, et al. PICALM mod-
Strom AL, Kemppainen R, Cox N, et al. Genetic inac- ulates autophagy activity and tau accumulation. Nat
68 M. Tang et al.

Commun. 2014;5:4998. PubMed PMID: 25241929. autophagy in a cell model of pathological tau aggre-
Pubmed Central PMCID: 4199285 gation. Autophagy. 2010;6(1):182–3. PubMed PMID:
81. Tekirdag K, Cuervo AM. Chaperone-mediated autoph- 20023429
agy and endosomal microautophagy: joint by a chap- 86. Caballero B, Wang Y, Diaz A, Tasset I, Juste YR,
erone. J Biol Chem. 2018;293(15):5414–24. PubMed Stiller B, et  al. Interplay of pathogenic forms of
PMID: 29247007. Pubmed Central PMCID: 5900761 human tau with different autophagic pathways. Aging
82. Chiang HL, Terlecky SR, Plant CP, Dice JF.  A role Cell. 2018;17(1). PubMed PMID: 29024336. Pubmed
for a 70-kilodalton heat shock protein in lysosomal Central PMCID: 5770880
degradation of intracellular proteins. Science. 87. Vaz-Silva J, Gomes P, Jin Q, Zhu M, Zhuravleva V,
1989;246(4928):382–5. PubMed PMID: 2799391 Quintremil S, et  al. Endolysosomal degradation of
83. Dice JF, Chiang HL. Lysosomal degradation of micro- Tau and its role in glucocorticoid-driven hippocampal
injected proteins. Revis Biol Celular. 1989;20:13–33. malfunction. EMBO J. 2018 Aug 30. PubMed PMID:
PubMed PMID: 2700096 30166454.
84. Wang Y, Garg S, Mandelkow EM, Mandelkow
88. Evans LD, Wassmer T, Fraser G, Smith J, Perkinton
E. Proteolytic processing of tau. Biochem Soc Trans. M, Billinton A, et  al. Extracellular monomeric and
2010;38(4):955–61. PubMed PMID: 20658984 aggregated Tau efficiently enter human neurons
85. Wang Y, Martinez-Vicente M, Kruger U, Kaushik S, through overlapping but distinct pathways. Cell Rep.
Wong E, Mandelkow EM, et al. Synergy and antago- 2018;22(13):3612–24. PubMed PMID: 29590627.
nism of macroautophagy and chaperone-mediated Pubmed Central PMCID: 5896171
 Mechanisms of Axonal Sorting
of Tau and Influence of the Axon 6
Initial Segment on Tau Cell Polarity

Hans Zempel and Eckhard Mandelkow

 au Is an Axonally Targeted
T expression is upregulated during neuronal differ-
Microtubule Associated Protein entiation together with tubulin [10]. In normally
in the Central Nervous System matured healthy neurons in the Central Nervous
System (CNS), Tau is predominantly present in
The protein Tau (UniProt: P10636, with isoform neuronal axons and only very little present in
F being the largest isoform in the human CNS dendrites [5, 34]. Human Tau is encoded on chro-
with 441 residues) is a microtubule-associated mosome 17q21 [40], and in the human CNS
protein (MAP). MAPs are proteins that bind to comprises six major alternatively spliced iso-
assembled microtubules, and can promote the forms, with different isoform ratios depending on
assembly of tubulin dimers into microtubules. the developmental stage, cell type, and brain
Besides Tau, other MAPs such as STOP/MAP 6 region [1, 11, 51].
(stable tubule only proteins), MAP 1a and MAP The major CNS isoforms differ by the pres-
1b are also present in neuronal axons, while MAP ence or absence of two near N-terminal inserts of
2 is exclusively present in the somatodendritic 29 residues each (N1, N2), encoded by exons 2
compartment. and 3, and by the second of four repeats (R2, 31
Tau was originally discovered by Marc residues) in the repeat domain, encoded by exon
Kirschner and colleagues in their quest for micro- 10. Domains can be subdivided into (i) the
tubule assembly-promoting factors [55]. Tau ‘assembly domain’ in the C-terminal half, com-
prising the repeat domain plus flanking regions,
which supports microtubule assembly, (ii) the
H. Zempel (*) ‘projection domain’, which does not bind to
Institute of Human Genetics, University of Cologne,
Cologne, Germany
microtubules and projects away from microtu-
bules (N-Terminal half), and (iii) the ‘proline-­rich
Faculty of Medicine, University Hospital Cologne,
Cologne, Germany
domain’ in the middle part (aa 150–240), which
contains seven PXXP motifs, an interaction motif
Center for Molecular Medicine Cologne (CMMC),
University of Cologne, Cologne, Germany
for binding proteins with SH3 domains (for
e-mail: Hans.Zempel@uk-koeln.de review see Zempel and Mandelkow [57]. Tau is a
E. Mandelkow (*)
highly soluble and natively unfolded protein
German Center for Neurodegenerative Diseases which normally resists aggregation. However,
(DZNE), Bonn, Germany repeats R2 and R3 contain two hexapeptide
CAESAR Research Center, Max-Planck-Society, motifs with increased propensity for β-structure,
Bonn, Germany which can form the nucleus for amyloid-­ like
e-mail: Eckhard.Mandelkow@dzne.de

© Springer Nature Singapore Pte Ltd. 2019 69

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_6
70 H. Zempel and E. Mandelkow

aggregation [53]. Otherwise the content of sec- restricted to soma and dendrites). The predomi-
ondary structure is unusually low [38]. nance of axonal sorting of Tau may be lost per-
manently in certain neurons in AD and other
tauopathies where Tau accumulates in the
 orting of Tau During Neuronal
S somatodendritic compartment (‘missorting’), or
Development temporarily in the brains of hibernating animals
(for review of missorting of Tau in pathological
Tau is a commonly used marker for axons and for conditions see Zempel and Mandelkow [57]).
neuronal cell polarity. Cell polarity is pronounced
for most neurons in the CNS, with usually one
long axon and several highly branched dendrites,  roposed Sorting Mechanisms
P
and a corresponding asymmetric distribution of of Tau
the cytoskeleton and other cellular components.
Although most protein synthesis in mature mam- Several models have been proposed to explain
malian neurons occurs in the somatodendritic the polarized cellular distribution of Tau. The
compartment, the axon represents up to 99% of suggested Tau sorting mechanisms can be broadly
the volume of a neuron. The asymmetry of high divided into mRNA based and preferred transla-
axonal volume versus protein synthesis in the tion in axons (mechanisms 1,2), protein turnover
soma requires efficient long-haul transport mech- based and preferred compartment specific degra-
anisms, mainly based on microtubules and their dation (mechanism 3), and preferential retention
motor proteins (kinesin, dynein). The process of and active sorting (mechanisms 4–6; for a graphi-
cell differentiation, neuritic outgrowth and estab- cal depiction of the different mechanisms see
lishment of neuronal cell polarity can be subdi- Fig. 6.1, for a summary and important references
vided into five ‘Banker’ stages [34]: (i) formation see Table 6.1).
of lamellipodia; (ii) outgrowth of minor pro-
cesses; (iii) formation and growth of the axon; 1. Axonal targeting of Tau mRNA: In a first
(iv) growth of the dendrites; and (v) maturation. attempt to understand the differential localiza-
With increasing maturation, Tau becomes sorted tion of Tau and MAP 2, the Ginzburg lab
into axons (in contrast to MAP 2, which is found that in primary rat neurons Tau mRNA

Fig. 6.1  Graphical depiction of proposed sorting ubiquitously present Tau mRNA. 3 Degradation of Tau
mechanisms for axonal targeting of Tau protein. protein specifically in dendrites, but not in axons. 4 Active
Compartments indicated from top to bottom are: axonal sorting of somatically translated Tau protein into
Dendrites, soma, axon hillock/AIS, axon, synapse. Red the axon by an AIS localized pumping function. 5 Axonal
colors indicate the localization of the presumed active enrichment of Tau protein by specific association of Tau
region of the proposed sorting mechanism. 1 Targeting of with axonal microtubules. 6 Barrier function of the AIS:
Tau mRNA into the axon hillock resulting in axonal pres- The Tau Diffusion Barrier (TDB) restricts Tau diffusion
ence of Tau protein. 2 Preferred axonal translation of out of the axon
6  Mechanisms of Axonal Sorting of Tau and Influence of the Axon Initial Segment on Tau Cell Polarity 71

Table 6.1  Overview of currently discussed sorting mechanisms resulting in axonal presence of Tau protein
Sorting mechanism Supporting arguments Counter-arguments
1. Axonal targeting Tau mRNA visible in the proximal axon [32]; Tau mRNA also identified in
of Tau mRNA 3′-UTR of Tau mRNA with axon localization sequence other compartments (soma,
[2] proximal dendrites) [3, 27]; Tau
protein expressed without
3′-UTR successfully sorted [60],
also when microinjected [17]
2. Preferred axonal 5′-UTR oligopyrimidine tract of Tau mRNA mTOR Little evidence of sufficient
translation of Tau governed protein synthesis in developing axons [37] protein synthesis in mature axons
mRNA
3. Dendritic Blocking protein degradation results in appearance of Expression of Tau with little
degradation of Tau dendritic Tau, blocking dendritic protein synthesis turnover and/or traceable Tau
prevents appearance of dendritic Tau [3] (pulse-chase) results in
redistribution of Tau from the
soma into the axon [31, 60]
4. Active axonal Tau is transported along microtubules into the axon Tau preferentially binds to axonal
sorting of [25]; microtubules in fixed cells [22];
somatically Blocking microtubule dynamics with nocodazole taxol inactivation of microtubule
translated Tau results in somatic appearance of Tau [31, 59] dynamics does not result in
somatodendritic appearance of
Tau [58, 59]
5. Axonal retention Tau preferentially binds axonal microtubules in fixed, Hypothesis not sufficiently
of Tau protein by permeabilized cells [22]; favorable kinase/phosphatase tested, indirect evidence
specific association distribution for immobilization of axonal Tau (low
of Tau with axonal KxGS phosphorylation of Tau (immobilizes Tau) in
microtubules axons [58] due to preferential localization of PP2a in
axons [61], and Tau-KxGS kinases in dendrites
(MARK, SAD [15, 23]; differential PTMs of
microtubules [33, 60]
6. Barrier function TDB retains Tau in the axon [31]; specific Knockdown TDB works both antero- (soma-­
of the AIS: The Tau of AIS components results in impaired function of the to-­axon) and retrogradely
diffusion barrier TDB [60]; establishment of AIS/TDB in parallel to (axon-to-soma) [60]
(TDB) establishment of sorting of Tau [34, 43]

localizes to the proximal region of the axon. without the 3′ UTR also results in axonal tar-
Whereas tubulin mRNA was mainly detect- geting of the injected or transfected Tau [17,
able in the soma, MAP 2 mRNA was present 60]. In developing neurons and in neurons
mainly in dendrites [32]. In neuronally differ- from non-mammals axonal protein synthesis
entiated p19 cells, an embryonic carcinoma appears to be important for several processes
cell line that retains pluripotency, the same such as axonal guidance and outgrowth (see
group later identified a cis-acting sequence at e.g. for Xenopus laevis, Cagnetta et  al. [7]);
the 3′ UTR of Tau mRNA that acts as a zip however, evidence that axonal protein synthe-
code in its targeting to the axon, and also for sis in mature mammalian neurons can main-
mRNA stabilization. Importantly, swapping tain Tau protein levels is missing.
this sequence with the dendrite targeting 3′ 2. Preferred axonal translation of Tau: In
UTR of MAP 2 resulted in dendritic targeting young primary neurons, a 5′ UTR of Tau (and
of Tau mRNA and Tau protein, as well as axo- CRMP2) regulates axonal translation of Tau
nal targeting of the normally dendritically tar- mRNA in the growing axon. This process is
geted MAP 2 mRNA and MAP 2 protein [2]. mTOR dependent, and insertion of this 5′ UTR
Other studies showed that a small fraction of upstream of the coding sequence of luciferase
Tau mRNA localizes also in the proximal also resulted in axonal localization of this
regions of dendrites [3, 27]. Further, injected marker protein when myristoylated to prevent
Tau protein as well as transfected Tau DNA diffusion out of the axon [37]. However, the
72 H. Zempel and E. Mandelkow

evidence for axonal translation in this study is evidence that Tau can slide one-dimensionally
indirect, and the study was limited to young along microtubules [46], and with predomi-
developing neurons. Recently, using microflu- nant anterograde polymerization of microtu-
idic chambers, dendritic presence and transla- bules this would enable Tau to propagate into
tion of Tau mRNA was detected [3], the axon by 1D-diffusion [52]. However,
questioning the generalizability of the study. within the axon initial segment stable micro-
Direct injection of Tau protein [17], as well as tubules are scarce, so that active Tau transport
transfection of Tau without the 5′ UTR [60] would have to overcome the microtubule gap
also results in axonal targeting of Tau, indicat- within the AIS [60]. As nocodazole acts as an
ing that preferred axonal translation of Tau inhibitor of polymerization of microtubules,
might only be important in immature neurons. rather than destroying existing microtubules,
3. Degradation of dendritic Tau: The first pro- one could assume that microtubule dynamics
tein based mechanism of axonal sorting of Tau might be essential for Tau sorting. Yet in sev-
was proposed by Hirokawa and colleagues. eral studies the microtubule stabilizer Taxol,
They observed that after 4  days of injecting which inhibits microtubule dynamic instabil-
Tau or MAP 2 into neurons, Tau remained ity, did not prevent axonal targeting of Tau
only in axons, while MAP 2 remained only in [31, 58]. Furthermore, Tau protein is able to
the soma and dendrites [17]. More recently, associate with axonal microtubules even in
specifically blocking proteasomal or autopha- fixed cells [22] (see also the following mecha-
gic degradation (with wortmannin and epoxi- nisms 5), indicating that active transport
micin) in dendrites resulted in increased mechanisms or processes depending on
presence of dendritic Tau to ~4-fold, while microtubule dynamics cannot be the only sort-
enhancement of proteasome or autophagy ing mechanisms.
activity (with trehalose and rolipram) resulted 5 . Preferred axonal retention of Tau due to
in decrease of dendritic Tau by~ 3-fold [3]. axonal binding sites: Axonal stabilization of
Compartment-specific degradation of Tau Tau or retention of Tau in axons could be
could be explained by differential interactions explained by a high affinity of Tau for axonal
with degradation pathways (proteasome, microtubules vs. low affinity for dendritic
autophagy [8, 28, 42, 54] or folding pathways microtubules, as shown with fixed permeabi-
(chaperones), depending on post-translational lized cells [22]. Swapping the microtubule
modifications (phosphorylation) or chaperone binding domain of Tau and MAP 2 resulted in
components (e.g., C-terminus of Hsc70 inter- increased axonal targeting of MAP 2 and den-
acting protein, CHIP) [9, 26]. dritic targeting of Tau [17], demonstrating
4 . Active axonal sorting of somatically trans- specific association of certain microtubule
lated Tau protein: Tau is a microtubule asso- binding domains with compartment specific
ciated protein, but can also be transported microtubules. This goes in line with the idea
along microtubules, typically at the rate of of differential compartment specific post-
slow axonal transport, 0.2–0.4  mm/day [36]. translational modifications of microtubules
Over short distances, Tau can be transported [19, 33]. Not only microtubules, but also Tau
anterogradely at the typical rates of kinesin itself could be differentially modified depend-
(1  μm/s) along microtubules [25]. ing on its localization in axons or in the
Microtubules appear to be essential for antero- somatodendritic compartment. This would in
grade sorting, as the microtubule depolymer- turn change Tau‘s interaction pattern with
izer nocodazole results in loss of anterograde microtubules and other structures. For exam-
sorting [31, 59]. Since the dwell-time of Tau ple, missorted dendritic Tau is preferentially
on microtubules is low (sub-second range), phosphorylated at the KXGS motifs (Ser262,
the protein can distribute locally by rapid dif- Ser356), which makes Tau more diffusible
fusion (“kiss and hop”, [20]). There is also and less able to bind to microtubules [45].
6  Mechanisms of Axonal Sorting of Tau and Influence of the Axon Initial Segment on Tau Cell Polarity 73

Axonal Tau is less phosphorylated at these AIS is defined by local enrichment of MAP 2
sites [58], probably because of the preferential and TRIM46 [13]. The AIS itself is the major
localization of PP2a in axons, the main phos- unit responsible for generating action potentials
phatase responsible for the dephosphorylation and is a crucial player for the establishment and
of the KXGS motifs [61]. That way, dephos- maintenance of neuronal cell polarity [43]. The
phorylated Tau could be enriched on axonal AIS was initially described to contain only
microtubules. All mechanisms described sparse microtubules [41], but was later shown to
above have been demonstrated not only in contain dense bundles of (unstable) microtu-
developing immature neurons, but also in ter- bules [21]. The different results obtained in these
minally polarized neurons, making this mech- studies can be explained by the different experi-
anism a likely contributor to axonal targeting mental cell fixation conditions.
of Tau protein. However, most of the evidence Since most protein synthesis takes place in the
is indirect, and no specific microtubule modi- cell body, Tau must transit through the AIS on its
fication has yet been identified that particu- way from the cell body to the axon. The AIS is
larly attracts Tau. enriched with cytoskeletal scaffolding proteins,
6 . Restricted Diffusion of Tau due to selective notably AnkyrinG and βIV spectrin, which
entry at the AIS: The Tau Diffusion Barrier: together cluster Nav and Kv channels and link
In our research we discovered a Tau Diffusion them to the actin cytoskeleton [4]. Apart from
Barrier (TDB) that is located within the Axon generating action potentials, the AIS is also an
Initial Segment (AIS), the proximal region of essential element of neuronal cell polarity [6].
the axon where action potentials are gener- Several studies described the barrier function of
ated. The TDB serves as a diffusion restrictor the AIS to be based either on lipids, membranes,
for Tau protein, and its efficiency roughly cor- F-actin or microtubules, depending on the com-
relates with the gradual establishment of the ponents studied [24, 31, 39, 47, 48, 56]. The
AIS [31]. This barrier was originally described internal structure of the AIS and its relationship
as a retrograde barrier (preventing axonal Tau to barrier functions remain disputed since its dis-
from diffusing back into the soma). More covery by electron microscopy in 1968 by Palay
recently we found that the TDB can differen- and colleagues, but recent studies revealed a
tially regulate anterograde trafficking of Tau dense, fibrillar/globular submembranous coat
isoforms, and that it can be disrupted by containing the classical AIS proteins (e.g.,
Alzheimer Disease-like stress conditions AnkyrinG, βIVspectrin) which covers microtu-
resulting in somatic accumulation of some bules bundles, actin rings similar to axonal actin
Tau isoforms [60]. structures, but also actin patches, and enrichment
in EB proteins [21, 29, 60].

The TDB Within the AIS Regulates

Tau Transit into the Axon  fficiency of the Tau Diffusion
E
Barrier Depends
As the TDB is currently the best hypothesis for on the Composition of the AIS
regulation of axonal entry of Tau protein, and as
the TDB precisely colocalizes with the AIS [31, The TDB is precisely located within the AIS, and
60], we will in the next section describe possible functionally depends on particular microtubule
influences of the AIS on the polarized distribu- dynamics. Thus, microtubule destabilization by
tion of Tau. nocodazol results in Tau missorting and break-
The Axon Initial Segment (AIS) is localized down of the TDB, whereas microtubule
between the cell body/axon hillock and the axon ­stabilization by taxol prevents Tau missorting,
proper and stretches over 20–50  μm [29]. e.g. after Aβ-insult [31, 58]. Microtubules within
Structurally, the area between the soma and the the AIS are extremely dynamic: In the AIS, mark-
74 H. Zempel and E. Mandelkow

ers of stable microtubules, such as acetylation essential for normal function of the TDB, and
and polyglutamylation are rare, and microtubules particular constellation of microtubule dynamics
cannot be stained without using molecular densi- and f-actin promote the anterograde targeting of
fier polyethylene glycol (PEG) and taxol [21, 60]. Tau [60]. By contrast, few studies have addressed
EB3-based live-imaging of growing microtubule whether Tau also modulates AIS properties. This
tips further reveals a hub of EB3-comet spawning possibility is suggested by the many “atypical”
within the AIS, consistent with observations of functions of Tau (besides stabilizing axonal
enrichment of EB-proteins in the AIS [30, 60]. microtubules), such as regulating presynaptic or
Knockdown of important structural AIS com- postsynaptic functions in neurotransmitter shut-
ponents results in partially impaired neuronal cell tling [18, 35], cargo and vesicle release in the
polarity, e.g. with the appearance of spines at the presynapse, or stabilizing DNA integrity (for
proximal axonal shaft [16]. We recently knocked review see Sotiropoulos et  al. [49]). A direct
down some key components of the AIS and modulation of neuronal activity in the axon itself
investigated the effects on Tau distribution and was until recently elusive.
the TDB [60]. This revealed that knockdown of The Gan laboratory showed recently that Tau,
structural components of the AIS results in weak- when acetylated at certain Lysines (K274 and
ening of the TDB. Tau transit through the TDB is K281), destabilizes the AIS via reduction of
facilitated when classical AIS components (e.g. important structural elements of the AIS, namely
AnkG, EB1) are knocked down via shRNA. All AnkyrinG and βIV-spectrin, similar to what
knockdowns also induced increased invasion of occurs in the brain of Alzheimer Disease patients.
the dendritic marker MAP 2 into the proximal Overexpression of Tau mimicking acetylated Tau
part of the axon. This indicates that the classical also resulted in increased microtubule dynamics
AIS components contribute to the TDB, but also in the AIS and impairment of the TDB, both con-
to the nonspecific filter function of the AIS. firmed by live-imaging, resulting in somatic
The protein kinase GSK3β, a major player accumulation of Tau [47].
anchored within the AIS [50] is also an impor- The AIS localization influences the activity
tant kinase for the phosphorylation of Tau, but and excitability of neurons [12]. Hatch and col-
does not phosphorylate the KXGS motifs of leagues showed that Tau phosphorylated at spe-
Tau, which regulate microtubule binding. cific sites (12E8 and AT180 epitopes), induced a
GSK3β induces accumulation of endogenous distal shift of AnkyrinG localization in transfected
and exogenous Tau in the somatodendritic com- primary neurons, which also correlated with a
partment, and thus impairs the normal Tau cel- decrease in excitability of the cells. The same was
lular distribution. This effect is independent of true for neurons from young rTg4510 and old pR5
direct phosphorylation of Tau by GSK3β, mice, models that show neurodegeneration due to
because even non-­phosphorylatable Tau shows overexpression of FTLD-derived P301L-Tau [14].
the same retrograde propagation from the axon It is thus clear that at least certain components
into the soma. This indicates that the interplay (AnkyrnG, EB1, microtubule dynamics) of the
of GSK3β with components of the AIS/TDB is AIS influence the TDB and overall Tau sorting
essential for the maintenance of proper TDB into the axon. In overexpression models of patho-
function and cellular polarity. logically modified Tau (phosphorylation, acety-
lation), Tau is also able to influence AIS structure
and function. Of note, there is also evidence sug-
 odulation of AIS Plasticity by Tau
M gesting that antibody mediated reduction of
Expression and Phosphorylation AnkyrnG in healthy aged individuals and in
actively vaccinated AD-model mice protects
We described so far that the AIS is instrumental from Alzheimer Disease, indicating that at least
in establishing neuronal cell polarity [43], that structural integrity of the AIS is not a prerequisite
certain structural elements (EB1, AnkyrinG) are for preserved cognitive function during aging
6  Mechanisms of Axonal Sorting of Tau and Influence of the Axon Initial Segment on Tau Cell Polarity 75

[44]. Future studies will have to show whether 7. Cagnetta R, Frese CK, Shigeoka T, Krijgsveld J, Holt
CE.  Rapid cue-specific remodeling of the nascent
Tau can influence AIS properties also when not axonal proteome. Neuron. 2018;99:29–46.
overexpressed, and whether this change contrib- 8. Chesser AS, Pritchard SM, Johnson GVW.  Tau
utes to AIS and cognitive function. clearance mechanisms and their possible role in the
pathogenesis of Alzheimer disease. Front Neurol.
2013;4:1–12.
9. Dickey CA, Yue M, Lin WL, Dickson DW, Dunmore
Conclusion JH, Lee WC, Zehr C, West G, Cao S, Clark AM, et al.
Deletion of the ubiquitin ligase CHIP leads to the
Several mechanisms for preferred axonal targeting accumulation, but not the aggregation, of both endog-
enous phospho- and caspase-3-cleaved tau species.
of Tau have been proposed, both mRNA and pro- J Neurosci. 2006;26:6985–96.
tein based (for a summary see Table  6.1 and 10. Drubin DG, Kirschner MW.  Tau protein function in
Fig.  6.1). Most of these mechanisms have been living cells. J Cell Biol. 1986;103:2739–46.
addressed in separate studies, using cells of differ- 11. Goedert M, Spillantini MG, Jakes R, Rutherford

D, Crowther RA.  Multiple isoforms of human
ent origins and of different developmental (Banker-) microtubule-­ associated protein tau: sequences and
stages. While the individual evidence supporting localization in neurofibrillary tangles of Alzheimer’s
the different mechanisms is strong, no study has disease. Neuron. 1989;3:519–26.
addressed the real contribution of each mechanism 12. Grubb MS, Burrone J. Activity-dependent relocation
of the axon initial segment fine-tunes neuronal excit-
in each developmental stage. In more mature neu- ability. Nature. 2010;465:1070–4.
rons with full establishment of the AIS the Tau- 13. Gumy LF, Hoogenraad CC. Local mechanisms regu-
Diffusion-Barrier within the AIS plays a major role lating selective cargo entry and long-range trafficking
in regulating the transit of Tau through the AIS. in axons. Curr Opin Neurobiol. 2018;51:23–8.
14. Hatch RJ, Wei Y, Xia D, Götz J. Hyperphosphorylated
tau causes reduced hippocampal CA1 excitability by
Acknowledgments This work was supported by the relocating the axon initial segment. Acta Neuropathol.
Cologne Fortune Program/Faculty of Medicine, 2017;133:717–30.
University of Cologne (HZ) and the German Center for 15. Hayashi K, Suzuki A, Ohno S.  PAR-1/MARK: a

Neurodegenerative Diseases (DZNE) and CAESAR/ kinase essential for maintaining the dynamic state of
Max-Planck-Society (MPG) (EM). We are grateful to microtubules. Cell Struct Funct. 2011;37:21–5.
Julia Lüdtke for excellent technical assistance and Dr. 16. Hedstrom KL, Ogawa Y, Rasband MN. AnkyrinG is
Eva-Maria Mandelkow for stimulating discussions required for maintenance of the axon initial segment
throughout this project. and neuronal polarity. J Cell Biol. 2008;183:635–40.
17. Hirokawa N, Funakoshi T, Sato-Harada R, Kanai

Y.  Selective stabilization of tau in axons and
microtubule-­ associated protein 2C in cell bodies
References and dendrites contributes to polarized localization of
cytoskeletal proteins in mature neurons. J Cell Biol.
1. Andreadis A.  Tau splicing and the intricacies of 1996;132:667–79.
dementia. J Cell Physiol. 2012;227:1220–5. 18. Ittner L, Ke YD, Delerue F, Bi M, Gladbach A, van
2. Aronov S, Aranda G, Behar L, Ginzburg I.  Axonal Eersel J, Wölfing H, Chieng BC, Christie MJ, Napier
tau mRNA localization coincides with tau protein in IA, Eckert A, Staufenbiel M, Hardeman E, Götz
living neuronal cells and depends on axonal targeting J.  Dendritic function of tau mediates amyloid-beta
signal. J Neurosci. 2001;21:6577–87. toxicity in Alzheimer’s disease mouse models. Cell.
3. Balaji V, Kaniyappan S, Mandelkow E, Wang 2010;142:387–97.
Y, Mandelkow EM.  Pathological missorting of 19. Janke C, Kneussel M. Tubulin post-translational modi-
endogenous MAPT/Tau in neurons caused by fail- fications: encoding functions on the ­neuronal microtu-
ure of protein degradation systems. Autophagy. bule cytoskeleton. Trends Neurosci. 2010;33:362–72.
2018;14(12):2139–54. 20. Janning D, Igaev M, Sundermann F, Bruhmann J,
4. Bender KJ, Trussell LO. The physiology of the axon Beutel O, Heinisch J, Bakota L, Piehler J, Junge W,
initial segment. Annu Rev Neurosci. 2012;35:249–65. Brandt R. Single-molecule tracking of tau reveals fast
5. Binder LI, Frankfurter A, Rebhun LI.  The distribu- kiss-and-hop interaction with microtubules in living
tion of tau in the mammalian central nervous system. neurons. Mol Biol Cell. 2014;25:3541–51.
J Cell Biol. 1985;101:1371–8. 21. Jones SL, Korobova F, Svitkina T.  Axon initial seg-
6. Buffington SA, Rasband MN.  The axon initial seg- ment cytoskeleton comprises a multiprotein submem-
ment in nervous system disease and injury. Eur branous coat containing sparse actin filaments. J Cell
J Neurosci. 2011;34:1609–19. Biol. 2014;205:67–81.
76 H. Zempel and E. Mandelkow

22. Kanai Y, Hirokawa N. Sorting mechanisms of tau and 38. Mukrasch MD, Bibow S, Korukottu J, Jeganathan S,
MAP 2 in neurons: suppressed axonal transit of MAP Biernat J, Griesinger C, Mandelkow E, Zweckstetter
2 and locally regulated microtubule binding. Neuron. M.  Structural polymorphism of 441-residue tau at
1995;14:421–32. single residue resolution. PLoS Biol. 2009;7:e34.
23. Kishi M, Pan YA, Crump JG, Sanes JR. Mammalian 39. Nakata T, Hirokawa N.  Microtubules provide direc-
SAD kinases are required for neuronal polarization. tional cues for polarized axonal transport through
Science. 2005;307:929–32. interaction with kinesin motor head. J  Cell Biol.
24. Kobayashi T, Storrie B, Simons K, Dotti CG. A func- 2003;162:1045–55.
tional barrier to movement of lipids in polarized neu- 40. Neve RL, Harris P, Kosik KS, Kurnit DM, Donlon
rons. Nature. 1992;359:647–50. TA.  Identification of cDNA clones for the human
25. Konzack S, Thies E, Marx A, Mandelkow EM,
microtubule-associated protein tau and chromosomal
Mandelkow E.  Swimming against the tide: mobility localization of the genes for tau and microtubule-­
of the microtubule-associated protein tau in neurons. associated protein 2. Brain Res. 1986;387:271–80.
J Neurosci. 2007;27:9916–27. 41. Palay SL, Sotelo C, Peters A, Orkand PM.  The

26. Kosik KS, Shimura H.  Phosphorylated tau and the axon hillock and the initial segment. J  Cell Biol.
neurodegenerative foldopathies. Biochim Biophys 1968;38:193–201.
Acta. 2005;1739:298–310. 42. Quinn JP, Corbett NJ, Kellett KAB, Hooper NM. Tau
27. Kosik KS, Crandall JE, Mufson EJ, Neve RL.  Tau proteolysis in the pathogenesis of tauopathies: neuro-
in situ hybridization in normal and Alzheimer brain: toxic fragments and novel biomarkers. J Alzheimers
localization in the somatodendritic compartment. Ann Dis. 2018;63(1):1–21.
Neurol. 1989;26:352–61. 43. Rasband MN.  The axon initial segment and the

28. Lee MJ, Lee JH, Rubinsztein DC. Tau degradation: the maintenance of neuronal polarity. Nat Rev Neurosci.
ubiquitin-proteasome system versus the autophagy-­ 2010;11:552–62.
lysosome system. Prog Neurobiol. 2013;105:49–59. 44. Santuccione AC, Merlini M, Shetty A, Tackenberg C,
29. Leterrier C.  The axon initial segment: an updated Bali J, Ferretti MT, McAfoose J, Kulic L, Bernreuther
viewpoint. J Neurosci. 2018;38(9):2135–45. C, Welt T, et  al. Active vaccination with Ankyrin G
30. Leterrier C, Vacher H, Fache MP, d’Ortoli SA, Castets reduces β-amyloid pathology in APP transgenic mice.
F, Autillo-Touati A, Dargent B. End-binding proteins Mol Psychiatry. 2013;18:358–68.
EB3 and EB1 link microtubules to Ankyrin G in 45. Schneider A, Biernat J, von Bergen M, Mandelkow
the axon initial segment. Proc Natl Acad Sci U S A. E, Mandelkow EM.  Phosphorylation that detaches
2011;108:8826–31. tau protein from microtubules (Ser262, Ser214) also
31. Li X, Kumar Y, Zempel H, Mandelkow EM, Biernat protects it against aggregation into Alzheimer paired
J, Mandelkow E.  Novel diffusion barrier for axonal helical filaments. Biochemistry. 1999;38:3549–58.
retention of Tau in neurons and its failure in neurode- 46. Scholz T, Mandelkow E. Transport and diffusion of Tau
generation. EMBO J. 2011;30:4825–37. protein in neurons. Cell Mol Life Sci. 2014;71:3139–50.
32.
Litman P, Barg J, Rindzoonski L, Ginzburg 47. Sohn PD, Tracy TE, Son HI, Zhou Y, Leite REP, Miller
I. Subcellular localization of tau mRNA in differenti- BL, Seeley WW, Grinberg LT, Gan L. Acetylated tau
ating neuronal cell culture: implications for neuronal destabilizes the cytoskeleton in the axon initial seg-
polarity. Neuron. 1993;10:627–38. ment and is mislocalized to the somatodendritic com-
33. Magiera MM, Singh P, Gadadhar S, Janke C. Tubulin partment. Mol Neurodegener. 2016;11:47.
posttranslational modifications and emerging links to 48. Song AH, Wang D, Chen G, Li Y, Luo J, Duan S, Poo
human disease. Cell. 2018;173:1323–7. MM. A selective filter for cytoplasmic transport at the
34. Mandell JW, Banker GA.  The microtubule cyto-
axon initial segment. Cell. 2009;136:1148–60.
skeleton and the development of neuronal polarity. 49. Sotiropoulos I, Galas MC, Silva JM, Skoulakis

Neurobiol Aging. 1995;16:229–38. E, Wegmann S, Maina MB, Blum D, Sayas CL,
35. McInnes J, Wierda K, Snellinx A, Bounti L, Wang Mandelkow EM, Mandelkow E, et  al. Atypical,
YC, Stancu IC, Apostolo N, Gevaert K, Dewachter ­non-­standard functions of the microtubule associated
I, Spires-Jones TL, De Strooper B, De Wit J, Zhou Tau protein. Acta Neuropathol Commun. 2017;5:91.
L, Verstreken P.  Synaptogyrin-3 mediates pre- 50. Tapia M, Del Puerto A, Puime A, Sanchez-Ponce D,
synaptic dysfunction induced by Tau. Neuron. Fronzaroli-Molinieres L, Pallas-Bazarra N, Carlier
2018;97:823–35. E, Giraud P, Debanne D, Wandosell F, et  al. GSK3
36. Mercken M, Fischer I, Kosik KS, Nixon RA.  Three and beta-catenin determines functional expression of
distinct axonal transport rates for tau, tubulin, and sodium channels at the axon initial segment. Cell Mol
other microtubule-associated proteins: evidence for Life Sci. 2013;70:105–20.
dynamic interactions of tau with microtubules in vivo. 51. Trabzuni D, Wray S, Vandrovcova J, Ramasamy A,
J Neurosci. 1995;15:8259–67. Walker R, Smith C, Luk C, Gibbs JR, Dillman A,
37. Morita T, Sobue K.  Specification of neuronal polar- Hernandez DG, et al. MAPT expression and splicing
ity regulated by local translation of CRMP2 and is differentially regulated by brain region: relation to
Tau via the mTOR-p70S6K pathway. J  Biol Chem. genotype and implication for tauopathies. Hum Mol
2009;284:27734–45. Genet. 2012;21:4094–103.
6  Mechanisms of Axonal Sorting of Tau and Influence of the Axon Initial Segment on Tau Cell Polarity 77

52. Van Den Bedem H, Kuhl E.  Tau-ism: the Yin and 58. Zempel H, Thies E, Mandelkow E, Mandelkow

Yang of microtubule sliding, detachment, and rupture. EM.  Abeta oligomers cause localized Ca(2+) eleva-
Biophys J. 2015;109:2215–7. tion, missorting of endogenous Tau into dendrites,
53. von Bergen M, Friedhoff P, Biernat J, Heberle J,
Tau phosphorylation, and destruction of microtubules
Mandelkow EM, Mandelkow E. Assembly of tau pro- and spines. J Neurosci. 2010;30:11938–50.
tein into Alzheimer paired helical filaments depends 59. Zempel H, Luedtke J, Kumar Y, Biernat J, Dawson
on a local sequence motif ((306)VQIVYK(311)) H, Mandelkow E, Mandelkow EM.  Amyloid-beta
forming beta structure. Proc Natl Acad Sci U S A. oligomers induce synaptic damage via Tau-dependent
2000;97:5129–34. microtubule severing by TTLL6 and spastin. EMBO
54. Wang Y, Mandelkow E. Tau in physiology and pathol- J. 2013;32:2920–37.
ogy. Nat Rev Neurosci. 2016;17:5–21. 60. Zempel H, Dennissen FJA, Kumar Y, Luedtke

55. Weingarten MD, Lockwood AH, Hwo SY, Kirschner J, Biernat J, Mandelkow EM, Mandelkow
MW. A protein factor essential for microtubule assem- E.  Axodendritic sorting and pathological missort-
bly. Proc Natl Acad Sci U S A. 1975;72:1858–62. ing of Tau are isoform-­specific and determined by
56. Winckler B, Forscher P, Mellman I. A diffusion bar- axon initial segment architecture. J  Biol Chem.
rier maintains distribution of membrane proteins in 2017;292:12192–207.
polarized neurons. Nature. 1999;397:698–701. 61. Zhu LQ, Zheng HY, Peng CX, Liu D, Li HL, Wang
57. Zempel H, Mandelkow E. Lost after translation: mis- Q, Wang JZ. Protein phosphatase 2A facilitates axo-
sorting of Tau protein and consequences for Alzheimer nogenesis by dephosphorylating CRMP2. J Neurosci.
disease. Trends Neurosci. 2014;37:721–32. 2010;30:3839–48.
 Part II
Tau Localization and Function
 Tau and Axonal Transport
Misregulation in Tauopathies 7
Benjamin Combs, Rebecca L. Mueller,
Gerardo Morfini, Scott T. Brady,
and Nicholas M. Kanaan

Introduction where communication with target cells takes

place. Such provision relies on the trafficking of
Neurons are unique cells because of their mor- material to and from the synapse by a complex set
phology and polarity, with the cell body extending of intracellular trafficking events collectively
two types of processes. These are the dendrites referred to as fast axonal transport (FAT)
that primarily receive signals from other cells and (reviewed in [6]). Microtubule motor-­based FAT
the surrounding environment and the axons that along axons is critical to the function and health
are long processes connecting neurons to their tar- of neurons, delivering organelles, vesicles, and
get cells. The longest neuronal process, the axon, other cellular materials that ultimately support
can extend up to a meter in length from the cell communication with target cells. FAT is also nec-
body, in the case of some motor and sensory neu- essary for returning damaged material to the cell
rons. One of the unique challenges of the neuron body for recycling, and for delivering neuro-
is the provision of materials to distant synapses, trophic signals received at the axon terminal to the
cell body, where they can affect gene transcription
Authors Benjamin Combs and Rebecca L. Mueller have (reviewed in [67]). The importance of this process
equally contributed to this chapter. is highlighted by examples of neurodegeneration
following its disruption. Mutations in the genes
B. Combs encoding proteins involved in transport along the
Department of Translational Neuroscience, College axon (including molecular motors, cytoskeletal
of Human Medicine, Michigan State University,
Grand Rapids, MI, USA tracks, and adaptor proteins) cause several neuro-
e-mail: combsben@msu.edu developmental and neurodegenerative diseases
R. L. Mueller (reviewed in [65, 79]). The tau protein is among
Department of Translational Neuroscience, College those implicated in affecting FAT in a group of
of Human Medicine, Michigan State University,
Grand Rapids, MI, USA
Neuroscience Program, Michigan State University, N. M. Kanaan (*)
East Lansing, MI, USA Department of Translational Neuroscience, College
e-mail: dangrem1@msu.edu of Human Medicine, Michigan State University,
G. Morfini · S. T. Brady Grand Rapids, MI, USA
Department of Anatomy and Cell Biology, University Neuroscience Program, Michigan State University,
of Illinois at Chicago, Chicago, IL, USA East Lansing, MI, USA
Marine Biological Laboratory, Hauenstein Neuroscience Center, Mercy Health Saint
Woods Hole, MA, USA Mary’s, Grand Rapids, MI, USA
e-mail: gmorfini@uic.edu; stbrady@uic.edu e-mail: nkanaan@msu.edu
© Springer Nature Singapore Pte Ltd. 2019 81
A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_7
82 B. Combs et al.

neurodegenerative diseases known as tauopathies tau was heavily phosphorylated in these inclu-
(reviewed in [42, 45]). In this chapter, we will dis- sions [29, 30, 107]. Interest in the tau protein has
cuss the relationship between tau and FAT under continued to grow as a central role in AD and sev-
normal conditions and how disruptions to this eral other neurodegenerative disorders became
process may be a toxic mechanism in multiple apparent. The discovery of inherited mutations
neurodegenerative diseases. within the tau gene that lead to early-onset fron-
totemporal dementias (FTDs) was the first line of
evidence demonstrating that pathological tau is
Tau Protein and Disease sufficient to cause neurodegenerative disease
[35]. In these diseases, tau undergoes a number
From its initial discovery in 1975, the tau protein of pathological post-translational modifications
was closely associated with microtubules and is (reviewed in [55, 84]) and forms a variety of mor-
somewhat enriched in axons [5, 105]. It is pri- phologically different aggregates that range from
marily expressed in the brain with higher levels small oligomers to much larger filamentous
of mRNA and protein in the cortex and hippo- inclusions (reviewed in [46]). Pathological tau is
campus than white matter and the cerebellum highly modified compared to its normal state
[95]. The protein exists as six major isoforms in with some of the most prominent modifications
the adult human brain, generated through alterna- including increased levels of phosphorylation,
tive mRNA splicing. The isoforms differ in their changes in overall conformation, and truncation
inclusion of 2, 1 or 0 of exons 2 and 3  in the of the protein, among many others [7, 15, 29, 31,
N-terminus and their inclusion of exon 10 which 34]. These modified versions of tau accumulate
determines whether the protein contains either 4 in the somatodendritic and axonal compartments
or 3 microtubule-binding repeat regions (+exon of neurons and often display impaired microtu-
10 = 4R tau, −exon 10 = 3R tau; Fig. 7.1a). The bule binding [8, 12, 28, 82], fueling the notion
microtubule-binding regions in the C-terminal that reduced stability of the microtubule cyto-
half of the protein are positively-charged, while skeleton represents a critical pathogenic event in
the N-terminus is enriched in acidic amino acids tauopathies. However, almost 50  years after its
leaving the protein isoforms with a relatively low discovery, the exact mechanisms of tau toxicity
net charge. Directly upstream of the microtubule-­ continue to be debated.
binding domains is a proline-rich region which Some potential clues may be found by eluci-
includes several phosphorylation sites and PXXP dating tau’s normal cellular functions as well as
motifs that may bind to SH3 domains (Fig. 7.1b). examining the neurodegenerative phenotypes
Other functional domains include a phosphatase-­ associated with tauopathies. As mentioned
activating domain (PAD, discussed in greater above, tau was traditionally linked to microtu-
detail below) at the extreme N-terminus [40]. bule-based functions involving stabilization and
Structural studies indicate that tau is a highly dynamics. However, the protein’s putative func-
dynamic protein, capable of multiple conforma- tions have expanded to include regulation of
tions that may underlie its diverse role in multiple FAT, scaffolding for phosphotransferases, syn-
cellular functions. The protein contains regions aptic plasticity, neuronal activity, actin bundling,
of secondary structure and acquires global con- and mediating interactions between cellular
formations such as the “paperclip”, which components (reviewed in [45, 103]). These
involves the N- and C-termini interacting with functions may be reliant on the local tau isoform
each other and the C-terminus interacting with composition as well as specific sets of regula-
the microtubule-binding repeat regions [38, 69]. tory post-­ translational modifications. Thus,
Approximately a decade after the discovery of many, if not all, of these potential functions of
tau, a series of studies demonstrated that tau was tau are likely affected in disease as the protein
the primary component of the hallmark tangle undergoes abnormal modification, misfolding
pathology in Alzheimer’s disease (AD) and that and aggregation.
7  Tau and Axonal Transport Misregulation in Tauopathies 83

Fig. 7.1  Tau isoforms and selected modifications. (a) FTDP-17 mutations can lead to inherited early-­onset
In the adult human brain tau primarily exists as six iso- frontotemporal dementias (reviewed in [53]). Several
forms generated through alternative mRNA splicing. of these mutations have been linked to FAT dysfunc-
The isoforms differ based on the inclusion of exons 2 tion including A152T, P301L, K369I, and R406W
and 3  in the N-terminus of the protein (2  N, 1  N, or (red). Functional domains associated with transport
0 N) and exon 10. Exon 10 contains the second of four include the phosphatase activating domain (yellow), a
potential microtubule-­binding repeat regions. Isoforms motif that is conformationally displayed in disease-
are referred to as 3R or 4R based on the number of related forms of tau and linked to changes in signaling
repeat regions they contain. (b) Modifications to the pathways that regulate transport. The tau molecule
protein can affect its function and induce dysfunction contains many phosphorylation sites throughout the
in disease. Some selected modifications discussed here sequence, some of which are found in healthy tissues
include phosphorylation sites (green) at tyrosine 18 and others are associated with tau pathology.
and the AT8 sites (serine 202 and threonine 205). (Reviewed in [84])

Axonal Transport Transport in the axon occurs along the axonal

microtubules [49] while actin microfilaments
Projection neurons affected in tauopathies face a and neurofilaments have roles that are more
unique challenge in maintaining their health and structural; although microfilaments are impor-
functional abilities due to their cellular architec- tant in short distance movements such as secre-
ture. On one end of a neuron is the somatoden- tion and neurotransmitter release. Two major
dritic compartment, the primary location of protein molecular motor families are responsible for
synthesis and degradation, and on the other end are bidirectional transport along microtubules in the
the synapses where inter-neuronal communication axon, kinesins and dyneins (Fig. 7.2). The kine-
takes place. The transport of material in both direc- sin superfamily is organized into 14 families,
tions is therefore vital to the cell and defects in this from which 46 kinesins are expressed in the
process contribute to a variety of neurodegenera- human brain (reviewed in [59]). Conventional
tive disorders (reviewed in [9]). kinesin, or kinesin-­ 1, is the most abundant
84 B. Combs et al.

Fig. 7.2  Neurons depend on robust microtubule-based driven anterograde transport is necessary for the delivery
transport in axons. A healthy, functional neuron is depen- of synaptic components, including mitochondria and ves-
dent on the molecular motor complexes kinesin (a) and icles, to the axon terminal where they will aid in cell sig-
dynein (b), whose roles are to transport material along naling as well as delivery of channels to axon to support
microtubules in the plus- (anterograde) or minus-end (ret- propagations of the action potential. Dynein-driven retro-
rograde) direction, respectively. Materials synthesized in grade transport is necessary for the transport of signaling
the soma (e.g., cytoskeletal components, mitochondria endosomes and material undergoing breakdown and recy-
and membrane-bound organelles) rely on kinesin for their cling, like damaged mitochondria, multivesicular bodies
delivery to the correct axonal compartment (c). Kinesin-­ and lysosomes (d) back to the neuronal soma

member of the kinesin superfamily and exists as ficking appears to be mediated by conventional
a heterotetramer composed of two heavy chains kinesin [3, 10, 88]. Cytoplasmic dynein is a
(involved in ATP and microtubule binding and large multisubunit (two proteins each of a heavy
movement) and two light chains (involved in chain, intermediate chain, light intermediate
cargo binding) that produce plus-end directed chain, and three light chains) motor complex
movement of cargoes toward the axon terminal responsible for minus-­end directed movement,
[20], a process known as anterograde fast axo- or retrograde FAT (rFAT) (reviewed in [76]). In
nal transport (aFAT) (reviewed in [67]). Other contrast to the FAT of membrane-bound organ-
members of the kinesin superfamily, kinesin-2, elles, other cytoplasmic and cytoskeletal pro-
kinesin-3 and kinesin-4 may also be involved in teins move through slow axonal transport, a
a subset of aFAT [75, 86], but the bulk of traf- process that likely involves kinesin-1 and dynein
7  Tau and Axonal Transport Misregulation in Tauopathies 85

but is not fully understood yet ([51, 98] and Axonal Degeneration
reviewed in [78]). in Tauopathies
Normal FAT is regulated through a compli-
cated system involving the composition of differ- There are 27 different tauopathies described to date
ent motor protein complexes and regulatory including Alzheimer’s disease (AD), chronic trau-
post-translational modifications of the motor matic encephalopathy (CTE), Pick’s disease (PiD),
complexes (e.g. phosphorylation) (reviewed in progressive supranuclear palsy (PSP), frontotem-
[9, 67]). Several studies have suggested that vari- poral dementia with parkinsonism linked to chro-
ation in kinesin-1 motor protein subunits can lead mosome 17 (FTDP-17), and corticobasal
to specificity in the cargoes being transported degeneration (CBD) (reviewed in [53]). Tauopathies
[20, 90, 92]. Phosphorylation is the best studied typically display pathological features consistent
post-translational modification in the context of with a “dying back” pattern of neuronal degenera-
regulating multisubunit motor complexes. tion. These include dystrophic axons and spher-
Studies in various experimental systems, most oids, as well as evidence of disrupted FAT (Fig. 7.3)
notably the isolated squid axoplasm preparation (reviewed in [42, 45]). Accordingly, synaptic dys-
[44, 85], revealed that signaling pathways involv- function and loss also occur very early in disease
ing several kinases or phosphatases can tightly and synaptic loss correlates closely with cognitive
regulate FAT by inhibiting binding to microtu- deficits in AD and other tauopathies [21]. Studies
bules or promoting cargo dissociation [19, 61, using post-­mortem human tissue samples suggest
62]. In fact, several regulatory kinase and phos- that tau inclusions appear as neuropil threads or
phatase pathways for FAT are disrupted in tauop- dystrophic neurites early during the progressive
athies and other neurodegenerative diseases accumulation of tau pathology in AD brains [25,
providing another potential link between FAT, 47]. Advanced brain imaging studies confirm
loss of axonal connectivity, and neuronal degen- degeneration of white matter containing axonal
eration (reviewed in [42, 65]). projections during the mild cognitive impairment

Fig. 7.3  Dystrophic axons containing tau pathology is a ogy is seen in severe Alzheimer’s disease brains (b; AD
prominent feature in multiple tauopathies. The TNT1 anti- (Braak stage V-VI). C-F) Axonal tau pathology in the sub-
body detects exposure of the phosphatase activating cortical white matter displays PAD exposure (i.e. TNT1
domain (PAD). (a, b) TNT1 pathology-containing axons reactivity) in chronic traumatic encephalopathy (c; CTE),
are observed in the subcortical white matter in non-­ Pick’s disease (d; PiD), progressive supranuclear palsy (e;
demented aged patients with early stages of tau deposition PSP) and corticobasal degeneration (f; CBD) as well.
(a; ND (Braak stage I-II)) and robust TNT1 axonal pathol- Scale bar is 50 μm
86 B. Combs et al.

stage, prior to onset of AD, and continued degen-

based transport of organelles along axons. The
eration in other regions with disease progression.
squid isolated axoplasm model of FAT was
Tau pathology and degeneration are also prominent
instrumental in several significant discoveries of
features of other tauopathies. In PSP, pathological
the biology of FAT, including the initial discov-
tau is observed in neuropil threads and in subcorti-
ery of conventional kinesin, the most abundant
cal white matter along with dystrophic axons andmember of the kinesin superfamily [11]. In this
spheroids indicative of FAT disruptions and degen-
model, the giant axons are removed and the axo-
eration in the same regions [2, 33, 36, 70]. Tau-
plasm extruded, which allows the study of axon-­
positive neuropil threads are also observed in CBD
autonomous molecular events including FAT [44,
and closely associated with white matter degenera-
85]. Video-enhanced contrast-differential inter-
tion detected by advanced imaging [36, 111]. Pick
ference contrast microscopy is used to visualize
bodies are the defining feature of PiD but tau and measure the velocity of membrane-bound
pathology can also be detected in axonal terminals
organelle cargoes undergoing FAT along micro-
along with neuritic threads and axonal spheroids in
tubules [85]. Over the past several years the squid
mossy fibers and cerebellar white matter [74]. axoplasm assay also was used to elucidate a num-
Again, this is associated with degeneration of white
ber of signaling pathways regulating aFAT and
matter regions that can be quite severe [108]. CTE
rFAT through different mechanisms. For exam-
brains present with neuropil threads and axonal ple, activation of the p38α MAPK pathway
atrophy within subcortical white matter and trans-
results in inhibition of kinesin-based aFAT, while
ported proteins accumulate in axons after traumatic
activation of CK2 or JNK3 causes inhibition of
brain injuries [48, 56, 99]. both aFAT and rFAT [61, 64, 66, 72]. The squid
Together, these and other studies demonstrateaxoplasm preparation was particularly important
that the most common tauopathies are character- in elucidating tau’s role in regulating FAT through
ized by several features that point to a significant
the PP1-GSK3 signaling pathway. This pathway
role for axonal dysfunction that may be caused involves activation of PP1 and subsequent
by deficits in FAT. Pathological tau modifications
dephosphorylation of Ser9  in GSK3β, an event
and axonal degeneration are closely associated that results in activation of this kinase (Fig. 7.4)
with each other, beginning with the earliest stages
[62]. Active GSK3 regulates aFAT by phosphory-
of tauopathy pathogenesis. Axonal swellings and lating kinesin light chains and causing release of
accumulation of vesicular organelles also suggest
conventional kinesin holoenzymes from their
abnormalities in FAT.  Collectively, these obser-
transported cargo [61, 97]. Importantly, several
vations suggest that pathological forms of tau disease-relevant pathological modifications of
may exert toxic effects through disruption of nor-
tau were shown to disrupt FAT through altera-
mal processes in the axon. In fact, as discussedtions of this signaling cascade [40, 50].
below, evidence from several model systems sup- Initial studies using the squid axoplasm
ports this hypothesis and indicates that regulation
focused on testing effects of tau monomers on
of FAT appears to be disturbed by several disease-­
FAT, which had been proposed to compete with
associated pathological changes to tau. conventional kinesin for microtubule binding
[81]. Introduction of soluble tau monomers at
concentrations from 1-25 μM tau had no apparent
Tau-Based Effects on Fast Axonal effect on FAT in either direction, while much
Transport lower concentration of a dynein microtubule-­
binding domain peptide blocked both directions
Tau and Regulatory Signaling [63]. These studies raised questions about
Pathways for Fast Axonal Transport whether tau aggregates would affect
FAT.  Recombinant wild-type tau aggregates (a
Accumulating evidence clearly implicates a mixed population of oligomers and filaments)
number of kinase and phosphatase signaling specifically inhibited aFAT but did not affect
pathways in the regulation of motor complex-­ rFAT, while monomeric tau had no effect on FAT
7  Tau and Axonal Transport Misregulation in Tauopathies 87

Fig. 7.4 Tau alters kinesin-cargo interactions through tion or specific phosphorylation events that can aber-
modulation of a PP1-GSK3β signaling pathway. Tau rantly expose the PAD epitope. In the proposed model,
contains a phosphatase-activating domain (PAD) within these forms of tau can disrupt normal kinesin-based
amino acids 2–18 at the extreme N-terminus of tau transport by activating protein phosphatase 1 (PP1).
(shown in red). Under normal conditions this epitope is This in turn dephosphorylates an inhibitory phosphate
obscured allowing for normal kinesin-based transport on GSK3β in order to activate it. GSK3β then phosphor-
along the microtubules. In disease conditions tau can ylates kinesin light chain inducing a release of cargo
undergo a variety of modifications including aggrega- and disruption of FAT

in either direction [18, 50]. The toxic effect of directly linked to a specific molecular pathway of
aggregated tau on aFAT was blocked when tau toxicity (i.e. the PP1-GSK3 pathway) that results
aggregates were co-perfused into isolated axo- in FAT impairment. Tau contains three putative
plasm with either PP1-specific or GSK3-specific RVxF PP1-interacting domains (including one
inhibitors demonstrating activation of a PP1-­ within PAD) and is able to localize the phospha-
GSK3 pathway as the underlying mechanism tase to microtubules, which is consistent with a
[40, 50]. Additional work found that tau-­mediated role for tau in regulating this pathway [24, 52].
transport toxicity was dependent upon a 17 amino The biological activity of the PAD suggests that
acid motif in the extreme amino terminus of tau physiological changes in tau could lead to tightly
(i.e. aa 2–18), which was since termed the regulated exposure, which may allow delivery of
phosphatase-­ activating domain (PAD). selected organelles at specific axonal subdomains
Additional experiments showed that the PAD was under normal conditions (reviewed in [67]).
sufficient to disrupt aFAT through this pathway Further studies tested the PAD-dependent
[40, 50]. Relevant to tauopathies, the PAD epit- mechanism of toxicity using two specific tau
ope was normally sequestered in control tau, but modifications that were predicted to facilitate
exposed in pathological forms of tau. This was PAD exposure by impairing tau paperclip folding.
true for both isolated tau and tau aggregates in First, a key set of studies showed that phospho-
vitro [40] and in vivo for all tauopathies exam- mimicking triple phosphorylation at AT8 sites in
ined to date [16, 17, 43]. Collectively, this body tau (i.e. Ser199, Ser202 and Thr205) disrupted the
of work provided a new framework for under- paperclip conformation causing extension of the
standing tau-mediated toxicity. Specifically, extreme amino-terminus [39] and subsequently
disease-­associated tau modifications (e.g. aggre- PAD exposure. Consistent with enhanced PAD
gation, specific phosphorylation events, etc.) exposure, when applied to squid axoplasm the
alter tau structure in a way that leads to phosphomimetic AT8 tau m ­ onomers disrupted
conformation-­ dependent exposure of aFAT [40]. Second, an FTDP-17 mutation was
PAD.  Subsequently, aberrant PAD exposure is described that causes deletion of exons 6–9 com-
88 B. Combs et al.

prising the first microtubule-­binding region and aFAT defects that were proposed to be associated
the proline-rich region, the domain of tau that with tau interacting with JIP1 [37]. However, this
allows flexibility for the N-terminus to fold onto suggestion was based on overexpression of
the C-terminus in the paperclip conformation mutant tau and cannot rule out a role for PAD
[77]. The prediction that the monomeric form of exposure in the FAT changes. Expression of
this protein would significantly impair aFAT was another FTDP-17 mutant tau, A152T, in C. ele-
confirmed in squid axoplasm as well [40]. A role gans neurons led to abnormal localization of syn-
of tau-mediated FAT toxicity in AD was consis- aptic vesicles that may be due to minor disruptions
tent with studies in mammalian neurons, where in aFAT and rFAT independent of any tau aggre-
toxic effects of oligomeric Aβ on FAT were gation or changes to microtubule binding [14,
dependent on both tau and GSK3β [102]. Finally, 73]. These studies do not identify a specific sig-
PAD-dependent impairments in synaptic trans- naling pathway, but the conclusions may be simi-
mission were found in the squid giant synapse lar to observed impairment of aFAT and rFAT by
[60] and other studies demonstrate oligomeric tau tau filaments phosphorylated at S422 [93], a
(a multimeric form with increased PAD exposure) disease-­specific phosphorylation event that
is toxic to synaptic function suggesting PAD occurs early in pre-tangles and robustly labels
exposure may disrupt both axonal and synaptic neuropil threads in areas of the brain involved in
functions in neurons. These studies not only con- memory and cognition [32, 94]. Such observa-
firmed this novel PAD-mediated mechanism of tions suggest that certain modifications in tau
tau toxicity but also indicated that modifications may expose PAD and other domain(s) that in turn
of monomeric tau can confer toxicity indepen- activate alternative signaling pathways.
dently of aggregation.
Human tissue studies further confirmed this
working model of tau toxicity in multiple tauopa-  au May Physically Interfere
T
thies, demonstrating that tau species shown to with Kinesin Binding to Microtubules
impair FAT manifest early during disease progres-
sion. For example, the TNT1 and TNT2 antibod- Several studies reported that increased levels of
ies are markers of conformation-dependent PAD tau can alter the behavior of motor proteins, while
exposure (Fig.  7.3) [17, 40]. These antibodies others have failed to see such effects.
label the earliest forms of tau deposition, pre-­ Overexpression of fluorescently-tagged tau in pri-
tangles, and robustly label neuropil threads early mary neurons was reported to inhibit aFAT of
in AD as well as the hallmark pathologies in AD amyloid precursor protein (APP) and similar
and other tauopathies. Importantly, these antibod- effects on kinesin-based transport of mitochon-
ies show little to no detection of normal tau in dria were observed in other cell lines [23, 87]. In
post-mortem human brains or in non-denaturing complementary in vitro studies, the presence of
in vitro assays [16–18, 40] (see Fig. 7.3). This pat- tau reduced the processivity of multiple motor
tern of staining is very similar to the robust label- proteins but did not affect their overall speed [96].
ing of pre-tangle inclusions and neuropil threads All of these effects were proposed to occur
observed early during disease progression with through tau-based hindrance of kinesin’s binding
the AT8 phospho-tau antibody [8, 40] and TOC1, sites [54]. A similar mechanism was proposed
a tau oligomer-specific antibody [18, 71, 104]. based on tau’s interference with kinesin activity
The conformation- and disease-specific labeling and dynein reversals on stabilized microtubules
with these specific tau antibodies suggests that [22]. However, other studies have not supported
changes in the global conformation of tau leads to this hypothesis. The kinesin- and dynein-binding
exposure of the PAD motif early in disease. sites on microtubules are overlapping, so a­ ddition
Activation of other pathways by mutant forms of the dynein microtubule-binding domain effec-
of tau have also been proposed. Transgenic mice tively blocks both kinesin- and dynein-based
expressing a K369I FTDP-17 mutation displayed motility [63]. In contrast, levels of tau as high as 1
7  Tau and Axonal Transport Misregulation in Tauopathies 89

tau per tubulin dimer had no effect on FAT in the isoform in an in vitro assay, an effect that was
squid axoplasm preparation, providing strong evi- dependent upon the number of motor proteins
dence against the notion that tau competes with bound to the beads and was interpreted to be a
kinesin for microtubule binding [63]. Also, tau result of tau reducing kinesin binding to microtu-
binding to microtubules did not alter kinesin bules [100]. Therefore, the local tau isoform com-
speed or run length and only marginally affected position was suggested to act as a regulatory
microtubule-binding at tau:tubulin concentration factor in influencing cargo travel and final desti-
ratios much higher than physiological levels [50, nations within the axon, although a mechanism
63, 81]. Studies in tau-overexpressing transgenic for creating such differential distributions of tau
mice also found that increasing levels of normal isoforms in cells remains to be identified.
tau had no effect on kinesin-based transport in The functional implications of having different
vivo [110]. Efforts to explain these discordant tau isoforms remain unclear and the pathogno-
results suggested that tau may interact differen- monic inclusions of different tauopathies are typi-
tially with microtubules under varying conditions. cally composed of specific isoforms. For example,
For example, the nucleotide-binding state of AD and CTE are a mixture of 4R and 3R inclu-
microtubules were reported to alter tau effects as sions, while PSP and CBD are primarily 4R inclu-
tau inhibited kinesin-based transport on GDP-­ sions and PiD is primarily a 3R inclusion disease
microtubules but not on GTP-microtubules and in [13, 27, 83, 109]. To evaluate the effects of differ-
an isoform-dependent manner [57]. In this case ent tau isoforms, preparations of monomeric and
the shortest 3R tau induced a more severe pheno- aggregated tau were generated for each of the 6
type than the longest 4R isoform. However, it is isoforms found in human CNS.  Although there
difficult to find a physiological role for these were differences in their relative toxicity, aggre-
effects and the high levels of tau and the distinc- gated forms of the six tau isoforms similarly
tive conditions involved are not known to exist in impaired aFAT in the squid axoplasm preparation,
any disease condition. Overall, definitive evidence suggesting the PAD-dependent PP1-GSK3 mech-
for direct physical interference of tau with micro- anism of toxicity may be a common element
tubule-based motor proteins in vivo is lacking, but among tauopathies independent of the tau iso-
more complicated interactions under non-physio- forms that comprise the pathology [18]. Further
logical conditions cannot be ruled out. work is required to fully appreciate the normal
and pathological functions of each tau isoform.

 au Isoforms Differentially Affect Fast

T
Axonal Transport  ther Mechanisms of Fast Axonal
O
Transport Regulation by Tau
Other studies have also examined isoform-­specific
effects of tau. Expression of wild-type human 3R Tau may also alter FAT through direct effects on
tau, but not 4R tau, induced an accumulation of microtubule organization. Tau, in an isoform-­
vesicles in the axons of Drosophila larva motor specific manner, may act as a microtubule-spacer
neurons [80]. This effect was exacerbated by tau that simultaneously bundles microtubules and
phosphorylation and reversed upon inhibition of prevents them from overcrowding which could
GSK3β, potentially implicating it in the mecha- facilitate transport under normal cellular condi-
nistic pathway [68]. In human induced pluripotent tions [58]. Microtubule-based effects were also
stem cell (iPSC) dopaminergic cultures, shRNA- examined by overexpressing or deleting the fly
mediated knockdown of only 4R isoforms reduced homologue of human tau in Drosophila melano-
the average velocity of axonal mitochondria com- gaster neurons [91]. Increasing tau expression
pared to control and total tau knockdown neurons levels resulted in increased pause times of vesi-
[4]. The shortest 3R tau isoform shortened kinesin cles being transported in aFAT and rFAT.  This
run length to a greater degree than the longest 4R was associated with changes to microtubule den-
90 B. Combs et al.

sity and axon caliber as well, which may have Phosphorylation at this site may affect other
influenced FAT.  Unlike humans, flies do not mechanisms of tau-mediated transport regula-
express redundant MAPs so it is difficult to know tion. For example, pseudophosphorylation at
how these findings translate to mammalian neu- Tyr18 rescued inhibition of kinesin-1 motility
rons. Other studies have failed to see a significant induced by the shortest 3R isoform [89].
effect of tau overexpression on either direction
transport in mouse axons [110].
Tau may also mediate effects of other AD-related Conclusion
molecules. In  addition to tau pathology, diseases
such as AD are characterized by the accumulation Based upon the evidence presented in this chap-
of pathological amyloid-β (Aβ) peptides in plaques ter, pathological forms of the tau protein found in
[106]. Treatment of primary neurons with Aβ oligo- AD and other tauopathies may induce toxicity by
mers inhibited FAT of mitochondria and TrkA and disrupting FAT. Tau modifications, such as con-
the effect was dependent upon presence of tau formational changes and increased phosphoryla-
[101]. A cross of MAPTP301L mice, an FTDP-17 tau tion, are detected within axons very early in the
line, with TgCRND8, a mouse line expressing course of disease progression. In addition, the
mutant amyloid precursor protein, resulted in more initial signs of neurodegeneration typically occur
severe decrease of mitochondrial transport while in with synaptic dysfunction followed by a “dying
the tau line alone aFAT was slightly increased back” degeneration of axons. Given that patho-
before decreasing with age [1, 26]. Synergistic logical tau and neurodegeneration first occur in
effects of oligomeric Aβ have been noted in a vari- the axon, it is natural to ask if the protein may
ety of other studies as well. In particular, GSK3 exert toxic effects through disruption of a critical
phosphorylation of kinesin light chains requires a cellular process like FAT.  In fact, many patho-
priming phosphorylation event [61], which is effi- logical forms of tau were found to disrupt FAT
ciently provided via activation of CK2 by oligo- across multiple model systems ranging from in
meric Aβ [72]. The result of activating both vitro biochemical systems to squid axoplasm and
pathways at the same time would exacerbate inhi- mouse models. Several potential mechanisms of
bition of FAT in AD (reviewed in [9]). tau toxicity, not mutually exclusive, are proposed
Reinforcing the notion of tau playing a role and supported experimentally. Given the discov-
in FAT regulation, a post-translational modifica- ery of biological activity for the  PAD, it is rea-
tion within the  PAD was shown to modulate sonable to speculate that tau may help regulate
tau’s effects. Specifically, phosphorylation at FAT under normal conditions and exert toxicity
Tyr18 rescued PAD-mediated aFAT impairment in disease through a hypermorphic gain-of-­
in the isolated squid axoplasm model [41]. function mechanism. Tau’s normal function may
Further, phospho-Tyr18 co-localized with TNT1 take the form of modulating signaling pathways
immunoreactive tau inclusions in human dis- that regulate FAT by modulating kinesin-cargo
ease [41]. Thus, Tyr18 phosphorylation may interactions and/or kinesin-microtubule interac-
represent a normal regulatory mechanism that tions. Further studies will provide a better under-
attenuates tau’s ability to cause dissociation of standing of tau’s normal and pathological roles,
kinesin from cargoes for delivery, and in patho- potentially providing a specific mechanistic
logical contexts may be neuroprotective and framework for the development of effective ther-
block the aberrant activation the PP1-GSK3 apeutic strategies to treat tauopathies.
pathway caused by PAD-exposed tau species.
Accordingly, surviving neurons at late AD Acknowledgments  Supported by NIH grants (AG044372
stages display strong immunoreactivity when and NS082730), the BrightFocus Foundation, the Secchia
Family Foundation and the Rainwater Foundation.
using an anti-phospho-Tyr18 antibody [41].
7  Tau and Axonal Transport Misregulation in Tauopathies 91

References Mandelkow EM, Encalada SE, Kao AW. Tau/MAPT

disease-associated variant A152T alters tau function
and toxicity via impaired retrograde axonal trans-
1. Adalbert R, Milde S, Durrant C, Ando K, Stygelbout
port. Hum Mol Genet. 2018;
V, Yilmaz Z, Gould S, Brion JP, Coleman
15. Carmel G, Mager EM, Binder LI, Kuret J. The struc-
MP.  Interaction between a MAPT variant causing
tural basis of monoclonal antibody Alz50’s selectiv-
frontotemporal dementia and mutant APP affects
ity for Alzheimer’s disease pathology. J Biol Chem.
axonal transport. Neurobiol Aging. 2018;68:68–75.
1996;271:32789–95.
2. Ahmed Z, Josephs KA, Gonzalez J, DelleDonne
16. Combs B, Kanaan NM.  Exposure of the amino
A, Dickson DW.  Clinical and neuropathologic fea-
terminus of tau is a pathological event in multiple
tures of progressive supranuclear palsy with severe
Tauopathies. Am J Pathol. 2017;187:1222–9.
pallido-nigro-luysial degeneration and axonal dys-
17. Combs B, Hamel C, Kanaan NM.  Pathological
trophy. Brain. 2008;131:460–72.
conformations involving the amino terminus of tau
3. Amaratunga A, Morin PJ, Kosik KS, Fine
occur early in Alzheimer’s disease and are differen-
RE. Inhibition of kinesin synthesis and rapid antero-
tially detected by monoclonal antibodies. Neurobiol
grade axonal transport in vivo by an antisense oligo-
Dis. 2016;94:18–31.
nucleotide. J Biol Chem. 1993;268:17427–30.
18. Cox K, Combs B, Abdelmesih B, Morfini G, Brady
4. Beevers JE, Lai MC, Collins E, Booth HDE, Zambon
ST, Kanaan NM.  Analysis of isoform-specific
F, Parkkinen L, Vowles J, Cowley SA, Wade-Martins
tau aggregates suggests a common toxic mecha-
R, Caffrey TM.  MAPT genetic variation and neu-
nism involving similar pathological conformations
ronal maturity Alter isoform expression affecting
and axonal transport inhibition. Neurobiol Aging.
axonal transport in iPSC-derived dopamine neurons.
2016;47:113–26.
Stem Cell Rep. 2017;9:587–99.
19. DeBerg HA, Blehm BH, Sheung J, Thompson AR,
5. Binder LI, Frankfurter A, Rebhun LI. The distribu-
Bookwalter CS, Torabi SF, Schroer TA, Berger CL,
tion of tau in the mammalian central nervous system.
Lu Y, Trybus KM, Selvin PR. Motor domain phos-
J Cell Biol. 1985;101:1371–8.
phorylation modulates kinesin-1 transport. J  Biol
6. Black MM.  Axonal transport: the orderly
Chem. 2013;288:32612–21.
motion of axonal structures. Methods Cell Biol.
20. DeBoer SR, You Y, Szodorai A, Kaminska A, Pigino
2016;131:1–19.
G, Nwabuisi E, Wang B, Estrada-Hernandez T, Kins
7. Braak H, Braak E. Staging of Alzheimer’s disease-­
S, Brady ST, Morfini G. Conventional kinesin holo-
related neurofibrillary changes. Neurobiol Aging.
enzymes are composed of heavy and light chain
1995;16:271–8.. discussion 278-284
homodimers. Biochemistry. 2008;47:4535–43.
8. Braak E, Braak H, Mandelkow EM. A sequence of
21. DeKosky ST, Scheff SW. Synapse loss in frontal cor-
cytoskeleton changes related to the formation of
tex biopsies in Alzheimer’s disease: correlation with
neurofibrillary tangles and neuropil threads. Acta
cognitive severity. Ann Neurol. 1990;27:457–64.
Neuropathol (Berl). 1994;87:554–67.
22. Dixit R, Ross JL, Goldman YE, Holzbaur
9. Brady ST, Morfini GA. Regulation of motor proteins,
EL.  Differential regulation of dynein and kinesin
axonal transport deficits and adult-onset neurodegen-
motor proteins by tau. Science. 2008;319:1086–9.
erative diseases. Neurobiol Dis. 2017;105:273–82.
23. Ebneth A, Godemann R, Stamer K, Illenberger S,
10. Brady ST, Sperry AO.  Biochemical and functional
Trinczek B, Mandelkow E.  Overexpression of tau
diversity of microtubule motors in the nervous sys-
protein inhibits kinesin-dependent trafficking of
tem. Curr Opin Neurobiol. 1995;5:551–8.
vesicles, mitochondria, and endoplasmic reticulum:
11. Brady ST, Lasek RJ, Allen RD.  Fast axonal trans-
implications for Alzheimer’s disease. J  Cell Biol.
port in extruded axoplasm from squid giant axon.
1998;143:777–94.
Science. 1982;218:1129–31.
24. Egloff MP, Johnson DF, Moorhead G, Cohen
12. Bramblett GT, Goedert M, Jakes R, Merrick SE,
PT, Cohen P, Barford D.  Structural basis for the
Trojanowski JQ, Lee VM. Abnormal tau phosphory-
recognition of regulatory subunits by the cata-
­
lation at Ser396 in Alzheimer’s disease recapitulates
lytic subunit of protein phosphatase 1. EMBO
development and contributes to reduced microtubule
J. 1997;16:1876–87.
binding. Neuron. 1993;10:1089–99.
25. Ghoshal N, Garcia-Sierra F, Wuu J, Leurgans S,
13. Buee L, Delacourte A. Comparative biochemistry of
Bennett DA, Berry RW, Binder LI.  Tau conforma-
tau in progressive supranuclear palsy, corticobasal
tional changes correspond to impairments of epi-
degeneration, FTDP-17 and Pick’s disease. Brain
sodic memory in mild cognitive impairment and
Pathol. 1999;9:681–93.
Alzheimer’s disease. Exp Neurol. 2002;177:475–93.
14. Butler VJ, Salazar DA, Soriano-Castell D, Alves-­
26. Gilley J, Seereeram A, Ando K, Mosely S, Andrews
Ferreira M, Dennissen FJA, Vohra M, Oses-Prieto
S, Kerschensteiner M, Misgeld T, Brion JP, Anderton
JA, Li KH, Wang AL, Jing B, Li B, Groisman A,
B, Hanger DP, Coleman MP.  Age-dependent axo-
Gutierrez E, Mooney S, Burlingame AL, Ashrafi K,
92 B. Combs et al.

nal transport and locomotor changes and tau hypo- induces a compaction of the paperclip folding of tau
phosphorylation in a “P301L” tau knockin mouse. and generates a pathological (MC-1) conformation.
Neurobiol Aging. 2012;33:621 e621–15. J Biol Chem. 2008;283:32066–76.
27. Goedert M, Spillantini MG, Cairns NJ, Crowther 40. Kanaan NM, Morfini GA, LaPointe NE, Pigino GF,
RA.  Tau proteins of Alzheimer paired helical fila- Patterson KR, Song Y, Andreadis A, Fu Y, Brady ST,
ments: abnormal phosphorylation of all six brain Binder LI. Pathogenic forms of tau inhibit kinesin-­
isoforms. Neuron. 1992;8:159–68. dependent axonal transport through a mechanism
28. Gong CX, Lidsky T, Wegiel J, Zuck L, Grundke-­ involving activation of axonal phosphotransferases.
Iqbal I, Iqbal K.  Phosphorylation of microtubule-­ J Neurosci. 2011;31:9858–68.
associated protein tau is regulated by protein 41. Kanaan NM, Morfini G, Pigino G, LaPointe NE,
phosphatase 2A in mammalian brain. Implications Andreadis A, Song Y, Leitman E, Binder LI, Brady
for neurofibrillary degeneration in Alzheimer’s dis- ST.  Phosphorylation in the amino terminus of tau
ease. J Biol Chem. 2000;275:5535–44. prevents inhibition of anterograde axonal transport.
29. Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M, Neurobiol Aging. 2012;33(826):e815–30.
Wisniewski HM, Binder LI. Abnormal phosphoryla- 42. Kanaan NM, Pigino GF, Brady ST, Lazarov O,
tion of the microtubule-associated protein tau (tau) Binder LI, Morfini GA.  Axonal degeneration in
in Alzheimer cytoskeletal pathology. Proc Natl Acad Alzheimer’s disease: when signaling abnormali-
Sci U S A. 1986a;83:4913–7. ties meet the axonal transport system. Exp Neurol.
30. Grundke-Iqbal I, Iqbal K, Quinlan M, Tung YC, 2013;246:44–53.
Zaidi MS, Wisniewski HM. Microtubule-associated 43. Kanaan NM, Cox K, Alvarez VE, Stein TD, Poncil
protein tau. A component of Alzheimer paired heli- S, McKee AC.  Characterization of early patho-
cal filaments. J Biol Chem. 1986b;261:6084–9. logical tau conformations and phosphorylation in
31. Guillozet-Bongaarts AL, Garcia-Sierra F, Reynolds chronic traumatic encephalopathy. J  Neuropathol
MR, Horowitz PM, Fu Y, Wang T, Cahill ME, Bigio Exp Neurol. 2016;75:19–34.
EH, Berry RW, Binder LI.  Tau truncation during 44. Kang M, Baker L, Song Y, Brady ST, Morfini
neurofibrillary tangle evolution in Alzheimer’s dis- G.  Biochemical analysis of axon-specific phos-
ease. Neurobiol Aging. 2005;26:1015–22. phorylation events using isolated squid axoplasms.
32. Guillozet-Bongaarts AL, Cahill ME, Cryns Methods Cell Biol. 2016;131:199–216.
VL, Reynolds MR, Berry RW, Binder 45. Kneynsberg A, Combs B, Christensen K, Morfini G,
LI. Pseudophosphorylation of tau at serine 422 inhib- Kanaan NM.  Axonal degeneration in Tauopathies:
its caspase cleavage: in vitro evidence and implica- disease relevance and underlying mechanisms. Front
tions for tangle formation in  vivo. J  Neurochem. Neurosci. 2017;11:572.
2006;97:1005–14. 46. Kovacs GG.  Invited review: neuropathology of
33. Hauw JJ, Verny M, Delaere P, Cervera P, He Y, tauopathies: principles and practice. Neuropathol
Duyckaerts C.  Constant neurofibrillary changes in Appl Neurobiol. 2015;41:3–23.
the neocortex in progressive supranuclear palsy. 47. Kowall NW, Kosik KS. Axonal disruption and aber-
Basic differences with Alzheimer’s disease and rant localization of tau protein characterize the neu-
aging. Neurosci Lett. 1990;119:182–6. ropil pathology of Alzheimer’s disease. Ann Neurol.
34. Horowitz PM, Patterson KR, Guillozet-Bongaarts 1987;22:639–43.
AL, Reynolds MR, Carroll CA, Weintraub ST, 48. Kraus MF, Susmaras T, Caughlin BP, Walker CJ,
Bennett DA, Cryns VL, Berry RW, Binder LI. Early Sweeney JA, Little DM. White matter integrity and
N-terminal changes and caspase-6 cleavage of tau in cognition in chronic traumatic brain injury: a diffu-
Alzheimer’s disease. J Neurosci. 2004;24:7895–902. sion tensor imaging study. Brain. 2007;130:2508–19.
35. Hutton M, et  al. Association of missense and 49. Kreutzberg GW.  Neuronal dynamics and axonal
5′-splice-site mutations in tau with the inherited flow. IV.  Blockage of intra-axonal enzyme trans-
dementia FTDP-17. Nature. 1998;393:702–5. port by colchicine. Proc Natl Acad Sci U S A.
36. Ikeda K, Akiyama H, Haga C, Kondo H, Arima K, 1969;62:722–8.
Oda T. Argyrophilic thread-like structure in cortico- 50. LaPointe NE, Morfini G, Pigino G, Gaisina IN,
basal degeneration and supranuclear palsy. Neurosci Kozikowski AP, Binder LI, Brady ST. The amino ter-
Lett. 1994;174:157–9. minus of tau inhibits kinesin-dependent ­axonal trans-
37. Ittner LM, Ke YD, Gotz J. Phosphorylated tau inter- port: implications for filament toxicity. J  Neurosci
acts with c-Jun N-terminal kinase-interacting pro- Res. 2009;87:440–51.
tein 1 (JIP1) in Alzheimer disease. J  Biol Chem. 51. Lasek RJ, Garner JA, Brady ST.  Axonal trans-
2009;284:20909–16. port of the cytoplasmic matrix. J  Cell Biol.
38. Jeganathan S, von Bergen M, Brutlach H, Steinhoff 1984;99:212s–21s.
HJ, Mandelkow E. Global hairpin folding of tau in 52. Liao H, Li Y, Brautigan DL, Gundersen GG. Protein
solution. Biochemistry. 2006;45:2283–93. phosphatase 1 is targeted to microtubules by the
39. Jeganathan S, Hascher A, Chinnathambi S, Biernat microtubule-associated protein tau. J  Biol Chem.
J, Mandelkow EM, Mandelkow E. Proline-directed 1998;273:21901–8.
pseudo-phosphorylation at AT8 and PHF1 epitopes
7  Tau and Axonal Transport Misregulation in Tauopathies 93

53. Ling H.  Untangling the tauopathies: current transport defects in neurodegenerative diseases.
concepts of tau pathology and neurodegenera- J Neurosci. 2009b;29:12776–86.
tion. Parkinsonism Relat Disord. 2018;46(Suppl 66. Morfini GA, Bosco DA, Brown H, Gatto R, Kaminska
1):S34–8. A, Song Y, Molla L, Baker L, Marangoni MN, Berth
54. Mandelkow EM, Stamer K, Vogel R, Thies E, S, Tavassoli E, Bagnato C, Tiwari A, Hayward LJ,
Mandelkow E.  Clogging of axons by tau, inhibi- Pigino GF, Watterson DM, Huang CF, Banker G,
tion of axonal traffic and starvation of synapses. Brown RH Jr, Brady ST.  Inhibition of fast axonal
Neurobiol Aging. 2003;24:1079–85. transport by pathogenic SOD1 involves activation of
55. Martin L, Latypova X, Terro F.  Post-translational p38 MAP kinase. PLoS One. 2013;8:e65235.
modifications of tau protein: implications 67. Morfini G, Schmidt N, Weissmann C, Pigino G, Kins
for Alzheimer’s disease. Neurochem Int. S. Conventional kinesin: biochemical heterogeneity
2011;58:458–71. and functional implications in health and disease.
56. McKee AC, Cantu RC, Nowinski CJ, Hedley-Whyte Brain Res Bull. 2016;126:347–53.
ET, Gavett BE, Budson AE, Santini VE, Lee HS, 68. Mudher A, Shepherd D, Newman TA, Mildren P,
Kubilus CA, Stern RA.  Chronic traumatic enceph- Jukes JP, Squire A, Mears A, Drummond JA, Berg S,
alopathy in athletes: progressive tauopathy after MacKay D, Asuni AA, Bhat R, Lovestone S. GSK-­
repetitive head injury. J  Neuropathol Exp Neurol. 3beta inhibition reverses axonal transport defects
2009;68:709–35. and behavioural phenotypes in Drosophila. Mol
57. McVicker DP, Chrin LR, Berger CL.  The Psychiatry. 2004;9:522–30.
nucleotide-­binding state of microtubules modu- 69. Mukrasch MD, Bibow S, Korukottu J, Jeganathan S,
lates kinesin processivity and the ability of tau to Biernat J, Griesinger C, Mandelkow E, Zweckstetter
inhibit kinesin-mediated transport. J  Biol Chem. M.  Structural polymorphism of 441-residue tau at
2011;286:42873–80. single residue resolution. PLoS Biol. 2009;7:e34.
58. Mephon-Gaspard A, Boca M, Pioche-Durieu C, 70. Padovani A, Borroni B, Brambati SM, Agosti C,
Desforges B, Burgo A, Hamon L, Pietrement O, Broli M, Alonso R, Scifo P, Bellelli G, Alberici A,
Pastre D. Role of tau in the spatial organization of Gasparotti R, Perani D.  Diffusion tensor imaging
axonal microtubules: keeping parallel microtubules and voxel based morphometry study in early pro-
evenly distributed despite macromolecular crowd- gressive supranuclear palsy. J  Neurol Neurosurg
ing. Cell Mol Life Sci. 2016;73:3745–60. Psychiatry. 2006;77:457–63.
59. Miki H, Okada Y, Hirokawa N. Analysis of the kine- 71. Patterson KR, Remmers C, Fu Y, Brooker S,
sin superfamily: insights into structure and function. Kanaan NM, Vana L, Ward S, Reyes JF, Philibert
Trends Cell Biol. 2005;15:467–76. K, Glucksman MJ, Binder LI.  Characterization of
60. Moreno H, Morfini G, Buitrago L, Ujlaki G, Choi S, prefibrillar tau oligomers in vitro and in Alzheimer
Yu E, Moreira JE, Avila J, Brady ST, Pant H, Sugimori disease. J Biol Chem. 2011;286:23063–76.
M, Llinas RR. Tau pathology-mediated presynaptic 72. Pigino G, Morfini G, Atagi Y, Deshpande A, Yu
dysfunction. Neuroscience. 2016;325:30–8. C, Jungbauer L, LaDu M, Busciglio J, Brady
61. Morfini G, Szebenyi G, Elluru R, Ratner N, Brady S. Disruption of fast axonal transport is a pathogenic
ST.  Glycogen synthase kinase 3 phosphorylates mechanism for intraneuronal amyloid beta. Proc
kinesin light chains and negatively regulates kinesin-­ Natl Acad Sci U S A. 2009;106:5907–12.
based motility. EMBO J. 2002;21:281–93. 73. Pir GJ, Choudhary B, Mandelkow E, Mandelkow
62. Morfini G, Szebenyi G, Brown H, Pant HC, Pigino EM. Tau mutant A152T, a risk factor for FTD/PSP,
G, DeBoer S, Beffert U, Brady ST. A novel CDK5-­ induces neuronal dysfunction and reduced lifes-
dependent pathway for regulating GSK3 activity pan independently of aggregation in a C. elegans
and kinesin-driven motility in neurons. EMBO Tauopathy model. Mol Neurodegener. 2016;11:33.
J. 2004;23:2235–45. 74. Probst A, Tolnay M, Langui D, Goedert M,
63. Morfini G, Pigino G, Mizuno N, Kikkawa M, Brady Spillantini MG. Pick’s disease: hyperphosphorylated
ST.  Tau binding to microtubules does not directly tau protein segregates to the somatoaxonal compart-
affect microtubule-based vesicle motility. J Neurosci ment. Acta Neuropathol (Berl). 1996;92:588–96.
Res. 2007;85:2620–30. 75. Ray K, Perez SE, Yang Z, Xu J, Ritchings BW,
64. Morfini GA, You YM, Pollema SL, Kaminska A, Liu Steller H, Goldstein LS.  Kinesin-II is required for
K, Yoshioka K, Bjorkblom B, Coffey ET, Bagnato axonal transport of choline acetyltransferase in
C, Han D, Huang CF, Banker G, Pigino G, Brady Drosophila. J Cell Biol. 1999;147:507–18.
ST. Pathogenic huntingtin inhibits fast axonal trans- 76. Reck-Peterson SL, Redwine WB, Vale RD, Carter
port by activating JNK3 and phosphorylating kine- AP.  The cytoplasmic dynein transport machin-
sin. Nat Neurosci. 2009a;12:864–71. ery and its many cargoes. Nat Rev Mol Cell Biol.
65. Morfini GA, Burns M, Binder LI, Kanaan NM, 2018;19:382–98.
LaPointe N, Bosco DA, Brown RH Jr, Brown H, 77. Rovelet-Lecrux A, Lecourtois M, Thomas-Anterion
Tiwari A, Hayward L, Edgar J, Nave KA, Garberrn C, Le Ber I, Brice A, Frebourg T, Hannequin D,
J, Atagi Y, Song Y, Pigino G, Brady ST.  Axonal Campion D.  Partial deletion of the MAPT gene:
94 B. Combs et al.

a novel mechanism of FTDP-17. Hum Mutat. 91. Talmat-Amar Y, Arribat Y, Parmentier ML. Vesicular
2009;30:E591–602. axonal transport is modified in vivo by tau deletion
78. Roy S.  Seeing the unseen: the hidden world of or overexpression in Drosophila. Int J  Mol Sci.
slow axonal transport. Neuroscientist RevJ Bring 2018;19
Neurobiol Neurol Psychiat. 2014;20:71–81. 92. Tanaka Y, Kanai Y, Okada Y, Nonaka S, Takeda
79. Roy S, Zhang B, Lee VM, Trojanowski JQ. Axonal S, Harada A, Hirokawa N.  Targeted disruption of
transport defects: a common theme in neurodegen- mouse conventional kinesin heavy chain, kif5B,
erative diseases. Acta Neuropathol. 2005;109:5–13. results in abnormal perinuclear clustering of mito-
80. Sealey MA, Vourkou E, Cowan CM, Bossing T, chondria. Cell. 1998;93:1147–58.
Quraishe S, Grammenoudi S, Skoulakis EMC, 93. Tiernan CT, Combs B, Cox K, Morfini G, Brady ST,
Mudher A. Distinct phenotypes of three-repeat and Counts SE, Kanaan NM. Pseudophosphorylation of
four-repeat human tau in a transgenic model of tau at S422 enhances SDS-stable dimer formation
tauopathy. Neurobiol Dis. 2017;105:74–83. and impairs both anterograde and retrograde fast
81. Seitz A, Kojima H, Oiwa K, Mandelkow EM, Song axonal transport. Exp Neurol. 2016;283:318–29.
YH, Mandelkow E. Single-molecule investigation of 94. Tiernan CT, Ginsberg SD, He B, Ward SM,
the interference between kinesin, tau and MAP 2c. Guillozet-Bongaarts AL, Kanaan NM, Mufson EJ,
EMBO J. 2002;21:4896–905. Counts SE.  Pretangle pathology within cholinergic
82. Sengupta A, Kabat J, Novak M, Wu Q, Grundke-­ nucleus basalis neurons coincides with neurotrophic
Iqbal I, Iqbal K. Phosphorylation of tau at both Thr and neurotransmitter receptor gene dysregulation
231 and Ser 262 is required for maximal inhibi- during the progression of Alzheimer’s disease.
tion of its binding to microtubules. Arch Biochem Neurobiol Dis. 2018;117:125–36.
Biophys. 1998;357:299–309. 95. Trabzuni D, Wray S, Vandrovcova J, Ramasamy A,
83. Sergeant N, Wattez A, Delacourte A. Neurofibrillary Walker R, Smith C, Luk C, Gibbs JR, Dillman A,
degeneration in progressive supranuclear palsy and Hernandez DG, Arepalli S, Singleton AB, Cookson
corticobasal degeneration: tau pathologies with MR, Pittman AM, de Silva R, Weale ME, Hardy J,
exclusively “exon 10” isoforms. J  Neurochem. Ryten M. MAPT expression and splicing is differen-
1999;72:1243–9. tially regulated by brain region: relation to genotype
84. Simic G, Babic Leko M, Wray S, Harrington C, and implication for tauopathies. Hum Mol Genet.
Delalle I, Jovanov-Milosevic N, Bazadona D, Buee 2012;21:4094–103.
L, de Silva R, Di Giovanni G, Wischik C, Hof 96. Trinczek B, Ebneth A, Mandelkow EM, Mandelkow
PR. Tau protein hyperphosphorylation and aggrega- E. Tau regulates the attachment/detachment but not
tion in Alzheimer’s disease and other Tauopathies, the speed of motors in microtubule-dependent trans-
and possible neuroprotective strategies. Biomol port of single vesicles and organelles. J  Cell Sci.
Ther. 2016;6:6. 1999;112(Pt 14):2355–67.
85. Song Y, Kang M, Morfini G, Brady ST. Fast axonal 97. Tsai MY, Morfini G, Szebenyi G, Brady ST. Release
transport in isolated axoplasm from the squid giant of kinesin from vesicles by hsc70 and regula-
axon. Methods Cell Biol. 2016;131:331–48. tion of fast axonal transport. Mol Biol Cell.
86. Soppina V, Verhey KJ.  The family-specific K-loop 2000;11:2161–73.
influences the microtubule on-rate but not the super- 98. Uchida A, Alami NH, Brown A.  Tight functional
processivity of kinesin-3 motors. Mol Biol Cell. coupling of kinesin-1A and dynein motors in the
2014;25:2161–70. bidirectional transport of neurofilaments. Mol Biol
87. Stamer K, Vogel R, Thies E, Mandelkow E, Cell. 2009;20:4997–5006.
Mandelkow EM.  Tau blocks traffic of organ- 99. Uryu K, Chen XH, Martinez D, Browne KD,
elles, neurofilaments, and APP vesicles in neu- Johnson VE, Graham DI, Lee VM, Trojanowski JQ,
rons and enhances oxidative stress. J  Cell Biol. Smith DH. Multiple proteins implicated in neurode-
2002;156:1051–63. generative diseases accumulate in axons after brain
88. Stenoien DL, Brady ST.  Immunochemical analy- trauma in humans. Exp Neurol. 2007;208:185–92.
sis of kinesin light chain function. Mol Biol Cell. 100. Vershinin M, Carter BC, Razafsky DS, King SJ,
1997;8:675–89. Gross SP.  Multiple-motor based transport and
89. Stern JL, Lessard DV, Hoeprich GJ, Morfini GA, its regulation by tau. Proc Natl Acad Sci U S A.
Berger CL.  Phosphoregulation of tau modulates 2007;104:87–92.
inhibition of kinesin-1 motility. Mol Biol Cell. 101. Vossel KA, Zhang K, Brodbeck J, Daub AC, Sharma
2017;28:1079–87. P, Finkbeiner S, Cui B, Mucke L.  Tau reduction
90. Szodorai A, Kuan YH, Hunzelmann S, Engel U, prevents Abeta-induced defects in axonal transport.
Sakane A, Sasaki T, Takai Y, Kirsch J, Muller U, Science. 2010;330:198.
Beyreuther K, Brady S, Morfini G, Kins S.  APP 102. Vossel KA, Xu JC, Fomenko V, Miyamoto T,
anterograde transport requires Rab3A GTPase activ- Suberbielle E, Knox JA, Ho K, Kim DH, Yu GQ,
ity for assembly of the transport vesicle. J Neurosci. Mucke L.  Tau reduction prevents Abeta-induced
2009;29:14534–44. axonal transport deficits by blocking activation of
GSK3beta. J Cell Biol. 2015;209:419–33.
7  Tau and Axonal Transport Misregulation in Tauopathies 95

103. Wang Y, Mandelkow E.  Tau in physiology and

108. Yamakawa K, Takanashi M, Watanabe M, Nakamura
pathology. Nat Rev Neurosci. 2016;17:5–21. N, Kobayashi T, Hasegawa M, Mizuno Y, Tanaka
104. Ward SM, Himmelstein DS, Lancia JK, Fu Y,
S, Mori H.  Pathological and biochemical studies
Patterson KR, Binder LI.  TOC1: characterization on a case of pick disease with severe white matter
of a selective oligomeric tau antibody. J Alzheimers atrophy. Neuropathol Off J  Jpn Soc Neuropathol.
Dis. 2013;37:593–602. 2006;26:586–91.

105. Weingarten MD, Lockwood AH, Hwo SY, 109. Yoshida M. Cellular tau pathology and immunohis-
Kirschner MW.  A protein factor essential for tochemical study of tau isoforms in sporadic tauop-
microtubule assembly. Proc Natl Acad Sci U S A. athies. Neuropathol Off J  Jpn Soc Neuropathol.
1975;72:1858–62. 2006;26:457–70.
106. Wong CW, Quaranta V, Glenner GG.  Neuritic
110. Yuan A, Kumar A, Peterhoff C, Duff K, Nixon
plaques and cerebrovascular amyloid in Alzheimer RA. Axonal transport rates in vivo are unaffected by
disease are antigenically related. Proc Natl Acad Sci tau deletion or overexpression in mice. J Neurosci.
U S A. 1985;82:8729–32. 2008;28:1682–7.

107. Wood JG, Mirra SS, Pollock NJ, Binder 111. Zhang Y, Walter R, Ng P, Luong PN, Dutt S, Heuer
LI.  Neurofibrillary tangles of Alzheimer disease H, Rojas-Rodriguez JC, Tsai R, Litvan I, Dickerson
share antigenic determinants with the axonal BC, Tartaglia MC, Rabinovici G, Miller BL, Rosen
microtubule-associated protein tau (tau) [published HJ, Schuff N, Boxer AL. Progression of microstruc-
erratum appears in Proc Natl Acad Sci U S A 1986 tural degeneration in progressive Supranuclear palsy
Dec;83(24):9773]. Proc Natl Acad Sci U S A. and Corticobasal syndrome: a longitudinal diffusion
1986;83:4040–3. tensor imaging study. PLoS One. 2016;11:e0157218.
 Presynaptic Pathophysiology
Encoded in Different Domains 8
of Tau – Hyper-Versus
Hypoexcitability?

Jochen Martin Decker and Eva-Maria Mandelkow

Introduction presynaptic sites, but unexpectedly the effects on

physiological parameters are opposite [12, 13].
Hyperphosphorylated and aggregated forms of
Tau are indicators of progressive neurodegenera-
tion in AD and several tauopathies (Braak stages Results
[1]), but certain forms of Tau can enhance toxicity
(e.g. FTDP-17 mutations [2]) and have therefore The repeat domain (RD, repeats 1–4, Fig.  8.1)
been used to generate animal models of Tau controls both the microtubule-binding and the
pathology [3, 4]. Various mechanisms of Tau-­ aggregation of Tau. This is largely due to two
induced toxicity have been proposed, operating hexapeptide motifs at the beginning of repeats R2
either in the cytosol (e.g. effects on MT-dependent and R3 [33]. The deletion mutation ΔK280
transport [5], protein degradation [6]), in the (related to frontotemporal dementia and AD, [2])
nucleus (interferences with DNA, RNA, or nuclear is one of several mutations that render Tau more
pores, [7, 8]) or at synapses [12, 13]. Loss of hip- pro-aggregant by enhancing its propensity for
pocampal synapses is an early hallmark of mem- β-structure, thus promoting Tau’s toxic effects.
ory loss in AD, and therefore much attention has Conversely, insertions of β-breaking prolines pre-
been paid to possible roles of Tau in pre- and post- vent β-structure and thus prevent aggregation and
synaptic compartments [9–11]. Our recent focus is toxicity in transgenic animals [14]. This argues
on electrophysiological effects of two mutations for a close correlation between Tau toxicity and
that lie in distinct domains of Tau and have distinct neurodegeneration. By contrast, the mutation
effects on Tau aggregation, strong (ΔK280) or A152T (a risk factor for FTD-spectrum disorders,
weak (A152T). We find that both affect mainly e.g. PSP [15]) is located in the N-terminal “pro-
jection” domain which does not bind to microtu-
J. M. Decker bules, at the beginning of the “proline-­ rich”
German Center for Neurodegenerative Diseases domain. This region interacts with signalling mol-
(DZNE), Bonn, Germany ecules containing SH3 domains (e.g. tyrosine
E.-M. Mandelkow (*) kinase fyn [16]) but is not directly involved in
German Center for Neurodegenerative Diseases aggregation (Fig.  8.1b). Both mutations cause
(DZNE), Bonn, Germany
severe changes of synaptic structures and function
Max-Planck-Institute for Metabolism Research, in the hippocampus, the “gate to memory”. But
Cologne, Germany
remarkably the effects of both mutations on syn-
Caesar Research Center, Bonn, Germany apses point in opposite directions.
e-mail: Eva.Mandelkow@dzne.de

© Springer Nature Singapore Pte Ltd. 2019 97

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_8
98 J. M. Decker and E.-M. Mandelkow

Tau-ΔK280 Tau-A152T
a b
ΔK280 A152T

Fibril formation Oligomer formation

Synaptic transmission Synaptic transmission

Fig. 8.1  Bar diagrams of human Tau (2N4R), showing repeat domain is responsible for the pathological aggrega-
domain structure, locations of Tau mutations, and tion of Tau. The mutation ΔK280 is located at the begin-
effects on synapses. (a) Full length human Tau (441 resi- ning of R2 and leads to pronounced aggregation of Tau.
dues, Uniprot 10,636–8, Tau-F) representing the 2N4R This pathological process causes a reduced basal synaptic
isoform of the protein. The N-terminal projection domain transmission in the hippocampus. (b) In contrast the
points away from microtubules, the C-terminal microtu- mutation A152T is located in the first proline rich domain
bule assembly domain is responsible for microtubule (P1) towards the N-terminus. This mutation leads to an
binding. The assembly domain contains the repeat domain increased oligomer formation thereby enhancing basal
(R1-R4) and is flanked by proline-rich domains. The synaptic transmission

The ΔK280 mutation leads to a decrease in The pronounced reduction of basal synaptic
presynaptic bouton density in the mossy fiber transmission in mice overexpressing TauRDΔK280
pathway, the synaptic connection of granule cells may point to a general network hypoexcitability,
from the dentate gyrus to pyramidal cells in area whereas the mutation A152T causes network
CA3 of the hippocampus [12]. Moreover these hyperexcitability and excitotoxicity as demon-
presynaptic boutons contain less synaptic vesicles strated both in vitro and in vivo [13, 18]. The fact
after expressing TauRDΔK280, and their calcium that Tau knockout mice show a similar impair-
dynamics is impaired as well. This severe presyn- ment of basal synaptic transmission as mice
aptic phenotype is accompanied by a decrease of expressing TauRDΔK280 [12] suggests that Tau
dendritic spines at the postsynaptic side [12]. may be an important regulator of basal synaptic
The pathological phenotype shows up as a transmisson in the mossy fiber pathway in a nor-
reduced basal synaptic transmission (Fig.  8.1a) mal physiological environment.
and impaired synaptic plasticity in transgenic It remains to be elucidated why a mutation
mice overexpressing TauRDΔK280. This patho- located in the N-terminal proline-rich domain has
physiological picture is similar in mice overex- an opposite effect than the one in the repeat
pressing full-length Tau with the ΔK280 domain. It is likely that the mutations have differ-
mutation, TauΔK280 [17]. ent effects on synapses. In case of the ΔK280
In contrast the A152T mutation leads to an mutation there is reduced release of synaptic ves-
increase in basal synaptic transmission (Fig. 8.1b) icles and a gradual loss of hippocampal spines
but has no effect on synaptic plasticity [13]. In and neurons, whereas in the case of the A152T
this model the expression of full-length TauA152T mutation there is an increase of glutamate release.
causes an increase in extracellular glutamate, This is consistent with the observed hyperexcit-
NR2B-dependent increase in intracellular cal- ability and mossy fiber sprouting in this region
cium and finally to neuronal death by excitotox- [13]. In older animals (20  month) expressing
icity [13]. A152T the number of dendritic spines is finally
8  Presynaptic Pathophysiology Encoded in Different Domains of Tau – Hyper-Versus Hypoexcitability? 99

decreased in area CA1 and CA3 when increased ing in organotypic hippocampal slices (Fig. 8.2a,
cell death becomes dominant [19]. b). Slices expressing the ΔK280 mutation showed
An underlying mechanism of Tau’s effect on less calcium influx after depolarization (Fig. 8.2a,
network activity may be calcium dynamics. In c). By contrast, slices expressing the A152T
area CA3, where mossy fiber synapses are mutation had a pronounced increase in calcium
located, we observed impaired calcium dynamics levels after KCl depolarization (Fig. 8.2b, d). The
after high KCl stimulation, using calcium imag- hyperexcitable phenotype (A152T mutation) can

a c
hTau/∆K280

hTau/A152T

b d
700 700 A152T [nM]
600 600
[Ca++]i [nM]
[Ca++]i [nM]

550
500 500

400
Ctrl [nM]
400

300 KCl 280

300
KCl
270
200 200
100
120 Ctrl
100 100
ΔK280 100
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
sec sec

Fig. 8.2  Hypo- and hyperactivity of intracellular cal- cellular calcium concentrations (nM). At baseline condi-
cium dynamics after chemical depolarisation (KCl) tions the control slices and hTauΔK280 overexpressing
due to the overexpression of human TauΔK280 and slices have similar intracellular calcium concentrations of
human TauA152T as observed in organotypic hippo- ~80  nM.  After chemical stimulation (KCl) intracellular
campal slices of transgenic mice. (a) Microphotograph calcium concentration rises up to 280 nM in control slices
showing area CA3 of organotypic hippocampal slices whereas slices overexpressing TauΔK280 reach only
expressing TauΔK280 loaded with Fura2-AM. Ratiometric ~120  nM. (c) Microphotograph showing area CA3 of
color-coding shows intracellular low calcium concentra- organotypic hippocampal slices expressing Tau-A152T
tions in dark blue, blue and green regions and high cal- loaded with Fura2-AM.  Note that at baseline level neu-
cium concentrations in yellow, red and white regions. In rons have higher calcium concentrations than control
imaging experiments slices are recorded in unstimulated, slices in (a). After KCl stimulation calcium concentrations
basal conditions for 30  seconds. Subsequently a high rise to high levels (red and green). (d) Diagram showing
potassium chloride (KCl) stimulation is applied and mem- intracellular calcium concentrations over time. Before
brane depolarization induced. Calcium rises rapidly, stimulation calcium concentrations range around 90  nM
reflecting neuronal activity. (b) Diagram showing the in control slices. In slices expressing Tau-A152T intracel-
quantification of experiments described in (a). X-axis lular calcium is already increased at baseline conditions.
denotes time in seconds, the Y-axis shows absolute intra- However after KCl stimulation calcium rises up to 550 nM
100 J. M. Decker and E.-M. Mandelkow

hTau∆ K280 hTauA152T

Phosporylation Phosporylation

Presynaptic transmitter Presynaptic transmitter

release (Glu) release (Glu)
Bb14,Rolofylline Ceftriaxone

Hypoactivity/ Hyperactivity/ Excitotoxicity

Postsynaptic decay Postsynaptic decay
Neuronal loss Neuronal loss

Cognitive decline
Fig. 8.3  Flow chart of events leading to cognitive overspill and excitotoxicity. Both mutations finally
decline due to the overexpression of Tau with the decrease postsynaptic structures leading to neuronal
mutations ΔK280 or A152T. Initially the overexpres- loss and cognitive decline. These scenarios can be pre-
sion of hTauΔK280 and hTauA152T causes pathologi- vented in the case of the ΔK280 mutation by applica-
cal phosphorylation. Thereafter presynaptic release is tion of aggregation inhibitor (bb14) or by the adenosine
affected. Presynaptic release can be decreased due to A1 receptor antagonist (rolofylline). In the case of the
structural and functional impairment of presynapses by A152T mutation the application of ceftriaxone,
hTauΔK280. By contrast, hTauA152T expression memantine, or ifenprodil can prevent excitotoxicity
enhances presynaptic release leading to glutamate and neuronal loss

be reversed by application of the drug ceftriax- ogy affected neurotransmitter release [26], arguing
one, which causes an increased glutamate uptake that pathological Tau can change the tethering of
by astrocytes via the glutamate transporter synaptic vesicles to actin filaments via synaptogy-
EAAT1/2 (Fig. 8.3). The hypoactivity phenotype rin-3, thereby slowing down presynaptic transmitter
(ΔK280 mutation) was ameliorated by the Tau release [27]. By analogy, other Tau mutations like
aggregation inhibitor bb14 [20] and by rolofyl- A152T could increase presynaptic transmitter
line, an antagonist of the adenosine A1 receptor release and thus network excitability. In both cases,
(Fig. 8.3 [21]). a presynaptic origin of early pathology would be
consistent with the axonal localization of Tau [28].
The ΔK280 mutation of Tau has a direct
Discussion impact on structure and function of presynapses
[12], which would be consistent with the effect of
Much work has been done in recent years to under- this variant on Tau conformation and aggrega-
stand Tau’s impact on hippocampal neurophysiol- tion. In the case of the A152T mutation an effect
ogy [22–24]. The focus was on postsynaptic on structure cannot be readily explained, but the
structures and dysfunction caused by pathological proximity to the proline-rich domain suggests an
Tau [10, 11]. More recently some studies focussed influence on signalling pathways. An increase of
on the evaluation of pathophysiological features at presynaptic vesicle release (and thus of gluta-
presynapses during the progression of tauopathies mate) was shown via tetanus toxin application
[9, 12, 25]. Some mechanisms of presynaptic pathol- [13] and may well be an initial mechanism of
8  Presynaptic Pathophysiology Encoded in Different Domains of Tau – Hyper-Versus Hypoexcitability? 101

A152T toxicity (Fig. 8.3). A direct comparison of The A152T mutation in Tau may trigger
presynaptic effects of different Tau mutations pathophysiological events at an early stage where
remains to be studied in the future. neuronal malfunction is triggered by oligomer
Finally it is important to consider Tau’s impact formation and network hyperexcitability/hyper-
on functional and structural plasticity in the brain. activity. By contrast, the aggregation-prone
For instance, Tau’s regulatory role in h­ ibernating ΔK280 mutation may reflect a later stage of
animals is well described [29]. In this case Tau tauopathy where aggregated, fibrillary Tau causes
adopts “Alzheimer-like” features (hyperphosphory- network hypoexcitability/hypoactivity. This
lation, somatodendritic mislocalization) which have “later stage” of tauopathy may reflect a kind of
transient physiological roles. Tau-dependent modu- “protection mechanism” in order to “calm down”
lation of synaptic plasticity has proven to be impor- increased network activity which may be trig-
tant in the regulation of hippocampal LTP/LTD, as gered by different pathological conditions like
demonstrated by mouse models overexpressing Aß, ROS and others.
human Tau [17, 30] or in Tau knockout models [31].
Indeed Tau was also found to impact Delta/theta Acknowledgements We are grateful to Dr. Lars
network oscillations important for cognition [32]. Krüger and Dr. Astrid Sydow for their important con-
tributions to the generation and analysis of the trans-
Changes of intrinsic neuronal excitability [22], pre- genic mouse lines and organotypic slices, and Dr.
and postsynaptic dysfunction and deficits in cal- Eckhard Mandelkow for stimulating discussions. This
cium dynamics/signaling due to the expression of work was supported by the German Center for
pathological Tau may underlie its role as a modifier Neurodegenerative Diseases (DZNE) and the Max-
Planck-Society (MPG).
of network excitability (Fig. 8.4).

Presynapse

Ceftriaxone

Tetanus Toxin
APV
NMDAR
EAAT1/2
Ca++

Astrocyte Postsynapse

Memantine
Ifenprodil

Fig. 8.4  Model of synaptic pathophysiology induced synaptic NR2B-containing NMDA receptors. The impair-
by TauA152T at the tripartite synapse. Expression of ment can be reduced by low concentrations of memantine
Tau-A152T in the presynapse leads to enhanced presyn- or ifenprodile (inhibition of NR2B) or by ceftriaxone
aptic release of glutamate which causes excitotoxicity (upregulation of astrocytic glutamate transporters
which is mediated by excess Ca++ influx through extra- (EAAT1/2))
102 J. M. Decker and E.-M. Mandelkow

References 14. Sydow A, Van der Jeugd A, Zheng F, Ahmed T,

Balschun D, Petrova O, et al. Tau-induced defects in
synaptic plasticity, learning, and memory are revers-
1. Braak H, Braak E.  Neuropathological stageing of
ible in transgenic mice after switching off the toxic
Alzheimer-related changes. Acta Neuropathol.
tau mutant. J Neurosci. 2011;31(7):2511–25.
1991;82(4):239–59.
15. Coppola G, Chinnathambi S, Lee JJ, Dombroski BA,
2. Hutton M, Lendon CL, Rizzu P, Baker M, Froelich
Baker MC, Soto-Ortolaza AI, et  al. Evidence for a
S, Houlden H, et  al. Association of missense and
role of the rare p.A152T variant in MAPT in increas-
5′-splice-site mutations in tau with the inherited
ing the risk for FTD-spectrum and Alzheimer’s dis-
dementia FTDP-17. Nature. 1998;393(6686):702–5.
eases. Hum Mol Genet. 2012;21(15):3500–12.
3. Ramsden M, Kotilinek L, Forster C, Paulson J,
16. Reynolds CH, Garwood CJ, Wray S, Price C, Kellie
McGowan E, SantaCruz K, et al. Age-dependent neu-
S, Perera T, et  al. Phosphorylation regulates tau
rofibrillary tangle formation, neuron loss, and mem-
interactions with Src homology 3 domains of phos-
ory impairment in a mouse model of human tauopathy
phatidylinositol 3-kinase, phospholipase Cgamma1,
(P301L). J Neurosci. 2005;25(46):10637–47.
Grb2, and Src family kinases. J  Biol Chem.
4. Eckermann K, Mocanu MM, Khlistunova I, Biernat
2008;283(26):18177–86.
J, Nissen A, Hofmann A, et  al. The beta-propensity
17. Van der Jeugd A, Hochgrafe K, Ahmed T, Decker
of tau determines aggregation and synaptic loss in
JM, Sydow A, Hofmann A, et  al. Cognitive defects
inducible mouse models of tauopathy. J  Biol Chem.
are reversible in inducible mice expressing pro-­
2007;282(43):31755–65.
aggregant full-length human tau. Acta Neuropathol.
5. Stamer K, Vogel R, Thies E, Mandelkow E,
2012;123(6):787–805.
Mandelkow EM. Tau blocks traffic of organelles, neu-
18. Maeda S, Djukic B, Taneja P, Yu GQ, Lo I, Davis A,
rofilaments, and APP vesicles in neurons and enhances
et  al. Expression of A152T human tau causes age-­
oxidative stress. J Cell Biol. 2002;156(6):1051–63.
dependent neuronal dysfunction and loss in trans-
6. Dickey CA, Kamal A, Lundgren K, Klosak N, Bailey
genic mice. EMBO Rep. 2016;17(4):530–51.
RM, Dunmore J, et  al. The high-affinity HSP90-­
19. Sydow A, Hochgrafe K, Konen S, Cadinu D, Matenia
CHIP complex recognizes and selectively degrades
D, Petrova O, et  al. Age-dependent neuroinflamma-
phosphorylated tau client proteins. J  Clin Invest.
tion and cognitive decline in a novel Ala152Thr-­
2007;117(3):648–58.
tau transgenic mouse model of PSP and AD.  Acta
7. Violet M, Delattre L, Tardivel M, Sultan A,
Neuropathol Commun. 2016;4:17.
Chauderlier A, Caillierez R, et al. A major role for tau
20. Messing L, Decker JM, Joseph M, Mandelkow E,
in neuronal DNA and RNA protection in vivo under
Mandelkow EM. Cascade of tau toxicity in inducible
physiological and hyperthermic conditions. Front Cell
hippocampal brain slices and prevention by aggrega-
Neurosci. 2014;8:84.
tion inhibitors. Neurobiol Aging. 2013;34(5):1343–54.
8. Sultan A, Nesslany F, Violet M, Begard S, Loyens A,
21.
Dennissen FJ, Anglada-Huguet M, Sydow A,
Talahari S, et al. Nuclear tau, a key player in neuronal
Mandelkow E, Mandelkow EM.  Adenosine A1
DNA protection. J Biol Chem. 2011;286(6):4566–75.
receptor antagonist rolofylline alleviates axonopathy
9. Polydoro M, Dzhala VI, Pooler AM, Nicholls SB,
caused by human tau DeltaK280. Proc Natl Acad Sci
McKinney AP, Sanchez L, et  al. Soluble pathologi-
U S A. 2016;113(41):11597–602.
cal tau in the entorhinal cortex leads to presynaptic
22. Booth CA, Witton J, Nowacki J, Tsaneva-Atanasova
deficits in an early Alzheimer’s disease model. Acta
K, Jones MW, Randall AD, et  al. Altered intrinsic
Neuropathol. 2014;127(2):257–70.
pyramidal neuron properties and pathway-specific
10. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A,
synaptic dysfunction underlie aberrant hippocampal
van Eersel J, et al. Dendritic function of tau mediates
network function in a mouse model of Tauopathy.
amyloid-beta toxicity in Alzheimer’s disease mouse
J Neurosci. 2016;36(2):350–63.
models. Cell. 2010;142(3):387–97.
23. Fu H, Rodriguez GA, Herman M, Emrani S, Nahmani
11. Hoover BR, Reed MN, Su J, Penrod RD, Kotilinek
E, Barrett G, et al. Tau pathology induces excitatory
LA, Grant MK, et al. Tau mislocalization to dendritic
neuron loss, grid cell dysfunction, and spatial mem-
spines mediates synaptic dysfunction independently
ory deficits reminiscent of early Alzheimer’s disease.
of neurodegeneration. Neuron. 2010;68(6):1067–81.
Neuron. 2017;93(3):533–41 e5.
12. Decker JM, Kruger L, Sydow A, Zhao S, Frotscher
24. Tamagnini F, Walsh DA, Brown JT, Bondulich MK,
M, Mandelkow E, et  al. Pro-aggregant tau impairs
Hanger DP, Randall AD. Hippocampal neurophysiology
mossy fiber plasticity due to structural changes and
is modified by a disease-associated C-terminal fragment
Ca(++) dysregulation. Acta Neuropathol Commun.
of tau protein. Neurobiol Aging. 2017;60:44–56.
2015;3:23.
25. Harris JA, Koyama A, Maeda S, Ho K, Devidze N,
13. Decker JM, Kruger L, Sydow A, Dennissen FJ,

Dubal DB, et al. Human P301L-mutant tau expression
Siskova Z, Mandelkow E, et al. The tau/A152T muta-
in mouse entorhinal-hippocampal network causes tau
tion, a risk factor for frontotemporal-spectrum disor-
aggregation and presynaptic pathology but no cogni-
ders, leads to NR2B receptor-mediated excitotoxicity.
tive deficits. PLoS One. 2012;7(9):e45881.
EMBO Rep. 2016;17(4):552–69.
8  Presynaptic Pathophysiology Encoded in Different Domains of Tau – Hyper-Versus Hypoexcitability? 103

26. Zhou L, McInnes J, Wierda K, Holt M, Herrmann AG, 30. Sydow A, Van der Jeugd A, Zheng F, Ahmed T,

Jackson RJ, et al. Tau association with synaptic ves- Balschun D, Petrova O, et  al. Reversibility of tau-­
icles causes presynaptic dysfunction. Nat Commun. related cognitive defects in a regulatable FTD mouse
2017;8:15295. model. J Mol Neurosci. 2011;45(3):432–7.
27. McInnes J, Wierda K, Snellinx A, Bounti L, Wang 31. Regan P, Piers T, Yi JH, Kim DH, Huh S, Park

YC, Stancu IC, et  al. Synaptogyrin-3 mediates SJ, et  al. Tau phosphorylation at serine 396 resi-
presynaptic dysfunction induced by tau. Neuron. due is required for hippocampal LTD.  J Neurosci.
2018;97(4):823–35 e8. 2015;35(12):4804–12.
28. Binder LI, Frankfurter A, Rebhun LI.  The distribu- 32. Das M, Maeda S, Hu B, Yu GQ, Guo W, Lopez I,
tion of tau in the mammalian central nervous system. et  al. Neuronal levels and sequence of tau modu-
J Cell Biol. 1985;101(4):1371–8. late the power of brain rhythms. Neurobiol Dis.
2 9.
Arendt T, Bullmann T.  Neuronal plastic- 2018;117:181–8.
ity in hibernation and the proposed role of the 33. von Bergen M, Friedhoff P, Biernat J, Heberle J,

microtubule-­associated protein tau as a “master Mandelkow EM, Mandelkow E. Assembly of tau pro-
switch” regulating synaptic gain in neuronal net- tein into Alzheimer paired helical filaments depends
works. Am J Physiol Regul Integr Comp Physiol. on a local sequence motif ((306)VQIVYK(311))
2013;305(5):R478–89. forming beta structure. Proc Natl Acad Sci U S A.
2000;97:5129–34.
 Synaptic Localisation of Tau
9
Diane P. Hanger, Despoina Goniotaki,
and Wendy Noble

Introduction kainate, with the specific receptor subunit type

being both brain region and cell-type specific [3].
Chemical synapses are highly specialised struc- Synaptic plasticity enables a rapid response to
tures in which neuronal activity at presynaptic electrical stimulation resulting in changes in the
boutons results in the release of neurotransmit- morphology of synapses. Synaptic plasticity and
ters from synaptic vesicles (Fig.  9.1). the formation of dendritic spines are considered
Neurotransmitter release is stimulated by an the structural correlates of learning and memory
increase in intracellular calcium ions caused by [4]. Long-term potentiation (LTP) is induced by
depolarisation of presynaptic voltage-gated cal- recurring synaptic activity that activates NMDA
cium channels, leading to fusion of synaptic ves- receptors, resulting in increased synaptic
icles with the presynaptic plasma membrane [1]. strength. In contrast, long-term depression
Release of neurotransmitters from synaptic vesi- (LTD), whilst also activity-dependent, weakens
cles into the synaptic cleft activates postsynaptic synaptic activity through phosphorylation of
receptors, resulting in the transduction of either AMPA receptors which can affect their density at
excitatory (glutamate) or inhibitory (gamma-­ the postsynaptic membrane [5]. Thus, function-
aminobutyric acid/glycine) signalling. The post- ing synaptic connections are required for the
synaptic density (PSD), an electron-dense region regulated transmission of chemical information
in the receiving neuron, is a specialised scaffold- between neurons and dysfunctional signalling
ing structure that anchors primarily excitatory, can lead to deficits in synaptic function and ulti-
ionotropic and metabotropic glutamate receptors mately loss of connectivity between neurons and
to the postsynaptic membrane of dendritic spines, neurodegeneration.
forming a signalling complex [2]. Ionotropic glu- Synaptic loss is the earliest indication of neu-
tamate receptors are activated by α-amino-3-­ ronal malfunction and the best biological corre-
hydroxy-5-methylisoxazole-4-propionate late of disease progression in Alzheimer’s disease
(AMPA), N-methyl-D-aspartate (NMDA), or and related tauopathies [6–8]. Tau has long been
regarded as an axonal microtubule-associated
protein [9, 10], but recently it has become appar-
ent that tau is also associated with other neuronal
D. P. Hanger (*) · D. Goniotaki · W. Noble
Department of Basic and Clinical Neuroscience, cellular compartments, including the plasma
Institute of Psychiatry, Psychology & Neuroscience, membrane and the synapse [11, 12]. Notably,
King’s College London, London, UK overexpression of tau in transgenic mice leads to
e-mail: diane.hanger@kcl.ac.uk; despoina. loss of synapses [13], and co-expression of tau
goniotaki@kcl.ac.uk; wendy.noble@kcl.ac.uk

© Springer Nature Singapore Pte Ltd. 2019 105

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_9
106 D. P. Hanger et al.

Microtubules

Tau
Presynapse
Synaptic vesicles

Neurotransmitters

Fyn

Postsynapse AMPA/NMDA receptors

Phosphorylated tau

Post-synaptic density

Dendritic spine F-actin

Fig. 9.1  Structural organisation and tau localisation in Tau binds to microtubules in axons and to synaptic vesi-
excitatory synapses. Neuronal activity at the presynaptic cles in the presynaptic compartment. Tau is also present
terminal results in the calcium-mediated fusion of synap- extracellularly, as well as at the postsynaptic terminal and
tic vesicles with the presynaptic membrane, releasing in dendritic spines. Accumulation of phosphorylated and/
neurotransmitters into the synaptic cleft. Interaction of the or acetylated tau, including that associated with fyn, in the
neurotransmitters with postsynaptic receptors causes postsynapse occurs in the tauopathies leading to detrimen-
membrane depolarisation, increasing the likelihood of tal effects on synaptic function
action potential propagation in the postsynaptic neuron.

with amyloid beta peptide (Aβ) suppresses neu- overlap of labelling of the dendritic proteins, dre-
ronal activity [14]. It is important therefore, to brin and microtubule-associated protein 2, with
gain a better understanding of synaptic dysfunc- endogenous mouse tau using highly specific tau
tion and neuronal loss in the tauopathies since antibodies [18]. However, since reducing tau
this will aid elucidation of the molecular mecha- expression results in loss of synapses [16], this
nisms that lead to tau-induced cognitive dysfunc- suggests a critical role for tau in maintaining syn-
tion and will inform the development of strategies aptic integrity. Under pathological conditions,
to protect synapses. such as in human tauopathy and when tau is over-
expressed in animal models of disease, the pres-
ence of tau  at synapses is more apparent,
Synaptic Localisation of Tau suggesting a role for tau in disease pathogenesis
[12, 15, 18–22]. Synaptic tau has been demon-
In the axon, tau functions as a cytoskeletal pro- strated in cell and animal models that overexpress
tein with roles in microtubule stabilisation and tauopathy-associated tau mutations, including in
axonal transport [10], whereas synaptic tau cultured hippocampal neurons [19], in rodents,
appears to be involved in neuronal signalling and and in Drosophila [17, 23, 24]. Importantly, tau
synaptic plasticity [15, 16]. The presence of tau spread from presynaptic to postsynaptic sites pre-
in synapses under physiological conditions has cedes synapse and neuronal loss in transgenic
been somewhat controversial, with some studies mice expressing mutant human tau [25].
detecting presynaptic tau in human and animal Mislocalisation of tau to dendrites is a neuro-
brain tissue [17] and others failing to detect any pathological feature of Alzheimer’s disease brain
9  Synaptic Localisation of Tau 107

that occurs early in disease pathogenesis, possi- Tau clusters synaptic vesicles by interacting
bly even pre-clinically and prior to tau aggrega- with both F-actin and synaptogyrin-3, which
tion [26, 27]. Both presynaptic and postsynaptic binds directly to the amino terminal region of tau,
compartments have been identified as sites of tau reducing the efficiency of synaptic vesicle release
accumulation in Alzheimer’s disease [20, 28–30] [24, 33]. Notably, synaptogyrin-3 preferentially
and the appearance of synaptic tau correlates binds 0N tau isoforms rather than 1N or
with the onset of cognitive decline [31, 32]. 2N-containing tau isoforms, which is compatible
Human brain preparations enriched in synap- with the identification of residues 1–111 (num-
tic vesicles from control and Alzheimer’s disease bering of the 0N4R human tau isoform used in
contain tau, with disease-associated phosphory- that study), which is equivalent to residues 1–44
lated tau detectable only in Alzheimer brain [33]; and 103–169  in the longest CNS isoform of
the abundance of phosphorylated tau oligomers human tau of (2N4R tau, 441 amino acids) [24,
at synapses is particularly associated with demen- 37]. The tau-induced clustering of synaptic vesi-
tia [34]. Accumulation of phosphorylated tau cles at the presynaptic terminal, mediated by
results in neuronal injury that is mediated by synaptogyrin-­ 3, impacts on neurotransmitter
damage to synapses, which increases with dis- release [33]. These data suggest that the amino
ease progression in the tauopathies [30, 35]. terminal region of tau has an important influence
Whilst the molecular mechanisms underlying on neuronal function in the presynaptic compart-
synaptic loss are incompletely understood, there ment by reducing the mobility of synaptic vesi-
are several potential routes through which patho- cles and release of neurotransmitters, resulting in
logical forms of tau might prove to be toxic to synaptic toxicity.
neurons either directly or indirectly, and at both
presynaptic and postsynaptic sites.
Post-Translational Modifications
of Tau Affect Synaptic Function
Tau Binds to Synaptic Vesicles
Tau is subject to a wide variety of post-­
The observation that axonal dystrophy occurs in translational modifications, several of which are
advance of the appearance of both neurofibrillary altered in the tauopathies, and these modifica-
tangles and amyloid plaques has led to the pro- tions can have deleterious effects on the neuronal
posal that loss of presynaptic terminals leads to a localisation and function of tau [38]. For exam-
retrograde “dying back” of neurons in Alzheimer’s ple, tau becomes increasingly phosphorylated
disease [32, 36]. The presence of soluble (mono- during the  development and progression of
meric or oligomeric), pathological tau species in tauopathy. Indeed, one of the key neuropatholog-
the presynaptic compartment has been linked to ical features of the tauopathies is the develop-
disease pathogenesis in Alzheimer’s disease [35] ment of neurofibrillary tau inclusions comprised
and to synaptic vesicle release deficits in fly and of highly phosphorylated tau fibrils, the appear-
rat neurons [24]. Several studies have demon- ance of which likely leads to neuronal dysfunc-
strated the association of tau with presynaptic tion and loss of cognitive ability as disease
proteins, including synaptophysin, synapsin 1, progresses [38].
synaptotagmin, synaptogyrin-3, syntaxin-1B, Although tau phosphorylation is a physiologi-
α-synuclein, and β-synuclein, all of which co-­ cal process, aberrantly increased tau phosphory-
immunoprecipitate with tau [37]. More recently, lation may be a forerunner of tau-induced
the synaptic vesicle transmembrane protein syn- neurofibrillary degeneration. Changes in tau
aptogyrin-­3 has been identified as a key tau inter- phosphorylation have significant impacts on tau
actor and a  regulator of tau-induced synaptic function, including microtubule binding and axo-
dysfunction [33]. nal transport, as well as bundling presynaptic
actin filaments, which is reduced by increased tau
108 D. P. Hanger et al.

phosphorylation [39]. Phosphorylation-­ Structural Consequences

mimicking tau mutants expressed in cultured of Aggregated Tau
neurons reduce the number of AMPA receptors at
synapses [19]. Furthermore, increased tau phos- Oligomeric and fibrillar tau but not monomeric
phorylation results in tau mislocalisation to post- tau, have detrimental effects at synapses because
synaptic sites in Alzheimer’s disease brain and in these tau species have been shown to be capable
transgenic mice overexpressing P301S tau, a of diffusing laterally through membranes and
mutation that causes frontotemporal dementia forming clusters that are proposed to selectively
[21, 30]. A potential mechanism that may be increase calcium-impermeable GluA2 AMPA
responsible for the observed increase in phos- receptors and to reduce sodium-potassium
phorylated tau in the tauopathies comes from ATPase [52]. Notably, this clustering of tau is
studies of the association of tau with the tyrosine enhanced by the presynaptic protein α-synuclein,
kinase fyn [15, 40–43]. Src homology-3 (SH3) which is frequently found in association with tau
domains in fyn bind to Pro-X-X-Pro motifs in the deposits in the tauopathies [53]. However, others
proline-rich region of tau [41–43] thereby direct- have shown a reduction in GluR1 AMPA recep-
ing tau to the postsynaptic density [15, 44]. tors in neurons cultured from P301L tau trans-
Importantly, the interaction of fyn and tau, and genic mice [19, 54] and directly in P301L tau
hence tau localisation is regulated by tau phos- mouse brain [55]. Moreover, soluble tau can
phorylation [41, 42, 45–47], suggesting that this induce synaptic loss in mouse brain [56], sug-
may be a component of the molecular mecha- gesting that multiple mechanisms may be
nism leading to increased postsynaptic localisa- involved in tau-induced disruption of synapses. It
tion of phosphorylated tau. Dendritic tau likely is plausible therefore, that exposure to tau induces
mediates the toxicity of Aβ, the major component mild structural changes at the synapse and the
of amyloid plaques in Alzheimer’s disease, since accumulation of tau gradually progresses to
neurons cultured from tau knockout mice are result in  significant functional abnormalities  at
resistant to Aβ toxicity [42, 44]. Importantly, the synapse [55, 57].
translocation of tau to the postsynaptic density of Oligomeric and fibrillar forms of tau can also
excitatory synapses is driven by Aβ, which could be generated by proteolytic cleavage, and a large
be a key event inducing synaptotoxicity in number of tau fragments have been identified in
dementia [48]. tauopathy brain [10]. Reduced clustering of
Acetylation of tau also blocks activity-induced AMPA receptors has also been noted in mice
actin polymerisation, through a mechanism expressing a mutant form of tau (tauΔ314), that
involving the postsynaptic memory-associated corresponds to caspase-cleaved tau [58]. Thus,
KIdney/BRAin protein (KIBRA), which is deposition and mislocalisation of tau at postsyn-
encoded by the WWC1 gene [49]. Notably aptic sites could damage neurons by interfering
genetic variation in WWC1 is associated with the with their ability to regulate synaptic membrane
development of late-onset dementia [50] and depolarisation and by interfering with the mem-
KIBRA protein is reduced in Alzheimer’s disease brane organisation and function of AMPA recep-
brain, whereas tau acetylation is increased [51]. tors, resulting in cognitive impairment in the
The finding that increased tau acetylation reduces tauopathies.
KIBRA and impairs LTP by impeding actin poly-
merisation and thereby affecting postsynaptic
membrane localisation of AMPA receptors, adds
to the evidence implicating the importance of tau
in regulating synaptic plasticity [49].
9  Synaptic Localisation of Tau 109

 unctional Impacts of Synaptic

F tions from human brain that release tau in
Localisation of Tau response to potassium-induced depolarisation
[20]. Notably, in a mouse model of Alzheimer’s
I mpact of Tau on Neuronal Activity disease (3xTg-AD mice) in which exogenous tau
and Neuronal Circuitry is both mutated and highly phosphorylated,
potassium chloride-induced depolarisation was
Tau appears to have a role in modulating synaptic unable to induce tau release from organotypic
activity since exposure of neurons to tau inhibits brain slices [67]. These findings suggest that tau
hippocampal LTP in rat hippocampal synapses mutations and/or increased phosphorylation as
[59]. Similarly, inhibition of LTP by extracts of found in human tauopathy brain, could poten-
Alzheimer’s disease brain could be prevented by tially obstruct the elevated release of tau in
co-injection with an antibody to tau, indicating response to neuronal activity.
the tau dependence of LTP inhibition in this
model [59]. However, tau has also been reported
to suppress cortical activity in P301L tau mice Conclusions
[14] and when P301L tau is targeted to the ento-
rhinal cortex, one of the earliest sites of tau Tau is present in multiple intracellular domains
pathology in Alzheimer’s disease [60], indicating within neurons, including being associated with
a negative impact of tau on neuronal the microtubule cytoskeleton and plasma mem-
connectivity. brane, as well as being present in a cytosolic pool
There is increasing evidence of a close rela- and as a secreted protein. These findings have led
tionship between tau and epileptic activity, with to increased scrutiny of the role of tau in these
an increased prevalence of seizures in Alzheimer’s differing locations, including at synapses, where
disease [61]. Several tau mutations that increase tau may either directly or indirectly affect the
the risk of developing frontotemporal dementia response of neurons to changes in neuronal activ-
are associated with seizures and abnormal net- ity (Fig. 9.1). During the development of tauopa-
work activity, some of which can be ameliorated thy, phosphorylated tau accumulates in
with anti-epileptic drugs, such as levetiracetam postsynaptic regions, although it is not yet clear
[62]. Seizures are also linked to cognitive impair- whether this is a cause or consequence of disease,
ment and tau pathology in temporal lobe epilepsy or precisely how this mis-localisation might dam-
[63], suggesting the possibility of some com- age neurons, leading to loss of synaptic dysfunc-
monalities in the mechanisms underlying epi- tion and neuronal connectivity. Synaptic tau
lepsy and dementia, that could be related to tau interacts with both presynaptic and postsynaptic
deposition. components and the strength of these associa-
tions may be modulated by disease-relevant post-­
translational modifications of tau. Thus, increased
Impact on Neuronal Tau Release tau phosphorylation, acetylation and/or the
occurrence of pathogenic mutations in tau, can
Increased neuronal activity induces tau secretion all lead to its mislocalisation at both presynaptic
from neurons [64, 65] and therefore, epileptic and postsynaptic sites, potentially contributing to
activity could enhance the spread of tau in the disease progression in the tauopathies.
tauopathies [66]. The precise region of the neu- Understanding the role of tau at the synapse and
ron from which tau is released has not yet been the impact on neuronal viability and function
established but one possibility is that tau could may provide new avenues for therapeutic inter-
dock at or near the synaptic membrane prior to vention in the tauopathies.
release. This view is supported by the finding of
equivalent amounts of tau in the soluble and Acknowledgements  Work in the authors’ laboratories is
detergent-soluble fractions of synaptic prepara- supported by Alzheimer’s Research UK, the Alzheimer’s
110 D. P. Hanger et al.

Society, and the Biotechnology and Biological Sciences amyloid-beta toxicity in Alzheimer’s disease mouse
Research Council. models. Cell. 2010;142(3):387–97.
16. Chen Q, Zhou Z, Zhang L, Wang Y, Zhang YW,

Zhong M, et  al. Tau protein is involved in morpho-
logical plasticity in hippocampal neurons in response
References to BDNF. Neurochem Int. 2012;60(3):233–42.
17. Jadhav S, Cubinkova V, Zimova I, Brezovakova V,
1. Catterall WA, Few AP. Calcium channel regulation and Madari A, Cigankova V, et  al. Tau-mediated synap-
presynaptic plasticity. Neuron. 2008;59(6):882–901. tic damage in Alzheimer’s disease. Transl Neurosci.
2. Sheng M, Kim E.  The postsynaptic organiza- 2015;6(1):214–26.
tion of synapses. Cold Spring Harb Perspect Biol. 18. Kubo A, Misonou H, Matsuyama M, Nomori A,

2011;3(12): a005678. Wada-Kakuda S, Takashima A, et al. Distribution of
3. Traynelis SF, Wollmuth LP, McBain CJ, Menniti endogenous normal tau in the mouse brain. J  Comp
FS, Vance KM, Ogden KK, et  al. Glutamate recep- Neurol. 2019;527(5):985–98.
tor ion channels: structure, regulation, and function. 19. Hoover BR, Reed MN, Su J, Penrod RD, Kotilinek
Pharmacol Rev. 2010;62(3):405–96. LA, Grant MK, et al. Tau mislocalization to dendritic
4. Kandel ER, Dudai Y, Mayford MR.  The molec- spines mediates synaptic dysfunction independently
ular and systems biology of memory. Cell. of neurodegeneration. Neuron. 2010;68(6):1067–81.
2014;157(1):163–86. 20. Sokolow S, Henkins KM, Bilousova T, Gonzalez

5. Tatavarty V, Sun Q, Turrigiano GG.  How to B, Vinters HV, Miller CA, et  al. Pre-synaptic
scale down postsynaptic strength. J  Neurosci. C-terminal truncated tau is released from corti-
2013;33(32):13179–89. cal synapses in Alzheimer’s disease. J  Neurochem.
6. Terry RD, Masliah E, Salmon DP, Butters N, 2015;133(3):368–79.
DeTeresa R, Hill R, et al. Physical basis of cognitive 21. Dejanovic B, Huntley MA, De Maziere A, Meilandt
alterations in Alzheimer’s disease: synapse loss is the WJ, Wu T, Srinivasan K, et  al. Changes in the syn-
major correlate of cognitive impairment. Ann Neurol. aptic proteome in Tauopathy and Rescue of tau-­
1991;30(4):572–80. Induced Synapse Loss by C1q antibodies. Neuron.
7. Masliah E, Hansen L, Albright T, Mallory M, Terry 2018;100(6):1322–36 e7.
RD.  Immunoelectron microscopic study of synaptic 22. Ji C, Tang M, Zeidler C, Hohfeld J, Johnson GV. BAG3
pathology in Alzheimer’s disease. Acta Neuropathol. and SYNPO (synaptopodin) facilitate phospho-­
1991;81(4):428–33. MAPT/tau degradation via autophagy in neuronal
8. Scheff SW, Price DA, Schmitt FA, Mufson processes. Autophagy. 2019;15(7):1199–1213.
EJ.  Hippocampal synaptic loss in early Alzheimer’s 23. Sahara N, Murayama M, Higuchi M, Suhara T,

disease and mild cognitive impairment. Neurobiol Takashima A. Biochemical distribution of tau protein
Aging. 2006;27(10):1372–84. in Synaptosomal fraction of transgenic mice express-
9. Binder LI, Frankfurter A, Rebhun LI.  The distribu- ing human P301L tau. Front Neurol. 2014;5:26.
tion of tau in the mammalian central nervous system. 24. Zhou L, McInnes J, Wierda K, Holt M, Herrmann AG,
J Cell Biol. 1985;101:1371–8. Jackson RJ, et al. Tau association with synaptic ves-
10. Guo T, Noble W, Hanger DP.  Roles of tau pro-
icles causes presynaptic dysfunction. Nat Commun.
tein in health and disease. Acta Neuropathol. 2017;8:15295.
2017;133(5):665–704. 25. Pickett EK, Henstridge CM, Allison E, Pitstick R,
11. Maas T, Eidenmuller J, Brandt R.  Interaction of tau Pooler A, Wegmann S, et al. Spread of tau down neu-
with the neural membrane cortex is regulated by phos- ral circuits precedes synapse and neuronal loss in the
phorylation at sites that are modified in paired helical rTgTauEC mouse model of early Alzheimer’s disease.
filaments. J Biol Chem. 2000;275(21):15733–40. Synapse. 2017;71(6).
12. Pooler AM, Noble W, Hanger DP.  A role for tau at 26. Braak H, Del Tredici K.  The pathological process
the synapse in Alzheimer’s disease pathogenesis. underlying Alzheimer’s disease in individuals under
Neuropharmacology. 2014;76(Pt A):1–8. thirty. Acta Neuropathol. 2011;121(2):171–81.
13. Jackson RJ, Rudinskiy N, Herrmann AG, Croft S, 27. DeVos SL, Corjuc BT, Oakley DH, Nobuhara CK,
Kim JM, Petrova V, et al. Human tau increases amy- Bannon RN, Chase A, et al. Synaptic tau seeding pre-
loid beta plaque size but not amyloid beta-mediated cedes tau pathology in human Alzheimer’s disease
synapse loss in a novel mouse model of Alzheimer’s brain. Front Neurosci. 2018;12:267.
disease. Eur J Neurosci. 2016;44(12):3056–66. 28. Fein JA, Sokolow S, Miller CA, Vinters HV, Yang F,
14. Busche MA, Wegmann S, Dujardin S, Commins
Cole GM, et al. Co-localization of amyloid beta and
C, Schiantarelli J, Klickstein N, et  al. Tau impairs tau pathology in Alzheimer’s disease synaptosomes.
neural circuits, dominating amyloid-beta effects, Am J Pathol. 2008;172(6):1683–92.
in Alzheimer models in  vivo. Nat Neurosci. 29. Henkins KM, Sokolow S, Miller CA, Vinters HV,
2019;22(1):57–64. Poon WW, Cornwell LB, et al. Extensive p-tau pathol-
15. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A, ogy and SDS-stable p-tau oligomers in Alzheimer’s
van Eersel J, et al. Dendritic function of tau mediates cortical synapses. Brain Pathol. 2012;22(6):826–33.
9  Synaptic Localisation of Tau 111

30. Tai HC, Serrano-Pozo A, Hashimoto T, Frosch MP, 44. Ittner LM, Gotz J. Amyloid-beta and tau–a toxic pas
Spires-Jones TL, Hyman BT.  The synaptic accu- de deux in Alzheimer’s disease. Nat Rev Neurosci.
mulation of hyperphosphorylated tau oligomers in 2011;12(2):65–72.
Alzheimer disease is associated with dysfunction 45. Reynolds CH, Garwood CJ, Wray S, Price C, Kellie
of the ubiquitin-proteasome system. Am J  Pathol. S, Perera T, et  al. Phosphorylation regulates tau
2012;181(4):1426–35. interactions with Src homology 3 domains of phos-
31. DeKosky ST, Scheff SW. Synapse loss in frontal cor- phatidylinositol 3-kinase, phospholipase Cgamma1,
tex biopsies in Alzheimer’s disease: correlation with Grb2, and Src family kinases. J  Biol Chem.
cognitive severity. Ann Neurol. 1990;27(5):457–64. 2008;283(26):18177–86.
32. Yoshiyama Y, Higuchi M, Zhang B, Huang SM, Iwata 46. Pooler AM, Usardi A, Evans CJ, Philpott KL, Noble
N, Saido TC, et al. Synapse loss and microglial acti- W, Hanger DP. Dynamic association of tau with neu-
vation precede tangles in a P301S tauopathy mouse ronal membranes is regulated by phosphorylation.
model. Neuron. 2007;53(3):337–51. Neurobiol Aging. 2012;33(2):431.e27–38.
33. McInnes J, Wierda K, Snellinx A, Bounti L, Wang 47. Xia D, Li C, Götz J.  Pseudophosphorylation of tau
YC, Stancu IC, et  al. Synaptogyrin-3 mediates at distinct epitopes or the presence of the P301L
presynaptic dysfunction induced by tau. Neuron. mutation targets the microtubule-associated pro-
2018;97(4):823–35. tein tau to dendritic spines. Biochim Biophys Acta.
34. Perez-Nievas BG, Stein TD, Tai HC, Dols-Icardo O, 2015;1852(5):913–24.
Scotton TC, Barroeta-Espar I, et al. Dissecting pheno- 48. Frandemiche ML, De Seranno S, Rush T, Borel E,
typic traits linked to human resilience to Alzheimer’s Elie A, Arnal I, et  al. Activity-dependent tau pro-
pathology. Brain 2013;136(Pt 8):2510–2526. tein translocation to excitatory synapse is disrupted
35. Tai HC, Wang BY, Serrano-Pozo A, Frosch MP,
by exposure to amyloid-beta oligomers. J  Neurosci.
Spires-Jones TL, Hyman BT. Frequent and symmetric 2014;34(17):6084–97.
deposition of misfolded tau oligomers within presyn- 49. Tracy TE, Gan L. Acetylated tau in Alzheimer’s dis-
aptic and postsynaptic terminals in Alzheimer’s dis- ease: an instigator of synaptic dysfunction underlying
ease. Acta Neuropathol Commun. 2014;2:146. memory loss: increased levels of acetylated tau blocks
36. Kowall NW, Kosik KS. Axonal disruption and aber- the postsynaptic signaling required for plasticity and
rant localization of tau protein characterize the neu- promotes memory deficits associated with tauopathy.
ropil pathology of Alzheimer’s disease. Ann Neurol. BioEssays. 2017;39(4).
1987;22:639–43. 50. Rodriguez-Rodriguez E, Infante J, Llorca J, Mateo
37.
Liu C, Song X, Nisbet R, Götz I, Sanchez-Quintana C, Garcia-Gorostiaga I, et  al.
J. Co-immunoprecipitation with tau isoform-specific Age-dependent association of KIBRA genetic varia-
antibodies reveals distinct protein interactions and tion and Alzheimer’s disease risk. Neurobiol Aging.
highlights a putative role for 2N tau in disease. J Biol 2009;30(2):322–4.
Chem. 2016;291(15):8173–88. 51. Tracy TE, Sohn PD, Minami SS, Wang C, Min SW, Li
38. Hanger DP, Anderton BH, Noble W. Tau phosphoryla- Y, et al. Acetylated tau obstructs KIBRA-mediated sig-
tion: the therapeutic challenge for neurodegenerative naling in synaptic plasticity and promotes Tauopathy-­
disease. Trends Mol Med. 2009;15(3):112–9. related memory loss. Neuron. 2016;90(2):245–60.
39. Cabrales Fontela Y, Kadavath H, Biernat J, Riedel D, 52. Shrivastava AN, Redeker V, Pieri L, Bousset L,

Mandelkow E, Zweckstetter M.  Multivalent cross-­ Renner M, Madiona K, et  al. Clustering of tau
linking of actin filaments and microtubules through fibrils impairs the synaptic composition of alpha3-­-
the microtubule-associated protein tau. Nat Commun. Na(+)/K(+)-ATPase and AMPA receptors. EMBO
2017;8(1):1981. J. 2019;38:e99871.
40. Lee G, Newman ST, Gard DL, Band H, Panchamoorthy 53. Moussaud S, Jones DR, Moussaud-Lamodiere EL,
G. Tau interacts with src-family non-receptor tyrosine Delenclos M, Ross OA, McLean PJ. Alpha-synuclein
kinases. J Cell Sci. 1998;111(21):3167–77. and tau: teammates in neurodegeneration? Mol
41. Bhaskar K, Yen SH, Lee G. Disease-related modifica- Neurodegener. 2014;9:43.
tions in tau affect the interaction between Fyn and tau. 54. Miller EC, Teravskis PJ, Dummer BW, Zhao X,

J Biol Chem. 2005;280(42):35119–25. Huganir RL, Liao D.  Tau phosphorylation and tau
42. Usardi A, Pooler AM, Seereeram A, Reynolds CH, mislocalization mediate soluble Abeta oligomer-­
Derkinderen P, Anderton B, et al. Tyrosine phosphor- induced AMPA glutamate receptor signaling deficits.
ylation of tau regulates its interactions with Fyn SH2 Eur J Neurosci. 2014;39(7):1214–24.
domains, but not SH3 domains, altering the cellular 55. Kopeikina KJ, Polydoro M, Tai HC, Yaeger E,

localization of tau. FEBS J. 2011;278(16):2927–37. Carlson GA, Pitstick R, et  al. Synaptic alterations
43. Lau DH, Hogseth M, Phillips EC, O’Neill MJ, Pooler in the rTg4510 mouse model of tauopathy. J  Comp
AM, Noble W, et al. Critical residues involved in tau Neurol. 2013;521(6):1334–53.
binding to fyn: implications for tau phosphorylation 56. Bolos M, Pallas-Bazarra N, Terreros-Roncal J, Perea
in Alzheimer’s disease. Acta Neuropathol Commun. JR, Jurado-Arjona J, Avila J, et  al. Soluble tau has
2016;4(1):49. devastating effects on the structural plasticity of
112 D. P. Hanger et al.

hippocampal granule neurons. Transl Psychiatry. late the power of brain rhythms. Neurobiol Dis.
2017;7(12):1267. 2018;117:181–8.
57. Jurado S.  AMPA receptor trafficking in natural and 63. Sen A, Capelli V, Husain M.  Cognition and

pathological aging. Front Mol Neurosci. 2017;10:446. dementia in older patients with epilepsy. Brain.
58. Zhao X, Kotilinek LA, Smith B, Hlynialuk C, Zahs K, 2018;141(6):1592–608.
Ramsden M, et al. Caspase-2 cleavage of tau revers- 64. Pooler AM, Phillips EC, Lau DH, Noble W,

ibly impairs memory. Nat Med. 2016;22(11):1268–76. Hanger DP.  Physiological release of endogenous
59. Ondrejcak T, Klyubin I, Corbett GT, Fraser G, Hong tau is stimulated by neuronal activity. EMBO Rep.
W, Mably AJ, et al. Cellular prion protein mediates the 2013;14(4):389–94.
disruption of hippocampal synaptic plasticity by solu- 65. Yamada K, Holth JK, Liao F, Stewart FR, Mahan TE,
ble tau in vivo. J Neurosci. 2018;38(50):10595–606. Jiang H, et al. Neuronal activity regulates extracellu-
60. Angulo SL, Orman R, Neymotin SA, Liu L, Buitrago lar tau in vivo. J Exp Med. 2014;211(3):387–93.
L, Cepeda-Prado E, et  al. Tau and amyloid-related 66. Wu JW, Hussaini SA, Bastille IM, Rodriguez GA,
pathologies in the entorhinal cortex have divergent Mrejeru A, Rilett K, et al. Neuronal activity enhances
effects in the hippocampal circuit. Neurobiol Dis. tau propagation and tau pathology in  vivo. Nat
2017;108:261–76. Neurosci. 2016;19(8):1085–92.
61. Vossel KA, Tartaglia MC, Nygaard HB, Zeman AZ, 67. Croft CL, Wade MA, Kurbatskaya K, Mastrandreas
Miller BL.  Epileptic activity in Alzheimer’s dis- P, Hughes MM, Phillips EC, et  al. Membrane asso-
ease: causes and clinical relevance. Lancet Neurol. ciation and release of wild-type and pathological tau
2017;16(4):311–22. from organotypic brain slice cultures. Cell Death Dis.
62. Das M, Maeda S, Hu B, Yu GQ, Guo W, Lopez I, 2017;8(3):e2671.
et  al. Neuronal levels and sequence of tau modu-
 The Role of Tau
in the Post-synapse 10
Philip Regan and Kwangwook Cho

Abbreviations logical hallmark of Alzheimer’s disease (AD)

consisting of insoluble and aggregated protein
Aβ Amyloid beta tau [1]. To date, the majority of tau biology
AD Alzheimer’s disease research has focused on the processes of tau
AMPAR α-amino-3-hydroxy-5-methyl-4-­ aggregation and its subsequent toxicity in neu-
isoxazolepropionic acid receptor rons. Recently, the post-synaptic localisation of
GSK-3 Glycogen synthase kinase-3 tau has been documented, suggesting that tau has
LTD Long-term depression physiological functions not only in the axon but
LTP Long-term potentiation as a synaptic protein [2–4]. Synaptic tau, how-
NMDAR N-methyl-D-aspartate receptor ever, remains a debatable subject and its precise
pTau Phosphorylated Tau synaptic function is currently uncertain.
Tau Microtubule associated protein tau Studies of tau knockout mice have provided
significant but inconclusive input into our under-
standing of the synaptic role of tau. Several phys-
iologically ‘normal’ phenotypes of tau knockout
I ntroduction mice have been reported, including normal cog-
nitive function in spatial learning tasks (see
Tau mostly localises in the axon and regulates review, Ke et  al. [5]). In contrast, some studies
axonal transport of cargo along microtubules. have described tau knockout mice with aging-­
Aberrant hyperphosphorylation of tau causes dependent memory deficits [6], impaired contex-
detachment of tau from microtubules and pre- tual fear conditioning, yet enhanced spatial
cipitates neurofibrillary tangles (NFTs), a patho- learning [7] and deficits in spatial reversal learn-
ing [4]. Mixed effects of tau deletion on synaptic
P. Regan plasticity are also evident, with reports of impair-
Henry Wellcome Laboratories for Integrative ments to long-term potentiation (LTP) of synap-
Neuroscience and Endocrinology, Bristol Medical tic transmission [6, 7] that are not evident in
School, Faculty of Health Sciences, University of another study, in which specific impairments to
Bristol, Bristol, UK
long-term depression (LTD) are observed [3]. To
K. Cho (*) add to this confusing picture, tau deletion alters
UK-Dementia Research Institute, Department of
Basic and Clinical Neuroscience, Maurice Wohl dendritic architecture and spine density in some
Clinical Neuroscience Institute, King’s College studies [8, 9] but not others [10, 11], the latter
London, London, UK being more consistent with reports of unaltered
e-mail: kei.cho@kcl.ac.uk

© Springer Nature Singapore Pte Ltd. 2019 113

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_10
114 P. Regan and K. Cho

basal synapse function following tau deletion [3,

6, 7, 12]. What is clear from the sum of these
studies is that tau must play a complex role at the
synapse, differentially affecting specific forms of
synaptic plasticity that may support specific cog-
nitive functions.
To delineate the role of tau in synaptic plastic-
ity, Kimura et al. [3] extended their in vivo findings
to an in vitro model in which confounding com-
pensatory and pre-synaptic variables were reduced. Fig. 10.1  Using immunogold electron microscopy (EM)
Corroborating their in vivo data, tau shRNA trans- and labelling tau antibody (JM, rabbit polyclonal anti-tau
fection prevented the induction of LTD, while LTP antibody), tau is found in both the pre and postsynaptic
compartment of hippocampal tissue obtained from
was unperturbed. Replacement of endogenous tau MAPT+/+ (left panel; 4-month-old) but there is no tau sig-
with human tau rescued the LTD phenotype, sup- nal (JM anti-tau antibody) in MAPT−/− (right panel;
porting a necessary role for tau in the hippocampal 4-month-old) mice. Arrow (S) shows postsynaptic density
expression of an NMDAR-­ dependent form of and arrowheads indicate tau. (Figures are taken and
adapted from Kimura et al. (2014), Phil. Trans. R. Soc. B
LTD.  Several studies have now linked spatial 369: 20130144 [3])
reversal learning, a form of behavioural flexibility,
to the molecular mechanisms of hippocampal
NMDAR-LTD [13–15]. The requirement for tau Translation of synaptic protein tau may occur
in hippocampal LTD is therefore quite striking in locally within dendrites [22, 23], or phosphory-
the context of the selective deficits in spatial rever- lated isoforms of tau may overcome an axonal
sal learning that have been observed in tau knock- diffusion barrier allowing diffusion into the
out mice [4]. somatodendritic compartment [24]. Alternatively,
These findings, together with the alterations to tau released from pre-synaptic terminals may be
synapse function that are consistently observed trans-synaptically internalized into post-synaptic
in experimental models of tauopathies [11, 16, regions following neuronal activity [25–27].
17] and the requirement for tau during BDNF-­ Spreading of tau via synaptic contacts might con-
driven synaptic plasticity [8], clearly highlight tribute to both physiological and pathological
the numerous pathophysiological implications tau-mediated consequences in neuronal function
for tau’s function at the synapse. Uncovering the [28, 29]. In this respect, neuronal activity and
precise nature of this function is a necessary external stimuli appear to be critical regulators of
milestone for tau biology research. tau subcellular location. For example, increased
neuronal activity or treatment with amyloid-beta
(Aβ) or glucocorticoids can increase the dendritic
 ocalization of Tau: Axon vs.
L and synaptic localization of tau [30, 31].
Dendrite

Tau is generally considered an axonally segre-  ow Does Tau Affect Synaptic

H
gated protein [18]. Indeed, some recent research Function?
agrees with this notion [19], supporting early
findings that tau is only present in dendrites and The post-synaptic mechanisms by which tau can
synapses either during neuronal development or affect synapse function are, by the nature of accu-
under pathological conditions [20, 21]. mulating evidence, either complex or difficult to
Nevertheless, overwhelming evidence that tau resolve. Except for muscarinic acetylcholine
exists in dendrites and pre- and post-synaptic receptors (mAChRs) [32, 33], tau is not known to
structures of healthy mature neurons must now bind directly to any synaptic receptors and is
also be considered [3, 4] (Fig. 10.1). therefore likely affecting synapse function
10  The Role of Tau in the Post-synapse 115

through interactions with signalling pathways, of Tau-shRNA transfected neurons have higher lev-
which some possibilities are outlined below. els of intra-dendritic GluA2 [37], consistent with
Tau has been demonstrated to bind to the Src impaired AMPAR trafficking at extrasynaptic
family kinase Fyn, and one hypothesis is that this sites where PICK1 primarily operates [39]. Given
binding is necessary for the synaptic localization that PICK1-driven AMPAR internalization is a
of Fyn [2]. Indeed, Fyn fails to localize to the key post-synaptic step in the manifestation of
synapse in tau-depleted neurons [2, 9]. Tau, Fyn, LTD [40], these findings are consistent with the
PSD-95 and NMDARs are predicted to form a LTD deficits observed upon tau depletion.
protein complex at the synapse [34], wherein However, PICK1-GluA2 interactions do appear
Fyn-mediated tyrosine phosphorylation of to have complex effects on AMPAR trafficking,
NMDAR GluN2B subunits facilitates synaptic including synaptic AMPAR delivery [39] and
NMDAR function, with both physiological and intracellular AMPAR retention [41], which may
pathological implications [2, 35]. Thus, tau and go some way to explaining the mixed reports of
Fyn may work in tandem at the synapse to regu- the effects of tau depletion upon synaptic plastic-
late a post-synaptic scaffolding complex that ity. A direct interaction between tau and PICK1
links GluN2B containing NMDARs to intracel- has not been reported, and therefore the precise
lular signaling cascades that affect synapse func- manner by which tau may regulate PICK1-GluA2
tion. Whether tau-Fyn binding is truly necessary interactions is currently unclear (Fig. 10.2).
for these effects is disputed [36] and their inher- Though no direct evidence currently exists to
ent physiological relevance is difficult to recon- support it, tau regulation of the cytoskeleton may
cile with observations of unaltered basal NMDAR have important implications for the regulation of
function in tau knockout synapses [3]. synapse function. By binding to microtubules
Tau-shRNA transfected neurons display and regulating their turnover, tau may modulate
impaired AMPAR removal from the synapse fol- the dynamic movement of microtubules into and
lowing LTD-inducing stimuli [3, 37], and mount- out of dendritic spines  – an important facet of
ing evidence supports a role for tau in the LTD [42]. Similarly, through direct binding to
regulation of an AMPAR internalization mecha- actin [43] or regulation of microtubule entry into
nism involving protein interacting with protein spines, tau could thus affect the cross-talk
kinase C (PICK1). Tau depletion impairs the between the microtubule and actin cytoskeletal
interaction between the AMPAR subunit GluA2 networks [44] – the latter being an important reg-
and PICK1 [4] and the presence of tau enhances ulator of synapse function and AMPAR traffick-
GluA2-PICK1 interaction [38]. Furthermore, ing [45, 46]. Clearly, further work is required to

Fig. 10.2  NMDA treatment (25 μM for 3 min) increases blot data. (b) Pooled data, all bars represent the
GluA2-PICK1 association in hippocampal slices of wild- mean ± SEM (∗p < 0.05). (Figures are taken and adapted
type mice (MAPT+/+) but there is no effect in tau knockout from Regan et al. (2015), J Neurosci 35, 4804–4812 [4])
mice (MAPT−/−). (a) Representative example of western
116 P. Regan and K. Cho

establish which mechanism(s) enable tau to exert pTau has been implicated in synapse weakening
its effects on synapse function. in Alzheimer’s disease [55, 56].
Since the identification of tau within presyn- The activity of one of the most prominent tau
aptic regions [3, 47], there is a notion that its kinases, GSK-3, is strongly influenced by pat-
regulation may have an important role in presyn- terns of neuronal activity that induce synaptic
aptic function. For instance, the accumulation of plasticity, and in turn, is a key driver of plasticity
pathological tau reduces synaptic vesicles in the induction. For example, an LTP-inducing stimu-
mossy fibers and dysregulates basal synaptic lus inhibits GSK-3, while its activity is critically
transmission, and certain pathological forms of up-regulated during LTD [57]. This correlates
tau interact with synapse vesicles to impair pre- with findings that LTD-inducing stimuli cause
synaptic function [48, 49]. The synaptic vesicle is phosphorylation of tau at GSK-3 substrate resi-
an essential part of synaptic transmission [50], dues in an NMDAR-dependent manner [3, 4, 34],
making it plausible that the presynaptic role of raising the possibility that LTD is associated with
tau indirectly affects postsynaptic function asso- pTau downstream of GSK-3 activity. Similarly,
ciated with learning and memory processes [51]. depotentiation of LTP involves hyperphosphory-
However, there is no evidence of differences in lation of tau at residues downstream of ERK1/2
presynaptic-mediated transmission in wild type activation [58]. These data indicate that the phos-
and tau knockout mice (MAPT−/−) [3]. This sug- phorylation status of tau is dynamically modified
gests that normal, or physiological, forms of tau by neuronal and synaptic stimuli. Could pTau
have no role in presynaptic-mediated transmis- therefore be key mechanistic step in plasticity
sion, but pathological forms of tau dysregulate processes?
presynaptic function. Therefore, the regulation of Mutagenesis studies, in which,
tau in presynaptic and postsynaptic regions may phosphorylation-­prone tau residues are mutated
be essential for both the physiological and patho- to prevent their phosphorylation (serine to ala-
physiological weakening of synaptic nine mutation), have gone some way to address-
transmission. ing the synaptic effects of tau phosphorylation.
Preventing pTau at serine residue 396 (relative to
full length human tau) effectively blocked neuro-
Phosphorylated Tau at the Synapse nal induction of LTD but not LTP, while prevent-
and Functional Consequences ing phosphorylation at other residues had no
effect [4] (Fig. 10.3).
Tau can be phosphorylated by multiple enzymes These results indicate that an individual phos-
and 45 sites of phosphorylation have been identi- phorylation site of tau plays an important role in
fied (Y18, S46, S68, S69, T71, S113, T123, affecting synapse function, though the reasoning
T153, T175, T181, S184, S185, S191, Y197, for such a specific requirement in LTD is not
S198, S199, S202, T205, S208, S210, T212, clear. pTau has been shown to affect its interac-
S214, T217, T231, S235, S237, S238, S258, tions with synaptic proteins [34] and can be nec-
S262, S289, S356, Y394, S396, S400, T403, essary for the synaptic recruitment of tau [30],
S404, S409, S412, S413, T414, S416, S422, though these effects are not known to be directly
T427, S433, S435) (see review, Hanger et  al. linked to tau serine 396 phosphorylation.
[52]). Multiple kinases can regulate the status of The pathophysiological relevance of the above
tau phosphorylation and, thus, tau function. findings is demonstrable through evidence of ser-
Certain phosphorylation sites of tau are closely ine 396 phosphorylation of tau in pathologies [59–
associated with neuropathology, including neuro- 61], including the early stages of AD. Given that
fibrillary tangles [53] and aberrant hyperphos- synapse loss is widely considered to underly early
phorylation is implicated in numerous forms of cognitive deficits in AD [62] it is possible that
tauopathy [54]. Specifically, GSK-3 mediated elevated serine 396 phosphorylation of tau is rep-
resentative of hyperactive LTD, or synapse weak-
10  The Role of Tau in the Post-synapse 117

Fig. 10.3 (a) Schematic diagram of the microdissection circles) but not in neighbouring neurons transfected with
procedure to separate the rat P24 – 28 CA1 somatic and rat tau-shRNA plus S396/404A or S396A phosphomutant
dendritic regions after 1  Hz electric stimulation at CA3 human tau (closed circles). In contrast, LTD is inducible in
synapse. 1  Hz stimulation induces pTau at the S396 and S/T199/202/205A or S404A phosphomutant human tau
S404 residues, an effect blocked by the NMDA antagonist transfected neurons. All data represent the mean  ±  SEM
AP5. B, LFS (black bar; 200 pulses at 1 Hz, holding volt- (∗∗∗p  <  0.001). (Figures are taken and adapted from
age −40 mV) induces LTD in untransfected neurons (open Regan et al. (2015), J Neurosci 35, 4804–4812 [4])

ening. Credence for this theory comes from AD logical synapse weakening signaling. Indeed, tau
models, where a caspase-3–Akt–GSK-3 signalling deletion is protective against the deleterious effects
cascade, which drives NMDAR-­dependent LTD, of Aβ on LTP [56]. pTau is also observed in other
is necessary for the Aβ-mediated inhibition of LTP stages of synapse weakening or degeneration, such
[55, 56, 63]. Thus, pTau, as a GSK-3 substrate, is a as exposure to chronic stress, diabetic retinopathy
putative downstream effector of this pathophysio- or hibernation [64–67]. Similar to Aβ, chronic
118 P. Regan and K. Cho

stress and elevated glucocorticoids inhibit hippo- implications for diagnosing and treating AD
campal LTP in a tau and serine-396 effectively.
phosphorylation-­dependent manner [66, 68].
Clearly, pTau may prove to be a key pathophysio-
logical synapse effector for synapse weakening. References
1. Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M,
Wisniewski HM, Binder LI. Abnormal phosphoryla-
Tau-Mediated Synapse Weakening tion of the microtubule-associated protein tau (tau) in
and Readout of AD Pathology Alzheimer cytoskeletal pathology. Proc Natl Acad Sci
U S A. 1986;83:4913–7.
Central to synapse weakening, several studies 2. Ittner LM, et  al. Dendritic function of tau mediates
amyloid-β toxicity in Alzheimer’s disease mouse
demonstrate a molecular mechanism whereby the models. Cell. 2010;142:387–97.
endocytosis of synaptic AMPARs appears to be 3. Kimura T, Whitcomb DJ, Jo J, Regan P, Piers T, Heo
pathologically enhanced in AD.  Persistence of S, Brown C, Hashikawa T, Murayama M, Seok H,
this aberrant synapse weakening disrupts physi- Sotiropoulos I, Kim E, Collingridge GL, Takashima
A, Cho K.  Microtubule-associated protein tau is
ological synaptic plasticity [55, 56] and may be a essential for long-term depression in the hippo-
key pathophysiological entity resulting in AD campus. Philos Trans R Soc Lond Ser B Biol Sci.
mediated cognitive deficits. In this context, 2014;369(1633):20130144.
GSK-3 mediated pTau is a favorable hypothesis 4. Regan P, Piers T, Yi J-H, Kim D-H, Huh S, Park SJ,
Ryu JH, Whitcomb DJ, Cho K. Tau phosphorylation
for synapse weakening. In animal studies, GSK-3 at serine 396 residue is required for hippocampal
inhibitors prevent the induction of LTD [57] and LTD. J Neurosci. 2015;35:4804–12.
neurodegeneration [69]. In clinical trials, a 5. Ke YD, Suchowerska AK, van der Hoven J, De Silva
GSK-3 inhibitor has shown cognitive improve- DM, Wu CW, van Eersel J, Ittner A, Ittner LM. Lessons
from tau-deficient mice. Int J  Alzheimers Dis.
ments when treated for long periods in AD 2012;2012:873270.
patients with MCI [70]. This indicates that aber- 6. Biundo F, Del Prete D, Zhang H, Arancio O,
rant GSK-3 mediated pTau causes pathological D’Adamio L. A role for tau in learning, memory and
synapse weakening and this inhibits physiologi- synaptic plasticity. Sci Rep. 2018;8(1):3184.
7. Ahmed T, Van der Jeugd A, Blum D, Galas MC,
cal plasticity (e.g., inhibition of LTP). Indeed, D’Hooge R, Buee L, Balschun D. Cognition and hippo-
LTP is impaired in AD models [55, 71], and dele- campal synaptic plasticity in mice with a homozygous
tion of tau rescues this physiological synaptic tau deletion. Neurobiol Aging. 2014;35(11):2474–8.
plasticity [56]. 8. Chen Q, Zhou Z, Zhang L, Wang Y, Zhang YW,
Zhong M, Xu SC, Chen CH, Li L, Yu ZP. Tau protein
Tau also interacts with the AD-related proteins is involved in morphological plasticity in hippocam-
such as ApoE, BIN1 and Aβ [72–74]. The rela- pal neurons in response to BDNF.  Neurochem Int.
tionship between tau and such AD-related pro- 2012;60(3):233–42.
teins reveals a possible mechanism for 9. Sapir T, Frotscher M, Levy T, Mandelkow EM, Reiner
O.  Tau’s role in the developing brain: ­implications
clinicopathological development of sporadic for intellectual disability. Hum Mol Genet.
AD. This suggests that tau plays a key role in the 2012;21(8):1681–92.
progress of pathology and subsequently cogni- 10. Tackenberg C, Grinschgl S, Trutzel A, Santuccione
tive impairment through functional modification AC, Frey MC, Konietzko U, Grimm J, Brandt R,
Nitsch RM.  NMDA receptor subunit composition
of synapses. Aβ deposition is an early pathologi- determines beta-amyloid-induced neurodegeneration
cal signature of AD, but in itself is not sufficient and synaptic loss. Cell Death Dis. 2013;4:e608.
to produce the clinical symptom (see review, Jack 11. Zempel H, Luedtke J, Kumar Y, Biernat J, Dawson
Jr. et  al. [75]). Arguably, the development of H, Mandelkow E, Mandelkow EM. Amyloid-β oligo-
mers induce synaptic damage via Tau-dependent
tauopathy is much more closely associated with microtubule severing by TTLL6 and spastin. EMBO
neuronal silencing [76] and progressive cognitive J. 2013;32(22):2920–37.
impairment [77]. Therefore, tau-mediated modi- 12. Roberson ED, Halabisky B, Yoo JW, Yao J, Chin J,
fication of synapse function may have important Yan F, Wu T, Hamto P, Devidze N, Yu GQ, Palop JJ,
Noebels JL, Mucke L.  Amyloid-β/Fyn-induced syn-
10  The Role of Tau in the Post-synapse 119

aptic, network, and cognitive impairments depend on 25. Liu L, Drouet V, Wu JW, Witter MP, Small SA,

tau levels in multiple mouse models of Alzheimer’s Clelland C, Duff K.  Trans-synaptic spread of tau
disease. J Neurosci. 2011;31(2):700–11. pathology in vivo. PLoS One. 2012;7(2):e31302.
13. Nicholls RE, Alarcon JM, Malleret G, Carroll RC, 26. Pooler AM, Phillips EC, Lau DH, Noble W,

Grody M, Vronskaya S, Kandel ER. Transgenic mice Hanger DP.  Physiological release of endogenous
lacking NMDAR-dependent LTD exhibit deficits in tau is stimulated by neuronal activity. EMBO Rep.
behavioral flexibility. Neuron. 2008;58(1):104–17. 2013;14(4):389–94.
14. Kim JI, Lee HR, Sim SE, Baek J, Yu NK, Choi JH, 27. Yamada K, Holth JK, Liao F, Stewart FR, Mahan
Ko HG, Lee YS, Park SW, Kwak C, Ahn SJ, Choi SY, TE, Jiang H, Cirrito JR, Patel TK, Hochgräfe K,
Kim H, Kim KH, Backx PH, Bradley CA, Kim E, Mandelkow EM, Holtzman DM.  Neuronal activ-
Jang DJ, Lee K, Kim SJ, Zhuo M, Collingridge GL, ity regulates extracellular tau in  vivo. J  Exp Med.
Kaang BK.  PI3Kγ is required for NMDA receptor-­ 2014;211(3):387–93.
dependent long-term depression and behavioral flex- 28. Calafate S, et  al. Synaptic contacts enhance cell-­

ibility. Nat Neurosci. 2011;14(11):1447–54. to-­cell Tau pathology propagation. Cell Rep.
15. Dong Z, Bai Y, Wu X, Li H, Gong B, Howland JG, 2015;11:1176–83.
Huang Y, He W, Li T, Wang YT. Hippocampal long-­ 29. de Calignon A, et  al. Propagation of Tau pathology
term depression mediates spatial reversal learn- in a model of early Alzheimer’s disease. Neuron.
ing in the Morris water maze. Neuropharmacology. 2012;73:685–97.
2013;64:65–73. 30. Frandemiche ML, De Seranno S, Rush T, Borel E, Elie
16. Biundo F, d’Abramo C, Tambini MD, Zhang H, Del A, Arnal I, Lanté F, Buisson A.  Activity-dependent
Prete D, Vitale F, Giliberto L, Arancio O, D’Adamio tau protein translocation to excitatory synapse is
L.  Abolishing Tau cleavage by caspases at aspar- disrupted by exposure to amyloid-beta oligomers.
tate(421) causes memory/synaptic plasticity defi- J Neurosci. 2014;34:6084–97.
cits and pre-pathological Tau alterations. Transl 31. Pinheiro S, Silva J, Mota C, Vaz-Silva J, Veloso A,
Psychiatry. 2017;7(8):e1198. Pinto V, Sousa N, Cerqueira J, Sotiropoulos I.  Tau
17. Jackson JS, Witton J, Johnson JD, Ahmed Z, Ward M, Mislocation in glucocorticoid-triggered hippocampal
Randall AD, Hutton ML, Isaac JT, O’Neill MJ, Ashby pathology. Mol Neurobiol. 2016;53(7):4745–53.
MC.  Altered synapse stability in the early stages of 32. Gómez-Ramos A, Díaz-Hernández M, Rubio A,

tauopathy. Cell Rep. 2017;18(13):3063–8. Miras-Portugal MT, Avila J.  Extracellular tau pro-
18. Binder LI, Frankfurter A, Rebhun LI.  The distribu- motes intracellular calcium increase through M1 and
tion of tau in the mammalian central nervous system. M3 muscarinic receptors in neuronal cells. Mol Cell
J Cell Biol. 1985;101(4):1371–8. Neurosci. 2008;37(4):673–81.
19. Kubo A, Misonou H, Matsuyama M, Nomori A,
33. Gómez-Ramos A, Díaz-Hernández M, Rubio A,

Wada-Kakuda S, Takashima A, Kawata M, Murayama Díaz-Hernández JI, Miras-Portugal MT, Avila
S, Ihara Y, Miyasaka T.  Distribution of endogenous J.  Characteristics and consequences of mus-
normal tau in the mouse brain. J  Comp Neurol. carinic receptor activation by tau protein. Eur
2019;527(5):985–98. Neuropsychopharmacol. 2009;19(10):708–17.
20. Kosik KS, Finch EA. MAP 2 and tau segregate into 34. Mondragón-Rodríguez S, Trillaud-Doppia E, Dudilot
dendritic and axonal domains after the elaboration of A, Bourgeois C, Lauzon M, Leclerc N, Boehm
morphologically distinct neurites: an immunocyto- J.  Interaction of endogenous tau protein with syn-
chemical study of cultured rat cerebrum. J Neurosci. aptic proteins is regulated by N-methyl-D-aspartate
1987;7(10):3142–53. receptor-­dependent tau phosphorylation. J Biol Chem.
21. Kowall NW, Kosik KS. Axonal disruption and aber- 2012;287(38):32040–53.
rant localization of tau protein characterize the neu- 35. Nakazawa T, Komai S, Tezuka T, Hisatsune C,

ropil pathology of Alzheimer’s disease. Ann Neurol. Umemori H, Semba K, Mishina M, Manabe T,
1987;22(5):639–43. Yamamoto T. Characterization of Fyn-mediated tyro-
22. Malmqvist T, Anthony K, Gallo JM.  Tau mRNA is sine phosphorylation sites on GluR epsilon 2 (NR2B)
present in axonal RNA granules and is associated with subunit of the N-methyl-D-aspartate receptor. J Biol
elongation factor 1A. Brain Res. 2014;1584:22–7. Chem. 2001;276(1):693–9.
23. Li C, Götz J.  Somatodendritic accumulation of Tau 36. Miyamoto T, Stein L, Thomas R, Djukic B, Taneja P,
in Alzheimer’s disease is promoted by Fyn-mediated Knox J, Vossel K, Mucke L. Phosphorylation of tau
local protein translation. EMBO J.  2017 Nov at Y18, but not tau-fyn binding, is required for tau
2;36(21):3120–313. to modulate NMDA receptor-dependent excitotoxic-
24. Li X, Kumar Y, Zempel H, Mandelkow EM, Biernat ity in primary neuronal culture. Mol Neurodegener.
J, Mandelkow E.  Novel diffusion barrier for axonal 2017;12(1):41.
retention of Tau in neurons and its failure in neurode- 37.
Suzuki M, Kimura T.  Microtubule-associated
generation. EMBO J. 2011;30(23):4825–37. tau contributes to intra-dendritic trafficking of
AMPA receptors in multiple ways. Neurosci Lett.
2017;653:276–82.
120 P. Regan and K. Cho

38. Yagishita S, Murayama M, Ebihara T, Maruyama 52. Hanger DH, Anderson BH, Noble W. Tau phosphory-
K, Takashima A.  Glycogen synthase kinase lation: the therapeutic challenge for neurodegenera-
3β-mediated phosphorylation in the most tive disease. Trends Mol Med. 2009;15:112–9.
C-terminal region of protein interacting with C 53. Bramblett GT, et al. Abnormal tau phosphorylation at
kinase 1 (PICK1) regulates the binding of PICK1 Ser396 in Alzheimer’s disease recapitulates develop-
to glutamate receptor subunit GluA2. J Biol Chem. ment and contributes to reduced microtubule binding.
2015;290(49):29438–48. Neuron. 1993;10:1089–99.
39. Gardner SM, Takamiya K, Xia J, Suh JG, Johnson R, 54. Iqbal K, Liu F, Gong CX, Grundke-Iqbal I.  Tau in
Yu S, Huganir RL. Calcium-permeable AMPA receptor Alzheimer disease and related tauopathies. Curr
plasticity is mediated by subunit-­specific interactions Alzheimer Res. 2010;7(8):656–64.
with PICK1 and NSF. Neuron. 2005;45(6):903–15. 55. Jo J, Whitcomb DJ, Olsen KM, Kerrigan TL, Lo SC,
40. Citri A, Bhattacharyya S, Ma C, Morishita W,
Bru-Mercier G, Dickinson B, Scullion S, Sheng M,
Fang S, Rizo J, Malenka RC.  Calcium binding to Collingridge G, Cho K. Aβ(1-42) inhibition of LTP is
PICK1 is essential for the intracellular retention of mediated by a signaling pathway involving caspase-3,
AMPA receptors underlying long-term depression. Akt1 and GSK-3β. Nat Neurosci. 2011;14(5):545–7.
J Neurosci. 2010;30(49):16437–52. 56. Shipton OA, Leitz JR, Dworzak J, Acton CE,

41. Lin DT, Huganir RL.  PICK1 and phosphoryla-
Tunbridge EM, Denk F, Dawson HN, Vitek MP,
tion of the glutamate receptor 2 (GluR2) AMPA Wade-Martins R, Paulsen O, Vargas-Caballero
receptor subunit regulates GluR2 recycling after M. Tau protein is required for amyloid {beta}-induced
NMDA receptor-induced internalization. J Neurosci. impairment of hippocampal long-term potentiation.
2007;27(50):13903–8. J Neurosci. 2011;31(5):1688–92.
42. Kapitein LC, Yau KW, Gouveia SM, van der Zwan 57. Peineau S, Taghibiglou C, Bradley C, Wong TP, Liu
WA, Wulf PS, Keijzer N, Demmers J, Jaworski J, L, Lu J, Lo E, Wu D, Saule E, Bouschet T, Matthews
Akhmanova A, Hoogenraad CC.  NMDA receptor P, Isaac JT, Bortolotto ZA, Wang YT, Collingridge
activation suppresses microtubule growth and spine GL. LTP inhibits LTD in the hippocampus via regula-
entry. J Neurosci. 2011;31(22):8194–209. tion of GSK3beta. Neuron. 2007;53(5):703–17.
43. Fulga TA, Elson-Schwab I, Khurana V, Steinhilb
58. Babür E, Tan B, Delibaş S, Yousef M, Dursun N, Süer
ML, Spires TL, Hyman BT, Feany MB.  Abnormal C. Depotentiation of long-term potentiation is associ-
bundling and accumulation of F-actin mediates tau-­ ated with epitope-specific Tau hyper-/hypophosphory-
induced neuronal degeneration in vivo. Nat Cell Biol. lation in the hippocampus of adult rats. J Mol Neurosci.
2007;9(2):139–48. 2018;67 https://doi.org/10.1007/s12031-018-1224-x.
44. Jaworski J, Kapitein LC, Gouveia SM, Dortland
59. Hu YY, He SS, Wang X, Duan QH, Grundke-Iqbal
BR, Wulf PS, Grigoriev I, Camera P, Spangler SA, I, Iqbal K, Wang J.  Levels of nonphosphorylated
Di Stefano P, Demmers J, Krugers H, Defilippi P, and phosphorylated tau in cerebrospinal fluid of
Akhmanova A, Hoogenraad CC.  Dynamic microtu- Alzheimer’s disease patients: an ultrasensitive
bules regulate dendritic spine morphology and synap- bienzyme-­substrate-recycle enzyme-linked immuno-
tic plasticity. Neuron. 2009;61(1):85–100. sorbent assay. Am J Pathol. 2002;160(4):1269–78.
45. Zhou Q, Xiao M, Nicoll RA.  Contribution of cyto- 60. Alonso Adel C, Mederlyova A, Novak M, Grundke-­
skeleton to the internalization of AMPA receptors. Iqbal I, Iqbal K. Promotion of hyperphosphorylation
Proc Natl Acad Sci U S A. 2001;98(3):1261–6. by frontotemporal dementia tau mutations. J  Biol
46. Zhou Q, Homma KJ, Poo MM. Shrinkage of dendritic Chem. 2004;279(33):34873–81.
spines associated with long-term depression of hip- 61. Mondragón-Rodríguez S, Perry G, Luna-Muñoz J,
pocampal synapses. Neuron. 2004;44(5):749–57. Acevedo-Aquino MC, Williams S. Phosphorylation of
47. Sokolow S, et  al. Pre-synaptic C-terminal truncated tau protein at sites Ser(396-404) is one of the earliest
tau is released from cortical synapses in Alzheimer’s events in Alzheimer’s disease and Down syndrome.
disease. J Neurochem. 2015;133:368–79. Neuropathol Appl Neurobiol. 2014;40(2):121–35.
48. Decker JM, et  al. Pro-aggregant Tau impairs mossy 62. Selkoe DJ. Alzheimer’s disease is a synaptic failure.
fiber plasticity due to structural changes and Ca++ dys- Science. 2002;298(5594):789–91.
regulation. Acta Neuropathol Commun. 2015;3:23. 63. Li Z, Jo J, Jia JM, Lo SC, Whitcomb DJ, Jiao S, Cho
49. Zhou L, McInnes J, Wierda K, Holt M, Herrmann AG, K, Sheng M. Caspase-3 activation via mitochondria is
Jackson RJ, Wang YC, Swerts J, Beyens J, Miskiewicz required for long-term depression and AMPA receptor
K, Vilain S, Dewachter I, Moechars D, De Strooper internalization. Cell. 2010;141(5):859–71.
B, Spires-Jones TL, De Wit J, Verstreken P. Tau asso- 64. Sotiropoulos I, Catania C, Pinto LG, Silva R,

ciation with synaptic vesicles causes presynaptic dys- Pollerberg GE, Takashima A, Sousa N, Almeida
function. Nat Commun. 2017;8:15295. OF. Stress acts cumulatively to precipitate Alzheimer’s
50. Lin RC, Scheller RH. Mechanisms of synaptic vesicle disease-like tau pathology and cognitive deficits.
exocytosis. Annu Rev Cell Dev Biol. 2000;16:19–49. J Neurosci. 2011;31(21):7840–7.
51. DeKosky ST, Scheff SW. Synapse loss in frontal cor- 65. Arendt T, Bullmann T.  Neuronal plasticity in hiber-
tex biopsies in Alzheimer’s disease: correlation with nation and the proposed role of the microtubule-­
cognitive severity. Ann Neurol. 1990;27:457e464. associated protein tau as a “master switch” regulating
10  The Role of Tau in the Post-synapse 121

synaptic gain in neuronal networks. Am J  Physiol M, Schmechel D, Roses AD. Isoform-specific interac-

Regul Integr Comp Physiol. 2013;305(5):R478–89. tions of apolipoprotein E with microtubule-associated
66. Yi JH, Brown C, Whitehead G, Piers T, Lee YS, Perez protein tau: implications for Alzheimer disease. Proc
CM, Regan P, Whitcomb DJ, Cho K. Glucocorticoids Natl Acad Sci U S A. 1994;91(23):11183–6.
activate a synapse weakening pathway culminating in 73. Manczak M, Reddy PH.  Abnormal interaction of

tau phosphorylation in the hippocampus. Pharmacol oligomeric amyloid-β with phosphorylated tau: impli-
Res. 2017;121:42–51. cations to synaptic dysfunction and neuronal damage.
67. Zhu H, Zhang W, Zhao Y, Shu X, Wang W, Wang D, J Alzheimers Dis. 2013;36(2):285–95.
Yang Y, He Z, Wang X, Ying Y. GSK3β-mediated tau 74. Sottejeau Y, Bretteville A, Cantrelle FX, Malmanche
hyperphosphorylation triggers diabetic retinal neuro- N, Demiaute F, Mendes T, Delay C, Alves Dos Alves
degeneration by disrupting synaptic and mitochon- H, Flaig A, Davies P, Dourlen P, Dermaut B, Laporte
drial functions. Mol Neurodegener. 2018;13(1):62. J, Amouyel P, Lippens G, Chapuis J, Landrieu I,
68. Lopes S, Vaz-Silva J, Pinto V, Dalla C, Kokras N, Lambert JC.  Tau phosphorylation regulates the
Bedenk B, Mack N, Czisch M, Almeida OF, Sousa interaction between BIN1’s SH3 domain and Tau’s
N, Sotiropoulos I. Tau protein is essential for stress-­ proline-rich domain. Acta Neuropathol Commun.
induced brain pathology. Proc Natl Acad Sci U S A. 2015;3:58.
2016;113(26):E3755–63. 75. Jack CR Jr, et  al. Tracking pathophysiological pro-
69. Noble W, et  al. Inhibition of glycogen synthase
cesses in Alzheimer’s disease: an updated hypotheti-
kinase-3 by lithium correlates with reduced tauopathy cal model of dynamic biomarkers. Lancet Neurol.
and degeneration in vivo. Proc Natl Acad Sci U S A. 2013;12:207–16.
2005;102:6990–5. 76. Busche MA, Wegmann S, Dujardin S, Commins C,
70. Forlenza OV, et  al. Disease-modifying properties of Schiantarelli J, Klcikstein N, Kamath TV, Carlson
long-term lithium treatment for amnestic mild cog- GA, Nelken I, Hyman BT. Tau impairs neural circuits,
nitive impairment: randomised controlled trial. Br dominating amyloid-β effects, n Alzheimer models
J Psychiatry. 2011;198:351–6. in vivo. Nat Neurosci. 2018;22:57–64.
71. Shankar GM, et al. Amyloid-beta protein dimers iso- 77. Nelson PT, et  al. Correlation of Alzheimer disease
lated directly from Alzheimer’s brains impair synaptic neuropathologic changes with cognitive status: a
plasticity and memory. Nat Med. 2008;14:837–42. review of the literature. J  Neuropathol Exp Neurol.
72. Strittmatter WJ, Saunders AM, Goedert M, Weisgraber 2012;71:362–81.
KH, Dong LM, Jakes R, Huang DY, Pericak-Vance
 Tau Secretion
11
Zhi Ruan and Tsuneya Ikezu

Introduction transfer from a “donor cell” to a “recipient cell”

and recruitment of endogenous Tau proteins in the
Pathogenic aggregation of microtubule-­associated latter to generate new abnormal Tau seeds [36, 43,
protein Tau, which also been known as intracellu- 122], which is similar to the mechanism of prion
lar neurofibrillary tangles (NFTs), is a hallmark of protein propagation [14, 70]. Based on these find-
multiple neurodegenerative disorders including ings, the way of Tau secretion and then uptake by
AD and FTDs. In AD, the severity of dementia is recipient neurons has become important mecha-
most closely linked with the distribution of Tau nism for understanding the spreading of Tau
aggregates in a hierarchical pattern [5]. The earli- pathology. For the prion spread, it has been shown
est stages of the disease show accumulation of to not only spread along neuroanatomical path-
abnormal Tau in the layer II of entorhinal cortex ways but also occur irrespective of the anatomical
(EC II) whereas later stages show accumulation in route or physical location of the prion inoculation
the hippocampus followed by neocortical areas. site [23]. During the past decades, many pathways
Based on this stereotypical sequential appearance, were proposed for Tau secretion, but it is unclear
tau pathology is diagnostically classified into six how efficient these are for the secretion of aggre-
Braak stages [11]. The EC II is monosynaptically gated Tau seeds. For example, though 90% of
connected to other hippocampal subregions, unaggregated Tau is secreted in a free form with
mainly dentate granular cells through perforant only a minority being secreted from extracellular
path, and it is trans-­synaptically connected with vesicles (EVs), the latter are more efficient for the
affected regions in the temporal and parietal lobes transfer of Tau aggregates. This highlights the
[114, 121]. One of the most interesting questions importance of understanding which mechanism(s)
in the field is whether pathology of the EC initiates of trans-cellular Tau spread plays a role in a physi-
anatomical progression of the disease, or whether ological condition and which may become more
pathology dysfunction in hippocampal areas prominent in pathological conditions. On one
develops independently, and is unrelated to events hand, there are a number of interesting observa-
occurring in the EC.  So far, in vitro and in vivo tions that support the trans-synaptic spread hypoth-
studies showed that the abnormal Tau seeds could esis for the spreading of Tau pathology in AD [27,
71]. On the other hand, the presence of Tau in the
Z. Ruan (*) · T. Ikezu (*) interstitial fluid and cerebrospinal fluid (CSF) of
Department of Pharmacology and Experimental Tau transgenic mice and human before neurode-
Therapeutics, Boston University School of Medicine, generation indicates that extracellular Tau can be
Boston, MA, USA released by an active process of secretion in vivo
e-mail: zhiruan@bu.edu; tikezu@bu.edu

© Springer Nature Singapore Pte Ltd. 2019 123

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_11
124 Z. Ruan and T. Ikezu

[9, 123]. In vitro, recombinant Tau was shown to Tau were detected in neurons downstream in the
be secreted by several non-­neuronal and neuronal synaptic circuit such as the dentate gyrus, CA fields
cell lines [78]. In those studies, Tau secretion into of the hippocampus, and cingulate cortex [27].
the extracellular space could also occurs through Because no expression of human Tau mRNA was
exo-synaptic spread way. detected in those regions, the accumulated human
Tau in those areas should come from the EC, sug-
gesting that Tau pathology is capable of traversing
Trans-Synaptic Spread Way neural networks in the brains. In addition, abnor-
mally phosphorylated, filamentous Tau derived
Considering Tau proteins is present in presynaptic from the brains of human P301S Tau transgenic
and postsynaptic terminals especially abnormal mice was sufficient to induce the formation of sil-
Tau enriched in synaptic sites [105], it suggested ver-positive Tau inclusions in ALZ17 mice that
that Tau could spread through the brain via a trans- overexpress wildtype human Tau, but do not
synaptic pathway. Such spread pattern of Tau develop Tau inclusions [20], and Tau inclusions as
means a trans-neuronal propagation of Tau aggre- early as 2  weeks after unilaterally infused and
gates from one brain region to the next using the showed contralateral hippocampal spread after
synaptic circuits, which has been first observed by 1 month [4]. More recently, intracerebral inocula-
Braak and Braak, whereby Tau pathology mani- tion of Tau fibrils purified from AD brains could
fests largely in a consistent spatiotemporal pattern result in the formation of abundant Tau inclusions
in human AD brains [11]. In vitro studies have also in anatomically connected brain regions in non-
documented that Tau aggregates have the ability of transgenic mice [44]. Additionally, these findings
to move along axons and dendrites in neuronal cul- suggested that Tau spread through the whole brain
tures, both antero- and retrograde, as well as to was dependent on synaptic connectivity. This trans-­
leave one neuron and be taken up by another [16, synaptic spreading of Tau pathology most likely
106, 122]. Moreover, the ability of Tau aggregates results from the secretion of Tau by presynaptic
has been demonstrated to behave similar to prions neurons and its uptake by postsynaptic ones.
in that they possess the ability to recruit and ulti- Since new Tau positron emission tomography
mately template the naïve monomeric version of (PET) imaging has developed, additional studies
the protein [16]. Based on these in  vitro results, have been performed to look at the accumulation
in vivo models have been generated to better study of Tau pathology in association with brain atro-
such spread mode of Tau propagation along synap- phy using in live patients. The formation of Tau
tically connected networks in the brain. Aggregates pathology correlates with neuronal dysfunction
of Tau injected into mouse brains appear to move and ultimately loss in the brain [50, 52, 54, 83,
from the injection site to synaptically connected 99, 101]. However, one caveat of this propagation
regions [16, 19, 53, 60, 79, 98]. Further, studies pattern is that EC II project to dentate gyrus,
in vivo and in vitro have demonstrated that patho- where tauopathies are frequently observed in
logical Tau transfers between cells and in between FTDs but not in AD, which show more frequent
anatomically connected brain regions, leading to tauopathy in the CA1/2 region in the early stage.
the spread of Tau in neurons and oligodendrocytes This discrepancy is not fully understood with the
[20, 21]. Studies also showed that inoculation of current trans-synaptic mechanism.
Tau induces time-dependent spreading of Tau
pathology from the inoculation site to synaptically
connected brain regions in human Tau transgenic Exo-Synaptic Spread Pathway
mice or even in wild type mice [71]. In two lines of
Tau transgenic mice, the expression of human Tau Extracellular Tau Secretory Pathway
under the control of the neuropsin promoter was
restricted to the EC, yet Tau aggregates composed The vast majority of conventional secretory pro-
of transgenic human Tau and endogenous mouse teins contain a secretory signal sequence that
11  Tau Secretion 125

directs their sorting to the endoplasmic reticulum  esicular Mediated Secretory

V
(ER) from where proteins are transported through Pathway
the lumen of the ER–Golgi-dependent secretory
pathway and then to the cell surface, either the The protein degradation of most cellular proteins
extracellular space or the plasma membrane [67]. in eukaryotic cell takes place by two major path-
However, emerging evidences have documented ways: through the proteasomal system or
that there is an “unconventional secretion path- autophagolysosomal system. In case of Tau pro-
ways” for the proteins including Tau without tein, evidences have showed that Tau protein
such kind of sequence to be secreted into extra- degrades through both systems [119, 124].
cellular space [81]. The presence of extracellular However, the protein secretion through EVs
Tau in the AD brain was revealed by its accumu- recently has been proposed to be an alternative
lation in CSF during the progression of the dis- pathway to compensate protein clearance in order
ease [47]. In three lines of Tau transgenic mice, to regulate a diverse range of biological processes
they all showed significant age-dependent including facilitating protein discharge, cell dif-
increases of Tau in the CSF [9]. Additionally, the ferentiation or intercellular communication but
presence of extracellular Tau in the absence of also in pathological process [46, 101]. At least
neurodegeneration was then detected by in vivo two types of EVs, based on their intracellular ori-
microdialysis, indicating that Tau was secreted gin, have been showed to be involved in the Tau
by neurons in vivo [122]. All above interesting secretion: microvesicles (MVs), which are origi-
findings suggested that the Tau secreted into nated by outward budding from the plasma mem-
extracellular space could play a role in the spread brane [22] and more frequently exosomes,
of Tau pathology that cause the neurodegenera- originated by inward budding of the multivesicu-
tive process in the AD brain. lar bodies (late endocytic compartments). MVs
As the exist of extracellular Tau was con- and exosomes share many protein markers, mak-
firmed, their role in the process of tauopathies ing it difficult to distinguish between them in the
has also been documented by many groups. In extracellular space or after vesicles have been
vitro, adding Tau to the culture medium could purified [63]. In addition, both types of EVs use
cause a robust increase of intracellular calcium the Endosomal Sorting Complex Required for
and cell death of human neuronal SH-SY5Y cells Transport (ESCRT) machinery in their formation
[31]. More recently, Tau oligomers isolated from [1]. One of ESCRT-0 proteins hepatocyte growth
the AD brain decreased long-term potentiation factor-regulated tyrosine kinase substrate (HRS)
(LTP) in hippocampal slices [66]. These studies and signal transducing adapter molecule 1
indicated that extracellular Tau can induce neuro- (STAM) are exclusively involved in the initial
nal dysfunction. In addition, several recent stud- phase of exosome biogenesis [24] (reviewed in
ies has reported that there was an propagation of Colombo et  al. [24]). Exosomes may also be
Tau pathology after injection of exogenous tau formed in an ESCRT-independent mechanism,
into the mouse brain [20]. mainly ceramide-initiated biogenesis [38, 55,
During the past decade, accumulating evi- 103, 110] and tetraspanin-mediated endosomal
dence has also indicated that these extracellular sorting [37, 114, 115]. Due to their different bio-
secreted pathological Tau could be taken up by logical origins, exosomes and MVs may play
neighboring neurons and glia [25, 80, 92], divergent roles in physiological and pathological
largely depending on the generation of EVs conditions; it is important to differentiate the role
such as exosomes [24, 99]. Meanwhile, other in Tau secretion between these two classes of
groups also demonstrated that the transfer of EVs until this point is fully understood.
such Tau between cells could also occur through Both MVs and exosomes have been reported
intercellular tubular connections [2, 3, 95]. We for most cell types if not all types of cells in the
will review these alternative pathways in the body, including the central nervous system (CNS)
following sections. [85], such as neurons [65], astrocytes [107], and
126 Z. Ruan and T. Ikezu

microglia [91], and have been proposed to play a nately secreted in MVs, which are plasma mem-
role in the development of neurodegenerative dis- brane-originating vesicles [33]. It suggested that
eases, such as prion diseases [117] and AD [93]. such specific vesicles, directly emerging from the
As larger EVs that directly shed from cells by plasma membrane, enabled cytosolic Tau to be
plasma membrane budding, Tau could be secreted shuttled to the extracellular medium.
in EVs [33]. In addition, EVs now have served as
important mediators of intercellular communica- Exosomes
tion due to their capacity to exchange contents Another major type of EVs are exosomes, which
between cell and act as biological signaling vehi- are a specific subset of EVs approximately
cles of pathological developments [116]. 50–150 nm in size and are released upon depolar-
ization of plasma membrane [30]. Beyond size,
Microvesicles exosomes are distinguished from other EVs by
As one of major types of EVs involved in inter- their enrichment of the tetraspanin proteins, such
cellular communication, MVs are generated by as CD63, CD81 and CD9 [64]. Exosomes are
outward, budding of the plasma membrane [116]. often characterized by their high concentration of
These types of EVs come in many sizes, although lipid raft components, such as ceramide and
not always, but are generally 150–1000  nm in sphingomyelin [28, 94]. Microtubule-unbound
size [24]. First, studies have documented that Tau was predisposed to be associated with intra-
MVs could play multiple roles in altering the cellular vesicles. The vesicular association of Tau
extracellular environment, intercellular signal- was especially prevalent in cellular compart-
ing, and facilitating cell invasion through cell-­ ments with disorganized microtubules and phos-
independent matrix proteolysis [17, 116]. phorylated Tau [68]. Furthermore, vesicular tau
Furthermore, MVs are also able to modify the has been reported to be colocalized with the lipid
extracellular milieu, proximal, and distal recipi- raft-associated tyrosine kinase fyn which has
ent cells through their ability to transfer bioactive been identified in exosomes [68], allowing for the
molecules, including proteins, DNAs [108] and potential of secretion of exosome containing Tau.
RNAs [112]. In addition, Tau was found in exosomes isolated
The association of Tau with the plasma mem- from the CSF [96] and blood of AD patients [35].
brane has been known for many years [12, 61, 68, More recently, Tau protein released by cultured
109], indicating that Tau could direct the vesicle primary neurons or by mouse neuronal N2a cells
shedding. As is a soluble cytoplasmic protein, Tau overexpressing Tau through the way of exosome
is not directed to the classical secretory endoplas- secretion [118].
mic reticulum-Golgi secretory pathway in physi- Injection of exosome-enriched fraction iso-
ological conditions. In addition, considering that lated from human P301L Tau transgenic rTg4510
MVs are released through cell membrane activa- mouse brains caused increased Tau phosphoryla-
tion by mediators such as intracellular levels of tion and the oligomerization of endogenous Tau
calcium, inflammatory molecules or oxidative [7]. The initiation of endogenous Tau misfolding
stress, which are involved in the pathology of and aggregation by Tau-containing exosomes was
Tauopathies [8, 32, 86], MVs could be good can- in a threshold-dependent manner [89]. Our recent
didates as the mechanism of secreting Tau protein work has also shown that microglia-­derived exo-
under pathological conditions. Recently, studies somes can efficiently transmit Tau to neurons in
by using in vitro and in vivo in a rat model of spo- mouse brain. Furthermore, injection of microglia-
radic Tauopathies, have demonstrated that Tau derived exosomes in the outer molecular layer of
protein is present in MVs in the extracellular fluid. dentate gyrus results in the spread of tau accumu-
Under basal conditions (rat neuronal primary cul- lation to the dentate granular cells. Inhibition of
tures), major free forms of Tau are actively cremaide-dependent exosome biogenesis by
secreted through secretory pathways, predomi- blocking of neutral sphingomyelinase-2 activity
11  Tau Secretion 127

significantly reduced Tau propagation in this tion and endosomal trafficking are overlapped,
model, suggesting that exosomes are indeed a both involving the components of exocyst com-
pathogenic agent in the spread of pathogenic Tau plex which regulates vesicular transport from
[6]. Taken together, these studies indicate that Golgi apparatus to the plasma membrane.
Tau-­containing exosomes may play an important M-sec, a part of the exocyst complex, interacts
role in the propagation of Tau pathology through with Ras-­ related protein-A (RalA, small
the brain, which accumulates not only in cell GTPase) and is required for TNT formation.
soma but also synapses. Microglia phagocytose M-Sec in cooperation with RalA and the exocyst
Tau-­containing inactive synapses from neurons complex serves as key factor for the formation
and may eventually secrete them in exosomes of functional TNTs and therefore M-Sec is con-
when their intracellular clearance is insufficient in sidered TNT marker [79]. Other studies demon-
an attempt to remain healthy. These pathological strate that formation of some TNTs might be
protein-laden exosomes could in turn be uptaken actinomyosin-dependent [15, 45]. Perhaps not
by neurons, which are especially vulnerable, surprising, motor proteins are required for the
resulting in their final death. Neurons can also generation of some forms of TNTs. For exam-
transmit Tau to each other via exosomes, although ple, myosin10 (Myo10) is required for TNT for-
this has only been shown in vitro [42, 88, 118]. mation from filopodia, where the overexpression
Exosomes created by these processes also spread of Myo10 results in increased TNT formation
pathological forms of Tau throughout the brain, and vesicle transfer between cells [40]. Elevation
and may eventually secreted in CSF or blood, of Eps8 (an actin regulatory protein) inhibits the
where their detection can serve as a potential bio- extension of filopodia in neurons and increases
marker. Moreover, exosome-associated Tau is TNT formation as well as intercellular vesicle
also present in human CSF samples and is phos- transfer [29]. Altogether, these observations
phorylated at Thr-181 (AT270), an established indicate that cells may use motor proteins as
phosphor-Tau biomarker for AD, in CSF samples component of both TNTs and EVs for shipping
from patients with mild (Braak stage 3) AD [96]. their cargo.
Overall, above results suggest that exosomes Emerging data showed that the microtubule-­
in the case of Tau transmission under pathologi- associated protein Tau is a specific constitutive
cal condition may be functioning more as a marker of TNTs. This is important because Tau
waste-disposal method rather than as a means of appears beside filamentous actin and myosin 10
cell-cell communication in an attempt by cells to as a specific marker of these fine protrusions of
protect themselves. membranes and cytosol that are difficult to visu-
alize. Furthermore, extracellular Tau species
(monomers, oligomers and fibrils) activate the
Tunneling Nanotubes (TNTs) formation of TNTs that subsequently facilitate
fibrillar Tau transfer from neuron to neuron and
TNTs are actin-based transient cytoplasmic such extracellular Tau may likely act as a signal
extensions which are stretched between cells in for TNT formation as they are found in patho-
the form of open ended nanotubular channels logical conditions [106]. In addition, exogenous
(50–200 nm), not always linked to the substrate, Tau species could also increase the number of
and forming bridges that connect remote cells, TNTs established between primary neurons,
discovered by Rustom and colleagues [26, 95]. thereby facilitating the intercellular transfer of
Like EVs, TNTs also represent subtypes and Tau fibrils [106]. In conclusion, Tau may contrib-
heterogeneous morphological structures [10]. ute to the formation and function of the highly
However, biosynthesis of TNTs differs from dynamic TNTs that may be involved in the prion-­
EVs and is attributed to f-actin polymerization like propagation of Tau assemblies.
[83]. The regulatory pathways of TNT forma-
128 Z. Ruan and T. Ikezu

Species of Tau That Are Secreted fibrils  were taken up and transported in axons
toward the axonal terminals [121].
Accumulating data shown that Tau is physiologi-
cally present in the extracellular fluid in the form
of full-length [59] or truncated forms [13, 57] not The Regulation on Tau Secretion
only in the culture media of cells over-expressing
human Tau [18], primary neurons [18], and iPSC The amount of soluble monomeric Tau released
derived neurons [13], but also in the brain intersti- from cells is altered by the various types of exter-
tial fluid and CSF of mice [122] and human [71, nal stimuli. For example, starvation or lysosomal
72]. However, there are many controversies dysfunction increases Tau release [73]. In addi-
around which specific Tau species are secreted to tion, stimulating neuronal activity in either cul-
the extracellular space under pathological condi- tured neurons or in vivo also enhances Tau release
tions. Some studies have shown that Tau propaga- [34, 90, 123]. This process was calcium-­dependent
tion could involve species ranging from small and modulated by phosphorylation. Given that
soluble monomers to large insoluble fibrils in vivo extracellular Tau influences neuronal activity
or in vitro [62, 66, 105]. Comparative analysis of [13], this observation suggests that the activity-
CSF from AD and healthy subjects showed a dependent release of Tau participates in a positive
clear increase of amino-terminal (N-terminal) feedback loop on neuronal activity. More recently,
Tau fragments in AD, with no evidence of full- Golgi dynamics were linked to a modulation of
length or carboxyl-terminal (C-terminal) Tau Tau secretion by both primary cortical neurons
fragments [72]. On the other hand, cell lines such and HeLa cells [74]. Differences in Tau species or
as M1C and Hela cells overexpressing human Tau isoforms also impact its  secretion: phosphoryla-
release a cleaved form of the protein in its tion, truncation and mutations. These finding of
C-terminal end [18, 58, 90]. The phosphorylation variable Tau secretion suggest the presence of
status of Tau has been also examined. Depending active cellular mechanisms regulating its
on the cell type, secretion of overexpressed secretion.
human Tau by non-neuronal cell lines is either
phosphorylated or importantly dephosphorylated
at several sites [18, 87]. It was reported that intra-  ole of Phosphorylation on Tau
R
cellular Tau fibrils could be directly released into Secretion
extracellular space in culture cells and then be
taken up by the co-cultured cells in the medium In AD, there are more than 40 phosphorylated
via cell-cell transfer in exosomes or tunneling sites of Tau compared to 9 sites in normal patients
nanotubes [62]. In vivo studies also shown that [49], which means that Tau phosphorylation
intracerebral inoculation of Tau fibrils purified plays a key role in the Tau pathology. In addition,
from AD brains could result in the formation of studies showed that the amounts of phosphory-
abundant Tau inclusions in non-­transgenic mice lated Tau of T181 and T231 in the CSF of AD
[44]. Above findings suggest the possibility that patients were significantly higher than that of
Tau fibrils could act as a seed to propagate pathol- normal patients [55], although phosphorylation
ogy between neurons in vivo. However, some other at other sites such as S199, S202, T205 S396 and
groups have identified non-­fibrillary Tau forms as S404 remains controversial [71]. It is reasonable
necessary for propagation [66, 111]. Tau propaga- to propose that the phosphorylation level might
tion has occurred after injected human Tau oligo- take an effect on the Tau secretion. Recently, an
mers into the hippocampus of wild type mice in vitro study has demonstrated that the mimick-
[66], as well as Tau aggregation induced by Tau ing of phosphorylation at 12 sites known to be
oligomers in human neurons [111]. Additionally, phosphorylated in AD enhanced Tau secretion by
an in vitro study showed that small misfolded Tau Hela cells [83]. Moreover, the Tau secreted by
aggregates and  short fibrils but not the long exosomes was shown to be phosphorylated at
11  Tau Secretion 129

several sites found in AD [47] while Tau in the Tau and a 20-22 kDa fragment of Tau compared to
extracellular space would be dephosphorylated in normal samples, and the majority of the tau lacked
AD brain by tissue nonspecific alkaline phospha- the C-terminal. Depolarization induced by potas-
tases [31]. Another study also confirmed that sium chloride could significantly enhanced the
M1C cells secrete selectively phosphorylated Tau Tau secretion from synaptosomes in a concentra-
which is also present in human CSF samples and tion-dependent manner [98]. In summary, it seems
is phosphorylated at Thr-181 [92]. More recently, that both full-length and proteolytic fragments of
expressing the different tau variants CHO cell Tau can be secreted by different mechanisms.
lines preferentially secreted more phosphoryla-
tion tau, which provokes microtubule detachment
and increases the availability of free protein Role of Mutations on Tau Secretion
inside cells. However, Tau secretion in primary
cortical neurons was controversial for that some Mutations in the tau gene (MAPT) were demon-
studies showed the release of unphosphorylated strated for the first time to be associated with
Tau in control conditions [86], other groups FTDP-17T (frontotemporal dementia with par-
reported that both phosphorylated and unphos- kinsonism linked to chromosome 17 and spe-
phorylated could be secreted [83]. Further study cifically characterized by tau pathology), and a
is necessary to elucidate the role of Tau phos- mutation in the MAPT alone could cause neuro-
phorylation on Tau secretion. degenerative disease and strongly suggested
that aberrant Tau plays a pathogenic role in
other tauopathies, including AD [50]. To date, at
Role of Truncation on Tau Secretion least 107 mutations associated with FTDP-17T
or related disorders have been identified, and
Full-length and various fragments of Tau cleavage most of the mutations in the MAPT coding
have been shown to be secreted in a variety of dif- regions are found in in the portion encoding the
ferent cells and animal models [76]. Tau fragments C-terminal region whereas intronic mutations
lacking the proline-region are either not secreted are located near the alternative splicing site of
or are secreted in a distinct manner to the full- intron 10, leading to increase the 4R/3R ratio.
length molecule by COS-7 kidney fibroblast cells The C-terminal missense mutations reduce the
[84]. On the other hand, studies have observed the ability of Tau to interact with microtubules. This
fragments of tau cleaved at the C-terminal present- partial loss of Tau function may be responsible
ing in the CSF collected from human AD subjects for inducing the abnormal aggregation of the
and from tau transgenic mice [62, 92]. In fact, protein. Several FTDP-17 mutations also could
C-terminal cleavage of Tau at D421 increased Tau promote the assembly of Tau protein into fila-
secretion [83]. Another study using three distinct ments [77]. In vitro studies demonstrated that
neuronal cultures including conditioned medium Tau mutations not only influenced the rate of
of N2a cells, induced pluripotent stem cell-derived Tau secretion, but also affect the form of Tau
human cortical neurons, and primary rat cortical secretion from the cell, which consisted of the N
neurons observed that the majority of tau released terminal and microtubule binding repeat length.
from healthy neurons was C-terminally truncated Moreover, such mutations also produce signifi-
whereas the microtubule-binding region-contain- cantly less extracellular total Tau without alter-
ing tau presenting outside of cells was solely due ing intracellular total Tau levels [57]. In addition,
to cell death [56]. In addition, an in vitro study CSF Tau levels in symptomatic FTD patients
showed that Tau protein fragments secretion to the expressing a MAPT mutation are slightly ele-
extracellular space required the presence of the vated as compared to controls but are signifi-
Tau N-terminal domain in a tauopathy model via cantly lower than in AD patients [57]. These
two distinct mechanisms [62]. The synaptosomes results indicate Tau mutations could affect the
from the AD brains had higher levels of dimerised secretion of certain Tau isoforms from the cell.
130 Z. Ruan and T. Ikezu

Concluding Remarks 7. Baker S, Polanco JC, Gotz J. Extracellular vesicles

containing P301L mutant Tau accelerate pathologi-
cal Tau phosphorylation and oligomer formation but
Tau can be secreted as is or via EV-dependent do not seed mature neurofibrillary tangles in ALZ17
pathways, including MVs and exosomes. The mice. J Alzheimers Dis. 2016;54:1207–17.
secreted Tau can impair synaptic function and 8. Baron M, Boulanger CM, Staels B, Tailleux A. Cell-­
derived microparticles in atherosclerosis: biomark-
cause memory deficits in animal models. This ers and targets for pharmacological modulation?
highlights the possibility that the accumulation J Cell Mol Med. 2012;16:1365–76.
of Tau in extracellular space could exert detri- 9. Barten DM, Cadelina GW, Hoque N, DeCarr LB,
mental effects on neuron function in the tauopa- Guss VL, Yang L, Sankaranarayanan S, Wes PD,
Flynn ME, Meredith JE, et al. Tau transgenic mice
thy brain. It will be interesting to investigate in as models for cerebrospinal fluid tau biomarkers.
which way of these extracellular Tau secreted J Alzheimers Dis. 2011;24(Suppl 2):127–41.
actively, and which pathway will be dominant 10. Benard M, Schapman D, Lebon A, Monterroso B,
especially under the pathological conditions. Bellenger M, Le Foll F, Pasquier J, Vaudry H, Vaudry
D, Galas L.  Structural and functional analysis of
Most importantly, Tau can be not secreted and tunneling nanotubes (TnTs) using gCW STED and
uptaken by adjacent neurons calls for the devel- gconfocal approaches. Biol Cell. 2015;107:419–25.
opment of novel strategies to prevent the propa- 11. Braak H, Braak E.  Neuropathological stageing
gation of Tau pathology via extracellular of Alzheimer-related changes. Acta Neuropathol.
1991;82:239–59.
pathways. Future studies are needed to elucidate 12. Brandt R, Leger J, Lee G.  Interaction of tau with
the respective roles of different Tau secretion the neural plasma membrane mediated by tau’s
pathways for the development of tauopathies. amino-terminal projection domain. J  Cell Biol.
1995;131:1327–40.
13. Bright J, Hussain S, Dang V, Wright S, Cooper B,
Byun T, Ramos C, Singh A, Parry G, Stagliano N,
References et  al. Human secreted tau increases amyloid-beta
production. Neurobiol Aging. 2015;36:693–709.
1. Abels ER, Breakefield XO.  Introduction to extra- 14. Brundin P, Melki R, Kopito R. Prion-like transmis-
cellular vesicles: biogenesis, RNA cargo selection, sion of protein aggregates in neurodegenerative dis-
content, release, and uptake. Cell Mol Neurobiol. eases. Nat Rev Mol Cell Biol. 2010;11:301–7.
2016;36:301–12. 15. Bukoreshtliev NV, Wang X, Hodneland E, Gurke S,
2. Abounit S, Wu JW, Duff K, Victoria GS, Zurzolo Barroso JFV, Gerdes HH. Selective block of tunnel-
C. Tunneling nanotubes: a possible highway in the ing nanotube (TNT) formation inhibits intercellular
spreading of tau and other prion-like proteins in neu- organelle transfer between PC12 cells. FEBS Lett.
rodegenerative diseases. Prion. 2016;10:344–51. 2009;583:1481–8.
3. Agnati LF, Fuxe K. Extracellular-vesicle type of vol- 16. Calafate S, Buist A, Miskiewicz K, Vijayan V,
ume transmission and tunnelling-nanotube type of Daneels G, de Strooper B, de Wit J, Verstreken
wiring transmission add a new dimension to brain P, Moechars D.  Synaptic contacts enhance cell-­
neuro-glial networks. Philos Trans R Soc Lond Ser to-­cell Tau pathology propagation. Cell Rep.
B Biol Sci. 2014;369:20130505. 2015;11:1176–83.
4. Ahmed Z, Cooper J, Murray TK, Garn K, 17. Candelario KM, Steindler DA.  The role of extra-
McNaughton E, Clarke H, Parhizkar S, Ward MA, cellular vesicles in the progression of neurode-
Cavallini A, Jackson S, et al. A novel in vivo model generative disease and cancer. Trends Mol Med.
of tau propagation with rapid and progressive neu- 2014;20:368–74.
rofibrillary tangle pathology: the pattern of spread 18. Chai X, Dage JL, Citron M.  Constitutive secretion
is determined by connectivity, not proximity. Acta of tau protein by an unconventional mechanism.
Neuropathol. 2014;127:667–83. Neurobiol Dis. 2012;48:356–66.
5. Arriagada PV, Growdon JH, Hedleywhyte ET, 19. Clavaguera F, Akatsu H, Fraser G, Crowther RA,
Hyman BT.  Neurofibrillary tangles but not senile Frank S, Hench J, Probst A, Winkler DT, Reichwald
plaques parallel duration and severity of Alzheimers-­ J, Staufenbiel M, et  al. Brain homogenates from
disease. Neurology. 1992;42:631–9. human tauopathies induce tau inclusions in mouse
6. Asai H, Ikezu S, Tsunoda S, Medalla M, Luebke J, brain. Proc Natl Acad Sci U S A. 2013;110:9535–40.
Haydar T, Wolozin B, Butovsky O, Kugler S, Ikezu 20. Clavaguera F, Bolmont T, Crowther RA, Abramowski
T.  Depletion of microglia and inhibition of exo- D, Frank S, Probst A, Fraser G, Stalder AK, Beibel
some synthesis halt tau propagation. Nat Neurosci. M, Staufenbiel M, et  al. Transmission and spread-
2015;18:1584–93. ing of tauopathy in transgenic mouse brain. Nat Cell
Biol. 2009;11:909–13.
11  Tau Secretion 131

21. Clavaguera F, Grueninger F, Tolnay M. Intercellular pathogenic proteins in neurally derived blood exo-
transfer of tau aggregates and spreading of tau somes: a case-control study. Alzheimers Dement.
pathology: implications for therapeutic strategies. 2015;11:600–7.
Neuropharmacology. 2014;76 Pt A:9–15. 36. Frost B, Jacks RL, Diamond MI. Propagation of tau
22. Cocucci E, Racchetti G, Meldolesi J.  Shedding misfolding from the outside to the inside of a cell.
microvesicles: artefacts no more. Trends Cell Biol. J Biol Chem. 2009;284:12845–52.
2009;19:43–51. 37. Gauthier SA, Perez-Gonzalez R, Sharma A, Huang
23. Collinge J.  Mammalian prions and their wider FK, Alldred MJ, Pawlik M, Kaur G, Ginsberg SD,
relevance in neurodegenerative diseases. Nature. Neubert TA, Levy E.  Enhanced exosome secretion
2016;539:217–26. in Down syndrome brain – a protective mechanism
24. Colombo M, Raposo G, Thery C. Biogenesis, secre- to alleviate neuronal endosomal abnormalities. Acta
tion, and intercellular interactions of exosomes and Neuropathol Commun. 2017;5:65.
other extracellular vesicles. Annu Rev Cell Dev Biol. 38. Ghossoub R, Lembo F, Rubio A, Gaillard CB,
2014;30:255–89. Bouchet J, Vitale N, Slavik J, Machala M,
25. Cruz-Garcia D, Malhotra V, Curwin Zimmermann P.  Syntenin-ALIX exosome bio-
AJ.  Unconventional protein secretion triggered genesis and budding into multivesicular bodies
by nutrient starvation. Semin Cell Dev Biol. are controlled by ARF6 and PLD2. Nat Commun.
2018;83:22–8. 2014;5:3477.
26. Davis DM, Sowinski S.  Membrane nanotubes: 39. Goedert M, Spillantini MG.  Pathogenesis of the
dynamic long-distance connections between animal tauopathies. J Mol Neurosci. 2011;45:425–31.
cells. Nat Rev Mol Cell Biol. 2008;9:431–6. 40. Gousset K, Marzo L, Commere PH, Zurzolo
27. de Calignon A, Polydoro M, Suarez-Calvet M, C.  Myo10 is a key regulator of TNT formation in
William C, Adamowicz DH, Kopeikina KJ, neuronal cells. J Cell Sci. 2013;126:4424–35.
Pitstick R, Sahara N, Ashe KH, Carlson GA, et al. 41. Green AJE, Harvey RJ, Thompson EJ, Rossor
Propagation of tau pathology in a model of early MN.  Increased tau in the cerebrospinal fluid
Alzheimer’s disease. Neuron. 2012;73:685–97. of patients with frontotemporal dementia and
28. de Gassart A, Geminard C, Fevrier B, Raposo G, Alzheimer’s disease. Neurosci Lett. 1999;259:133–5.
Vidal M. Lipid raft-associated protein sorting in exo- 42. Guix FX, Corbett GT, Cha DJ, Mustapic M, Liu
somes. Blood. 2003;102:4336–44. W, Mengel D, Chen Z, Aikawa E, Young-Pearse
29. Delage E, Cervantes DC, Penard E, Schmitt C, Syan T, Kapogiannis D, et  al. Detection of aggregation-­
S, Disanza A, Scita G, Zurzolo C. Differential iden- competent Tau in neuron-derived extracellular vesi-
tity of filopodia and tunneling nanotubes revealed cles. Int J Mol Sci. 2018;19:pii: E663.
by the opposite functions of actin regulatory com- 43. Guo JL, Lee VM.  Seeding of normal Tau by
plexes. Sci Rep. 2016;6:39632. pathological Tau conformers drives pathogen-
30. Deleo AM, Ikezu T.  Extracellular vesicle biol- esis of Alzheimer-like tangles. J  Biol Chem.
ogy in Alzheimer’s disease and related tauopathy. 2011;286:15317–31.
J Neuroimmune Pharmcol. 2018;13:292–308. 44. Guo JL, Narasimhan S, Changolkar L, He ZH,
31. Diaz-Hernandez M, Gomez-Ramos A, Rubio A, Stieber A, Zhang B, Gathagan RJ, Iba M, McBride
Gomez-Villafuertes R, Naranjo JR, Miras-Portugal JD, Trojanowski JQ, et  al. Unique pathological
MT, Avila J.  Tissue-nonspecific alkaline phospha- tau conformers from Alzheimer’s brains transmit
tase promotes the neurotoxicity effect of extracellu- tau pathology in nontransgenic mice. J  Exp Med.
lar Tau. J Biol Chem. 2010;285:32539–48. 2016;213:2635–54.
32. Doeuvre L, Plawinski L, Toti F, Angles-Cano 45. Gurke S, Barroso JFV, Hodneland E, Bukoreshtliev
E.  Cell-derived microparticles: a new challenge in NV, Schlicker O, Gerdes HH.  Tunneling nano-
neuroscience. J Neurochem. 2009;110:457–68. tube (TNT)-like structures facilitate a constitu-
33. Dujardin S, Begard S, Caillierez R, Lachaud C, tive, actomyosin-­dependent exchange of endocytic
Delattre L, Carrier S, Loyens A, Galas MC, Bousset organelles between normal rat kidney cells. Exp Cell
L, Melki R, et al. Ectosomes: a new mechanism for Res. 2008;314:3669–83.
non-exosomal secretion of tau protein. PLoS One. 46. Gyorgy B, Szabo TG, Pasztoi M, Pal Z, Misjak P,
2014;9:e100760. Aradi B, Laszlo V, Pallinger E, Pap E, Kittel A, et al.
34. Fa M, Puzzo D, Piacentini R, Staniszewski A, Zhang Membrane vesicles, current state-of-the-art: emerg-
H, Baltrons MA, Li Puma DD, Chatterjee I, Li J, ing role of extracellular vesicles. Cell Mol Life Sci.
Saeed F, et al. Extracellular Tau oligomers produce 2011;68:2667–88.
an immediate impairment of LTP and memory. Sci 47. Hampel H, Blennow K, Shaw LM, Hoessler YC,
Rep. 2016;6:19393. Zetterberg H, Trojanowski JQ.  Total and phos-
35. Fiandaca MS, Kapogiannis D, Mapstone M, Boxer phorylated tau protein as biological markers of
A, Eitan E, Schwartz JB, Abner EL, Petersen RC, Alzheimer’s disease. Exp Gerontol. 2010;45:30–40.
Federoff HJ, Miller BL, et  al. Identification of 48. Hampel H, Teipel SJ, Fuchsberger T, Andreasen N,
preclinical Alzheimer’s disease by a profile of Wiltfang J, Otto M, Shen Y, Dodel R, Du Y, Farlow
132 Z. Ruan and T. Ikezu

M, et al. Value of CSF beta-amyloid(1-42) and tau activity and strains from formaldehyde-fixed tissue.
as predictors of Alzheimer’s disease in patients Acta Neuropathol Commun. 2017;5:41.
with mild cognitive impairment. Mol Psychiatry. 61. Kempf M, Clement A, Faissner A, Lee G, Brandt
2004;9:705–10. R. Tau binds to the distal axon early in development
49. Hanger DP, Byers HL, Wray S, Leung KY, Saxton of polarity in a microtubule- and microfilament-­
MJ, Seereeram A, Reynolds CH, Ward MA, dependent manner. J Neurosci. 1996;16:5583–92.
Anderton BH. Novel phosphorylation sites in tau 62. Kfoury N, Holmes BB, Jiang H, Holtzman DM,
from Alzheimer brain support a role for casein Diamond MI.  Trans-cellular propagation of Tau
kinase 1 in disease pathogenesis. J Biol Chem. aggregation by fibrillar species. J  Biol Chem.
2007;282:23645–54. 2012;287:19440–51.
50. Hoenig MC, Bischof GN, Seemiller J, Hammes 63. Kim W, Lee S, Jung C, Ahmed A, Lee G, Hall
J, Kukolja J, Onur OA, Jessen F, Fliessbach GF.  Interneuronal transfer of human Tau between
K, Neumaier B, Fink GR, et  al. Networks of lamprey central neurons in situ. J  Alzheimers Dis.
tau distribution in Alzheimer’s disease. Brain. 2010;19:647–64.
2018;141:568–81. 64. Kowal J, Arras G, Colombo M, Jouve M, Morath
51. Hutton M, Lendon CL, Rizzu P, Baker M, Froelich JP, Primdal-Bengtson B, Dingli F, Loew D, Tkach
S, Houlden H, Pickering-Brown S, Chakraverty S, M, Thery C.  Proteomic comparison defines novel
Isaacs A, Grover A, et  al. Association of missense markers to characterize heterogeneous populations
and 5′-splice-site mutations in tau with the inherited of extracellular vesicle subtypes. Proc Natl Acad Sci
dementia FTDP-17. Nature. 1998;393:702–5. U S A. 2016;113:E968–77.
52. Iaccarino L, Tammewar G, Ayakta N, Baker SL, 65. Lachenal G, Pernet-Gallay K, Chivet M, Hemming
Bejanin A, Boxer AL, Gorno-Tempini ML, Janabi FJ, Belly A, Bodon G, Blot B, Haase G, Goldberg Y,
M, Kramer JH, Lazaris A, et  al. Local and distant Sadoul R. Release of exosomes from differentiated
relationships between amyloid, tau and neurodegen- neurons and its regulation by synaptic glutamatergic
eration in Alzheimer’s disease. Neuroimage Clin. activity. Mol Cell Neurosci. 2011;46:409–18.
2018;17:452–64. 66. Lasagna-Reeves CA, Castillo-Carranza DL,
53. Iba M, Guo JL, McBride JD, Zhang B, Trojanowski Sengupta U, Guerrero-Munoz MJ, Kiritoshi T,
JQ, Lee VM. Synthetic tau fibrils mediate transmis- Neugebauer V, Jackson GR, Kayed R.  Alzheimer
sion of neurofibrillary tangles in a transgenic mouse brain-derived tau oligomers propagate pathology
model of Alzheimer’s-like tauopathy. J  Neurosci. from endogenous tau. Sci Rep. 2012;2:700.
2013;33:1024–37. 67. Lee MCS, Miller EA, Goldberg J, Orci L, Schekman
54. Ishiki A, Okamura N, Furukawa K, Furumoto R.  Bi-directional protein transport between the ER
S, Harada R, Tomita N, Hiraoka K, Watanuki S, and Golgi. Annu Rev Cell Dev Biol. 2004;20:87–123.
Ishikawa Y, Tago T, et al. Longitudinal assessment of 68. Lee S, Kim W, Li Z, Hall GF.  Accumulation of
Tau pathology in patients with Alzheimer’s disease vesicle-associated human tau in distal dendrites
using [18F]THK-5117 positron emission tomogra- drives degeneration and tau secretion in an in situ
phy. PLoS One. 2015;10:e0140311. cellular tauopathy model. Int J  Alzheimers Dis.
55. Kajimoto T, Okada T, Miya S, Zhang L, Nakamura 2012;2012:172837.
S.  Ongoing activation of sphingosine 1-phosphate 69. Lee SJ, Desplats P, Sigurdson C, Tsigelny I, Masliah
receptors mediates maturation of exosomal multive- E.  Cell-to-cell transmission of non-prion protein
sicular endosomes. Nat Commun. 2013;4:2712. aggregates. Nat Rev Neurol. 2010;6:702–6.
56. Kandimalla RJL, Prabhakar S, Wani WY, Kaushal 70. Liu L, Drouet V, Wu JW, Witter MP, Small SA,
A, Gupta N, Sharma DR, Grover VK, Bhardwaj N, Clelland C, Duff K.  Trans-synaptic spread of tau
Jain K, Gill KD. CSF p-Tau levels in the prediction pathology in vivo. PLoS One. 2012;7:e31302.
of Alzheimer’s disease. Biol Open. 2013;2:1119–24. 71. Magnoni S, Esparza TJ, Conte V, Carbonara
57. Kanmert D, Cantlon A, Muratore CR, Jin M, M, Carrabba G, Holtzman DM, Zipfel GJ,
O’Malley TT, Lee G, Young-Pearse TL, Selkoe DJ, Stocchetti N, Brody DL.  Tau elevations in the
Walsh DM. C-terminally truncated forms of Tau, but brain extracellular space correlate with reduced
not full-length Tau or its C-terminal fragments, are amyloid-beta levels and predict adverse clini-
released from neurons independently of cell death. cal outcomes after severe traumatic brain injury.
J Neurosci. 2015;35:10851–65. Brain. 2012;135:1268–80.
58. Karch CM, Jeng AT, Goate AM.  Extracellular 72. Meredith JE Jr, Sankaranarayanan S, Guss V, Lanzetti
Tau levels are influenced by variability in Tau AJ, Berisha F, Neely RJ, Slemmon JR, Portelius E,
that is associated with tauopathies. J  Biol Chem. Zetterberg H, Blennow K, et al. Characterization of
2012;287:42751–62. novel CSF Tau and pTau biomarkers for Alzheimer’s
59. Karch CM, Jeng AT, Goate AM.  Calcium phos- disease. PLoS One. 2013;8:e76523.
phatase calcineurin influences tau metabolism. 73. Mohamed N, Plouffe V, Remillard-Labrosse G,
Neurobiol Aging. 2013;34:374–86. Planel E, Leclerc NM. Starvation and inhibition of
60. Kaufman SK, Thomas TL, Del Tredici K, Braak H, lysosomal function increased tau secretion by pri-
Diamond MI. Characterization of tau prion seeding mary cortical neurons. Mol Biol Cell. 2014;4:5715.
11  Tau Secretion 133

74. Mohamed NV, Plouffe V, Rémillard-Labrosse G, somal pathway to spread to interconnected neurons.
Planel E, Leclerc N. Starvation and inhibition of Acta Neuropathol Commun. 2018;6:10.
lysosomal function increased tau secretion by pri- 89. Polanco JC, Scicluna BJ, Hill AF, Gotz
mary cortical neurons. Sci Rep. 2014;4:5715. J.  Extracellular vesicles isolated from the brains
75. Mohamed NV, Herrou T, Plouffe V, Piperno N, of rTg4510 mice seed Tau protein aggregation
Leclerc N. Spreading of tau pathology in Alzheimer’s in a threshold-dependent manner. J  Biol Chem.
disease by cell-to-cell transmission. Eur J Neurosci. 2016;291:12445–66.
2013;37:1939–48. 90. Pooler AM, Phillips EC, Lau DHW, Noble W,
76. Mudher A, Colin M, Dujardin S, Medina M, Hanger DP.  Physiological release of endogenous
Dewachter I, Alavi Naini SM, Mandelkow EM, tau is stimulated by neuronal activity. EMBO Rep.
Mandelkow E, Buee L, Goedert M, et al. What is 2013;14:389–94.
the evidence that tau pathology spreads through 91. Potolicchio I, Carven GJ, Xu X, Stipp C, Riese RJ,
prion like propagation? Acta Neuropathol Commun. Stern LJ, Santambrogio L.  Proteomic analysis of
2017;5:99. microglia-derived exosomes: metabolic role of the
77. Nacharaju P, Lewis J, Easson C, Yen S, Hackett J, aminopeptidase CD13  in neuropeptide catabolism.
Hutton M, Yen SH. Accelerated filament formation J Immunol. 2005;175:2237–43.
from tau protein with specific FTDP-17 missense 92. Rabouille C, Malhotra V, Nickel W.  Diversity
mutations. FEBS Lett. 1999;447:195–9. in unconventional protein secretion. J  Cell Sci.
78. Narasimhan S, Guo JL, Changolkar L, Stieber A, 2012;125:5251–5.
McBride JD, Silva LV, He Z, Zhang B, Gathagan 93. Rajendran L, Honsho M, Zahn TR, Keller P,
RJ, Trojanowski JQ, et  al. Pathological Tau Geiger KD, Verkade P, Simons K.  Alzheimer’s
strains from human brains recapitulate the diver- disease beta-amyloid peptides are released in asso-
sity of Tauopathies in nontransgenic mouse brain. ciation with exosomes. Proc Natl Acad Sci U S A.
J Neurosci. 2017;37:11406–23. 2006;103:11172–7.
79. Nawaz M, Fatima F.  Extracellular vesicles, tunnel- 94. Record M, Carayon K, Poirot M, Silvente-Poirot
ing nanotubes, and cellular interplay: synergies and S.  Exosomes as new vesicular lipid transport-
missing links. Front Mol Biosci. 2017;4:50. ers involved in cell-cell communication and vari-
80. Ng F, Tang BL. Unconventional protein secretion in ous pathophysiologies. BBA-Mol Cell Biol L.
animal cells. Methods Mol Biol. 2016;1459:31–46. 2014;1841:108–20.
81. Nickel W, Rabouille C.  Mechanisms of regulated 95. Rustom A, Saffrich R, Markovic I, Walther P, Gerdes
unconventional protein secretion. Nat Rev Mol Cell HH.  Nanotubular highways for intercellular organ-
Biol. 2009;10:148–55. elle transport. Science. 2004;303:1007–10.
82. Ossenkoppele R, Schonhaut DR, Scholl M, Lockhart 96. Saman S, Kim W, Raya M, Visnick Y, Miro S,
SN, Ayakta N, Baker SL, O’Neil JP, Janabi M, Saman S, Jackson B, McKee AC, Alvarez VE, Lee
Lazaris A, Cantwell A, et  al. Tau PET patterns NCY, et  al. Exosome-associated Tau is secreted in
mirror clinical and neuroanatomical variability in tauopathy models and is selectively phosphorylated
Alzheimer’s disease. Brain. 2016;139:1551–67. in cerebrospinal fluid in early Alzheimer disease.
83. Osteikoetxea-Molnar A, Szabo-Meleg E, Toth EA, J Biol Chem. 2012;287:3842–9.
Oszvald A, Izsepi E, Kremlitzka M, Biri B, Nyitray 97. Sanders DW, Kaufman SK, DeVos SL, Sharma AM,
L, Bozo T, Nemeth P, et al. The growth determinants Mirbaha H, Li A, Barker SJ, Foley AC, Thorpe JR,
and transport properties of tunneling nanotube net- Serpell LC, et al. Distinct tau prion strains propagate
works between B lymphocytes. Cell Mol Life Sci. in cells and mice and define different tauopathies.
2016;73:4531–45. Neuron. 2014;82:1271–88.
84. Perez M, Cuadros R, Hernandez F, Avila J. Secretion 98. Scholl M, Lockhart SN, Schonhaut DR, O’Neil JP,
of full-length tau or tau fragments in a cell culture Janabi M, Ossenkoppele R, Baker SL, Vogel JW,
model. Neurosci Lett. 2016;634:63–9. Faria J, Schwimmer HD, et  al. PET imaging of
85. Perez-Gonzalez R, Gauthier SA, Kumar A, Levy tau deposition in the aging human brain. Neuron.
E. The exosome secretory pathway transports amy- 2016;89:971–82.
loid precursor protein carboxyl-terminal fragments 99. Sedgwick AE, D’Souza-Schorey C. The biology of
from the cell into the brain extracellular space. J Biol extracellular microvesicles. Traffic. 2018;19:319–27.
Chem. 2012;287:43108–15. 100. Sepulcre J, Schultz AP, Sabuncu M, Gomez-Isla T,
86. Piccin A, Murphy WG, Smith OP. Circulating mic- Chhatwal J, Becker A, Sperling R, Johnson KA. In
roparticles: pathophysiology and clinical implica- vivo Tau, amyloid, and gray matter profiles in the
tions. Blood Rev. 2007;21:157–71. aging brain. J Neurosci. 2016;36:7364–74.
87. Plouffe V, Mohamed NV, Rivest-McGraw J, Bertrand 101. Simons M, Raposo G. Exosomes – vesicular carri-
J, Lauzon M, Leclerc N. Hyperphosphorylation and ers for intercellular communication. Curr Opin Cell
cleavage at D421 enhance tau secretion. PLoS One. Biol. 2009;21:575–81.
2012;7:e36873. 102. Sokolow S, Henkins KM, Bilousova T, Gonzalez
88. Polanco JC, Li CZ, Durisic N, Sullivan R, Gotz B, Vinters HV, Miller CA, Cornwell L, Poon WW,
J.  Exosomes taken up by neurons hijack the endo- Gylys KH. Pre-synaptic C-terminal truncated tau is
134 Z. Ruan and T. Ikezu

released from cortical synapses in Alzheimer’s dis- 113. van Groen T, Miettinen P, Kadish I. The entorhinal
ease. J Neurochem. 2015;133:368–79. cortex of the mouse: organization of the projec-
103. Stuffers S, Wegner CS, Stenmark H, Brech tion to the hippocampal formation. Hippocampus.
A.  Multivesicular endosome biogenesis in the 2003;13:133–49.
absence of ESCRTs. Traffic. 2009;10:925–37. 114. van Niel G, Bergam P, Di Cicco A, Hurbain I, Lo
104. Tai HC, Serrano-Pozo A, Hashimoto T, Frosch MP, Cicero A, Dingli F, Palmulli R, Fort C, Potier MC,
Spires-Jones TL, Hyman BT.  The synaptic accu- Schurgers LJ, et al. Apolipoprotein E regulates amy-
mulation of hyperphosphorylated tau oligomers in loid formation within endosomes of pigment cells.
Alzheimer disease is associated with dysfunction Cell Rep. 2015;13:43–51.
of the ubiquitin-proteasome system. Am J  Pathol. 115. van Niel G, Charrin S, Simoes S, Romao M, Rochin
2012;181:1426–35. L, Saftig P, Marks MS, Rubinstein E, Raposo G. The
105. Takeda S, Wegmann S, Cho H, DeVos SL, Commins tetraspanin CD63 regulates ESCRT-independent and
C, Roe AD, Nicholls SB, Carlson GA, Pitstick R, -dependent endosomal sorting during melanogen-
Nobuhara CK, et al. Neuronal uptake and propaga- esis. Dev Cell. 2011;21:708–21.
tion of a rare phosphorylated high-molecular-weight 116. van Niel G, D’Angelo G, Raposo G. Shedding light
tau derived from Alzheimer’s disease brain. Nat on the cell biology of extracellular vesicles. Nat Rev
Commun. 2015;6:8490. Mol Cell Biol. 2018;19:213–28.
106. Tardivel M, Begard S, Bousset L, Dujardin S, Coens 117. Vella LJ, Sharples RA, Lawson VA, Masters CL,
A, Melki R, Buee L, Colin M. Tunneling nanotube Cappai R, Hill AF.  Packaging of prions into exo-
(TNT)-mediated neuron-to neuron transfer of patho- somes is associated with a novel pathway of PrP
logical Tau protein assemblies. Acta Neuropathol processing. J Pathol. 2007;211:582–90.
Commun. 2016;4:117. 118. Wang Y, Balaji V, Kaniyappan S, Kruger L, Irsen S,
107. Taylor AR, Robinson MB, Gifondorwa DJ, Tytell Tepper K, Chandupatla R, Maetzler W, Schneider
M, Milligan CE. Regulation of heat shock protein 70 A, Mandelkow E, et  al. The release and trans-­
release in astrocytes: role of signaling kinases. Dev synaptic transmission of Tau via exosomes. Mol
Neurobiol. 2007;67:1815–29. Neurodegener. 2017;12:5.
108. Torralba D, Baixauli F, Villarroya-Beltri C, 119. Wang Y, Martinez-Vicente M, Kruger U, Kaushik S,
Fernandez-Delgado I, Latorre-Pellicer A, Acin-­ Wong E, Mandelkow EM, Cuervo AM, Mandelkow
Perez R, Martin-Cofreces NB, Jaso-Tamame AL, E. Tau fragmentation, aggregation and clearance: the
Iborra S, Jorge I, et al. Priming of dendritic cells by dual role of lysosomal processing. Hum Mol Genet.
DNA-containing extracellular vesicles from acti- 2009;18:4153–70.
vated T cells through antigen-driven contacts. Nat 120. Witter MP, Wouterlood FG, Naber PA, Van Haeften
Commun. 2018;9:2658. T. Anatomical organization of the parahippocampal-­
109. Torres-Cruz FM, Rodriguez-Cruz F, Escobar-­ hippocampal network. Ann N Y Acad Sci.
Herrera J, Barragan-Andrade N, Basurto-Islas G, 2000;911:1–24.
Ripova D, Avila J, Garcia-Sierra F.  Expression of 121. Wu JW, Herman M, Liu L, Simoes S, Acker CM,
Tau produces aberrant plasma membrane blebbing in Figueroa H, Steinberg JI, Margittai M, Kayed R,
glial cells through RhoA-ROCK-dependent F-actin Zurzolo C, et  al. Small misfolded Tau species are
remodeling. J Alzheimers Dis. 2016;52:463–82. internalized via bulk endocytosis and anterogradely
110. Trajkovic K, Hsu C, Chiantia S, Rajendran L, and retrogradely transported in neurons. J  Biol
Wenzel D, Wieland F, Schwille P, Brugger B, Chem. 2013;288:1856–70.
Simons M.  Ceramide triggers budding of exosome 122. Yamada K, Cirrito JR, Stewart FR, Jiang H, Finn
vesicles into multivesicular endosomes. Science. MB, Holmes BB, Binder LI, Mandelkow EM,
2008;319:1244–7. Diamond MI, Lee VM, et al. In vivo microdialysis
111. Usenovic M, Niroomand S, Drolet RE, Yao LH, reveals age-dependent decrease of brain interstitial
Gaspar RC, Hatcher NG, Schachter J, Renger JJ, fluid tau levels in P301S human tau transgenic mice.
Parmentier-Batteur S.  Internalized Tau oligomers J Neurosci. 2011;31:13110–7.
cause neurodegeneration by inducing accumula- 123. Yamada K, Holth JK, Liao F, Stewart FR, Mahan
tion of pathogenic Tau in human neurons derived TE, Jiang H, Cirrito JR, Patel TK, Hochgrafe
from induced pluripotent stem cells. J  Neurosci. K, Mandelkow EM, et  al. Neuronal activity
2015;35:14234–50. regulates extracellular tau in  vivo. J  Exp Med.
112. Valadi H, Ekstrom K, Bossios A, Sjostrand M, 2014;211:387–93.
Lee JJ, Lotvall JO.  Exosome-mediated transfer 124. Zhang JY, Liu SJ, Li HL, Wang JZ.  Microtubule-­
of mRNAs and microRNAs is a novel mechanism associated protein tau is a substrate of ATP/Mg(2+)-
of genetic exchange between cells. Nat Cell Biol. dependent proteasome protease system. J  Neural
2007;9:654–9. Transm (Vienna). 2005;112:547–55.
 Emerging Connections Between
Tau and Nucleic Acids 12
Marie-Christine Galas, Eliette Bonnefoy, Luc Buee,
and Bruno Lefebvre

Introduction the tau species described as interacting with DNA

in the nucleus is dephosphorylated at the epitope
Tau has long been defined as a microtubule-­ S195–S202, and recognized by the tau1 antibody
associated protein (MAP) although early studies (Fig. 12.1).
have reported its capacity to bind to DNA and The sequences of both tau and DNA, involved
RNA.  The links between tau and nucleic acids in the Tau-DNA complex formation, have been
have been largely underestimated until recently recently examined in more detail.
when several reports highlighted new key roles of
tau in relation with DNA and RNA structure, Tau Sequences Involved in DNA
metabolism and integrity. This chapter will focus Binding
on recent research on tau and nucleic acids con- The identification, by electromobility shift assays
nections, and their implications in tauopathies. (EMSA) and nuclear magnetic resonance (NMR),
of tau sequences interacting with DNA demon-
strated the importance of amino acids localized
Tau and DNA Connections in the proline-rich domain (PRD) and the second
repeat (R2) of the microtubule-binding domain
Tau-DNA Complex Formation (MTBD) [5, 6]. A consensus of research studies
supports the hypothesis broad phosphorylation of
The interaction of tau with DNA was first tau reduces the capacity of tau protein to bind
reported in vitro in 1980 [1]. Later, nuclear local- DNA [6–8]. However, the biochemical studies
ization of tau and tau-DNA complex formation done to date have investigated whether tau inter-
were observed in various cellular models includ- acts with DNA only in vitro, which leaves open
ing neurons (for reviews see [2–4]). In neurons, the question of whether tau is able to directly
bind DNA in vivo.

 NA Sequences Targeted by Tau

D
M.-C. Galas (*) · L. Buee · B. Lefebvre Using gel retardation techniques, recombinant
University of Lille, INSERM, CHU-Lille, UMR-S
human tau protein was shown to bind and form
1172, Alzheimer & Tauopathies, LabEx DISTALZ,
Lille, France protein-DNA complexes with double-stranded
e-mail: marie-christine.galas@inserm.fr DNA (either supercoiled, relaxed closed circular
E. Bonnefoy or linear) [9], and tau was reported to bind to
Institut Cochin, Inserm U1016, Paris, France single-strand DNA by capillary electrophoresis

© Springer Nature Singapore Pte Ltd. 2019 135

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_12
136 M.-C. Galas et al.

Tau pathology-induced nucleic acid alterations Tau functions

DNA damage DNA protection

Transposable elements activation

Pericentomeric Heterochromatin
Aneuploïdy structuration
transcription into lncRNA
Heterochromatin alteration DNA integrity
Perinucleolar Heterochromatin
rRNA DNA stability
rRNA transcription
Nucleus

Cytoplasm RNA damage RNA protection

Tau-RNA interaction
Tau aggregation
Tau-RNA binding Protein
TIA1,…..

Tau-DDX6
mRNA biogenesis/silencing alteration miRNA silencing activity

RNA damage RNA protection

Fig. 12.1  Schematic representation of tau and tau pathology connections with nucleic acids

methods [10]. Further in vitro gel retardations tau1 antibody followed by DNA tiling array
assays revealed a preference of tau binding to hybridization (ChIP-on-chip) assays. Tau protein
A/T rich sequences compared with G/C-rich [5]. showed a preference for DNA regions overlap-
However, this preference was not confirmed by ping with DNA sequences coding for long non-­
NMR [6]. In neuronal cultures, results obtained coding (lnc)RNAs and for DNA regions
with netropsin, which is a polyamide with antibi- comprising an AG-rich GAGA-like motifs. The
otic and antiviral activity, suggested that in situ, intragenic tau-interacting DNA regions also
tau may also interact with DNA through the exhibit functional enrichment for diseases associ-
minor groove of A/T-rich DNA sequences [11]. ated with neurological disorders [12].
This work is supported by the analysis of the Interestingly, under heat stress conditions in adult
interaction of tau protein with plasmid DNA by mouse cortex and hippocampus neurons, tau
electron microscopy, which showed the forma- behaved as a repressor of the expression of a sub-
tion of regularly spaced bead like structures in a set of tau-interacting genes, although under basal
necklace manner [9]. conditions tau protein displayed DNA-binding
Recently, the first large scale genome-wide characteristics different from those expected for a
analysis of tau interaction with neuronal DNA conventional transcription factor. This stress-­
was published [12]. Specific intragenic and inter- linked transcriptional repressor role of tau pro-
genic DNA sequences interacting with tau were tein was exacerbated by the nuclear accumulation
identified in differentiated neurons using genome-­ of pathological oligomerized forms of tau protein
wide chromatin immunoprecipitation with the in hippocampal neurons [12].
12  Emerging Connections Between Tau and Nucleic Acids 137

 au and the Maintenance of Genome
T in some cancer cells, it opens up possibility for
Integrity relevant predictive markers of susceptibility to
anti-­cancer treatments.
DNA Protection
The capacity of tau protein to protect DNA from Chromatin Organization
free radical induced damage and heat induced Several reports have highlighted the implication
denaturation was first reported in vitro [5, 9]. In of tau in chromatin organization and a genome
vitro experiments also showed that hyperphos- 3D structure in neurons, altered in the context of
phorylation reduced or altered tau DNA binding tau pathology and AD [17–20].
and folding capacity, impairing the DNA protec-
tive effect of tau [13]. These studies led to the Pericentromeric Heterochromatin
hypothesis that the protective activity against free Pericentromeric heterochromatin (PCH) is com-
radical attack was linked to DNA folding pro- posed of highly repeated major satellite DNA
moted by tau binding to double stranded (ds) sequences. PCH displays a highly ordered
DNA. nucleosome distribution rich in particular epi-
Later, the capacity of tau to bind DNA and to genetic marks such as the trimethylated form of
protect DNA integrity was highlighted in murine lysine 9 of histone H3 (H3K9me3), giving rise to
neurons, in primary neuronal cultures and in vivo compact chromatin regions that influence
in the hippocampus and cortex, under physiolog- genome stability and gene expression regulation.
ical and stress conditions [11, 14]. These results Using neurons from wild type (WT) and tau
suggested that the DNA protective function of deficient (KOTau) mice, Mansuroglu et al. dem-
tau in neurons could rely partly on its capacity to onstrated that tau protein interacts with PCH
interact with DNA.  However, these studies did major satellite sequences and regulates PCH
not exclude an alternative mechanism in which structure integrity in differentiated neurons [18].
tau contributes to this process by regulating other Also, PCH structures were shown to be altered in
proteins in the DNA repair pathways [14]. Drosophila and murine models of tauopathies
DNA damage is an early feature of neurons and in AD brain [17, 18] and a widespread tran-
developing tau pathology in AD brain. In vivo in scriptional increase of genes heterochromatically
hippocampal neurons, nucleic acid vulnerability silenced under physiological conditions, was
has been linked to the nuclear accumulation of observed in AD brains [17]. Altogether, these
prefibrillar tau oligomers that form in the process observations suggest that tau pathology affects
of tau pathophysiology [15]. However, the mech- the physiological PCH regulatory function of tau
anism responsible of the loss of the protective leading to an alteration of PCH structure and
function of monomeric tau in the presence of tau function in neurons.
oligomerization has not been deciphered. Small PCH DNA sequences can be transcribed into
tau oligomers retain the capacity to interact with non-coding RNAs (ncRNAs) that exist in both,
DNA in vitro [8]. Although not demonstrated in sense and antisense, orientations. Tau has been
vivo, a potential role of tau oligomers in compet- shown to repress the transcription of the anti-
ing with tau monomers for DNA-binding has sense PCH ncRNAs [18]. This repression could
been hypothesized to explain this loss of protec- modulate the ratio of sense:antisense PCH
tive function. ncRNAs, which itself would lead to the regula-
The role of tau in maintaining genome integ- tion of the ratio of single- versus double-stranded
rity is not limited to neurons. Cells from Bloom’s conformations of PCH ncRNAs in neurons [18].
syndrome (BS) patients characterized by muta- In the nucleus, the binding of the single strand
tions in the BLM gene, a 3′–5′ DNA helicase, sense PCH ncRNA to its complementary PCH
display a strong genome instability. Interestingly DNA sequence is necessary for the establish-
BS cells use tau to restrict DNA damage contrib- ment of the trimethylated form of lysine 9 of his-
uting to cell survival [16]. As tau is overexpressed tone H3 (H3K9me3) and the recruitment of
138 M.-C. Galas et al.

HP1alpha protein on major satellite DNA to localize at the periphery of the nucleolus par-
sequences required to initiate the heterochroma- tially colocalizing with constitutive heterochro-
tization of pericentromeric major satellite matin [24]. Studies in neuroblastoma cells
DNA.  Dysregulation of the physiological equi- suggest that tau regulates rRNA transcription
librium between sense and antisense PCH [25]. A nucleolar localization was also described
ncRNAs, induced by tau deletion or tau pathol- for the phosphorylated epitope (Ser212/Thr214)
ogy, could contribute to the change in PCH and abnormal conformation of tau protein recog-
structure observed in KO tau neurons and AD nized by the AT100 antibody in human neurons
brain [18]. from the CA1 region that varies during cell aging
In Drosophila and mammalian cells, repair of [19]. The recruitment of tau protein to rRNA
PCH DSBs occurs through two steps. In the first DNA loci, along with upstream binding factor
step, DNA damage response proteins, such as the (UBTF), has been shown in cells from BS
phosphorylated form of the histone variant H2AX patients. In this case, a protective role for tau pro-
(γH2AX) are recruited within the core of clusters tein was described, necessary to preserve rRNA
of PCH structures named chromocenters. In the DNA stability and modulate rRNA transcription
second step, the damaged DNA sequences are [16]. Interestingly, important alterations in the
relocated from the inside to the periphery of protein synthesis machinery involving the nucle-
chromocenters, leading to the formation of long-­ olus have been observed in the hippocampus of
lasting γH2AX foci positioned at the periphery of advance stages of AD [26]. Tau-mediated ribo-
chromocenters. Tau deletion prevented the transi- somal dysfunction caused by pathological tau has
tion from the first to the second step in stressed also been suggested as a pathogenic process
hippocampal neurons in vivo leading to the per- potentially leading to cognitive impairment [27].
manent accumulation of γH2AX foci and DSBs Overall these results suggested that tau protein
within PCH structures [18]. These results show could be involved in rRNA transcription, ribo-
that alongside with a role in the organization of somal biogenesis and ribosomal function.
PCH, tau protein also participates in the control
of the double strand break (DSB) DNA repair  enomic Instability in Tauopathies
G
(DDR) process of neuronal PCH DNA [18]. Affected neurons in AD are characterized by
genomic instability including changes in nucleic
Perinucleolar Heterochromatin and rDNA acid sequences, chromosomal rearrangements or
Genes aneuploidy [28]. A growing body of evidence
During the 1990’s, the laboratory of Lester brought to light the involvement of tau-mediated
Binder described the presence of unphosphory- mechanisms leading to neuronal genomic
lated tau protein recognized by the tau1 antibody instability.
in the nucleolus of interphase human neuroblas-
toma cells as well as other non-neuronal human Transposable Elements
cell lines. This nucleolar localization of tau was In contrast to the initial idea that the neuronal
observed also in monkey cell lines, but not in genome was a static entity, there is now abun-
non-primate cells [21]. Further analyses local- dant data showing that rearrangement of trans-
ized tau protein to the fibrillar region of the posable DNA elements occurs and modulate
nucleolus that contains rRNA genes and corre- genome stability and expression in neurons.
sponds to the region where transcription of rRNA Transposable elements (TE) are mobile genetic
genes occurs. In agreement with these results, in sequences that can jump from one location to
mitotic cells, tau localized to the nucleolar orga- another in the genome. Rearrangement of ret-
nizer region (NOR) that contains the tandemly rotransposons, transposons with RNA interme-
repeated DNA sequences coding for rRNAs [22, diates, have been documented in adult neurons,
23]. In human non-neuronal cells, tau was shown in human brains [29]. TE expression is tightly
12  Emerging Connections Between Tau and Nucleic Acids 139

regulated and limited by heterochromatin con- Tau and RNA Connections

densation that transcriptionally silences TE loci,
and post-transcriptional events such as clearance A clear relationship between tau and RNA has
via piwi-interacting RNAs (piRNA) [30]. been demonstrated by numerous investigations
Transposition of TE is abnormally activated in over the past decades. Although most of them
neurological disorders such as TDP43-mediated were discovered in the context of tauopathies, it
neurodegeneration [31], and may lead to further suggests that tau could be involved in
increased genome instability and transcriptional some of these mechanisms in physiological
deregulation. Recently, two different groups conditions.
reported abnormal TE activation in AD brains
and in Drosophila transgenic models expressing
wild type or mutated (R406W) tau [32, 33], sug- Tau-RNA Interaction
gesting that pathological forms of tau deregulate
TE expression, possibly via heterochromatin Early studies demonstrated that RNA could
relaxation and/or piRNA depletion. bind to tau in vitro and promote nucleation of
Consequently, tau pathology-related TE activa- tau oligomers in the same way as other nega-
tion may contribute to genomic instability in tively charged cofactors such as heparin [41,
tauopathies. 42]. Zhang et  al. reported recently that tau
associated with tRNA in different cell types
Aneuploidy (HEK-293 and human-induced pluripotent
Aneuploidy, defined as an abnormal number of stem cells (IPSCs) derived neurons). Mixing
chromosomes, represents another form of tau and tRNA in vitro led to the formation of
genomic instability and is strongly associated droplets, known as complex coacervates, and
with tau dysfunction. Cells from patients and creating conditions in which tau may become
mice carrying tau mutations exhibit aneuploidy vulnerable to aggregation (see Chaps. 25 and
[34–36]. Drosophila overexpressing 4R (4 26) [43]. Although the existence of these drop-
repeats) tau isoform, but not 3R (3 repeats), lets has not been demonstrated in vivo, trans-
exhibit a mitotic block characterized by spindle fection of tau in neuronally derived IPSCs
defects that lead to mosaic aneuploidy [37, 38]. increased the presence of tau in the sarkosyl
Overexpressing wild type or mutant 4R-tau in insoluble fraction. It is not yet known if the
neuroblastoma cells also induced monopolar interaction of tau with tRNA is involved in
spindles [38]. The relative distribution of 3R and other pathological mechanisms, such as the
4R isoforms is crucial during brain development. translational inhibition observed with the path-
An imbalance in the expression of tau isoforms ological forms of tau [27].
may lead to the genesis of aneuploid cells during In a more general way, we observed that tau
brain development that survive in adult life that deficiency, either from knock-out or pathologi-
could present a vulnerability leading to age-­ cal conditions, triggers alterations in RNA integ-
related neuronal degeneration. Altogether these rity in mouse hippocampal neurons under
results suggest the intriguing hypothesis that a physiological, heat-stress and ROS-producing
developmental origin of tauopathies may exist. conditions [14, 15]. RNA export machinery
DNA damage and DNA repair dysfunction seems also to be involved in tau-induced neuro-
also appear to alter neuronal chromatin stability toxicity. It has been reported that tau reduced the
in AD [39, 40], suggesting that dysregulation of protein level of Lamin protein to cause nucleo-
the roles of tau in maintaining genome integrity plasmic reticulum expansion in Drosophila and
(cf. parts 13.2.1. and 13.2.2.) might contribute to human patients with Alzheimer’s disease [44].
the development of genomic instability early in The authors showed recently that messenger
the course of AD. RNA (mRNA) accumulated in these expansions.
140 M.-C. Galas et al.

Interestingly, pharmacological inhibition or first translational repression and second mRNA

genetic knock-­down in Drosophila suppressed decay, comprising deadenylation, decapping
tau-induced neurotoxicity [45]. and 5′-3′ decay of mRNA. Although the mecha-
nisms have been not fully elucidated, it has been
demonstrated that CCR4-NOT present in the
RNA Binding Proteins miRISC complex recruits DDX6 and thus pro-
vides a link between translational inhibition and
The relationship between abnormal aggregation decapping [50].
of tau and RNA metabolism is not limited to The Tau/DDX6 interaction increased the
RNA molecules alone but could involve also translation repression by miRNAs. Importantly,
RNA binding proteins (RBPs). For example, the tau mutations (P301S, P301L) found in the inher-
stress granules nucleating protein T cell intracel- ited tauopathies, frontotemporal dementia and
lular antigen 1 (TIA1) promotes tau misfolding parkinsonism linked to chromosome 17, disrupt
[46] (see Chap. 27 for further details). Part of the tau/DDX6 interaction and impair miRNA silenc-
translational stress response involves the forma- ing [49]. Furthermore, DDX6 accumulates in the
tion of stress granules (SGs) in the soma and den- cytoplasm of neurons in AD and primary tauopa-
drites of neurons. SGs are cytoplasmic aggregates thy Corticobasal Degeneration (CBD) human
composed of proteins and RNA molecules that brains, mainly colocalized with hyperphosphory-
repress translation of a subset of RNAs. In patho- lated tau, indicating that the level of DDX6 is
logical conditions, it has been shown that the increased in neurons from tauopathy brains [49].
interaction between tau and TIA1 accelerates the DDX6 also accumulated in the cytoplasm of
formation and the size of SGs, promoting tau some neurons devoid of pathological phosphory-
aggregation. Interestingly, the authors showed lated tau, suggesting that an increase in DDX6
that TIA1 depletion or pharmacological inhibi- abundance is an early event in the development of
tion reduced SGs formation and pathological tau tau pathology. Several miRNAs have been
misfolding. Further work from the same group involved in the progression of AD or tauopathies.
demonstrated that the interaction between patho- We found in particular that tau/DDX6 interaction
logical tau and RBPs is not limited to TIA1 but increased the translational repression miR-21 and
could involve several proteins such as hnRNPA0, miR-124. miR-21 restored the cognitive deficits
EWSR1, PABP, RPL7 and DDX6 [47]. in APP/PS1 mice and prevented pathologic fea-
These observations raised questions about tures [52]. Recent studies demonstrated that miR-­
whether tau/RBPs interaction is limited to path- 124 decreased in AD [53]. Of interest, BACE1,
ological conditions and linked to pathological an enzyme involved in Aβ production, is a direct
tau or could be extended to physiological condi- target of this miRNA [53]. In addition, recent
tions. Indeed, proteomic studies revealed an report showed an alteration in several miRNA
interaction of tau with several complexes expression (miR-10a-5p, miR-142a-5p, miR-­
involved in RNA metabolism, including the 146a-­ 5p, miR-155-5p, miR-211-5p and miR-­
DEAD-box RNA helicase DDX6 [48, 49]. 455-­5p) in tau transgenic mice [54]. However, it
DDX6 is involved in translational repression, is still unknown whether tau, through the interac-
RNA decay and miRNA pathway [50]. It is also tion with RBPs, regulates miRNA biogenesis or
required for the assembly of processing bodies whether these changes are a consequence of the
(P-bodies) a class of RNA granules found in pathology. The identification of several RBPs
somatic cells [51]. Tau increased miRNA silenc- that interact with tau suggest a more general role
ing activities through DDX6 interaction [49]. in RNA metabolism. However, further studies are
Post-transcriptional silencing by miRNAs is needed to elucidate the role of tau/RBP com-
thought to occur by two distinct mechanisms, plexes in this evolving field.
12  Emerging Connections Between Tau and Nucleic Acids 141

Conclusion 9. Hua Q, He RQ, Haque N, Qu MH, del Carmen Alonso

A, Grundke-Iqbal I, et  al. Microtubule associated
protein tau binds to double-stranded but not single-­
Nucleic acids are key macromolecules which stranded DNA. Cell Mol Life Sci. 2003;60(2):413–21.
hold essential information to maintain normal 10. Krylova SM, Musheev M, Nutiu R, Li Y, Lee G,
functioning cells. The growing body of evidence Krylov SN.  Tau protein binds single-stranded DNA
sequence specifically--the proof obtained in vitro with
implicating tau and tau pathology in mecha- non-equilibrium capillary electrophoresis of equilib-
nisms regulating genome integrity, chromatin rium mixtures. FEBS Lett. 2005;579(6):1371–5.
organization and RNA metabolism, suggests 11. Sultan A, Nesslany F, Violet M, Begard S, Loyens A,
that tau is a major player of neuronal homeosta- Talahari S, et al. Nuclear tau, a key player in neuronal
DNA protection. J Biol Chem. 2011;286(6):4566–75.
sis. Interestingly, the links between tau and 12. Benhelli-Mokrani H, Mansuroglu Z, Chauderlier A,
nucleic acids are not restricted to neurons and Albaud B, Gentien D, Sommer S, et  al. Genome-­
may play also important roles in cancer cells. wide identification of genic and intergenic neu-
Clearly, what we currently know about connec- ronal DNA regions bound by Tau protein under
physiological and stress conditions. Nucleic Acids
tions between the multifunctional tau protein Res. 2018;46(21):11405–22.
and DNA/RNA is still an emerging but fascinat- 13. Lu Y, He HJ, Zhou J, Miao JY, Lu J, He YG, et  al.
ing research field that definitely needs further Hyperphosphorylation results in tau dysfunction
investigations. in DNA folding and protection. J  Alzheimers Dis.
2013;37(3):551–63.
14. Violet M, Delattre L, Tardivel M, Sultan A,

Chauderlier A, Caillierez R, et al. A major role for Tau
References in neuronal DNA and RNA protection in vivo under
physiological and hyperthermic conditions. Front Cell
1. Villasante A, Corces VG, Manso-Martinez R, Avila Neurosci. 2014;8:84.
J. Binding of microtubule protein to DNA and chro- 15. Violet M, Chauderlier A, Delattre L, Tardivel M,
matin: possibility of simultaneous linkage of micro- Chouala MS, Sultan A, et  al. Prefibrillar Tau oligo-
tubule to nucleic and assembly of the microtubule mers alter the nucleic acid protective function of
structure. Nucleic Acids Res. 1981;9(4):895–908. Tau in hippocampal neurons in vivo. Neurobiol Dis.
2. Multhaup G, Huber O, Buee L, Galas MC. Amyloid 2015;82:540–51.
precursor protein (APP) metabolites APP intracellular 16. Bou Samra E, Buhagiar-Labarchede G, Machon C,
fragment (AICD), Abeta42, and tau in nuclear roles. Guitton J, Onclercq-Delic R, Green MR, et al. A role
J Biol Chem. 2015;290(39):23515–22. for Tau protein in maintaining ribosomal DNA stabil-
3. Bukar Maina M, Al-Hilaly YK, Serpell LC. Nuclear ity and cytidine deaminase-deficient cell survival. Nat
tau and its potential role in Alzheimer’s disease. Commun. 2017;8(1):693.
Biomol Ther. 2016;6(1):9. 17. Frost B, Hemberg M, Lewis J, Feany MB.  Tau pro-
4. Sotiropoulos I, Galas MC, Silva JM, Skoulakis E, motes neurodegeneration through global chromatin
Wegmann S, Maina MB, et al. Atypical, non-standard relaxation. Nat Neurosci. 2014;17(3):357–66.
functions of the microtubule associated Tau protein. 18. Mansuroglu Z, Benhelli-Mokrani H, Marcato V,
Acta Neuropathol Commun. 2017;5(1):91. Sultan A, Violet M, Chauderlier A, et al. Loss of Tau
5. Wei Y, Qu MH, Wang XS, Chen L, Wang DL, Liu protein affects the structure, transcription and repair
Y, et al. Binding to the minor groove of the double-­ of neuronal pericentromeric heterochromatin. Sci
strand, tau protein prevents DNA from damage by Rep. 2016;6:33047.
peroxidation. PLoS One. 2008;3(7):e2600. 19. Gil L, Federico C, Pinedo F, Bruno F, Rebolledo AB,
6. Qi H, Cantrelle FX, Benhelli-Mokrani H, Smet-Nocca Montoya JJ, et al. Aging dependent effect of nuclear
C, Buee L, Lippens G, et al. Nuclear magnetic reso- tau. Brain Res. 2017;1677:129–37.
nance spectroscopy characterization of interaction of 20. Winick-Ng W, Rylett RJ. Into the fourth dimension:
Tau with DNA and its regulation by phosphorylation. dysregulation of genome architecture in aging and
Biochemistry. 2015;54(7):1525–33. Alzheimer’s disease. Front Mol Neurosci. 2018;11:60.
7. Hua Q, He RQ. Effect of phosphorylation and aggre- 21. Loomis PA, Howard TH, Castleberry RP, Binder
gation on tau binding to DNA.  Protein Pept Lett. LI.  Identification of nuclear tau isoforms in human
2002;9(4):349–57. neuroblastoma cells. Proc Natl Acad Sci U S A.
8. Camero S, Benitez MJ, Cuadros R, Hernandez F, 1990;87(21):8422–6.
Avila J, Jimenez JS. Thermodynamics of the interac- 22. Thurston VC, Zinkowski RP, Binder LI.  Tau as a
tion between Alzheimer’s disease related tau protein nucleolar protein in human nonneural cells in  vitro
and DNA. PLoS One. 2014;9(8):e104690. and in vivo. Chromosoma. 1996;105(1):20–30.
142 M.-C. Galas et al.

23. Thurston VC, Pena P, Pestell R, Binder LI. Nucleolar 37. Bouge AL, Parmentier ML. Tau excess impairs mito-
localization of the microtubule-associated protein sis and kinesin-5 function, leading to aneuploidy and
tau in neuroblastomas using sense and anti-sense cell death. Dis Model Mech. 2016;9(3):307–19.
transfection strategies. Cell Motil Cytoskeleton. 38. Malmanche N, Dourlen P, Gistelinck M, Demiautte
1997;38(1):100–10. F, Link N, Dupont C, et  al. Developmental expres-
24. Sjoberg MK, Shestakova E, Mansuroglu Z, Maccioni sion of 4-repeat-tau induces neuronal aneuploidy in
RB, Bonnefoy E. Tau protein binds to pericentromeric Drosophila tauopathy models. Sci Rep. 2017;7:40764.
DNA: a putative role for nuclear tau in nucleolar orga- 39. Madabhushi R, Pan L, Tsai LH.  DNA dam-

nization. J Cell Sci. 2006;119(Pt 10):2025–34. age and its links to neurodegeneration. Neuron.
25. Maina MB, Bailey LJ, Wagih S, Biasetti L, Pollack 2014;83(2):266–82.
SJ, Quinn JP, et al. The involvement of tau in nucle- 40. Hou Y, Song H, Croteau DL, Akbari M, Bohr

olar transcription and the stress response. Acta VA.  Genome instability in Alzheimer disease. Mech
Neuropathol Commun. 2018;6(1):70. Ageing Dev. 2017;161(Pt A):83–94.
26. Hernandez-Ortega K, Garcia-Esparcia P, Gil L, Lucas 41. Dinkel PD, Holden MR, Matin N, Margittai M. RNA
JJ, Ferrer I.  Altered machinery of protein synthesis binds to tau fibrils and sustains template-assisted
in Alzheimer’s: from the nucleolus to the ribosome. growth. Biochemistry. 2015;54(30):4731–40.
Brain Pathol. 2016;26(5):593–605. 42. Kampers T, Friedhoff P, Biernat J, Mandelkow

27. Meier S, Bell M, Lyons DN, Rodriguez-Rivera J,
EM, Mandelkow E.  RNA stimulates aggrega-
Ingram A, Fontaine SN, et  al. Pathological tau pro- tion of microtubule-associated protein tau into
motes neuronal damage by impairing ribosomal func- Alzheimer-like paired helical filaments. FEBS Lett.
tion and decreasing protein synthesis. J  Neurosci. 1996;399(3):344–9.
2016;36(3):1001–7. 43. Zhang X, Lin Y, Eschmann NA, Zhou H, Rauch JN,
28. Zekanowski C, Wojda U.  Aneuploidy, chromo-
Hernandez I, et al. RNA stores tau reversibly in com-
somal missegregation, and cell cycle reentry in plex coacervates. PLoS Biol. 2017;15(7):e2002183.
Alzheimer's disease. Acta Neurobiol Exp (Wars). 44. Frost B, Bardai FH, Feany MB.  Lamin dysfunction
2009;69(2):232–53. mediates neurodegeneration in tauopathies. Curr Biol.
29. Baillie JK, Barnett MW, Upton KR, Gerhardt DJ, 2016;26(1):129–36.
Richmond TA, De Sapio F, et al. Somatic retrotrans- 45. Cornelison GL, Levy SA, Jenson T, Frost B.  Tau-­
position alters the genetic landscape of the human induced nuclear envelope invagination causes a toxic
brain. Nature. 2011;479(7374):534–7. accumulation of mRNA in Drosophila. Aging Cell.
30. Brennecke J, Aravin AA, Stark A, Dus M, Kellis
2018;18:e12847.
M, Sachidanandam R, et  al. Discrete small RNA-­ 46. Apicco DJ, Ash PEA, Maziuk B, LeBlang C,

generating loci as master regulators of transposon Medalla M, Al Abdullatif A, et  al. Reducing the
activity in Drosophila. Cell. 2007;128(6):1089–103. RNA binding protein TIA1 protects against tau-
31. Li W, Jin Y, Prazak L, Hammell M, Dubnau
mediated neurodegeneration in vivo. Nat Neurosci.
J. Transposable elements in TDP-43-mediated neuro- 2018;21(1):72–80.
degenerative disorders. PLoS One. 2012;7(9):e44099. 47. Maziuk BF, Apicco DJ, Cruz AL, Jiang L, Ash

32. Sun W, Samimi H, Gamez M, Zare H, Frost
PEA, da Rocha EL, et al. RNA binding proteins co-­
B.  Pathogenic tau-induced piRNA depletion pro- localize with small tau inclusions in tauopathy. Acta
motes neuronal death through transposable element Neuropathol Commun. 2018;6(1):71.
dysregulation in neurodegenerative tauopathies. Nat 48. Gunawardana CG, Mehrabian M, Wang X, Mueller
Neurosci. 2018;21(8):1038–48. I, Lubambo IB, Jonkman JE, et  al. The human tau
33. Guo C, Jeong HH, Hsieh YC, Klein HU, Bennett interactome: binding to the ribonucleoproteome, and
DA, De Jager PL, et  al. Tau activates transpos- impaired binding of the proline-to-leucine mutant at
able elements in Alzheimer’s disease. Cell Rep. position 301 (P301L) to chaperones and the protea-
2018;23(10):2874–80. some. Mol Cell Proteomics. 2015;14(11):3000–14.
34. Rossi G, Conconi D, Panzeri E, Redaelli S, Piccoli 49. Chauderlier A, Gilles M, Spolcova A, Caillierez R,
E, Paoletta L, et al. Mutations in MAPT gene cause Chwastyniak M, Kress M, et  al. Tau/DDX6 interac-
chromosome instability and introduce copy number tion increases microRNA activity. Biochim Biophys
variations widely in the genome. J  Alzheimers Dis. Acta Gene Regul Mech. 2018;1861(8):762–72.
2013;33(4):969–82. 50. Mathys H, Basquin J, Ozgur S, Czarnocki-Cieciura
35. Rossi G, Conconi D, Panzeri E, Paoletta L, Piccoli M, Bonneau F, Aartse A, et  al. Structural and bio-
E, Ferretti MG, et  al. Mutations in MAPT give chemical insights to the role of the CCR4-NOT com-
rise to aneuploidy in animal models of tauopathy. plex and DDX6 ATPase in microRNA repression.
Neurogenetics. 2014;15(1):31–40. Mol Cell. 2014;54(5):751–65.
36. Rossi G, Dalpra L, Crosti F, Lissoni S, Sciacca FL, 51. Minshall N, Kress M, Weil D, Standart N.  Role of
Catania M, et  al. A new function of microtubule-­ p54 RNA helicase activity and its C-terminal domain
associated protein tau: involvement in chromosome in translational repression, P-body localization and
stability. Cell Cycle. 2008;7(12):1788–94. assembly. Mol Biol Cell. 2009;20(9):2464–72.
12  Emerging Connections Between Tau and Nucleic Acids 143

52. Cui GH, Wu J, Mou FF, Xie WH, Wang FB, 5 3. An F, Gong G, Wang Y, Bian M, Yu L, Wei C. MiR-­

Wang QL, et  al. Exosomes derived from hypoxia-­ 124 acts as a target for Alzheimer's disease by regulat-
preconditioned mesenchymal stromal cells ameliorate ing BACE1. Oncotarget. 2017;8(69):114065–71.
cognitive decline by rescuing synaptic dysfunction 54. Sierksma A, Lu A, Salta E, Vanden Eynden E,
and ­regulating inflammatory responses in APP/PS1 Callaerts-Vegh Z, D'Hooge R, et al. Deregulation of
mice. FASEB J. 2018;32(2):654–68. neuronal miRNAs induced by amyloid-beta or TAU
pathology. Mol Neurodegener. 2018;13(1):54.
 Tau Interacting Proteins: Gaining
Insight into the Roles of Tau 13
in Health and Disease

Ilie-Cosmin Stancu, Mattia Ferraiolo, Dick Terwel,

and Ilse Dewachter

Tauopathies Position Tau different symptoms governed by the affected

as a Crucial Protein in Health brain regions. The most common Tauopathies are
and Disease corticobasal degeneration (CBD), Pick’s disease,
progressive supranuclear palsy (PSP) and fronto-
Tau is most intensely studied in relation to its temporal dementias with parkinsonism linked to
executive role in Tauopathies, a family of neuro- chromosome 17 (FTDP-17). However a growing
degenerative disorders characterized by the accu- number of diseases are characterized by Tau
mulation of Tau aggregates [15, 21, 38, 75, 89, aggregation amounting to a large family of more
111, 121, 135, 175, 176, 192]. Tau aggregation in than 20 disorders [176]. Most Tauopathies are
the different Tauopathies differs in the affected sporadic, and are hence linked to a combination
cell type, the structure of aggregates and Tau iso- of environmental and genetic risk factors.
form composition. However, in all Tauopathies, However, mutations in MAPT have been identi-
accumulation of pathological Tau in well-charac- fied which are autosomal dominantly linked to
terized and well-defined brain regions, correlates Tauopathies, including FTDP, PSP and CBD [94,
strongly with symptoms associated with the dys- 163, 185] (Alzforum, https://www.alzforum.org/
function of this brain region. Hence, symptoms mutations/mapt). More than 80 mutations have
of neurodegenerative Tauopathies can range from been identified in MAPT, both in intronic and
motoric to cognitive and behavioral symptoms, exonic regions of the human MAPT. These muta-
even extending to deterioration of vital functions tions can be classified as missense mutations or
when the disease progresses, or combinations of splicing mutations. Most missense mutations
cluster in or near the microtubule binding site of
I.-C. Stancu · D. Terwel Tau, while most splicing mutations affect the
BIOMED, Alzheimer Research Group, splicing of exon 10 (encoding the R2 domain),
Hasselt University, Hasselt, Belgium and hence affect the 3R/4R ratio. While
M. Ferraiolo Alzheimer’s disease (AD), is the most prevalent
Institute of Neuroscience, Catholic University Tauopathy, no mutations in MAPT associated
of Louvain, Brussels, Belgium with AD have been identified. Brains of AD
I. Dewachter (*) patients are pathologically characterized by the
BIOMED, Alzheimer Research Group, combined presence of amyloid plaques and neu-
Hasselt University, Hasselt, Belgium
rofibrillary tangles [171]. Familial forms of AD,
Institute of Neuroscience, Catholic University termed early onset familial AD (EOFAD) with
of Louvain, Brussels, Belgium
e-mail: ilse.dewachter@uhasselt.be clinical mutations in APP or PS1/2, have an early

© Springer Nature Singapore Pte Ltd. 2019 145

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_13
146 I.-C. Stancu et al.

onset, and are invariably characterized by the ogy is a highly compelling explanation for the
combined presence of amyloid and Tau pathol- progressive and characteristic development of
ogy [24, 80, 170]. These EOFAD cases, identify Tau pathology, remarkably strongly correlated
a causal link between APP/PS1 misprocessing with symptom progression in AD.  Importantly,
and the development of Tau pathology and neuro- the presence of Tau seeds has been demonstrated
degeneration [80, 170]. Furthermore, combined not only in brains of Tau transgenic mice, but also
genetic, pathological, biomarker and in  vivo in brains of patients using a sensitive Tau seeding
modelling data, indicate that amyloid pathology detector assay [55, 88]. Hence, Tau misfolding
precedes Tau pathology, and support a role for Aβ can initiate a self-propagating process, spreading
as initiator and Tau as executor in the pathoge- along functionally connected brain regions.
netic process of AD [80, 96, 97]. Hence, AD is Based on the above arguments for an execu-
often considered as a secondary Tauopathy (simi- tive role of Tau, Tau represents a key therapeutic
lar as for Down syndrome patients), in contrast to target for primary as well as for secondary
the primary Tauopathies described above. Tau Tauopathies (including Alzheimer’s disease,
aggregates in Tauopathies vary with respect to Down’s syndrome, dementia pugilistica) [75,
the ratio of different Tau isoforms (3R/4R), to the 111]. A detailed understanding of the role of Tau
cell types displaying Tau aggregation and the in health and disease hence is crucial to identify
structure of the aggregates. However, in all therapeutic targets for Tauopathies, which can
Tauopathies a strong correlation between pro- inhibit its pathological role leaving its physiolog-
gressive development of pathological Tau accu- ical role (relatively) intact [75]. Importantly, the
mulation and the loss of the respective brain exact mechanism as to how pathological changes
functions is observed. in Tau induce neuronal dysfunction and neurode-
Although the presence of Tau aggregates in generation are still not understood and are inten-
Tauopathies suggests its involvement in the sively studied. In general, three different
pathogenetic process, an executive role for Tau in mechanisms can be considered: Tau can induce
Tauopathies, including AD is substantiated by neuronal dysfunction and neurodegeneration
several arguments listed here. These arguments through: (i) a loss of its physiological function,
include: (i) the existence of a family of neurode- (ii) a toxic gain of function of the altered soluble
generative disorders all characterized by Tau forms of Tau (ranging from monomers to oligo-
aggregation [38, 176]; (ii) the strong correlation mers) or (iii) a toxicity of the larger insoluble Tau
of the presence and progression of Tau-pathology aggregates [19, 122]. While the above hypotheses
with disease symptoms in Tauopathies [59, 97, are intensively studied, for either hypothesis it is
171]; and most importantly, (iii) the identification important to understand the role of physiological
of MAPT mutations autosomal dominantly and pathological Tau forms. One way to pursue
linked to Tauopathies, which indicates that Tau this is to identify the interacting proteins of phys-
dysfunction is sufficient to drive or cause neuro- iological and pathological Tau forms to gain
degeneration (https://www.alzforum.org/muta- insight in the mechanisms involved.
tions/mapt). And interestingly, recent findings We have recently performed a genome wide
have indicated that (iv) Tau pathology can propa- unbiased identification of the Tau interactome.
gate in a prion-like way from one cell to another, Hereto, we performed isobaric tags for relative
and can propagate to functionally connected and absolute quantitation (iTRAQ)-based Tau
brain regions [52, 59, 66, 102, 143]. This prion- interactome mapping to gain unbiased insight
like or templated propagation of Tau pathology into Tau pathophysiology and to identify novel
has been consistently and reproducibly shown in Tau-directed therapeutic targets. This yielded
in  vitro and in  vivo models, proving the self- some interesting starting points for further analy-
propagating nature of Tau-misfolding and Tau- sis of the physiological and pathological role of
pathology once initiated [6, 26, 52–57, 59, 66, Tau. Within this Tau interactome mapping we
72–74, 87, 101, 102, 110, 119, 147, 167, 180, identified proteins that can be subcategorized in
190, 191]. Prion-like propagation of Tau pathol- different functional groups, including ­cytoskeletal
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 147

proteins, proteins involved in post-translational 6 isoforms differing by the presence of 0,1 or

modifications of Tau, synaptic proteins, signaling 2  N-terminal inserts and the presence of 3 or 4
molecules [193]. Similar ubiased screening microtubule binding domains (MTBD) [61–64,
approaches have been performed by other groups 75, 135]. Differing roles of different Tau isoforms
aiming at the identification of Tau interacting pro- are suggested by the predominant expression of
teins (directly and indirectly) [27, 71, 130, 193]. the shortest isoform (0N3R) during development
In addition unbiased screening approaches of Tau shifting to nearly equal 3R/4R ratios and by dif-
modifiers in small animal models encompassing ferences in their regional expression in the adult
C. elegans and Drosophila have been performed human brain [60]. And differing proteins have
and provided an interesting starting point for fur- been shown to interact with Tau in interactome
ther analysis [42, 116]. In this chapter we will mapping approaches [71, 130]. While the differ-
review the Tau interactome as known from ing physiological roles of these different isoforms
hypothesis based studies and from the identifica- remain less clear, changes in the 3R/4R ratio of
tion of the Tau interactome by broad unbiased aggregated Tau are observed in Tauopathies [8,
experimental approaches and unbiased functional 75, 121, 176], even to the extent that Tauopathies
screens as performed by us and others, comple- display a fixed ratio of 3R/4R Tau characteristic
mented by hypothesis based studies. Combined for the specific Tauopathy, highlighting the impor-
these approaches have yielded important insights tance of a correct balance of these isoforms in the
in Tau’s physiological and pathological role. This brain to maintain healthy conditions.
chapter does not attempt to be exhaustive or to Tau protein is a natively unfolded protein with a
cover all important Tau-protein interactions. We remarkably high hydrophylic profile, and hence
merely aim but to give an overall picture of Tau water solubility [99, 135, 192]. Tau (4R/2N) pre-
and its interactors and the potential of the combi- dominantly consists of hydrophilic amino acids,
nation of hypothesis based and unbiased screen- containing 80 Ser or Thr residues, 56 negative (Asp
ing approaches and functional screens to uncover or Glu) residues, 58 positive (Lys or Arg) residues
novel as yet unexplored functions of Tau, or novel and 8 aromatic (5 Tyr and 3 Phe, but no Trp) resi-
interacting proteins of Tau. Excellent reviews dues. Although Tau is predominantly a basic protein,
about Tau in health and disease have been pub- it can be subdivided in 4 subdomains, based on their
lished (e.g. [8, 20, 75, 121, 135, 176, 192]). charge and respective motifs [75, 135, 192]. These
domains are crucial determinants in the interaction
with other proteins and encompass, the N-terminal
 he Structure of Protein Tau
T domain (AA 1–150), a Proline-rich domain
as a Target for Protein-Protein (AA151–243), the microtubule binding domain
Interactions consisting of 3 or 4 imperfectly repeated motifs and
a C-terminal tail (AA370–441). The N-terminal
Interactions between proteins are driven and regu- region is predominantly acidic, while the C-terminus
lated by specific AA sequences or domains within is neutral. This charge asymmetry and different
the protein, their post-translational modifica- ‘domain characteristics’ enable specific protein-pro-
tions  – including phosphorylation -, the protein tein interactions with other proteins but also intra-
isoform, the environmental conditions and the molecular folding of Tau and Tau aggregation (Fig.
cellular and subcellular localisation of the protein. 13.1). Furthermore, the specific sequences in the
Tau protein is predominantly expressed in neu- microtubule binding domains (MTBD) are respon-
rons, in axons, while it has also been detected in sible for interaction with specific sets of proteins.
lower concentrations in dendrites, in the nucleus The microtubule binding domain of Tau (4R/2N)
and in synapses (pre- and postsynaptically) [155]. contains 4 imperfectly repeated regions (R1 to R4,
Tau is encoded by the MAPT gene, located on encoded by exons 9–12), separated by 13 to 14
chromosome 17, comprising 16 exons [148]. amino acid long spacer regions [67, 120]. The
Alternative splicing of exon 2, 3 and 10, generates repeats consist of highly conserved 18 amino acids
148 I.-C. Stancu et al.

Fig. 13.1  Tau interacting proteins

stretches, which include KXGS motifs, a motif  au:Tau Binding and Tau

T
also conserved in other microtubule associated Aggregation
­proteins. The interaction between Tau and microtu-
bules heavily depends on positively charged Lys As discussed above, Tau aggregation is a core
residues on Tau within the microtubule binding pathological feature of Tauopathies [15] and
domains and the negative charge of microtubules considered important for their pathogenetic pro-
[23, 48, 103–105, 145]. Furthermore, the microtu- cess. Tau can aggregate into paired helical fila-
bule binding region is important in Tau aggregation. ments or straight filaments [15, 61]. Interestingly,
While Tau is a natively unfolded protein, the second the domains in Tau involved in Tau-Tau and Tau-
and third MTBDs exhibit a high propensity to form microtubule interaction seem to overlap [15, 23,
an ordered β-sheet structure, and are important in 48, 103, 105, 145]. Binding of Tau to microtu-
Tau aggregation [10, 31, 33, 48, 49, 144, 145]. The bules thereby prevents Tau from self-aggrega-
Proline-rich domain  (PRD) contains seven Pro-X- tion, or conversely detachment of Tau from
X-Pro (PXXP) motifs, rendering this part involved microtubules renders Tau and its MTB domain
in signaling pathways, and a target for Proline- available for Tau aggregation. The repeat
directed kinases, as further outlined below. Although domains of Tau form the core of PHFs, while the
Tau is an intrinsically disordered protein, interaction C-terminal and the N-terminal regions form the
between the different domains may result in dynamic fuzzy coat. The loss of interaction of the MTBD
folding of Tau. In this respect, a paperclip structure with microtubules, thereby allows Tau to acquire
has been proposed for Tau [98, 99]. In this paperclip highly ordered β-sheet structures, facilitating
structure, the C terminus folds over the microtubule Tau aggregation  (Fig. 13.1). The hexapeptide
binding domain and the N terminus folds back over motifs of Tau, VQIINK (275–280) and VQIVYK
the C terminus, bringing both termini in close prox- (306–311) in R2 and R3 of the MTBD, are likely
imity [98, 99]. This folding of Tau is modulated by to be important for Tau aggregation due to their
binding to interacting proteins, such as microtubules strong β-sheet forming properties, which has
and many others, and by post-translational modifica- been confirmed in in  vitro and in  vivo studies.
tions. Tau interactions with other proteins are mostly Indeed, disruption of these motifs (e.g., by Pro
described with reference to these domains. mutations) abrogates Tau’s tendency to aggregate,
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 149

not only in  vitro, but also in cell and animal for the polarity of the microtubules with a respec-
models [112, 140]. Conversely, strengthening tive plus and minus end [17].
the β-sheet-propensity by mutations (e.g., As described above, Tau contains either 3 or 4
ΔK280 or P301L) accelerates aggregation in MTBDs, depending on the isoform (3R/4R). The
in vitro and in animal models. However, recently, MTBD, containing the KXGS motif, which is con-
cryo-EM has identified the core of Tau filaments served among microtubule associated proteins, and
of AD brains. It was found that the core does not its flanking basic Pro-rich region, have been shown
consist of the complete MBD but rather, only important for microtubule binding [23, 48, 103,
amino acid residues 306–378 (i.e., spanning R3, 105, 145]. The exact identification of the interacting
R4, and a part of C-terminal region containing a sequences in the binding between Tau and tubulin
number of Lys residues), and the remaining por- has been hampered by the intrinsically disordered
tion functions as the fuzzy coat [49]. How nature of Tau and the dynamic nature of the interac-
β-sheet propensity of R2 domain, and how other tion between Tau and tubulin. A variety of models
domains of Tau affect its aggregation to form have been proposed, including (i) binding of Tau to
this core, as exemplified in the studies listed the outer surface of microtubules connecting tubu-
above, remains to be further analyzed. The struc- lin subunits either across or along protofilaments [1]
ture of 3R Tau only ­aggregates, as observed in as well as (ii) binding of Tau to the interior of the
Pick’s disease was found to be different from microtubule wall [107]. Using a combination of
Tau aggregates isolated from AD brains. The NMR spectroscopy and mass spectrometry short
part of Tau not belonging to the core of Tau fila- sequences within the microtubule binding domain
ments forms the fuzzy coat, which is highly were demonstrated to be important for the interac-
mobile and can even in its aggregated state inter- tion between Tau and tubulin. These sequences
act with other proteins. included the imperfect repeat domains 224–237,
245–253, 275–284 and 300–317, while the flanking
sequences remain flexible [103, 105, 144, 145].
 au Interactions with Cytoskeletal
T These regions of Tau were found to bind to the
Proteins interface between α−  and β-tubulin heterodimers.
Combined NMR methods optimized for ligand-
Tau and Tubulins: The Microtubule receptor interactions, demonstrated the competitive
Associated Protein Tau binding of the flanking domain downstream of the
four microtubule binding repeats of Tau to a site on
Even before its identification as major component the α-tubulin surface [103, 105, 144, 145].
of NFTs, Tau was identified as a microtubule asso- Binding of Tau to microtubules, promotes
ciated protein [194]. Dynamic regulation of micro- self-assembly of microtubules and their stabiliza-
tubules is essential for cellular structure and shape, tion [14]. Also more subtle effects on microtubule
for intracellular organization, organelle trafficking dynamics have been reported, with overstabiliza-
and chromosome segregation [17]. In adult brain, tion of microtubules by Tau decreasing dynamics
Tau is predominantly localized axonally in neu- and being detrimental for cell viability in  vitro
rons, where it binds to microtubules, which exert [90]. Interestingly, the domains in Tau involved in
crucial functions in neuronal and axonal structure microtubule stabilization are also involved in Tau
as well as axonal transport. Microtubules are tubu- aggregation (Fig. 13.1), which may suggest a loss
lar polymers composed of α- and β-tubulin sub- of physiological function following aggregation
units assembled into linear protofilaments. of Tau [103, 105, 144, 145]. In this respect it is
Microtubules consist of 10 to 15 protofilaments, interesting to note, that many modifications asso-
generated by end-to-end polymerization of α/β ciated with Tauopathies (phosphorylation, clini-
tubulin dimers, associated laterally to form one cal mutations, altered 3R/4R ratio) alter the
hollow cylinder or microtubule, which can further ability of Tau to bind to microtubules [76, 77,
be extended by addition of additional α/β tubulin 136]. While the N-terminal projection domain
dimers. This specific organization also accounts does not bind to microtubules, its long acidic
150 I.-C. Stancu et al.

domain provides repulsive forces allowing it to other proteins for axonal transport. Hence differ-
exert a spacer function. Similar spacer functions ent mechanisms have been shown to implicate
have been noted for domains of other microtu- Tau in dysregulation of axonal transport in vitro.
bule associated proteins of the MAP family However, also here in vivo compensating mecha-
(MAPs: MAP2, MAP2c, ….) [76, 77, 136]. nisms may exist, as Tau deletion did not affect
These roles for Tau have been demonstrated axonal transport in vivo [196, 197].
unequivocally in cell free systems and in  vitro Interestingly, using an unbiased approach to
[76, 77, 136]. Subsequent antisense experiments identify Tau interacting proteins (iTRAQ)  also
and experiments with primary cultures from Tau identified these well characterized Tau interacting
knockout mice, have indicated a role for Tau in proteins within the Tau interactome [193]. Within
axonal outgrowth, consistent with a role of Tau in the Tau interactome mapping, α-tubulin and
microtubule stabilization and dynamics [35, 47, β-tubulin subunits were identified, as well as pro-
135, 154, 192]. However, while these roles for teins from the dynactin complex [130, 193]. This
Tau have been identified in in vitro experiments validates the power of these unbiased screening
(albeit with some controversy), it must be noted approaches to identify previously characterized
Tau knockout mice display only subtle pheno- and novel Tau interacting proteins. Indirect inter-
typic alterations and are surprisingly healthy [93, actors, i.e. proteins bound in a protein complex,
152, 183]. This indicates that Tau may regulate hence in an indirect way, are also identified using
these functions but is not essential, although more these approaches. Dynein and kinesin have been
detailed analysis is still ongoing and will yield picked up as Tau interacting proteins, probably due
more detailed insights. Importantly and most par- to their interaction with tubulins. Hence, this fur-
simoniously, redundancy between different ther validates unbiased interactome mapping as an
MAPs may account for the absence of gross interesting experimental approach. Similarly, sev-
developmental and structural abnormalities in eral of these proteins have also been identified in a
Tau deficient mice. These findings should be functional unbiased screening for Tau-modifiers.
taken into account when considering loss of func- Several unbiased screening approaches have been
tion hypothesis in the context of the pathogenetic performed in Drosophila and C. elegans [3, 22,
process for Tauopathies. Recently, it has been 79]. Within these screens proteins involved in axo-
reported that acute knockdown of Tau in the hip- nal transport and cytoskeletal dynamics have been
pocampus of adult mice, and thus bypassing identified as modifiers of Tau-induced degenera-
developmental compensation, causes learning tive phenotypes. This highlights the interest of
and memory deficits, which are alleviated when these complementary approaches: proteomics
knockdown is lifted [188]. Binding of Tau to based identification of Tau interacting proteins,
microtubules is not only important for regulation unbiased genetic screening for phenotypic modu-
of microtubule stability and dynamics, but also lators of Tau-induced phenotypes, and hypothesis
indirectly affects microtubule dependent func- based identification of Tau interacting proteins.
tions, including axonal transport and cellular
localization [43, 75, 135, 178, 192]. By binding
to microtubules Tau interferes with the roles of Tau and Actin Cytoskeleton
kinesin and dynein motors involved in antero-
and retrograde axonal transport. When encoun- Besides the microtubules, actin is an important
tering patches of Tau, dynein tended to reverse cytoskeletal protein, involved in modulation of
direction, whereas kinesin to detach at patches of cell shape in response to signals, including axo-
bound Tau [40]. Interestingly, Tau also binds to nal outgrowth [91, 153, 157]. In fact, the coordi-
the p150 subunit of dynactin, a multisubunit pro- nated action of microtubules and actin, is crucial
tein complex, which binds directly to dynein, and to respond in an adequate way to the cellular
supports dynein dependent axonal transport [133]. environment [41]. Microtubule-associated pro-
Furthermore, Tau as a cargo itself competes with teins regulate microtubule dynamics, but can also
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 151

bundle and cross-link actin filaments hence rep- brains [54, 56]. Interestingly, starting from an
resenting crucially important proteins for cellular unbiased functional screening to rescue Tau tox-
responses [23, 69]. Actin filaments are the small- icity in Drosophila, several proteins interacting
est type of filaments, consisting of actin polymers with the actin and tubulin cytoskeleton were
organized as a long spiral chain of about 6  nm identified [11]. In a different study, it was shown
diameter. Like microtubules, actin filaments have in a Drosophila Tauopathy model that abnormal
plus and minus ends, with ATP-dependent growth bundling and accumulation of F-actin mediated
occurring at the positive end predominantly [91, Tau-induced neuronal degeneration, indicating a
153, 157]. Actin filaments are often closely situ- pathological role for the interaction between Tau
ated to the membrane, and involved in ­modulating and actin [54, 56]. Together these data support a
or maintaining particular cell shapes [91, 153, role of the Tau-actin and even Tau-actin-
157]. More importantly, actin dynamics are microtubule interaction in health and disease
important in synaptic contacts and their matura- conditions [23, 41, 45]. Moreover, within the
tion, and consequently important for neuronal Tau interactome mapping actin and actin cyto-
function [29]. skeleton associated proteins have been identified
While the interaction of Tau and microtubules as Tau interacting proteins [71, 130, 193], fur-
has been intensively studied leading to clearcut ther underscoring this interaction and reassuring
insights, the interaction between Tau and actin the relevance of unbiased screens to uncover Tau
has received less attention. Tau has been found protein interactors.
to directly interact with actin. Both sequences
within the microtubule binding domain and the
Pro-rich domain were identified to interact with  au Interactions Inducing Post-
T
actin directly, thereby regulating dynamic actin Translational Modifications
polymerization and stability [83, 195]. The
important role of actin dynamics at the synapse Tau Phosphorylation
and the role of Tau in actin dynamics, suggests a
role of Tau in synaptic function and remodeling. Post-translational modifications result from
The interaction Tau-actin, could thereby play a interaction of Tau with other proteins, and subse-
role in dendritic spine morphology and postsyn- quently affect the interaction of Tau with still
aptic reorganizations [92]. A recent study dem- other proteins. In NFTs, Tau is invariably hyper-
onstrated that Tau uses several short helical phosphorylated [15, 16, 70], rendering post-
segments to bind dynamically to the hydropho- translational modifications of Tau, and
bic pocket between subdomains 1 and 3 of actin particularly its hyperphosphorylation an inten-
[23]. While a single Tau segment is sufficient for sively studied topic. Interaction of Tau with
binding, minimally two helical segments are kinases and phosphatases alters its phosphoryla-
required for actin bundling [23]. Phosphorylation tion status, altering its interaction with other pro-
of Tau at Ser262 attenuates binding of Tau to teins and its aggregation. Protein Tau (4R/2N)
filamentous actin, in line with a structural model contains approximately 80 Ser/Thr residues and
of Tau repeat sequences in complex with actin 5 Tyr residues, providing 85 potential phosphory-
filaments. The interaction of Tau with actin fila- lation sites. Approximately 45 of these residues
ments, may play a role in Tau–induced dysregu- haven been experimentally observed to be phos-
lation of neuronal function and synaptic phorylated [78]. Within the Pro-rich domain of
dysfunction, in view of the crucial role of actin Tau many of these Ser and Thr sites are followed
filament dynamics in synaptic remodeling. by a Pro residue, and can hence be phosphory-
Furthermore, Hirano bodies, which are actin- lated by Pro-directed Ser/Thr protein kinases.
rich inclusions, containing different proteins, Many different kinases have been demonstrated
including Tau, have been identified within AD in vitro to phosphorylate Tau, including GSK-3β
152 I.-C. Stancu et al.

and CDK5. There is substantial support for the accounts for ~70% of the human brain Tau phos-
Pro-directed kinases CDK5 and GSK-3 being phatase activity [75, 78, 129]. In vitro and in vivo
relevant kinases in Tauopathies. In vivo analysis studies in transgenic mice, further have demon-
in animal models with Tau pathology, demon- strated increased Tau phosphorylation and somato-
strated respectively aggravated and rescued Tau dendritic translocation of Tau following PP2A
pathology following their increased and inhibition [75, 78, 114, 131]
decreased activation respectively, following Tau phosphorylation is very strongly regu-
genetic modification or drug based targeting [16, lated during development, with fetal Tau display-
18, 46, 85, 124, 125, 132, 150, 151, 177, 184]. In ing high levels of phosphorylation [13, 65, 106].
addition, in brains of AD patients, dysregulation Tau phosphorylation modulates also cellular
of these kinases has been demonstrated [124, localization of Tau within the axon (proximal or
135]. Most of the phosphorylation sites cluster in distal), its somatodendritic localization (although
the flanking regions of the MTBD – the Pro-rich this is not the sole mechanism [115], its binding
domain and the C-terminal domain  – but not to microtubules, its binding to the plasma mem-
exclusively  (Fig. 13.1). Tau can be phosphory- brane and its interaction with other proteins.
lated by Pro-directed Ser/Thr-protein kinases, Increased somatodentritic localization and aggre-
non-Pro-directed Ser/Thr-protein kinases and gation of Tau has also been detected following
protein kinases specific for Tyr residues. increased activation of Tau-kinases in in  vivo
In addition, some kinases phosphorylate Tau models. Aggregated Tau accumulating in
in or near the repeat domain, these kinases Tauopathies is invariably hyperphosphorylated,
include MARK (microtubule affinity-regulating and Tau phosphorylation increases Tau aggrega-
kinase), CamKII (calmodulin dependent kinase tion  [2, 39, 186], but other post-translational
II) and PKA (protein kinase A or cAMP depen- modifications may also play a role both pro and
dent protein kinase) [78]. Particularly phosphor- contra aggregation (see below). Tau phosphoryla-
ylation of Ser262 of the KXGS motif within the tion is hence not only considered important for
MTBD, has been shown to regulate its binding to regulating its physiological role, but also in the
microtubules, and to be phosphorylated by pathogenetic process leading to Tau aggregation
microtubule affinity-regulating kinases (MARKs, [75, 78, 135, 192].
also known as PAR1 kinases). Tau can become An unbiased approach to identify Tau kinases
phosphorylated at Tyr18, Tyr29, Tyr197, Tyr310, that phosphorylate specific Tau phosphorylation
and Tyr394, with Tyr18 and Tyr394 residues of sites has uncovered novel Tau kinases and novel
Tau been detected in NFTs in AD patient brains. pathways that affect Tau phosphorylation [27].
Src family kinases, such as Src, Lck, Syk, Fyn Interestingly, unbiased functional screens in C.
and c-Abl have been shown to phosphorylate Tau elegans have identified several of these kinases
at these Tyr residues. Tyr phosphorylation, has and phosphatases as modulators of Tau toxicity
been shown to occur in brains of Tauopathy [116]. Similarly kinases and phosphatases have
patients and animal models of Tauopathies, and been picked up in interactome mapping
to exert diverse effects in the neurodegenerative approaches of Tau [130, 193]. The process of Tau
process in animal models [75, 78]. phosphorylation and its role in physiological and
Obviously, the converse process -Tau dephos- pathological processes is described in depth in
phorylation- is crucial in the regulation of its physi- Chap. 6. We here point out the interaction of Tau
ological and pathological roles. Protein phosphatase with this important class of Tau interacting and
1/2A/2B/2C and 5 (PP1, PP2A, PP2B, PP2C and Tau modifying proteins. These interactions and
PP5) have been implicated in Tau dephosphoryla- post-translational modifications not only affect
tion. Accumulating evidence supports an important the physiological role of Tau but also affect its
role for PP2A as relevant Tau phosphatase. PP2A aggregation and pathological role.
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 153

Other Post-Translational has been found to be impaired in post-mortem

Modifications: Acetylation, human AD brains [109, 146]. Together these
Glycosylation, Glycation, findings indicate the importance to understand
Ubiquitination, … Tau homeostasis and its regulation by ubiquitina-
tion and the ubiquitin-proteasome system in the
While Tau phosphorylation traditionally has brain [34, 109, 146]. Notably experimental evi-
been the most intensively studied post-transla- dence has been obtained for the notion that larger
tional Tau modification, Tau is subject to sev- Tau aggregates are degraded by the autophagy-
eral other post-translational modifications, lysosomal system, while more soluble, and
including glycosylation, glycation, deamida- smaller Tau forms and Tau oligomers are
tion, isomerization, nitration, methylation, degraded by the UPS [36, 75, 123, 156].
ubiquitylation, sumoylation and truncation [75, Poly-ubiquitination occurs by different ubiq-
135, 156, 192]. All these modifications are the uitin ligases (E3) in combination with a ubiquitin
result of the interaction of Tau with modifying conjugating enzyme (E2), and defines the cellular
enzymes and they alter the interaction of Tau fate of proteins including their degradation [168].
with other Tau interacting proteins, potentially Ubiquitin can be ubiquitinated on different Lys
affecting its role in physiological and patholog- residues, leading to polyubiquitin chains with
ical conditions. Furthermore, these post-trans- complex topologies. For Tau differently linked
lational modification may influence one another, (Lys6, Lys11, Lys48, and Lys63) polyubiquiti-
as Tau phosphorylation may inhibit its subse- nated Tau forms have been identified in AD brains
quent post-translational modification, whereas and models [37, 117]. Several E3 ligases have
acetylation of Tau has been shown to affect its been identified to ubiquitinate Tau, which include
ubiquitiation etc.… . A more extensive discus- C-terminus Hsp70 interacting protein (CHIP),
sion of the post-translational modifications of tumor necrosis factor receptor-associated factor 6
Tau extends beyond the scope of the current (TRAF6), and axotrophin/MARCH7 [7, 50, 159,
chapter (has been provided in [75, 135, 156, 166, 172]. Convincing evidence indicates a mod-
192]). Tau acetylation is discussed in depth in ulatory role for CHIP as Tau modulator and Tau-
Chap. 7 and Tau truncation in Chap. 8. We here pathology modulator in vitro and in vivo. UbcH5/
would like to focus on the process of Tau ubiq- UBE2D acts as E2 enzyme and Hsp70 as a coen-
uitination and deubiquitination. zyme, recognizing the MTBD and PRD of Tau
The link between Tau and ubiquitination in [82]. In line with these findings, deletion of the
AD has been intriguing, ever since the identifica- ubiquitin ligase CHIP leads to the accumulation
tion of ubiquitin (Ub) in the senile plaques of AD of phosphorylated Tau species [166]. Furthermore,
patients and the finding that Tau proteins are the high-affinity HSP90-CHIP complex recog-
highly ubiquitinated in the brain of human AD nizes and selectively targets phosphorylated Tau
patients [30, 142, 156, 158]. Increased levels of proteins for degradation. In line with this, overex-
Ub-Tau were also found in cerebrospinal fluid pression of CHIP attenuates the toxicity of hyper-
(CSF) of AD patients [36, 123]. Accumulation of phosphorylated Tau and maintains neuronal
Ub-Tau suggested inefficient proteasomal degra- survival [82, 172]. In mouse and human brains
dation, as the ubiquitin-proteasome system (UPS) levels of CHIP and Hsp70 inversely correlate
controls degradation of abnormally folded cyto- with the levels of insoluble Tau aggregates,
solic proteins [34]. This was further corroborated strongly suggesting that CHIP-mediated ubiquiti-
by the finding that PHF-Tau isolated from human nation of Tau is an important negative regulator
AD brains co-immunoprecipitates with various of Tauopathies [166]. CHIP, along with its stress-
proteasome subunits, suggesting inefficient pro- activated E2s and Hsp70, has diverse target pro-
cessing by the proteasome [109, 146]. teins, contributing as an important regulator of
Furthermore, proteasome degradative activity the cellular protein folding-refolding machinery
154 I.-C. Stancu et al.

and its degradation machinery. Taken together, ical Tau but also and importantly on the develop-
targeting Tau degradation by modulation of the ment of Tau pathology. These have been described
HSP90-CHIP complex  represents an interesting in detail in Chap. 20.
approach for Tauopathies. We furthermore would like to highlight that
Ubiquitination of proteins is highly dynamic within the Tau interactome mapping also many
due to the balanced action of ubiquitinases and Tau interacting proteins involved in its post-
deubiquitinating enzymes known as deubiquitin- translational modification have been identified
ases (DUBs) [149]. While Tau ubiquitination has [11, 130, 193]. These include proteins with
been studied intensively, its deubiquitination already known effects on Tau, but also novel Tau
remains less well explored. Starting from the Tau modifiers may still be uncovered by analyzing
interactome mapping we have recently identified, the role of novel identified Tau interacting pro-
several potential Tau deubiquitinases, and identi- teins. Similarly, within functional genome wide
fied Otub1 as a novel Tau deubiquitinating unbiased phenotypic screens, enzymes which
­
enzyme, with modifying effects on Tau pathol- post-translationally modify Tau have been identi-
ogy in vivo [193]. Otub1 directly affected Lys48- fied [11, 116].
linked Tau deubiquitination, impairing Tau
degradation, dependent on its catalytically active
cysteine, but independent on its noncanonical Tau Interactions in the Nucleus
action modulated by its N-terminal domain in
primary neurons. Otub1 strongly increased AT8- Besides its localization in axons, dendrites and
positive Tau and oligomeric Tau forms and synapses, Tau has been detected in the nucleus
increased Tau-seeded Tau aggregation in primary of neuronal and non-neuronal cells. In vitro and
neurons. Finally, we demonstrated that expres- in  vivo data in animal models, and data in
sion of Otub1 but not its catalytically inactive humans support a nuclear role of Tau, in pre-
form induced pathological Tau forms in Tau serving DNA integrity and potentially other
transgenic mice in  vivo, including AT8-positive nuclear roles of Tau [137, 173, 181, 189]. Tau
and oligomeric Tau forms [193]. has also been shown to protect cytoplasmic and
Taken together these findings indicate that a nuclear RNA. In a Drosophila model Tau expres-
detailed understanding of the degradation of Tau sion was shown to affect mitosis, mitotic spindle
and different Tau forms by ubiquitination and formation and resulted in aneuploidy in
deubiquitination, yields important insights in the Drosophila and cell culture [134]. In addition,
accumulation and degradation of Tau in health Tau aggregates have been detected in the nucleus
and disease. While the UPS selectively degrades of AD patients and Huntington disease patients,
normal and abnormally folded soluble proteins, whereas oligomers have been shown to affect
which are tagged by ubiquitin for elimination, the the protective function of Tau in the nucleus
autophagic–lysosomal system (ALS) mainly [137, 173, 181, 189]. The nuclear role of Tau
degrades large protein aggregates or inclusions and the binding proteins involved have been
and organelles [36, 123]. The UPS is thereby reviewed in detail [173] and are discussed in
considered interesting for removal of smaller detail in a separate chapter of the book.
soluble oligomeric forms of Tau, often consid- Interactions of Tau in the nucleus can be directly
ered as the toxic Tau forms [123]. The ALS con- with DNA or RNA, but also with DNA/RNA
versely can also be involved in the removal of associated proteins. For instance, an interaction
larger Tau aggregates. This topic has been ele- between Tau and TIA1 has been found [187].
gantly reviewed in [169] and is discussed in TIA1 interacts with proteins linked to RNA
Chap. 20 of this book. Similar as for phosphory- metabolism. Tau appeared to be required for
lation and ubiquitination/deubiquitination, the normal interaction of TIA1 with these proteins,
post-translational modifications listed above have whereas conversely TIA1was shown to induce
profound effects on the availability of physiolog- Tau misfolding. Most recently Tau has been
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 155

found to associate with the RNA binding protein a presynaptic function were identified . These
Nup98 [44]. Nup98 is part of the nuclear pore proteins included, several proteins involved in
complex. Pathological Tau was shown to impair presynaptic vesicle recycling including dyna-
nuclear import and export. Furthermore, Nup98 min, amphiphsyin, clathrin, actin and PP2A,
was shown to accelerate Tau aggregation synaptotagmin, SV2A among others [130, 193].
in  vitro. It is interesting to note that different While these interactions can be indirect they
RNA binding proteins appear to be able to suggest a presynaptic role of Tau in synaptic
induce Tau misfolding and aggregation. vesicle dynamics [130, 193]. Detailed analysis
demonstrated that in pathological conditions,
Tau dissociates from axonal microtubules and
 au Interacts with Synaptic Proteins
T missorts to pre- and postsynaptic terminals
and Affects Synaptic Function [182]. Interestingly, a novel modulatory presyn-
aptic role for Tau was recently demonstrated.
While Tauopathies have been initially charac- Pathogenic Tau was shown to bind to synaptic
terized by large predominantly intracellular vesicles via its N-terminal domain, thereby
aggregates, Tau aggregates range from small interfering with presynaptic functions, includ-
soluble oligomers to large Tau aggregates. ing synaptic vesicle mobility and release rate,
Smaller soluble oligomeric Tau forms are (also) lowering neurotransmission in fly and rat neu-
considered important in the neurotoxic action of rons [138, 198]. Pathological Tau mutants lack-
Tau. The mechanisms by which this toxic fea- ing the vesicle binding domain still localize to
ture is acquired imply most likely the interac- the presynaptic compartment but do not impair
tions of Tau with neuronal proteins, directly or synaptic function in fly neurons. Moreover, an
indirectly linked to synaptic functions. For this exogenously applied membrane-permeable pep-
reason, great effort has been dedicated in trying tide that competes for Tau-vesicle binding sup-
to find Tau  interacting proteins, eventually presses Tau-induced synaptic toxicity in rat
unraveling several binding partners potentially neurons [138, 198]. These data demonstrate a
involved in a complex picture underlying neuro- role for accumulating Tau at the presynaptic ter-
degeneration. Interestingly, in pathological con- minal in dysregulated neuronal functions. In a
ditions Tau is increasingly found in pre- and follow-up paper, the N-terminal domain of Tau
postsynaptic terminals [182]. Hence, important was shown to bind to synaptic vesicles through
focus is on understanding the interacting pro- its interaction with the transmembrane vesicle
teins and the pathophysiological consequences protein synaptogyrin-3 [138]. Synaptogyrin-3
of this interaction. and synaptic vesicle associated proteins previ-
ously had been identified in Tau interactome
mapping indicating direct and indirect interac-
 au Affects Presynaptic Function,
T tions of Tau with synaptic vesicle associated
through Binding with Presynaptic proteins [130, 193]. In fly and mouse derived
Proteins models of Tauopathy, reduction of synaptogy-
rin-3 prevents the association of presynaptic
Interestingly, although Tau has been identified Tau with vesicles, alleviates Tau-induced
predominantly in neuronal axons, it has been defects in vesicle mobility, and restores neu-
also unequivocally identified in synaptic termi- rotransmitter release [138]. Hence Tau binds
nals of human neurons, both pre- and postsynap- through its N-terminal domain to synaptogy-
tically. The role of Tau at the presynaptic site rin-3 associated with synaptic vesicles, altering
has remained relatively long unexplored. Within their mobility. The identification of this interac-
the Tau interactome mapping many different tion may be further exploited for therapeutic
proteins were found with a synaptic function strategies aiming to counteract Tau induced syn-
[28, 193]. More particularly many proteins with aptic dysfunction.
156 I.-C. Stancu et al.

 au Binds to Postsynaptically
T Taken together, both pre- and postsynaptic
Localized Proteins roles (and protein-protein interactions) of Tau,
have been demonstrated and are increasingly
Although Tau was originally described as a neu- investigated to understand their roles in health and
ronal, predominantly axonal protein, its pres- disease in more detail. Broad spectrum, unbiased
ence in glial cells, but also in the neuronal approaches as well as hypothesis based approaches
somatodendritic compartment and at pre- and have in a combined way contributed to these
postynaptic sites has been identified in physio- insights.
logical and pathological conditions [182]. A syn-
aptic role for Tau first became apparent from the
finding that Tau deficiency could rescue Aβ Tau and Proteins Identified in GWAS
induced synaptic defects mediated at the post-
synaptic site [95, 164]. Oligomeric Aβ was Identification of causal mutations for Tauopathies
shown to induce synaptic deficits and excitotox- and related neurodegenerative disorders has pro-
icity [126]. An elegant publication demonstrated vided important insights and cornerstones for
that Aβ dependent excitotoxicity was dependent their analysis. However, the majority of
on Tau expression, but was also mediated by Tauopathies, including AD, are sporadic forms,
Tau-dependent recruitment of the SRC family driven by interaction between risk genes and
tyrosine kinase Fyn to postsynaptic NMDAR environmental factors. Hence there is an impor-
complexes [95, 164]. Tau was shown to interact tant interest in understanding the contribution of
with Fyn through the PXXP motifs in its Pro- risk factors and risk genes in the pathogenetic
rich domain. At the postsynapse, the Tau/Fyn process, which led to the identification of risk
complex interacts with the PDZ (postsynaptic genes for AD. The ApoE4 allele has been identi-
density, discs-large, zona occludens protein-1) fied as important risk gene for AD, and subse-
domain of the protein PSD-95 [95, 140], a key quent GWAS studies have identified additional
scaffolding protein for postsynaptic receptors AD risk genes [81, 86, 108, 118]. Genome wide
[113]. While Tau is essential for recruiting Fyn associated studies (GWAS) have identified poly-
to the PSD-95 complex, it is the localized activ- morphisms in or near several genes that are asso-
ity of Fyn that is critical for NMDAR-mediated ciated with AD risk: ABCA7, CLU,CR1, CD33,
excitotoxicity [95, 139]. Finally, Fyn can phos- CD2AP, EPHA1, BIN1, PICALM, MS4A.
phorylate Tau at tyrosine-18. This phosphoryla- Additional loci were identified in a meta-analy-
tion of Tau is important for the interaction of Tau sis of these large LOAD consortium datasets:
with Fyn [9], adding another means for regulat- CASS4, CELF1, DSG2, FERMT2, HLA DRB5
ing the formation of PSD-95/Tau/Fyn com- DBR1, INPP5D, MEF2C, NME8, PTK2B,
plexes (Fig. 13.1). Taken together these findings SLC24H4 RIN3, SORL1, ZCWPW1 [81, 86,
indicate a role for Tau in mediating Aβ toxicity 108, 118]. Two genes with moderate to large
at the postsynaptic level. effects on LOAD risk: PLD3 and TREM2 [32,
In addition, Tau may exert a physiological 100] have been identified in smaller datasets.
role at the postsynaptic site. In line with such a The identification of common variants with
role, induction of LTP induces increased levels diverging effect sizes on AD risk has begun to
of postsynaptic Tau [51]. Furthermore, activa- create a broader picture of the processes and
tion of synaptic glutamate receptors induced pathways involved in AD risk. Variants in genes
translocation of Tau from dendritic shafts into involved in lipid metabolism, the inflammatory
postsynaptic densities, together with Fyn kinase response, and endocytosis have been identified
[165]. In addition, Tau has been shown to be through these GWAS [81, 86, 108, 118]. Several
released by neurons, following stimulation of of the gene products of genes picked up in
neuronal activity, in line with a regulation of Tau GWAS may belong to the Tau interactome.
by synaptic activity and a role for Tau in its In this respect an interaction between Bin1
modulation [161]. and Tau has been identified [28]. GWAS have
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 157

identified a region upstream of the BIN1 gene genes for AD. This interesting approach identified
as the most important genetic susceptibility several BIN1 and PTKB2 as Tau modifiers [42].
locus in Alzheimer’s disease (AD) after
APOE. BIN1 belongs to the family of proteins
capable of influencing endocytotic processes  au and Pathologically Aggregating
T
by influencing membrane curvature and recruit- Proteins in Neurodegenerative
ing dynamin, modulating membrane trafficking Disorders
and actin polymerization: these proteins are
BIN1, amphiphysin and RUS167, also called Brains of AD patients are invariably character-
“the BARs” [25, 28]. BIN1 transcript levels ized by the presence of amyloid plaques and neu-
were shown to be increased in AD brains and a rofibrillary tangles, composed of aggregated
novel 3 bp insertion allele ∼28  kb upstream of amyloid peptides and aggregated hyperphosphor-
BIN1 was identified, which increased: (i) tran- ylated Tau, respectively. Furthermore, biomarker
scriptional activity in  vitro, (ii) BIN1 expres- analysis indicated that changes in amyloid
sion levels in human brain and (iii) AD risk in pathology precede Tau pathology, providing a
three independent case-control cohorts [28]. therapeutic window [96, 97]. Amyloid positive
Furthermore, decreased expression of Amph PET scans provide a high risk for conversion to
(BIN1 ortholog in Drosophila) suppressed Tau- AD in the following decade [96, 97]. Most impor-
mediated ­neurotoxicity. Accordingly, Tau and tantly, early onset familial Alzheimer’s disease
BIN1 colocalized and interacted in human neu- patients, carrying EOFAD mutations in APP,
roblastoma cells and in mouse brain [28]. which cause Alzheimer’s disease, invariably
Interestingly, BIN1 variants correlated with develop Tau pathology [80]. This proofs a link
Tau pathology in AD brains, but not with amy- between altered APP processing giving rise to
loid pathology [28]. Subsequent studies dem- increased amyloidogenic processing and the
onstrated that (i) BIN1 and Tau bind through an development of Tau pathology. This finding is
SH3-PRD interaction and (ii) the interaction is further strengthened by similar findings in
downregulated by phosphorylation of Tau in patients with presenilin mutations, APP duplica-
primary neurons [174]. tion, and Down syndrome [80].
Similarly, clathrin adaptor CALM/PICALM, Importantly, studies in animal models have
which has been identified as an AD risk locus, invariably proven a link between amyloid and Tau
has been shown to be associated with Tau inclu- pathology [179]. This includes the first in  vivo
sions in AD, PSP and Pick disease [4, 5]. studies highlighting aggravated Tau pathology in
PICALM is a key component of clathrin-medi- the presence of amyloid pathology, in (multiple)
ated endocytosis, and levels of PICALM were transgenic mice [179]. This included the study by
shown to correlate with levels of phospho-Tau Jada Lewis and colleagues, by generating crosses
and autophagy-related proteins [4, 5], suggesting of mutant Tau and APP transgenic mice, demon-
a potential role as Tau modulator. strating for the first time aggravation of Tau
Along the same line, PTKB2 encodes Pyk2, a pathology in this model compared to the parental
tyrosine kinase, which has been shown to act as a strain [127]. This was further supported by data of
direct Tau kinase [128] and as a suppressor of Tau Jurgen Gotz and colleagues, demonstrating that
toxicity in the Drosophila eye assay [42] and injection of aggregated Aβ induced Tau aggrega-
improves a murine AD model [58]. However, tion in Tau transgenic mice [68]. Follow-up
Pyk2 may also affect APP processing [160]. studies demonstrated that injection of brain
The relative role of PTKB2  in AD related pro- extracts of mice with amyloid pathology in mice
cesses involving APP or Tau has been discussed expressing wild type Tau, could induce Tau
in Polis and Henn [160]. pathology in mice which normally do not develop
Interestingly, an unbiased modifier screen in NFTs [12]. The group of Hyman very elegantly
Drosophila has been used to analyse the Tau mod- demonstrated that amyloid pathology could facil-
ifying potential of several GWAS identified risk itate spreading of Tau pathology in a model with
158 I.-C. Stancu et al.

entorhinal expression of Tau when crossed to a of studies has already yielded important insights
model with amyloid pathology [162]. These data in the role of Tau and is anticipated to yield novel
were further strengthened by studies performed and more detailed insights into the role of Tau in
by different groups demonstrating clearcut aggra- health and disease. Importantly and stressed here
vation of Tau pathology in the presence of accu- again, comparing the outcome of different types
mulated Aβ (reviewed in [179]). Most recently of studies is more than the sum of parts. It yields
these studies were elegantly complemented by a novel insights into the physiological role of Tau
new milestone paper demonstrating that amyloid and the complex mechanisms that are at the basis
pathology facilitated Tau seeded aggregation of of Tauopathies, and in turn inspires new hypoth-
endogenous Tau [84]. Taken together, these stud- eses for further investigation (see Fig. 13.1).
ies indicate a link between amyloid and Tau
pathology, which needs to be understood in detail,
to understand the pathogenetic process of AD References
[179].
The underlying mechanisms linking amyloid 1. Al-Bassam J, Ozer RS, Safer D, Halpain S, Milligan
RA.  MAP2 and tau bind longitudinally along the
and Tau pathology are being studied in detail. outer ridges of microtubule protofilaments. J Cell
Several processes have been demonstrated which Biol. 2002;157:1187–96.
could be involved in the link between Aβ and Tau 2. Alonso A, Zaidi T, Novak M, Grundke-Iqbal I,
pathology in AD. A detailed discussion extends Iqbal K.  Hyperphosphorylation induces self-
assembly of tau into tangles of paired helical fila-
beyond the scope of this chapter. We have previ- ments/straight filaments. Proc Natl Acad Sci USA.
ously reviewed this topic in a detailed way [179], 2001;98:6923–8.
and a detailed discussion of the link between Aβ 3. Ambegaokar SS, Jackson GR.  Functional genomic
and Tau pathology is also described in Chap. 20 screen and network analysis reveal novel modifiers
of tauopathy dissociated from tau phosphorylation.
of this book on Tau biology. While this interac- Hum Mol Genet. 2011;20:4947–77.
tion may or may not involve a direct interaction 4. Ando K, Brion JP, Stygelbout V, Suain V, Authelet M,
between Tau and Aβ, it is crucial for the pathoge- Dedecker R, Chanut A, Lacor P, Lavaur J, Sazdovitch
netic process of AD. V, Rogaeva E, Potier MC, Duyckaerts C.  Clathrin
adaptor CALM/PICALM is associated with neurofi-
Similarly, pathological interactions between brillary tangles and is cleaved in Alzheimer's brains.
α-synuclein and Tau, and huntingtin and Tau are Acta Neuropathol. 2013;125:861–78.
described in separate chapters in this book. 5. Ando K, Tomimura K, Sazdovitch V, Suain V, Yilmaz
Z, Authelet M, Ndjim M, Vergara C, Belkouch
M, Potier MC, Duyckaerts C, Brion JP.  Level of
PICALM, a key component of clathrin-mediated
Conclusion endocytosis, is correlated with levels of phosphotau
and autophagy-related proteins and is associated
The original role ascribed to Tau was stabiliza- with tau inclusions in AD, PSP and Pick disease.
Neurobiol Dis. 2016;94:32–43.
tion of microtubules regulated by phosphoryla- 6. Audouard E, Houben S, Masaracchia C, Yilmaz Z,
tion and dephosphorylation of Tau. This view has Suain V, Authelet M, De Decker R, Buee L, Boom
been drastically changed in the past decade A, Leroy K, Ando K, Brion JP.  High-molecular-
through hypothesis based studies on interactions weight paired helical filaments from Alzheimer
brain induces seeding of wild-type mouse tau into an
of Tau with itself and other proteins, proteomics Argyrophilic 4R tau pathology in vivo. Am J Pathol.
analysis in different physiological and pathologi- 2016;186:2709–22.
cal relevant conditions and cell types, functional 7. Babu JR, Geetha T, Wooten MW. Sequestosome 1/
screens for identification of Tau modifiers (which p62 shuttles polyubiquitinated tau for proteasomal
degradation. J Neurochem. 2005;94:192–203.
may directly or indirectly interact with Tau). 8. Ballatore C, Lee VM, Trojanowski JQ. Tau-mediated
Interestingly, studies of the latter type have been neurodegeneration in Alzheimer’s disease and
performed in C. elegans and Drosophila and related disorders. Nat Rev Neurosci. 2007;8:663–72.
related small animal model organisms. 9. Bhaskar K, Yen SH, Lee G. Disease-related modifi-
cations in tau affect the interaction between Fyn and
Combination of results from these different types Tau. J Biol Chem. 2005;280:35119–25.
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 159

10. Bibow S, Mukrasch MD, Chinnathambi S, Biernat 23. Cabrales Fontela Y, Kadavath H, Biernat J, Riedel
J, Griesinger C, Mandelkow E, Zweckstetter M. The D, Mandelkow E, Zweckstetter M.  Multivalent
dynamic structure of filamentous tau. Angew Chem cross-linking of actin filaments and microtubules
Int Ed Engl. 2011;50:11520–4. through the microtubule-associated protein tau. Nat
11. Blard O, Feuillette S, Bou J, Chaumette B, Frebourg Commun. 2017;8:1981.
T, Campion D, Lecourtois M.  Cytoskeleton pro- 24. Cacace R, Sleegers K, Van Broeckhoven C. Molecular
teins are modulators of mutant tau-induced neu- genetics of early-onset Alzheimer’s disease revis-
rodegeneration in Drosophila. Hum Mol Genet. ited. Alzheimers Dement. 2016;12:733–48.
2007;16:555–66. 25. Calafate S, Flavin W, Verstreken P, Moechars
12. Bolmont T, Clavaguera F, Meyer-Luehmann M, D.  Loss of Bin1 promotes the propagation of Tau
Herzig MC, Radde R, Staufenbiel M, Lewis J, pathology. Cell Rep. 2016;17:931–40.
Hutton M, Tolnay M, Jucker M.  Induction of tau 26. Castillo-Carranza DL, Gerson JE, Sengupta U,
pathology by intracerebral infusion of amyloid- Guerrero-Munoz MJ, Lasagna-Reeves CA, Kayed
beta -containing brain extract and by amyloid-beta R. Specific targeting of tau oligomers in Htau mice
deposition in APP x Tau transgenic mice. Am J
­ prevents cognitive impairment and tau toxicity
Pathol. 2007;171:2012–20. following injection with brain-derived tau oligo-
13. Bramblett GT, Goedert M, Jakes R, Merrick SE, meric seeds. J Alzheimers Dis. 2014;40(Suppl
Trojanowski JQ, Lee VM. Abnormal tau phosphory- 1):S97–S111.
lation at Ser396 in Alzheimer's disease recapitulates 27. Cavallini A, Brewerton S, Bell A, Sargent S, Glover
development and contributes to reduced microtubule S, Hardy C, Moore R, Calley J, Ramachandran
binding. Neuron. 1993;10:1089–99. D, Poidinger M, Karran E, Davies P, Hutton M,
14. Brandt R, Hundelt M, Shahani N.  Tau altera- Szekeres P, Bose S. An unbiased approach to iden-
tion and neuronal degeneration in tauopathies: tifying tau kinases that phosphorylate tau at sites
mechanisms and models. Biochim Biophys Acta. associated with Alzheimer disease. J Biol Chem.
2005;1739:331–54. 2013;288:23331–47.
15. Brion JP, Couck AM, Passareiro E, Flament-Durand 28. Chapuis J, Hansmannel F, Gistelinck M, Mounier
J. Neurofibrillary tangles of Alzheimer’s disease: an A, Van Cauwenberghe C, Kolen KV, Geller F,
immunohistochemical study. J Submicrosc Cytol. Sottejeau Y, Harold D, Dourlen P, Grenier-Boley B,
1985;17:89–96. Kamatani Y, Delepine B, Demiautte F, Zelenika D,
16. Brion JP, Anderton BH, Authelet M, Dayanandan Zommer N, Hamdane M, Bellenguez C, Dartigues
R, Leroy K, Lovestone S, Octave JN, Pradier L, JF, Hauw JJ, Letronne F, Ayral AM, Sleegers K,
Touchet N, Tremp G. Neurofibrillary tangles and tau Schellens A, Broeck LV, Engelborghs S, De Deyn
phosphorylation. Biochem Soc Symp. 2001:81–8. PP, Vandenberghe R, O’Donovan M, Owen M,
17. Brouhard GJ, Rice LM.  Microtubule dynamics: an Epelbaum J, Mercken M, Karran E, Bantscheff M,
interplay of biochemistry and mechanics. Nat Rev Drewes G, Joberty G, Campion D, Octave JN, Berr
Mol Cell Biol. 2018;19:451–63. C, Lathrop M, Callaerts P, Mann D, Williams J,
18. Brownlees J, Irving NG, Brion JP, Gibb BJ, Buee L, Dewachter I, Van Broeckhoven C, Amouyel
Wagner U, Woodgett J, Miller CC.  Tau phosphor- P, Moechars D, Dermaut B, Lambert JC, G. con-
ylation in transgenic mice expressing glycogen sortium. Increased expression of BIN1 mediates
synthase kinase-3beta transgenes. Neuroreport. Alzheimer genetic risk by modulating tau pathology.
1997;8:3251–5. Mol Psychiatry. 2013;18:1225–34.
19. Brunden KR, Trojanowski JQ, Lee VM.  Advances 29. Cingolani LA, Goda Y. Actin in action: the interplay
in tau-focused drug discovery for Alzheimer’s dis- between the actin cytoskeleton and synaptic efficacy.
ease and related tauopathies. Nat Rev Drug Discov. Nat Rev Neurosci. 2008;9:344–56.
2009;8:783–93. 30. Cripps D, Thomas SN, Jeng Y, Yang F, Davies P,
20. Brunden KR, Ballatore C, Crowe A, Smith AB 3rd, Yang AJ.  Alzheimer disease-specific conformation
Lee VM, Trojanowski JQ. Tau-directed drug discov- of hyperphosphorylated paired helical filament-
ery for Alzheimer's disease and related tauopathies: Tau is polyubiquitinated through Lys-48, Lys-11,
a focus on tau assembly inhibitors. Exp Neurol. and Lys-6 ubiquitin conjugation. J Biol Chem.
2010;223:304–10. 2006;281:10825–38.
21. Buee L, Bussiere T, Buee-Scherrer V, Delacourte A, 31. Crowther T, Goedert M, Wischik CM.  The repeat
Hof PR. Tau protein isoforms, phosphorylation and region of microtubule-associated protein tau forms
role in neurodegenerative disorders. Brain Res Brain part of the core of the paired helical filament of
Res Rev. 2000;33:95–130. Alzheimer's disease. Ann Med. 1989;21:127–32.
22. Butzlaff M, Hannan SB, Karsten P, Lenz S, Ng J, 32. Cruchaga C, Karch CM, Jin SC, Benitez BA,
Vossfeldt H, Prussing K, Pflanz R, Schulz JB, Rasse Cai Y, Guerreiro R, Harari O, Norton J, Budde J,
T, Voigt A.  Impaired retrograde transport by the Bertelsen S, Jeng AT, Cooper B, Skorupa T, Carrell
Dynein/Dynactin complex contributes to tau-induced D, Levitch D, Hsu S, Choi J, Ryten M, Sassi C,
toxicity. Hum Mol Genet. 2015;24:3623–37. Bras J, Gibbs RJ, Hernandez DG, Lupton MK,
160 I.-C. Stancu et al.

Powell J, Forabosco P, Ridge PG, Corcoran CD, implications for Alzheimer’s disease. J Cell Biol.
Tschanz JT, Norton MC, Munger RG, Schmutz 1998;143:777–94.
C, Leary M, Demirci FY, Bamne MN, Wang X, 44. Eftekharzadeh B, Daigle JG, Kapinos LE, Coyne
Lopez OL, Ganguli M, Medway C, Turton J, Lord A, Schiantarelli J, Carlomagno Y, Cook C, Miller
J, Braae A, Barber I, Brown K, U.K.C. Alzheimer’s SJ, Dujardin S, Amaral AS, Grima JC, Bennett
Research, Pastor P, Lorenzo-Betancor O, Brkanac Z, RE, Tepper K, DeTure M, Vanderburgh CR, Corjuc
Scott E, Topol E, Morgan K, Rogaeva E, Singleton BT, DeVos SL, Gonzalez JA, Chew J, Vidensky S,
A, Hardy J, Kamboh MI, George-Hyslop PS, Cairns Gage FH, Mertens J, Troncoso J, Mandelkow E,
N, Morris JC, Kauwe JSK, Goate AM. Rare coding Salvatella X, Lim RYH, Petrucelli L, Wegmann
variants in the phospholipase D3 gene confer risk for S, Rothstein JD, Hyman BT.  Tau protein dis-
Alzheimer’s disease. Nature. 2014;505:550–4. rupts Nucleocytoplasmic transport in Alzheimer’s
33. Daebel V, Chinnathambi S, Biernat J, Schwalbe Disease. Neuron. 2018;99(925–940):e927.
M, Habenstein B, Loquet A, Akoury E, Tepper K, 45. Elie A, Prezel E, Guerin C, Denarier E, Ramirez-Rios
Muller H, Baldus M, Griesinger C, Zweckstetter M, S, Serre L, Andrieux A, Fourest-Lieuvin A, Blanchoin
Mandelkow E, Vijayan V, Lange A. Beta-sheet core L, Arnal I.  Tau co-organizes dynamic microtubule
of tau paired helical filaments revealed by solid-state and actin networks. Sci Rep. 2015;5:9964.
NMR. J Am Chem Soc. 2012;134:13982–9. 46. Engel T, Goni-Oliver P, Lucas JJ, Avila J, Hernandez
34. David DC, Layfield R, Serpell L, Narain Y, Goedert F.  Chronic lithium administration to FTDP-17 tau
M, Spillantini MG. Proteasomal degradation of tau and GSK-3beta overexpressing mice prevents tau
protein. J Neurochem. 2002;83:176–85. hyperphosphorylation and neurofibrillary tangle for-
35. Dawson HN, Ferreira A, Eyster MV, Ghoshal N, mation, but pre-formed neurofibrillary tangles do not
Binder LI, Vitek MP. Inhibition of neuronal matura- revert. J Neurochem. 2006;99:1445–55.
tion in primary hippocampal neurons from tau defi- 47. Feinstein SC, Wilson L.  Inability of tau to prop-
cient mice. J Cell Sci. 2001;114:1179–87. erly regulate neuronal microtubule dynamics: a
36. de Vrij FM, Fischer DF, van Leeuwen FW, Hol loss-of-function mechanism by which tau might
­
EM.  Protein quality control in Alzheimer’s disease mediate neuronal cell death. Biochim Biophys Acta.
by the ubiquitin proteasome system. Prog Neurobiol. 2005;1739:268–79.
2004;74:249–70. 48. Fischer D, Mukrasch MD, von Bergen M, Klos-
37. Deger JM, Gerson JE, Kayed R.  The interrelation- Witkowska A, Biernat J, Griesinger C, Mandelkow
ship of proteasome impairment and oligomeric E, Zweckstetter M.  Structural and microtubule
intermediates in neurodegeneration. Aging Cell. binding properties of tau mutants of frontotemporal
2015;14:715–24. dementias. Biochemistry. 2007;46:2574–82.
38. Delacourte A, Buee L.  Tau pathology: a marker of 49. Fitzpatrick AWP, Falcon B, He S, Murzin AG,
neurodegenerative disorders. Curr Opin Neurol. Murshudov G, Garringer HJ, Crowther RA, Ghetti
2000;13:371–6. B, Goedert M, Scheres SHW.  Cryo-EM structures
39. Despres C, Byrne C, Qi H, Cantrelle FX, Huvent I, of tau filaments from Alzheimer’s disease. Nature.
Chambraud B, Baulieu EE, Jacquot Y, Landrieu I, 2017;547:185–90.
Lippens G, Smet-Nocca C. Identification of the Tau 50. Flach K, Ramminger E, Hilbrich I, Arsalan-Werner
phosphorylation pattern that drives its aggregation. A, Albrecht F, Herrmann L, Goedert M, Arendt T,
Proc Natl Acad Sci USA. 2017;114:9080–5. Holzer M.  Axotrophin/MARCH7 acts as an E3
40. Dixit R, Ross JL, Goldman YE, Holzbaur ubiquitin ligase and ubiquitinates tau protein in vitro
EL.  Differential regulation of dynein and kinesin impairing microtubule binding. Biochim Biophys
motor proteins by tau. Science. 2008;319:1086–9. Acta. 2014;1842:1527–38.
41. Dogterom M, Koenderink GH.  Actin-microtubule 51. Frandemiche ML, De Seranno S, Rush T, Borel
crosstalk in cell biology. Nat Rev Mol Cell Biol. E, Elie A, Arnal I, Lante F, Buisson A.  Activity-
2018; dependent tau protein translocation to excitatory
42. Dourlen P, Fernandez-Gomez FJ, Dupont C, synapse is disrupted by exposure to amyloid-beta
Grenier-Boley B, Bellenguez C, Obriot H, Caillierez oligomers. J Neurosci. 2014;34:6084–97.
R, Sottejeau Y, Chapuis J, Bretteville A, Abdelfettah 52. Frost B, Diamond MI.  Prion-like mechanisms in
F, Delay C, Malmanche N, Soininen H, Hiltunen M, neurodegenerative diseases. Nat Rev Neurosci.
Galas MC, Amouyel P, Sergeant N, Buee L, Lambert 2010;11:155–9.
JC, Dermaut B. Functional screening of Alzheimer 53. Frost B, Jacks RL, Diamond MI. Propagation of tau
risk loci identifies PTK2B as an in vivo modulator misfolding from the outside to the inside of a cell. J
and early marker of Tau pathology. Mol Psychiatry. Biol Chem. 2009;284:12845–52.
2017;22:874–83. 54. Fulga TA, Elson-Schwab I, Khurana V, Steinhilb
43. Ebneth A, Godemann R, Stamer K, Illenberger S, ML, Spires TL, Hyman BT, Feany MB.  Abnormal
Trinczek B, Mandelkow E.  Overexpression of tau bundling and accumulation of F-actin mediates tau-
protein inhibits kinesin-dependent trafficking of induced neuronal degeneration in  vivo. Nat Cell
vesicles, mitochondria, and endoplasmic reticulum: Biol. 2007;9(2):139–48.
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 161

55. Furman JL, Vaquer-Alicea J, White CL 3rd, Cairns 68. Gotz J, Chen F, van Dorpe J, Nitsch RM. Formation
NJ, Nelson PT, Diamond MI. Widespread tau seed- of neurofibrillary tangles in P301l tau trans-
ing activity at early Braak stages. Acta Neuropathol. genic mice induced by Abeta 42 fibrils. Science.
2017;133:91–100. 2001;293:1491–5.
56. Gallo G. Tau is actin up in Alzheimer's disease. Nat 69. Griffith LM, Pollard TD.  The interaction of actin
Cell Biol. 2007;9(2):133–4. filaments with microtubules and microtubule-associ-
57. Gibbons GS, Banks RA, Kim B, Xu H, Changolkar ated proteins. J Biol Chem. 1982;257:9143–51.
L, Leight SN, Riddle DM, Li C, Gathagan RJ, Brown 70. Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M,
HJ, Zhang B, Trojanowski JQ, Lee VM. GFP-mutant Wisniewski HM, Binder LI. Abnormal phosphoryla-
human tau transgenic mice develop Tauopathy fol- tion of the microtubule-associated protein tau (tau)
lowing CNS injections of Alzheimer's brain-derived in Alzheimer cytoskeletal pathology. Proc Natl Acad
pathological tau or synthetic mutant human tau Sci U S A. 1986;83:4913–7.
fibrils. J Neurosci. 2017;37:11485–94. 71. Gunawardana CG, Mehrabian M, Wang X, Mueller
58. Gibbons GS, Lee VMY, Trojanowski I, Lubambo IB, Jonkman JE, Wang H, Schmitt-
JQ. Mechanisms of cell-to-cell transmission of path- Ulms G.  The human tau Interactome: binding to
ological tau: A review. JAMA Neurol. 2018. the Ribonucleoproteome, and impaired binding
59. Giralt A, de Pins B, Cifuentes-Diaz C, Lopez-Molina of the Proline-to-Leucine mutant at position 301
L, Farah AT, Tible M, Deramecourt V, Arold ST, (P301L) to chaperones and the proteasome. Mol Cell
Gines S, Hugon J, Girault JA.  PTK2B/Pyk2 over- Proteomics. 2015;14:3000–14.
expression improves a mouse model of Alzheimer's 72. Guo JL, Lee VM. Seeding of normal tau by pathologi-
disease. Exp Neurol. 2018;307:62–73. cal tau conformers drives pathogenesis of Alzheimer-
60. Goedert M. NEURODEGENERATION. Alzheimer like tangles. J Biol Chem. 2011;286:15317–31.
’s and Parkinson’s diseases: the prion concept in rela- 73. Guo JL, Buist A, Soares A, Callaerts K, Calafate S,
tion to assembled Abeta, tau, and alpha-synuclein. Stevenaert F, Daniels JP, Zoll BE, Crowe A, Brunden
Science. 2015;349:1255555. KR, Moechars D, Lee VM. The dynamics and turn-
61. Goedert M, Jakes R. Expression of separate isoforms over of tau aggregates in cultured cells: INSIGHTS
of human tau protein: correlation with the tau pat- INTO THERAPIES FOR TAUOPATHIES.  J Biol
tern in brain and effects on tubulin polymerization. Chem. 2016a;291:13175–93.
EMBO J. 1990;9:4225–30. 74. Guo JL, Narasimhan S, Changolkar L, He Z, Stieber
62. Goedert M, Wischik CM, Crowther RA, Walker A, Zhang B, Gathagan RJ, Iba M, McBride JD,
JE, Klug A.  Cloning and sequencing of the cDNA Trojanowski JQ, Lee VM.  Unique pathological
encoding a core protein of the paired helical fila- tau conformers from Alzheimer's brains transmit
ment of Alzheimer disease: identification as the tau pathology in nontransgenic mice. J Exp Med.
microtubule-associated protein tau. Proc Natl Acad 2016b;213:2635–54.
Sci USA. 1988;85:4051–5. 75. Guo T, Noble W, Hanger DP.  Roles of tau pro-
63. Goedert M, Spillantini MG, Jakes R, Rutherford D, tein in health and disease. Acta Neuropathol.
Crowther RA. Multiple isoforms of human microtu- 2017;133:665–704.
bule-associated protein tau: sequences and localiza- 76. Gustke N, Steiner B, Mandelkow EM, Biernat
tion in neurofibrillary tangles of Alzheimer's disease. J, Meyer HE, Goedert M, Mandelkow E.  The
Neuron. 1989a;3:519–26. Alzheimer-like phosphorylation of tau protein
64. Goedert M, Spillantini MG, Potier MC, Ulrich J, reduces microtubule binding and involves Ser-pro
Crowther RA. Cloning and sequencing of the cDNA and Thr-pro motifs. FEBS Lett. 1992;307:199–205.
encoding an isoform of microtubule-associated pro- 77. Gustke N, Trinczek B, Biernat J, Mandelkow
tein tau containing four tandem repeats: differential EM, Mandelkow E.  Domains of tau protein and
expression of tau protein mRNAs in human brain. interactions with microtubules. Biochemistry.
EMBO J. 1989b;8:393–9. 1994;33:9511–22.
65. Goedert M, Jakes R, Crowther RA, Six J, Lubke 78. Hanger DP, Anderton BH, Noble W. Tau phosphory-
U, Vandermeeren M, Cras P, Trojanowski JQ, Lee lation: the therapeutic challenge for neurodegenera-
VM.  The abnormal phosphorylation of tau protein tive disease. Trends Mol Med. 2009;15:112–9.
at Ser-202 in Alzheimer disease recapitulates phos- 79. Hannan SB, Drager NM, Rasse TM, Voigt A, Jahn
phorylation during development. Proc Natl Acad Sci TR.  Cellular and molecular modifier pathways in
USA. 1993;90:5066–70. tauopathies: the big picture from screening inverte-
66. Goedert M, Eisenberg DS, Crowther RA. Propagation brate models. J Neurochem. 2016;137:12–25.
of tau aggregates and Neurodegeneration. Annu Rev 80. Hardy J, Selkoe DJ.  The amyloid hypothesis of
Neurosci. 2017;40:189–210. Alzheimer's disease: progress and problems on the
67. Goode BL, Feinstein SC.  Identification of a novel road to therapeutics. Science. 2002;297:353–6.
microtubule binding and assembly domain in the 81. Harold D, Abraham R, Hollingworth P, Sims R,
developmentally regulated inter-repeat region of tau. Gerrish A, Hamshere ML, Pahwa JS, Moskvina
J Cell Biol. 1994;124:769–82. V, Dowzell K, Williams A, Jones N, Thomas C,
162 I.-C. Stancu et al.

Stretton A, Morgan AR, Lovestone S, Powell J, Mayo K, Sleegers K, Bettens K, Engelborghs S, De

Proitsi P, Lupton MK, Brayne C, Rubinsztein DC, Deyn PP, Van Broeckhoven C, Livingston G, Bass
Gill M, Lawlor B, Lynch A, Morgan K, Brown NJ, Gurling H, McQuillin A, Gwilliam R, Deloukas
KS, Passmore PA, Craig D, McGuinness B, Todd P, Al-Chalabi A, Shaw CE, Tsolaki M, Singleton AB,
S, Holmes C, Mann D, Smith AD, Love S, Kehoe Guerreiro R, Muhleisen TW, Nothen MM, Moebus
PG, Hardy J, Mead S, Fox N, Rossor M, Collinge S, Jockel KH, Klopp N, Wichmann HE, Pankratz VS,
J, Maier W, Jessen F, Schurmann B, Heun R, van Sando SB, Aasly JO, Barcikowska M, Wszolek ZK,
den Bussche H, Heuser I, Kornhuber J, Wiltfang J, Dickson DW, Graff-Radford NR, Petersen RC, et al.
Dichgans M, Frolich L, Hampel H, Hull M, Rujescu Common variants at ABCA7, MS4A6A/MS4A4E,
D, Goate AM, Kauwe JS, Cruchaga C, Nowotny EPHA1, CD33 and CD2AP are associated with
P, Morris JC, Mayo K, Sleegers K, Bettens K, Alzheimer’s disease. Nat Genet. 2011;43:429–35.
Engelborghs S, De Deyn PP, Van Broeckhoven C, 87. Holmes BB, Diamond MI.  Prion-like proper-
Livingston G, Bass NJ, Gurling H, McQuillin A, ties of tau protein: the importance of extracel-
Gwilliam R, Deloukas P, Al-Chalabi A, Shaw CE, lular tau as a therapeutic target. J Biol Chem.
Tsolaki M, Singleton AB, Guerreiro R, Muhleisen 2014;289:19855–61.
TW, Nothen MM, Moebus S, Jockel KH, Klopp 88. Holmes BB, Furman JL, Mahan TE, Yamasaki TR,
N, Wichmann HE, Carrasquillo MM, Pankratz VS, Mirbaha H, Eades WC, Belaygorod L, Cairns NJ,
Younkin SG, Holmans PA, O'Donovan M, Owen MJ, Holtzman DM, Diamond MI. Proteopathic tau seed-
Williams J. Genome-wide association study identi- ing predicts tauopathy in  vivo. Proc Natl Acad Sci
fies variants at CLU and PICALM associated with USA. 2014;111:E4376–85.
Alzheimer’s disease. Nat Genet. 2009;41:1088–93. 89. Holtzman DM, Carrillo MC, Hendrix JA, Bain LJ,
82. Hatakeyama S, Matsumoto M, Kamura T, Murayama Catafau AM, Gault LM, Goedert M, Mandelkow
M, Chui DH, Planel E, Takahashi R, Nakayama KI, E, Mandelkow EM, Miller DS, Ostrowitzki S,
Takashima A.  U-box protein carboxyl terminus of Polydoro M, Smith S, Wittmann M, Hutton M. Tau:
Hsc70-interacting protein (CHIP) mediates poly- from research to clinical development. Alzheimers
ubiquitylation preferentially on four-repeat tau and Dement. 2016;12:1033–9.
is involved in neurodegeneration of tauopathy. J 90. Hoogenraad CC, Bradke F.  Control of neuronal
Neurochem. 2004;91:299–307. polarity and plasticity–a renaissance for microtu-
83. He HJ, Wang XS, Pan R, Wang DL, Liu MN, bules? Trends Cell Biol. 2009;19:669–76.
He RQ.  The proline-rich domain of tau plays a 91. Horwitz AR, Parsons JT. Cell migration–movin’ on.
role in interactions with actin. BMC Cell Biol. Science. 1999;286:1102–3.
2009;10:81. 92. Hotulainen P, Hoogenraad CC.  Actin in dendritic
84. He Z, Guo JL, McBride JD, Narasimhan S, Kim spines: connecting dynamics to function. J Cell Biol.
H, Changolkar L, Zhang B, Gathagan RJ, Yue C, 2010;189:619–29.
Dengler C, Stieber A, Nitla M, Coulter DA, Abel T, 93. Ikegami S, Harada A, Hirokawa N.  Muscle weak-
Brunden KR, Trojanowski JQ, Lee VM.  Amyloid- ness, hyperactivity, and impairment in fear con-
beta plaques enhance Alzheimer’s brain tau-seeded ditioning in tau-deficient mice. Neurosci Lett.
pathologies by facilitating neuritic plaque tau aggre- 2000;279:129–32.
gation. Nat Med. 2018;24:29–38. 94. Ingram EM, Spillantini MG.  Tau gene mutations:
85. Hernandez F, Borrell J, Guaza C, Avila J, Lucas dissecting the pathogenesis of FTDP-17. Trends Mol
JJ.  Spatial learning deficit in transgenic mice Med. 2002;8:555–62.
that conditionally over-express GSK-3beta in the 95. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A,
brain but do not form tau filaments. J Neurochem. van Eersel J, Wolfing H, Chieng BC, Christie MJ,
2002;83:1529–33. Napier IA, Eckert A, Staufenbiel M, Hardeman E,
86. Hollingworth P, Harold D, Sims R, Gerrish A, Gotz J. Dendritic function of tau mediates amyloid-
Lambert JC, Carrasquillo MM, Abraham R, beta toxicity in Alzheimer's disease mouse models.
Hamshere ML, Pahwa JS, Moskvina V, Dowzell K, Cell. 2010;142:387–97.
Jones N, Stretton A, Thomas C, Richards A, Ivanov 96. Jack CR Jr, Knopman DS, Jagust WJ, Shaw LM,
D, Widdowson C, Chapman J, Lovestone S, Powell Aisen PS, Weiner MW, Petersen RC, Trojanowski
J, Proitsi P, Lupton MK, Brayne C, Rubinsztein DC, JQ.  Hypothetical model of dynamic biomarkers
Gill M, Lawlor B, Lynch A, Brown KS, Passmore of the Alzheimer’s pathological cascade. Lancet
PA, Craig D, McGuinness B, Todd S, Holmes Neurol. 2010;9:119–28.
C, Mann D, Smith AD, Beaumont H, Warden D, 97. Jack CR Jr, Knopman DS, Jagust WJ, Petersen
Wilcock G, Love S, Kehoe PG, Hooper NM, Vardy RC, Weiner MW, Aisen PS, Shaw LM, Vemuri P,
ER, Hardy J, Mead S, Fox NC, Rossor M, Collinge J, Wiste HJ, Weigand SD, Lesnick TG, Pankratz VS,
Maier W, Jessen F, Ruther E, Schurmann B, Heun R, Donohue MC, Trojanowski JQ.  Tracking patho-
Kolsch H, van den Bussche H, Heuser I, Kornhuber physiological processes in Alzheimer’s disease: an
J, Wiltfang J, Dichgans M, Frolich L, Hampel H, updated hypothetical model of dynamic biomarkers.
Gallacher J, Hull M, Rujescu D, Giegling I, Goate Lancet Neurol. 2013;12:207–16.
AM, Kauwe JS, Cruchaga C, Nowotny P, Morris JC,
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 163

98. Jeganathan S, von Bergen M, Brutlach H, Steinhoff 113. Kim E, Sheng M. PDZ domain proteins of synapses.
HJ, Mandelkow E. Global hairpin folding of tau in Nat Rev Neurosci. 2004;5:771–81.
solution. Biochemistry. 2006;45:2283–93. 114. Kins S, Crameri A, Evans DR, Hemmings BA,
99. Jeganathan S, von Bergen M, Mandelkow EM, Nitsch RM, Gotz J. Reduced protein phosphatase 2A
Mandelkow E.  The natively unfolded character of activity induces hyperphosphorylation and altered
tau and its aggregation to Alzheimer-like paired heli- compartmentalization of tau in transgenic mice. J
cal filaments. Biochemistry. 2008;47:10526–39. Biol Chem. 2001;276:38193–200.
100. Jonsson T, Stefansson H, Steinberg S, Jonsdottir I, 115. Kobayashi S, Tanaka T, Soeda Y, Almeida OFX,
Jonsson PV, Snaedal J, Bjornsson S, Huttenlocher Takashima A.  Local Somatodendritic transla-
J, Levey AI, Lah JJ, Rujescu D, Hampel H, tion and hyperphosphorylation of tau protein trig-
Giegling I, Andreassen OA, Engedal K, Ulstein I, gered by AMPA and NMDA receptor stimulation.
Djurovic S, Ibrahim-Verbaas C, Hofman A, Ikram EBioMedicine. 2017;20:120–6.
MA, van Duijn CM, Thorsteinsdottir U, Kong A, 116. Kraemer BC, Burgess JK, Chen JH, Thomas JH,
Stefansson K.  Variant of TREM2 associated with Schellenberg GD. Molecular pathways that influence
the risk of Alzheimer’s disease. N Engl J Med. human tau-induced pathology in Caenorhabditis ele-
2013;368:107–16. gans. Hum Mol Genet. 2006;15:1483–96.
101. Jucker M, Walker LC. Pathogenic protein seeding in 117. Kulathu Y, Komander D.  Atypical ubiquitylation  –
Alzheimer disease and other neurodegenerative dis- the unexplored world of polyubiquitin beyond
orders. Ann Neurol. 2011;70:532–40. Lys48 and Lys63 linkages. Nat Rev Mol Cell Biol.
102. Jucker M, Walker LC.  Self-propagation of patho- 2012;13:508–23.
genic protein aggregates in neurodegenerative dis- 118. Lambert JC, Heath S, Even G, Campion D,
eases. Nature. 2013;501:45–51. Sleegers K, Hiltunen M, Combarros O, Zelenika
103. Kadavath H, Hofele RV, Biernat J, Kumar S, Tepper D, Bullido MJ, Tavernier B, Letenneur L, Bettens
K, Urlaub H, Mandelkow E, Zweckstetter M.  Tau K, Berr C, Pasquier F, Fievet N, Barberger-Gateau
stabilizes microtubules by binding at the interface P, Engelborghs S, De Deyn P, Mateo I, Franck A,
between tubulin heterodimers. Proc Natl Acad Sci Helisalmi S, Porcellini E, Hanon O, I.  European
USA. 2015a;112:7501–6. Alzheimer’s Disease Initiative, de Pancorbo MM,
104. Kadavath H, Jaremko M, Jaremko L, Biernat J, Lendon C, Dufouil C, Jaillard C, Leveillard T,
Mandelkow E, Zweckstetter M.  Folding of the tau Alvarez V, Bosco P, Mancuso M, Panza F, Nacmias
protein on microtubules. Angew Chem Int Ed Engl. B, Bossu P, Piccardi P, Annoni G, Seripa D,
2015b;54:10347–51. Galimberti D, Hannequin D, Licastro F, Soininen
105. Kadavath H, Cabrales Fontela Y, Jaremko M, H, Ritchie K, Blanche H, Dartigues JF, Tzourio C,
Jaremko L, Overkamp K, Biernat J, Mandelkow Gut I, Van Broeckhoven C, Alperovitch A, Lathrop
E, Zweckstetter M.  The binding mode of a tau M, Amouyel P.  Genome-wide association study
peptide with tubulin. Angew Chem Int Ed Engl. identifies variants at CLU and CR1 associated with
2018;57:3246–50. Alzheimer’s disease. Nat Genet. 2009;41:1094–9.
106. Kanemaru K, Takio K, Miura R, Titani K, Ihara 119. Lasagna-Reeves CA, Castillo-Carranza DL,
Y.  Fetal-type phosphorylation of the tau in paired Sengupta U, Guerrero-Munoz MJ, Kiritoshi T,
helical filaments. J Neurochem. 1992;58:1667–75. Neugebauer V, Jackson GR, Kayed R.  Alzheimer
107. Kar S, Fan J, Smith MJ, Goedert M, Amos LA. Repeat brain-derived tau oligomers propagate pathology
motifs of tau bind to the insides of microtubules in from endogenous tau. Sci Rep. 2012;2:700.
the absence of taxol. EMBO J. 2003;22:70–7. 120. Lee G, Cowan N, Kirschner M. The primary struc-
108. Karch CM, Goate AM.  Alzheimer's disease risk ture and heterogeneity of tau protein from mouse
genes and mechanisms of disease pathogenesis. Biol brain. Science. 1988;239:285–8.
Psychiatry. 2015;77:43–51. 121. Lee VM, Goedert M, Trojanowski
109. Keck S, Nitsch R, Grune T, Ullrich O. Proteasome JQ.  Neurodegenerative tauopathies. Annu Rev
inhibition by paired helical filament-tau in brains Neurosci. 2001;24:1121–59.
of patients with Alzheimer’s disease. J Neurochem. 122. Lee VM, Brunden KR, Hutton M, Trojanowski

2003;85:115–22. JQ.  Developing therapeutic approaches to tau,
110. Kfoury N, Holmes BB, Jiang H, Holtzman DM, selected kinases, and related neuronal pro-
Diamond MI.  Trans-cellular propagation of tau tein targets. Cold Spring Harb Perspect Med.
aggregation by fibrillar species. J Biol Chem. 2011;1:a006437.
2012;287:19440–51. 123. Lee MJ, Lee JH, Rubinsztein DC.  Tau degrada-
111. Khanna MR, Kovalevich J, Lee VM, Trojanowski tion: the ubiquitin-proteasome system versus the
JQ, Brunden KR.  Therapeutic strategies for the autophagy-lysosome system. Prog Neurobiol.
treatment of tauopathies: hopes and challenges. 2013;105:49–59.
Alzheimers Dement. 2016;12:1051–65. 124. Leroy K, Yilmaz Z, Brion JP.  Increased level of
112. Khlistunova I, Pickhardt M, Biernat J, Wang Y, active GSK-3beta in Alzheimer’s disease and accu-
Mandelkow EM, Mandelkow E.  Inhibition of mulation in argyrophilic grains and in neurones at
tau aggregation in cell models of tauopathy. Curr different stages of neurofibrillary degeneration.
Alzheimer Res. 2007;4:544–6. Neuropathol Appl Neurobiol. 2007;33:43–55.
164 I.-C. Stancu et al.

125. Leroy K, Ando K, Heraud C, Yilmaz Z, Authelet 137. Mansuroglu Z, Benhelli-Mokrani H, Marcato
M, Boeynaems JM, Buee L, De Decker R, Brion V, Sultan A, Violet M, Chauderlier A, Delattre
JP.  Lithium treatment arrests the development of L, Loyens A, Talahari S, Begard S, Nesslany F,
neurofibrillary tangles in mutant tau transgenic Colin M, Soues S, Lefebvre B, Buee L, Galas MC,
mice with advanced neurofibrillary pathology. J Bonnefoy E. Loss of tau protein affects the structure,
Alzheimers Dis. 2010;19:705–19. transcription and repair of neuronal pericentromeric
126. Lesne S, Koh MT, Kotilinek L, Kayed R, Glabe CG, heterochromatin. Sci Rep. 2016;6:33047.
Yang A, Gallagher M, Ashe KH. A specific amyloid- 138. McInnes J, Wierda K, Snellinx A, Bounti L, Wang
beta protein assembly in the brain impairs memory. YC, Stancu IC, Apostolo N, Gevaert K, Dewachter
Nature. 2006;440:352–7. I, Spires-Jones TL, De Strooper B, De Wit J,
127. Lewis J, Dickson DW, Lin WL, Chisholm L, Corral Zhou L, Verstreken P.  Synaptogyrin-3 mediates
A, Jones G, Yen SH, Sahara N, Skipper L, Yager presynaptic dysfunction induced by tau. Neuron.
D, Eckman C, Hardy J, Hutton M, McGowan 2018;97(823–835):e828.
E.  Enhanced neurofibrillary degeneration in trans- 139. Miyamoto T, Stein L, Thomas R, Djukic B, Taneja P,
genic mice expressing mutant tau and APP. Science. Knox J, Vossel K, Mucke L. Phosphorylation of tau
2001;293:1487–91. at Y18, but not tau-fyn binding, is required for tau
128. Li C, Gotz J. Pyk2 is a novel tau tyrosine kinase that to modulate NMDA receptor-dependent excitotoxic-
is regulated by the tyrosine kinase Fyn. J Alzheimers ity in primary neuronal culture. Mol Neurodegener.
Dis. 2018;64:205–21. 2017;12:41.
129. Liu F, Grundke-Iqbal I, Iqbal K, Gong 140. Mocanu MM, Nissen A, Eckermann K, Khlistunova
CX.  Contributions of protein phosphatases PP1, I, Biernat J, Drexler D, Petrova O, Schonig K, Bujard
PP2A, PP2B and PP5 to the regulation of tau phos- H, Mandelkow E, Zhou L, Rune G, Mandelkow
phorylation. Eur J Neurosci. 2005;22:1942–50. EM.  The potential for beta-structure in the repeat
130. Liu C, Song X, Nisbet R, Gotz J. domain of tau protein determines aggregation, syn-
Co-immunoprecipitation with tau isoform-specific aptic decay, neuronal loss, and coassembly with
antibodies reveals distinct protein interactions and endogenous tau in inducible mouse models of tauop-
highlights a putative role for 2N tau in Disease. J athy. J Neurosci. 2008;28:737–48.
Biol Chem. 2016;291:8173–88. 141. Mondragon-Rodriguez S, Trillaud-Doppia E,
131. Louis JV, Martens E, Borghgraef P, Lambrecht C, Dudilot A, Bourgeois C, Lauzon M, Leclerc N,
Sents W, Longin S, Zwaenepoel K, Pijnenborg R, Boehm J.  Interaction of endogenous tau protein
Landrieu I, Lippens G, Ledermann B, Gotz J, Van with synaptic proteins is regulated by N-methyl-D-
Leuven F, Goris J, Janssens V.  Mice lacking phos- aspartate receptor-dependent tau phosphorylation. J
phatase PP2A subunit PR61/B’delta (Ppp2r5d) Biol Chem. 2012;287:32040–53.
develop spatially restricted tauopathy by deregula- 142. Mori H, Kondo J, Ihara Y. Ubiquitin is a component
tion of CDK5 and GSK3beta. Proc Natl Acad Sci of paired helical filaments in Alzheimer's disease.
USA. 2011;108:6957–62. Science. 1987;235:1641–4.
132. Lovestone S, Davis DR, Webster MT, Kaech S, Brion 143. Mudher A, Colin M, Dujardin S, Medina M,
JP, Matus A, Anderton BH.  Lithium reduces tau Dewachter I, Alavi Naini SM, Mandelkow EM,
phosphorylation: effects in living cells and in neu- Mandelkow E, Buee L, Goedert M, Brion JP. What
rons at therapeutic concentrations. Biol Psychiatry. is the evidence that tau pathology spreads through
1999;45:995–1003. prion-like propagation? Acta Neuropathol Commun.
133. Magnani E, Fan J, Gasparini L, Golding M, Williams 2017;5:99.
M, Schiavo G, Goedert M, Amos LA, Spillantini 144. Mukrasch MD, Biernat J, von Bergen M, Griesinger
MG.  Interaction of tau protein with the dynactin C, Mandelkow E, Zweckstetter M.  Sites of tau
complex. EMBO J. 2007;26:4546–54. important for aggregation populate {beta}-structure
134. Malmanche N, Dourlen P, Gistelinck M, Demiautte and bind to microtubules and polyanions. J Biol
F, Link N, Dupont C, Vanden Broeck L, Werkmeister Chem. 2005;280:24978–86.
E, Amouyel P, Bongiovanni A, Bauderlique H, 145. Mukrasch MD, von Bergen M, Biernat J, Fischer D,
Moechars D, Royou A, Bellen HJ, Lafont F, Griesinger C, Mandelkow E, Zweckstetter M.  The
Callaerts P, Lambert JC, Dermaut B. Developmental “jaws” of the tau-microtubule interaction. J Biol
expression of 4-repeat-tau induces neuronal aneu- Chem. 2007;282:12230–9.
ploidy in Drosophila Tauopathy models. Sci Rep. 146. Myeku N, Clelland CL, Emrani S, Kukushkin NV,
2017;7:40764. Yu WH, Goldberg AL, Duff KE. Tau-driven 26S pro-
135. Mandelkow EM, Mandelkow E. Biochemistry and cell teasome impairment and cognitive dysfunction can
biology of tau protein in neurofibrillary degeneration. be prevented early in disease by activating cAMP-
Cold Spring Harb Perspect Med. 2012;2:a006247. PKA signaling. Nat Med. 2016;22:46–53.
136. Mandelkow EM, Biernat J, Drewes G, Gustke N, 147. Narasimhan S, Guo JL, Changolkar L, Stieber A,
Trinczek B, Mandelkow E. Tau domains, phosphory- McBride JD, Silva LV, He Z, Zhang B, Gathagan
lation, and interactions with microtubules. Neurobiol RJ, Trojanowski JQ, Lee VMY.  Pathological tau
Aging. 1995;16:355–62.. discussion 362–353 strains from human brains recapitulate the diver-
13  Tau Interacting Proteins: Gaining Insight into the Roles of Tau in Health and Disease 165

sity of Tauopathies in nontransgenic mouse brain. J 161. Pooler AM, Phillips EC, Lau DH, Noble W,
Neurosci. 2017;37:11406–23. Hanger DP.  Physiological release of endogenous
148. Neve RL, Harris P, Kosik KS, Kurnit DM, Donlon tau is stimulated by neuronal activity. EMBO Rep.
TA.  Identification of cDNA clones for the human 2013;14:389–94.
microtubule-associated protein tau and chromosomal 162. Pooler AM, Polydoro M, Maury EA, Nicholls SB,
localization of the genes for tau and microtubule- Reddy SM, Wegmann S, William C, Saqran L,
associated protein 2. Brain Res. 1986;387:271–80. Cagsal-Getkin O, Pitstick R, Beier DR, Carlson
149. Nijman SM, Luna-Vargas MP, Velds A, GA, Spires-Jones TL, Hyman BT. Amyloid acceler-
Brummelkamp TR, Dirac AM, Sixma TK, Bernards ates tau propagation and toxicity in a model of early
R. A genomic and functional inventory of deubiqui- Alzheimer’s disease. Acta Neuropathol Commun.
tinating enzymes. Cell. 2005;123:773–86. 2015;3:14.
150. Noble W, Olm V, Takata K, Casey E, Mary O, 163. Poorkaj P, Bird TD, Wijsman E, Nemens E, Garruto
Meyerson J, Gaynor K, LaFrancois J, Wang L, RM, Anderson L, Andreadis A, Wiederholt WC,
Kondo T, Davies P, Burns M, Veeranna R, Nixon Raskind M, Schellenberg GD.  Tau is a candidate
D, Dickson Y, Matsuoka M, Ahlijanian LFL, Duff gene for chromosome 17 frontotemporal dementia.
K. Cdk5 is a key factor in tau aggregation and tangle Ann Neurol. 1998;43:815–25.
formation in vivo. Neuron. 2003;38:555–65. 164. Roberson ED, Scearce-Levie K, Palop JJ, Yan
151. Noble W, Planel E, Zehr C, Olm V, Meyerson F, Cheng IH, Wu T, Gerstein H, Yu GQ, Mucke
J, Suleman F, Gaynor K, Wang L, LaFrancois J, L.  Reducing endogenous tau ameliorates amyloid
Feinstein B, Burns M, Krishnamurthy P, Wen Y, beta-induced deficits in an Alzheimer’s disease
Bhat R, Lewis J, Dickson D, Duff K. Inhibition of mouse model. Science. 2007;316:750–4.
glycogen synthase kinase-3 by lithium correlates 165. Roberson ED, Halabisky B, Yoo JW, Yao J, Chin J,
with reduced tauopathy and degeneration in  vivo. Yan F, Wu T, Hamto P, Devidze N, Yu GQ, Palop JJ,
Proc Natl Acad Sci USA. 2005;102:6990–5. Noebels JL, Mucke L.  Amyloid-beta/Fyn-induced
152. Oyama F, Kotliarova S, Harada A, Ito M, Miyazaki synaptic, network, and cognitive impairments
H, Ueyama Y, Hirokawa N, Nukina N, Ihara Y. Gem depend on tau levels in multiple mouse models of
GTPase and tau: morphological changes induced by Alzheimer’s disease. J Neurosci. 2011;31:700–11.
gem GTPase in cho cells are antagonized by tau. J 166. Sahara N, Murayama M, Mizoroki T, Urushitani
Biol Chem. 2004;279:27272–7. M, Imai Y, Takahashi R, Murata S, Tanaka K,
153. Pak CW, Flynn KC, Bamburg JR.  Actin-binding Takashima A.  In vivo evidence of CHIP up-regu-
proteins take the reins in growth cones. Nat Rev lation attenuating tau aggregation. J Neurochem.
Neurosci. 2008;9:136–47. 2005;94:1254–63.
154. Panda D, Samuel JC, Massie M, Feinstein SC, 167. Sanders DW, Kaufman SK, DeVos SL, Sharma AM,
Wilson L.  Differential regulation of microtubule Mirbaha H, Li A, Barker SJ, Foley AC, Thorpe JR,
dynamics by three- and four-repeat tau: implications Serpell LC, Miller TM, Grinberg LT, Seeley WW,
for the onset of neurodegenerative disease. Proc Natl Diamond MI.  Distinct tau prion strains propagate
Acad Sci USA. 2003;100:9548–53. in cells and mice and define different tauopathies.
155. Papasozomenos SC, Binder LI. Phosphorylation deter- Neuron. 2014;82:1271–88.
mines two distinct species of tau in the central nervous 168. Schulman BA, Harper JW.  Ubiquitin-like protein
system. Cell Motil Cytoskeleton. 1987;8:210–26. activation by E1 enzymes: the apex for down-
156. Park S, Lee JH, Jeon JH, Lee MJ.  Degradation or stream signalling pathways. Nat Rev Mol Cell Biol.
aggregation: the ramifications of post-translational 2009;10:319–31.
modifications on tau. BMB Rep. 2018;51:265–73. 169. Scrivo A, Bourdenx M, Pampliega O, Cuervo
157. Parsons JT, Horwitz AR, Schwartz MA. Cell adhe- AM.  Selective autophagy as a potential therapeu-
sion: integrating cytoskeletal dynamics and cellular tic target for neurodegenerative disorders. Lancet
tension. Nat Rev Mol Cell Biol. 2010;11:633–43. Neurol. 2018;17:802–15.
158. Perry G, Friedman R, Shaw G, Chau V. Ubiquitin is 170. Selkoe DJ, Hardy J.  The amyloid hypothesis of
detected in neurofibrillary tangles and senile plaque Alzheimer's disease at 25 years. EMBO Mol Med.
neurites of Alzheimer disease brains. Proc Natl Acad 2016;8:595–608.
Sci USA. 1987;84:3033–6. 171. Serrano-Pozo A, Frosch MP, Masliah E, Hyman
159. Petrucelli L, Dickson D, Kehoe K, Taylor J, Snyder BT.  Neuropathological alterations in Alzheimer
H, Grover A, De Lucia M, McGowan E, Lewis J, disease. Cold Spring Harb Perspect Med.
Prihar G, Kim J, Dillmann WH, Browne SE, Hall 2011;1:a006189.
A, Voellmy R, Tsuboi Y, Dawson TM, Wolozin B, 172. Shimura H, Schwartz D, Gygi SP, Kosik KS. CHIP-
Hardy J, Hutton M.  CHIP and Hsp70 regulate tau Hsc70 complex ubiquitinates phosphorylated
ubiquitination, degradation and aggregation. Hum tau and enhances cell survival. J Biol Chem.
Mol Genet. 2004;13:703–14. 2004;279:4869–76.
160. Polis B, Gil-Henn H. Commentary on Giralt et al.: 173. Sotiropoulos I, Galas MC, Silva JM, Skoulakis
PTK2B/Pyk2 overexpression improves a mouse E, Wegmann S, Maina MB, Blum D, Sayas CL,
model of Alzheimer’s disease. Exp Neurol. 2018. Mandelkow EM, Mandelkow E, Spillantini MG,
166 I.-C. Stancu et al.

Sousa N, Avila J, Medina M, Mudher A, Buee 186. Vandebroek T, Vanhelmont T, Terwel D, Borghgraef
L.  Atypical, non-standard functions of the micro- P, Lemaire K, Snauwaert J, Wera S, Van Leuven F,
tubule associated tau protein. Acta Neuropathol Winderickx J. Identification and isolation of a hyper-
Commun. 2017;5:91. phosphorylated, conformationally changed inter-
174. Sottejeau Y, Bretteville A, Cantrelle FX, Malmanche mediate of human protein tau expressed in yeast.
N, Demiaute F, Mendes T, Delay C, Alves Dos Alves Biochemistry. 2005;44:11466–75.
H, Flaig A, Davies P, Dourlen P, Dermaut B, Laporte 187. Vanderweyde T, Apicco DJ, Youmans-Kidder K,
J, Amouyel P, Lippens G, Chapuis J, Landrieu I, Ash PEA, Cook C, Lummertz da Rocha E, Jansen-
Lambert JC.  Tau phosphorylation regulates the West K, Frame AA, Citro A, Leszyk JD, Ivanov
interaction between BIN1’s SH3 domain and Tau’s P, Abisambra JF, Steffen M, Li H, Petrucelli L,
proline-rich domain. Acta Neuropathol Commun. Wolozin B. Interaction of tau with the RNA-binding
2015;3:58. protein TIA1 regulates tau pathophysiology and tox-
175. Spillantini MG, Goedert M.  Tau protein pathology icity. Cell Rep. 2016;15:1455–66.
in neurodegenerative diseases. Trends Neurosci. 188. Velazquez R, Ferreira E, Tran A, Turner EC, Belfiore
1998;21:428–33. R, Branca C, Oddo S. Acute tau knockdown in the
176. Spillantini MG, Goedert M. Tau pathology and neu- hippocampus of adult mice causes learning and
rodegeneration. Lancet Neurol. 2013;12:609–22. memory deficits. Aging Cell. 2018; e12775.
177. Spittaels K, Van den Haute C, Van Dorpe J, Geerts H, 189. Violet M, Chauderlier A, Delattre L, Tardivel M,
Mercken M, Bruynseels K, Lasrado R, Vandezande Chouala MS, Sultan A, Marciniak E, Humez S,
K, Laenen I, Boon T, Van Lint J, Vandenheede J, Binder L, Kayed R, Lefebvre B, Bonnefoy E, Buee
Moechars D, Loos R, Van Leuven F. Glycogen syn- L, Galas MC.  Prefibrillar tau oligomers alter the
thase kinase-3beta phosphorylates protein tau and nucleic acid protective function of tau in hippocam-
rescues the axonopathy in the central nervous sys- pal neurons in vivo. Neurobiol Dis. 2015;82:540–51.
tem of human four-repeat tau transgenic mice. J Biol 190. Walker LC, Jucker M. Neurodegenerative diseases:
Chem. 2000;275:41340–9. expanding the prion concept. Annu Rev Neurosci.
178. Stamer K, Vogel R, Thies E, Mandelkow E, 2015;38:87–103.
Mandelkow EM.  Tau blocks traffic of organ- 191. Walker LC, Schelle J, Jucker M.  The prion-like
elles, neurofilaments, and APP vesicles in neu- properties of amyloid-beta assemblies: implications
rons and enhances oxidative stress. J Cell Biol. for Alzheimer’s Disease. Cold Spring Harb Perspect
2002;156:1051–63. Med. 2016;6
179. Stancu IC, Vasconcelos B, Terwel D, Dewachter 192. Wang Y, Mandelkow E.  Tau in physiology and
I.  Models of beta-amyloid induced tau-pathology: pathology. Nat Rev Neurosci. 2016;17:5–21.
the long and “folded” road to understand the mecha- 193. Wang P, Joberty G, Buist A, Vanoosthuyse A, Stancu
nism. Mol Neurodegener. 2014;9:51. IC, Vasconcelos B, Pierrot N, Faelth-Savitski M,
180. Stopschinski BE, Diamond MI. The prion model for Kienlen-Campard P, Octave JN, Bantscheff M,
progression and diversity of neurodegenerative dis- Drewes G, Moechars D, Dewachter I.  Tau inter-
eases. Lancet Neurol. 2017;16:323–32. actome mapping based identification of Otub1 as
181. Sultan A, Nesslany F, Violet M, Begard S, Loyens tau deubiquitinase involved in accumulation of
A, Talahari S, Mansuroglu Z, Marzin D, Sergeant pathological tau forms in  vitro and in  vivo. Acta
N, Humez S, Colin M, Bonnefoy E, Buee L, Galas Neuropathol. 2017;133:731–49.
MC. Nuclear tau, a key player in neuronal DNA pro- 194. Weingarten MD, Lockwood AH, Hwo SY,
tection. J Biol Chem. 2011;286:4566–75. Kirschner MW.  A protein factor essential for
182. Tai HC, Wang BY, Serrano-Pozo A, Frosch MP, microtubule assembly. Proc Natl Acad Sci U S A.
Spires-Jones TL, Hyman BT.  Frequent and sym- 1975;72:1858–62.
metric deposition of misfolded tau oligomers 195. Yu JZ, Rasenick MM. Tau associates with actin in dif-
within presynaptic and postsynaptic terminals in ferentiating PC12 cells. FASEB J. 2006;20:1452–61.
Alzheimer’s disease. Acta Neuropathol Commun. 196. Yuan A, Kumar A, Peterhoff C, Duff K, Nixon
2014;2:146. RA. Axonal transport rates in vivo are unaffected by
183. Takei Y, Teng J, Harada A, Hirokawa N. Defects in tau deletion or overexpression in mice. J Neurosci.
axonal elongation and neuronal migration in mice 2008;28:1682–7.
with disrupted tau and map1b genes. J Cell Biol. 197. Yuan A, Kumar A, Sasaki T, Duff K, Nixon
2000;150:989–1000. RA.  Global axonal transport rates are unaltered in
184. Terwel D, Muyllaert D, Dewachter I, Borghgraef P, htau mice in vivo. J Alzheimers Dis. 2013;37:579–86.
Croes S, Devijver H, Van Leuven F. Amyloid acti- 198. Zhou L, McInnes J, Wierda K, Holt M, Herrmann
vates GSK-3beta to aggravate neuronal tauopathy in AG, Jackson RJ, Wang YC, Swerts J, Beyens J,
bigenic mice. Am J Pathol. 2008;172:786–98. Miskiewicz K, Vilain S, Dewachter I, Moechars D,
185. van Slegtenhorst M, Lewis J, Hutton M. The molec- De Strooper B, Spires-Jones TL, De Wit J, Verstreken
ular genetics of the tauopathies. Exp Gerontol. P. Tau association with synaptic vesicles causes pre-
2000;35:461–71. synaptic dysfunction. Nat Commun. 2017;8:15295.
 Part III
Tau and Disease-Related Proteins
 Relationship Between Tau, β
Amyloid and α-Synuclein 14
Pathologies

Lauren Walker and Johannes Attems

Frequencies Increase with Age cases, age at death 77–87), comprehensive map-

ping of abnormal protein depositions reported all
It has been shown that up to 55% of older adults cases showed some degree of HP-T burden, with
that undergo post-mortem examination exhibit 68.7% of cases harboring Aβ pathology, and evi-
pathologies associated with more than one neuro- dence of α-synuclein deposits in 24.9% of cases
degenerative disease [22, 43]. The largest con- [23].
tributing factor for the prevalence of mixed
pathologies is age (Fig. 14.1).
The observation that multiple pathologies of  ixed Pathology vs Mixed
M
differing aetiologies co-exist in the same indi- Dementia
vidual brain has induced a number of cross-­
sectional studies to establish the prevalence of The consensus diagnosis of neurodegenerative
mixed pathologies in the ageing brain. Most diseases is based on the concordance of the clini-
notably, pathologies associated with Alzheimer’s cal symptoms observed during life and the signa-
disease (intracellular hyperphosphorylated tau ture pathological lesions present at post-mortem
(HP-T), and extracellular β amyloid (Aβ)), and examination, with the neuropathological element
Lewy body disease (intracellular inclusions of based on the most prevalent pathology, regardless
α-synuclein), which are the most common neuro- of the presence of other abnormal misfolded pro-
degenerative diseases known to cause dementia, teins. The term mixed or concomitant pathology
together accounting for up to 80% of neurode- describes further pathological changes in addi-
generative dementias [21]. Neuropathological tion to the primary pathological lesion and should
studies have also emphasised the presence of not be mistaken for mixed dementia, where the
pathological protein aggregates in the brains of burden of 2 or more pathological lesions is
asymptomatic patients who may represent the enough to fulfill the neuropathological criteria
prodromal phase of the disease and may account for more than one neurodegenerative disease (i.e.
for disease heterogeneity and often complex clin- NIA:AA criteria, high AD neuropathologic
ical presentations [22, 23]. Irrespective of clinical change, and neocortical Lewy body disease
diagnosis, in a community-based study (233 (termed mixed AD/LBD)) [3, 19, 33, 34, 51, 55].
Interestingly neuropathologically mixed AD/
L. Walker (*) · J. Attems LBD cases can present with differing clinical
Institute of Neuroscience, Newcastle University, phenotypes including AD, DLB, and Parkinson’s
Newcastle upon Tyne, UK disease dementia (PDD).
e-mail: Lauren.walker1@ncl.ac.uk

© Springer Nature Singapore Pte Ltd. 2019 169

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_14
170 L. Walker and J. Attems

Concomitant tau pathology in LBD Concomitant Aβ pathology in LBD

A cases (NFT Braak stages) B cases (Thal phases)
100 100
Braak I Thal 0
Braak II Thal 1

80 Braak III 80 Thal 2

Braak IV Thal 3
Prevalence

Prevalence
Thal 4
60 60
Thal 5

40 40

20 20

65-69 70-74 75-79 80-84 85-89 65-69 70-74 75-79 80-84 85-89
Age Age

Concomitant neuritic plaque pathology in D Concomitant α-synuclein pathology in

C LBD cases (CERAD scores) AD cases (McKeith criteria)
100 Negative 100
Negative
Sparse
Brainstem
Moderate
80 80 Limbic
Frequent
Neocortical
Prevalence
Prevalence

60 60

40 40

20 20

65-69 70-74 75-79 80-84 85-89 65-69 70-74 75-79 80-84 85-89
Age Age

Fig. 14.1  Concomitant pathologies increase in frequency pathology as determined by McKeith criteria [33] (d) was
with age. Age related concomitant tau pathology as strati- observed in n = 61 AD cases. Abbreviations: LBD Lewy
fied by NFT Braak stage [3] (a), Aβ pathology as stratified body disease, NFT neurofibrillary tangle, Aβ β amyloid,
by Thal phase [51] (b), and neuritic plaque pathology as CERAD Consortium to Establish a Registry for
measured by CERAD score [34] (c) was observed in Alzheimer’s disease, AD Alzheimer’s disease
n = 46 LBD cases. Whilst the prevalence of α-synuclein

 linical Implications of Mixed

C subjects were correctly identified as having
Dementia mixed AD/DLB, and by final diagnosis this
increased to just 23%, with the only distinguish-
The simultaneous presence of multiple patholo- ing clinical symptom to detect DLB in an AD
gies in the brain is a clinically diagnostic chal- context being complex visual hallucinations [52].
lenge, as the secondary pathology (which on its However, in another study, Savica and colleagues
own is enough to underlie clinical dementia) can found the overall score for UPDRS II and III and
often remain undetected. Several studies using individual components of routine neuropsycho-
neuropathologically confirmed mixed AD/LBD logical assessment, specifically the Auditory
cases have tried to address this; in a cohort of 22 Verbal Learning Test and Boston Naming Test,
mixed AD/LBD cases, at baseline diagnosis, no were able to detect Lewy body pathology in 54
14  Relationship Between Tau, β Amyloid and α-Synuclein Pathologies 171

pathologically confirmed AD cases [41]. Of note, in the entorhinal cortex and α-synuclein initially
it is important to detect the presence of LBs in affecting the medulla oblongata), clinical demen-
such cases to identify patients that will experi- tia in both AD and LBD is associated with the
ence a more rapid decline in cognition [24, 40, progression of both pathologies to the neocortex,
47, 55], and will have adverse effects to the with widespread affliction of many anatomically
administration of anti-psychotics [32]. connected brain regions [2, 3, 33]. Of note, the
hierarchical deposition of Aβ deposition initially
affects the neocortex, however Aβ is only associ-
 europathological Burden of Mixed
N ated with clinical AD when it has progressed to
Dementia the brain stem and cerebellum, and only in the
presence of neocortical HP-T pathology, which in
As neuropathologically mixed AD/LBD cases turn is only seen in the presence of Aβ [35, 51].
can present clinically as either AD or LBD, we Disentangling the underlying aetiology of cogni-
have previously investigated the burden of patho- tive impairment in the presence of multiple
logical protein aggregates (HP-T, Aβ, and pathologies is a complex undertaking, however
α-synuclein) to shed light on the temporal there are a plethora of studies investigating the
sequence of abnormal protein deposition. Using putative synergistic relationship between patho-
quantitative techniques mixed AD/DLB cases logical protein aggregates, and the combined
with a clinical diagnosis of AD exhibited a greater effect they have on clinical phenotype.
burden of HP-T pathology, and lower burden of
α-synuclein compared to mixed AD/LBD cases
with a PDD clinical phenotype [55]. This may HP-T and Aβ in AD
suggest that clinical phenotype represents the pri-
mary underlying pathology and initial cause of Before we address the putative relationships
the dementia. between pathological proteins in cases exhibit-
Another key finding from our study is that the ing mixed pathology/mixed dementia it is impor-
topographical distribution of pathological pro- tant to understand the neurobiology of individual
tein aggregates differs between distinct neuro- neurodegenerative diseases which is crucial for
pathological phenotypes; when comparing the development of effective treatments. Both
mixed AD/LBD cases to ‘pure’ DLB cases, the HP-T and Aβ are required for a neuropathologi-
mixed AD/LBD cases exhibited considerable cal diagnosis of AD, and the relationship
higher α-synuclein burden in the temporal cor- between these proteins has been scrutinized [7,
tex. The temporal cortex also harboured the 46, 53, 56].
highest HP-T loads, and hence HP-T may have However there has been growing interest in
promoted the aggregation and accumulation of post-translational modifications of Aβ, and one
α-synuclein [55]. species of particular interest is Aβ pyroglu-
tamylated at position 3 (Aβ3(pE)-40 and
Aβ3(pE)-42). Recent data suggests Aβ3(pE)-42
 utative Synergistic Relationships
P plays a crucial role in AD pathogenesis [1], and
Between Pathological Proteins where similar amounts of non-pyroglutamate
modified Aβ are seen in AD and controls (e.g.
Whether concomitant pathologies interact or isoaspartate Aβ), Aβ3(pE)-42 is more abundant,
independently affect clinical disease progression and accounts for up to 25% of total Aβ in AD
is uncertain, with several studies reporting con- [17, 36].
flicting results [4, 44]. Although the temporal A functional connection has been established
sequence of the pathological spread of HP-T and between Aβ3(pE)-42 and HP-T as Aβ oligomers
α-synuclein differ in the neurodegenerative dis- containing Aβ3(pE) have been suggested to initi-
eases they are associated with (HP-T originating ate tau-dependent cytotoxicity [38]. Furthermore,
172 L. Walker and J. Attems

a study using human post-mortem brain tissue is a topic of some debate [4, 44]. However, a
demonstrated an association between Aβ3(pE)-42 large study using statistical modeling to exten-
and HP-T, whilst no associated was observed sively analyse whether LB related pathology
between HP-T and non- Aβ3(pE)-42 [28]. reacts with AD related pathology revealed the
effect of additional pathologies on clinical dis-
ease progression was greater in those with inter-
HP-T and α-Synuclein mediate AD related pathology compared to
patients with high levels of AD related pathology,
As HP-T is frequently seen as a concomitant which could have implications in the targeted
pathology particularly in LBD, research to eluci- design of therapeutics [4].
date a putative interaction between HP-T and
α-synuclein is of great interest. In a seminal paper
in 2003, α-synuclein was shown to initiate the Aβ and α-Synuclein
polymerization of HP-T in vitro [13]. Supporting
evidence from animal models demonstrated HP-T The potential relationship between Aβ and
deposits accumulate in transgenic mice overex- α-synuclein has been studied [11, 29, 30]. Using
pressing human α-synuclein, which increases in post-mortem brain tissue from 40 PD cases and 20
an age-dependant manner [15, 26], whilst phos- controls Lashley and colleagues used a combina-
phorylated α-synuclein aggregates have been tion of semi-quantitative and morphometric
detected in the brain of rTg4510 mice that over- assessment to determine Aβ plaque load and LB
express human P301L mutant tau [50]. A poten- density, and found a significant correlation
tial seeding mechanism has been suggested as tau between Aβ plaque load and overall LB burden
oligomers derived from well-characterised pro- [25]. More recently Swirski and colleagues dem-
gressive supranuclear palsy cases (a 4-repeat onstrated insoluble levels of α-synuclein phos-
tauopathy) and complexes of brain-derived phorylated at serine 129 (α-synuclein pSer129)
α-synuclein/tau oligomers isolated from cases positively correlated with soluble and insoluble
with Parkinson’s disease (PD), enhanced endog- levels of Aβ in a cohort of LBD cases. In parallel
enous tau oligomer formation in Htau mice, par- experiments, exposure of SH-SY5Y cells trans-
allel to increasing cell loss [5]. At the cellular fected with wild-type α-synuclein cDNA to aggre-
level, the co-occurrence of HP-T and α-synuclein gated Aβ42 significantly increased α-synuclein
has been shown using double immunofluores- pSer129, suggesting the levels of both proteins are
cence [10, 20]. closely related to each other [49].
In addition to exacerbating the aggregation
properties of each other, the presence of AD
related pathology has been demonstrated to HP-T, Aβ, and α-Synuclein – A Toxic
accelerate the clinical progression of LBD. In a Triad?
large clinicopathological correlative study Irwin
and colleagues reveal only HP-T neurofibrillary It is essential to examine the relationships
tangle pathology independently predicted sur- between individual pathological protein aggre-
vival time and a shorter interval between motor gates, and the cumulative affect they exert on
symptoms and the onset of dementia [19]. This clinical symptoms. However, an important caveat
corroborates findings of a PET imaging study, of such studies is that in vivo these pathological
which demonstrated a high frequency of cortical proteins do not exist in isolation, and this does
HP-T pathology in the brains of patients with not accurately reflect the mixed pathology load
LBD, which associated with cognitive decline in frequently observed in the brains of the demented
these patients [14]. Whether pathological pro- elderly, in particular LBDs. This limitation has
teins of differing aetiologies act individually or been addressed by the generation of a mouse
synergistically to alter the clinical disease course model that expresses HP-T, Aβ, and α-synuclein
14  Relationship Between Tau, β Amyloid and α-Synuclein Pathologies 173

pathologies by crossing 3xTg-AD mice with Indirect Interactions

A53T mice overexpressing human α-synuclein.
The resulting AD-DLB mice exhibited all 3 path- In addition to hallmark pathological lesions
ological lesions at a higher load, and an acceler- observed in AD and LBD, extensive synapse loss
ated cognitive decline compared to single in the neocortex is associated with the clinical
transgenic animals [9]. Studies that fully recapit- and pathological profile of both neurodegenera-
ulate these findings in human brains are conflict- tive disorders. Tau neuropathology is associated
ing; neuropathologically mixed AD/LBD cases with a loss of functional synapses in a mouse
experience an accelerated rate of cognitive model of AD [42], and Masliah and colleagues
decline compared to AD and LBD cases without demonstrated a 45% decrease of pre-synaptic
significant concomitant pathology [55], and a boutons in AD cases [31]. Accumulation of
combined pathology score of HP-T, Aβ, and pathologic α-synuclein has also been shown to
α-synuclein being a major determining factor in lead to selective decreases in synaptic proteins
the development of dementia in a cohort of LBD [54]. These data suggest vulnerable synpases
cases [18]. However, in a separate cohort of LBD may be more susceptible to further insults in the
cases a correlation between Aβ, and α-synuclein presence of both HP-T, and α-synuclein.
was observed, but not between HP-T and Phosphorylation of tau and α-synuclein are
α-synuclein [10]. Future studies in large, clini- key events in the pathogenesis of AD and LBD,
cally characterised cohorts are warranted to elu- and numerous kinases have been implicated in
cidate the mechanistic link between multiple the phosphorylation of both proteins including
pathologies. the tyrosine kinase Fyn, and casein kinase 1 [27,
37, 39, 48]. α-synuclein has also been shown to
contribute to glycogen synthase-3β mediated
Potential Mechanisms phosphorylation of tau [12].
for Interaction

Direct Interactions Stratification for Clinical Trials

In healthy neurons HP-T, Aβ, and α-synuclein do Identifying the presence of multiple pathological
not exist in the same subcellular compartment lesions in human subjects will have important
therefore limiting the potential for a direct inter- implications for the design of effective therapeu-
action. However, in neuropathologic states they tic targets for neurodegenerative diseases. Firstly,
have been found in mitochondria, lysosomes, and it will allow for the correct stratification of
autophagosomes [6, 8, 16], although whether the patients into groups that have similar baseline
proteins are co-localized is yet to be established. characteristics, which will strengthen the out-
An emerging body of evidence is pointing to come and confirm endpoints are attributed to any
the oligomeric forms of HP-T, and α-synuclein as potential drug of interest without any confound-
being the major culprits in a putative interaction ing factors that may exacerbate the disease pro-
between the pathologic proteins, with one recent cess. Secondly, due to the heterogeneous nature
study using novel conformational antibodies of neurodegenerative diseases, identification of
demonstrating toxic oligomeric species of HP-T, concomitant pathologies (even if clinically
and α-synuclein can accumulate in LBD [45]. silent), and elucidation of basic molecular mech-
Another study by Castillo-Carranza and col- anisms governing the deposition of hallmark
leagues further supports this proposed mechanis- pathologies may allow the development of per-
tic link, as complexes of α-synuclein/tau sonalised treatment options for patients affected
oligomers enhanced tau oligomer formation with by neurodegenerative diseases associated with
associated neuronal loss [5]. clinical dementia.
174 L. Walker and J. Attems

Acknowledgments  Tissue for this study which was pro- in dementia with Lewy bodies. J  Neuropathol Exp
vided by the Newcastle Brain Tissue Resource, which is Neurol. 2013;72:1203–12.
funded in part by a grant from the UK Medical Research 11. Compta Y, Parkkinen L, O'Sullivan SS, Vandrovcova
Council (Grant Number G0400074) and by Brains for J, Holton JL, Collins C, Lashley T, Kallis C, Williams
Dementia research, a joint venture between Alzheimer’s DR, De Silva R, Lees AJ, Revesz T.  Lewy- and
Society and Alzheimer’s Research UK. Alzheimer-type pathologies in Parkinson’s dis-
ease dementia: which is more important? Brain.
2011;134:1493–505.
12. Duka T, Duka V, Joyce JN, Sidhu A. Alpha-Synuclein
References contributes to GSK-3beta-catalyzed Tau phos-
phorylation in Parkinson's disease models. FASEB
1. Acero G, Manoutcharian K, Vasilevko V, Munguia J. 2009;23:2820–30.
ME, Govezensky T, Coronas G, Luz-Madrigal A, 13. Giasson BI, Forman MS, Higuchi M, Golbe LI,

Cribbs DH, Gevorkian G.  Immunodominant epi- Graves CL, Kotzbauer PT, Trojanowski JQ, Lee
tope and properties of pyroglutamate-modified VM. Initiation and synergistic fibrillization of tau and
Abeta-specific antibodies produced in rabbits. alpha-synuclein. Science. 2003;300:636–40.
J Neuroimmunol. 2009;213:39–46. 14. Gomperts SN, Locascio JJ, Makaretz SJ, Schultz

2. Braak H, Alafuzoff I, Arzberger T, Kretzschmar H, Del A, Caso C, Vasdev N, Sperling R, Growdon JH,
Tredici K.  Staging of Alzheimer disease-­associated Dickerson BC, Johnson K.  Tau Positron Emission
neurofibrillary pathology using paraffin sections Tomographic Imaging in the Lewy Body Diseases.
and immunocytochemistry. Acta Neuropathol. JAMA Neurol. 2016;73:1334–41.
2006;112:389–404. 15. Haggerty T, Credle J, Rodriguez O, Wills J, Oaks
3. Braak H, Braak E.  Neuropathological stageing of AW, Masliah E, Sidhu A.  Hyperphosphorylated
Alzheimer-related changes. Acta Neuropathol. Tau in an alpha-synuclein-overexpressing trans-
1991;82:239–59. genic model of Parkinson’s disease. Eur J Neurosci.
4. Brenowitz WD, Hubbard RA, Keene CD, Hawes SE, 2011;33:1598–610.
Longstreth WTJR, Woltjer RL, Kukull WA.  Mixed 16. Hansson Petersen CA, Alikhani N, Behbahani H,

neuropathologies and estimated rates of clinical Wiehager B, Pavlov PF, Alafuzoff I, Leinonen V,
progression in a large autopsy sample. Alzheimers Ito A, Winblad B, Glaser E, Ankarcrona M.  The
Dement. 2017;13:654–62. amyloid beta-peptide is imported into mitochon-
5. Castillo-Carranza DL, Guerrero-Munoz MJ, Sengupta dria via the TOM import machinery and localized
U, Gerson JE, Kayed R. alpha-synuclein oligomers to mitochondrial cristae. Proc Natl Acad Sci U S A.
induce a unique toxic tau strain. Biol Psychiatry. 2008;105:13145–50.
2018;84:499–508. 17. Harigaya Y, Saido TC, Eckman CB, Prada CM, Shoji
6. Chinta SJ, Mallajosyula JK, Rane A, Andersen M, Younkin SG. Amyloid beta protein starting pyro-
JK.  Mitochondrial alpha-synuclein accumulation glutamate at position 3 is a major component of the
impairs complex I function in dopaminergic neurons amyloid deposits in the Alzheimer’s disease brain.
and results in increased mitophagy in vivo. Neurosci Biochem Biophys Res Commun. 2000;276:422–7.
Lett. 2010;486:235–9. 18. Howlett DR, Whitfield D, Johnson M, Attems J,

7. Choi SH, Kim YH, Hebisch M, Sliwinski C, Lee S, O'Brien JT, Aarsland D, Lai MK, Lee JH, Chen C,
D'Avanzo C, Chen H, Hooli B, Asselin C, Muffat Ballard C, Hortobagyi T, Francis PT.  Regional mul-
J, Klee JB, Zhang C, Wainger BJ, Peitz M, Kovacs tiple pathology scores are associated with cogni-
DM, Woolf CJ, Wagner SL, Tanzi RE, Kim DY.  A tive decline in Lewy body dementias. Brain Pathol.
three-dimensional human neural cell culture model of 2014;25(4):401–8. [Epub ahead of print]
Alzheimer’s disease. Nature. 2014;515:274–8. 19. Irwin DJ, Grossman M, Weintraub D, Hurtig HI, Duda
8. Cieri D, Vicario M, Vallese F, D'Orsi B, Berto P, JE, Xie SX, Lee EB, Van Deerlin VM, Lopez OL,
Grinzato A, Catoni C, De Stefani D, Rizzuto R, Kofler JK, Nelson PT, Jicha GA, Woltjer R, Quinn
Brini M, Cali T.  Tau localises within mitochondrial JF, Kaye J, Leverenz JB, Tsuang D, Longfellow K,
sub-compartments and its caspase cleavage affects Yearout D, Kukull W, Keene CD, Montine TJ,
ER-mitochondria interactions and cellular Ca(2+) Zabetian CP, Trojanowski JQ. Neuropathological and
handling. Biochim Biophys Acta Mol basis Dis. genetic correlates of survival and dementia onset in
2018;1864(10):3247–56. synucleinopathies: a retrospective analysis. Lancet
9. Clinton LK, Blurton-Jones M, Myczek K, Neurol. 2017;16:55–65.
Trojanowski JQ, Laferla FM. Synergistic Interactions 20. Ishizawa T, Mattila P, Davies P, Wang D, Dickson
between Abeta, tau, and alpha-synuclein: acceleration DW.  Colocalization of tau and alpha-synuclein epi-
of neuropathology and cognitive decline. J Neurosci. topes in Lewy bodies. J  Neuropathol Exp Neurol.
2010;30:7281–9. 2003;62:389–97.
10. Colom-Cadena M, Gelpi E, Charif S, Belbin O, Blesa 21. Kane JPM, Surendranathan A, Bentley A, Barker

R, Marti MJ, Clarimon J, Lleo A.  Confluence of SAH, Taylor JP, Thomas AJ, Allan LM, Mcnally RJ,
alpha-synuclein, tau, and beta-amyloid pathologies James PW, Mckeith IG, Burn DJ, O'Brien JT. Clinical
14  Relationship Between Tau, β Amyloid and α-Synuclein Pathologies 175

prevalence of Lewy body dementia. Alzheimers Res Plotkin A, Duda JE, El-Agnaf O, Feldman H, Ferman
Ther. 2018;10:19. TJ, Ffytche D, Fujishiro H, Galasko D, Goldman JG,
22. Kovacs GG, Alafuzoff I, Al-Sarraj S, Arzberger T, Gomperts SN, Graff-Radford NR, Honig LS, Iranzo
Bogdanovic N, Capellari S, Ferrer I, Gelpi E, Kovari A, Kantarci K, Kaufer D, Kukull W, Lee VMY,
V, Kretzschmar H, Nagy Z, Parchi P, Seilhean D, Leverenz JB, Lewis S, Lippa C, Lunde A, Masellis
Soininen H, Troakes C, Budka H.  Mixed brain M, Masliah E, McLean P, Mollenhauer B, Montine
pathologies in dementia: the BrainNet Europe con- TJ, Moreno E, Mori E, Murray M, O'Brien JT, Orimo
sortium experience. Dement Geriatr Cogn Disord. S, Postuma RB, Ramaswamy S, Ross OA, Salmon
2008;26:343–50. DP, Singleton A, Taylor A, Thomas A, Tiraboschi
23. Kovacs GG, Milenkovic I, Wohrer A, Hoftberger
P, Toledo JB, Trojanowski JQ, Tsuang D, Walker Z,
R, Gelpi E, Haberler C, Honigschnabl S, Reiner-­ Yamada M, Kosaka K. Diagnosis and management of
Concin A, Heinzl H, Jungwirth S, Krampla W, dementia with Lewy bodies: fourth consensus report
Fischer P, Budka H. Non-Alzheimer neurodegenera- of the DLB Consortium. Neurology. 2017;89:88–100.
tive pathologies and their combinations are more fre- 34. Mirra SS, Heyman A, McKeel D, Sumi SM, Crain
quent than commonly believed in the elderly brain: a BJ, Brownlee LM, Vogel FS, Hughes JP, Van
community-based autopsy series. Acta Neuropathol. Belle G, Berg L.  The consortium to establish a
2013;126:365–84. registry for Alzheimer’s disease (CERAD). Part
24. Kraybill ML, Larson EB, Tsuang DW, Teri L,
II. Standardization of the neuropathologic assessment
McCormick WC, Bowen JD, Kukull WA, Leverenz of Alzheimer’s disease. Neurology. 1991;41:479–86.
JB, Cherrier MM. Cognitive differences in dementia 35. Montine T, Phelps C, Beach T, Bigio E, Cairns N,
patients with autopsy-verified AD, Lewy body pathol- Dickson D, Duyckaerts C, Frosch M, Masliah E,
ogy, or both. Neurology. 2005;64:2069–73. Mirra S, Nelson P, Schneider J, Thal D, Trojanowski
25. Lashley T, Holton JL, Gray E, Kirkham K, O'Sullivan J, Vinters H, Hyman B. National Institute on Aging–
SS, Hilbig A, Wood NW, Lees AJ, Revesz T. Cortical Alzheimer’s Association guidelines for the neuro-
alpha-synuclein load is associated with amyloid-­ pathologic assessment of Alzheimer’s disease: a
beta plaque burden in a subset of Parkinson's disease practical approach. Acta Neuropathol. 2012;123:1–11.
patients. Acta Neuropathol. 2008;115:417–25. 36. Moro ML, Phillips AS, Gaimster K, Paul C, Mudher
26. Lee VM, Giasson BI, Trojanowski JQ. More than just A, Nicoll JAR, Boche D. Pyroglutamate and isoaspar-
two peas in a pod: common amyloidogenic proper- tate modified amyloid-beta in ageing and Alzheimer’s
ties of tau and alpha-synuclein in neurodegenerative disease. Acta Neuropathol Commun. 2018;6:3.
diseases. Trends Neurosci. 2004;27:129–34. 37. Nakamura T, Yamashita H, Takahashi T, Nakamura
27. Li G, Yin H, Kuret J. Casein kinase 1 delta phosphor- S.  Activated Fyn phosphorylates alpha-synuclein at
ylates tau and disrupts its binding to microtubules. tyrosine residue 125. Biochem Biophys Res Commun.
J Biol Chem. 2004;279:15938–45. 2001;280:1085–92.
28. Mandler M, Walker L, Santic R, Hanson P, Upadhaya 38. Nussbaum JM, Schilling S, Cynis H, Silva A, Swanson
AR, Colloby SJ, Morris CM, Thal DR, Thomas E, Wangsanut T, Tayler K, Wiltgen B, Hatami A,
AJ, Schneeberger A, Attems J.  Pyroglutamylated Ronicke R, Reymann K, Hutter-Paier B, Alexandru A,
amyloid-beta is associated with hyperphosphory- Jagla W, Graubner S, Glabe CG, Demuth HU, Bloom
lated tau and severity of Alzheimer’s disease. Acta GS.  Prion-like behaviour and tau-dependent cyto-
Neuropathol. 2014;128:67–79. toxicity of pyroglutamylated amyloid-beta. Nature.
29. Marsh SE, Blurton-Jones M.  Examining the mecha- 2012;485:651–5.
nisms that link beta-amyloid and alpha-synuclein 39. Okochi M, Walter J, Koyama A, Nakajo S, Baba M,
pathologies. Alzheimers Res Ther. 2012;4:11. Iwatsubo T, Meijer L, Kahle PJ, Haass C. Constitutive
30. Masliah E, Rockenstein E, Veinbergs I, Sagara Y,
phosphorylation of the Parkinson’s disease associated
Mallory M, Hashimoto M, Mucke L. beta-amyloid alpha-synuclein. J Biol Chem. 2000;275:390–7.
peptides enhance alpha-synuclein accumulation and 40. Olichney JM, Galasko D, Salmon DP, Hofstetter CR,
neuronal deficits in a transgenic mouse model link- Hansen LA, Katzman R, Thal LJ. Cognitive decline is
ing Alzheimer's disease and Parkinson's disease. Proc faster in Lewy body variant than in Alzheimer’s dis-
Natl Acad Sci U S A. 2001;98:12245–50. ease. Neurology. 1998;51:351–7.
31.
Masliah E, Terry RD, Deteresa RM, Hansen 41. Savica R, Beach TG, Hentz JG, Sabbagh MN, Serrano
LA.  Immunohistochemical quantification of the GE, Sue LI, Dugger BN, Shill HA, Driver-Dunckley
synapse-­related protein synaptophysin in Alzheimer E, Caviness JN, Mehta SH, Jacobson SA, Belden CM,
disease. Neurosci Lett. 1989;103:234–9. Davis KJ, zamrini E, Shprecher DR, Adler CH. Lewy
32. Mckeith I, Fairbairn A, Perry R, Thompson P, Perry body pathology in Alzheimer’s disease: a clinicopath-
E.  Neuroleptic sensitivity in patients with senile ological prospective study. Acta Neurol Scand. 2018;
dementia of Lewy body type. BMJ. 1992;305:673–8. https://doi.org/10.1111/ane.13028.
33. Mckeith IG, Boeve BF, Dickson DW, Halliday G, 42. Schindowski K, Bretteville A, Leroy K, Begard S,
Taylor JP, Weintraub D, Aarsland D, Galvin J, Attems Brion JP, Hamdane M, Buee L. Alzheimer’s disease-­
J, Ballard CG, Bayston A, Beach TG, Blanc F, Bohnen like tau neuropathology leads to memory deficits and
N, Bonanni L, Bras J, Brundin P, Burn D, Chen-­ loss of functional synapses in a novel mutated tau
176 L. Walker and J. Attems

transgenic mouse without any motor deficits. Am 50. Takaichi Y, Ano Y, Chambers JK, Uchida K,

J Pathol. 2006;169:599–616. Takashima A, Nakayama H.  Deposition of phos-
43. Schneider JA, Arvanitakis Z, Bang W, Bennett
phorylated alpha-­ synuclein in the rTg4510 mouse
DA.  Mixed brain pathologies account for most model of tauopathy. J  Neuropathol Exp Neurol.
dementia cases in community-dwelling older persons. 2018;77:920–8.
Neurology. 2007;69:2197–204. 51. Thal DR, Rub U, Orantes M, Braak H. Phases of A beta-
44. Schneider JA, Arvanitakis Z, Yu L, Boyle PA, Leurgans deposition in the human brain and its relevance for the
SE, Bennett DA. Cognitive impairment, decline and development of AD. Neurology. 2002;58:1791–800.
fluctuations in older community-­ dwelling subjects 52. Thomas AJ, Saidi M, Mahin-Babaei F, Lett DJ,

with Lewy bodies. Brain. 2012;135:3005–14. Taylor JP, Walker L, Attems J.  Improving the diag-
45. Sengupta U, Guerrero-Munoz MJ, Castillo-Carranza nosis of dementia with Lewy bodies in the context of
DL, Lasagna-Reeves CA, Gerson JE, Paulucci-­ an Alzheimer’s type dementia. Alzheimers Res Ther.
Holthauzen AA, Krishnamurthy S, Farhed M, Jackson 2018;10(1):27.
GR, Kayed R. Pathological interface between oligo- 53. Umeda T, Maekawa S, Kimura T, Takashima A,

meric alpha-synuclein and tau in synucleinopathies. Tomiyama T, Mori H. Neurofibrillary tangle formation
Biol Psychiatry. 2015;78:672–83. by introducing wild-type human tau into APP trans-
46. Sepulcre J, Schultz AP, Sabuncu M, Gomez-Isla T, genic mice. Acta Neuropathol. 2014;127(5):685–98.
Chhatwal J. In vivo tau, amyloid, and gray matter pro- 54. Volpicelli-Daley LA, Luk KC, Patel TP, Tanik SA,
files in the aging brain. J Neurosci. 2016;36:7364–74. Riddle DM, Stieber A, Meaney DF, Trojanowski JQ,
47. Serby M, Brickman AM, Haroutunian V, Purohit DP, Lee VM.  Exogenous alpha-synuclein fibrils induce
Marin D, Lantz M, Mohs RC, Davis KL.  Cognitive Lewy body pathology leading to synaptic dysfunction
burden and excess Lewy-body pathology in the Lewy-­ and neuron death. Neuron. 2011;72:57–71.
body variant of Alzheimer disease. Am J  Geriatr 55. Walker L, Mcaleese KE, Thomas AJ, Johnson M,
Psychiatry. 2003;11:371–4. Martin-Ruiz C, Parker C, Colloby SJ, Jellinger K,
48. Shirazi SK, Wood JG.  The protein tyrosine kinase, Attems J.  Neuropathologically mixed Alzheimer’s
fyn, in Alzheimer’s disease pathology. Neuroreport. and Lewy body disease: burden of pathological pro-
1993;4:435–7. tein aggregates differs between clinical phenotypes.
49. Swirski M, Miners JS, De Silva R, Lashley T, Ling Acta Neuropathol. 2015;129:729–48.
H, Holton J, Revesz T, Love S.  Evaluating the rela- 56. Zempel H, Thies E, Mandelkow E, Mandelkow

tionship between amyloid-beta and alpha-synuclein EM.  Abeta oligomers cause localized Ca(2+) eleva-
phosphorylated at Ser129  in dementia with Lewy tion, missorting of endogenous Tau into dendrites,
bodies and Parkinson’s disease. Alzheimers Res Ther. Tau phosphorylation, and destruction of microtubules
2014;6:77. and spines. J Neurosci. 2010;30:11938–50.
 Associations Between APOE
Variants, Tau and α-Synuclein 15
Elena Rodriguez-Vieitez and Henrietta M. Nielsen

Introduction folded protein is associated to a higher frequency

of other types of protein aggregates. For exam-
A common feature of neurodegenerative diseases ple, in a transgenic mouse model of AD, it was
is the aggregation and deposition of misfolded shown that Aβ, tau, and α-syn interact in vivo,
proteins in the brain. The two most common neu- leading to a synergistic accumulation of brain
rodegenerative diseases leading to dementia are pathology and cognitive dysfunction [5]. These
Alzheimer’s disease (AD), neuropathologically findings add support to the hypothesis that the
defined by the presence of amyloid-β (Aβ) and presence of co-pathologies may be due to the
hyperphosphorylated tau aggregates as hallmark seeding or template effect of one type of mis-
pathologies, and dementia with Lewy bodies folded protein that promotes the aggregation of
(DLB) characterized by α-synuclein (α-syn) other types of protein. In addition, the results
inclusions. Other synucleinopathies include suggest the possibility of common biological
Parkinson’s disease (PD) and Parkinson’s disease mechanisms partly due to shared genetic factors
with dementia (PDD). While each disease is influencing the development of these diseases.
defined by either tau or α-syn as hallmarks, the The ε4 allele of the apolipoprotein (APOE)
presence of co-pathologies is very frequent, gene is the most common genetic risk factor for
which leads to a large degree of pathological sporadic AD, and it was also found to increase
overlap among diseases [1, 2]. the risk of DLB, as well as increasing the likeli-
In particular, an overlap of Aβ, tau and α-syn hood of AD pathology in healthy aging [6].
is found in both AD and DLB, which may explain Across many neurodegenerative diseases, the
the observed commonalities in their phenotypic presence of APOE ε4 allele is, together with
clinical presentations, making differential diag- age, a factor that increases the frequency and
nosis challenging [3, 4]. There is also evidence burden of co-pathologies [1], suggesting that
that Aβ, tau and α-syn may act synergistically APOE ε4 may be a common genetic factor
and that a greater brain load of one type of mis- implicated in the biological mechanisms under-
lying and linking the different proteinopathies.
E. Rodriguez-Vieitez However, tauopathies and synucleinopathies
Department of Neurobiology, Care Sciences and also have differential patterns of co-pathologi-
Society, Karolinska Institutet, Stockholm, Sweden cal burden, suggesting that tau and α-syn strains
H. M. Nielsen (*) may have a distinct influence on the presence of
Department of Biochemistry and Biophysics, co-pathologies.
Stockholm University, Stockholm, Sweden
e-mail: henrietta.nielsen@dbb.su.se

© Springer Nature Singapore Pte Ltd. 2019 177

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_15
178 E. Rodriguez-Vieitez and H. M. Nielsen

The apolipoprotein E (apoE) protein, coded APOE ε4 enable six possible genotypes: ε2/ε2,
by the APOE gene, is involved in lipid transport, ε2/ε3, ε2/ε4, ε3/ε3, ε3/ε4 and ε4/ε4. The frequen-
as well as in processes related to Aβ aggregation cies of the APOE alleles in the general population
and clearance in the brain, and is abundantly vary slightly between different populations in the
expressed in astrocytes. Several apoE isoforms world, however the APOE ε3 is the most common
with major structural differences have been with a frequency of between 54 and 91% across
shown to influence brain lipid transport, neuroin- populations and the APOE ε2 is the least common
flammation, glucose metabolism, neuronal sig- variant with frequencies of nearly zero in native
naling and mitochondrial function in various Americans to 15% in Papuans [8]. The APOE ε4
ways. A large body of evidence supports the exis- allele with a frequency of 9–41% across popula-
tence of  links between APOE ε4 and brain Aβ tions [8], is considered the ancestral APOE vari-
burden, in particular APOE ε4 was reported to ant which was proposed to initially have ‘acted’
promote Aβ fibrillization directly, but it may like the APOE ε3 variant and that due to changes
also  contribute to AD pathogenesis via in diet and physical activity evolved to APOE ε2
Aβ-independent mechanisms including tau phos- and APOE ε3 by some 200,000 years ago [9]. The
phorylation or alterations in α-syn processing. APOE ε4 variant with the amino acid sequence
There is also evidence that neuroinflammation, positions 112 and 158 occupied by Arg is the only
increasingly recognized as an early and key variant to be found in other mammals like mice,
mechanism involved in the pathogenesis of many rats, orangutans, chimps and gorillas [10] and the
neurodegenerative diseases, may underlie the evolutionary development leading to APOE ε2
links between APOE ε4 and proteinopathies [7]. and APOE ε3  in humans was suggested to be
This chapter describes current experimental linked to the increased lifespan in humans com-
evidence on the relationships between APOE pared to other mammals [9].
variants, tau and α-syn, from clinical studies on Apolipoprotein E does not cross the blood
fluid biomarkers and positron emission tomogra- brain barrier and it is produced mainly by the
phy (PET) imaging, and from basic experimental liver in the periphery and by glial cells in the cen-
studies using cellular/molecular biology and ani- tral nervous system [11–13]. The apoE in the
mal models. The experimental data summarized plasma differs slightly from the one found in
here motivates further research using multimodal cerebrospinal fluid (CSF) due to less sialylation
imaging, fluid biomarkers and genetic factors [14]. In both compartments apoE serves as an
towards understanding the biological mecha- important lipid carrier which, enabled by its
nisms underlying these proteinopathies, and to lipid-binding domain, carries cholesterol and
contribute to differential diagnosis and patient through interactions with receptors of the low-­
stratification for clinical trials. density lipoprotein receptor (LDLR) family con-
tributes to the cholesterol homeostasis [15].
Isoform-specific differences in the capacity to
 POE and the Gene Product
A carry lipids have been reported with apoE4 pro-
Apolipoprotein E posed to be hypolipidated [15].
In humans, the fluid levels of apoE in plasma
In humans the APOE gene is polymorphic and and CSF are relatively high (μg/mL range) with
gives rise to three major isoforms of the 299 ten times higher concentrations found in the
amino acid apolipoprotein E (apoE); apoE2, plasma. The turnover rate of apoE4 is higher than
apoE3 and apoE4 which vary by only one or two those of apoE3 and apoE2 in plasma [16–18] and
amino acid residues at positions 112 and 158 carriers of the APOE ε4 variant are known to
(apoE2: Cys112, Cys158, apoE3: Cys112, Arg158 exhibit low plasma apoE levels [19] which may
and apoE4: Arg112, Arg158). The three alleles contribute to the increased risk of neurodegenera-
encoding these variants, APOE ε2, APOE ε3 and tive disease in these individuals [20].
15  Associations Between APOE Variants, Tau and α-Synuclein 179

 POE Variants and the Risk

A enzyme-linked immunosorbent assays (ELISAs)
of Tauopathy and Synucleinopathy have similar affinity for the different isoforms
and only one of the isoforms is normally used to
The APOE ε4 variant is to date the strongest produce a standard curve. The technical difficul-
described genetic risk factor for sporadic onset of ties to quantify apoE, which furthermore differs
AD and DLB, and it increases the likelihood of slightly in post-translational modifications
disease up to 15 versus sixfold respectively. In between the central nervous system compartment
contrast, the APOE ε2 allele appears protective and the periphery, are evident from the discrep-
against both AD and DLB [21]. It was further ancy in reported apoE levels [27–31]. Recently,
proposed that APOE ε4 increases the risk and mass-spectrometry based approaches have there-
lowers the age of onset for PD [22] and that it fore been employed to in an unbiased manner
increases the risk of dementia in patients with quantify not only total apoE levels but also the
pure synucleinopathies [23]. Other studies how- individual apoE isoform levels in both plasma
ever found no associations between APOE and and CSF [32]. Indeed, previous studies reported
PD [24], hence more studies are needed to estab- very poor correlation between mass-­spectrometry
lish APOE ε4 as a risk factor for PD. One study generated apoE fluid levels and data generated by
to date in which a potential association between assays like ELISA (R2 = 0.22 and 0.59) [27, 32]
the presence of APOE ε4 and multiple system and although stronger also the correlation
atrophy (MSA) was investigated reported no sig- between mass-spectrometry data and data from
nificant association between APOE ε4 and the rules based medicine (RBM) was insufficient
disease, a synucleinopathy affecting mainly oli- (R2  =  0.79) [27]. Data generated from mass-­
godendrocytes [25]. spectrometry assays showed that apoE levels in
Whereas the APOE ε2 allele appears to be plasma are directly associated with the APOE
protective against both AD and DLB, a recent genotype in which APOE ε4-carriers exhibited a
study proposed a novel link between APOE ε2 strong apoE plasma deficiency attributed specifi-
homozygosity and the brain load of tau pathol- cally to low levels of the apoE4 isoform, as deter-
ogy in subjects with progressive supranuclear mined in APOE heterozygous individuals [19,
palsy (PSP). The same study further demon- 33]. Cerebrospinal fluid levels of apoE appeared
strated an increased risk of the primary tauopa- to not differ amongst the APOE genotypes and
thies PSP and corticobasal degeneration (CBD) also not amongst AD and non-AD patients [19].
in homozygous APOE ε2 carriers [26]. The rele- Studies in which individual apoE isoforms
vance of the different APOE variants to the risk were quantified enabled the investigation of
of frontotemporal lobe dementia however potential correlations between individual apoE
remains unclear. Taken together, so far the APOE isoform fluid levels and concentrations of neuro-
ε2 and ε4 variants have individually been associ- degenerative disease-associated biomarkers
ated with the risk of diseases involving Aβ, tau including fluid levels of tau, hyperphosphory-
and α-syn pathology. lated tau (p-tau), Aβ peptides, and α-syn.
Cerebrospinal fluid levels of the apoE3 were
found to be positively associated with both total
 vidence from Clinical Studies
E tau (t-tau) and p(Thr181)-tau levels in APOE ε3/
Linking Apolipoprotein E to Tau ε3 and APOE ε3/ε4 individuals whereas these
and α-Synuclein Fluid Levels associations were absent for the fluid levels of
apoE4  in APOE ε3/ε4 subjects [19]. The direc-
Fluid levels of apoE have long been controversial tion of the observed correlations suggests that
since conventional immuno-based assays fail to higher levels of CSF t-tau and p-tau were paral-
distinguish between the apoE isoforms. Total leled by higher apoE in the CSF. Similar results
apoE levels have frequently been reported assum- were reported by Toledo and colleagues who fur-
ing that the antibody-pairs in for example ther speculated that higher CSF levels of at least
180 E. Rodriguez-Vieitez and H. M. Nielsen

apoE2 and apoE3 may be due to a protective jective memory complainers, in addition to a posi-
response to neurodegenerative processes associ- tive association between CSF α-syn concentrations
ated with AD [34]. Interestingly, a study that uti- and brain Aβ deposition, a positive association
lized transcranial magnetic stimulation over the was also reported between CSF α-syn and both
primary motor cortex to assess long-term-­ CSF t-tau and p(Thr181)-tau concentrations [37]
potentiation (LTP)-like cortical plasticity in AD suggesting an involvement of CSF α-syn in AD
patients with or without the APOE ε4 allele dem- pathophysiological mechanisms in preclinical
onstrated higher CSF t-tau levels in the APOE stages of the disease; the possible influence of
ε4-carriers versus the APOE ε3-carriers [35]. The APOE ε4 was however not investigated. In spo-
study further showed that only in APOE radic MCI and AD groups, a positive association
ε4-carrying AD patients CSF tau levels were cor- was observed between higher CSF levels of α-syn
related with impaired cortical plasticity, cogni- and both t-tau and p-tau [36], as illustrated in
tive decline and astrocyte survival. Whether the Fig. 15.1d, e; this type of positive association was
observations were related to the fluid levels of also reported in familial ADAD [36].
apoE4 was not assessed. Many studies in different diagnostic groups
Correlations between fluid levels of apoE iso- have found a relationship between APOE ε4 and
forms and α-syn remain to be investigated. low concentrations of CSF Aβ42 [38], especially
However it was recently demonstrated that CSF in cognitively normal individuals [39] or early
levels of α-syn were dose-dependently associated stages of AD, but less evident at later stages of
with the APOE ε4 allele in patients with sporadic the disease. The connection between APOE ε4
mild cognitive impairment (MCI) and who within and tau levels seems weaker with some contra-
2 years after initial diagnosis fulfilled the clinical dictory findings, and dependent upon the disease
criteria for an AD dementia diagnosis [36], as stage. In cognitively normal individuals, no asso-
illustrated in Fig.  15.1a. Furthermore, the same ciation was found between APOE ε4 and tau lev-
study exhibited a positive correlation between els in the CSF [39]. However, a positive
increased CSF α-syn levels and brain Aβ plaque association between APOE ε4 and both t-tau and
load in Aβ-positive asymptomatic subjects carry- p-tau levels in the CSF was reported in MCI and
ing autosomal dominant AD (ADAD) mutations AD groups [40].
in the APP, PSEN1 and PSEN2 genes (Fig. 15.1b). APOE ε2 is a protective factor leading to
Interestingly, when stratifying all mutation carri- lower Aβ pathology as measured either in the
ers into APOE ε4-positive and APOE ε4-negative, brain or CSF, but it did not affect tau levels in the
a significant positive association between CSF CSF [41], so its protective effect seems to be
α-syn levels and brain Aβ plaque load was found restricted to the Aβ pathology. The protective
in the APOE ε4-positive ADAD mutation carriers action of APOE ε2 has been ascribed to its anti-­
only (Fig.  15.1c). The described link between oxidative and anti-inflammatory effects. There is
CSF α-syn levels and the build-up of brain Aβ still very little evidence whether APOE ε2 has
plaque load was absent in APOE ε4-negative benefits in terms of reduced tau load, and the
mutation carriers [36]. Together, these results mechanism on how these two factors are related
propose a novel link between the APOE ε4 allele, is not known [42].
higher CSF α-syn levels, and the development of
both sporadic and familial AD.  The molecular
relevance of this relationship remains to be  rain Imaging-Based Evidence
B
investigated. Linking Apolipoprotein E to Tau
Other studies have investigated the association and α-Synuclein
between CSF levels of α-syn, t-tau and p-tau in
different neurodegenerative diseases, however the Recently, tau PET imaging has become available
influence of APOE ε4 on those relationships is and applied to visualize tau deposits in different
still not well known. In cognitively normal sub- neurodegenerative diseases [3, 4]. No α-syn PET
15  Associations Between APOE Variants, Tau and α-Synuclein 181

Fig. 15.1  Relationships between APOE ε4, α-syn, tau tions between CSF α-syn and regional Aβ uptake as mea-
and Aβ in the development of sporadic and familial AD. sured by PiB in APOE ε4-positive familial AD mutation
(a) CSF α-syn concentrations as a function of APOE ε4 in carriers; (d) Higher CSF α-syn concentrations in CSF
sporadic AD patients; (b) Positive associations between t-tau positive MCI and AD patients; (e) Higher CSF α-syn
CSF α-syn and regional Aβ uptake as measured by concentrations in CSF p-tau positive MCI and AD
11
C-Pittsburgh compound B (PiB) in PiB-positive familial patients. (Figure adapted from Twohig et al. [36])
AD asymptomatic mutation carriers; (c) Positive associa-

tracer is yet available, although it is subject of PET tracers especially in healthy aging [38], con-
intense research and development efforts [43]. sistent with APOE ε4 being a risk factor for
Tau PET imaging studies comparing the developing AD pathology with aging. In MCI
brain  regional uptake patterns in synucleinopa- and AD groups, previous studies have reported
thies versus in other proteinopathies  are scarce either a positive association between APOE ε4
[44, 45], and much is still unknown regarding the and Aβ plaque deposition [38], or no correlation
relationships between APOE variants and tau [47].
PET uptake in different proteinopathies [46]. In cognitively normal individuals, the relation-
Previous to tau PET imaging, a vast literature ship between APOE ε4 and tau PET uptake as
had reported positive associations between APOE measured by the  18F-AV1451 PET tracer was not
ε4 and Aβ plaque deposition as measured by Aβ significant, despite a positive association between
182 E. Rodriguez-Vieitez and H. M. Nielsen

APOE ε4 and Aβ PET uptake in the same group PET  uptake involving the posterior temporo-­
[48]. When investigating MCI and AD groups, parietal and occipital regions in DLB compared
few studies have looked at the differences in topo- to AD [44, 45], as illustrated in Fig. 15.2. No PET
graphical patterns of tau PET uptake depending imaging studies to our knowledge have yet
on the APOE ε4 genotype. The limited evidence reported on the relationship between APOE ε4
available points to APOE ε4 carriers displaying a and tau PET retention in synucleinopathies, or on
more typical AD-like pattern of tau deposits the associations between APOE ε4, tau PET
involving the entorhinal cortex, medial temporal retention and CSF α-syn in different
lobe and parts of the neocortex, consistent with a proteinopathies.
more amnestic clinical presentation [46, 49]. In
contrast, APOE ε4 non-carriers had a more atypi-
cal tau topography, showing high tau load in the  ell and Molecular Observations
C
neocortex including parietal and frontal regions Proposing Associations
and relatively lower levels of tau in the entorhinal Between APOE, Tau and α-Synuclein
cortex [46, 49]. The total tau load was found to be
higher in APOE ε4 carriers compared to non-car- The bulk evidence tying APOE to neurodegener-
riers in a small group of mostly atypical AD ative disease supports a strong association of
patients, even after controlling for global Aβ load, mainly APOE ε4 and the apoE4 isoform to AD
suggesting a direct and Aβ-independent effect of and Aβ pathology [52]. It is well-described that
APOE ε4 on tau deposition [50], consistent with carriers of the APOE ε4 allele silently, in the
findings that APOE ε4 facilitates the phosphoryla- absence of cognitive symptoms, develop Aβ
tion of tau [51]. pathology already in their third and fourth decade
Due to the recent development of tau PET of life [53, 54]. Numerous studies have by now
imaging, few studies have yet compared the pat- shown that apoE may hamper cellular clearance
terns of tau PET upake across different pro- of Aβ [55–57]. The underlying mechanism was
teinopathies. The evidence so far has shown proposed to be competition between Aβ and
higher brain tau load in AD compared to DLB apoE, a heparin-binding protein, for cell surface
[44, 45], also showing different topographies, heparin sulfate proteoglycans (HSPG) [58]. The
with greater tau load within AD-typical regions HSPG pathway was previously suggested to be a
such as the medial temporal lobe in AD, and a common route for cellular uptake of both tau and
more atypical (non-Braak) pattern of tau α-syn [59], hence in theory apoE may be able to

Fig. 15.2  Surface-based comparisons of the regional pat- represent−log10P values for each comparison. (Figure
terns of tau (18F-AV1451) PET retention in DLB and AD adapted from  Fig. 2  in Lee et  al. [44]  with permission
patient groups, compared to healthy controls. Color bars from John Wiley & Sons Inc. © 1999-2019)
15  Associations Between APOE Variants, Tau and α-Synuclein 183

also alter the cellular uptake of both these mole- Ser202 sites by the activation of GSK-3β in the
cules. Indeed, whereas apoE appeared to inter- murine neuroblastoma N2a cell line [66].
fere with cellular Aβ uptake in an apoE The notion of a potentially pathological role
isoform-unspecific manner in vitro [58], specifi- of neuronal apoE4 expression was supported by
cally the apoE4 isoform reduced oligodendro- results generated by overexpression of the apoE4
cytic uptake of α-syn in vitro [25]. Whether the isoform specifically in neurons (under the Thy1
apoE4 effect on cellular uptake of both tau and promoter) resulting in tau hyperphosphorylation
α-syn is significant enough to cause pathology in two independent Thy1-ApoE4 transgenic
remains to be investigated. Interestingly, an mouse lines [67]. More recent studies in which
increase in extracellular levels of α-syn, possibly P301S tau mice were crossed to human APOE
due to an apoE4-mediated shift in the distribution knock-out (KO) or knock-in (KI) mice showed
of the protein between the intra- and extracellular higher tau levels and a redistribution of tau from
compartments, may in turn promote tau phos- the axons to the neuronal soma in P301S/E4 mice
phorylation by glycogen synthase kinase-3β compared to P301S/E2, P301S/E3 and P301S/
(GSK-3β) [60]. EKO [68]. The authors attributed their findings to
A separate line of evidence proposes a direct a toxic gain of function of the apoE4 isoform,
link between the apoE4 isoform and tau pathol- which led to tau pathogenesis, neuroinflamma-
ogy. Apolipoprotein E immunoreactivity was tion and tau-mediated neurodegeneration inde-
described in association with tau neurofibrillary pendent from Aβ. Whether the toxic gain of
tangles [61] and specifically an 18  kDa amino-­ function observed in this particular animal model
terminal fragment of the apoE4 isoform was resulted from the generation of toxic apoE4 frag-
shown to colocalize with the neurofibrillary tan- ments remains to be investigated.
gle marker PHF-1 in frontal cortex sections from Taken together, experimental evidence sup-
AD patients [62]. It is well-known that the apoE4 ports a direct association between the apoE4 iso-
isoform is more susceptible to proteolysis than the form and GSK-3β mediated tau
apoE3 isoform and in addition to the identifica- hyperphosphorylation leading to subsequent tau
tion of the amino-terminal fragment of apoE4 in pathology. This cascade may be further amplified
AD patients, Rohn and colleagues also identified if apoE4 affects the distribution between intra-
the same fragment in Pick bodies in the brains of and extracellular pools of α-syn where increased
patients suffering from the tauopathy Pick’s dis- levels of the latter would also promote tau phos-
ease. Based on their results the authors suggested phorylation via activation of GSK-3β.
that apoE may have a broader role beyond the
association with AD [63]. Other apoE fragment
species including C-terminal truncated 29  kDa Conclusions
and 14–20 kDa fragments were shown to accumu-
late in the brains of AD patients who were APOE Neurodegenerative diseases are multifactorial,
ε4-carriers, exhibiting higher levels than those of characterized by abnormal protein deposition,
subjects with other APOE alleles [51, 64]. and showing high degrees of co-pathologies and
Bioactive C-terminal truncated apoE4 fragments overlaps in clinical presentation, which presents
were generated in neurons and mainly in brain a challenge for differential diagnosis. The rela-
regions susceptible to AD-associated neurodegen- tionships between genetic risk factors and the
eration. This neuron-specific C-terminal fragmen- different proteinopathies are still not well known.
tation of apoE4 was associated with increased tau In particular, tau PET imaging is still a young
phosphorylation which the authors proposed may research field  [69], and no PET tracer is yet
play a key role in AD-associated neuronal deficits available to visualize α-syn inclusions, although
[65]. Another study showed that both truncated this is a very active field of research. In addition,
apoE4 (delta279-299) and full-length apoE4 PET imaging of activated astrocytes and microg-
overexpression caused tau phosphorylation at lia will help to characterize the role of inflamma-
184 E. Rodriguez-Vieitez and H. M. Nielsen

tion and genetic risk factors on the presence and 14. Pitas RE, et al. Astrocytes synthesize apolipoprotein E
and metabolize apolipoprotein E-containing lipopro-
brain distribution of proteinopathies. Future teins. Biochim Biophys Acta. 1987;917(1):148–61.
studies should incorporate multiple fluid and 15. Bu G.  Apolipoprotein E and its receptors in

imaging biomarkers in longitudinal studies of Alzheimer’s disease: pathways, pathogenesis and
different neurodegenerative disease cohorts. therapy. Nat Rev Neurosci. 2009;10(5):333–44.
16. Gregg RE, et  al. Abnormal in  vivo metabolism

These studies will help to elucidate the underly- of apolipoprotein E4  in humans. J  Clin Invest.
ing mechanisms by which APOE ε4 and other 1986;78(3):815–21.
genetic factors may predispose individuals to 17. Wildsmith KR, et  al. In vivo human apolipoprotein
neurodegenerative dementias. E isoform fractional turnover rates in the CNS. PLoS
One. 2012;7(6):e38013.
18. Ikewaki K, et al. Comparative in vivo metabolism of
apolipoproteins E2 and E4 in heterozygous apoE2/4
References subjects. J Lab Clin Med. 2002;140(5):369–74.
19. Martinez-Morillo E, et al. Total apolipoprotein E lev-
1. Robinson JL, et  al. Neurodegenerative disease con- els and specific isoform composition in cerebrospinal
comitant proteinopathies are prevalent, age-related fluid and plasma from Alzheimer’s disease patients
and APOE4-associated. Brain. 2018;141(7):2181–93. and controls. Acta Neuropathol. 2014;127(5):633–43.
2. Irwin DJ, et  al. Neuropathological and genetic cor- 20. Rasmussen KL, et al. Plasma levels of apolipoprotein
relates of survival and dementia onset in synucle- E and risk of dementia in the general population. Ann
inopathies: a retrospective analysis. Lancet Neurol. Neurol. 2015;77(2):301–11.
2017;16(1):55–65. 21. Berge G, et  al. Apolipoprotein E epsilon2 genotype
3. Villemagne VL, et al. Imaging tau and amyloid-beta delays onset of dementia with Lewy bodies in a
proteinopathies in Alzheimer disease and other condi- Norwegian cohort. J  Neurol Neurosurg Psychiatry.
tions. Nat Rev Neurol. 2018;14(4):225–36. 2014;85(11):1227–31.
4. Villemagne VL, et al. Tau imaging: early progress and 22. Li YJ, et  al. Apolipoprotein E controls the risk

future directions. Lancet Neurol. 2015;14(1):114–24. and age at onset of Parkinson disease. Neurology.
5. Clinton LK, et  al. Synergistic interactions between 2004;62(11):2005–9.
Abeta, tau, and alpha-synuclein: acceleration of 23. Tsuang D, et  al. APOE epsilon4 increases risk for
neuropathology and cognitive decline. J  Neurosci. dementia in pure synucleinopathies. JAMA Neurol.
2010;30(21):7281–9. 2013;70(2):223–8.
6. Van Cauwenberghe C, Van Broeckhoven C, Sleegers 24.
Verghese PB, Castellano JM, Holtzman
K.  The genetic landscape of Alzheimer disease: DM.  Apolipoprotein E in Alzheimer’s disease
clinical implications and perspectives. Genet Med. and other neurological disorders. Lancet Neurol.
2016;18(5):421–30. 2011;10(3):241–52.
7. Shi Y, Holtzman DM. Interplay between innate immu- 25. Ogaki K, et al. Multiple system atrophy and apolipo-
nity and Alzheimer disease: APOE and TREM2 in the protein E. Mov Disord. 2018;33:647–50.
spotlight. Nat Rev Immunol. 2018;18:759–72. 26. Zhao N, et  al. APOE epsilon2 is associated with

8. Corbo RM, Scacchi R.  Apolipoprotein E (APOE) increased tau pathology in primary tauopathy. Nat
allele distribution in the world. Is APOE∗4 a ‘thrifty’ Commun. 2018;9(1):4388.
allele? Ann Hum Genet. 1999;63(Pt 4):301–10. 27. Cruchaga C, et al. Cerebrospinal fluid APOE levels:
9. Raichlen DA, Alexander GE. Exercise, APOE geno- an endophenotype for genetic studies for Alzheimer’s
type, and the evolution of the human lifespan. Trends disease. Hum Mol Genet. 2012;21(20):4558–71.
Neurosci. 2014;37(5):247–55. 28. Darreh-Shori T, et al. Differential levels of apolipopro-
10. Leduc V, et al. Function and comorbidities of apolipo- tein E and butyrylcholinesterase show strong associa-
protein e in Alzheimer’s disease. Int J Alzheimers Dis. tion with pathological signs of Alzheimer’s disease in
2011;2011:974361. the brain in vivo. Neurobiol Aging. 2011;32(12):2320
11. Kraft HG, et  al. Changes of genetic apolipoprotein e15-32.
phenotypes caused by liver-transplantation  – impli- 29. Riddell DR, et al. Impact of apolipoprotein E (ApoE)
cations for apolipoprotein synthesis. J  Clin Investig. polymorphism on brain ApoE levels. J  Neurosci.
1989;83(1):137–42. 2008;28(45):11445–53.
12. Linton MF, et al. Phenotypes of apolipoprotein B and 30. Simon R, et al. Total ApoE and ApoE4 isoform assays
apolipoprotein E after liver transplantation. J  Clin in an Alzheimer’s disease case-control study by
Invest. 1991;88(1):270–81. ­targeted mass spectrometry (n=669): a pilot assay for
13. Kockx M, Traini M, Kritharides L. Cell-specific pro- methionine-containing proteotypic peptides. Mol Cell
duction, secretion, and function of apolipoprotein E. J Proteomics. 2012;11(11):1389–403.
Mol Med (Berl). 2018;96(5):361–71. 31. Song F, et al. Plasma apolipoprotein levels are asso-
ciated with cognitive status and decline in a com-
15  Associations Between APOE Variants, Tau and α-Synuclein 185

munity cohort of older individuals. PLoS One. 47. Rabinovici GD, et al. Increased metabolic vulnerabil-
2012;7(6):e34078. ity in early-onset Alzheimer’s disease is not related to
32. Martinez-Morillo E, et  al. Assessment of peptide
amyloid burden. Brain. 2010;133(Pt 2):512–28.
chemical modifications on the development of an 48. Johnson KA, et al. Tau positron emission tomographic
accurate and precise multiplex selected reaction imaging in aging and early Alzheimer disease. Ann
monitoring assay for apolipoprotein e isoforms. Neurol. 2016;79(1):110–9.
J Proteome Res. 2014;13(2):1077–87. 49. Whitwell JL, et  al. [(18) F]AV-1451 clustering of

33. Nielsen HM, et al. Peripheral apoE isoform levels in entorhinal and cortical uptake in Alzheimer’s disease.
cognitively normal APOE epsilon3/epsilon4 individu- Ann Neurol. 2018;83(2):248–57.
als are associated with regional gray matter volume 50. Ossenkoppele R, et al. Tau PET patterns mirror clini-
and cerebral glucose metabolism. Alzheimers Res cal and neuroanatomical variability in Alzheimer’s
Ther. 2017;9(1):5. disease. Brain. 2016;139(Pt 5):1551–67.
34. Toledo JB, et al. CSF Apo-E levels associate with cog- 51. Harris FM, et  al. Carboxyl-terminal-truncated apoli-
nitive decline and MRI changes. Acta Neuropathol. poprotein E4 causes Alzheimer’s disease-like neurode-
2014;127:621–32. generation and behavioral deficits in transgenic mice.
35. Koch G, et  al. CSF tau is associated with impaired Proc Natl Acad Sci U S A. 2003;100(19):10966–71.
cortical plasticity, cognitive decline and astrocyte sur- 52. Kanekiyo T, Xu H, Bu G.  ApoE and Abeta in

vival only in APOE4-positive Alzheimer’s disease. Alzheimer’s disease: accidental encounters or part-
Sci Rep. 2017;7(1):13728. ners? Neuron. 2014;81(4):740–54.
36. Twohig D, et  al. The relevance of cerebrospinal
53. Kok E, et al. Apolipoprotein E-dependent accumula-
fluid alpha-synuclein levels to sporadic and familial tion of Alzheimer disease-related lesions begins in
Alzheimer’s disease. Acta Neuropathol Commun. middle age. Ann Neurol. 2009;65(6):650–7.
2018;6(130):1–19. 54. Pletnikova O, et  al. The spectrum of preclinical

37. Vergallo A, et  al. Association of cerebrospinal fluid Alzheimer’s disease pathology and its modulation by
alpha-synuclein with total and phospho-tau181 protein ApoE genotype. Neurobiol Aging. 2018;71:72–80.
concentrations and brain amyloid load in cognitively 55. Nielsen HM, et  al. Astrocytic Abeta1-42 uptake

normal subjective memory complainers stratified by is determined by Abeta-aggregation state and the
Alzheimer’s disease biomarkers. Alzheimers Dement. presence of amyloid-associated proteins. Glia.
2018;14:1623–31. 2010;58(10):1235–46.
38. Veitch DP, et  al. Understanding disease progres-
56. Mulder SD, et  al. Apolipoproteins E and J  interfere
sion and improving Alzheimer’s disease clinical with amyloid-beta uptake by primary human astrocytes
trials: recent highlights from the Alzheimer’s dis- and microglia in vitro. Glia. 2014;62(4):493–503.
ease neuroimaging initiative. Alzheimers Dement. 57. Verghese PB, et  al. ApoE influences amyloid-beta
2019;15:106–52. (Abeta) clearance despite minimal apoE/Abeta asso-
39. Morris JC, et al. APOE predicts amyloid-beta but not ciation in physiological conditions. Proc Natl Acad
tau Alzheimer pathology in cognitively normal aging. Sci U S A. 2013;110(19):E1807–16.
Ann Neurol. 2010;67(1):122–31. 58. Fu Y, et  al. Apolipoprotein E lipoprotein particles
40. Risacher SL, et  al. The role of apolipoprotein E
inhibit amyloid-beta uptake through cell surface
(APOE) genotype in early mild cognitive impairment heparan sulphate proteoglycan. Mol Neurodegener.
(E-MCI). Front Aging Neurosci. 2013;5:11. 2016;11(1):37.
41. Grothe MJ, et al. Multimodal characterization of older 59. Holmes BB, et  al. Heparan sulfate proteoglycans

APOE2 carriers reveals selective reduction of amy- mediate internalization and propagation of spe-
loid load. Neurology. 2017;88(6):569–76. cific proteopathic seeds. Proc Natl Acad Sci U S A.
42. Suri S, et  al. The forgotten APOE allele: a review 2013;110(33):E3138–47.
of the evidence and suggested mechanisms for the 60. Gassowska M, et  al. Extracellular alpha-synuclein

protective effect of APOE varepsilon2. Neurosci leads to microtubule destabilization via GSK-3beta-­
Biobehav Rev. 2013;37(10 Pt 2):2878–86. dependent Tau phosphorylation in PC12 cells. PLoS
43. Mathis CA, et  al. Small-molecule PET tracers
One. 2014;9(4):e94259.
for imaging proteinopathies. Semin Nucl Med. 61. Namba Y, et  al. Apolipoprotein E immunoreactiv-

2017;47(5):553–75. ity in cerebral amyloid deposits and neurofibril-
44. Lee SH, et al. Distinct patterns of amyloid-dependent lary tangles in Alzheimer’s disease and kuru plaque
tau accumulation in Lewy body diseases. Mov Disord. amyloid in Creutzfeldt-Jakob disease. Brain Res.
2018;33(2):262–72. 1991;541(1):163–6.
45. Kantarci K, et al. AV-1451 tau and beta-amyloid posi- 62. Rohn TT, et  al. Identification of an amino-terminal
tron emission tomography imaging in dementia with fragment of apolipoprotein E4 that localizes to neu-
Lewy bodies. Ann Neurol. 2017;81(1):58–67. rofibrillary tangles of the Alzheimer’s disease brain.
46. Mattsson N, et al. Greater tau load and reduced corti- Brain Res. 2012;1475:106–15.
cal thickness in APOE epsilon4-negative Alzheimer’s 63. Rohn TT, et  al. Immunolocalization of an amino-­

disease: a cohort study. Alzheimers Res Ther. terminal fragment of apolipoprotein E in the Pick’s
2018;10(1):77. disease brain. PLoS One. 2013;8(12):e80180.
186 E. Rodriguez-Vieitez and H. M. Nielsen

64. Huang Y, et  al. Apolipoprotein E fragments present 67. Tesseur I, et al. Expression of human apolipoprotein
in Alzheimer’s disease brains induce neurofibrillary E4  in neurons causes hyperphosphorylation of pro-
tangle-like intracellular inclusions in neurons. Proc tein tau in the brains of transgenic mice. Am J Pathol.
Natl Acad Sci U S A. 2001;98(15):8838–43. 2000;156(3):951–64.
65. Brecht WJ, et  al. Neuron-specific apolipoprotein e4 68. Shi Y, et  al. ApoE4 markedly exacerbates tau-­

proteolysis is associated with increased tau phos- mediated neurodegeneration in a mouse model of
phorylation in brains of transgenic mice. J Neurosci. tauopathy. Nature. 2017;549(7673):523–7.
2004;24(10):2527–34. 69. Leuzy A, et  al. Tau PET imaging in neurodegenera-
66. Zhou J, Chen J, Feng Y.  Effect of truncated-ApoE4 tive tauopathies—still a challenge. Mol Psychiatry.
overexpression on tau phosphorylation in cultured 2019;24(8):1112–34.
N2a cells. J Huazhong Univ Sci Technolog Med Sci.
2006;26(3):272–4.
 Amyloid-β and Tau
at the Crossroads of Alzheimer’s 16
Disease

Gilbert Gallardo and David M. Holtzman

Introduction as the “amyloid hypothesis”, which postulates

that amyloid-β is the predominant driving factor
Alzheimer’s disease (AD) is the most common in AD development. Nonetheless, more recent
form of dementia characterized neuropathologi- advances in imaging analysis, biomarkers and
cally by senile plaques and neurofibrillary tangles mouse models are now redefining this original
(NFTs). Early breakthroughs in AD research led hypothesis, as it is likely amyloid-β, tau and other
to the discovery of amyloid-β as the major com- pathophysiological mechanism such as inflam-
ponent of senile plaques and tau protein as the mation, come together at a crossroads that ulti-
major component of NFTs. Shortly following the mately leads to the development of AD.
identification of the amyloid-β (Aβ) peptide was
the discovery that a genetic mutation in the amy-
loid precursor protein (APP), a type1 transmem-  lzheimer’s Disease Introduction
A
brane protein, can be a cause of autosomal (History)
dominant familial AD (fAD). These discoveries,
coupled with other breakthroughs in cell biology In 1901, Alois Alzheimer first examined Auguste
and human genetics, have led to a theory known D., a 51-year-old woman at the Frankfurt hospital
who was admitted after presenting symptoms of
disorientation, paranoia, delusions, hallucina-
G. Gallardo tions, psychosocial incompetence and cognitive
Department of Neurology, Washington University impairment [1]. Over the ensuing years, Dr.
School of Medicine, St. Louis, MO, USA Alzheimer continued to evaluate Auguste D. until
Hope Center for Neurological Disorders, Washington her death on April 8, 1906  in Frankfurt. In
University, St. Louis, MO, USA November of 1906, at the 37th Conference of
e-mail: gallardog@neuro.wustl.edu
South-West German Psychiatrists in Tubingen,
D. M. Holtzman (*) Dr. Alzheimer presented his clinical studies on
Department of Neurology, Washington University
School of Medicine, St. Louis, MO, USA Auguste D., where he correlated her neuropsy-
chological symptoms with the neuropathology
Hope Center for Neurological Disorders, Washington
University, St. Louis, MO, USA that consisted of “miliary foci” and “neurofibrils”
in the cerebral cortex [1]. These pathological
Charles F. and Joanne Knight Alzheimer’s Disease
Research Center, Washington University, hallmarks would later become known as amyloid
St. Louis, MO, USA plaques and neurofibrillary tangles. The disease
e-mail: holtzman@wustl.edu itself would become known as Alzheimer’s

© Springer Nature Singapore Pte Ltd. 2019 187

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_16
188 G. Gallardo and D. M. Holtzman

d­ isease (AD) used initially to describe presenile ease causing agent and have essentially nullified
dementia. the amyloid cascade hypothesis. On the other
For the better part of the century, research into hand, the failure of multiple clinical trials to pre-
AD lay dormant until the mid-1980s. Around this vent cognitive decline or improve cognitive
time, it was discovered that the Aβ protein was parameters when targeting Aβ suggest several
the predominant pathological component in amy- possibilities. First, the treatments used in trials
loid plaques isolated from AD and Down syn- that have gone through phase III without hitting
drome patients [2–4]. Because of this link, it was their primary endpoints have not effectively
subsequently hypothesized that the gene encod- removed Aβ from the brain. Second, they tar-
ing the Aβ protein might be located on chromo- geted Aβ ~25 years after it started accumulating
some 21, given that complete trisomy of when it is likely other mechanisms driving dis-
chromosome 21 was known to be the cause of ease [11]. Third, while Aβ may be critical in ini-
Down syndrome. It was also known that Down tiating the process that we ultimately know of as
syndrome patients develop early amyloid plaque AD, it may serve as a catalyst that initiates a
pathology, with the vast majority developing series of pathophysiological events in a time-­
dementia by the age of 50 [5, 6]. This hypothesis dependent manner that also includes neurofibril-
held to be true as it was discovered that the APP lary tangles. Therefore, consistent with the Aβ
gene encoding a larger type 1 transmembrane hypothesis, Aβ may still be the predominant initi-
protein, from which the Aβ peptide is excised, ating factor but targeting Aβ production or elimi-
was indeed located on chromosome 21 [7, 8]. nating amyloid plaques alone may not be
These early breakthroughs in AD research led to sufficient to halt downstream events once signifi-
the origins of the “amyloid cascade hypothesis” cant Aβ accumulation has already taken place.
[9]. The amyloid cascade hypothesis posits that We will review AD pathology and discuss the
amyloid-β aggregation and/or deposition is the relevance of Aβ, its influences on tau pathology,
starting point of AD development that initiates all and the evidence that either supports or refutes
associated pathophysiological events. These the amyloid cascade hypothesis.
pathophysiological events include tau neurofi-
brillary tangles, synaptic dysfunction, inflamma-
tion, neurodegeneration, vascular abnormalities Neuropathological Features of AD
and dementia. Following this logic, therapeutic
interventions targeting Aβ should ultimately pre- Amyloid Plaques
vent or cure AD-related dementia.
Although the amyloid cascade hypothesis for The defining neuropathological features of AD
AD has been highly influential in guiding include senile plaques extracellularly and the
research toward understanding the biological and intracellular neurofibrillary tangles (NFTs). The
pathological roles of Aβ peptide and its precursor primary component of amyloid plaques is the
protein, its dominance in the field over the past accumulation of aggregated non-fibrillar, fibril-
two decades has been criticized. Because of this lar, or oligomeric Aβ peptide. In AD, there are
dominance, the majority of research has empha- two types of Aβ plaques; diffuse Aβ plaques and
sized therapeutic strategies aimed at either the dense-core plaques. The differences between
decreasing Aβ production and aggregation or the two are based on their staining properties
enhancing Aβ clearance and plaque burden. To with the dyes Thioflavin-S and Congo Red that
date, large clinical trials targeting Aβ have not are specific for the β-pleated sheet conformation
met their primary endpoints [10]. To some peo- [12–14]. Diffuse amyloid-β plaques are
ple, these failures in clinical trials have brought Thioflavin-S negative plaques and are commonly
to question if Aβ is genuinely the preeminent dis- present in the brains of cognitively healthy adults.
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 189

The dense-core plaques are composed of Aβ include pretangles, intracellular tangles (globose
fibrils, which are Thioflavin-S positive and are or flame shape) or extraneuronal (ghost) [26].
most often found in AD patients [12]. Dense-core Pretangles are defined by diffuse, punctate tau
plaques are associated with neurotoxicity as they staining within the cytoplasm of neurons that
are commonly surrounded by synaptic loss, reac- have well-preserved processes. The pretangles
tive astrocytes, and microglia. The dense-core may give rise to the formation of mature intra-
plaques also associate with dystrophic neurites, neuronal fibril tangles consisting of aggregated
which are distortions in neuronal axons or den- pathological tau in the neuronal soma that is
dritic processes that contain tau paired helical accompanied with dendritic abnormalities. Ghost
filament (PHF), also referred to as neuritic tangles are a result of the death of neurons con-
plaques. Aβ peptide not only accumulates in taining NFTs and distinguished by the absence of
senile plaques but also in arteriolar blood vessel a nucleus or cytoplasm [26].
walls in the form of cerebral amyloid angiopathy
(CAA) [15]. While the contribution of CAA
pathology in the development or progression of  he Genetic Basis of AD
T
dementia in AD is not completely clear, it some and the Amyloid Cascade
cases it leads to lobar hemorrhages and likely Hypothesis
contributes to cognitive decline in AD when
present. The Aβ peptide or alternative APP fragments
derive from sequential proteolytic cleavage of the
amyloid precursor protein (APP), by either
Neurofibrillary Tangles α-secretase or β-secretase generating a soluble
carboxyterminal fragment (CTF) through either
The neurofibrillary tangles are mainly composed an amyloidogenic processing pathway or a non-­
of aggregated hyperphosphorylated tau that ultra- amyloidogenic processing pathway respectively.
structurally appears as PHF [16, 17]. Tau is a Under normal physiological conditions, 90% of
microtubule binding protein that is predomi- APP is cleaved by the α-secretase, which is non-­
nately localized to axons of neurons under nor- amyloidogenic as it generates a sAβPPα and an
mal conditions [18]. Under pathological APP CTF83. The CTF83 is cleavable by
conditions, tau forms PHF in both AD and a host γ-secretase, which generates a small p3 peptide.
of neurodegenerative diseases known as tauopa- The amyloidogenic processing pathway is gener-
thies [19]. There is strong evidence that the ated by β-secretase cleavage, resulting in a sAPPβ
hyperphosphorylation of tau plays a pivotal role and a CTF99 fragment that is targeted by
in the pathogenesis of tau, as phosphorylation γ-secretase at distinctly different sites than the
liberates tau from microtubules, enabling it to CTF83. This generates different Aβ peptide iso-
aggregate and translocate to the neuronal pro- forms 38–43 amino acids in length. Under non-­
cesses or cell body [20–22]. In AD, PHF aggre- pathological conditions, the predominant isoform
gation in the neuronal processes that is associated is the Aβ40 peptide whereas in AD there is an
with neuritic plaques is referred to as neuropil increase of the Aβ42 peptide, which displays a
threads [23, 24]. These neuropil threads are higher propensity to aggregate.
linked to profound alteration in the neuronal The discovery that the APP gene localizes to
cytoskeleton [25]. As opposed to the tau pathol- chromosome 21 and the striking resemblance of
ogy in the dendritic and axon processes, the accu- AD-like neuropathology seen in Down’s syn-
mulation of PHF in the cell bodies of neurons drome played a pivotal role in providing the first
form the neurofibrillary tangles (NFT) [16]. In genetic evidence for Aβ as the primary agent in
AD, the different types of NFTs that are present AD development. This knowledge lead to the
190 G. Gallardo and D. M. Holtzman

speculation that perhaps either increased APP are associated with early-onset fAD [43–46]. The
expression, mutations in the APP gene, or altera- PSEN1 and PSEN2 genes encode for the poly-
tion of APP processing leads to higher Aβ levels topic transmembrane protein that forms the cata-
and may be the initiator of AD development. lytic subunits of the γ-secretase complex [47].
Subsequently, missense mutations in the APP This complex is a heterotrimeric complex com-
gene were indeed identified to be associated with posed of nicastrin, PEN2, and anterior pharynx
AD and CAA [27–31]. Moreover, mutations in defective 1 [48, 49]. The γ-secretase complex
the APP flanking the C-terminus cleavage sites belongs to a family of intramembranous aspartyl
that regulate where Aβ is cleaved from APP lead proteases that cleave transmembrane proteins.
to an increase of Aβ42 to Aβ40 ratio [32, 33]. Among the γ-secretase complex substrates is the
This evidence for a higher ratio of the Aβ42 vs. APP transmembrane protein from which
shorter species supported the amyloid cascade γ-secretase activity is necessary for the produc-
hypothesis as the Aβ42 species is more fibrillo- tion of Aβ peptide. Similar to mutations in APP,
genic and is the predominant form in amyloid point mutations in the PSEN1 and PSEN2 either
plaques because of it’s propensity to aggregate. increase Aβ peptide production or alter the ratio
Coinciding with the discovery that an increase of Aβ species being made and produced in favor
in Aβ production and its aggregation are the initi- of fibrillogenic Aβ42 species.
ating steps of AD was the discovery of mutations Aside from mutations related to Aβ process-
that lead to an excessive production of all Aβ spe- ing, the most significant genetic modifier of the
cies and are associated with both AD and CAA late-onset AD is the ε4 allele of apolipoprotein E
[34–37]. In addition to mutations, duplication of (APOE) [50]. There is strong evidence support-
the APP gene was also discovered in five families ing that the ApoE protein influences amyloid-β
with an autosomal dominant early-onset AD as plaque deposition [51]. In humans, there are three
well as a mutation identified near the β-secretase common alleles of the APOE gene (ApoE2,
cleavage sites of APP found in a Swedish family ApoE3, and ApoE4). Carriers of the ApoE4 allele
[37]. In striking contrast, a mutation only two display a dosage response for increased risk of
residues away from the β-secretase cleavage site developing AD (~3.7 fold increase for 1 copy of
of APP (Iceland mutation A673T) attenuates the apoE4 and ~12 fold increase risk for 2 copies of
β-secretase interaction [38]. This attenuation sig- apoE4 relative to 2 copies of the apoE3 allele,
nificantly lowers Aβ production and thus is pro- www.alzgene.org) [52, 53]. ApoE4 carriers dis-
tective against sporadic AD (sAD). play an earlier age of onset for clinical dementia
In addition to mutations that alter the produc- and these patients also show an increased amy-
tion or the ratio of the amyloidogenic Aβ42 spe- loid plaque burden, which is in alignment with
cies, other mutations located within the the amyloid cascade hypothesis [50, 52]. Studies
Aβ-peptide coding region have also been identi- in genetically manipulated mice expressing
fied in fAD [39]. These point mutations within human APP mutations and ApoE isoforms pro-
the Aβ-peptide which are C-terminal to the vided the first evidence that the ApoE protein
α-secretase cleavage site increase the propensity influences amyloid-β fibrillogenesis in an
of Aβ to form protofibrils or enhances oligomer- isoform-­specific mechanism [54, 55]. The expres-
ization. Some evidence suggest these Aβ species sion of ApoE4 displays the highest propensity to
are more toxic to neurons [40–42]. induce Aβ plaque deposition whereas endoge-
Additional support for the involvement of the nous ApoE mouse knockout significantly dimin-
amylogenic pathway in AD development comes ished the fibrillary Aβ plaque deposition [54].
from genetic linkage studies that identified muta- Intriguingly, effects of ApoE do not appear to be
tions in the presenilin 1 and its homolog, preseni- restricted to Aβ deposition, as recent studies have
lin 2, genes (PSEN1 and PSEN2), both of which demonstrated ApoE4 influences tau pathogenesis
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 191

and tau-mediated neurodegeneration indepen- with cognitive decline, hippocampal atrophy, and
dent of amyloid-β by exacerbating a neuroin- cortical hypometabolism [11, 61, 62]. These
flammatory response [56]. However, in agreement analyses infer that a genetic mutation in one of
with the amyloid cascade hypothesis, the genetic the three genes known to cause fAD predisposes
evidence appears to support that Aβ deposition is individuals to abnormal Aβ processing that is first
the preeminent initiating factor that leads to AD detectable by a marked decrease in Aβ42 in the
dementia. The summation of genetic studies sug- CSF followed by the appearance of amyloid
gests either overproduction, altered processing of plaque deposition, then by change in tau that
Aβ42 or increased Aβ/decreased Aβ clearance as coincide with neurodegeneration. Thus, abnor-
the initiators of AD. mal Aβ production is the first of many pathophys-
iological events in AD development that spans
over several years.
 he Correlations of Amyloid-β
T In additional longitudinal studies, the amyloid
and AD Development burden in MCI patients seems to predict cogni-
tive decline over the ensuing years [63–65]. This
Cross-sectional studies in fAD patients have also is probably because it reflects those whose cogni-
provided evidence that abnormal production of tive impairment is due to AD vs. other causes
Aβ is the initiating factor of AD development. which in general, don’t progress as fast. In addi-
These studies were pioneered particularly by the tion to this evidence of a link between amyloid
Dominantly Inherited Alzheimer’s Network presence and cognitive decline in AD, amyloid-­
(DIAN), which took advantage of the fact that the PET imaging studies have revealed a strong cor-
age at which clinical onset occurs for dementia in relation between amyloid-β burden and brain
fAD patients is similar between generations [57]. atrophy in AD [66–69]. Moreover, individuals
This parity in the age of onset across generations that are clinically normal but display amyloid
enabled a pathological timeline to be formed and deposition by PET imaging tend to perform lower
comparisons of brain atrophy, biochemical mea- in cognitive assessments and are associated with
surements of Aβ peptide and tau protein in the grey matter atrophy within the hippocampus and
cerebrospinal fluid (CSF), positron emission the posterior cingulate [70–73]. This finding sug-
tomography (PET) for fibrillary Aβ-deposition, gests Aβ deposition is linked with a pathological
and cognitive assessment to be made between state, which primes vulnerable neurons for tau
carriers and non-carriers within families. The pathogenesis that ultimately drives disease
first abnormality detected was the levels of progression.
Aβ42 in the CSF, which begin to decline from an
elevated level ~25  years before the expected
onset of dementia and continue to decrease over Amyloid-β Is Unlikely to Drive AD
the ensuing years as the disease progresses. Next Development Alone
in the pathological time-line of AD is Aβ deposi-
tion, as visualized by amyloid-PET imaging that While these studies illustrate that abnormalities
occurs within a 10-year period from the decline in Aβ processing precedes all other AD-related
of Aβ42-CSF levels. The decline in Aβ42 levels pathologies, several other studies have demon-
in CSF levels is thought to reflect amyloid-β42 strated there are disconnects between Aβ plaque
sequestration in plaques in the brain [58]. The deposition and the other pathological events in
deposition of Aβ closely coincides with elevated AD.  These studies argue Aβ may be necessary
tau and phosphorylated tau levels in the CSF but not sufficient to lead to marked synaptic and
approximately 3–5 years before cognitive decline neuronal loss. During the course of AD, neuronal
[59, 60]. The changes in tau CSF levels coincide cell death begins in the entorhinal cortex and
192 G. Gallardo and D. M. Holtzman

spreads to the hippocampus and then the neocor- Additional evidence that argues Aβ produc-
tex, whereas Aβ plaque deposition originates tion and or deposition is not the preeminent fac-
within the neocortex and spreads inwardly from tor leading to dementia is found in fAD patients.
there [14, 74]. Differing from the pathology of In fAD, individuals are predisposed to abnormal
Aβ plaque deposition, postmortem histopatho- production of Aβ from birth, yet it seemingly
logical studies have revealed that tau pathology takes several decades before the complex symp-
closely parallels neuronal loss, as tauopathy first tomatology of AD to manifest. This disparity in
appears in the entorhinal cortex and spreads into age with the onset of disease does not imply Aβ is
the hippocampus and then the neocortex [75–77]. not neurotoxic but instead implies other patho-
There is also a steady increase in tauopathy with physiological factors in the aging process are
normal aging, at least in parts of the medial tem- necessary to decrease neuronal health. However,
poral lobe, thus correlating increased tauopathy the age of onset is much earlier in fAD cases as
with the rate and progression of age-related cog- opposed to sporadic AD cases, which suggests
nitive decline whereas Aβ plaque deposition pla- that there may be a threshold of the amyloid bur-
teaus predominantly in the preclinical period of den that is necessary before the disease is
AD [78]. Supporting this, more recent PET initiated.
­imaging tau studies have also revealed tau pathol- Perhaps most perplexing and most challeng-
ogy and cognition are much more closely linked ing for the amyloid cascade hypothesis is the fact
than cognition and amyloid-β plaque deposition that normal individuals may display substantial
[79–81]. These studies suggest tau pathology and Aβ plaque deposition [83, 84]. This may be
its progression is likely responsible for cognitive because it takes time for Aβ accumulation to lead
decline in AD. In addition, the severity of tempo- to enough downstream events and damage to lead
ral lobe tau pathology is sufficient to predict the to cognitive decline. These observations set an
levels of cognitive dysfunction in early disease intriguing conundrum; on one hand, for a definite
stages of AD, unlike Aβ plaque deposition [80]. diagnosis of a dementia-related to AD, it is nec-
Intriguingly, phosphorylated tau is also present essary to have Aβ plaque deposition. On the other
within the brainstem nuclei and the coeruleus/ hand, it is possible for otherwise healthy individ-
subcoeruleus complex in otherwise healthy uals to be amyloid positive without the presence
young adults under the age of 30 with the absence of dementia either at autopsy or by PET imaging.
of Aβ plaque pathology [82]. Thus, some tau However, AD is an age-dependent neurodegen-
pathology occurs in everyone with normal aging erative disease and one might be poised to con-
so that the progression of tau pathology from the clude healthy individuals that present amyloid-β
brainstem and medial temporal lobe to other plaque pathology are on course to developing
brain regions may be driven by Aβ pathology in AD-related dementia which in fact appears to be
AD. Although, it is important to note the basis of the case from longitudinal biomarker studies
these studies are neuropathological and are lack- [85]. Coincidently, cross-sectional biomarker
ing biochemical analysis for oligomeric Aβ or studies spurred from the Alzheimer’s Disease
other amyloid-β species known to influence tau Neuroimaging Imitative (ADNI) suggest that the
phosphorylation [40–42]. Nevertheless, these presence of amyloid-β plaque deposition in oth-
anatomical and temporal mismatches between erwise healthy individuals are not necessarily
Aβ pathology and neuronal cell death indeed asymptomatic, as the presence of Aβ plaques
demonstrates that there is much more complexity deposits in these individuals increases the risk for
in pathophysiological events of AD, supplement- developing dementia by ~four-fold in the ensuing
ing the argument that Aβ plaque deposition alone years [86–88]. Elderly individuals that are Aβ
is not sufficient to initiate the neurodegenerative plaque-positive but otherwise cognitively normal
process as has been proposed. also display accelerated brain atrophy in the cor-
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 193

tex and hippocampus [67]. In a more recent study, phosphatase activation domain is negatively
Alzheimer Disease Neuroimaging (ADNI) charged and generally hidden in the native pro-
enrolled participants that displayed normal cog- tein. However, under pathological conditions, the
nition but were positive for brain amyloid. These N-terminal domain of tau becomes exposed
patients showed a faster decline in cognition, which is thought to be an early event in tau patho-
brain volume, and brain glucose metabolism than genesis [90, 91]. Studies of the C-terminal
individuals without amyloid pathology within a domain suggest it may play a role in inhibiting
four-year span [89]. These studies suggest the tau polymerization, whereas truncation of the
amount of amyloid plaque in otherwise cogni- C-terminal domain facilitates tau pathology [92].
tively normal individuals have not yet reached a The MBD domain and the proline-rich domain
certain amyloid threshold or damage downstream are involved in protein-protein interactions and
of amyloid that is necessary to induce the clinical signaling through the PxxP motifs that are within
symptoms. It appears that tau pathology is likely the proline-rich domain and that interact with
one of the critical downstream factors. SRC kinases.
During the course of AD, several post-­
translational modifications of tau occur, the
Tau Biology and Tau Pathogenesis most notable of which is phosphorylation at
numerous sites [20–22]. The phosphorylation
As stated above, tau is a microtubule-associated of serine and threonine residues at sites imme-
protein that plays an essential role in microtubule diately adjacent to or within the MBD alters its
stability. Phosphorylated tau has been identified confirmation. This alteration liberates tau from
as the main component of NFT in AD as well as the microtubules, leading to the accumulation
in a host of neurodegenerative diseases such as in the somatodendritic compartment of pair
progressive supranuclear palsy, corticobasal helical filaments and other abnormal confor-
degeneration (PSP), Pick’s disease (PiD), cortical mations. Neuropathological studies of post-
basal degeneration (CBD) and certain forms of mortem AD patients provided the first evidence
frontotemporal dementia (FTDP). The tau gene that tau tangles occur hierarchically, with the
consists of 16 exons and the alternative splicing first appearance in the transentorhinal cortex
of exons 2, 3, and 10 generates six major tau pro- (Braak stages I-II) [14, 74]. Neurons associ-
tein isoforms. Exons 9–12 encode the four micro- ated with these tangles give rise to the perfo-
tubule binding repeats, giving rise to tau protein rant pathway, the major projections to the
isoforms with four or three repeats (4R and 3R). hippocampus where tau pathology gradually
Tau isoforms also differ by the presence of zero, advances into CA1 region (Braak II). Next, tau
one or two inserts of a 29-amino-acid sequence tangles develop in the limbic structures of the
(0 N, 1 N, and 2 N) in the amino-terminal half. In subiculum and inferior temporal neocortex
the adult human brain, there is an equal molar (Braak III) the amygdala and thalamus (Braak
ratio in the three-3R and the three-4R isoforms. IV) finally spreading into the neocortex (Braak
In AD and FTDP, there is an equal 3R:4R ratio, V-VI) during the process of AD.
whereas PSP and CBD predominately display 4R The hierarchical pattern of neurofibrillary tau
tau inclusions. accumulation is consistent with transmission of
Structurally, tau is composed of four domains tau fibrils demonstrated in mouse models of
the C-terminal, N-terminal, a proline-rich tauopathy. Injection of brain extracts from dis-
domain, and microtubule binding domain eased tau transgenic mice into the cortex or hip-
(MBD). These domains all play a role in physio- pocampus of mice expressing human wild-type
logical and pathological conditions. In normal tau induced the formation of wild-type tau fibrils
conditions, the N-terminal domain known as the that spread to distant neurons [93]. Similarly, the
194 G. Gallardo and D. M. Holtzman

injection of tau oligomers isolated from AD brain ogy despite an anatomical discord. However, it is
subjects or preformed synthetic tau fibrils into important to emphasize that mutations in the tau
the hippocampus or striatum of tau transgenic gene are capable of leading to neurodegenerative
mice induce neurofibrillary inclusions that propa- diseases independent of Aβ although tau filament
gated from the injected site to connected brain ultrastructure and the cell type in which tau
regions in a time-dependent manner [94]. These pathology occurs as well as anatomical distribu-
studies suggest that pathological tau can propa- tion differs from AD [19].
gate and spreads pathology through intercon-
nected brain regions.
Amyloid-β Deposition Favors Tau
Pathology
Amyloid-β Deposition Influences
Tau Pathology It is evident from the human neuropathological
data that AD is a dual proteinopathy, needing
Although tau pathology is more closely associ- both amyloid plaque deposition together with
ated with cognitive decline and neuronal cell loss tauopathy to progress. In early disease, the NFT
then amyloid-β plaque deposition in AD, tau tan- formation is restricted to the medial temporal
gles are commonly observed in the brainstem and lobe. As the disease progresses, tau pathology
medial temporal lobe of older individuals who spreads outside the temporal lobe into the neo-
are cognitively normal with no evidence of neu- cortex including many areas filled with Aβ
rodegeneration. These findings suggest tau plaques. These findings tend to support that Aβ
­tangles alone do not necessarily lead to the devel- deposition creates an environment that promotes
opment of AD [82]. In AD, where Aβ deposition the formation of tau pathology in line with the
and tau tangles are both present, neurodegenera- amyloid cascade hypothesis for AD.
tion does occur throughout several brain regions, Further evidence for Aβ deposition influenc-
including the hippocampus [95, 96]. This further ing tau pathology has come from mouse models
enforces the suggestion that the combination of of tauopathy. The injection of synthetic Aβ fibrils
Aβ plaques with tauopathy are necessary for the into the hippocampus of transgenic mice express-
AD degenerative process. Pathologically, there is ing human mutant tau-induced tau pathology in
also evidence to suggest Aβ plaque burden exac- the hippocampus that spread to brain connected
erbates downstream tau pathology, as there is a regions including the entorhinal cortex and
strong correlation between the severity of Aβ amygdala [98]. Similarly, the injection of Aβ
pathology and the severity of tau pathology in containing extracts from aged transgenic mice
clinically demented AD patients [80]. overexpressing human APP (hAPP) which har-
Additionally, neuritic plaques that consist of Aβ bor mutations that cause fAD, also induced tau
plaques associated with neurofibrillary tau pathology in transgenic tau mice [99].
pathology correlate with neuronal loss and Characteristic to human AD pathology, the injec-
dementia [97]. There is also evidence suggesting tion of Aβ induced tau pathology spatially sepa-
the Aβ plaque burden in non-demented cases is rate from the Aβ injected region. These findings
associated with the development of dystrophic also coincide with the notion that tau protein dis-
neurites within neuritic plaques [78]. These cases plays the phenomenon of “spreading” tau pathol-
display an increase in both the number of tangles ogy in a non-cell autonomous fashion. This
and rate tangle formation with age. Collectively, “spreading” progresses from cell to cell through
the human neuropathological data is supportive the brain in a well-defined pattern, depending on
of the amyloid cascade hypothesis as it suggests the location where it starts and the specific dis-
the Aβ plaque deposition exacerbates tau pathol- ease in which it occurs.
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 195

Additional evidence that supports the sug- by the intracerebral injections of purified
gestion that Aβ influences tau pathology has human tau from sAD subjects [103, 104]. The
come from double transgenic mice that express intracerebral injections of purified human tau
a human mutant tau together with a human from AD subjects into plaque-bearing trans-
APP gene harboring a double mutation. The genic hAPP mice lead to the formation of tau
double transgenic mice displayed amyloid pathology. Moreover, the tau pathology induced
plaque formation that was indistinguishable resembled human AD pathology because it
from the single hAPP transgenic line, suggest- exhibited the accumulation of tau aggregates in
ing tau mutations do not necessarily augment dystrophic neurites surrounding amyloid
Aβ pathology [100]. This finding is consistent plaques, NFTs, and neuropil threads. The dis-
with human primary tauopathies that display tinct tau pathologies also displayed a temporal
tau pathology independent of amyloid plaque onset with tau pathology first appearing in the
pathology. However, the double transgenic dystrophic neurites, followed by NFTs, and
mice displayed enhanced NFT in the limbic subsequently neuropil threads. Additional neu-
system and olfactory cortex. Additional studies ropathological and behavioral analysis sug-
in double transgenic mice expressing human gested the tau aggregates in dystrophic neurites
mutant tau and hAPP mutations revealed the potentially promote the spreading of tau pathol-
NFT formation displays an earlier onset with a ogy to distal regions whereas NFTs impair neu-
more consistent manner of staging that ema- ronal activity leading behavioral abnormalities
nated in the entorhinal cortex and proceeds like anxiety, for example. However, the mecha-
into hippocampus [101](Lee). The double nism by which amyloid plaques induce distinct
transgenic line also displayed enhancement of tau pathologies and the role of non-neuronal
argyrophilic ThioS-positive NFT in the hippo- cell types including microglia and astrocytes
campus and cortical regions which contain an remains mostly unknown.
abundance of Aβ plaque deposition. Additional
studies have also suggested Aβ potentially pro-
motes an environment that is ideal for the At the Crossroads of AD
development of a distinct tau species that dis-
plays enhanced tau seeding [102]. Collectively, The influence of Aβ deposition on tau pathology
these studies in mouse models of tauopathy does not go only one way, as there is evidence
and amyloidosis suggest Aβ deposition is that amyloid-induced neurotoxicity is depen-
upstream of NFT formation. Aβ plaques favor dent on tau protein. Primary hippocampal neu-
an environment for tau pathology and once tau ronal cultures from rodents exposed to
pathology initiates, it is capable of self-propa- synthetic Aβ fibrils develop hyperphosphory-
gating through neuronal projections. lation of tau protein that accumulated in the
While these studies demonstrate that Aβ somatodendritic compartment, display axonal
plaques influence tau pathology, they rely on retraction, loss of synapses and dendritic atro-
double transgenic mice overexpressing human phy within days [105, 106]. However, primary
tau mutations found in autosomal dominant neurons derived from tau knockout mice or
frontotemporal degeneration. Therefore, the antisense oligonucleotide knockdown of tau
relevance of tau mutations in AD-dependent expression protects neurons from the Aβ
dementia remains unclear. In a more recent induced abnormalities [107, 108]. In addition
study, to eliminate potential confounding to fibrillar amyloid-induced neurotoxicity,
effects of tau mutations, investigators utilized a several studies have demonstrated soluble
paradigm that induces tau seeding and spread- oligomers of Aβ display enhanced neurotoxic-
ing in wild-type mice (endogenous mouse tau) ity in cell culture that are also dependent on
196 G. Gallardo and D. M. Holtzman

tau protein [108, 109]. The Aβ-induced abnor- Other Pathophysiological

malities have also been demonstrated in vivo Mechanisms That Contribute to AD
as the intracerebroventricular injection of
purified Aβ oligomers into adult rodents In addition to Aβ and tau, emerging evidence
blocked long-term potentiation (LTP), a model suggests other neuronal stressors such as neuro-
of memory [40]. The Aβ mediated blockage of inflammation play a role in AD development.
LTP appears to be dependent on tau protein as Reactive astrocytes, microglia, and elevated lev-
acute slice preparation from tau knockout els of proinflammatory molecules span across the
mice prevented the Aβ mediated blockage of course of the disease that may influence either Aβ
LTP [110]. These studies suggest Aβ and tau plaque clearance or deposition and promote tau
work together to drive the neuronal abnormali- pathogenesis. Further evidence has come from
ties, and the combination of both induces syn- whole exome sequencing studies that have impli-
aptic dysfunction in vivo. cated the innate immune system as a critical com-
The dependence of tau in Aβ-induced neu- ponent of AD pathogenesis. Most notably, are the
rotoxicity is also evident in transgenic mice recent findings of rare variants in the microglial-­
expressing familial AD mutations in hAPP expressed gene TREM2 (Triggering Receptor
[111, 112]. One line of transgenic hAPP mice Expressed on Myeloid Cells2) that increase the
display aged dependent deposition of Aβ risk of developing AD by 2–4 fold [114, 115].
plaques, spatial learning deficits, hyperactive, The expression of TREM2 in microglial appears
and premature death. The genetic crossing of to regulate the microglial-mediated phagocytic
hAPP mice for either homozygous or heterozy- clearance of cellular debris and their responsive-
gous null of the tau allele protects against these ness to brain insults [116, 117]. Several studies
behavioral abnormalities. There is a restoration from mouse models of amyloidosis and postmor-
of learning and memory deficits and survival is tem human brain sections have demonstrated
prolonged without altering the Aβ deposition. TREM2 variants associated with AD, attenuate
These studies imply that over-expression of the microglia responsiveness to Aβ plaques sug-
APP or one of its products such as Aβ initiates gesting these variants results in a loss of function
a pathophysiological cascade of events leading of TREM2 [118–121]. Notably, the loss of
to behavioral abnormalities that are dependent microglia surrounding the plaques seems to pro-
on tau protein. In addition, tau knockout mice mote diffuse star-shaped amyloid fibrillary
appear to be ­protected against PTZ induced plaques that tend to be associated with neuritic
seizure activity in wild-type APP mice, consis- dystrophy [120, 121]. Thus, potentially early in
tent with the notion that tau protein is capable disease the microglial responsiveness to Aβ
of modulating synaptic activity. In a more plaques plays a protective role by lessening the
recent study, the hAPP mice expressing human impact of Aβ-induced neurotoxicity by trimming
mutant PS1 that display memory impairments, plaques or forming a barrier between the plaques
loss of synapses and premature death were also and the surrounding neurons. However, addi-
rescued by homozygous tau knockout. tional studies have suggested microglia poten-
However, this particular study differed in tially influence the spread of tau pathology by
which deletion of the tau gene decreased the phagocytosing and exocytosing tau protein [122].
Aβ deposition, suggesting tau pathology may Moreover, microglia deficient in the microglia
influence the amyloid burden [113]. fractalkine receptor that promotes microglia acti-
vation appears to facilitate tau pathology in tau
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 197

transgenic mice [123, 124]. While the role of genic mice expressing human tau [56, 132].
microglial response in AD is complex, microglial Although, these studies demonstrate a compel-
appear to be influencing Aβ deposition early in ling association between AD and reactive astro-
AD development potentially in a protective role, cytes, their impact on disease pathogenesis and
as the disease progresses the amyloid plaque bur- their therapeutic potential remains largely
den exceeds a threshold that induces tau pathol- unknown.
ogy promoting a pathological role for microglia
in the spreading of tau pathology.
Reactive astrocytes are also a well-known fea- Concluding Remarks
ture of AD that co-localize with Aβ plaques,
NFT, and brain atrophy in postmortem brain tis- Overwhelming evidence suggests AD is a dual
sue [119, 125]. Both Aβ and tau pathology are proteinopathy in which Aβ deposition and the
also commonly found intracellularly in astro- accumulation of aggregated tau drive AD patho-
cytes and in non-AD dementias, astrocytes may genesis. Although there is much clarity needed
contain tau pathology with distinct characteris- before fully understanding the relationship
tics that are pathological hallmarks for many of between Aβ and tau, it still appears amyloid-β
the primary tauopathies [126, 127]. Although, the pathophysiological effects occur early in AD
mechanisms for either Aβ/tau uptake or the cel- development, which seems necessary but not suf-
lular effects for the accumulation of pathological ficient for priming vulnerable brain regions for
Aβ/tau in astrocytes remains mostly unexplored. the intrusion of tau pathology. It is likely the Aβ
However, studies have illustrated the exposure of deposition exacerbates tau aggregation despite
Aβ fibrils to astrocytes activates an NFƙB-­ the temporal and spatial separation, which plays
mediated response that releases the complement an important role in damaging axons, dendrites,
protein C3, which is capable of disrupting den- synapses, and cellular function and that ulti-
dritic morphology [128]. Both the activation of mately leads to neuronal cell death.
NFƙB and C3 protein are associated with AD In addition, the recent advances in basic
development and the inhibition of C3  in mouse research, genetics, and biomarkers are demon-
models of amyloidosis reduce synapse loss and strating AD is a multifaceted degenerative pro-
rescues cognitive impairments [128–130]. cess with potentially several pathophysiological
Additionally, recent studies have identified a mechanisms. Likely, targeting one protein, one
minimum of two different forms of reactive gene or one cell type that is selective for one
astrocytes based on gene expression profiles pathophysiological mechanism may not suffice
[131]. One subtype of reactive astrocytes appears to prevent the degenerative process of
to display a potential reparative role, while a sec- AD. Therefore, it is imperative for understanding
ond subtype adopts an aggressive toxic gain of the molecular mechanism that regulates Aβ depo-
function characterized by an upregulation of sev- sition, its influences on tau pathology and the
eral complement cascade genes and neurotoxins spreading of tauopathy. Moreover, as emerging
detrimental to synaptic integrity, neuronal health, evidence now suggests neuroinflammation plays
and survival. These disease-associated astrocytes a prominent role in the neurodegenerative pro-
termed “A1”, are present in postmortem AD brain cess of AD and altering the neuroinflammatory
tissue and have been recently described in trans- response may be beneficial (Fig. 16.1).
198 G. Gallardo and D. M. Holtzman

Fig. 16.1  Model of Alzheimer’s disease pathology and ease progresses the Aβ plaque deposition reaches a
progression. (a) The first pathological event in AD devel- threshold forming the dense-core plaques that are associ-
opment is the production of the Aβ42 peptide produced ate with dystrophic neurites and influences tau pathogen-
from the transmembrane protein APP via cleavage by two esis. In late-stage AD, disease-associated microglia may
enzymes, β-secretase and γ-secretase. (b, c) Extracellular facilitate tau pathogenesis by phagocytosing and exocy-
Aβ42 peptide forms Aβ42 oligomers and initiates Aβ tosing tau protein and promoting a neurotoxic inflamma-
plaque deposition. The Aβ plaque formation leads to a tory response. In addition, during the late-stage of disease
microglial response, which surround the plaque and the disease-associated astrocytes exacerbate tau pathol-
recruits astrocytes. The glia response appears to be protec- ogy and neurodegeneration by potentially secreting neu-
tive by altering Aβ plaque structure and forming a barrier rotoxic factors. (e) Tau pathology spreads to connect brain
between the plaques and the surrounding neurons lessen- regions, which drives the neurodegenerative process
ing the impact Aβ-induced neurotoxicity. (d) As the dis-

6. Wisniewski KE, Wisniewski HM, Wen

References GY.  Occurrence of neuropathological changes and
dementia of Alzheimer’s disease in Down’s syn-
1. Maurer K, Volk S, Gerbaldo H.  Auguste drome. Ann Neurol. 1985;17(3):278–82.
D and Alzheimer’s disease. Lancet. 7. Goldgaber D, Lerman MI, McBride OW, Saffiotti
1997;349(9064):1546–9. U, Gajdusek DC.  Characterization and chro-
2. Glenner GG, Wong CW.  Alzheimer’s disease and mosomal localization of a cDNA encoding
Down’s syndrome: sharing of a unique cerebrovas- brain amyloid of Alzheimer’s disease. Science.
cular amyloid fibril protein. Biochem Biophys Res 1987;235(4791):877–80.
Commun. 1984;122(3):1131–5. 8. Kang J, Lemaire HG, Unterbeck A, Salbaum
3. Glenner GG, Wong CW, Quaranta V, Eanes ED. The JM, Masters CL, Grzeschik KH, et  al. The pre-
amyloid deposits in Alzheimer’s disease: their nature cursor of Alzheimer’s disease amyloid A4 pro-
and pathogenesis. Appl Pathol. 1984;2(6):357–69. tein resembles a cell-surface receptor. Nature.
4. Masters CL, Simms G, Weinman NA, Multhaup G, 1987;325(6106):733–6.
McDonald BL, Beyreuther K. Amyloid plaque core 9. Hardy JA, Higgins GA.  Alzheimer’s dis-
protein in Alzheimer disease and Down syndrome. ease: the amyloid cascade hypothesis. Science.
Proc Natl Acad Sci U S A. 1985;82(12):4245–9. 1992;256(5054):184–5.
5. Lai F, Williams RS.  A prospective study of 10. Karran E, Mercken M, De Strooper B.  The amy-
Alzheimer disease in Down syndrome. Arch Neurol. loid cascade hypothesis for Alzheimer’s disease: an
1989;46(8):849–53.
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 199

appraisal for the development of therapeutics. Nat 24. Braak H, Braak E.  Neuropil threads occur in den-
Rev Drug Discov. 2011;10(9):698–712. drites of tangle-bearing nerve cells. Neuropathol
11. Perrin RJ, Fagan AM, Holtzman Appl Neurobiol. 1988;14(1):39–44.
DM.  Multimodal techniques for diagnosis 25. Perry G, Kawai M, Tabaton M, Onorato M,
and prognosis of Alzheimer’s disease. Nature. Mulvihill P, Richey P, et  al. Neuropil threads
2009;461(7266):916–22. of Alzheimer’s disease show a marked altera-
12. Dickson TC, Vickers JC.  The morphological phe- tion of the normal cytoskeleton. J  Neurosci.
notype of beta-amyloid plaques and associated neu- 1991;11(6):1748–55.
ritic changes in Alzheimer’s disease. Neuroscience. 26. Schmidt ML, Gur RE, Gur RC, Trojanowski
2001;105(1):99–107. JQ.  Intraneuronal and extracellular neurofibrillary
13. Gomez-Isla T, Spires T, De Calignon A, Hyman tangles exhibit mutually exclusive cytoskeletal anti-
BT. Neuropathology of Alzheimer’s disease. Handb gens. Ann Neurol. 1988;23(2):184–9.
Clin Neurol. 2008;89:233–43. 27. Van Broeckhoven C, Haan J, Bakker E, Hardy
14. Serrano-Pozo A, Frosch MP, Masliah E, Hyman JA, Van Hul W, Wehnert A, et  al. Amyloid beta
BT.  Neuropathological alterations in Alzheimer protein precursor gene and hereditary cerebral
disease. Cold Spring Harb Perspect Med. hemorrhage with amyloidosis (Dutch). Science.
2011;1(1):a006189. 1990;248(4959):1120–2.
15. Glenner GG, Wong CW. Alzheimer’s disease: initial 28. Levy E, Carman MD, Fernandez-Madrid IJ, Power
report of the purification and characterization of a MD, Lieberburg I, van Duinen SG, et al. Mutation
novel cerebrovascular amyloid protein. Biochem of the Alzheimer’s disease amyloid gene in heredi-
Biophys Res Commun. 1984;120(3):885–90. tary cerebral hemorrhage, Dutch type. Science.
16. Kidd M.  Paired helical filaments in electron 1990;248(4959):1124–6.
microscopy of Alzheimer’s disease. Nature. 29. Goate A, Chartier-Harlin MC, Mullan M, Brown J,
1963;197:192–3. Crawford F, Fidani L, et  al. Segregation of a mis-
17. Joachim CL, Morris JH, Selkoe DJ, Kosik KS. Tau sense mutation in the amyloid precursor protein
epitopes are incorporated into a range of lesions in gene with familial Alzheimer’s disease. Nature.
Alzheimer’s disease. J  Neuropathol Exp Neurol. 1991;349(6311):704–6.
1987;46(6):611–22. 30. Chartier-Harlin MC, Crawford F, Houlden H,
18. Witman GB, Cleveland DW, Weingarten MD, Warren A, Hughes D, Fidani L, et  al. Early-onset
Kirschner MW. Tubulin requires tau for growth onto Alzheimer’s disease caused by mutations at codon
microtubule initiating sites. Proc Natl Acad Sci U S 717 of the beta-amyloid precursor protein gene.
A. 1976;73(11):4070–4. Nature. 1991;353(6347):844–6.
19. Lee VM, Goedert M, Trojanowski 31. Murrell J, Farlow M, Ghetti B, Benson MD.  A
JQ.  Neurodegenerative tauopathies. Annu Rev mutation in the amyloid precursor protein associ-
Neurosci. 2001;24:1121–59. ated with hereditary Alzheimer’s disease. Science.
20. Biernat J, Mandelkow EM, Schroter C, Lichtenberg-­ 1991;254(5028):97–9.
Kraag B, Steiner B, Berling B, et  al. The switch 32. Haass C, Hung AY, Selkoe DJ, Teplow
of tau protein to an Alzheimer-like state includes DB.  Mutations associated with a locus for familial
the phosphorylation of two serine-proline motifs Alzheimer’s disease result in alternative process-
upstream of the microtubule binding region. EMBO ing of amyloid beta-protein precursor. J Biol Chem.
J. 1992;11(4):1593–7. 1994;269(26):17741–8.
21. Goedert M, Jakes R, Crowther RA, Cohen P, 33. Suzuki N, Cheung TT, Cai XD, Odaka A, Otvos
Vanmechelen E, Vandermeeren M, et  al. Epitope L Jr, Eckman C, et  al. An increased percentage of
mapping of monoclonal antibodies to the paired long amyloid beta protein secreted by familial amy-
helical filaments of Alzheimer’s disease: identifica- loid beta protein precursor (beta APP717) mutants.
tion of phosphorylation sites in tau protein. Biochem Science. 1994;264(5163):1336–40.
J. 1994;301(Pt 3):871–7. 34. Mullan M, Crawford F, Axelman K, Houlden H,
22. Matsuo ES, Shin RW, Billingsley ML, Van de Lilius L, Winblad B, et  al. A pathogenic mutation
Voorde A, O’Connor M, Trojanowski JQ, et  al. for probable Alzheimer’s disease in the APP gene
Biopsy-­derived adult human brain tau is phosphor- at the N-terminus of beta-amyloid. Nat Genet.
ylated at many of the same sites as Alzheimer’s 1992;1(5):345–7.
disease paired helical filament tau. Neuron. 35. Cai XD, Golde TE, Younkin SG. Release of excess
1994;13(4):989–1002. amyloid beta protein from a mutant amyloid beta
23. Braak H, Braak E, Grundke-Iqbal I, Iqbal protein precursor. Science. 1993;259(5094):514–6.
K.  Occurrence of neuropil threads in the senile 36. Citron M, Oltersdorf T, Haass C, McConlogue L,
human brain and in Alzheimer’s disease: a third Hung AY, Seubert P, et  al. Mutation of the beta-­
location of paired helical filaments outside of neuro- amyloid precursor protein in familial Alzheimer’s
fibrillary tangles and neuritic plaques. Neurosci Lett. disease increases beta-protein production. Nature.
1986;65(3):351–5. 1992;360(6405):672–4.
200 G. Gallardo and D. M. Holtzman

37. Rovelet-Lecrux A, Hannequin D, Raux G, Le Meur Notch pathway signaling, gamma-secretase cleavage
N, Laquerriere A, Vital A, et  al. APP locus dupli- of betaAPP, and presenilin protein accumulation.
cation causes autosomal dominant early-onset Dev Cell. 2002;3(1):85–97.
Alzheimer disease with cerebral amyloid angiopa- 50. Strittmatter WJ, Saunders AM, Schmechel D,
thy. Nat Genet. 2006;38(1):24–6. Pericak-Vance M, Enghild J, Salvesen GS, et  al.
38. Jonsson T, Atwal JK, Steinberg S, Snaedal J, Jonsson Apolipoprotein E: high-avidity binding to beta-­
PV, Bjornsson S, et al. A mutation in APP protects amyloid and increased frequency of type 4 allele
against Alzheimer’s disease and age-related cogni- in late-onset familial Alzheimer disease. Proc Natl
tive decline. Nature. 2012;488(7409):96–9. Acad Sci U S A. 1993;90(5):1977–81.
39. Nilsberth C, Westlind-Danielsson A, Eckman CB, 51. Schmechel DE, Saunders AM, Strittmatter WJ,
Condron MM, Axelman K, Forsell C, et  al. The Crain BJ, Hulette CM, Joo SH, et al. Increased amy-
‘Arctic’ APP mutation (E693G) causes Alzheimer’s loid beta-peptide deposition in cerebral cortex as a
disease by enhanced Abeta protofibril formation. consequence of apolipoprotein E genotype in late-­
Nat Neurosci. 2001;4(9):887–93. onset Alzheimer disease. Proc Natl Acad Sci U S A.
40. Walsh DM, Klyubin I, Fadeeva JV, Cullen WK, 1993;90(20):9649–53.
Anwyl R, Wolfe MS, et al. Naturally secreted oligo- 52. Corder EH, Saunders AM, Strittmatter WJ,
mers of amyloid beta protein potently inhibit hip- Schmechel DE, Gaskell PC, Small GW, et al. Gene
pocampal long-term potentiation in  vivo. Nature. dose of apolipoprotein E type 4 allele and the risk of
2002;416(6880):535–9. Alzheimer’s disease in late onset families. Science.
41. Dodart JC, Bales KR, Gannon KS, Greene SJ, 1993;261(5123):921–3.
DeMattos RB, Mathis C, et  al. Immunization 53. Genin E, Hannequin D, Wallon D, Sleegers K,
reverses memory deficits without reducing brain Hiltunen M, Combarros O, et  al. APOE and
Abeta burden in Alzheimer’s disease model. Nat Alzheimer disease: a major gene with semi-­dominant
Neurosci. 2002;5(5):452–7. inheritance. Mol Psychiatry. 2011;16(9):903–7.
42. Lesne S, Koh MT, Kotilinek L, Kayed R, Glabe 54. Holtzman DM, Fagan AM, Mackey B, Tenkova T,
CG, Yang A, et  al. A specific amyloid-beta protein Sartorius L, Paul SM, et al. Apolipoprotein E facili-
assembly in the brain impairs memory. Nature. tates neuritic and cerebrovascular plaque forma-
2006;440(7082):352–7. tion in an Alzheimer’s disease model. Ann Neurol.
43. Scheuner D, Eckman C, Jensen M, Song X, Citron 2000;47(6):739–47.
M, Suzuki N, et  al. Secreted amyloid beta-protein 55. Fagan AM, Watson M, Parsadanian M, Bales KR,
similar to that in the senile plaques of Alzheimer’s Paul SM, Holtzman DM. Human and murine ApoE
disease is increased in vivo by the presenilin 1 and markedly alters A beta metabolism before and after
2 and APP mutations linked to familial Alzheimer’s plaque formation in a mouse model of Alzheimer’s
disease. Nat Med. 1996;2(8):864–70. disease. Neurobiol Dis. 2002;9(3):305–18.
44. Sherrington R, Rogaev EI, Liang Y, Rogaeva 56. Shi Y, Yamada K, Liddelow SA, Smith ST, Zhao
EA, Levesque G, Ikeda M, et  al. Cloning of L, Luo W, et  al. ApoE4 markedly exacerbates tau-­
a gene bearing missense mutations in early-­ mediated neurodegeneration in a mouse model of
onset familial Alzheimer’s disease. Nature. tauopathy. Nature. 2017;549(7673):523–7.
1995;375(6534):754–60. 57. Bateman RJ, Xiong C, Benzinger TL, Fagan AM,
45. Rogaev EI, Sherrington R, Rogaeva EA, Levesque Goate A, Fox NC, et  al. Clinical and biomarker
G, Ikeda M, Liang Y, et al. Familial Alzheimer’s dis- changes in dominantly inherited Alzheimer’s dis-
ease in kindreds with missense mutations in a gene ease. N Engl J Med. 2012;367(9):795–804.
on chromosome 1 related to the Alzheimer’s disease 58. Fagan AM, Mintun MA, Mach RH, Lee SY, Dence
type 3 gene. Nature. 1995;376(6543):775–8. CS, Shah AR, et al. Inverse relation between in vivo
46. Citron M, Westaway D, Xia W, Carlson G, Diehl T, amyloid imaging load and cerebrospinal fluid
Levesque G, et al. Mutant presenilins of Alzheimer’s Abeta42 in humans. Ann Neurol. 2006;59(3):512–9.
disease increase production of 42-residue amyloid 59. Jack CR Jr, Vemuri P, Wiste HJ, Weigand SD, Aisen
beta-protein in both transfected cells and transgenic PS, Trojanowski JQ, et  al. Evidence for ordering
mice. Nat Med. 1997;3(1):67–72. of Alzheimer disease biomarkers. Arch Neurol.
47. Wolfe MS, Xia W, Ostaszewski BL, Diehl TS, 2011;68(12):1526–35.
Kimberly WT, Selkoe DJ.  Two transmembrane 60. Fagan AM, Roe CM, Xiong C, Mintun MA,
aspartates in presenilin-1 required for presenilin Morris JC, Holtzman DM.  Cerebrospinal fluid tau/
endoproteolysis and gamma-secretase activity. beta-­amyloid(42) ratio as a prediction of cognitive
Nature. 1999;398(6727):513–7. decline in nondemented older adults. Arch Neurol.
48. Yu G, Nishimura M, Arawaka S, Levitan D, Zhang 2007;64(3):343–9.
L, Tandon A, et al. Nicastrin modulates presenilin-­ 61. Jack CR Jr, Knopman DS, Jagust WJ, Shaw LM,
mediated notch/glp-1 signal transduction and beta- Aisen PS, Weiner MW, et al. Hypothetical model of
APP processing. Nature. 2000;407(6800):48–54. dynamic biomarkers of the Alzheimer’s pathological
49. Francis R, McGrath G, Zhang J, Ruddy DA, Sym cascade. Lancet Neurol. 2010;9(1):119–28.
M, Apfeld J, et al. aph-1 and pen-2 are required for
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 201

62. Jagust WJ, Landau SM, Shaw LM, Trojanowski but exceeds neurofibrillary tangles in Alzheimer’s
JQ, Koeppe RA, Reiman EM, et  al. Relationships disease. Ann Neurol. 1997;41(1):17–24.
between biomarkers in aging and dementia. 76. Arriagada PV, Marzloff K, Hyman BT. Distribution
Neurology. 2009;73(15):1193–9. of Alzheimer-type pathologic changes in nonde-
63. Doraiswamy PM, Sperling RA, Coleman RE, mented elderly individuals matches the pattern in
Johnson KA, Reiman EM, Davis MD, et  al. Alzheimer’s disease. Neurology. 1992;42(9):1681–8.
Amyloid-beta assessed by florbetapir F 18 PET and 77. Price JL, Davis PB, Morris JC, White DL. The distri-
18-month cognitive decline: a multicenter study. bution of tangles, plaques and related immunohisto-
Neurology. 2012;79(16):1636–44. chemical markers in healthy aging and Alzheimer’s
64. Doraiswamy PM, Sperling RA, Johnson K, Reiman disease. Neurobiol Aging. 1991;12(4):295–312.
EM, Wong TZ, Sabbagh MN, et  al. Florbetapir F 78. Price JL, Morris JC. Tangles and plaques in nonde-
18 amyloid PET and 36-month cognitive decline: mented aging and "preclinical" Alzheimer’s disease.
a prospective multicenter study. Mol Psychiatry. Ann Neurol. 1999;45(3):358–68.
2014;19(9):1044–51. 79. Johnson KA, Schultz A, Betensky RA, Becker JA,
65. Ong K, Villemagne VL, Bahar-Fuchs A, Lamb F, Sepulcre J, Rentz D, et  al. Tau positron emission
Chetelat G, Raniga P, et al. (18)F-florbetaben Abeta tomographic imaging in aging and early Alzheimer
imaging in mild cognitive impairment. Alzheimers disease. Ann Neurol. 2016;79(1):110–9.
Res Ther. 2013;5(1):4. 80. Brier MR, Gordon B, Friedrichsen K, McCarthy J,
66. Sabuncu MR, Desikan RS, Sepulcre J, Yeo BT, Liu Stern A, Christensen J, et  al. Tau and Abeta imag-
H, Schmansky NJ, et  al. The dynamics of cortical ing, CSF measures, and cognition in Alzheimer’s
and hippocampal atrophy in Alzheimer disease. disease. Sci Transl Med. 2016;8(338):338ra66.
Arch Neurol. 2011;68(8):1040–8. 81. Bejanin A, Schonhaut DR, La Joie R, Kramer JH,
67. Chetelat G, Villemagne VL, Villain N, Jones G, Ellis Baker SL, Sosa N, et al. Tau pathology and neuro-
KA, Ames D, et al. Accelerated cortical atrophy in degeneration contribute to cognitive impairment in
cognitively normal elderly with high beta-amyloid Alzheimer’s disease. Brain. 2017;140(12):3286–300.
deposition. Neurology. 2012;78(7):477–84. 82. Braak H, Del Tredici K.  The pathological process
68. Chetelat G, Villemagne VL, Bourgeat P, Pike KE, underlying Alzheimer’s disease in individuals under
Jones G, Ames D, et al. Relationship between atro- thirty. Acta Neuropathol. 2011;121(2):171–81.
phy and beta-amyloid deposition in Alzheimer dis- 83. Price JL, McKeel DW Jr, Buckles VD, Roe CM,
ease. Ann Neurol. 2010;67(3):317–24. Xiong C, Grundman M, et  al. Neuropathology
69. Archer HA, Edison P, Brooks DJ, Barnes J, Frost of nondemented aging: presumptive evidence for
C, Yeatman T, et  al. Amyloid load and cerebral preclinical Alzheimer disease. Neurobiol Aging.
atrophy in Alzheimer’s disease: an 11C-PIB posi- 2009;30(7):1026–36.
tron emission tomography study. Ann Neurol. 84. Nelson PT, Braak H, Markesbery
2006;60(1):145–7. WR.  Neuropathology and cognitive impairment in
70. Andrews KA, Frost C, Modat M, Cardoso MJ, Rowe Alzheimer disease: a complex but coherent relation-
CC, Villemagne V, et al. Acceleration of hippocam- ship. J Neuropathol Exp Neurol. 2009;68(1):1–14.
pal atrophy rates in asymptomatic amyloidosis. 85. Vos SJ, Xiong C, Visser PJ, Jasielec MS, Hassenstab
Neurobiol Aging. 2016;39:99–107. J, Grant EA, et  al. Preclinical Alzheimer’s disease
71. Dore V, Villemagne VL, Bourgeat P, Fripp J, Acosta and its outcome: a longitudinal cohort study. Lancet
O, Chetelat G, et al. Cross-sectional and longitudinal Neurol. 2013;12(10):957–65.
analysis of the relationship between Abeta deposi- 86. Roe CM, Fagan AM, Grant EA, Hassenstab J,
tion, cortical thickness, and memory in cognitively Moulder KL, Maue Dreyfus D, et  al. Amyloid
unimpaired individuals and in Alzheimer disease. imaging and CSF biomarkers in predicting cogni-
JAMA Neurol. 2013;70(7):903–11. tive impairment up to 7.5 years later. Neurology.
72. Sperling RA, Johnson KA, Doraiswamy PM, 2013;80(19):1784–91.
Reiman EM, Fleisher AS, Sabbagh MN, et  al. 87. Villemagne VL, Burnham S, Bourgeat P, Brown B,
Amyloid deposition detected with florbetapir F 18 Ellis KA, Salvado O, et al. Amyloid beta deposition,
((18)F-AV-45) is related to lower episodic memory neurodegeneration, and cognitive decline in sporadic
performance in clinically normal older individuals. Alzheimer’s disease: a prospective cohort study.
Neurobiol Aging. 2013;34(3):822–31. Lancet Neurol. 2013;12(4):357–67.
73. Andrews KA, Modat M, Macdonald KE, Yeatman T, 88. Chen X, Li M, Wang S, Zhu H, Xiong Y, Liu
Cardoso MJ, Leung KK, et al. Atrophy rates in asymp- X. Pittsburgh compound B retention and progression
tomatic amyloidosis: implications for Alzheimer of cognitive status--a meta-analysis. Eur J  Neurol.
prevention trials. PLoS One. 2013;8(3):e58816. 2014;21(8):1060–7.
74. Braak H, Braak E.  Neuropathological stageing 89. Donohue MC, Sperling RA, Petersen R, Sun
of Alzheimer-related changes. Acta Neuropathol. CK, Weiner MW, Aisen PS.  Association between
1991;82(4):239–59. elevated brain amyloid and subsequent cognitive
75. Gomez-Isla T, Hollister R, West H, Mui S, Growdon decline among cognitively normal persons. JAMA.
JH, Petersen RC, et al. Neuronal loss correlates with 2017;317(22):2305–16.
202 G. Gallardo and D. M. Holtzman

90. Combs B, Hamel C, Kanaan NM.  Pathological 103. Guo JL, Narasimhan S, Changolkar L, He Z,
conformations involving the amino terminus of tau Stieber A, Zhang B, et  al. Unique pathological
occur early in Alzheimer’s disease and are differen- tau conformers from Alzheimer’s brains transmit
tially detected by monoclonal antibodies. Neurobiol tau pathology in nontransgenic mice. J  Exp Med.
Dis. 2016;94:18–31. 2016;213(12):2635–54.
91. Kanaan NM, Morfini G, Pigino G, LaPointe NE, 104. He Z, Guo JL, McBride JD, Narasimhan S, Kim H,
Andreadis A, Song Y, et  al. Phosphorylation in Changolkar L, et al. Amyloid-beta plaques enhance
the amino terminus of tau prevents inhibition of Alzheimer’s brain tau-seeded pathologies by facili-
anterograde axonal transport. Neurobiol Aging. tating neuritic plaque tau aggregation. Nat Med.
2012;33(4):826.e15–30. 2018;24(1):29–38.
92. Abraha A, Ghoshal N, Gamblin TC, Cryns V, Berry 105. Busciglio J, Lorenzo A, Yeh J, Yankner BA. beta-­
RW, Kuret J, et  al. C-terminal inhibition of tau amyloid fibrils induce tau phosphorylation and loss
assembly in vitro and in Alzheimer’s disease. J Cell of microtubule binding. Neuron. 1995;14(4):879–88.
Sci. 2000;113(Pt 21):3737–45. 106. Arioka M, Tsukamoto M, Ishiguro K, Kato R, Sato
93. Clavaguera F, Bolmont T, Crowther RA, Abramowski K, Imahori K, et al. Tau protein kinase II is involved
D, Frank S, Probst A, et al. Transmission and spread- in the regulation of the normal phosphorylation state
ing of tauopathy in transgenic mouse brain. Nat Cell of tau protein. J Neurochem. 1993;60(2):461–8.
Biol. 2009;11(7):909–13. 107. Rapoport M, Dawson HN, Binder LI, Vitek MP,
94. Iba M, Guo JL, McBride JD, Zhang B, Trojanowski Ferreira A.  Tau is essential to beta -amyloid-­
JQ, Lee VM. Synthetic tau fibrils mediate transmis- induced neurotoxicity. Proc Natl Acad Sci U S A.
sion of neurofibrillary tangles in a transgenic mouse 2002;99(9):6364–9.
model of Alzheimer’s-like tauopathy. J  Neurosci. 108. Zempel H, Luedtke J, Kumar Y, Biernat J, Dawson
2013;33(3):1024–37. H, Mandelkow E, et  al. Amyloid-beta oligomers
95. Gomez-Isla T, Price JL, McKeel DW Jr, Morris induce synaptic damage via Tau-dependent micro-
JC, Growdon JH, Hyman BT.  Profound loss tubule severing by TTLL6 and spastin. EMBO
of layer II entorhinal cortex neurons occurs J. 2013;32(22):2920–37.
in very mild Alzheimer’s disease. J  Neurosci. 109. King ME, Kan HM, Baas PW, Erisir A, Glabe CG,
1996;16(14):4491–500. Bloom GS. Tau-dependent microtubule disassembly
96. West MJ, Coleman PD, Flood DG, Troncoso initiated by prefibrillar beta-amyloid. J  Cell Biol.
JC. Differences in the pattern of hippocampal neu- 2006;175(4):541–6.
ronal loss in normal ageing and Alzheimer’s disease. 110. Shipton OA, Leitz JR, Dworzak J, Acton CE,
Lancet. 1994;344(8925):769–72. Tunbridge EM, Denk F, et  al. Tau protein is
97. Mattsson N, Insel PS, Nosheny R, Tosun D, required for amyloid {beta}-induced impairment
Trojanowski JQ, Shaw LM, et  al. Emerging beta-­ of hippocampal long-term potentiation. J Neurosci.
amyloid pathology and accelerated cortical atrophy. 2011;31(5):1688–92.
JAMA Neurol. 2014;71(6):725–34. 111. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A,
98. Gotz J, Chen F, van Dorpe J, Nitsch RM. Formation van Eersel J, et al. Dendritic function of tau mediates
of neurofibrillary tangles in P301l tau trans- amyloid-beta toxicity in Alzheimer’s disease mouse
genic mice induced by Abeta 42 fibrils. Science. models. Cell. 2010;142(3):387–97.
2001;293(5534):1491–5. 112. Roberson ED, Scearce-Levie K, Palop JJ, Yan
99. Bolmont T, Clavaguera F, Meyer-Luehmann F, Cheng IH, Wu T, et  al. Reducing endogenous
M, Herzig MC, Radde R, Staufenbiel M, et  al. tau ameliorates amyloid beta-induced deficits in
Induction of tau pathology by intracerebral infu- an Alzheimer’s disease mouse model. Science.
sion of amyloid-­beta -containing brain extract and 2007;316(5825):750–4.
by amyloid-beta deposition in APP x Tau transgenic 113. Leroy K, Ando K, Laporte V, Dedecker R, Suain
mice. Am J Pathol. 2007;171(6):2012–20. V, Authelet M, et  al. Lack of tau proteins rescues
100. Lewis J, Dickson DW, Lin WL, Chisholm L, Corral neuronal cell death and decreases amyloidogenic
A, Jones G, et  al. Enhanced neurofibrillary degen- processing of APP in APP/PS1 mice. Am J Pathol.
eration in transgenic mice expressing mutant tau and 2012;181(6):1928–40.
APP. Science. 2001;293(5534):1487–91. 114. Guerreiro R, Wojtas A, Bras J, Carrasquillo M,
101. Hurtado DE, Molina-Porcel L, Iba M, Aboagye AK, Rogaeva E, Majounie E, et  al. TREM2 vari-
Paul SM, Trojanowski JQ, et al. A{beta} accelerates ants in Alzheimer’s disease. N Engl J  Med.
the spatiotemporal progression of tau pathology and 2013;368(2):117–27.
augments tau amyloidosis in an Alzheimer mouse 115. Jonsson T, Stefansson H, Steinberg S, Jonsdottir I,
model. Am J Pathol. 2010;177(4):1977–88. Jonsson PV, Snaedal J, et al. Variant of TREM2 asso-
102. Bennett RE, DeVos SL, Dujardin S, Corjuc B, Gor ciated with the risk of Alzheimer’s disease. N Engl
R, Gonzalez J, et  al. Enhanced Tau Aggregation J Med. 2013;368(2):107–16.
in the Presence of Amyloid beta. Am J  Pathol. 116. Takahashi K, Rochford CD, Neumann H. Clearance
2017;187(7):1601–12. of apoptotic neurons without inflammation by
16  Amyloid-β and Tau at the Crossroads of Alzheimer’s Disease 203

microglial triggering receptor expressed on myeloid 125. Nagele RG, Wegiel J, Venkataraman V, Imaki H,
cells-2. J Exp Med. 2005;201(4):647–57. Wang KC.  Contribution of glial cells to the devel-
117. Piccio L, Buonsanti C, Mariani M, Cella M, Gilfillan opment of amyloid plaques in Alzheimer’s disease.
S, Cross AH, et al. Blockade of TREM-2 exacerbates Neurobiol Aging. 2004;25(5):663–74.
experimental autoimmune encephalomyelitis. Eur 126. Kovacs GG, Ferrer I, Grinberg LT, Alafuzoff I,
J Immunol. 2007;37(5):1290–301. Attems J, Budka H, et al. Aging-related tau astrogli-
118. Ulrich JD, Finn MB, Wang Y, Shen A, Mahan TE, opathy (ARTAG): harmonized evaluation strategy.
Jiang H, et al. Altered microglial response to Abeta Acta Neuropathol. 2016;131(1):87–102.
plaques in APPPS1-21 mice heterozygous for 127. Kovacs GG, Robinson JL, Xie SX, Lee EB, Grossman
TREM2. Mol Neurodegener. 2014;9:20. M, Wolk DA, et al. Evaluating the patterns of aging-­
119. Wang Y, Cella M, Mallinson K, Ulrich JD, Young related tau astrogliopathy unravels novel insights
KL, Robinette ML, et al. TREM2 lipid sensing sus- into brain aging and neurodegenerative diseases.
tains the microglial response in an Alzheimer’s dis- J Neuropathol Exp Neurol. 2017;76(4):270–88.
ease model. Cell. 2015;160(6):1061–71. 128. Lian H, Yang L, Cole A, Sun L, Chiang AC, Fowler
120. Wang Y, Ulland TK, Ulrich JD, Song W, Tzaferis SW, et al. NFkappaB-activated astroglial release of
JA, Hole JT, et al. TREM2-mediated early microg- complement C3 compromises neuronal morphology
lial response limits diffusion and toxicity of amyloid and function associated with Alzheimer’s disease.
plaques. J Exp Med. 2016;213(5):667–75. Neuron. 2015;85(1):101–15.
121. Yuan P, Condello C, Keene CD, Wang Y, Bird TD, 129. Kaltschmidt B, Uherek M, Volk B, Baeuerle PA,
Paul SM, et al. TREM2 haplodeficiency in mice and Kaltschmidt C.  Transcription factor NF-kappaB is
humans impairs the microglia barrier function lead- activated in primary neurons by amyloid beta pep-
ing to decreased amyloid compaction and severe tides and in neurons surrounding early plaques from
axonal dystrophy. Neuron. 2016;92(1):252–64. patients with Alzheimer disease. Proc Natl Acad Sci
122. Asai H, Ikezu S, Tsunoda S, Medalla M, Luebke J, U S A. 1997;94(6):2642–7.
Haydar T, et al. Depletion of microglia and inhibi- 130. Hong S, Beja-Glasser VF, Nfonoyim BM, Frouin
tion of exosome synthesis halt tau propagation. Nat A, Li S, Ramakrishnan S, et  al. Complement and
Neurosci. 2015;18(11):1584–93. microglia mediate early synapse loss in Alzheimer
123. Bhaskar K, Konerth M, Kokiko-Cochran ON, mouse models. Science. 2016;352(6286):712–6.
Cardona A, Ransohoff RM, Lamb BT. Regulation of 131. Zamanian JL, Xu L, Foo LC, Nouri N, Zhou L,
tau pathology by the microglial fractalkine receptor. Giffard RG, et  al. Genomic analysis of reactive
Neuron. 2010;68(1):19–31. astrogliosis. J Neurosci. 2012;32(18):6391–410.
124. Maphis N, Xu G, Kokiko-Cochran ON, Jiang S, 132. Liddelow SA, Guttenplan KA, Clarke LE, Bennett
Cardona A, Ransohoff RM, et al. Reactive microglia FC, Bohlen CJ, Schirmer L, et al. Neurotoxic reac-
drive tau pathology and contribute to the spreading tive astrocytes are induced by activated microglia.
of pathological tau in the brain. Brain. 2015;138(Pt Nature. 2017;541(7638):481–7.
6):1738–55.
 Part IV
Tauopathies; Pathology, Drivers,
and Marker
 Myotonic Dystrophy: an RNA Toxic
Gain of Function Tauopathy? 17
Francisco Fernandez-Gomez, Helene Tran,
Claire-Marie Dhaenens, Marie-Laure 
Caillet-Boudin, Susanna Schraen-Maschke,
David Blum, Bernard Sablonnière, Valérie 
Buée-Scherrer, Luc Buee, and Nicolas Sergeant

 ystrophia Myotonica: an RNA Gain

D mal inherited genetic disease caused by an unsta-
of Toxic Function Disease ble expanded CTG repeat sequence located in the
3’UTR of the DMPK (dystrophia myotonica pro-
Myotonic Dystrophy (DM) also referred to as tein kinase) gene [8]. The normal allele contains
Steinert’s disease or DM of type 1 (DM1) is a 5–30 repeats whereas over 50 CTGs repeats the
rare inherited multisystemic genetic disease unstable mutation is pathogenic. A second gene
affecting multiple organs [33]. Worldwide in its mutation is responsible for myotonic dystrophy
most common form, the adult form, DM preva- of type 2 (DM2) also known as proximal myo-
lence is of 1 per 8000. Major symptoms include tonic myopathy (PROMM). The repeat expan-
myotonia, progressive muscle wasting, cardiac sion is a tetranucleotide motif (CCTG)n ranging
conduction defects, endocrine deficiencies as from 100 to 11,000 repeats in the first intron of
well as cognitive impairments. DM1 is an autoso- the CNBP gene also known as ZNF9 gene [53].
Both pathogenic repeat expansions are in non-­
coding region of unrelated genes. However,
DMPK and CNBP are ubiquitously expressed.
F. Fernandez-Gomez DMPK and CNBP expanded alleles are tran-
University of Lille, INSERM, CHU-Lille, UMR-S scribed into RNA but not translated into protein.
1172 – Alzheimer & Tauopathies, LabEx DISTALZ, Mutant RNAs containing expanded CUG(n > 50) or
Lille, France
CCUG(n > 75) repeats are defectively exported out-
Area of Pharmaceutical Technology. Group of side the cell nucleus and consequently accumu-
Cellular and Molecular Pharmacology, Department of
Pharmacology, University of Murcia. Murcia late into discrete intranuclear aggregates. Those
Research Institute of Health Sciences (IMIB-­ aggregates composed of mutated RNAs and pro-
Arrixaca), Murcia, Spain teins are referred to as nuclear foci. Expression of
H. Tran expanded CUG repeats in cells [4] or in trans-
Department of Neurology, Albert Sherman Center, genic animals [56, 76] is necessary and sufficient
University of Massachusetts Medical School,
to reproduce DM1 phenotype suggesting a “RNA
Worcester, MA, USA
toxic gain of function”.
C.-M. Dhaenens · M.-L. Caillet-Boudin
Nuclear foci composed of expanded CUGs
S. Schraen-Maschke · D. Blum · B. Sablonnière
V. Buée-Scherrer · L. Buee · N. Sergeant (*) sequester RNA-binding proteins belonging to the
University of Lille, INSERM, CHU-Lille, UMR-S MBNL family of splicing regulatory factors [61].
1172 – Alzheimer & Tauopathies, LabEx DISTALZ, Nuclear foci are thus labeled by fluorescent in
Lille, France
situ hybridization (FISH) using a CAG ­fluorescent
e-mail: nicolas.sergeant@inserm.fr

© Springer Nature Singapore Pte Ltd. 2019 207

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_17
208 F. Fernandez-Gomez et al.

RNA probes as well as using antibodies against sufficiency of DMPK expression, a defective
MBNL [24]. Sequestration of MBNL proteins expression of the surrounding genes such as
result in a loss of function and more especially in DMD and SIX5 gene expression, the expression
a splicing misregulation of specific transcripts of of antisense RNA, microRNA, … that together
tissues affected in DM1. Specific splicing contribute to the pathophysiology of DM1 (for
changes in DM1 affected tissues were mainly review see [83]). Moreover, in animal model of
attributable to the loss-of-function of MBNL1 in DM1 proteomics analyses of the brain also sug-
the skeletal muscle [42], and both MBNL1 and gest that several molecular deregulations are not
MBNL2 in the central nervous system and in the directly related to an RNA-gain of toxic function
heart supporting a functional loss of MBNL fam- [35]. Thus, the neurological determinism of DM1
ily of splicing factors as a central mechanism of and DM2 might be determined by pathological
RNA gain of toxic function of the CUG repeats in processes shared by these diseases but not yet
DM1 [13, 31]. completely unraveled.
This toxic gain-of-function at the RNA level is
considered as one of the main pathophysiological
hypotheses to explain the multisystemic develop-  Multisystemic Neuromuscular
A
ment of myotonic dystrophy. In addition to the Disease with Cognitive
sequestration inside foci of proteins such as Dysfunctions
members of the MBNL family of splicing factors
results in a partial loss-of-function of the MBNL Adie and collaborators reported in the beginning
family of splicing factors [42,  45], a gain-of-­ of the twentieth century the involvement of the
function of CUG-binding protein and ETR3-like central nervous system in myotonic dystrophy
factor 1 (CUGBP1/CELF1) in skeletal or heart [2]. It’s only in the 60’s that studies including a
muscle, through its stabilization and phosphory- systematic analysis of brain structures and defec-
lation by PKC, is also instrumental to DM1 tive cognitive functions were initiated  [28, 55].
pathophysiology [68, 84]. CELF1 and embryonic Clinical data available describe that DM1 patients
lethal abnormal vision-like RNA-binding pro- experience focused executive and visuoconstruc-
tein-­3 (ETR3/CELF2) like factors (CELF) belong tive progressive disabilities [55, 73, 92], frontal
to a family of RNA-binding proteins that includes cognitive impairment including attentional abil-
6 members. While both CELF1 and CELF2 are ity which are worsening with aging  [60, 72]
ubiquitously expressed, the latter is highly together with visual attention, verbal memory,
expressed in the brain and heart of rodents [47, language and executive functions  [25].
48]. The four other members, CELF3, 4, 5 and 6, Impairments of facial emotion recognition is a
are enriched in the rodent brain. However, less is cognitive deficit trait often reported in the adult
known about their expression in the human brain form of DM1 [46, 92]. Apart from being princi-
and their potential implication in DM1 pally a muscle disease, DM1 also include a heter-
pathophysiology. ogenous but systematic involvement of cognitive
Also related to the transcription of RNA bear- brain dysfunctions.
ing large tri- or tetranucleotide repeats, an aber-
rant translational process, named
“repeat-associated non-AUG translation” (RAN  rain Lesions in Myotonic
B
translation), lead to the translation of DM1 poly- Dystrophy
glutamine peptides, in human DM1 tissues and
mouse models (for review see [16]). However, in Neuroimaging in DM brain provided insights
DM1 the contribution of the polyamine to the into the effort to link the brain involvement with
pathology remains ill-defined. Other non-RNA the cognitive deficits. At the macroscopic level,
mediated toxic consequences are also reported in structural imaging studies using magnetic reso-
DM1. The expanded allele induces an haploin- nance (MRI) in adult DM1 revealed global
17  Myotonic Dystrophy: an RNA Toxic Gain of Function Tauopathy? 209

c­ erebral atrophy with dilated ventricular [6, 12]. neuropathological lesion found also in
Cortical gray matter loss is primarily observed in Alzheimer’s disease is described in DM1 brain.
the frontal, parietal, and occipital lobes and the This neuropathological lesion is the granovacuo-
superior and middle temporal gyrus, whereas the lar degeneration which are intracytoplasmic vac-
subcortical gray matter loss is detected in tha- uoles visualize insides neurons [64, 66, 95].
lamic and basal ganglia structures [6, 62, 63, 89]. Lewy bodies made of α-synuclein aggregates
There was no correlation between brain tissue were also reported in the substantia nigra of a
volumes and the grade of pathology, disease single case [64]. Interestingly, amyloid deposits
duration or even CTG expansion. However, the or amyloid angiopathy of diffuse amyloid deposi-
areas with grey matter loss are involved in cogni- tion have yet never been reported in DM1 or
tive dysfunctions and personality disorders, such DM2 [59].
as apathy, depression, anxiety, and deficits in
attention, memory, and visuospatial function [6].
MRI examination often reveal hyperintense white DM1 a Tau Missplicing Tauopahty
matter signals localized principally in the ante-
rior temporal lobe but were originally and etio- NFTs are intraneuronal fibrillar aggregates made
logically underestimated [30] and latter suggested of hyper- and abnormally phosphorylated iso-
to be correlated to cognitive impairments [1]. forms of microtubule-associated Tau [11, 75].
Improvement of brain imaging has provided Tau proteins belong to the family of microtubule-­
in depth and extensive description of brain lesions associated proteins. Tau functions are to promote
in DM1. Thus, a more widespread cortical and microtubule assembly, their stabilization and
subcortical involvement affecting gray and white their organization, necessary both for axonal
matter of all lobes, brainstem and cerebellum is transport and neurite outgrowth. Tau proteins are
reported (for review see [62]). Moreover, current encoded by a single MAPT gene located on chro-
structural and functional brain imaging network mosome 17. It is composed of of 15 exons num-
analyses also provide valuable information bered from 0 to 15 and a canonical and alternative
regarding the link between brain pathology and promoter [38]. In the central nervous system, six
cognitive impairments. However, up to now the major isoforms of Tau are expressed and pro-
relationship between brain imaging abnormali- duced by the alternative splicing of exons 2, 3
ties and cognitive deficits remains ill-defined and and 10 [11, 75]. Exon 10 splicing results in the
longitudinal studies such as those performed inclusion or exclusion of an additional
recently for cognitive impairments should be microtubule-­ binding sequences leading to Tau
considered [25]. isoforms with either three (3R-Tau) or four (4R-­
At the neuropathological level, intracellular Tau) microtubule binding repeats. Thus, Tau pro-
proteinaceous inclusions in neurons of the thala- teins with four microtubule-binding domains
mus have been reported in DM1 brain  [17, 67, (4R) bind more avidly and further stabilize
94]. However, the nature of these neuronal inclu- microtubules than those with three binding
sions remained unknown. Moreover, their poten- domains (3R).
tial incidence upon cognitive functions or other The alternative splicing of Tau is develop-
neurological symptoms also remained elusive. mentally regulated, tissue specific and deregu-
Neurofibrillary tangles (NFTs), a neuropatholog- lated in several Tauopathies [75]. A single
ical feature described by Alois Alzheimer [3] are isoform devoid of alternative of exon sequences
often reported in the limbic system of DM1 is express in the foetus whereas alternative splic-
patients, including the hippocampus and isocorti- ing occurs post-­natally. In DM1, we reported a
cal areas such as the temporal and frontal cortices modified expression pattern of Tau protein iso-
[44, 59, 74, 78, 85, 97]. Neurofibrillary tangles forms. Instead of the six adult brain isoforms
were also reported to contain heparan sulfates only two isoforms are expressed that are lacking
such as NFT of Alzheimer’s disease [78]. Another the inclusion of exons 2 and 3 [18, 75]. Tau exon
210 F. Fernandez-Gomez et al.

10 reduced inclusion has also been reported in aggregates found in Pick’s disease consist of
several brains from DM1 patients [40] as well as 3R-Tau isoforms mainly, whereas in supranuclear
in a transgenic animal model of DM1 [32]. palsy, corticobasal degeneration and argirophylic
However, molecular mechanism leading to the grain dementia, Tau aggregates are mainly com-
reduce Tau exon-10 inclusion in DM1 and its posed of 4R-Tau isoforms. It remains unknown
functional consequences remain unknown. This whether these specific pattern of pathological Tau
is of importance since the Tau exon-10 encodes proteins are associated with a change of Tau
for a fourth microtubule-­binding domain. The splicing or is it related to selective aggregation of
increased expression of 3R Tau isoforms in DM1 Tau protein isoforms expressed in sub-sets of
may modify the axonal transport and plasticity, neuronal populations. A differential splicing of
since Tau is suggested to regulate the motor pro- Tau transcript is also reported in Alzheimer’s dis-
teins transport along microtubules  [22]. ease [10, 96]. Although speculative, nucleic acid
Moreover, a growing body of evidences suggests integrity is regulated by Tau and more recently,
that Tau proteins may also localize to spines of the nucleotransport have been shown to be dis-
dendrites and regulate glutamatergic transmis- rupted by Tau proteins [23]. Nucleocytoplasmic
sion [37, 39]. The splicing of N-Methyl transport disruption is supposed to be a common
D-Aspartate receptor subunit one is altered in mechanism to several neurodegenerative diseases
DM1 brains [19, 40], which together with Tau including amyotrophic lateral sclerosis and fron-
mis splicing, may also deregulate the glutama- totemporal lobar degeneration [98]. Tau contribu-
tergic postsynaptic transmission. Thus, reduced tion to this defective nucleocytoplasmic transport
Tau protein isoforms expression in DM1 brains brings novel insights for understanding the
and the development of NFT may together con- pathophysiology of neurodegenerative diseases,
tribute the cognitive impairments in DM1. although the specific contribution of Tau iso-
Tau mRNA alternative splicing is tightly regu- forms remains to be determined.
lated during development. In the human brain, Several RNA splicing factors regulate alterna-
only 3R-Tau is expressed in the fetal stage, while tive splicing of Tau [5, 54] for review see [70]).
both 3R-Tau and 4R-Tau are expressed in a ratio They bind regulatory sequences (cis-elements)
of approximately 1:1 post-natally and in the adult within or around Tau exon 2 and 10 as well as
brain [82]. In DM1, the reversal of the adult Tau other specific regulatory proteins (trans-acting
splicing pattern to a fetal pattern of Tau splicing factors). Splicing factors regulating Tau belong
is supposed to occur with aging in consequence essentially to two major groups, the hnRNPs and
to the somatic instability of DM1 mutation. Large SR/SR-like proteins. The U1 sNRP has been
CTGs expansion are found in the adult brain tis- reported to be deregulated in Alzheimer’s dis-
sue which may lead to the progressive missplic- ease [7]. SRp75/hnRNPG complex regulate Tau
ing of Tau [74]. exon 10 skipping [87]. The loss of MBNL1 or
In frontotemporal dementia, MAPT gene gain of ETR3 function in DM1 [19, 52], gain of
mutations is associated with changes in Tau iso- DYRK1A [77], GSK3 [34], CLK2  [29] or
forms proportion, leading to the development of hnRNPE3 [87] in Down syndrome or AD is
NFTs in adulthood (for review see [75]). More linked to changes in the splicing of Tau exons 2,
than 35 mutations in MAPT gene are reported in 3 and 10.
patients suffering from rare forms of familial Whereas MBNL1 or MBNL1 both and inde-
frontotemporal dementia with parkinsonism-17 pendently regulate Tau exon inclusion, a syner-
(FTDP-17). Several of these MAPT gene muta- gistic effect of both MBNL1 and MBNL2 is
tions located downtstream exon 10 cause aber- necessary to regulated Tau exon 10 inclusion.
rant splicing of Tau and leading to a decrease This splicing regulatory mechanism may contrib-
ratio 3R: 4R-Tau [26]. In sporadic FTDP-17 and ute to explain why DM1 is more a muscle disease
other Tauopathies, imbalances in Tau isoform rather than brain disease since both the loss of
ratios are equally observed. For example, Tau MBNL1 and MBNL2 recapitulate the brain dys-
17  Myotonic Dystrophy: an RNA Toxic Gain of Function Tauopathy? 211

function observed in the disease [31]. Into the neuroaxonal degeneration. For tau, the situation is
complex landscape of Tau splicing regulatory promising but less clear. Plasma total-­tau concen-
mechanism, DM1 pathophysiology has enable tration in AD is increased but less so than in CSF
the discovery of master Tau splicing factors and there is no detectable increase in the Mild
which regulate Tau during development, post-­ Cognitive Impairment (MCI) stage of the disease
natally and in disease conditions. However, the [57, 58].
relationship between Tau missplicing, neurofi- The ability to measure Aβ in CSF or by using
brillary degeneration and cognitive deficits positron emission tomography (PET) imaging
remains a matter a debate but as in other has offered invaluable insights into the early
Tauopathies, Tau splicing can be directly (muta- stages of AD [80] and shown great potential as a
tions) or indirectly (splicing factors) deregulated diagnostic add-on to the clinical workup of
and hence ultimately lead to neurofibrillary patients with cognitive deficits  [69]. On the
degeneration. other hand, biomarker studies have shown that
the relationships between Aβ pathology and
most downstream processes such as glucose
Diagnostic and Therapeutic hypometabolism, brain atrophy, disease sever-
Perspective ity, progression, and clinical presentation are
modest. Parallel to this, worldwide interest in
Cognitive deficits suggest the involvement of the amyloid imaging has generated great momen-
cerebral cortex as well as subcortical areas, such tum to develop PET tracers that bind to non-Aβ
as the amygdala and the insula. In line with these processes such as the microtubule-­ associated
observations, several protein concentrations were protein tau aggregated as neurofibrillary tangles
altered in the cerebrospinal fluid of patient with in AD and other tau-­related diseases (e.g. FTLD
DM1 [36]. Moreover, biomarkers of Alzheimer or chronic traumatic encephalopathy [CTE]).
disease, including amyloid-beta peptide and Tau Indeed, tau pathology is known to have devas-
protein concentrations were significantly reduced tating effects on synaptic function [81], its tem-
and increased in the cerebrospinal spinal fluid, poral and spatial distribution correlates strongly
respectively, in DM1 when compared to aged-­ with the clinical evolution of AD [9], and post-
matched control patients [93]. Structural neuro- mortem tau aggregates are closely associated
imaging in DM1 patients reveals the presence of with cognitive performance during life. Over
brain atrophy and white matter lesions [20,  21, the past few years, several promising tau com-
43]. pounds have emerged, including 18F-AV-1451
Cerebrospinal fluid (CSF) biomarkers reflect- (or ‘flortaucipir’, previously known as T807,
ing the core pathological features of AD (total-­tau [15]). These tracers consistently demonstrated
reflecting degeneration, phospho-tau reflecting good in vivo brain penetration, tracer binding to
tau phosphorylation and tangle formation, and paired helical filaments of tau in AD brain tissue
Aβ42 which inversely correlates with plaque without labelling Aβ, deposits and patterns of
pathology in Alzheimer’s disease) have also been tracer retention on PET scans resembling tradi-
studied in DM1 (Concentration of Aβ species in tional Braak staging (see V. L. Villemagne et al.
CSF is therefore considered as a valuable periph- (2015), for a review [86]). These tracers would
eral biomarker of the central nervous system. A be valuable to investigate the natural history of
significant decrease in CSF levels of amyloid Tau pathology in DM1 or even DM2 with aging
Aß42 protein as well as increase in total tau pro- and its relationship if any with the cognitives
tein was reported in adult and juvenile DM1 [71, deficits.
93]. No correlation between the levels of these No curative treatment is available for DM
CSF biomarkers and neuropsychological impair- patients to date. The central hypothesis of DM1 is
ment were found, however. There are so far no the RNA gain of toxic function. Consequently
reliable blood biomarkers for NFT pathology and most of the therapeutic strategies aim at repress-
212 F. Fernandez-Gomez et al.

ing, delaying or even stopping this toxic effects Overexpression of MBNL1  in the HSA-LR
of CUGs or CCUGs large RNA repeats. Currently mouse model also reversed the DM-associated
under development are the antisense phenotypes [14, 41, 42] indicating that CUGexp-­
­oligonucleotides (ASO) or small compounds tar- RNA toxicity is mediated by the loss of func-
geting the large CUG repeats have which have tional MBNL in these DM1 mice. These results
shown promising beneficial effects in DM1 ani- suggest that MBNL overexpression could be a
mal models. Thus, two strategies are proposed. valid therapeutic option for DM disorders.
One is based on the degradation and the second However, the expression of MBNL1, 2 and 3 is
by blocking the consequences of CUGs expan- not ubiquitous but appears to be tissue-specific.
sion. Degradation of transcripts containing CUG Which MBNL protein, which isoform (since
expansions was achieved by using RNase there is at least 10 different isoforms for each
H-active ASO [90, 91] that recognize the RNA MBNL) and to what extent they have to be
target, hydrolyse the target by RNaseH1 and expressed to counteract CUGexp-RNA toxicity,
degrade the cleaved-RNA fragments by exonu- are questions to be answered. Novel tools or
clease activity. CUGS expansion degradation small molecules will be necessary to achieve a
was also achieved using siRNA oligonucleotides therapeutic efficacy.
targeting CUG repeats [50, 79]. Recently, a well-known anti-diabetic drug
The blockade of CUGs expansion toxic conse- Metformin, has been shown to correct some
quences was achieved either by the use of modi- splicing events in  vitro and in  vivo  [49] and to
fied ASO containing CAG-antisense sequences improve the muscle impairment the DMSXL
[51, 65, 90] or small compounds [27, 88] that mouse model of DM1. In patient this drug at the
were able to bind specifically to these structures. maximal tolerated dose provided positive effect
As a consequence, abnormal protein-­ CUG on the mobility and gait abilities of DM1 patients.
repeats interactions were disrupted and toxic Promising therapeutic strategies are under
consequences were modified. Both strategies clinical evaluation and also open avenues to
were assessed in transgenic mouse model of determine whether those therapeutic strategies
DM1 (expressing 220 CTG in the 3’UTR of the would be efficient against the brain dysfunction
Human Skeletal Actin gene; [56]) and showed and brain lesions.
significant to almost complete correction of
DM1-associated phenotypes such as splicing
defects and myotonia. References
Beyond these encouraging animal studies,
IONIS has recently completed a phase 1/2a placebo 1. Abe K, Fujimura H, Toyooka K, Yorifuji S, Nishikawa
Y, Hazama T, Yanagihara T. Involvement of the cen-
controlled study to assess safety and tolerability of tral nervous system in myotonic dystrophy. J Neurol
multiple doses of their lead compound in adults Sci. 1994;127(2):179–85.
with DM1 (clinicaltrial.gov NCT023412011). The 2. Adie WJ, Greenfield JG. Dystrophia myotonica (myo-
compound was delivered subcutaneously and was tonia atrophica). Brain. 1923;46:73–127.
3. Alzheimer A.  Uber eigenartige Krankheitsfalle
well tolerated at all doses tested. Despite some des spateren Alters. Zbl ges Neurol Psych. 1911;4:
modest changes in key biomarkers such as RNA 356–85.
splicing, the clinical benefits of the drug remain yet 4. Amack JD, Paguio AP, Mahadevan MS.  Cis and
to be proven. However recent advances in develop- trans effects of the myotonic dystrophy (DM)
mutation in a cell culture model. Hum Mol Genet.
ment of ASO generation 3.0, including LICA 1999;8(11):1975–84.
(Ligand Conjugated Antisense) or peptide-medi- 5. Andreadis A.  Tau gene alternative splicing: expres-
ated delivery technologies have shown tremendous sion patterns, regulation and modulation of function
improvement in delivery and potency in cell and in normal brain and neurodegenerative diseases.
Biochim Biophys Acta. 2005;1739:91–103.
animal models. How they translate in DM1 patient 6. Antonini G, Mainero C, Romano A, Giubilei F,
will be reported in the near future. Ceschin V, Gragnani F, Morino S, Fiorelli M, Soscia
F, Di Pasquale A, Caramia F. Cerebral atrophy in myo-
17  Myotonic Dystrophy: an RNA Toxic Gain of Function Tauopathy? 213

tonic dystrophy: a voxel based morphometric study. J Ishiura S, Buée L, Cooper TA, Caillet-Boudin ML,
Neurol Neurosurg Psychiatry. 2004;75(11):1611–3. Charlet-Berguerand N, Sablonnière B, Sergeant
7. Bai B, Hales CM, Chen PC, Gozal Y, Dammer EB, N.  Overexpression of MBNL1 fetal isoforms and
Fritz JJ, Wang X, Xia Q, Duong DM, Street C, modified splicing of Tau in the DM1 brain: two indi-
Cantero G, Cheng D, Jones DR, Wu Z, Li Y, Diner vidual consequences of CUG trinucleotide repeats.
I, Heilman CJ, Rees HD, Wu H, Lin L, Szulwach Exp Neurol. 2008;210(2):467–78.
KE, Gearing M, Mufson EJ, Bennett DA, Montine 19. Dhaenens CM, Tran H, Frandemiche ML, Carpentier
TJ, Seyfried NT, Wingo TS, Sun YE, Jin P, Hanfelt C, Schraen-Maschke S, Sistiaga A, Goicoechea M,
J, Willcock DM, Levey A, Lah JJ, Peng J. U1 small Eddarkaoui S, Van Brussels E, Obriot H, Labudeck
nuclear ribonucleoprotein complex and RNA splicing A, Gevaert MH, Fernandez-Gomez F, Charlet-
alterations in Alzheimer’s disease. Proc Natl Acad Sci Berguerand N, Deramecourt V, Maurage CA, Buée L,
U S A. 2013;110(41):16562–7. Lopez de Munain A, Sablonnière B, Caillet-Boudin
8. Brook JD, McCurrach ME, Harley HG, Buckler ML, Sergeant N. Mis-splicing of Tau exon 10 in myo-
AJ, Church D, Aburatani H, Hunter K, Stanton VP, tonic dystrophy type 1 is reproduced by overexpres-
Thirion JP, Hudson T, et al. Molecular basis of myo- sion of CELF2 but not by MBNL1 silencing. Biochim
tonic dystrophy: expansion of a trinucleotide (CTG) Biophys Acta. 2011;1812(7):732–42.
repeat at the 3′ end of a transcript encoding a protein 20. Di Costanzo A, Di Salle F, Santoro L, Bonavita V,
kinase family member. Cell. 1992;69(2):385. Tedeschi G.  Brain MRI features of congenital- and
9. Braak H, Del Tredici K.  Alzheimer’s pathogen- adult-form myotonic dystrophy type 1: case-control
esis: is there neuron-to-neuron propagation? Acta study. Neuromuscul Disord. 2002;12(5):476–83.
Neuropathol. 2011;121(5):589–95. 21. Di Costanzo A, Di Salle F, Santoro L, Tessitore A,
10. Caffrey TM, Joachim C, Paracchini S, Esiri MM, Bonavita V, Tedeschi G.  Pattern and significance of
Wade-Martins R.  Haplotype-specific expression of white matter abnormalities in myotonic dystrophy
exon 10 at the human MAPT locus. Hum Mol Genet. type 1: an MRI study. J Neurol. 2002;249(9):1175–82.
2006;15(24):3529–37. 22.
Dixit R, Ross JL, Goldman YE, Holzbaur
11. Caillet-Boudin ML, Fernandez-Gomez FJ, Tran H, EL. Differential regulation of dynein and kinesin motor
Dhaenens CM, Buee L, Sergeant N.  Brain pathol- proteins by tau. Science. 2008;319(5866):1086–9.
ogy in myotonic dystrophy: when tauopathy meets 23. Eftekharzadeh B, Daigle JG, Kapinos LE, Coyne A,
spliceopathy and RNAopathy. Front Mol Neurosci. Schiantarelli J, Carlomagno Y, Cook C, Miller SJ,
2014;6:57. Dujardin S, Amaral AS, Grima JC, Bennett RE, Tepper
12. Censori B, Provinciali L, Danni M, Chiaramoni
K, DeTure M, Vanderburg CR, Corjuc BT, DeVos SL,
L, Maricotti M, Foschi N, Del Pesce M, Salvolini Gonzalez JA, Chew J, Vidensky S, Gage FH, Mertens
U.  Brain involvement in myotonic dystrophy: MRI J, Troncoso J, Mandelkow E, Salvatella X, Lim RYH,
features and their relationship to clinical and cognitive Petrucelli L, Wegmann S, Rothstein JD, Hyman
conditions. Acta Neurol Scand. 1994;90(3):211–7. BT. Tau protein disrupts Nucleocytoplasmic transport
13. Chen G, Carter RE, Cleary JD, Reid TS, Ranum LP, in Alzheimer’s disease. Neuron. 2018;99(5):925–940.
Swanson MS, Ebner TJ. Altered levels of the splicing e7.
factor muscleblind modifies cerebral cortical function 24. Fardaei M, Larkin K, Brook JD, Hamshere MG.  In
in mouse models of myotonic dystrophy. Neurobiol vivo co-localisation of MBNL protein with DMPK
Dis. 2018;112:35–48. expanded-repeat transcripts. Nucleic Acids Res.
14. Chamberlain CM, Ranum LP.  Mouse model of
2001;29(13):2766–71.
muscleblind-like 1 overexpression: skeletal muscle 25. Gallais B, Gagnon C, Mathieu J. Richer L.Cognitive
effects and therapeutic promise. Hum Mol Genet. decline over time in adults with myotonic dystrophy
2012;21(21):4645–54. type 1: A 9-year longitudinal study. Neuromuscul
15. Chien DT, Bahri S, Szardenings AK, Walsh JC, Mu Disord. 2017;27(1):61–72.
F, Su MY, Shankle WR, Elizarov A, Kolb HC. Early 26. Gallo JM, Noble W, Martin TRRNA. protein-

clinical PET imaging results with the novel PHF- dependent mechanisms in tauopathies: conse-
tau radioligand [F-18]-T807. J Alzheimers Dis. quences for therapeutic strategies. Cell Mol Life Sci.
2013;34(2):457–68. 2007;64(13):1701–14. Review
16. Cleary JD, Pattamatta A, Ranum LPW. Repeat associ- 27. García-López A, Llamusí B, Orzáez M, Pérez-Payá
ated non-ATG (RAN) translation. J Biol Chem. 2018; E, Artero RD. In vivo discovery of a peptide that pre-
[Epub ahead of print] vents CUG-RNA hairpin formation and reverses RNA
17. Culebras A, Feldman RG, Merk FB.  Cytoplasmic
toxicity in myotonic dystrophy models. Proc Natl
inclusion bodies within neurons of the thalamus in Acad Sci U S A. 2011;108(29):11866–71.
myotonic dystrophy. A light and electron microscope 28. Gibson R. Dystrophia myotonica in association with
study. J Neurol Sci. 1973;19(3):319–29. mental defect. Can Med Assoc J. 1961;84:1234–7.
18.
Dhaenens CM, Schraen-Maschke S, Tran H, 29. Glatz DC, Rujescu D, Tang Y, Berendt FJ, Hartmann
Vingtdeux V, Ghanem D, Leroy O, Delplanque J, AM, Faltraco F, Rosenberg C, Hulette C, Jellinger
Vanbrussel E, Delacourte A, Vermersch P, Maurage K, Hampel H, Riederer P, Möller HJ, Andreadis A,
CA, Gruffat H, Sergeant A, Mahadevan MS, Henkel K, Stamm S.  The alternative splicing of tau
214 F. Fernandez-Gomez et al.

exon 10 and its regulatory proteins CLK2 and TRA2- 41. Kanadia RN, Johnstone KA, Mankodi A, Lungu C,
BETA1 changes in sporadic Alzheimer’s disease. J Thornton CA, Esson D, Timmers AM, Hauswirth WW,
Neurochem. 2006;96(3):635–44. Swanson MS. A muscleblind knockout model for myo-
30. Glantz RH, Wright RB, Huckman MS, Garron DC, tonic dystrophy. Science. 2003;302(5652):1978–80.
Siegel IM.  Central nervous system magnetic reso- 42. Kanadia RN, Shin J, Yuan Y, Beattie SG, Wheeler
nance imaging findings in myotonic dystrophy. Arch TM, Thornton CA, Swanson MS.  Reversal of
Neurol. 1988;45(1):36–7. RNA missplicing and myotonia after muscleblind
31. Goodwin M, Mohan A, Batra R, Lee KY, Charizanis overexpression in a mouse poly(CUG) model for
K, Fernández Gómez FJ, Eddarkaoui S, Sergeant N, myotonic dystrophy. Proc Natl Acad Sci U S A.
Buée L, Kimura T, Clark HB, Dalton J, Takamura K, 2006;103(31):11748–53.
Weyn-Vanhentenryck SM, Zhang C, Reid T, Ranum 43. Kornblum C, Reul J, Kress W, Grothe C, Amanatidis
LP, Day JW.  Swanson MS.  MBNL sequestration by N, Klockgether T, Schröder R. Cranial magnetic reso-
toxic RNAs and RNA misprocessing in the myotonic nance imaging in genetically proven myotonic dystro-
Dystrophy brain. Cell Rep. 2015;12(7):1159–68. phy type 1 and 2. J Neurol. 2004;251(6):710–4.
32. Gomes-Pereira M, Foiry L, Nicole A, Huguet A,
44. Kiuchi A, Otsuka N, Namba Y, Nakano I, Tomonaga
Junien C, Munnich A. Gourdon G.CTG trinucleotide M.  Presenile appearance of abundant Alzheimer’s
repeat “big jumps”: large expansions, small mice. neurofibrillary tangles without senile plaques in the
PLoS Genet. 2007;3(4):e52. Epub 2007 Feb 22 brain in myotonic dystrophy. Acta Neuropathol.
33. Harper PS.  Myotonic dystrophy. London: WB
1991;82:1–5.
Saunders; 2001. 45. Kuyumcu-Martinez NM, Cooper TA. Misregulation of
34. Hernández-Hernández O, Bermúdez-de-León M,
alternative splicing causes pathogenesis in myotonic
Gómez P, Velázquez-Bernardino P, García-Sierra dystrophy. Prog Mol Subcell Biol. 2006;44:133–59.
F, Cisneros B.  Myotonic dystrophy expanded CUG 46. Labayru G, Arenzana I, Aliri J, Zulaica M, López de
repeats disturb the expression and phosphorylation of Munain A, Sistiaga AA. Social cognition in myotonic
tau in PC12 cells. J Neurosci Res. 2006;84(4):841–51. dystrophy type 1: specific or secondary impairment?
35. Hernández-Hernández O, Guiraud-Dogan C, Sicot G, PLoS One. 2018;13(9):e0204227.
Huguet A, Luilier S, Steidl E, Saenger S, Marciniak 47. Ladd AN, Charlet N, Cooper TA. The CELF family
E, Obriot H, Chevarin C, Nicole A, Revillod L, of RNA binding proteins is implicated in cell-specific
Charizanis K, Lee KY, Suzuki Y, Kimura T, Matsuura and developmentally regulated alternative splicing.
T, Cisneros B, Swanson MS, Trovero F, Buisson B, Mol Cell Biol. 2001;21(4):1285–96.
Bizot JC, Hamon M, Humez S, Bassez G, Metzger F, 48.
Ladd AN, Nguyen NH, Malhotra K, Cooper
Buée L, Munnich A, Sergeant N, Gourdon G, Gomes- TA. CELF6, a member of the CELF family of RNA-
Pereira M.  Myotonic dystrophy CTG expansion binding proteins, regulates muscle-specific splicing
affects synaptic vesicle proteins, neurotransmission enhancer dependent alternative splicing. J Biol Chem.
and mouse behaviour. Brain. 2013;136(Pt 3):957–70. 2004;279(17):17756–64.
36. Hirase T, Araki S. Cerebrospinal fluid proteins in mus- 49. Laustriat D, Gide J, Barrault L, Chautard E, Benoit
cular dystrophy patients. Brain Dev. 1984;6(1):10–6. C, Auboeuf D, Boland A, Battail C, Artiguenave F,
37. Hoover BR, Reed MN, Su J, Penrod RD, Kotilinek Deleuze JF, Bénit P, Rustin P, Franc S, Charpentier
LA, Grant MK, Pitstick R, Carlson GA, Lanier G, Furling D, Bassez G, Nissan X, Martinat C,
LM, Yuan LL, Ashe KH, Liao D.  Tau mislocaliza- Peschanski M, Baghdoyan S.  In Vitro and In Vivo
tion to dendritic spines mediates synaptic dysfunc- Modulation of Alternative Splicing by the Biguanide
tion independently of neurodegeneration. Neuron. Metformin. Mol Ther Nucleic Acids. 2015;4:e262.
2010;68(6):1067–81. 50. Lee JE, Bennett CF, Cooper TA. RNase H-mediated
38. Huin V, Buée L, Behal H, Labreuche J, Sablonnière degradation of toxic RNA in myotonic dystrophy type
B, Dhaenens CM.  Alternative promoter usage gen- 1. Proc Natl Acad Sci U S A. 2012;109(11):4221–6.
erates novel shorter MAPT mRNA transcripts in 51. Leger AJ, Mosquea LM, Clayton NP, Wu IH, Weeden
Alzheimer’s disease and progressive supranuclear T, Nelson CA, Phillips L, Roberts E, Piepenhagen
palsy brains. Sci Rep. 2017;7(1):12589. PA, Cheng SH, Wentworth BM.  Systemic deliv-
39. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A, van ery of a Peptide-linked morpholino oligonucleotide
Eersel J, Wölfing H, Chieng BC, Christie MJ, Napier neutralizes mutant RNA toxicity in a mouse model
IA, Eckert A, Staufenbiel M, Hardeman E, Götz of myotonic dystrophy. Nucleic Acid Ther. 2013
J.  Dendritic function of tau mediates amyloid-beta Apr;23(2):109–17.
toxicity in Alzheimer’s disease mouse models. Cell. 52. Leroy O, Dhaenens CM, Schraen-Maschke S, Belarbi
2010;142(3):387–97. K, Delacourte A, Andreadis A, Sablonnière B, Buée
40. Jiang H, Mankodi A, Swanson MS, Moxley RT,
L, Sergeant N, Caillet-Boudin ML. ETR-3 represses
Thornton CA.  Myotonic dystrophy type 1 is asso- Tau exons 2/3 inclusion, a splicing event abnormally
ciated with nuclear foci of mutant RNA, seques- enhanced in myotonic dystrophy type I.  J Neurosci
tration of muscleblind proteins and deregulated Res. 2006;84(4):852–9.
alternative splicing in neurons. Hum Mol Genet. 53. Liquori CL, Ricker K, Moseley ML, Jacobsen JF,
2004;13(24):3079–88. Kress W, Naylor SL, Day JW, Ranum LP. Myotonic
17  Myotonic Dystrophy: an RNA Toxic Gain of Function Tauopathy? 215

dystrophy type 2 caused by a CCTG expansion in 67. Ono S.  A 53-year-old woman with muscular atro-
intron 1 of ZNF9. Science. 2001;293(5531):864–7. phy showing hypersomnia and respiratory failure.
54. Liu F, Gong CX. Tau exon 10 alternative splicing and Neuropathology. 2008;28(5):557–60.
tauopathies. Mol Neurodegener. 2008;3:8. 68. Orengo JP, Chambon P, Metzger D, Mosier DR, Snipes
55. Malloy P, Mishra SK, Adler SH. Neuropsychological GJ, Cooper TA.  Expanded CTG repeats within the
deficits in myotonic muscular dystrophy. J Neurol DMPK 3′ UTR causes severe skeletal muscle wasting
Neurosurg Psychiatry. 1990;53(11):1011–3. in an inducible mouse model for myotonic dystrophy.
56. Mankodi A, Logigian E, Callahan L, McClain
Proc Natl Acad Sci U S A. 2008;105(7):2646–51.
C, White R, Henderson D, Krym M, Thornton 69. Ossenkoppele R, Prins ND, Pijnenburg YA, Lemstra
CA. Myotonic dystrophy in transgenic mice AW, van der Flier WM, Adriaanse SF, Windhorst AD,
expressing an expanded CUG repeat. Science. Handels RL, Wolfs CA, Aalten P, Verhey FR, Verbeek
2000;289(5485):1769–73. MM, van Buchem MA, Hoekstra OS, Lammertsma
57. Mattsson N, Schöll M, Strandberg O, Smith R,
AA, Scheltens P, van Berckel BN. Impact of molecu-
Palmqvist S, Insel PS, Hägerström D, Ohlsson lar imaging on the diagnostic process in a memory
T, Zetterberg H, Jögi J, Blennow K, Hansson O. clinic. Alzheimers Dement. 2013;9(4):414–21.
18F-AV-1451 and CSF T-tau and P-tau as bio- 70. Park SA, Ahn SI, Gallo JM. Tau mis-splicing in the
markers in Alzheimer’s disease. EMBO Mol Med. pathogenesis of neurodegenerative disorders. BMB
2017;9(9):1212–23. Rep. 2016;49(8):405–13.
58. Mattsson N, Lönneborg A, Boccardi M, Blennow K, 71. Peric V, Jovanovic A, Sovtic S, Stolic R, Djikic D,
Hansson O.  Geneva task force for the roadmap of Otasevic P. Temporal changes in plasma brain natri-
Alzheimer’s biomarkers. Clinical validity of cerebro- uretic peptide levels during exercise stressechocar-
spinal fluid Aβ42, tau, and phospho-tau as biomarkers diography in patients with dilated cardiomyopathy.
for Alzheimer’s disease in the context of a structured Int Heart J. 2014;55(5):428–32.
5-phase development framework. Neurobiol Aging. 72. Sansone V, Gandossini S, Cotelli M, Calabria M,

2017;52:196–213. Review Zanetti O, Meola G.  Cognitive impairment in adult
59. Maurage CA, Udd B, Ruchoux MM, Vermersch P, myotonic dystrophies: a longitudinal study. Neurol
Kalimo H, Krahe R, Delacourte A, Sergeant N. Similar Sci. 2007;28(1):9–15.
brain tau pathology in DM2/PROMM and DM1/ 73. Sistiaga A, Urreta I, Jodar M, Cobo AM, Emparanza J,
Steinert disease. Neurology. 2005;65(10):1636–8. Otaegui D, Poza JJ, Merino JJ, Imaz H, Martí-Massó
60. Meola G, Sansone V.  Cerebral involvement in myo- JF, López de Munain A. Cognitive/personality pattern
tonic dystrophies. Muscle Nerve. 2007;36(3):294– and triplet expansion size in adult myotonic dystrophy
306. Review type 1 (DM1): CTG repeats, cognition and personal-
61.
Miller JW, Urbinati CR, Teng-Umnuay P, ity in DM1. Psychol Med. 2010;40(3):487–95.
Stenberg MG, Byrne BJ, Thornton CA, Swanson 74. Sergeant N, Sablonnière B, Schraen-Maschke S,

MS.  Recruitment of human muscleblind proteins to Ghestem A, Maurage CA, Wattez A, Vermersch
(CUG)(n) expansions associated with myotonic dys- P, Delacourte A.  Dysregulation of human brain
trophy. EMBO J. 2000;19(17):4439–48. microtubuleassociated tau mRNA maturation
62. Minnerop M, Gliem C, Kornblum C. Current progress in myotonic dystrophy type 1. Hum Mol Genet.
in CNS imaging of myotonic dystrophy. Front Neurol. 2001;10(19):2143–55.
2018;9:646. 75. Sergeant N, Bretteville A, Hamdane M, Caillet-Boudin
63. Minnerop M, Weber B, Schoene-Bake JC, Roeske ML, Grognet P, Bombois S, Blum D, Delacourte
S, Mirbach S, Anspach C, Schneider-Gold C, Betz A, Pasquier F, Vanmechelen E, Schraen-Maschke
RC, Helmstaedter C, Tittgemeyer M, Klockgether T, S, Buée L.  Biochemistry of Tau in Alzheimer’s dis-
Kornblum C. The brain in myotonic dystrophy 1 and ease and related neurological disorders. Expert
2: evidence for a predominant white matter disease. Rev Proteomics. 2008;5(2):207–24. https://doi.
Brain. 2011;134(Pt 12):3530–46. org/10.1586/14789450.5.2.207.Review.
64. Mizukami K, Sasaki M, Baba A, Suzuki T, Shiraishi 76. Seznec H, Lia-Baldini AS, Duros C, Fouquet C,

H. An autopsy case of myotonic dystrophy with men- Lacroix C, Hofmann-Radvanyi H, Junien C, Gourdon
tal disorders and various neuropathologic features. G.  Transgenic mice carrying large human genomic
Psychiatry Clin Neurosci. 1999;53(1):51–5. sequences with expanded CTG repeat mimic closely
65. Mulders SA, van den Broek WJ, Wheeler TM, Croes the DM CTG repeat intergenerational and somatic
HJ. van Kuik-Romeijn P, de Kimpe SJ, Furling D, instability. Hum Mol Genet. 2000;9(8):1185–94.
Platenburg GJ, Gourdon G, Thornton CA, Wieringa 77. Shi J, Zhang T, Zhou C, Chohan MO, Gu X, Wegiel
B, Wansink DG. Triplet-repeat oligonucleotide-medi- J, Zhou J, Hwang YW, Iqbal K, Grundke-Iqbal I,
ated reversal of RNA toxicity in myotonic dystrophy. Gong CX, Liu F. Increased dosage of Dyrk1A alters
Proc Natl Acad Sci U S A. 2009;106(33):13915–20. alternative splicing factor (ASF)-regulated alterna-
66. Nakamori M, Takahashi T, Yamazaki Y, Kurashige T, tive splicing of tau in Down syndrome. J Biol Chem.
Yamawaki T, Matsumoto M. Cyclin-dependent kinase 2008;283(42):28660–9.
5 immunoreactivity for granulovacuolar degeneration. 78.
Spillantini MG, Tolnay M, Love S, Goedert
Neuroreport. 2012;23(15):867–72. M.  Microtubule-associated protein tau, heparan sul-
216 F. Fernandez-Gomez et al.

phate and alpha-synuclein in several neurodegen- 89. Weber YG, Roebling R, Kassubek J, Hoffmann S,
erative diseases with dementia. Acta Neuropathol. Rosenbohm A, Wolf M, Steinbach P, Jurkat-Rott K,
1999;97(6):585–94. Walter H, Reske SN, Lehmann-Horn F, Mottaghy
79. Sobczak K, Wheeler TM, Wang W, Thornton
FM, Lerche H.  Comparative analysis of brain struc-
CA.  RNA interference targeting CUG repeats in ture, metabolism, and cognition in myotonic dystro-
a mouse model of myotonic dystrophy. Mol Ther. phy 1 and 2. Neurology. 2010;74(14):1108–17.
2013;21(2):380–7. 90. Wheeler TM, Sobczak K, Lueck JD, Osborne RJ,
80. Sperling RA, Aisen PS, Beckett LA, Bennett DA, Lin X, Dirksen RT, Thornton CA. Reversal of RNA
Craft S, Fagan AM, Iwatsubo T, Jack CR Jr, Kaye J, dominance by displacement of protein sequestered on
Montine TJ, Park DC, Reiman EM, Rowe CC, Siemers triplet repeat RNA. Science. 2009;325(5938):336–9.
E, Stern Y, Yaffe K, Carrillo MC, Thies B, Morrison- 91. Wheeler TM, Leger AJ, Pandey SK, MacLeod AR,
Bogorad M, Wagster MV, Phelps CH. Toward defin- Nakamori M, Cheng SH, Wentworth BM, Bennett
ing the preclinical stages of Alzheimer’s disease: CF, Thornton CA.  Targeting nuclear RNA for
recommendations from the National Institute on in  vivo correction of myotonic dystrophy. Nature.
Aging-Alzheimer’s Association workgroups on diag- 2012;488(7409):111–5. https://doi.org/10.1038/
nostic guidelines for Alzheimer’s disease. Alzheimers nature11362.
Dement. 2011;7(3):280–92. 92. Winblad S, Hellström P, Lindberg C, Hansen S. Facial
81. Spires-Jones TL, Hyman BT.  The intersection of
emotion recognition in myotonic dystrophy type
amyloid beta and tau at synapses in Alzheimer’s dis- 1 correlates with CTG repeat expansion. J Neurol
ease. Neuron. 2014;82(4):756–71. Neurosurg Psychiatry. 2006;77(2):219–23.
82. Takuma H, Arawaka S, Mori H. Isoforms changes of 93. Winblad MJE.  Blennow K, Jensen C, Samuelsson
tau protein during development in various species. L, Lindberg C.  Cerebrospinal fluid tau and amyloid
Brain Res Dev Brain Res. 2003;142(2):121–7. beta42 protein in patients with myotonic dystrophy
83. Thomas JD, Oliveira R, Sznajder ŁJ, Swanson
type 1. Eur J Neurol. 2008;15(9):947–52.
MS.  Myotonic Dystrophy and developmental 94. Wiśniewski HM, Berry K, Spiro AJ. Ultrastructure of
regulation of RNA processing. Compr Physiol. thalamic neuronal inclusions in myotonic dystrophy. J
2018;8(2):509–53. Neurol Sci. 1975;24(3):321–9.
84. Timchenko NA, Cai ZJ, Welm AL, Reddy S,
95. Yamazaki Y, Matsubara T, Takahashi T, Kurashige T,
Ashizawa T.  Timchenko LT.  RNA CUG repeats Dohi E, Hiji M, Nagano Y, Yamawaki T, Matsumoto
sequester CUGBP1 and alter protein levels and M. Granulovacuolar degenerations appear in relation
activity of CUGBP1. J Biol Chem. 2001;276(11): to hippocampal phosphorylated tau accumulation
7820–6. in various neurodegenerative disorders. PLoS One.
85. Vermersch P, Sergeant N, Ruchoux MM, Hofmann- 2011;6(11):e26996.
Radvanyi H, Wattez A, Petit H, Dwailly P, Delacourte 96. Yasojima K, McGeer EG, McGeer PL.  Tangled

A. Specific tau variants in the brains of patients with areas of Alzheimer brain have upregulated lev-
myotonic dystrophy. Neurology. 1996;47(3):711–7. els of exon 10 containing tau mRNA.  Brain Res.
86. Villemagne VL, Okamura N. Tau imaging in the study 1999;831(1–2):301–5.
of ageing, Alzheimer’s disease, and other neurodegen- 97. Yoshimura N.  Alzheimer’s neurofibrillary changes

erative conditions. Curr Opin Neurobiol. 2016;36:43– in the olfactory bulb in myotonic dystrophy. Clin
51. Review Neuropathol. 1990;9(5):240–3.
87. Wang Y, Gao L, Tse SW, Andreadis A. Heterogeneous 98. Zhang YJ, Gendron TF, Ebbert MTW, O’Raw AD,
nuclear ribonucleoprotein E3 modestly activates Yue M, Jansen-West K, Zhang X, Prudencio M, Chew
splicing of tau exon 10 via its proximal downstream J, Cook CN, Daughrity LM, Tong J, Song Y, Pickles
intron, a hotspot for frontotemporal dementia muta- SR, Castanedes-Casey M, Kurti A, Rademakers R,
tions. Gene. 2010;451(1–2):23–31. Oskarsson B, Dickson DW, Hu W, Gitler AD, Fryer
88. Warf MB, Nakamori M, Matthys CM, Thornton
JD, Petrucelli L. Poly(GR) impairs protein translation
CA, Berglund JA.  Pentamidine reverses the splicing and stress granule dynamics in C9orf72-associated
defects associated with myotonic dystrophy. Proc frontotemporal dementia and amyotrophic lateral
Natl Acad Sci U S A. 2009;106(44):18551–6. sclerosis. Nat Med. 2018;24(8):1136–42.
 Tau PET Imaging
18
Makoto Higuchi

Introduction is also expected to allow classification of FTLDs

into neuropathological subtypes at an early stage
Neurodegenerative dementias primarily consist and assessments of the disease severity in an
of Alzheimer’s disease (AD), dementia with objective fashion. In consideration of close asso-
Lewy bodies (DLB) and frontotemporal lobar ciations between tau depositions and neurode-
degeneration (FTLD), and are neuropathologi- generation, the  commencement of an anti-tau
cally characterized by depositions of aggregated treatment in an optimal time frame is supposed to
proteins in the brain, as exemplified by fibrillar be of critical significance for modifying the dis-
assemblies of tau, amyloid-β (Aβ), α-synuclein ease course, and would be enabled by tau imag-
and TDP-43 (Fig. 18.1). The major constituent of ing of individual subjects. The necessity of
protein deposits and their topology in the brain diagnostic and therapeutic approaches to prodro-
are specific for each illness, as exemplified by mal neurodegenerative dementias has been high-
accumulations of tau and Aβ in the neocortex and lighted in recent days, and it would be largely
limbic system of AD brains and a large subset of dependent on biological markers represented by
DLB brains. By contrast, neuropathological fea- imaging findings whether a disease-modifying
tures of FTLD are formations of tau, inclusions treatment such as an anti-tau therapeutic should
in the frontotemporal regions and subcortical be started antecedent to the disease onset.
structures including the basal ganglia and brain- Furthermore, it is virtually impossible to evaluate
stem without noticeable Aβ pathologies efficacies of a candidate drug in asymptomatic
(Fig. 18.1) [1]. In AD, tau fibrillogenesis is more individuals based on clinical manifestations, jus-
intimately linked to neuronal death than Aβ tifying the use of biomarkers as essential indices
aggregation, and is supposed to be initiated for therapeutic effects. In fact, US Food and Drug
decades before the clinical onset of the disease. Administration issued “Early Alzheimer’s
This raises the possibility that visualization of tau Disease: Developing Drugs for Treatment;
lesions in the living brains provides information Guidance for Industry” in February 2018, indi-
serving for diagnosis and staging of AD from an cating that a drug examined by a clinical trial in
asymptomatic period. Imaging of tau pathologies presymptomatic subjects would be provisionally
approved once its efficacy is proven with changes
M. Higuchi (*) in a biomarker measure that is capable of ratio-
National Institute of Radiological Sciences, National nally predicting progress to the clinical onset. In
Institutes for Quantum and Radiological Science and this context, the  effectiveness of this drug to
Technology, Chiba, Japan modify the clinical course of the disease would
e-mail: higuchi.makoto@qst.go.jp

© Springer Nature Singapore Pte Ltd. 2019 217

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_18
218 M. Higuchi

Tau Ab

TDP-43
FTLD

AD
Vascular
dementia

DLB

a-synuclein

Fig. 18.1  Major classification of dementing disorders these illnesses with several fluorescent β-sheet dyes,
and proteins deposited in neurodegenerative dementias. along with PET imaging of tau and Aβ deposits in AD
Histochemical staining of brain sections derived from cases are also displayed

be tested through a post-marketing clinical trial. tralized analog of thioflavin-­T was developed in
Imaging of tau lesions would offer such a bio- the  early 2000s for capturing Aβ plaques in the
logical index helping prediction of the clinical brain [2]. This compound was dubbed Pittsburgh
manifestations, but its utility in the trials as a reli- Compound-B (PiB), and PiB labeled with a radio-
able tool is greatly dependent on the selection of nuclide, 11C, was successfully applied to positron
an appropriate tau imaging agent coupled with a emission tomography (PET) of Aβ deposits in the
robust method for quantification of tau burdens in brains of AD patients [3]. Interestingly, PiB only
the brain. weakly binds to tau aggregates in contrast to its
strong reactivity with Aβ deposits [4], indicating
the potential selectivity of a β-sheet ligand for
 iscovery of a Selective Imaging
D either Aβ or tau assemblies. This notion raises the
Probe for Tau Pathologies possibility that structural modifications of such a
chemical lead to the  generation of an imaging
Pathogenic proteins such as Aβ, tau, α-synuclein, agent selectively capturing tau lesions in a manner
and TDP43 self-assemble by forming a β-pleated opposite to PiB. Amyloid dyes with higher selec-
sheet as a secondary structure and are deposited in tivity for tau versus Aβ pathologies are exempli-
the brain. It is known that thioflavin-­T, thioflavin- fied by thioflavin-S, and a major component of
S, Congo red, and several other fluorescent dyes thioflavin-S has a more extended backbone struc-
can detect amyloid fibrils by binding to β-sheets ture than PiB. In our attempt to discover a near-
contained in these molecules. Most of these infrared fluorescent tracer for amyloid deposits,
chemicals are electrically charged, which ham- we also found that a β-sheet ligand with a long
pers their entry into the brain through the blood- backbone produced a longer emission fluores-
brain barrier, and are therefore not qualified as cence wavelength and tended to show high affin-
in-vivo neuroimaging probes. Meanwhile, a neu- ity for tau rather than Aβ aggregates.
18  Tau PET Imaging 219

Chemical Core Core length AD Pick’s disease Model mouse

Name
structure M. W. (Å) Cmpd. pTau Cmpd. pTau Cmpld. pTau

216.2 10.9 PiB

208.2 11.1 BF-158

217.3 11.1 THK523

201.2 11.7 FDDNP

234.2 12.1 BF-227

240.2 13.2 DMSB

239.3 15.6 Curcumin

264.2 15.6 PBB1

264.2 16.0 PBB2

266.3 15.6 PBB3

264.2 15.6 PBB4

264.2 15.5 PBB5

264.2 16.6 FSB

346.4 17.3 Thioflavin-S

280.2 18.5 BF-189

356.5 20.5 DM-POTEB

Fig. 18.2  Relationships between the length of a core be assessed by fluorescence staining of a brain section.
(backbone) structure of a β-sheet ligand and its binding to Tau inclusions are doubly stained with the compounds
tau aggregates in AD, Pick’s disease, and tau transgenic (Cmpd.) and an antibody against phosphorylated tau
mice dubbed PS19. The core structure of a compound is (pTau), AT8. Compounds with a core length ranging from
highlighted in red. Ligands are listed in ascending order of 15 to 17 Å exhibit versatile reactivity with AD and non-
the core length. All these chemicals possess self-­ AD tau aggregates. (Modified from Ref. [5])
fluorescence, and their reactivity with tau aggregates can

We then revealed that a near-infrared laser compounds with a core structure approximating
dye, LDS750 (a. k. a. styryl 7), which possesses a 15 A in length detected a variety of tau patholo-
core structure seemingly formed by inserting a gies with high contrast. Since styryl 7 had a qua-
linker into the middle of a backbone of thiofla- ternary amine with an electric charge similar to
vin-­T or PiB, selectively bound to tau inclusions thioflavin-T, its permeability through BBB was
in diverse disorders such as AD and FTLD modest, impeding the use of this compound for
including Pick’s disease, progressive supranu- high-sensitivity in-vivo imaging of tau lesions.
clear palsy (PSP), corticobasal degeneration We accordingly remodeled the structure of styryl
(CBD) [5]. The backbone lengths of β-sheet 7, resulting in the creation of electrically neutral-
ligands were related to their reactivity with vari- ized chemicals termed PBB (Pyridinyl/Phenyl-
ous tau aggregates, as shown in Fig.  18.2, and Butadienyl-Benzothiazole/Benzothiazolium)
220 M. Higuchi

Fig. 18.3  Clinical PET imaging of tau and Aβ accumula- campal formation to the limbic system and subsequently to
tions with [11C]PBB3 and [11C]PiB, respectively. PET the neocortex. By contrast, a patient with suspected CBD is
images reflecting radioligand binding are superimposed on PiB-negative and displays tau accumulations in the basal
co-registered MRI data. Depositions of tau are noticeable in ganglia, subthalamic nucleus, and midbrain with left-right
the hippocampal formation of a cognitively normal individ- asymmetry. Volumetric analysis of MRI data with software
ual (second row). In the AD spectrum, Aβ pathology appears termed VSRAD indicates local atrophy in agreement with
to trigger the expansion of tau depositions from the hippo- tau lesions in this case. (Modified from Ref. [5])

(Fig. 18.2) [5]. Styryl 7 was also renamed PBB5 sions. Correspondingly, PET imaging of the tau
and was employed for additional assessments. transgenics with 11C-labeled PBB2, PBB3, and
PBBs were dubbed PBB1–5  in descending methoxy-PBB5 indicated the highest sensitivity
order of their lipophilicity, and all of these com- of tau detection obtained by [11C]PBB3 [5].
pounds tightly bound to tau deposits in brain sec- Supported by these nonclinical data, we initiated
tions derived from almost all tauopathies. We a clinical PET study with [11C]PBB3, resulting in
subsequently utilized PBBs as fluorescent probes the revelation that this radioligand was capable of
for intravital two-photon laser microscopic imag- visualizing tau pathologies in AD, PSP, CBD and
ing of a tau transgenic mouse expressing mutant several other tauopathies (Fig. 18.3) [5].
human tau to compare performances of these There have also been PET imaging agents for
compounds, and found that PBB3 yielded the tau lesions developed by several other industrial
highest in-vivo contrast for neuronal tau inclu- and academic research groups. A series of β-sheet
18  Tau PET Imaging 221

[11C]PBB3 [18F]THK5351

AD patient Elderly control AD patient Elderly control

(Maruyama et al. Neuron 2013) (Harada et al. J Nucl Med 2016)

[18F]T807

Elderly controls MCI patients AD patients

(Johnson et al. Ann Neurol 2016)

Fig. 18.4  Tau PET images of cognitively normal elderly controls and patients with AD spectrum pathologies with [11C]
PBB3, [18F]THK5351, and [18F]T807. (Modified from Refs. [5, 9, 17])

ligands created by BF Institute (occasionally roinflammatory gliosis in light of the expression

called BF compounds) were evaluated by scien- of MAO-B in astrocytes [12, 13]. Primarily for
tists at Tohoku University, and they demonstrated this reason, clinical trials for [18F]THK5351 con-
high-affinity binding of chemicals containing ducted by General Electric Company was discon-
quinoline as a part of its core structure to tau tinued, but there remains a possibility that
deposits. They accordingly promoted clinical additional structural modifications of BF com-
applications of these quinoline derivatives in col- pounds will lead to radiotracers with high selec-
laboration with University of Melbourne and tivity for either tau fibrils or astrocytes enriched
some other research sites, and proved that PET with MAO-B.
radioligands including [18F]THK523 [6], [18F] Other efforts to develop tau imaging agents
THK5105 [7], [18F]THK5117 [8] and [18F] are represented by the development of ligands for
THK5351 [9] were able to detect tau accumula- tau aggregates including, T807 and T808 by
tions in the brains of living AD patients. Among Siemens Medical Solution USA, Inc., on the
these BF compounds, [18F]THK5351 was structural basis of a BF compound, BF-126 [14–
reported to capture tau pathologies in AD [9], 16]. A notable feature of their discovery strategy
PSP [10], and CBD [11] with the highest contrast was that candidate chemicals were evaluated by
(Fig. 18.4). Meanwhile, the quinoline derivatives assessing their reactivity with native tau deposits
were also documented to have an  affinity for in AD brain tissues but not in-vitro assemblies of
monoamine oxidase B (MAO-B), and this may recombinant tau proteins. As most β-sheet ligands
hamper dissociation of tau depositions from neu- bind to Aβ, tau, and α-synuclein multimerized in
222 M. Higuchi

a test tube, the implementation of an assay sys- of Aβ plaques (Fig. 18.3); (2) Aβ aggregations in
tem with real brain samples may be more feasible AD spectrum likely trigger spreading of tau
for the screening of test compounds than the use depositions from the hippocampal formation to
of artificial models of the protein fibrillogenesis. the limbic system and subsequently to the neo-
These tau fibril ligands were then acquired by Eli cortex (Fig.  18.3). This expansion also occurs
Lilly and Company, and their performances have with slower advancement in the absence of Aβ
been further analyzed by Lilly’s subsidiary, Avid plaques, which is referred to as primary age-
Radiopharmaceuticals. Since metabolic and related tauopathy (PART) [23]. The extent of tau
pharmacokinetic properties of T807 were supe- accumulations intimately correlates with the
rior to those of T808, subsequent developments severity of MCI due to AD and AD dementia,
focused on 18F-labeled T807, [18F]T807, which bringing an objective index for the disease sever-
was also named [18F]AV-1451 in the development ity; and (3) Localizations of tau depositions agree
by Avid and flortaucipir F-18 in clinical trials by with those of atrophies in patients with MCI due
Lilly. [18F]T807 was extensively used in US, to AD and AD dementia, suggesting on-site neu-
Canadian and European research institutes rotoxicity provoked by tau aggregates.
(Fig. 18.4) [17], as exemplified by its application The close association of PET-detectable tau
to a multicenter imaging program in  the US deposits with neuronal deteriorations and symp-
termed AD Imaging Initiative (US-ADNI) for tomatic manifestations supports the view that tau
collection of tau PET data with a large sample PET imaging would enable objective and precise
size [18]. evaluations of candidate therapeutics modifying
AD processes from an MCI stage. In addition, the
presence of tau PET-positive individuals in
 tility of Tau PET Imaging
U the  non-demented elderly population indicates
Demonstrated by Clinical the possibility of classifying healthy aged sub-
Assessments jects into four categories according to Aβ and tau
PET findings (Table 18.1). A disease-modifying
PET imaging of plaque lesions revealed Aβ depo- therapy at a preclinical stage may be applied to
sitions in a significant subset of elderly popula- one or more of these categories, and tau PET
tion, providing a notion that cognitively normal might be employed for assessing biological out-
subjects and patients with mild cognitive impair- comes of the treatment in subjects lacking notice-
ment (MCI) and dementia exhibiting PET- able cognitive declines.
detectable Aβ pathologies are conceived to be To establish the utility of tau PET for thera-
afflicted with AD at different stages defined as peutic evaluations, time-course changes of tau
preclinical AD [19], MCI due to AD [20] and AD radioligand binding in the brains of the same
dementia [21], respectively, in a continuum individuals in transition from an early to an
called AD spectrum. The tau pathogenesis in advanced stage of the  AD spectrum need to be
aging and AD spectrum has been characterized clarified by a longitudinal clinical PET study. In a
by the following three features according to clini- cohort study conducted by a research group at
cal PET findings with [11C]PBB3, [18F]THK5351, Mayo Clinic, spatiotemporal alterations of tau
and [18F]T807 [5, 9, 17, 18, 22]: (1) Accumulations depositions were pursued in 126 subjects by PET
of tau aggregates in the hippocampal formation with [18F]T807. There was an increase of the tau
are detectable in a subpopulation of cognitively burden by 0.5% per year in non-demented elderly
intact elderly individuals in a manner indepen- people with PET-visible Aβ plaques, and this
dent of Aβ depositions. Aging-related tau pathol- increment was 3% per year in Aβ-PET-positive
ogies in the hippocampal formation have also cases with cognitive deficits [24]. This work also
been documented by neuropathological assays of documented that tau PET exhibited smaller vari-
autopsied brains, and often precede the formation abilities of annual changes among individuals
18  Tau PET Imaging 223

Table 18.1  Stratification of cognitively normal elderly individuals according to findings in tau and Abeta PET scans.
Tau-positive and Abeta-negative subjects may be conceived as having primary age-related tauopathy (PART)

Aβ- Aβ+

Preclinical
Tau- Normal aging
AD

Preclinical
Tau+ PART?
AD

than did psychometries, leading to a higher statis- cognitive functions, clinical trials for PSP rela-
tical power in assessing therapeutic efficacies of tive to AD may yield high precisions for deter-
a test drug. mining efficacies of an anti-tau drug candidate.
For proving the usefulness of a tau PET ligand These notions also rationalize the application
as a reliable diagnostic agent, the specificity of of a PET radioligand for PSP and CBD tau depos-
PET radiosignals for tau aggregates should be its to clinical settings, but it has been pointed out
demonstrated by analyzing correlations between that the reactivity of [18F]T807 with these aggre-
neuroimaging and neuropathological data in the gates is insufficient for in-vivo imaging
same subjects. Such assays on imaging-­ (Fig. 18.5) [26–28]. It is well known that tau pro-
neuropathology relationships with a small sam- teins in CNS are comprised of six isoforms,
ple size were reported with [18F]T807 and [18F] which are classified into four-repeat and three-­
THK5351 in patients with AD spectrum patholo- repeat isoforms according to the number of repeat
gies [25]. We have also carried out neuropatho- domains. While AD tau tangles are constituted of
logical examinations of the autopsied brains all six isoforms, PSP and CBD tau aggregates
derived from PSP and AD patients who had contain four-repeat isoforms only, in contrast to
undergone a PET scan with [11C]PBB3 at multi- Pick bodies in Pick’s disease formed by three-
ple imaging sites in Japan, and links between in- repeat tau isoforms. The difference in the isoform
vivo radioligand binding and composition may result in diversity in 3-D struc-
immunohistochemical and biochemical measures tures of tau aggregates, causing selectivity of a
of tau deposits are being investigated. PET ligand for tau assemblies in AD versus PSP/
FTLD tauopathies, including PSP and CBD, CBD or Pick’s disease as represented by [18F]
are characterized by accumulations of tau fibrils T807, which preferentially binds to AD-type tau
in the brain without prominent amyloid plaque fibrils.
formations, and an anti-tau treatment could be A neuropathological study of PSP patients
evaluated in these disorders without confounding following a [18F]T807-PET scan of these cases
effects of Aβ pathologies. Unlike AD, clinical also indicated a dissociation between topologies
outcome measures in a therapeutic trial for PSP of in-vivo PET signals and tau depositions [29].
patients are primarily composed of motor indices Although [18F]T807 binding in PET scans of
rather than cognitive scores. As motor perfor- CBD patients was found to be correlated with
mances could be assessed with smaller variabili- neuronal and glial tau inclusions assessed post-
ties, placebo effects, and individual biases than mortemly in the same individuals [28], in-vivo
224 M. Higuchi

Fig. 18.5  Autoradiograms of brain sections derived from regional distributions of [11C]PBB3 and [18F]T807 radio-
patients with AD, CBD, and frontotemporal dementia and signals differ from each other. Binding of [11C]PBB3  in
parkinsonism linked to chromosome 17 (FTDP-17) due to CBD and N279K mutant FTDP-17 brain tissues enriched
the N279K MAPT mutation. Adjacent slices are labeled with four-repeat tau aggregates is greater than that of [18F]
with [11C]PBB3 and [18F]T807 in a comparative manner. T807. (Modified from Ref. [26])
Both radioligands tightly bind to AD pathologies, but sub-

contrast for CBD tau pathologies produced by brains of PSP and CBD patients [5, 30], along
this radioligand was modest, often impeding dif- with symptomatic carriers of MAPT mutations
ferentiation between CBD and control subjects. causative of familial three-­repeat and four-repeat
[18F]THK5351 was demonstrated to accumulate tauopathies [31]. However, a signal-to-noise ratio
in the brains of PSP and CBD patients with its for capturing these non-AD tau lesions with [11C]
spatial distributions being characteristic of each PBB3 was not sufficient for determining tau posi-
disease [10, 11], but it remains likely that these tivity or negativity for each individual and for
radiosignals were mainly attributed to the bind- accurate quantification of longitudinal changes in
ing of this radioligand to MAO-B expressed in the tau load, necessitating a new tau PET ligand
activated astrocytes. allowing sensitive imaging assays.
PBB compounds were originally designed as
ligands for diverse tau aggregates composed of
all six isoforms or four- or three-repeat isoforms Development of Next-Generation
only, on the basis of structure-activity relation- Tau PET Probes
ships [5], and high-affinity binding of these
chemicals to PSP and CBD tau deposits was PET imaging with [11C]PBB3, [18F]THK5351,
shown by our in-vitro experiments (Fig.  18.5) and [18F]T807 has provided numerous insights
[26]. Moreover, clinical PET examinations have into the tau pathogenesis, while properties of
revealed elevated retention of [11C]PBB3  in the these compounds were yet to be improved.
18  Tau PET Imaging 225

Since [11C]PBB3 is labeled with 11C, which is a have been engaged in the development and evalu-
radionuclide with a half-life as short as 20 min, ation of imaging agents in the next generation
possible improvements of this ligand should with more preferred characteristics. Apart from
incorporate the use of 18F with a half-­life of [18F]PM-PBB3, most of these improved tracers
110  min instead of 11C for wider availabilities were designed based on the chemical structure of
and applicability to deliveries from a production T807. While THK5351 cross-reacts with MAO-
site to PET facilities. An additional issue on B, it has been reported that T807 exhibits off-
[11C]PBB3 is a prompt metabolic conversion of target binding to MAO-A. Accordingly, analogs
this ligand to a major metabolite, hampering of T807 with no or minimal binding to MAO-A
the efficient transfer of the unmetabolized com- and MAO-B have been selected for application
pound to the brain [5, 32]. Since a structural to clinical PET imaging. As summarized in
domain of [11C]PBB3 with a high propensity to Table  18.2, all next-generation tau imaging
biometabolism was identified [33], we have agents capture AD tau lesions without reacting
recently generated a new PBB3 analog, [18F] with MAO enzymes. Among these chemicals,
PM-PBB3, by substitution of this moiety with [18F]PM-PBB3 is hitherto the only radioligand
an 18F-containing side chain. Our exploratory capable of visualizing four-repeat tau pathologies
clinical work has shown that [18F]PM-PBB3 is in PSP and CBD with high contrast. [18F]PI-2620,
more resistant to metabolic changes than [11C] a compound developed by Piramal Imaging, was
PBB3, and that peak uptake of [18F]PM-PBB3 in also reported to sensitively detect PSP tau lesions,
the brain is approximately 1.5-fold higher than [34] but more recent investigations have indi-
that of [11C]PBB3 [34]. Owing to its high reac- cated that increased binding of this ligand in the
tivity with tau aggregates in the living brain and brainstem and basal ganglia is mostly derived
low background retentions, [18F]PM-PBB3 has from its reaction with non-tau components. It
been found to yield a more than two-fold higher should also be noted that a subset of tau radioli-
contrast for tau lesions in tau transgenic mice gands, including [11C]PBB3, [18F]T807, and [18F]
than [11C]PBB3 [34]. PM-PBB3, accumulate in choroid plexus to a
Following these non-clinical assessments, variable extent [5, 34, 40]. Exact mechanisms of
clinical evaluations of [18F]PM-PBB3 were com- this seemingly nonspecific radiosignals remain to
menced at the National Institute of Radiological be clarified but could be explained by interac-
Sciences as a principal investigator-­initiated PET tions of these chemicals with irons and pigments
study. In parallel with this program, a license on in choroid plexus [41]. It is also likely that the
a patent covering PBB compounds was granted compounds bind to Biondi bodies, which are
to a bioventure, APRINOIA Therapeutics, and amyloid-like protein aggregates deposited in
this company initiated a phase 0 clinical trial for choroid plexus with aging [42].
[18F]PM-PBB3 in USA [34]. These human assays Despite the superiority of the next-generation
have proven that contrasts for AD, PSP, and CBD tau PET probes to the first-generation compounds
tau deposits produced by [18F]PM-PBB3 are due to their low nonspecific binding, it is still uncer-
more than two-­fold of contrasts provided by [11C] tain whether these new ligands can indeed offer reli-
PBB3. It is of particular significance that four- able biological indices in the diagnosis and
repeat tau inclusions in PSP and CBD can be sen- therapeutic evaluations of AD spectrum patients in
sitively captured by [18F]PM-PBB3, allowing comparison with [18F]T807. To address this issue,
robust quantification of these pathologies at an longitudinal alterations of PET signals and their
individual level. correlations with postmortem observations would
[11C]PBB3, [18F]THK5351, and [18F]T807 are need to be examined with standardized protocols in
often referred to as “first-generation” tau radioli- a collaborative research program involving multiple
gands, and several pharmaceutical companies imaging and neuropathology sites.
 226

Table 18.2  Clinically available tau PET imaging agents. Next-generation radioligands are highlighted in bold
Non-AD tauopathy Off-target binding
PET tracer Developer Contrast for AD tau lesions Detectable types PSP tau contrast Tg mouse inclusions MAOs CP
[11C]PBB3 [5, 22] QST 1.2–1.4 4-repeat tauopathy 1.09–1.18 Yes − ±
3-repeat tauopathy
[18F]T807 [15, 17] Avid/Lilly 1.3–1.6 4-repeat tauopathy 1.03–1.16 MAO-A +
[18F]THK5351 [13] Tohoku Univ. 1.4–1.8 MAO-B −
[18F]PI-2620 [35] Piramal 2.0–2.5 − ±
[18F]GTP1 [36] Genentech 2.0–3.0 − −
[18F]RO6958948 [37, 38] Roche 2.0–2.5 − −
[18F] MK-6240 [39] Merck 2.0–3.0 − −
[18F]PM-PBB3 [34] QST APRINOIA 2.0–2.5 4-repeat tauopathy 1.25–2.25 Yes − +
3-repeat tauopathy
Contrasts for AD tau lesions are estimated by calculating ratios of standardized uptake value ratio (SUVR), distribution volume ratio (DVR), or non-displaceable binding potential
(BPND) plus 1.0 in neocortical and limbic areas between patients with MCI due to AD or early AD dementia and aged normal controls. Similarly, contrasts for PSP are determined by
calculating ratios of SUVR, DVR and BPND plus 1.0 in the brainstem, subthalamic nucleus, and globus pallidus between patients with PSP and age-matched controls. PBB3 and
PM-PBB3 are the only probes with specific binding to tau aggregates in several tau transgenic (Tg) mouse strains [5]. Some compounds exhibit off-target binding to MAO enzymes
and choroid plexus (CP)
M. Higuchi
18  Tau PET Imaging 227

Utilization for development 90-100% of degenerative dementias 70% of degenerative dementias

of blood biomarkers diagnosed by tau, TDP-43 &α-syn PET diagnosed by tau PET

Functional phenotypes assessed by fMRI

Mass screening by blood markers Workup by PET imaging Disease-modifying therapy

Ruling out 80% of subjects Construction of next-generation diagnostic & therapeutic workflows

Requiring new technical Clinically applicable in 5 - 10 Currently available in clinical

breakthroughs years settings

Fig. 18.6  Diagnostic and therapeutic workflows targeting neurodegenerative dementias utilizing proteinopathy PET
imaging as a pivotal element

Conclusions References

PET imaging of tau pathologies in the AD spec- 1. Kovacs GG. Molecular pathological classification of
neurodegenerative diseases: turning towards precision
trum and FTLD will cover approximately 70% of
medicine. Int J Mol Sci. 2016;17(2). pii: E189. https://
neurodegenerative dementias, and this coverage doi.org/10.3390/ijms17020189.
will reach 90–100% with imaging technologies 2. Mathis CA, Wang Y, Holt DP, Huang GF, Debnath ML,
to visualize α-synuclein and TDP-43 deposits. Klunk WE.  Synthesis and evaluation of 11C-labeled
6-substituted 2-arylbenzothiazoles as amyloid imag-
However, multiple PET scans for each elderly
ing agents. J Med Chem. 2003;46(13):2740–54.
subject will not be cost-effective and somewhat 3. Klunk WE, Engler H, Nordberg A, Wang Y, Blomqvist
time-consuming, and mass screening for demen- G, Holt DP, Bergström M, Savitcheva I, Huang GF,
tias would need to be conducted with a handier Estrada S, Ausén B, Debnath ML, Barletta J, Price
JC, Sandell J, Lopresti BJ, Wall A, Koivisto P, Antoni
method such as a blood test. Measurements of tau
G, Mathis CA, Långström B. Imaging brain amyloid
proteins in plasma could be performed using in Alzheimer’s disease with Pittsburgh compound-
ultra-sensitive ELISA and related innovations B. Ann Neurol. 2004;55(3):306–19.
[43], and validations of these liquid biopsies 4. Klunk WE, Wang Y, Huang GF, Debnath ML, Holt
DP, Shao L, Hamilton RL, Ikonomovic MD, DeKosky
would be facilitated by the use of tau and Aβ PET
ST, Mathis CA.  The binding of 2-(4′-methylamino-
scans as references. The major purpose of the phenyl) benzothiazole to postmortem brain homog-
mass screening with a blood biomarker would be enates is dominated by the amyloid component.
to exclude 80–90% of the subjects from further J Neurosci. 2003;23(6):2086–92.
5. Maruyama M, Shimada H, Suhara T, Shinotoh H, Ji
diagnostic workups, including PET examina-
B, Maeda J, Zhang MR, Trojanowski JQ, Lee VM,
tions. In the secondary tests for those who are Ono M, Masamoto K, Takano H, Sahara N, Iwata N,
positive in the initial screening, information on Okamura N, Furumoto S, Kudo Y, Chang Q, Saido
pathological protein depositions obtained by PET TC, Takashima A, Lewis J, Jang MK, Aoki I, Ito H,
Higuchi M. Imaging of tau pathology in a tauopathy
would be used for  the selection of appropriate
mouse model and in Alzheimer patients compared
disease-modifying therapies. The effects of such to normal controls. Neuron. 2013;79(6):1094–108.
treatments could also be assessed by PET scans, https://doi.org/10.1016/j.neuron.2013.07.037.
and functional MRI or similar non-radiological 6. Fodero-Tavoletti MT, Okamura N, Furumoto S,
Mulligan RS, Connor AR, McLean CA, Cao D,
imaging techniques might provide surrogate bio-
Rigopoulos A, Cartwright GA, O'Keefe G, Gong S,
logical indices reflecting neuronal integrities. In Adlard PA, Barnham KJ, Rowe CC, Masters CL, Kudo
an evolutional diagnostic and therapeutic system Y, Cappai R, Yanai K, Villemagne VL. 18F-THK523:
for dementing illnesses (Fig.  18.6), proteinopa- a novel in  vivo tau imaging ligand for Alzheimer’s
disease. Brain. 2011;134(Pt 4):1089–100. https://doi.
thy imaging represented by tau PET would play
org/10.1093/brain/awr038.
pivotal roles as an in-vivo test most reliably prob- 7. Okamura N, Furumoto S, Fodero-Tavoletti MT,
ing pathogenetic processes. Mulligan RS, Harada R, Yates P, Pejoska S, Kudo Y,
228 M. Higuchi

Masters CL, Yanai K, Rowe CC, Villemagne VL. Non- Dis. 2012;31(3):601–12. https://doi.org/10.3233/
invasive assessment of Alzheimer’s disease neurofi- JAD-2012-120712.
brillary pathology using 18F-THK5105 PET.  Brain. 15. Chien DT, Bahri S, Szardenings AK, Walsh JC, Mu
2014;137(Pt 6):1762–71. https://doi.org/10.1093/ F, Su MY, Shankle WR, Elizarov A, Kolb HC. Early
brain/awu064. clinical PET imaging results with the novel PHF-
8. Chiotis K, Saint-Aubert L, Savitcheva I, Jelic V, tau radioligand [F-18]-T807. J  Alzheimers Dis.
Andersen P, Jonasson M, Eriksson J, Lubberink M, 2013;34(2):457–68. https://doi.org/10.3233/
Almkvist O, Wall A, Antoni G, Nordberg A. Imaging JAD-122059.
in-vivo tau pathology in Alzheimer’s disease with 16. Chien DT, Szardenings AK, Bahri S, Walsh JC, Mu F,
THK5317 PET in a multimodal paradigm. Eur J Nucl Xia C, Shankle WR, Lerner AJ, Su MY, Elizarov A,
Med Mol Imaging. 2016;43(9):1686–99. https://doi. Kolb HC. Early clinical PET imaging results with the
org/10.1007/s00259-016-3363-z. novel PHF-tau radioligand [F18]-T808. J Alzheimers
9. Harada R, Okamura N, Furumoto S, Furukawa K, Dis. 2014;38(1):171–84. https://doi.org/10.3233/
Ishiki A, Tomita N, Tago T, Hiraoka K, Watanuki S, JAD-130098.
Shidahara M, Miyake M, Ishikawa Y, Matsuda R, 17. Johnson KA, Schultz A, Betensky RA, Becker

Inami A, Yoshikawa T, Funaki Y, Iwata R, Tashiro M, JA, Sepulcre J, Rentz D, Mormino E, Chhatwal J,
Yanai K, Arai H, Kudo Y. 18F-THK5351: a novel PET Amariglio R, Papp K, Marshall G, Albers M, Mauro
radiotracer for imaging neurofibrillary pathology in S, Pepin L, Alverio J, Judge K, Philiossaint M, Shoup
Alzheimer disease. J Nucl Med. 2016;57(2):208–14. T, Yokell D, Dickerson B, Gomez-Isla T, Hyman B,
https://doi.org/10.2967/jnumed.115.164848. Vasdev N, Sperling R.  Tau positron emission tomo-
10. Brendel M, Schönecker S, Höglinger G, Lindner S, graphic imaging in aging and early Alzheimer dis-
Havla J, Blautzik J, Sauerbeck J, Rohrer G, Zach ease. Ann Neurol. 2016;79(1):110–9. https://doi.
C, Vettermann F, Lang AE, Golbe L, Nübling G, org/10.1002/ana.24546.. Epub 2015 Dec 15
Bartenstein P, Furukawa K, Ishiki A, Bötzel K, Danek 18. Tosun D, Landau S, Aisen PS, Petersen RC, Mintun M,
A, Okamura N, Levin J, Rominger A. [18F]-THK5351 Jagust W, Weiner MW, Initiative A’s DN. Association
PET correlates with topology and symptom sever- between tau deposition and antecedent amyloid-β
ity in progressive supranuclear palsy. Front Aging accumulation rates in normal and early symptomatic
Neurosci. 2018;9:440. https://doi.org/10.3389/ individuals. Brain. 2017;140(5):1499–512.
fnagi.2017.00440. 19. Sperling RA, Aisen PS, Beckett LA, Bennett DA,
11. Kikuchi A, Okamura N, Hasegawa T, Harada R,
Craft S, Fagan AM, Iwatsubo T, Jack CR Jr, Kaye J,
Watanuki S, Funaki Y, Hiraoka K, Baba T, Sugeno Montine TJ, Park DC, Reiman EM, Rowe CC, Siemers
N, Oshima R, Yoshida S, Kobayashi J, Ezura M, E, Stern Y, Yaffe K, Carrillo MC, Thies B, Morrison-
Kobayashi M, Tano O, Mugikura S, Iwata R, Ishiki A, Bogorad M, Wagster MV, Phelps CH. Toward defin-
Furukawa K, Arai H, Furumoto S, Tashiro M, Yanai ing the preclinical stages of Alzheimer’s disease:
K, Kudo Y, Takeda A, Aoki M. In vivo visualization recommendations from the National Institute on
of tau deposits in corticobasal syndrome by 18F- Aging-Alzheimer’s Association workgroups on diag-
THK5351 PET. Neurology. 2016;87(22):2309–16. nostic guidelines for Alzheimer’s disease. Alzheimers
12. Ng KP, Pascoal TA, Mathotaarachchi S, Therriault Dement. 2011;7(3):280–92. https://doi.org/10.1016/j.
J, Kang MS, Shin M, Guiot MC, Guo Q, Harada jalz.2011.03.003.
R, Comley RA, Massarweh G, Soucy JP, Okamura 20. Albert MS, DeKosky ST, Dickson D, Dubois B,

N, Gauthier S, Rosa-Neto P.  Monoamine oxidase B Feldman HH, Fox NC, Gamst A, Holtzman DM,
inhibitor, selegiline, reduces 18F-THK5351 uptake in Jagust WJ, Petersen RC, Snyder PJ, Carrillo MC,
the human brain. Alzheimers Res Ther. 2017;9(1):25. Thies B, Phelps CH.  The diagnosis of mild cog-
https://doi.org/10.1186/s13195-017-0253-y. nitive impairment due to Alzheimer’s disease:
13. Harada R, Ishiki A, Kai H, Sato N, Furukawa K, recommendations from the National Institute on
Furumoto S, Tago T, Tomita N, Watanuki S, Hiraoka Aging-Alzheimer’s Association workgroups on diag-
K, Ishikawa Y, Funaki Y, Nakamura T, Yoshikawa T, nostic guidelines for Alzheimer’s disease. Alzheimers
Iwata R, Tashiro M, Sasano H, Kitamoto T, Yanai Dement. 2011;7(3):270–9. https://doi.org/10.1016/j.
K, Arai H, Kudo Y, Okamura N.  Correlations of jalz.2011.03.008.
18F-THK5351 PET with postmortem burden of 21. McKhann GM, Knopman DS, Chertkow H, Hyman
tau and Astrogliosis in Alzheimer disease. J  Nucl BT, Jack CR Jr, Kawas CH, Klunk WE, Koroshetz WJ,
Med. 2018;59(4):671–4. https://doi.org/10.2967/ Manly JJ, Mayeux R, Mohs RC, Morris JC, Rossor
jnumed.117.197426. MN, Scheltens P, Carrillo MC, Thies B, Weintraub
14.
Zhang W, Arteaga J, Cashion DK, Chen G, S, Phelps CH.  The diagnosis of dementia due to
Gangadharmath U, Gomez LF, Kasi D, Lam C, Alzheimer’s disease: recommendations from the
Liang Q, Liu C, Mocharla VP, Mu F, Sinha A, National Institute on Aging-Alzheimer’s Association
Szardenings AK, Wang E, Walsh JC, Xia C, Yu C, workgroups on diagnostic guidelines for Alzheimer’s
Zhao T, Kolb HC. A highly selective and specific PET disease. Alzheimers Dement. 2011;7(3):263–9.
tracer for imaging of tau pathologies. J  Alzheimers https://doi.org/10.1016/j.jalz.2011.03.005.
18  Tau PET Imaging 229

22. Shimada H, Kitamura S, Shinotoh H, Endo H, Niwa 18

F-flortaucipir tau positron emission tomography dis-
F, Hirano S, Kimura Y, Zhang MR, Kuwabara S, tinguishes established progressive supranuclear palsy
Suhara T, Higuchi M.  Association between Aβ and from controls and Parkinson disease: a multicenter
tau accumulations and their influence on clinical study. Ann Neurol. 2017;82(4):622–34. https://doi.
features in aging and Alzheimer’s disease spectrum org/10.1002/ana.25060.
brains: a [11C]PBB3-PET study. Alzheimers Dement 29. Marquié M, Normandin MD, Meltzer AC, Siao Tick
(Amst). 2016;6:11–20. https://doi.org/10.1016/j. Chong M, Andrea NV, Antón-Fernández A, Klunk
dadm.2016.12.009. WE, Mathis CA, Ikonomovic MD, Debnath M, Bien
23. Crary JF, Trojanowski JQ, Schneider JA, Abisambra EA, Vanderburg CR, Costantino I, Makaretz S, DeVos
JF, Abner EL, Alafuzoff I, Arnold SE, Attems J, Beach SL, Oakley DH, Gomperts SN, Growdon JH, Domoto-
TG, Bigio EH, Cairns NJ, Dickson DW, Gearing M, Reilly K, Lucente D, Dickerson BC, Frosch MP,
Grinberg LT, Hof PR, Hyman BT, Jellinger K, Jicha Hyman BT, Johnson KA, Gómez-Isla T. Pathological
GA, Kovacs GG, Knopman DS, Kofler J, Kukull correlations of [F-18]-AV-1451 imaging in non-
WA, Mackenzie IR, Masliah E, McKee A, Montine alzheimer tauopathies. Ann Neurol. 2017;81(1):117–
TJ, Murray ME, Neltner JH, Santa-Maria I, Seeley 28. https://doi.org/10.1002/ana.24844.
WW, Serrano-Pozo A, Shelanski ML, Stein T, Takao 30. Endo H, Shimada H, Sahara N, Ono M, Koga S,
M, Thal DR, Toledo JB, Troncoso JC, Vonsattel JP, Kitamura S, Niwa F, Hirano S, Kimura Y, Ichise M,
White CL 3rd, Wisniewski T, Woltjer RL, Yamada M, Shinotoh H, Zhang MR, Kuwabara S, Dickson DW,
Nelson PT.  Primary age-related tauopathy (PART): Toda T, Suhara T, Higuchi M. In-vivo binding of a tau
a common pathology associated with human aging. imaging probe, [11C]PBB3, in patients with progres-
Acta Neuropathol. 2014;128(6):755–66. https://doi. sive supranuclear palsy. Mov Diord. in press
org/10.1007/s00401-014-1349-0. 31. Ikeda A, Shimada H, Nishioka K, Takanashi M,

24. Jack CR Jr, Wiste HJ, Schwarz CG, Lowe VJ, Senjem Hayashida A, Li Y, Yoshino H, Funayama M, Ueno
ML, Vemuri P, Weigand SD, Therneau TM, Knopman Y, Hatano T, Sahara N, Suhara T, Higuchi M, Hattori
DS, Gunter JL, Jones DT, Graff-Radford J, Kantarci N. Clinical heterogeneity of frontotemporal dementia
K, Roberts RO, Mielke MM, Machulda MM, Petersen and parkinsonism linked to chromosome 17 caused
RC. Longitudinal tau PET in ageing and Alzheimer’s by MAPT N279K mutation in relation to tau positron
disease. Brain. 2018;141(5):1517–28. emission tomography features. Mov Disord. 2019;
25. Siderowf A, Keene C, Beach T, Montine T, Arora https://doi.org/10.1002/mds.27623.
A, Devous M, Navitsky M, Kennedy I, Joshi A, 32. Hashimoto H, Kawamura K, Igarashi N, Takei M,
Pontecorvo M, Lu M, Serrano G, Rose S, Wilson A, Fujishiro T, Aihara Y, Shiomi S, Muto M, Ito T,
Hellstern L, Coleman N, Mintun M.  Comparison of Furutsuka K, Yamasaki T, Yui J, Xie L, Ono M,
regional flortaucipir PET SUVr values to quantitative Hatori A, Nemoto K, Suhara T, Higuchi M, Zhang
tau histology and quantitative tau immunoassay in MR.  Radiosynthesis, photoisomerization, biodis-
patients with Alzheimer’s disease pathology: a clin- tribution, and metabolite analysis of 11C-PBB3 as a
ico-pathological study. J  Nucl Med. 2017;58(Suppl clinically useful PET probe for imaging of tau pathol-
1):629. ogy. J  Nucl Med. 2014;55(9):1532–8. https://doi.
26. Ono M, Sahara N, Kumata K, Ji B, Ni R, Koga S, org/10.2967/jnumed.114.139550.
Dickson DW, Trojanowski JQ, Lee VM, Yoshida M, 33. Hashimoto H, Kawamura K, Takei M, Igarashi

Hozumi I, Yoshiyama Y, van Swieten JC, Nordberg N, Fujishiro T, Shiomi S, Watanabe R, Muto M,
A, Suhara T, Zhang MR, Higuchi M.  Distinct bind- Furutsuka K, Ito T, Yamasaki T, Yui J, Nemoto K,
ing of PET ligands PBB3 and AV-1451 to tau fibril Kimura Y, Higuchi M, Zhang MR.  Identification of
strains in neurodegenerative tauopathies. Brain. a major radiometabolite of [11C]PBB3. Nucl Med
2017;140(3):764–80. https://doi.org/10.1093/brain/ Biol. 2015;42(12):905–10. https://doi.org/10.1016/j.
aww339. nucmedbio.2015.08.006.
27.
Marquié M, Normandin MD, Vanderburg CR, 34. Shimada H, Kitamura S, Ono M, Kimura Y, Ichise M,
Costantino IM, Bien EA, Rycyna LG, Klunk WE, Takahata K, Moriguchi S, Kubota M, Ishii T, Takado
Mathis CA, Ikonomovic MD, Debnath ML, Vasdev Y, Seki C, Hirano S, Shinotoh H, Sahara N, Tempest
N, Dickerson BC, Gomperts SN, Growdon JH, P, Tamagnan G, Seibyl J, Barret O, Alagille D, Zhang
Johnson KA, Frosch MP, Hyman BT, Gómez-Isla MR, Kuwabara S, Jang MK, Marek K, Suhara T,
T. Validating novel tau positron emission tomography Higuchi M. First-in-human PET study with 18F-AM-
tracer [F-18]-AV-1451 (T807) on postmortem brain PBB3 and 18F-PM-PBB3. Alzheimer Dement.
tissue. Ann Neurol. 2015;78(5):787–800. https://doi. 2017;7(Suppl 7):P1104.
org/10.1002/ana.24517. 35. Barret O, Seibyl J, Stephens A, Madonia J, Alagille
28. Schonhaut DR, McMillan CT, Spina S, Dickerson D, Mueller A, Berndt M, Kroth H, Capotosti F, Muhs
BC, Siderowf A, Devous MD Sr, Tsai R, Winer J, A, Pfeifer A, Tamagnan G, Dinkelborg L, Marek
Russell DS, Litvan I, Roberson ED, Seeley WW, K.  First-in-human PET studies with the next gen-
Grinberg LT, Kramer JH, Miller BL, Pressman P, eration tau agent 18-F PI-2620  in Alzheimer’s dis-
Nasrallah I, Baker SL, Gomperts SN, Johnson KA, ease, progressive supranuclear palsy, and controls.
Grossman M, Jagust WJ, Boxer AL, Rabinovici GD. Alzheimer Dement. 2017;13(7):P3–4.
230 M. Higuchi

36. Bohorquez SS, Barret O, Tamagnan G, Alagille

pii: jnumed.118.209650; https://doi.org/10.2967/
D, Seibyl J, Marek K, de Crespigny A, Weimer jnumed.118.209650.
R. Assessing optimal injected dose for tau PET imag- 40. Ikonomovic MD, Abrahamson EE, Price JC, Mathis
ing using [18F]GTP1 (Genentech Tau Probe 1). J Nucl CA, Klunk WE. [F-18]AV-1451 positron emission
Med. 2017;58(Suppl 1):848. tomography retention in choroid plexus: more than
37. Honer M, Gobbi L, Knust H, Kuwabara H, Muri "off-target" binding. Ann Neurol. 2016;80(2):307–8.
D, Koerner M, Valentine H, Dannals RF, Wong DF, https://doi.org/10.1002/ana.24706.
Borroni E. Preclinical evaluation of 18F-RO6958948, 41. Marquié M, Verwer EE, Meltzer AC, Kim SJW, Agüero
11
C-RO6931643, and 11C-RO6924963 as novel PET C, Gonzalez J, Makaretz SJ, Siao Tick Chong M,
radiotracers for imaging tau aggregates in Alzheimer Ramanan P, Amaral AC, Normandin MD, Vanderburg
disease. J Nucl Med. 2018;59(4):675–81. https://doi. CR, Gomperts SN, Johnson KA, Frosch MP, Gómez-­
org/10.2967/jnumed.117.196741. Isla T. Lessons learned about [F-18]-AV-1451 off-tar-
38. Wong DF, Borroni E, Kuwabara H, George N,
get binding from an autopsy-confirmed Parkinson’s
Rosenberg P, Lyketsos C, Resnick SM, Thambisetty case. Acta Neuropathol Commun. 2017;5(1):75.
M, Brasic J, Gapasin L, Willis W, Knust H, Guerard https://doi.org/10.1186/s40478-017-0482-0.
M, Belli S, Muri D, Carey T, Bedding A, Wandel 42. Wen GY, Wisniewski HM, Kascsak RJ.  Biondi ring
C, Hansrod T, Honer M, Moghekar A, Boess F, tangles in the choroid plexus of Alzheimer’s disease
Albert MS, Shaya E, Oh E, Ostrowitzki S, Dannals and normal aging brains: a quantitative study. Brain
RF, Comley RA.  First in-human PET study of 3 Res. 1999;832(1–2):40–6.
novel tau radiopharmaceuticals: [11C]RO6924963, 43. Tatebe H, Kasai T, Ohmichi T, Kishi Y, Kakeya

[11C]RO6931643, and [18F]RO6958948. Alzheimer T, Waragai M, Kondo M, Allsop D, Tokuda
Dement. 2015;11(7):P850–1. T.  Quantification of plasma phosphorylated tau
39. Betthauser TJ, Cody KA, Zammit MD, Murali D, to use as a biomarker for brain Alzheimer pathol-
Converse AK, Barnhart TE, Stone CK, Rowley HA, ogy: pilot case-control studies including patients
Johnson SC, Christian BT.  In vivo characterization with Alzheimer’s disease and down syndrome.
and quantification of neurofibrillary tau PET radioli- Mol Neurodegener. 2017;12(1):63. https://doi.
gand [18F]MK-6240 in humans from Alzheimer’s dis- org/10.1186/s13024-017-0206-8.
ease dementia to young controls. J Nucl Med. 2018;.
 Tau Accumulation and Network
Breakdown in Alzheimer’s Disease 19
Hirohisa Watanabe, Epifanio Bagarinao,
Takamasa Yokoi, Hiroshi Yamaguchi,
Shinsuke Ishigaki, Michihito Mausuda,
Masahisa Katsuno, and Gen Sobue

Introduction AD and other tauopathies [4, 5]. Intriguingly, it is

also clearly revealed that the process of Aβ accu-
The pathological hallmarks of Alzheimer’s dis- mulation starts 10–20 years before clinical signs
ease (AD) are the presence of extracellular senile of dementia become apparent in AD [6] in accor-
plaques containing amyloid β (Aβ) peptide, and dance with Nun study that pathologically proved
intracellular neurofibrillary tangles of the the presence of some subjects with good cogni-
microtubule-­ associated protein tau [1, 2]. tion during their lives irrespective of a significant
Astrogliosis associated with neurodegeneration increase in Aβ and tau accumulation correspond-
is also a significant hallmark of AD [3]. However, ing to advanced Braak’s stage V and VI [7].
the roles of Aβ, tau, and astrogliosis/neurodegen- Moreover, the healthy older subject can show
eration in developing dementia and their progres- increased tau tracer accumulation in the medial
sion pattern have not been fully elucidated. temporal lobe with aging [8]. These findings sug-
Recently, several studies have demonstrated gest that there are other factors or mechanisms
that Aβ PET tracers as well several tau PET trac- which act as resilience to or compensation against
ers including flortaucipir, THK5317, and PBB3 the AD-related pathological changes before the
will significantly accumulate in the specific areas development of dementia.
well correspond to previous pathological data in Advanced magnetic resonance imaging (MRI)
technique have provided the insightful mecha-
nisms with visualization of the large-scale func-
tional networks among brain regions [9–11]. A
H. Watanabe (*)
Brain and Mind Research Center, Nagoya University, functional network is generally presumed by the
Nagoya, Japan correlation between nodal endogenous low-­
Department of Neurology, Nagoya University frequency fluctuations of blood oxygenation
Graduate School of Medicine, Nagoya, Japan level–dependent (BOLD) signals using resting
Department of Neurology, Fujita Health University, state functional MRI (rs-fMRI). The rs-fMRI
Toyoake, Japan study is believed to have a potential to delineate
e-mail: nabe@med.nagoya-u.ac.jp the not only decreased but also increased connec-
E. Bagarinao · H. Yamaguchi · G. Sobue tivity that may be associated with resilient or
Brain and Mind Research Center, Nagoya University, compensatory process against dementia and
Nagoya, Japan aging process [12, 13]. On the contrary, Jones
T. Yokoi · S. Ishigaki · M. Mausuda · M. Katsuno et  al. proposed that such increased connectivity
Department of Neurology, Nagoya University can trigger the local overloads proliferating to
Graduate School of Medicine, Nagoya, Japan

© Springer Nature Singapore Pte Ltd. 2019 231

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_19
232 H. Watanabe et al.

downstream nodes eventually leading to orthogonal to the first component. Thus, it is

­widespread power or systems failures in AD [14]. expected to obtain not only disease-­specific SSM/
Since AD is considered to be a disconnection PCA patterns (patient’s patterns) but also normal
syndrome characterized by abnormalities in subject’s patterns. Besides, noise elements also
large-scale networks [10, 12, 15], it will be cru- will decrease if only the first few eigenvectors that
cial to understanding the relationship between show significant statistical differences between
RSNs and accumulation of Aβ and tau related to patient and healthy control are selected.
the maintenance of cognition or development of We applied SSM/PCA to the THK 5351 PET
dementia. images to identify the disease-specific spatial dis-
In this review, we summarized the recent tribution in a patient with AD [24]. To delineate a
advancement of spatial tau PET tracers’ retention specific AD-related topography, we limited the
pattern using data mining approach and disrup- analysis to the set of principal components that in
tion of RSNs on MRI in patients with AD. aggregate accounted for the top 50% of sub-
ject × voxel variability, and for which each prin-
cipal component contributed at least 10% to the
Spatial Distribution Pattern of Tau total variance in the scan data [25]. We also gen-
Retention in AD erated eigenimages and the associated subject-­
specific eigenimage scores which represented the
In patients with AD, all available tau tracers similarity of each’s preprocessed THK5351 con-
showed the retention predominantly in the infe- centration to the SSM/PCA-identified pattern.
rior temporal and parietal cortices corresponding Figure 19.1 shows a significant spatial
to the distribution of neurofibrillary tangles as THK5351 pattern corresponding to principal
well as the clinical phenotype of dementia [4, 6, component 1. This pattern accounted for 23.6%
16]. The severity of retention also related to cog- of the total subject voxel variance of the data and
nitive decline in AD [17]. However, first-­ showed 82.6% sensitivity and 79.1% specificity
generation of tau tracers also demonstrated for differentiating AD from healthy controls. We
off-target binding in the basal ganglia, midbrain, called this spatial pattern as AD-related THK5351
thalamus, hippocampus, choroid plexus, and pattern (ADRTP). The ADRTP was composed of
venous sinus. As for THK5351, binding to mono- three major clusters. Two clusters were mainly
amine oxidase B (MAO-B) presumably explains located in the bilateral inferior, middle, and supe-
most off-target binding but increased MAO-B is rior frontal gyrus. Cluster 3 was the major com-
also observed in disease-related brain regions ponent, which included the precuneus, posterior
linked to astrogliosis with misfolded protein cingulate cortex, inferior parietal lobule, inferior,
accumulation [18–20]. middle, and superior temporal gyrus, and fusi-
Principal component analysis (PCA) is one of form gyrus. Among the ADRTP-related regions,
the most popular data-driven multivariate statisti- the most prominent areas were the precuneus and
cal approaches to simplify several possible corre- PCC, followed by the lateral middle and superior
lated variables into a smaller number of frontal gyri corresponding to Braak stage IV and
uncorrelated variables. Scaled subprofile model V. The subject scores of ADRTP showed signifi-
(SSM) is PCA-based approach that can be used to cant correlation with general cognitive perfor-
extract disease-related spatial covariance patterns mance assessed by ACE-R and CDR score.
from brain imaging data such as cerebral blood Principal component 2 (PC 2) accounted for
flow and metabolism PET and SPECT [21–23]. It 11.8% of the total subject voxel variance com-
has the advantage to identify the most significant posed by bilateral basal ganglia, hippocampus,
sources of data variation associated with the dis- medial prefrontal cortex, and anterior cingulate
ease in the first few principal components. The cortex (Fig. 19.2). The distributions were consis-
second principal component has the direction tent with non-specific bindings sites of THK5351,
which maximizes variance among all directions which were previously reported [18, 19]. Subject
19  Tau Accumulation and Network Breakdown in Alzheimer’s Disease 233

Fig. 19.1  Spatial THK5351 pattern corresponding to retention in the precuneus, posterior cingulate cortex, dor-
principal component 1: Alzheimer’s disease-related solateral prefrontal cortex, and infra parietal lobule in
THK5351 distribution pattern. Scaled subprofile Alzheimer’s disease compared to control
model/principal component analysis showed specific tau

scores based on PC 2 had 70.8% sensitivity and Interestingly, SSM/PCA and THK5351 as
52.2% specificity in discriminating healthy con- well ICA and AV-1451 showed very similar spa-
trols from AD. There was no significant correla- tial distribution pattern in patients with
tion between PC 2 scores and ACE-R and CDR. AD.  Mainly, precuneus and posterior cingulate
Independent component analysis (ICA) is cortex were the core regions associated with the
also a data-driven approach and can separate progression of dementia. Data mining approach
the imaging signal into several temporally will provide for strong and novel aspects for the
correlated independent networks as spatial pathophysiology of tau PET retention in AD.
maps by multivariate decomposition without More recently, second-generation tau tracers
any a priori presumptions. Hoenig et al. inves- including [18F]RO-948, [18F]GTP1, [18F]PM-­
tigated AV-1451 PET network using ICA in PBB3, [18F]AM-PBB3, [18F]MK-6240, and [18F]
patients with AD [26]. The ICA showed in the PI-2620 have been developed [4, 6]. These tracers
detection of 10 independently coherent tau showed higher affinity for tau pathology, greater
pathology networks with highly functionally signal-to-background ratio, more excellent selectiv-
connected brain regions such as the precuneus ity, and less off-target binding to such as MAO-A
and cingulate cortex. The ten networks spa- and MAO-B than the first generation. However, the
tially resembled established language, frontal binding specificity to tau protein deposits must be
control, default mode, visuospatial, and hip- carefully validated in the development of new tau
pocampal networks. The total percent vari- tracers [4]. SSM/PCA and ICA will be beneficial for
ance explained by all ten components was identifying individual spatial tau tracer’s distribution
95.7%. with minimizing the effect of off-target bindings.
234 H. Watanabe et al.

Fig. 19.2  Spatial THK5351 pattern corresponding to retention in the basal ganglia, hippocampus, medial pre-
principal component 2: Healthy subjects-related frontal cortex, and anterior cingulate cortex in healthy
THK5351 distribution pattern. Scaled subprofile control
model/principal component analysis showed specific tau

Resting State Networks (RSNs) nection between two spatial regions of interest
which have the significant linear temporal corre-
Since neurons do not have potential to store inter- lation of the time-series supporting the idea that
nal reserves of glucose and oxygen, hemody- they are involved in the same underlying func-
namic response delivering glucose and oxygen in tional process. This low-frequency fluctuation
response to a demand for information processing phenomenon relates to some spontaneous neural
is crucial to their proper function. This cause activity [30] and thought to be related to
increasing oxyhemoglobin in the active area, giv- extremely excessive high resting energy con-
ing rise to a significant change in the regional sumption. There are many different measures to
ratio of oxy- to deoxyhemoglobin (dia- to para- analyze rs-fMRI data including seed-based func-
magnetic), providing a difference in magnetic tional connectivity analysis, independent compo-
susceptibility between the blood and the sur- nent analysis, and graph analysis.
rounding tissue. MRI can visualize the differ- Seed-based functional connectivity provides
ences of the deoxyhemoglobin content with the regions correlated with the time-series in a
image intensity that has been termed Blood seed region. Seed is generally determined based
Oxygenation Level Dependent (BOLD) [27–29]. on a hypothesis or prior results. The computation
Resting-state functional MRI measures the is a simple and intuitive understanding of the
intrinsic brain activity as low-frequency fluctua- result is easy.
tions (<0.1 Hz) in BOLD signals in the absence ICA is also frequently used for rs-fMRI analy-
of any explicit task or an input [9–11]. Functional sis. Using ICA, we can evaluate several canonical
connectivity (FC) is generally defined as the con- large-scale networks including default mode
19  Tau Accumulation and Network Breakdown in Alzheimer’s Disease 235

n­ etwork, salience network, executive control net- length, which is the average number of connec-
work, dorsal attention network, medial visual tions between all pairs of nodes measures func-
network, lateral visual network, sensorimotor tional integration (efficiency) in brain networks.
network, basal ganglia network, auditory net- Degree describes the number of connections of a
work and so on (Fig.  19.3). Classical default node. Highly connected nodes within the net-
mode network divided into three (precuneus net- work are called as “hub (nodes with higher cen-
work, dorsal default mode network and ventral trality)”. Hub will contribute to the large-scale
default mode network) or four (ventral (precu- effects of the network.
neus), posterior, anterior ventral, and anterior
dorsal default mode networks) parts depending
on the ICA results [14]. Relationship Between Tau
Graph theory, which has been developed to Retention and RSN Changes in AD
investigate the properties of complex networks
provides a theoretic framework of the local and According to the recent network degeneration
global brain networks organization. Graph theory hypothesis, tau pathology will stereotypically
has several essential indices. Clustering coeffi- propagate with disease progression spreading
cient describes the level of local neighborhood along RSNs in AD. We computed the similarity
clustering. High clustering coefficient corre- of the distribution between the THK5351 con-
sponds to well-connected areas with the same centration assessed by SSM/PCA, and canonical
functional specialization. Characteristic path RSNs generated from the group RSNs obtained

Fig. 19.3  Representative resting state networks. This figure shows the representative resting state networks on mag-
netic resonance imaging obtained by independent component analysis
236 H. Watanabe et al.

using group ICA [24]. THK5351 values and tral default mode network, visuospatial network,
intrinsic connectivity values were extracted from and language network.
all voxels within the RSNs. Then, the similarity Bilateral precuneus/PCC and the left dorsolat-
of the distribution between THK5351 intrinsic eral prefrontal cortex (DLPFC) corresponding to
connectivity was computed using Pearson’s cor- hubs showed the most significant difference in
relation. The estimated similarity values were THK5351 retention between early AD and
mostly negative in healthy controls indicating healthy controls in ADRTP.  According to seed-­
that the lower level of THK5351 concentration based connectivity analysis using two ROIs as
voxels is associated with higher intrinsic connec- the seed regions, the intrinsic connectivity of pre-
tivity within RSNs. cuneus/PCC significantly decreased in the left
On the contrary, in the patient with AD, a shift middle occipital gyrus, left superior temporal
towards a more positive association. Mainly, pre- gyrus, left amygdala/hippocampus, and right
cuneus/PCC network, right executive control net- fusiform gyrus (Fig. 19.5). That of left dorsolat-
work, ventral default mode network, visuospatial eral prefrontal cortex decreased to the left infe-
network, and language network showed signifi- rior parietal lobule.
cant similarity with THK5351 retention As described previously, Hoenig et  al. per-
(Fig. 19.4). The results of ICA and dual regres- formed seed-based connectivity analysis from
sion analyses also showed decreased connectivity regions which showed the maximum z-value of
within the right executive control network, ven- AV-1451  in each of the ten generated tau

Fig. 19.4  Relationship between THK5351 retention each canonical network, and the blue areas indicate
and canonical resting state networks. (a) Shows the decreased connectivity, as identified by independent com-
similarity of the THK5351 retention pattern and each ponent analysis of resting-state functional MRI data
canonical resting state network. (b) Is the results of inde- (Family wised error (FWE) at p < 0.05)
pendent component analysis. The green areas indicate
19  Tau Accumulation and Network Breakdown in Alzheimer’s Disease 237

A,D. B,E.

Fig. 19.5  Seed-based connectivity of the precuneus/ ease; (c) intrinsic connectivity of precuneus/PCC signifi-
posterior cingulate cortex. This figure shows seed-based cantly decreased to the hot areas (Family wised error
connectivity analysis from ROIs located in the precuneus/ correction (FWEc) p < 0.05, cluster defining threshold, p
PCC (a–c). (a) healthy control; (b), early Alzheimer’s dis- = 0.001, cluster size = 170).

p­ athology networks extracted from patients with dominantly in the inferior, medial, and lateral
mild-­to-­moderate AD by ICA in healthy adults temporal cortical areas, precuneus/posterior
(n  =  26) [26]. Then, they quantified the spatial cingulate, and lateral parts of the parietal and
overlap between the seed-based networks and the occipital cortex. Based on the standardized
corresponding tau pathology network using the maps of RSNs derived from large-scale rs-fMRI
similarity coefficient. Besides, they compare the data of a healthy adult population, they com-
tau-dependent seed-based networks and canoni- puted quantitative metrics including the mean
cal RSNs. As a result, the overlap between the tau Z-score of AV 1451 within each RSN template
pathology networks and similar seed-based net- and assessed the difference between the mean
works was a fair-to-moderate level. These net- Z-score value of voxels falling within a given
works also resembled well-known RSNs, ICN template and the mean Z-score value of
particularly the default mode network. cortical voxels outside the ICN template as a
Hansson et al. investigated the regional dis- goodness-of-fit (GOF) score. GOF showed that
tribution profile of AV 1451 pattern and canoni- AV 1451 retention pattern overlapped primarily
cal RSNs in two independent samples of with the dorsal attention, and to some extent
prodromal cases (the ADNI study, n = 35) and with higher visual, limbic and parts of the
manifest AD cases (the Swedish BioFINDER default-mode network. Prodromal AD showed
study, n  =  44) [31]. In manifest AD cases, the highly similar spatial distribution profile but
typical AV 1451 retention was observed pre- less pronounced.
238 H. Watanabe et al.

Hub Vulnerability in AD This compensatory mechanism will be important

to keep its function and associated with resil-
These three studies, which investigated the rela- ience to disease or aging process but be also
tionship between tau imaging networks assessed associated with overload of the hubs. If this
by THK5351 and AV 1451 and RSNs showed aberrant rerouting is critical and sustained, the
some differences but generally similar results hub nodes themselves may become affected
about the involvements of default mode network resulting in cascading network failure. It is con-
particularly precuneus and posterior cingulate sidered the one of the reasons why hub nodes are
cortex. Both the precuneus and posterior cingu- vulnerable to damage in various neurological
late cortex show high levels of metabolism, are diseases.
vascular boundaries, serve as hub, and key nodes
in the classical default mode network [32–35].
Pathologically, the precuneus/PCC is involved in Relationship Between Tau
Braak’s stage IV, leading to the appearance of Retention and Structural Network
initial clinical symptoms. Tau pathology can Changes in AD
spread along RSNs and extend to the precuneus,
and posterior cingulate cortex could keenly asso- Diffusion-weighted imaging (DWI) can non-­
ciate with widespread disruption of brain net- invasively provide the findings of white matter
works resulting in the manifestation of dementia fiber architecture in vivo. Mainly, diffusion ten-
in AD. sor imaging (DTI) has widely performed for
Cope et  al. investigated the relationship investigating the structural integrity of white
between retention pattern of AV-1451 and rs-­ matter in patients with AD.  Kitamura, et  al.
fMRI analyzed by graph theory in 17 patients showed that PBB3 retention in orbitofrontal cor-
with AD and 12 controls. This study showed that tex associated with cortical thickness in the area
strongly connected nodes displayed more tau and disruption of the orbitofrontal cortex – unci-
pathology in AD, independently of intrinsic con- nate fasciculus network, resulting in the develop-
nectivity network. Increasing tau burden in AD ment and progression of apathy in AD.  This
can be associated with decreasing the connectiv- result suggests the significant tau retention is
ity of these same nodes associated with reducing directly associated with neuronal loss and ana-
the weighted degree and local efficiency and tomical network changes [37].
weaker ‘small-world’ properties supporting the However, there are significant limitations in
idea of trans-neuronal spreading. In our study, the DTI including crossing-fiber and kissing-­
THK5351 retention was the most prominent in fiber if we particularly investigate the global
the precuneus/PCC and the dorsolateral prefron- brain network analysis. A recent technique that
tal cortex corresponding to hub. enables fiber tract-specific statistical analysis
Stam proposed that hub overload and failure using a specific fiber population within a voxel is
as a final common pathway. Under normal con- based on DWI data using constrained spherical
ditions, a brain network constitutes a multilay- deconvolution (fixel-based analysis, FBA). FBA
ered hierarchical structure with a hub as vertex. has a potential to characterize the multiple fiber
This is a well-designed organization for efficient orientations within voxels. Mito et  al. demon-
exchange of information locally and globally. strated that FBA showed significant white matter
Graph analysis showed that the brain organiza- involvement in specific fiber pathways related to
tion corresponds to an optimal balance between default mode network nodes in AD patients.
the segregation and integration for processing Intriguingly, patients with mild cognitive impair-
information [36]. Theoretically, nodes failure in ment patients only exhibited within the posterior
the lower level of hierarchy can cause the cingulum [38]. Future study will be needed to
enlargement of information processing load par- elucidate the relationship between tau retention
ticularly in the hubs with the highest centrality. and FBA abnormalities.
19  Tau Accumulation and Network Breakdown in Alzheimer’s Disease 239

Conclusions 2. Canter RG, Penney J, Tsai L-H. The road to restoring

neural circuits for the treatment of Alzheimer’s dis-
ease. Nature. 2016;539:187–96.
In this review, we summarized the results of the 3. Rodríguez-Arellano JJ, Parpura V, Zorec R,
spatial distribution pattern of cutting-edge tau Verkhratsky A.  Review astrocytes in physiological
PET tracer’ retention as well network disruption aging and Alzheimer’s disease. Neuroscience IBRO.
2016;323:170–82.
assessed by resting-state functional MRI and dif- 4. Okamura N, Harada R, Ishiki A, Kikuchi A, Nakamura
fusion tensor imaging in AD. T, Kudo Y.  The development and validation of tau
The data drive approach such as SSM/PCA PET tracers: current status and future directions. Clin
and ICA has great advantage to extract the Transl Imaging 6th ed., Springer. 2018;6:305–16.
5. Valotassiou V.  SPECT and PET imaging in
disease-­ specific stereotypically spatial pattern Alzheimer’s disease. Ann Nucl Med 3rd ed., Springer,
from brain tau tracer signals as multivariate Japan. 2018;32:583–93.
decomposition in AD.  SSM/PCA and ICA can 6. Villemagne VL, Doré V, Burnham SC, Masters CL,
also be useful for decreasing noise and off-­ Rowe CC. Imaging tau and amyloid-β proteinopathies
in Alzheimer disease and other conditions. Nat Rev
targeted bindings data. It is believed that these Neurol. 2018;14:225–36.
methods will be appropriate for not only diagno- 7. Snowdon DA. Aging and Alzheimer’s disease: lessons
sis but also better understanding the spatially dis- from the Nun Study. Gerontologist. 1997;37:150–6.
tribution pattern of tau in AD. 8. Schöll M, Lockhart SN, Schonhaut DR, et  al. PET
imaging of tau deposition in the aging human brain.
According to recent comparative studies of Neuron. 2016;89:971–82.
PET and RSNs, Aβ and tau retention can spread 9. Lv H, Wang Z, Tong E, et al. Resting-state functional
along specific RSNs including default mode MRI: everything that nonexperts have always wanted
network, executive control network, visuospa- to know. Am J Neuroradiol. 2018;60:1672–10.
10. Pievani M, de Haan W, Wu T, Seeley WW, Frisoni
tial network, and language network. Hub GB.  Functional network disruption in the degen-
lesions of the networks, which can demand erative dementias. Lancet Neurol Elsevier Ltd.
higher metabolic and have differential gene 2011;10:829–43.
expression showed the most prominent tau 11. van den Heuvel MP, Pol HEH.  Exploring the brain
network: a review on resting-state fMRI functional
retention among these networks. Particularly, connectivity. Eur Neuropsychopharmacol Elsevier
precuneus and PCC can be one of the most BV. 2010;20:519–34.
important vital nodes and disruptions of func- 12. Mormino EC, Smiljic A, Hayenga AO, et  al.
tional connectivity from these hubs to the pres- Relationships between beta-amyloid and func-
tional connectivity in different components of the
ence of dementia in AD. default mode network in aging. Cereb Cortex.
Although the reason why hub regions are vul- 2011;21:2399–407.
nerable to AD pathology is not still elucidated, 13. Anthony M, Lin F. A systematic review for functional
recent studies support the view that AD can be neuroimaging studies of cognitive reserve across the
cognitive aging Spectrum. Arch Clin Neuropsychol.
the disconnection syndrome with disruption of a 2017;33:937–48.
large-scale network. Longitudinal PET and RSNs 14. Jones DT, Knopman DS, Gunter JL, et al. Cascading
observation study from healthy aging to dementia network failure across the Alzheimer’s disease spec-
will shed light on the mechanism of not only trum. Brain. 2016;139:547–62.
15.
Delbeuck X, Van der Linden M, Collette
resiliency and compensatory mechanisms against F. Alzheimer’s disease as a disconnection syndrome?
Aβ and tau retention but also the picture of local Neuropsychol Rev Kluwer Academic Publishers-
overloads and failure of specific hubs. Plenum Publishers. 2003;13:79–92.
16. Saint-Aubert L, Lemoine L, Chiotis K, Leuzy A,

Rodriguez-Vieitez E, Nordberg A.  Tau PET imag-
ing: present and future directions. Mol Neurodegener
References Molecular Neurodegeneration. 2017;12:1–21.
17. Bejanin A, Schonhaut DR, La Joie R, et  al. Tau

1. Montine TJ, Phelps CH, Beach TG, et  al. National pathology and neurodegeneration contribute to cog-
Institute on Aging–Alzheimer’s Association nitive impairment in Alzheimer’s disease. Brain.
guidelines for the neuropathologic assessment of 2017;140:3286–300.
Alzheimer’s disease: a practical approach. Acta 18. Ng KP, Pascoal TA, Mathotaarachchi S, et  al.
Neuropathologica Springer-Verlag. 2011;123:1–11. Monoamine oxidase B inhibitor, selegiline, reduces
240 H. Watanabe et al.

18F-­THK5351 uptake in the human brain. Alzheimers 28. Handwerker DA, Bandettini PA.  Hemodynamic

Res Ther. 2017;9:1–9. signals not predicted? Not so: a comment on
19. Lemoine L, Gillberg P-G, Svedberg M, et  al. Sirotin and Das (2009). NeuroImage Elsevier BV.
Comparative binding properties of the tau PET tracers 2011;55:1409–12.
THK5117, THK5351, PBB3, and T807 in postmortem 29. Sirotin YB, Das A. Anticipatory haemodynamic sig-
Alzheimer brains. Alzheimers Res Ther. 2017;9:1–13. nals in sensory cortex not predicted by local neuronal
20. Harada R, Ishiki A, Kai H, et  al. Correlations of
activity. Nature. 2009;457:475–9.
18F-THK5351 PET with postmortem burden of Tau 30. Ma Y, Shaik MA, Kozberg MG, et  al. Resting-

and astrogliosis in Alzheimer disease. J  Nucl Med. state hemodynamics are spatiotemporally coupled
2018;59:671–4. to synchronized and symmetric neural activity
21. Moeller JR, Strother SC, Sidtis JJ, Rottenberg
in excitatory neurons. Proc Natl Acad Sci U S A.
DA. Scaled subprofile model: a statistical approach to 2016;113:E8463–71.
the analysis of functional patterns in positron emis- 31. Hansson O, Grothe MJ, Strandberg TO, et  al. Tau
sion tomographic data. J  Cereb Blood Flow Metab. pathology distribution in Alzheimer’s disease corre-
1987;7:649–58. sponds differentially to cognition-relevant functional
22. Alexander GE, Moeller JR. Application of the scaled brain networks. Front Neurosci. 2017;11:270–11.
subprofile model to functional imaging in neuropsy- 32. Cavanna AE, Trimble MR.  The precuneus: a review
chiatric disorders: a principal component approach to of its functional anatomy and behavioural correlates.
modeling brain function in disease. Hum Brain Mapp Brain. 2006;129:564–83.
Wiley-Blackwell. 1994;2:79–94. 33. Zhang S, C-SR L.  Functional connectivity map-

23. Spetsieris P, Ma Y, Peng S, et  al. Identification of ping of the human precuneus by resting state
disease-related spatial covariance patterns using neu- fMRI. NeuroImage Elsevier Inc. 2012;59:3548–62.
roimaging data. J Vis Exp. 2013;76:1–14. 34. Leech R, Sharp DJ.  The role of the posterior cin-
24. Yokoi T, Watanabe H, Yamaguchi H, et al. Involvement gulate cortex in cognition and disease. Brain.
of the precuneus/posterior cingulate cortex is signifi- 2013;137:12–32.
cant for the development of Alzheimer’s disease: a 35. Utevsky AV, Smith DV, Huettel SA.  Precuneus

PET (THK5351, PiB) and resting fMRI study. Front is a functional core of the default-mode network.
Aging Neurosci. 2018;10:e17–5. J Neurosci. 2014;34:932–40.
25. Niethammer M, Tang CC, Feigin A, et al. A disease-­ 36. Stam CJ.  Modern network science of neurological
specific metabolic brain network associated with cor- disorders. Nat Rev Neurosci. 2014;15:683–95.
ticobasal degeneration. Brain. 2014;137:3036–46. 37. Kitamura S, Shimada H, Niwa F, et  al. Tau-induced
26. Hoenig MC, Bischof GN, Seemiller J, et al. Networks focal neurotoxicity and network disruption related to
of tau distribution in Alzheimer’s disease. Brain. apathy in Alzheimer’s disease. J  Neurol Neurosurg
2018;141:568–81. Psychiatry. 2018;89:1208–14.
27. Ogawa S, Lee TM, Kay AR, Tank DW.  Brain mag- 38. Mito R, Raffelt D, Dhollander T, et al. Fibre-specific
netic resonance imaging with contrast dependent on white matter reductions in Alzheimer’s disease and
blood oxygenation. Proc Natl Acad Sci USA National mild cognitive impairment. Brain. 2018;141:888–902.
Academy of Sciences. 1990;87:9868–72.
 Stress and the Etiopathogenesis
of Alzheimer’s Disease 20
and Depression

Ioannis Sotiropoulos, Joana M. Silva,
Patricia Gomes, Nuno Sousa,
and Osborne F. X. Almeida

Introduction (GC) secretion, trigger the two main pathomech-

anisms of AD: (i) misprocessing of amyloid pre-
The brain is the most adaptive of all organs. It has cursor protein (APP) and the generation of
a remarkable capacity to respond to a variety of amyloid beta (Aβ) and (ii) Tau hyperphosphory-
internal and environmental stimuli and to mount lation and aggregation. Given that depression is a
pro-survival behavioral responses by orchestrat- well-known stress-related illness, and the evidence
ing multiple molecular and biochemical cas- that depression may precede AD, this chapter
cades. The latter changes are embraced by the also explores neurobiological mechanisms that
term neural plasticity, the cornerstone of learning may be common to depressive and AD patholo-
and memory [1]. Impairments in neuroplastic gies. This review also discusses emerging insights
mechanisms are commonly found during aging, into the role of Tau and its malfunction in disrupt-
the primary risk factor for Alzheimer’s disease ing neuronal cascades and neuroplasticity and,
(AD), a disorder characterized by memory defi- thus triggering brain pathology.
cits. Over their lifetime, individuals experience
both good and adverse (stressful) events and
notably, stressful events appear to accelerate Stress: A Physiological Tug-of-War –
brain aging [2].  Accumulating clinical and From Adaptive to Maladaptive
experimental evidence suggests a causal role of Responses
lifetime stress in AD.  This chapter summarizes
current knowledge about how chronic stress and Stress is defined as a challenge to homeostasis
its accompanying high levels of glucocorticoid (physiological and behavioral equilibrium) by
physical or psychological events [3]. When
challenged by endogenous or exogenous aversive
I. Sotiropoulos (*) · J. M. Silva · P. Gomes or threatening stimuli (stressors), a series of
N. Sousa
Life and Health Sciences Research Institute (ICVS), defensive systems and processes become acti-
School of Medicine, University of Minho – Campus vated; these include the release of monoamines
de Gualtar, Braga, Portugal and GC that initially promote a return to the
ICVS/3B’s – PT Government Associate Laboratory, homeostatic state [4, 5]. The “stress response”
Braga/Guimarães, Portugal normally terminates once homeostasis has been
e-mail: ioannis@med.uminho.pt restored, but when the organism is faced with an
O. F. X. Almeida insurmountable stress (high intensity, contextu-
Max Planck Institute of Psychiatry, ally inappropriate and/or chronic), it may take
Munich, Germany

© Springer Nature Singapore Pte Ltd. 2019 241

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_20
242 I. Sotiropoulos et al.

inappropriate – maladaptive – actions that result pared to GR and are believed to play an important
in chronically elevated GC secretion. Besides role in GC feedback under physiological (non-­
interfering with normal structural and plastic stressful) conditions [11]. In addition, MR have
arrangements within the brain, such inadequate been implicated in “protecting” the brain against
responses can have negative consequences for the GR-mediated cytotoxicity [12, 13] and behav-
immune and visceral systems that may ultimately ioral maladaptation [14]. Interestingly, the cen-
lead to multiple disorders, including neuropsy- tral expression of GR and MR is subject to
chiatric and neurological diseases [6–9]. regulation by stress and age [15, 16].
The endocrine response to stress is orches- While GC have been shown to modulate syn-
trated within the so-called hypothalamo–pituitary– aptic activity through non-genomic mechanisms
adrenal (HPA) axis (Fig. 20.1). Stress, perceived [17–19], GR and MR are better known as potent
by cortical areas of the brain, triggers the release ligand-dependent transcriptional regulators, i.e.
of corticotropin-releasing hormone (CRH) from control gene expression/repression [20–23]. The
the hypothalamic paraventricular nucleus (PVN) unliganded receptors are located in the cytoplasm
which, in turn, induces the secretion of adreno- in association with chaperone proteins (e.g. the
corticotropic hormone (ACTH) release from the heat shock proteins Hsp90 and 70 and the immu-
pituitary and GC (cortisol in humans and corti- nophilins FKBP51 and 52) [24]. Ligand binding
costerone in rodents) from the adrenal glands. results in conformational change of the
This sequence of events is normally curtailed by GR-chaperone complex and subsequently, recep-
negative feedback of GC at central sites; how- tor translocation to the nucleus and binding to
ever, the nature of the stressor and/or impair- specific regions of DNA containing glucocorti-
ments in negative feedback mechanisms (e.g. coid response elements (GRE) within the pro-
during aging) may block this crucial feedback moters of target genes [25]. Gene transcription or
loop, resulting in supraphysiological exposure to repression is then determined by the recruitment
GC. A key area among the brain regions involved of co-activators and repressors [26] as well as by
in the regulation of the HPA axis is the hippo- post-translational modifications of the receptor
campus; this area, which also plays a pivotal role [27–29].
in learning and memory, sends inhibitory projec- In the brain, MR and GR differentially regu-
tions to the PVN (and other hypothalamic nuclei). late the expression of genes, in a site-specific
Similarly, the frontal cortex mediates GC nega- manner; these include genes responsible for the
tive feedback effects on the HPA axis, whereas regulation of the HPA axis (CRH and CRH recep-
the GC activation of the amygdala results in a tors and pro-opiomelanocortin [POMC], from
positive drive on this axis (see Fig. 20.1). whose gene product ACTH is cleaved) as well as
pro- and anti-apoptotic genes [13] and, impor-
tantly, genes with roles in neural energy metabo-
Mechanisms and Consequences lism, structure and synaptic transmission, the
of GC Action in the Brain synthesis of rate-limiting neurotransmitter
enzymes and receptors as well of various neuro-
Corticosteroid actions in the brain are mediated peptide, growth factors and cell adhesion mole-
by glucocorticoid (GR) and mineralocorticoid cule [30–34]. While all of these GC-initiated
(MR) receptors. The previously-referred feed- transcriptional events contribute to neural plastic-
back effects of GC during stressful experiences ity, stress and GC result in the manifestation of
primarily depend on activation of GR, expressed visualizable effects, namely, alterations in neuro-
throughout the brain, but at highest density in the nal morphology. As reviewed by Lucassen [35],
hippocampus [10]. Indeed, almost all we know stress and GC lead to changes in the rate of neu-
about GC actions in the brain are GR-mediated. rogenesis, cell death, and neuronal connectivity
Less is known about the role of MR, although, as well as astroglia-neuronal interactions. In par-
they have a ten-fold higher affinity for GC com- ticular, numerous studies have highlighted how
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 243

Fig. 20.1  The hypothalamo-pituitary-adrenal (HPA) in humans and corticosterone in rodents] from the adre-
axis and the response to stress in the healthy state. nal cortex. The secreted GC access peripheral and central
Stressors perceived in higher brain centers trigger the tissues via the general circulation where they serve to
release of corticotropin-releasing hormone (CRH) from mount adaptive responses to the initiating stimulus
neurons in the paraventricular nucleus (PVN). Carried (stress) after binding to glucocorticoid receptors (GR).
via a portal vein system, CRH reaches the anterior pitu- Eventually, GC secretion and action is restrained by
itary where it stimulates the secretion of adrenocortico- inhibitory feedback of GC on central (chiefly, the frontal
tropin hormone (ACTH) which, in turn, stimulates the cortex, hippocampus, hypothalamus and pituitary) com-
production and release of glucocorticoids (GC) [cortisol ponents of the HPA axis
244 I. Sotiropoulos et al.

stress, acting through GC, impacts on dendritic of mossy fiber synapses, increased surface area
arborization and synaptic number; this aspect of of the post-synaptic density, and rearrangements
GC actions is considered in the following s­ ection. of synaptic mitochondria and vesicles at the pre-
First, in the context of AD as a disease that devel- synaptic terminals [54]. Further, dendritic spines,
ops as age progresses, it is important to briefly which have an important role in information stor-
mention the growing view that stress and GC age, are severely reduced by stress [55] but can
leave long-lasting “memories” of past experi- mostly be rapidly reversed after a recovery period
ences via epigenetic mechanisms; these are or subsequent training ([56, 57]; but see [58] for
thought to contribute importantly to the organ- exceptions).
ism’s physical and mental health trajectory [36– New work from our labs indicates that Tau, a
39]. Notably, epigenetic mechanisms have key factor in AD pathology, is essential for
recently been implicated in the lasting effects of chronic stress to disrupt neuroplasticity. Briefly,
lifetime adversity in humans [40, 41]. we showed that mice in which Tau has been
deleted are spared from the deleterious behav-
ioral (e.g. deficits in learning and memory,
 tress, Glucocorticoids and Neural
S depressive-like behavior and anxiety) and neuro-
Plasticity structural (namely, dendritic atrophy disconnec-
tion of the hippocampal-prefrontocortical
Functional plasticity in the brain is generally pre- pathway) of chronic stress and GC [59, 60].
ceded by structural plasticity, typically, dendritic As already alluded to, stress and GC also
and synaptic remodeling. Basal levels of GC are influence neuroplasticity by modulating the pro-
crucial for maintaining synaptic plasticity in the duction of new neurons in adult brain [61].
hippocampus in the form of long-term potentia- Several studies indicate that stress/GC related
tion (LTP) [42], a well-documented mechanism effects on neurogenesis have the potential to
involved in memory formation [43]. On the other affect mental health, including susceptibility to
hand, high levels of GC such as those experi- depression [62, 63] and AD [64, 65].
enced during stress impair LTP induction and
facilitate long-term depression (LTD) [44]. An
important role for N-methyl-d-aspartate (NMDA)  hronic Stress: Etiopathogenic Role
C
receptors and shifts in calcium flux has been sug- and Mechanisms in AD
gested in both, LTP and LTD modulation by
stress/GC [45–48]. A sizeable literature suggests that elevated GC
One of the best-described forms of stress-­ and chronic stress – a state that an increasing pro-
evoked structural plasticity is dendritic retrac- portion of the population finds itself in today  –
tion, with a pioneering study revealing that may increase the risk for developing AD
chronic stress interrupts connectivity between pathology and related dementias [66, 67] and
hippocampal CA1 neurons and neurons in the may even advance the age of onset of the familial
medial prefrontal cortex (PFC) [49, 50]. The lat- form of AD [68–70]. Indirect support for the link
ter work followed previous demonstrations that between high GC exposure and AD includes
chronic stress causes atrophy of apical (but not reports that AD patients produce and secrete
basal) dendritic complexity in CA3 pyramidal higher-than-normal levels of cortisol [67, 71–74].
neurons [51]. Meanwhile, other studies have Interestingly, transgenic mouse models of AD
reported that stress can also increase dendritic also display high levels of GC [75, 76].
length in certain brain regions such as the orbito- Nevertheless, while the direction of the cause-­
frontal cortex, amygdala and bed nucleus of the effect relationship between AD-like pathology
stria terminalis (BNST, also known as the and hypercorticalism remains unclear, it is worth
“extended amygdala”) [52, 53]. Interestingly, recalling (see previous section) that the hippo-
chronic stress has also been associated with a loss campus is responsible for mediating the negative
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 245

feedback effects of GC on the HPA axis; thus, memory deficits in mice expressing an aggressive
any damage to this brain region is likely to (human) mutant form of APP V717ICT-100 [76,
uncouple this control mechanism and unleash 83]. Similar observations were made when young
unrestrained GC secretion. 3xTg-AD mice (expressing APP Swedish,
To put these findings into context, it is worth P301L-Tau, and PSEN1 M146V mutations) were
noting that there is evidence that GC levels cor- treated with the synthetic GC, dexamethasone
relate with the rate of cognitive decline [35, 77] [76]. That the effects of stress are most likely
and the extent of neuronal remodeling in AD sub- transduced by GC was demonstrated by experi-
jects [78]. Such remodeling is especially marked ments on dexamethasone-treated neural cell lines
in the hippocampus, the area in which most stud- (N2A [76] and differentiated PC12 cells [84]).
ies on stress/GC effects on neuroplasticity have Consistent with these reports, our own studies in
been conducted in rodents, and the brain area wildtype rats demonstrated that chronic stress
which clearly displays the first signs of AD neu- and/or treated with GC increases APP mispro-
ropathology  – deposits of amyloid β (Aβ) and cessing along the amyloidogenic pathway by
accumulation and aggregation of hyperphosphor- upregulating BACE-1 and Nicastrin (a compo-
ylated Tau [79–81]. The hippocampal lesions nent of the γ-secretase complex) to produce neu-
induced by these deposits correlate with the rotoxic and cognition-impairing effects [85]; in
extent of deficits in declarative, spatial and con- this regard, it is worth noting that high exogenous
textual memory [82]. levels of GC upregulate the transcription of APP
and βACE-1, the promoters of which contain a
glucocorticoid response element (GRE) [76].
Consideration Regarding How Lastly, experiments that attempted to mimic
Chronic Stress and High GC Levels intermittent stressful events (the effects of which
May Contribute to AD Pathology may be cumulative over the lifetime) showed that
GC potentiate the APP misprocessing pathway
In this section, we will review some of the evi- [85].
dence for a link between GC/stress and AD and In recent years, an increasing amount of atten-
consider some of the possible underlying mecha- tion has turned to Tau pathology, especially its
nisms. After briefly considering stress/GC effects hyperphosphorylated forms, in a range of neuro-
on amyloidogenesis, our attention will focus on degenerative diseases. Among the first reports to
how chronic exposure to stress or high levels of indicate a relationship between stress/GC and
GC influence Tau biology, culminating in its mal- Tau was a study by Stein-Behrens et al. [86] who
function and dendro-synaptic toxicity. found that GC exacerbate kainic acid-induced
As noted earlier, AD neuropathology is char- hippocampal neuronal loss with a contemporane-
acterized by overproduction of Aβ that forms ous increase in Tau immunoreactivity. A later
deposits into senile (amyloid) plaques, and by study showed that chronic treatment of 3xTg AD
accumulation of hyperphosphorylated forms of mice with dexamethasone leads to the somato-
Tau protein that becomes insoluble, aggregates dendritic accumulation of Tau in the hippocam-
and forms neurofibrillary tangles (NFT) [79–81]. pus, amygdala and cortex [76]. Our own in vivo
Aβ is the proteolytic product of amyloid precur- studies demonstrated that chronic stress or GC
sor protein (APP), a large transmembrane protein increase the levels of aberrantly hyperphosphory-
called, that is sequentially cleaved by β-secretase lated Tau in the rat hippocampus and PFC, both
(BACE-1) and γ-secretase (a complex of in the presence and absence of exogenous Aβ
enzymes, including presenilin) to yield Aβ; this [87]. Importantly, the hyperphosphorylation
post-translational pathway is often called APP occurred at certain Tau epitopes that are strongly
misprocessing. Studies have shown that extended implicated in cytoskeletal dysfunction and syn-
exposure to immobilization stress increases the aptic loss (e.g., pSer262) [88, 89] and hippocam-
load of extracellular Aβ deposits and exacerbates pal atrophy (e.g., pThr231) [90] in AD patients.
246 I. Sotiropoulos et al.

Here, it is pertinent to note that the extent of While Kobayashi et al. [101] showed Tau may
phosphorylation at Thr231- and Ser262-Tau be synthesized de novo in the somato-dendritic
correlates strongly with severity of memory
­ compartment, earlier work by Ittner and col-
impairment, speed of mental processing, and leagues [102] demonstrated that hyperphosphor-
executive functioning in AD patients [91–93]. ylated Tau is missorted to synapses which
Although chronic stress and GC treatment exert subsequently become dysfunctional. The mis-
similar, but not identical, effects on individual sorting of Tau to synapses is now acknowledged
Tau phosphoepitopes in vivo and in vitro [84], the as an early event in AD, preceding the manifesta-
overall evidence points to GC as the key mediator tion of detectable neurodegenerative processes
of the AD-like pathology induced by stress. On [102–104]. It is important to note that this series
the other hand, some studies have suggested a of events depend on Tau hyperphosphorylation
role for at least one other stress-related molecule, [103, 105, 106] and results in the targeting of Fyn
namely, corticotrophin-releasing hormone (CRH) (a member of the Src kinase family) to postsyn-
as deletion of the CRH receptor 1 gene in mice aptic sites [102] where it selectively modulates
prevents the detrimental effects of stress on Tau the function of GluN2B-containing NMDAR
phosphorylation [94, 95]. Supporting these links, (GluN2BR), by phosphorylation of the GluN2B
are the results from in vitro experiments which at the Y1472 epitope [102, 107]. The latter stabi-
indicate that the GC effects on Tau involve acti- lizes GluN2B at postsynaptic sites, thus increas-
vation of glycogen synthase kinase 3 (GSK3) and ing the risk for excitotoxicity [102, 107].
cyclin-dependent kinase 5 (CDK5), two principal Since NMDAR are known to mediate stress-
Tau kinases [84]. and GC-driven neurotoxicity [108] and neuronal
Transgenic mice expressing human P301L-­ remodeling [109], we were prompted to examine
Tau (the most common Tau mutation), also whether the mechanistic scenario just described
helped strengthen the evidence that chronic stress also applies to the actions of stress and
can exacerbate Tau pathology. Briefly, we found GC. Indeed, we found that chronic stress and GC
that stress stimulates the aberrant hyperphos- also trigger Tau accumulation at synapses with
phorylation and aggregation of insoluble Tau subsequent increases of Fyn at postsynaptic sites
[96]. Further, we demonstrated in the latter work [59, 110] (see also Fig. 20.2).
that chronic stress enhances caspase 3-mediated Other mechanisms that may underlie the abil-
truncation of Tau at its C-terminal in the hippo- ity of stress/GC to contribute to AD pathology
campus, with the protein misfolding and adopt- have been coming to light in the last few years.
ing a conformation [96] that facilitates its One of these is autophagy. As the guardian of
nucleation and recruitment of other Tau mole- cellular homeostasis, autophagy is now seen to
cules into neurotoxic, pre-tangle aggregates of play a pivotal role in the pathology of a number
Tau (see [97–100]. Importantly, experiments also of neurodegenerative disorders [111, 112].
showed that GC contribute to AD pathology by Briefly, autophagic mechanisms are responsible
reducing the degradation of Tau, thereby increas- for the degradation of misfolded proteins and
ing its accumulation [84]. The latter is likely to aggregates such as Tau aggregates; interruption
result from dysregulation of molecular chaper- of autophagy leads to the accumulation of pro-
ones (e.g. Hsp90 and Hsp70) that are responsible tein aggregates, a pathological features shared
for Tau proteostasis [96]. As noted previously, by a range of neurodegenerative disorders [113,
these same heat shock proteins serve to maintain 114]. Our investigations demonstrated (summa-
GR in a high affinity state; thus, they may repre- rized in Fig.  20.3), for the first time, that both,
sent a point at which GC signaling intersects with chronic stress and high GC levels inhibit autoph-
the cellular machinery that regulates Tau agic process, thus explaining how these condi-
degradation. tions contribute to the accumulation and
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 247

Fig. 20.2  Multiple mechanisms contribute to the acetylation of tubulin and cortactin. As described in the
induction of Tau pathology and AD by chronic stress. text, GC induce hyperphosphorylation of Tau and its con-
The scheme summarizes the potential mechanisms sequent detachment from microtubules, leading to micro-
through which chronic stress and GC activate processes tubule destabilization and cytoskeletal disturbances that,
that result Tau accumulation, aggregation and neurotoxic- together with HDAC6, may contribute to: (i) the forma-
ity. Stress leads to increased activation of glucocorticoid tion of stress granules (SG) that promote Tau aggregation
receptors (GR) by GC; GR transcriptional activity and (ii) the inhibition of autophagic process that also con-
depends on an interplay of a variety of molecular chaper- tribute to Tau accumulation and aggregation. Interestingly,
ones (e.g. Hsp90, Hsp70, FKBP51) and HDAC6, a protein stress/GC inhibit mTOR, a crucial signaling molecule in
that may lead to cytoskeletal instability by reducing the the initial phases of autophagy

aggregation of Tau [115]. In fact, defective inhibition of mTOR blocked the GC-triggered
autophagy is suggested to be major player in AD Tau accumulation and aggregation [115].
pathology [116–118]; although Tau itself is a New research has implicated the endolyso-
proteosomal substrate [119, 120], it is thought somal pathway in neurodegenerative diseases
that Tau inclusions and aggregates may be inac- such as AD and Parkinson’s disease in which Tau
cessible to the ubiquitin-proteasome system accumulation is a pathological feature [128–130].
[121, 122]. Our results showing that chronic As shown in Fig. 20.2, Tau has been identified as
stress and GC increase mTOR signaling and a substrate of the endolysosomal degradation
reduces the ratio of the autophagic markers pathway [131]. We demonstrated that in vitro or
LC3II:LC3I and accumulation of p62 [115], in vivo exposure to elevated GC levels block this
indicate that chronic stress inhibits the autopha- pathway, accompanied by increases in the build-
gic process by activating the mTOR pathway; ­up of Tau, including that of specific phospho-Tau
these findings are in line with previous reports species. Further, we showed that the involvement
that chronic stress stimulates mTOR activity in of the small GTPase Rab35 and the endosomal
the hippocampus [123], an event associated with sorting complexes required for transport
increased total Tau levels in the brains of AD (ESCRT) machinery that delivers Tau to lyso-
subjects [124, 125]. In addition, support for our somes via early endosomes and multivesicular
interpretation comes from the finding that inhibi- bodies (MVBs). The ESCRT system mediates the
tion of mTOR signaling ameliorates Tau pathol- degradation of membrane-associated proteins
ogy [126, 127] while we demonstrated that such as epidermal growth factor receptor [132],
248 I. Sotiropoulos et al.

Fig. 20.3  Cumulative effects of stress and glucocorti- axis. The model assumes that elements of the HPA axis
coids on normal and pathological aging. In this hypotheti- serve as part of a threshold-regulator mechanism (repre-
cal representation of brain aging, cognitive and mood sented by thick line). Note that brain areas important for
status may decline over time. Chronic exposure to stress- regulation of HPA axis (e.g. hippocampus, prefrontal cor-
ful conditions, associated with higher exposure to GC, tex) also appear to be subject to impairments triggered by
lead to cumulative effects that accelerate brain aging by presymptomatic AD pathology (e.g. mild cognitive
imposing an increasing allostatic load on brain function impairment, depression), thus feeding into a vicious cycle
by causing neuronal atrophy and synaptic loss, modified that further drives GC secretion and neuronal damage.
by other factors such as genetics and sex – the latter also The shaded grey area represents the threshold-transition
influence the magnitude of the stress load by modulating area where a subject may progress from depression (with
the activity of the hypothalamo-pituitary-adrenal (HPA) or without MCI symptoms) to Alzheimer’s disease (AD)

but is also implicated in the degradation of cyto- (Fig.  20.2). Further, overexpression of Rab35
solic proteins GAPDH and aldolase [133]; these reverses GC-induced Tau accumulation and res-
findings are of particular relevance for Tau, which cues hippocampal neurons from the dystrophic
has both cytosolic and membrane-associated actions of chronic stress [131].
pools [134, 135], and has been shown to localize
to different neuronal sub-compartments, depend-
ing on its phosphorylation state [103, 110].  NA-Binding Proteins and Stress
R
Interestingly, not all phosphorylated Tau species Granules Facilitate Stress-Induced
are equally susceptible to degradation in the Tau Pathology
Rab35/ESCRT pathway. In particular, we found
that pSer396/404 and pSer262, but not pSer202, Stress granules (SG) have been recently impli-
phospho-Tau species undergo Rab35-mediated cated in the Tau pathology that accompanies AD
degradation, indicative of preferential sorting of and fronto-temporal dementia with parkinson-
specific phospho-Tau proteins into the Rab35/ ism-­17 (FTDP-17) in humans as well as in vari-
ESCRT pathway [131]. Importantly, we demon- ous transgenic mouse models of Tau-related
strated that high GC levels suppress Rab35 tran- disorders [136]. The eukaryotic stress response
scription, and thus, result in an accumulation of involves translational suppression of non-­
Tau due impaired degradation of the protein housekeeping proteins and the sequestration of
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 249

unnecessary mRNA transcripts by RNA-binding trates our current working model, designed to
proteins (RBP) into SG.  These macromolecular explore more about the biology of RNA-protein
complexes constitute a protective mechanism interactions in stress-related pathologies.
against cellular stress (e.g. oxidative stress) that
help protect mRNA species and enable the fast
production of cytoprotective proteins [136–138].  au and Its Malfunction in Stress-­
T
However, prolonged SG induction can become Related Brain Pathology:
pathological and neurotoxic; in n­ eurodegenerative Beyond Alzheimer’s Disease
diseases such as AD, SG promote the accumula-
tion of Tau aggregates [139–142]. In fact, SG are Stress pervades all our lives and most of us will
suggested to accelerate Tau aggregation in a respond to daily life stressors in an adaptive man-
vicious cycle wherein Tau stimulates SG for- ner. However, as noted by Selye as early as 1936,
mation, with the RNA binding protein TIA1 mounting a transient and adaptive response may
playing a lead role in Tau misfolding and aggre- not be possible in all circumstances and the
gation [143]. Notably, while hyperphosphoryla- stressful experience may become chronic and
tion and aggregation-prone mutations of Tau can maladaptive. The negative impact of chronic
enhance SG formation, they are not essential for stress (and the associated rise in circulating GC
this event [143]. levels) on brain structure and function (e.g. cog-
We recently showed that chronic stress and nition, mood, emotion) is now well recognized.
high GC upregulate various RBP and SG markers In addition to the role of chronic stress/GC in the
in soluble and insoluble fractions in the hippo- development of AD pathology, chronic stress is
campus of P301L-Tau Tg animals and primary causally related to major depression which, as in
neuron cultures. Specifically, tissues from ani- AD, may reflect defects in neuroplastic mecha-
mals exposed to chronic stress displayed nisms [1, 5, 146]. As briefly mentioned above,
increased cytoplasmic (soluble and insoluble) major depressive disorder appears to predispose
levels of several RBPs and SG-associated mark- to AD [147]. The body of evidence supporting
ers (e.g. TIA-1, PABP, G3BP, FUS, DDX5) that the latter clinical observation includes findings of
contributed to the formation of insoluble Tau potentially common neurobiological mecha-
inclusions and Tau accumulation. As noted nisms in the two disorders [84, 85, 148, 149].
above, TIA-1 plays a prominent role in Tau Given this, it is interesting that epidemiological
aggregation: under stressful conditions, TIA-1 is studies implicate depression as a risk factor for
trafficked the nucleus to the cytospasm where it the development of AD [5], with support for this
interacts directly with Tau (and other RBP such coming from the observation that previously
as PABP and EWSR1) to stimulate its aggrega- depressed subjects have increased amyloid
tion and accumulation [143–145]. plaque and neurofibrillary tangles (NFT) loads
In other recent work, we showed that Tau mis- [150]. Indeed, since clinicians are sometimes
sorting and accumulation in the dendritic com- faced with the challenge of distinguishing
partment, such as is found in AD pathology between patients suffering from depression and
[102], is also triggered by chronic stress/GC AD, several authors have attempted to develop
exposure [59, 110]. This is interesting because assays based on the detection of various APP
Tau missorting is hypothesized to facilitate for- cleavage products that might help such distinc-
mation of SG as part of the translational stress tion [151–154], albeit with little success.
response [143]. While the temporal profile and We previously referred to neurogenesis as a
precise mechanisms underlying stress/GC-evoked phenomenon that contributes neuroplasticity,
dysregulation of RBPs and the associated SG with impaired neurogenesis being implicated in
cascade remain to be elucidated, Fig. 20.2 illus- the pathogenesis of depression [155] as well as
250 I. Sotiropoulos et al.

AD [156]. Neurogenesis declines with age (also References

in humans [156]) and is disrupted by stress and
high GC levels [62]. In light of the previously-­ 1. Sousa N, Almeida OFX.  Disconnection and recon-
nection: the morphological basis of (mal)adaptation
referred interactions between stress/GC and Tau, to stress. Trends Neurosci. 2012;35:742–51. https://
it is therefore interesting that our recent research doi.org/10.1016/j.tins.2012.08.006.
suggests that Tau plays an essential role in stress-­ 2. Carroll BJ.  Ageing, stress and the brain. Novartis
driven suppression of birth of neurons (but not Found Symp. 2002;242:26–36.. discussion 36–45
3. McEwen BS.  Mood disorders and allostatic load.
astrocytes) in the adult dentate gyrus (DG, a hip- Biol Psychiatry. 2003;54:200–7.
pocampal subfield) [62]; specifically, chronic 4. McEwen BS. Physiology and neurobiology of stress
stress is unable to impair the propliferation of and adaptation: central role of the brain. Physiol
neuroblasts and newborn neurons in the DG of Rev. 2007;87:873–904. https://doi.org/10.1152/
physrev.00041.2006.
mice in which the tau gene is deleted. 5. de Kloet ER, Joëls M, Holsboer F.  Stress and the
Interestingly, tau ablation does not interfere with brain: from adaptation to disease. Nat Rev Neurosci.
stress-induced suppression of astrocyte prolifera- 2005;6:463–75. https://doi.org/10.1038/nrn1683.
tion. This finding is likely related to the differen- 6. Sapolsky RM.  Glucocorticoids and hippocampal
atrophy in neuropsychiatric disorders. Arch Gen
tial expression of Tau in neuronal vs. astrocytic Psychiatry. 2000;57:925–35.
precursor cells – Tau is expressed in neurons at 7. Sorrells SF, Sapolsky RM. An inflammatory review
much higher levels than in astrocytes [157]. of glucocorticoid actions in the CNS.  Brain Behav
These observations suggest a novel mechanism Immun. 2007;21:259–72. https://doi.org/10.1016/j.
bbi.2006.11.006.
through which stress can remodel the adult brain. 8. Sotiropoulos I, Cerqueira JJ, Catania C, Takashima
Interestingly, our investigations also showed that A, Sousa N, Almeida OFX.  Stress and glucocorti-
chronic stress increases the 4R-Tau:3R-Tau iso- coid footprints in the brain—the path from depres-
form ratio in the DG.  Given that 3R-Tau has a sion to Alzheimer’s disease. Neurosci Biobehav
Rev. 2008;32:1161–73. https://doi.org/10.1016/j.
lower affinity for MT than 4R-tau, neuroblasts neubiorev.2008.05.007.
may be endowed with greater cytoskeletal plas- 9. Vyas S, Rodrigues AJ, Silva JM, Tronche F, Almeida
ticity than mature neurons since 3R-Tau is more OFX, Sousa N, Sotiropoulos I.  Chronic stress and
abundant in the former. Here, it is also relevant to glucocorticoids: from neuronal plasticity to neuro-
degeneration. Neural Plast. 2016;2016:1–15. https://
note that it was recently shown that higher levels doi.org/10.1155/2016/6391686.
of 4R-Tau are associated with increased Tau 10. Reul JM, de Kloet ER.  Two receptor systems
phosphorylation and brain pathology [158]. for corticosterone in rat brain: microdistribu-
Moreover, an increased 4R/3R-Tau ratio is asso- tion and differential occupation. Endocrinology.
1985;117:2505–11. https://doi.org/10.1210/
ciated with cytoskeletal disturbances and tauopa- endo-117-6-2505.
thies such as AD [64]. 11. de Kloet ER, Reul JM, Sutanto W.  Corticosteroids
and the brain. J  Steroid Biochem Mol Biol.
1990;37:387–94.
12. Hassan AH, von Rosenstiel P, Patchev VK, Holsboer
Summary/Conclusions F, Almeida OF.  Exacerbation of apoptosis in the
dentate gyrus of the aged rat by dexamethasone
The evidence reviewed in this chapter suggest and the protective role of corticosterone. Exp
that deficits in Tau function and Tau proteostasis Neurol. 1996;140:43–52. https://doi.org/10.1006/
exnr.1996.0113.
may be critical and cooperative mechanisms 13. Almeida OF, Condé GL, Crochemore C, Demeneix
through which stress (whose actions are executed BA, Fischer D, Hassan AH, Meyer M, Holsboer F,
by GC) remodels neural circuits (cell birth and Michaelidis TM.  Subtle shifts in the ratio between
death, dendritic and synaptic atrophy/connectiv- pro- and antiapoptotic molecules after activation
of corticosteroid receptors decide neuronal fate.
ity), thus inducing impairments mood and cogni- FASEB J. 2000;14(5):779–90.
tion. Importantly, we suggest that Tau lies at the 14. ter Horst JP, van der Mark MH, Arp M, Berger S, de
core of a set of common neurobiological mecha- Kloet ER, Oitzl MS. Stress or no stress: mineralocor-
nisms that link stress with AD and other brain ticoid receptors in the forebrain regulate behavioral
adaptation. Neurobiol Learn Mem. 2012;98(1):33–
pathologies such as depression. 40. https://doi.org/10.1016/j.nlm.2012.04.006.
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 251

15. van Eekelen JA, Rots NY, Sutanto W, de Kloet Pharmacol. 2004;65:761–9. https://doi.org/10.1124/
ER. The effect of aging on stress responsiveness and mol.65.3.761.
central corticosteroid receptors in the brown Norway 27. Ballinger CA, Connell P, Wu Y, Hu Z, Thompson
rat. Neurobiol Aging. 1992;13:159–70. LJ, Yin L-Y, Patterson C.  Identification of CHIP,
16. Hassan AH, Patchev VK, von Rosenstiel P, Holsboer a novel tetratricopeptide repeat-containing protein
F, Almeida OF.  Plasticity of hippocampal cortico- that interacts with heat shock proteins and nega-
steroid receptors during aging in the rat. FASEB tively regulates chaperone functions. Mol Cell
J. 1999;13:115–22. Biol. 1999;19:4535–45. https://doi.org/10.1128/
17. Riedemann T, Patchev AV, Cho K, Almeida MCB.19.6.4535.
OFX.  Corticosteroids: way upstream. Mol Brain. 28. Tirard M, Jasbinsek J, Almeida OFX, Michaelidis
2010;3:2. https://doi.org/10.1186/1756-6606-3-2. TM. The manifold actions of the protein inhibitor of
18. Groeneweg FL, Karst H, de Kloet ER, Joëls activated STAT proteins on the transcriptional activ-
M.  Mineralocorticoid and glucocorticoid receptors ity of mineralocorticoid and glucocorticoid receptors
at the neuronal membrane, regulators of nongenomic in neural cells. J Mol Endocrinol. 2004;32:825–41.
corticosteroid signalling. Mol Cell Endocrinol. 29. Tirard M, Almeida OFX, Hutzler P, Melchior F,
2012;350:299–309. https://doi.org/10.1016/j. Michaelidis TM.  Sumoylation and proteasomal
mce.2011.06.020. activity determine the transactivation properties of
19. Yang S, Roselli F, Patchev AV, Yu S, Almeida the mineralocorticoid receptor. Mol Cell Endocrinol.
OFX.  Non-receptor-tyrosine kinases integrate fast 2007;268:20–9. https://doi.org/10.1016/j.
glucocorticoid signaling in hippocampal neurons. mce.2007.01.010.
J  Biol Chem. 2013;288:23725–39. https://doi. 30. Datson NA, Van Der Perk J, De Kloet ER, Vreugdenhil
org/10.1074/jbc.M113.470146. E.  Identification of corticosteroid-responsive genes
20. Trapp T, Rupprecht R, Castrén M, Reul JMHM, in rat hippocampus using serial analysis of gene
Holsboer F.  Heterodimerization between min- expression: corticosteroid-responsive genes in hip-
eralocorticoid and glucocorticoid receptor: a pocampus. Eur J Neurosci. 2001;14:675–89. https://
new principle of glucocorticoid action in the doi.org/10.1046/j.0953-816x.2001.01685.x.
CNS.  Neuron. 1994;13:1457–62. https://doi. 31. Schaaf MJM, de Jong J, de Kloet ER, Vreugdenhil
org/10.1016/0896-6273(94)90431-6. E.  Downregulation of BDNF mRNA and protein
21. Nishi M, Ogawa H, Ito T, Matsuda K-I, Kawata in the rat hippocampus by corticosterone. Brain
M.  Dynamic changes in subcellular localization of Res. 1998;813:112–20. https://doi.org/10.1016/
mineralocorticoid receptor in living cells: in com- S0006-8993(98)01010-5.
parison with glucocorticoid receptor using dual-­ 32. Hansson AC, Cintra A, Belluardo N, Sommer W,
color labeling with green fluorescent protein spectral Bhatnagar M, Bader M, Ganten D, Fuxe K. Gluco-
variants. Mol Endocrinol. 2001;15:1077–92. https:// and mineralocorticoid receptor-mediated regula-
doi.org/10.1210/mend.15.7.0659. tion of neurotrophic factor gene expression in
22. Gesing A, Bilang-Bleuel A, Droste SK, Linthorst the dorsal hippocampus and the neocortex of
AC, Holsboer F, Reul JM.  Psychological stress the rat: GR- and MR-mediated gene expression.
increases hippocampal mineralocorticoid receptor Eur J  Neurosci. 2000;12:2918–34. https://doi.
levels: involvement of corticotropin-releasing hor- org/10.1046/j.1460-9568.2000.00185.x.
mone. J Neurosci. 2001;21:4822–9. 33. Sandi C. Stress, cognitive impairment and cell adhe-
23. Chen Y, Rex CS, Rice CJ, Dube CM, Gall CM, sion molecules. Nat Rev Neurosci. 2004;5:917–30.
Lynch G, Baram TZ. Correlated memory defects and https://doi.org/10.1038/nrn1555.
hippocampal dendritic spine loss after acute stress 34. Sabban EL, Liu X, Serova L, Gueorguiev V,
involve corticotropin-releasing hormone signaling. Kvetnansky R.  Stress triggered changes in gene
Proc Natl Acad Sci. 2010;107:13123–8. https://doi. expression in adrenal medulla: transcriptional
org/10.1073/pnas.1003825107. responses to acute and chronic stress. Cell Mol
24. Rein T.  FK506 binding protein 51 integrates path- Neurobiol. 2006;26:843–54. https://doi.org/10.1007/
ways of adaptation: FKBP51 shapes the reactivity s10571-006-9069-1.
to environmental change. Bioessays. 2016;38:894– 35. Lucassen PJ, Pruessner J, Sousa N, Almeida OFX,
902. https://doi.org/10.1002/bies.201600050. Van Dam AM, Rajkowska G, Swaab DF, Czéh
25. Grad I, Picard D.  The glucocorticoid responses B.  Neuropathology of stress. Acta Neuropathol.
are shaped by molecular chaperones. Mol 2014;127:109–35. https://doi.org/10.1007/
Cell Endocrinol. 2007;275:2–12. https://doi. s00401-013-1223-5.
org/10.1016/j.mce.2007.05.018. 36. Harris A, Seckl J.  Glucocorticoids, prenatal
26. Obradović D, Tirard M, Némethy Z, Hirsch O, stress and the programming of disease. Horm
Gronemeyer H, Almeida OFX.  DAXX, FLASH, Behav. 2011;59:279–89. https://doi.org/10.1016/j.
and FAF-1 modulate mineralocorticoid and gluco- yhbeh.2010.06.007.
corticoid receptor-mediated transcription in hip- 37. Seckl JR.  Prenatal glucocorticoids and long-term
pocampal cells--toward a basis for the opposite programming. Eur J  Endocrinol. 2004;151(Suppl
actions elicited by two nuclear receptors? Mol 3):U49–62.
252 I. Sotiropoulos et al.

38. Moisiadis VG, Matthews SG.  Glucocorticoids Cereb Cortex. 2007;17(9):1998–2006. https://doi.
and fetal programming part 2: mechanisms. Nat org/10.1093/cercor/bhl108.
Rev Endocrinol. 2014;10:403–11. https://doi. 50. Sousa N, Lukoyanov NV, Madeira MD, Almeida
org/10.1038/nrendo.2014.74. OFX, Paula-Barbosa MM.  Reorganization of the
39. Patchev AV, Rodrigues AJ, Sousa N, Spengler morphology of hippocampal neurites and synapses
D, Almeida OFX.  The future is now: early life after stress-induced damage correlates with behav-
events preset adult behaviour. Acta Physiol ioral improvement. Neuroscience. 2000;97:253–66.
(Oxf). 2014;210:46–57. https://doi.org/10.1111/ https://doi.org/10.1016/S0306-4522(00)00050-6.
apha.12140. 51. Woolley CS, Gould E, McEwen BS.  Exposure to
40. McGowan PO, Sasaki A, D’Alessio AC, Dymov excess glucocorticoids alters dendritic morphology
S, Labonté B, Szyf M, Turecki G, Meaney of adult hippocampal pyramidal neurons. Brain Res.
MJ.  Epigenetic regulation of the glucocorticoid 1990;531:225–31.
receptor in human brain associates with childhood 52. Liston C, Miller MM, Goldwater DS, Radley
abuse. Nat Neurosci. 2009;12:342–8. https://doi. JJ, Rocher AB, Hof PR, Morrison JH, McEwen
org/10.1038/nn.2270. BS. Stress-induced alterations in prefrontal cortical
41. Oberlander TF, Weinberg J, Papsdorf M, Grunau R, dendritic morphology predict selective impairments
Misri S, Devlin AM. Prenatal exposure to maternal in perceptual attentional set-shifting. J  Neurosci.
depression, neonatal methylation of human gluco- 2006;26(30):7870–4. https://doi.org/10.1523/
corticoid receptor gene (NR3C1) and infant cortisol JNEUROSCI.1184-06.2006.
stress responses. Epigenetics. 2008;3:97–106. 53. Schwabe L, Wolf OT. Stress prompts habit behavior
42. Diamond DM, Bennett MC, Fleshner M, Rose in humans. J Neurosci. 2009;29:7191–8. https://doi.
GM.  Inverted-U relationship between the level of org/10.1523/JNEUROSCI.0979-09.2009.
peripheral corticosterone and the magnitude of hip- 54. Sandi C, Davies HA, Cordero MI, Rodriguez JJ,
pocampal primed burst potentiation. Hippocampus. Popov VI, Stewart MG.  Rapid reversal of stress
1992;2:421–30. https://doi.org/10.1002/ induced loss of synapses in CA3 of rat hippocam-
hipo.450020409. pus following water maze training. Eur J Neurosci.
43. Martin SJ, Grimwood PD, Morris RGM.  Synaptic 2003;17:2447–56.
plasticity and memory: an evaluation of the hypoth- 55. Tasker JG, Herman JP. Mechanisms of rapid gluco-
esis. Annu Rev Neurosci. 2000;23:649–711. https:// corticoid feedback inhibition of the hypothalamic–
doi.org/10.1146/annurev.neuro.23.1.649. pituitary–adrenal axis. Stress. 2011;14:398–406.
44. Kim JJ, Diamond DM.  The stressed hippocampus, https://doi.org/10.3109/10253890.2011.586446.
synaptic plasticity and lost memories. Nat Rev 56. Goldwater DS, Pavlides C, Hunter RG, Bloss EB,
Neurosci. 2002;3:453–62. https://doi.org/10.1038/ Hof PR, McEwen BS, Morrison JH.  Structural
nrn849. and functional alterations to rat medial prefrontal
45. Kim JJ, Foy MR, Thompson RF.  Behavioral stress cortex following chronic restraint stress and recov-
modifies hippocampal plasticity through N-methyl-­ ery. Neuroscience. 2009;164:798–808. https://doi.
D-aspartate receptor activation. Proc Natl Acad Sci org/10.1016/j.neuroscience.2009.08.053.
U S A. 1996;93:4750–3. 57. Radley JJ, Morrison JH.  Repeated stress and
46. Oitzl MS, de Kloet ER.  Selective corticosteroid structural plasticity in the brain. Ageing Res
antagonists modulate specific aspects of spatial ori- Rev. 2005;4:271–87. https://doi.org/10.1016/j.
entation learning. Behav Neurosci. 1992;106:62–71. arr.2005.03.004.
47. Sandi C, Woodson JC, Haynes VF, Park CR, Touyarot 58. Shansky RM, Morrison JH.  Stress-induced den-
K, Lopez-Fernandez MA, Venero C, Diamond dritic remodeling in the medial prefrontal cor-
DM.  Acute stress-induced impairment of spatial tex: effects of circuit, hormones and rest. Brain
memory is associated with decreased expression of Res. 2009;1293:108–13. https://doi.org/10.1016/j.
neural cell adhesion molecule in the hippocampus brainres.2009.03.062.
and prefrontal cortex. Biol Psychiatry. 2005;57:856– 59. Lopes S, Vaz-Silva J, Pinto V, Dalla C, Kokras N,
64. https://doi.org/10.1016/j.biopsych.2004.12.034. Bedenk B, Mack N, Czisch M, Almeida OFX,
48. Cacucci F, Wills TJ, Lever C, Giese KP, O’Keefe Sousa N, Sotiropoulos I.  Tau protein is essential
J. Experience-dependent increase in CA1 place cell for stress-induced brain pathology. Proc Natl Acad
spatial information, but not spatial reproducibil- Sci. 2016;113:E3755–63. https://doi.org/10.1073/
ity, is dependent on the autophosphorylation of the pnas.1600953113.
alpha-isoform of the calcium/calmodulin-dependent 60. Lopes S, Teplytska L, Vaz-Silva J, Dioli C, Trindade
protein kinase II.  J Neurosci. 2007;27(29):7854–9. R, Morais M, Webhofer C, Maccarrone G, Almeida
https://doi.org/10.1523/JNEUROSCI.1704-07.2007. OFX, Turck CW, Sousa N, Sotiropoulos I, Filiou
49. Cerqueira JJ, Taipa R, Uylings HBM, Almeida OFX, MD.  Tau deletion prevents stress-induced den-
Sousa N. Specific configuration of dendritic degen- dritic atrophy in prefrontal cortex: role of synaptic
eration in pyramidal neurons of the medial prefrontal mitochondria. Cereb Cortex. 2016;27(4):2580–91.
cortex induced by differing corticosteroid regimens. https://doi.org/10.1093/cercor/bhw057.
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 253

61. Heine VM, Maslam S, Zareno J, Joëls M, Lucassen 74. Hartmann A, Veldhuis JD, Deuschle M, Standhardt
PJ.  Suppressed proliferation and apoptotic changes H, Heuser I. Twenty-four hour cortisol release pro-
in the rat dentate gyrus after acute and chronic stress files in patients with Alzheimer’s and Parkinson’s
are reversible. Eur J Neurosci. 2004;19:131–44. disease compared to normal controls: ultradian
62. Dioli C, Patrício P, Trindade R, Pinto LG, Silva secretory pulsatility and diurnal variation. Neurobiol
JM, Morais M, Ferreiro E, Borges S, Mateus-­ Aging. 1997;18:285–9.
Pinheiro A, Rodrigues AJ, Sousa N, Bessa JM, 75. Touma C, Ambrée O, Görtz N, Keyvani K,
Pinto L, Sotiropoulos I. Tau-dependent suppression Lewejohann L, Palme R, Paulus W, Schwarze-Eicker
of adult neurogenesis in the stressed hippocam- K, Sachser N.  Age- and sex-dependent develop-
pus. Mol Psychiatry. 2017;22:1110–8. https://doi. ment of adrenocortical hyperactivity in a transgenic
org/10.1038/mp.2017.103. mouse model of Alzheimer’s disease. Neurobiol
63. Kempermann G, Krebs J, Fabel K. The contribution Aging. 2004;25:893–904. https://doi.org/10.1016/j.
of failing adult hippocampal neurogenesis to psychi- neurobiolaging.2003.09.004.
atric disorders. Curr Opin Psychiatry. 2008;21:290– 76. Green KN.  Glucocorticoids increase amyloid-beta
5. https://doi.org/10.1097/YCO.0b013e3282fad375. and tau pathology in a mouse model of Alzheimer’s
64. Mu Y, Gage FH. Adult hippocampal neurogenesis and disease. J  Neurosci. 2006;26:9047–56. https://doi.
its role in Alzheimer’s disease. Mol Neurodegener. org/10.1523/JNEUROSCI.2797-06.2006.
2011;6:85. https://doi.org/10.1186/1750-1326-6-85. 77. Lupien SJ, de Leon M, de Santi S, Convit A, Tarshish
65. Demars M, Hu Y-S, Gadadhar A, Lazarov C, Nair NPV, Thakur M, McEwen BS, Hauger
O. Impaired neurogenesis is an early event in the eti- RL, Meaney MJ.  Cortisol levels during human
ology of familial Alzheimer’s disease in transgenic aging predict hippocampal atrophy and memory
mice. J Neurosci Res. 2010;88:2103–17. https://doi. deficits. Nat Neurosci. 1998;1:69–73. https://doi.
org/10.1002/jnr.22387. org/10.1038/271.
66. O’Brien JT, Ames D, Schweitzer I, Mastwyk M, 78. Huang C-W, Lui C-C, Chang W-N, Lu C-H, Wang
Colman P.  Enhanced adrenal sensitivity to adreno- Y-L, Chang C-C.  Elevated basal cortisol level pre-
corticotrophic hormone (ACTH) is evidence of HPA dicts lower hippocampal volume and cognitive
axis hyperactivity in Alzheimer’s disease. Psychol decline in Alzheimer’s disease. J  Clin Neurosci.
Med. 1996;26:7–14. 2009;16:1283–6. https://doi.org/10.1016/j.
67. Rasmuson S, Andrew R, Näsman B, Seckl JR, jocn.2008.12.026.
Walker BR, Olsson T.  Increased glucocorticoid 79. Karran E, Mercken M, De Strooper B.  The amy-
production and altered cortisol metabolism in loid cascade hypothesis for Alzheimer’s disease: an
women with mild to moderate Alzheimer’s dis- appraisal for the development of therapeutics. Nat
ease. Biol Psychiatry. 2001;49:547–52. https://doi. Rev Drug Discov. 2011;10:698–712. https://doi.
org/10.1016/S0006-3223(00)01015-5. org/10.1038/nrd3505.
68. Simard M, Hudon C, van Reekum R. Psychological 80. Ballard C, Gauthier S, Corbett A, Brayne C,
distress and risk for dementia. Curr Psychiatry Rep. Aarsland D, Jones E.  Alzheimer’s disease. Lancet.
2009;11:41–7. 2011;377:1019–31. https://doi.org/10.1016/
69. Mejía S, Giraldo M, Pineda D, Ardila A, Lopera S0140-6736(10)61349-9.
F. Nongenetic factors as modifiers of the age of onset 81. Duyckaerts C, Delatour B, Potier M-C. Classification
of familial Alzheimer’s disease. Int Psychogeriatr. and basic pathology of Alzheimer disease. Acta
2003;15:337–49. Neuropathol (Berl). 2009;118:5–36. https://doi.
70. Rothman SM, Mattson MP.  Adverse stress, hip- org/10.1007/s00401-009-0532-1.
pocampal networks, and Alzheimer’s disease. 82. Dede AJO, Wixted JT, Hopkins RO, Squire
Neuromolecular Med. 2010;12:56–70. https://doi. LR.  Hippocampal damage impairs recognition
org/10.1007/s12017-009-8107-9. memory broadly, affecting both parameters in two
71. Hatzinger M, Z’Brun A, Hemmeter U, Seifritz prominent models of memory. Proc Natl Acad
E, Baumann F, Holsboer-Trachsler E, Heuser Sci. 2013;110:6577–82. https://doi.org/10.1073/
IJ.  Hypothalamic-pituitary-adrenal system func- pnas.1304739110.
tion in patients with Alzheimer’s disease. Neurobiol 83. Jeong YH, Park CH, Yoo J, Shin KY, Ahn S-M, Kim
Aging. 1995;16:205–9. H-S, Lee SH, Emson PC, Suh Y-H.  Chronic stress
72. Peskind ER, Wilkinson CW, Petrie EC, Schellenberg accelerates learning and memory impairments and
GD, Raskind MA.  Increased CSF cortisol in increases amyloid deposition in APP V717I-CT100
AD is a function of APOE genotype. Neurology. transgenic mice, an Alzheimer’s disease model.
2001;56:1094–8. FASEB J. 2006;20:729–31. https://doi.org/10.1096/
73. Hoogendijk WJG, Meynen G, Endert E, Hofman fj.05-4265fje.
MA, Swaab DF. Increased cerebrospinal fluid cortisol 84. Sotiropoulos I, Catania C, Riedemann T, Fry JP, Breen
level in Alzheimer’s disease is not related to depres- KC, Michaelidis TM, Almeida OFX. Glucocorticoids
sion. Neurobiol Aging. 2006;27:780.e1–2. https:// trigger Alzheimer disease-like ­ pathobiochemistry
doi.org/10.1016/j.neurobiolaging.2005.07.017. in rat neuronal cells expressing human tau: glu-
cocorticoids, human tau and Alzheimer’s dis-
254 I. Sotiropoulos et al.

ease. J  Neurochem. 2008;107:385–97. https://doi. 95. Rissman RA, Lee K-F, Vale W, Sawchenko
org/10.1111/j.1471-4159.2008.05613.x. PE.  Corticotropin-releasing factor receptors dif-
85. Catania C, Sotiropoulos I, Silva R, Onofri C, ferentially regulate stress-induced tau phosphory-
Breen KC, Sousa N, Almeida OFX.  The amy- lation. J  Neurosci. 2007;27:6552–62. https://doi.
loidogenic potential and behavioral correlates of org/10.1523/JNEUROSCI.5173-06.2007.
stress. Mol Psychiatry. 2009;14:95–105. https://doi. 96. Sotiropoulos I, Silva J, Kimura T, Rodrigues
org/10.1038/sj.mp.4002101. AJ, Costa P, Almeida OFX, Sousa N, Takashima
86. Stein-Behrens BA, Elliott EM, Miller CA, Schilling A.  Female hippocampus vulnerability to environ-
JW, Newcombe R, Sapolsky RM.  Glucocorticoids mental stress, a precipitating factor in Tau aggrega-
exacerbate kainic acid-induced extracellular accu- tion pathology. J Alzheimers Dis. 2015;43:763–74.
mulation of excitatory amino acids in the rat hippo- https://doi.org/10.3233/JAD-140693.
campus. J Neurochem. 1992;58:1730–5. 97. de Calignon A, Polydoro M, Suárez-Calvet M,
87. Sotiropoulos I, Catania C, Pinto LG, Silva R, William C, Adamowicz DH, Kopeikina KJ, Pitstick
Pollerberg GE, Takashima A, Sousa N, Almeida R, Sahara N, Ashe KH, Carlson GA, Spires-Jones TL,
OFX.  Stress acts cumulatively to precipitate Hyman BT. Propagation of tau pathology in a model
Alzheimer’s disease-like tau pathology and cogni- of early Alzheimer’s disease. Neuron. 2012;73:685–
tive deficits. J  Neurosci. 2011;31:7840–7. https:// 97. https://doi.org/10.1016/j.neuron.2011.11.033.
doi.org/10.1523/JNEUROSCI.0730-11.2011. 98. Wang YP, Biernat J, Pickhardt M, Mandelkow E,
88. Callahan LM, Vaules WA, Coleman PD. Progressive Mandelkow E-M. Stepwise proteolysis liberates tau
reduction of synaptophysin message in single neu- fragments that nucleate the Alzheimer-like aggre-
rons in Alzheimer disease. J  Neuropathol Exp gation of full-length tau in a neuronal cell model.
Neurol. 2002;61:384–95. Proc Natl Acad Sci. 2007;104:10252–7. https://doi.
89. Lauckner J, Frey P, Geula C. Comparative distribu- org/10.1073/pnas.0703676104.
tion of tau phosphorylated at Ser262 in pre-tangles 99. Rissman RA, Poon WW, Blurton-Jones M, Oddo S,
and tangles. Neurobiol Aging. 2003;24:767–76. Torp R, Vitek MP, LaFerla FM, Rohn TT, Cotman
90. Hampel H, Bürger K, Pruessner JC, Zinkowski R, CW.  Caspase-cleavage of tau is an early event
DeBernardis J, Kerkman D, Leinsinger G, Evans AC, in Alzheimer disease tangle pathology. J  Clin
Davies P, Möller H-J, Teipel SJ. Correlation of cere- Invest. 2004;114:121–30. https://doi.org/10.1172/
brospinal fluid levels of tau protein phosphorylated JCI20640.
at threonine 231 with rates of hippocampal atrophy 100. Weaver CL, Espinoza M, Kress Y, Davies
in Alzheimer disease. Arch Neurol. 2005;62(5):770– P.  Conformational change as one of the earliest
3. https://doi.org/10.1001/archneur.62.5.770. alterations of tau in Alzheimer’s disease. Neurobiol
91. Augustinack JC, Schneider A, Mandelkow E-M, Aging. 2000;21:719–27.
Hyman BT.  Specific tau phosphorylation sites cor- 101. Kobayashi S, Tanaka T, Soeda Y, Almeida OFX,
relate with severity of neuronal cytopathology in Takashima A.  Local somatodendritic translation
Alzheimer’s disease. Acta Neuropathol (Berl). and hyperphosphorylation of tau protein trig-
2002;103:26–35. gered by AMPA and NMDA receptor stimula-
92. Ewers M, Buerger K, Teipel SJ, Scheltens P, Schroder tion. EBioMedicine. 2017;20:120–6. https://doi.
J, Zinkowski RP, Bouwman FH, Schonknecht P, org/10.1016/j.ebiom.2017.05.012.
Schoonenboom NSM, Andreasen N, Wallin A, 102. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A,
DeBernardis JF, Kerkman DJ, Heindl B, Blennow van Eersel J, Wölfing H, Chieng BC, Christie MJ,
K, Hampel H.  Multicenter assessment of CSF-­ Napier IA, Eckert A, Staufenbiel M, Hardeman E,
phosphorylated tau for the prediction of conversion Götz J. Dendritic function of tau mediates amyloid-β
of MCI.  Neurology. 2007;69:2205–12. https://doi. toxicity in Alzheimer’s disease mouse models.
org/10.1212/01.wnl.0000286944.22262.ff. Cell. 2010;142:387–97. https://doi.org/10.1016/j.
93. van der Vlies AE, Verwey NA, Bouwman FH, cell.2010.06.036.
Blankenstein MA, Klein M, Scheltens P, van der 103. Hoover BR, Reed MN, Su J, Penrod RD, Kotilinek
Flier WM.  CSF biomarkers in relationship to cog- LA, Grant MK, Pitstick R, Carlson GA, Lanier
nitive profiles in Alzheimer disease. Neurology. LM, Yuan L-L, Ashe KH, Liao D. Tau mislocaliza-
2009;72:1056–61. https://doi.org/10.1212/01. tion to dendritic spines mediates synaptic dysfunc-
wnl.0000345014.48839.71. tion independently of neurodegeneration. Neuron.
94. Carroll JC, Iba M, Bangasser DA, Valentino RJ, 2010;68:1067–81. https://doi.org/10.1016/j.
James MJ, Brunden KR, Lee VM-Y, Trojanowski neuron.2010.11.030.
JQ.  Chronic stress exacerbates tau pathology, neu- 104. McKinney RA.  Excitatory amino acid involve-
rodegeneration, and cognitive performance through ment in dendritic spine formation, maintenance
a corticotropin-releasing factor receptor-dependent and remodelling: glutamate and dendritic spines.
mechanism in a transgenic mouse model of tauopa- J Physiol. 2010;588:107–16. https://doi.org/10.1113/
thy. J  Neurosci. 2011;31(40):14436–49. https://doi. jphysiol.2009.178905.
org/10.1523/JNEUROSCI.3836-11.2011. 105. Mondragon-Rodriguez S, Trillaud-Doppia E,
Dudilot A, Bourgeois C, Lauzon M, Leclerc N,
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 255

Boehm J.  Interaction of endogenous tau protein 117. Wang Y, Martinez-Vicente M, Krüger U, Kaushik S,
with synaptic proteins is regulated by N-methyl-­ Wong E, Mandelkow E-M, Cuervo AM, Mandelkow
D-aspartate receptor-dependent tau phosphoryla- E. Tau fragmentation, aggregation and clearance: the
tion. J Biol Chem. 2012;287:32040–53. https://doi. dual role of lysosomal processing. Hum Mol Genet.
org/10.1074/jbc.M112.401240. 2009;18:4153–70. https://doi.org/10.1093/hmg/
106. Miller EC, Teravskis PJ, Dummer BW, Zhao X, ddp367.
Huganir RL, Liao D.  Tau phosphorylation and 118. Inoue K, Rispoli J, Kaphzan H, Klann E, Chen EI, Kim
tau mislocalization mediate soluble Aβ oligomer-­ J, Komatsu M, Abeliovich A. Macroautophagy defi-
induced AMPA glutamate receptor signaling defi- ciency mediates age-dependent neurodegeneration
cits. Eur J  Neurosci. 2014;39:1214–24. https://doi. through a phospho-tau pathway. Mol Neurodegener.
org/10.1111/ejn.12507. 2012;7:48. https://doi.org/10.1186/1750-1326-7-48.
107. Nong Y, Huang Y-Q, Ju W, Kalia LV, Ahmadian G, 119. Brown MR, Bondada V, Keller JN, Thorpe J, Geddes
Wang YT, Salter MW. Glycine binding primes NMDA JW. Proteasome or calpain inhibition does not alter
receptor internalization. Nature. 2003;422:302–7. cellular tau levels in neuroblastoma cells or primary
https://doi.org/10.1038/nature01497. neurons. J Alzheimers Dis. 2005;7:15–24.
108. Yang C-H. Behavioral stress enhances hippocampal 120. Feuillette S, Blard O, Lecourtois M, Frébourg T,
CA1 long-term depression through the blockade of Campion D, Dumanchin C.  Tau is not normally
the glutamate uptake. J Neurosci. 2005;25:4288–93. degraded by the proteasome: tau not normally
https://doi.org/10.1523/JNEUROSCI.0406-05.2005. degraded by the proteasome. J  Neurosci Res.
109. Magariños AM, McEwen BS, Flügge G, Fuchs 2005;80:400–5. https://doi.org/10.1002/jnr.20414.
E.  Chronic psychosocial stress causes apical den- 121. Hara T, Nakamura K, Matsui M, Yamamoto A,
dritic atrophy of hippocampal CA3 pyramidal Nakahara Y, Suzuki-Migishima R, Yokoyama
neurons in subordinate tree shrews. J  Neurosci. M, Mishima K, Saito I, Okano H, Mizushima
1996;16:3534–40. N.  Suppression of basal autophagy in neural
110. Pinheiro S, Silva J, Mota C, Vaz-Silva J, Veloso A, cells causes neurodegenerative disease in mice.
Pinto V, Sousa N, Cerqueira J, Sotiropoulos I.  Tau Nature. 2006;441:885–9. https://doi.org/10.1038/
mislocation in glucocorticoid-triggered hippocam- nature04724.
pal pathology. Mol Neurobiol. 2016;53:4745–53. 122. Boland B, Kumar A, Lee S, Platt FM, Wegiel J, Yu
https://doi.org/10.1007/s12035-015-9356-2. WH, Nixon RA. Autophagy induction and autopha-
111. Nixon RA.  Autophagy, amyloidogenesis and gosome clearance in neurons: relationship to autoph-
Alzheimer disease. J  Cell Sci. 2007;120:4081–91. agic pathology in Alzheimer’s disease. J  Neurosci.
https://doi.org/10.1242/jcs.019265. 2008;28:6926–37. https://doi.org/10.1523/
112. Banerjee R, Beal MF, Thomas B.  Autophagy in JNEUROSCI.0800-08.2008.
neurodegenerative disorders: pathogenic roles 123. Polman JAE, Hunter RG, Speksnijder N, van den
and therapeutic implications. Trends Neurosci. Oever JME, Korobko OB, McEwen BS, de Kloet
2010;33:541–9. https://doi.org/10.1016/j. ER, Datson NA.  Glucocorticoids modulate the
tins.2010.09.001. mTOR pathway in the hippocampus: differential
113. Komatsu M, Waguri S, Ueno T, Iwata J, Murata S, effects depending on stress history. Endocrinology.
Tanida I, Ezaki J, Mizushima N, Ohsumi Y, Uchiyama 2012;153:4317–27. https://doi.org/10.1210/
Y, Kominami E, Tanaka K, Chiba T. Impairment of en.2012-1255.
starvation-induced and constitutive autophagy in 124. An W-L, Cowburn RF, Li L, Braak H,
Atg7-deficient mice. J Cell Biol. 2005;169:425–34. Alafuzoff I, Iqbal K, Iqbal I-G, Winblad B, Pei
https://doi.org/10.1083/jcb.200412022. J-J.  Up-regulation of phosphorylated/activated
114. Frake RA, Ricketts T, Menzies FM, Rubinsztein p70 S6 kinase and its relationship to neurofibril-
DC.  Autophagy and neurodegeneration. J  Clin lary pathology in Alzheimer’s disease. Am J Pathol.
Invest. 2015;125:65–74. https://doi.org/10.1172/ 2003;163:591–607. https://doi.org/10.1016/
JCI73944. S0002-9440(10)63687-5.
115. Silva JM, Rodrigues S, Sampaio-Marques B, Gomes 125. Pei J-J, Hugon J.  mTOR-dependent sig-
P, Neves-Carvalho A, Dioli C, Soares-Cunha C, nalling in Alzheimer’s disease. J  Cell
Mazuik BF, Takashima A, Ludovico P, Wolozin B, Mol Med. 2008;12:2525–32. https://doi.
Sousa N, Sotiropoulos I. Dysregulation of autophagy org/10.1111/j.1582-4934.2008.00509.x.
and stress granule-related proteins in stress-driven 126. Menzies FM, Huebener J, Renna M, Bonin M, Riess
Tau pathology. Cell Death Differ. 2018;26(8):1411– O, Rubinsztein DC.  Autophagy induction reduces
27. https://doi.org/10.1038/s41418-018-0217-1. mutant ataxin-3 levels and toxicity in a mouse model
116. Hamano T, Gendron TF, Causevic E, Yen S-H, Lin of spinocerebellar ataxia type 3. Brain. 2010;133:93–
W-L, Isidoro C, DeTure M, Ko L.  Autophagic-­ 104. https://doi.org/10.1093/brain/awp292.
lysosomal perturbation enhances tau aggregation 127. Jiang T, Yu J-T, Zhu X-C, Zhang Q-Q, Cao L,
in transfectants with induced wild-type tau expres- Wang H-F, Tan M-S, Gao Q, Qin H, Zhang Y-D,
sion. Eur J Neurosci. 2008;27:1119–30. https://doi. Tan L.  Temsirolimus attenuates tauopathy in  vitro
org/10.1111/j.1460-9568.2008.06084.x. and in  vivo by targeting tau hyperphosphorylation
256 I. Sotiropoulos et al.

and autophagic clearance. Neuropharmacology. Alzheimer-like paired helical filaments. FEBS

2014;85:121–30. https://doi.org/10.1016/j. Lett. 1996;399:344–9. https://doi.org/10.1016/
neuropharm.2014.05.032. S0014-5793(96)01386-5.
128. Rivero-Ríos P, Gómez-Suaga P, Fernández B, 140. Liu-Yesucevitz L, Bilgutay A, Zhang Y-J,
Madero-Pérez J, Schwab AJ, Ebert AD, Hilfiker Vanderwyde T, Citro A, Mehta T, Zaarur N, McKee
S.  Alterations in late endocytic trafficking related A, Bowser R, Sherman M, Petrucelli L, Wolozin
to the pathobiology of LRRK2-linked Parkinson’s B.  Tar DNA binding protein-43 (TDP-43) asso-
disease. Biochem Soc Trans. 2015;43:390–5. https:// ciates with stress granules: analysis of cultured
doi.org/10.1042/BST20140301. cells and pathological brain tissue. PLoS One.
129. Kett LR, Dauer WT. Endolysosomal dysfunction in 2010;5(10):e13250. https://doi.org/10.1371/journal.
Parkinson’s disease: recent developments and future pone.0013250.
challenges: Endolysosomal dysfunction in PD. Mov 141. Liu-Yesucevitz L, Lin AY, Ebata A, Boon JY, Reid
Disord. 2016;31:1433–43. https://doi.org/10.1002/ W, Xu Y-F, Kobrin K, Murphy GJ, Petrucelli L,
mds.26797. Wolozin B. ALS-linked mutations enlarge TDP-43-­
130. Small SA, Simoes-Spassov S, Mayeux R, Petsko enriched neuronal RNA granules in the dendritic
GA. Endosomal traffic jams represent a pathogenic arbor. J  Neurosci. 2014;34:4167–74. https://doi.
hub and therapeutic target in Alzheimer’s disease. org/10.1523/JNEUROSCI.2350-13.2014.
Trends Neurosci. 2017;40:592–602. https://doi. 142.
Vanderweyde T, Yu H, Varnum M, Liu-
org/10.1016/j.tins.2017.08.003. Yesucevitz L, Citro A, Ikezu T, Duff K, Wolozin
131. Vaz-Silva J, Gomes P, Jin Q, Zhu M, Zhuravleva B. Contrasting pathology of the stress granule pro-
V, Quintremil S, Meira T, Silva J, Dioli C, Soares-­ teins TIA-1 and G3BP in tauopathies. J Neurosci.
Cunha C, Daskalakis NP, Sousa N, Sotiropoulos I, 2012;32:8270–83. https://doi.org/10.1523/
Waites CL. Endolysosomal degradation of Tau and JNEUROSCI.1592-12.2012.
its role in glucocorticoid-driven hippocampal mal- 143. Vanderweyde T, Apicco DJ, Youmans-Kidder K,
function. EMBO J.  2018;37:e99084. https://doi. Ash PEA, Cook C, Lummertz da Rocha E, Jansen-­
org/10.15252/embj.201899084. West K, Frame AA, Citro A, Leszyk JD, Ivanov
132. Raiborg C, Stenmark H.  The ESCRT machinery P, Abisambra JF, Steffen M, Li H, Petrucelli L,
in endosomal sorting of ubiquitylated membrane Wolozin B. Interaction of tau with the RNA-binding
proteins. Nature. 2009;458:445–52. https://doi. protein TIA1 regulates tau pathophysiology and
org/10.1038/nature07961. toxicity. Cell Rep. 2016;15:1455–66. https://doi.
133. Sahu R, Kaushik S, Clement CC, Cannizzo ES, org/10.1016/j.celrep.2016.04.045.
Scharf B, Follenzi A, Potolicchio I, Nieves E, Cuervo 144. Apicco DJ, Ash PEA, Maziuk B, LeBlang C,
AM, Santambrogio L. Microautophagy of cytosolic Medalla M, Al Abdullatif A, Ferragud A, Botelho
proteins by late endosomes. Dev Cell. 2011;20:131– E, Ballance HI, Dhawan U, Boudeau S, Cruz AL,
9. https://doi.org/10.1016/j.devcel.2010.12.003. Kashy D, Wong A, Goldberg LR, Yazdani N, Zhang
134. Pooler AM, Hanger DP. Functional implications of C, Ung CY, Tripodis Y, Kanaan NM, Ikezu T, Cottone
the association of tau with the plasma membrane. P, Leszyk J, Li H, Luebke J, Bryant CD, Wolozin
Biochem Soc Trans. 2010;38:1012–5. https://doi. B. Reducing the RNA binding protein TIA1 protects
org/10.1042/BST0381012. against tau-mediated neurodegeneration in vivo. Nat
135. Georgieva ER, Xiao S, Borbat PP, Freed JH, Eliezer Neurosci. 2018;21:72–80. https://doi.org/10.1038/
D.  Tau binds to lipid membrane surfaces via short s41593-017-0022-z.
amphipathic helices located in its microtubule-­ 145. Maziuk B, Ballance HI, Wolozin B.  Dysregulation
binding repeats. Biophys J.  2014;107:1441–52. of RNA binding protein aggregation in neurodegen-
https://doi.org/10.1016/j.bpj.2014.07.046. erative disorders. Front Mol Neurosci. 2017;10:89.
136. Wolozin B.  Regulated protein aggregation: stress https://doi.org/10.3389/fnmol.2017.00089.
granules and neurodegeneration. Mol Neurodegener. 146. McEwen BS, Gray JD, Nasca C. 60 years of neu-
2012;7:56. https://doi.org/10.1186/1750-1326-7-56. roendocrinology: redefining neuroendocrinol-
137. Arimoto K, Fukuda H, Imajoh-Ohmi S, Saito H, ogy: stress, sex and cognitive and emotional
Takekawa M.  Formation of stress granules inhibits regulation. J Endocrinol. 2015;226:T67–83. https://
apoptosis by suppressing stress-responsive MAPK doi.org/10.1530/JOE-15-0121.
pathways. Nat Cell Biol. 2008;10:1324–32. https:// 147. Ownby RL, Crocco E, Acevedo A, John
doi.org/10.1038/ncb1791. V, Loewenstein D.  Depression and risk for
138. Arimoto-Matsuzaki K, Saito H, Takekawa M. TIA1 Alzheimer disease: systematic review, meta-­
oxidation inhibits stress granule assembly and analysis, and metaregression analysis. Arch Gen
sensitizes cells to stress-induced apoptosis. Nat Psychiatry. 2006;63:530. https://doi.org/10.1001/
Commun. 2016;7:10252. https://doi.org/10.1038/ archpsyc.63.5.530.
ncomms10252. 148. Heun R, Kockler M, Ptok U.  Depression in
139. Kampers T, Friedhoff P, Biernat J, Mandelkow Alzheimer’s disease: is there a temporal relation-
E-M, Mandelkow E.  RNA stimulates aggrega- ship between the onset of depression and the onset
tion of microtubule-­ associated protein tau into of dementia? Eur Psychiatry. 2002;17:254–8.
20  Stress and the Etiopathogenesis of Alzheimer’s Disease and Depression 257

149. Kim HK, Nunes PV, Oliveira KC, Young LT, Lafer 154. Sun X, Steffens DC, Au R, Folstein M, Summergrad
B.  Neuropathological relationship between major P, Yee J, Rosenberg I, Mwamburi DM, Qiu
depression and dementia: a hypothetical model WQ.  Amyloid-associated depression: a prodro-
and review. Prog Neuropsychopharmacol Biol mal depression of Alzheimer disease? Arch Gen
Psychiatry. 2016;67:51–7. https://doi.org/10.1016/j. Psychiatry. 2008;65:542. https://doi.org/10.1001/
pnpbp.2016.01.008. archpsyc.65.5.542.
150. Rapp MA, Schnaider-Beeri M, Grossman HT, 155.
Mateus-Pinheiro A, Pinto L, Sousa
Sano M, Perl DP, Purohit DP, Gorman JM, N.  Epigenetic (de)regulation of adult hippo-
Haroutunian V.  Increased hippocampal plaques campal neurogenesis: implications for depres-
and tangles in patients with Alzheimer disease with sion. Clin Epigenetics. 2011;3:5. https://doi.
a lifetime history of major depression. Arch Gen org/10.1186/1868-7083-3-5.
Psychiatry. 2006;63:161–7. https://doi.org/10.1001/ 156. Moreno-Jiménez EP, Flor-García M, Terreros-­
archpsyc.63.2.161. Roncal J, Rábano A, Cafini F, Pallas-Bazarra N,
151. Mayeux R, Honig LS, Tang M-X, Manly J, Stern Ávila J, Llorens-Martín M. Adult hippocampal neu-
Y, Schupf N, Mehta PD.  Plasma A[beta]40 and rogenesis is abundant in neurologically healthy sub-
A[beta]42 and Alzheimer’s disease: relation to age, jects and drops sharply in patients with Alzheimer’s
mortality, and risk. Neurology. 2003;61:1185–90. disease. Nat Med. 2019;25:554–60. https://doi.
152. Andreasen N, Blennow K. CSF biomarkers for mild org/10.1038/s41591-019-0375-9.
cognitive impairment and early Alzheimer’s disease. 157. Gu Y, Oyama F, Ihara Y. Tau is widely expressed in
Clin Neurol Neurosurg. 2005;107:165–73. https:// rat tissues. J Neurochem. 1996;67:1235–44.
doi.org/10.1016/j.clineuro.2004.10.011. 158. Schoch KM, DeVos SL, Miller RL, Chun SJ,
153. Evered L, Scott DA, Silbert B, Maruff Norrbom M, Wozniak DF, Dawson HN, Bennett
P.  Postoperative cognitive dysfunction is inde- CF, Rigo F, Miller TM.  Increased 4R-tau induces
pendent of type of surgery and anesthetic. Anesth pathological changes in a human-tau mouse model.
Analg. 2011;112:1179–85. https://doi.org/10.1213/ Neuron. 2016;90:941–7. https://doi.org/10.1016/j.
ANE.0b013e318215217e. neuron.2016.04.042.
 Tau, Diabetes and Insulin
21
Maud Gratuze, Aurélie Joly-Amado, Luc Buee,
Didier Vieau, and David Blum

Overview on Tau and Tauopathies [11, 133, 280]. Tau is also physiologically

released by neurons [233] even if the natural
Tau protein which was discovered in 1975 [310] function of extracellular Tau remains to be uncov-
became of great interest when it was identified as ered (see other chapters of the present book).
the main component of neurofibrillary tangles Tau protein sequence contains more than 85
(NFT), a pathological feature in the brain of phosphorylated or phosphorylable sites [266].
patients with Alzheimer’s disease (AD) [39, 110, As a result of hyperphosphorylation, conforma-
232]. Tau protein is expressed mainly in the brain tional changes in Tau protein notably impair its
as six isoforms generated by alternative splicing ability to bind to microtubules. Consequently,
[46, 97]. Tau is a microtubule associated proteins accretion of free monomers of misfolded Tau
(MAPs) and plays a role in microtubules assem- leads to accumulation, oligomerization and
bly and stability, as well as diverse cellular pro- aggregation. During the aggregation process,
cesses such as cell morphogenesis, cell division, repeated domains of Tau adopt a beta sheet con-
and intracellular trafficking [49]. Additionally, formation [28] and form filaments [155]. Tau
Tau is involved in much larger neuronal functions aggregates can deposit in NFT a pathological
particularly at the level of synapses and nuclei feature of a group of diseases called Tauopathies
[169, 232, 273], divided into primary Tauopathies
(Pick’s disease, progressive supranuclear palsy,
M. Gratuze
Department of Neurology, Washington University fronto-temporal dementia…) and secondary or
School of Medicine, St. Louis, MO, USA mixed Tauopathies (Alzheimer’s Disease…)
Hope Center for Neurological Disorders, Washington characterized by different clinical features  and
University School of Medicine, St. Louis, MO, USA pathological hallmarks reviewed elsewhere
Knight Alzheimer’s Disease Research Center, [168] (see other chapters of the present book).
Washington University School of Medicine, Tauopathies particularly differ from the cell
St. Louis, MO, USA types exhibiting NFT and Tau isoforms aggrega-
A. Joly-Amado tion [266]. Notably, in AD, NFT are observed
Department of Molecular Pharmacology early in life and increase during aging [37]. The
and Physiology, Byrd Alzheimer’s Institute, spatiotemporal progression of NFT from the
University of South Florida, Tampa, FL, USA
entorhinal cortex and the hippocampus to the
L. Buee · D. Vieau · D. Blum (*) isocortical areas has been shown correlated with
University of Lille, INSERM, CHU-Lille, UMR-S
1172 – JPArc, “Alzheimer & Tauopathies”, Lille, France cognitive deficits [73], supporting a pivotal role
e-mail: david.blum@inserm.fr for Tau pathology and spreading in AD-related

© Springer Nature Singapore Pte Ltd. 2019 259

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_21
260 M. Gratuze et al.

memory impairments [72]. Remarkably, some Physiology of Insulin

AD patients also present a­ lterations in hypothal-
amus, a brain structure that plays a key role in Peripheral Insulin
the central control of energy metabolism [193].
Indeed, in a study investigating 28 patients with Insulin Biosynthesis
AD, 22 showed plaques and NFTs in the hypo- Insulin is a peptide hormone composed of 51
thalamus [193]. Braak in its description of the amino acids which was discovered in 1921 by
different stages of AD reported plaques and tan- Banting, Best, Macleod and Collip from dog
gles in the hypothalamus at stages III and IV pancreatic extracts (for history Rosenfeld [254]).
[37]. Interestingly these stages correspond to Mature insulin is produced in the beta cells of
early-moderate AD, whereas disturbances in the islets of Langerhans and is formed of two
metabolism, such as weight loss, which is regu- peptide chains connected by two disulfide bonds:
lated by the hypothalamus, are often described to the A and B chains that are composed of 21 and
appear prior to cognitive impairments. The latter 30 amino acids, respectively. It derives from the
indicates that factors different from Tau and Aβ proteolytic processing of a precursor molecule
accumulation in the hypothalamus could play a called Proinsulin that is cleaved into mature
role in metabolic deregulation. Indeed, studies of insulin and C-peptide by the prohormone con-
early cases of AD show neuronal loss in several vertases PCSK1 and PCSK2 (for review Steiner
nuclei of the hypothalamus such as the supraop- et  al. [275]). After cleavage, mature insulin is
tic nucleus (SON), and the paraventricular stored in secretory granules before being secreted
nucleus (PVN), and more so in the suprachias- into the bloodstream by exocytosis (for review
matic nucleus (SCN), with minimal deposits of Fu et al. [85]).
Tau or Aβ [17, 281]. SCN degeneration in AD
was correlated with circadian rhythm deregula- Insulin Secretion
tion of body temperature [115]. Notably, deregu- The main stimulus for insulin secretion is the
lation of body temperature is well characterized increase in blood sugar. When glycemia
to promote Tau pathology in vivo [75, 104, 295, increases, glucose enters in β-cells of pancre-
303]. Hyperphosphorylated Tau as well as NFT atic islets through the glucose 2 receptor
[173] and atrophy [276] in the hypothalamus (GLUT2). Subsequently, glycolysis and Krebs
have also been reported in mouse models of cycle at the mitochondrial level occur leading to
Tauopathies. an increase of intracellular ATP.  This is fol-
Hyperphophorylated Tau is the result of an lowed by a membrane depolarization, caused
imbalance between kinases and phosphatases. by the closing of K+ channels, sensitive to
More than 30 kinases have been described as ATP.  Voltage dependent Ca2+ channels then
regulating Tau phosphorylation in vitro while open and induce a massive influx of calcium
only a few have been confirmed in vivo. Among ions leading to the secretion of insulin by exo-
them, GSK3-β is an important kinase which cytosis of the secretory granules (for review
phosphorylates Tau on more than 30 sites and [70, 122]). Other factors may also stimulate
seems to play a pivotal role in AD and NFT insulin secretion such as Glucagon-like pep-
development [14]. All serine/threonine phospha- tide-1 (GLP-1) and other digestive tract hor-
tases in the brain are capable of dephosphorylat- mones (secretin and gastrin) [163, 249]. Insulin
ing Tau in vitro except PP2C [100, 101, 177, 240, secretion and action are also regulated by the
319], with PP2A contributing to 71% of the autonomic nervous system (ANS), parasympa-
phosphatase activity on Tau [294]. Interestingly, thetic fibers directly stimulating insulin secre-
several kinases and phosphatases of the Tau pro- tion through muscarinic receptors M3 present at
tein are particularly involved in the signalling of the surface of the β-cell. The main effect of
insulin pathway such as GSK3-β, AMPK, ERK, sympathetic innervation is an inhibition of insu-
JNK, PP1 and PP2A. lin secretion through β adrenergic receptors.
21  Tau, Diabetes and Insulin 261

Insulin Signaling while adipose tissue captures only 10% of plasma

The insulin receptor (IR), which is widely glucose [66]. Glucose that is not used or stored as
expressed in peripheral tissues, is a heterotetra- glycogen can be converted into fatty acids in liver
meric membrane glycoprotein belonging to the and adipose tissue under the influence of insulin
tyrosine kinase receptor family. It is composed of (for review Newsholme and Dimitriadis [205]
two α- subunits at the extracellular level that bind and Smith [270]) through a process called de
insulin, and two β-subunits generating the intra- novo lipogenesis [296]. Additional functions of
cellular signaling cascade of insulin. When insu- the hormone have been evidenced in the brain,
lin binds to α subunits, a conformational change particularly in the regulation of food intake,
occurs, leading to the receptor autophosphoryla- reproduction, learning and memory and are
tion of tyrosines on the β-subunit. Insulin signal- detailed in section “Brain Insulin Physiological
ing through its receptor involves the tyrosine Functions”.
phosphorylation of cellular substrates, including
several insulin receptor substrates (IRSs) [313] as
well as other scaffold proteins such as Src-­ Insulin in the Brain
Homology2/a Collagen (SHC) [145] and Grb2-­
associated docking protein (Grb2-SOS) [125].  rigin of Brain Insulin
O
The IRSs ensure metabolic effects of insulin Insulin was first detected in the rat brain through
through the activation of PhosphatidylInositol immunostaining in 1978 [118]. In this study, lev-
3-Kinase (PI3K) which recruits both the Ser/Thr els of insulin were higher in the brain compared
3-phosphatidylinositol-dependent protein kinase to plasma and seemed to be independent of the
(PDK) and protein kinase B (PKB or Akt) to the peripheral concentration [117]. Immunoreactive
plasma membrane, where Akt is activated by insulin was also found in dogs brain and seemed
PDK1- and PDK2 [57]. At the end of this path- to be correlated with basal plasma levels [322]
way, mTOR links insulin signaling to nutrient and dependent on feeding state, being lower dur-
sensing [284]. Mitogenic effects of insulin ing fasting [279]. Insulin mRNA was then identi-
involve the recruitment of mitogen-activated pro- fied by in situ hybridization in various brain areas
tein kinases (MAPKs) through small G-protein including hypothalamus, hippocampus and olfac-
Ras leading to the translocation of ERK to the tory bulb in rodents [67, 324], suggesting a puta-
nucleus, where it controls gene expression [92]. tive role of the CNS in the production of the
hormone. While central insulin production is still
I nsulin Physiological Functions controversial, insulin transport from the periph-
The main insulin biological role is to lower the ery to the brain is well documented. Transport of
level of blood glucose in case of hyperglycemia, insulin through the blood brain barrier (BBB)
following a meal for example. For this, insulin was first reported in rabbits using radiolabeled
increases the use of glucose by peripheral tissues insulin [71]. Insulin transport necessitates bind-
such as skeletal muscle or adipose tissue, stimu- ing of the hormone to the IR and transcytosis of
lates hepatic glycogenogenesis and inhibits the IR insulin complex through brain endothelial
hepatic neoglucogenesis (for review Barthel and cells allowing insulin to cross the BBB without
Schmoll [24]). Glucose uptake by muscle and fat being degraded. This transport can be regulated
cells is made possible by the binding of insulin to by high fat diet, astrocyte stimulation or nitric
its receptor leading to the translocation of GLUT4 oxide inhibition [107].
glucose transporters to the cell membrane. Insulin
then stimulates the use of glucose by peripheral  rain Insulin Receptors
B
tissues as an energetic substrate for glycolysis Radiolabeled 125I-insulin allowed the detection of
or promotes its storage in the form of glycogen the receptor in the brain of monkeys, pigeons,
or triglycerides. Skeletal muscles capture the humans and rats [69, 144, 234]. IRs found in the
majority of glucose in case of hyperglycemia, brain and liver showed different patterns of
262 M. Gratuze et al.

migration during electrophoresis suggesting two signaling was also shown to play a role in reward
different isoforms [180]. In addition, conversely dopaminergic circuitry, since inactivation of IR
to peripheral IRs, brain receptors are not down- in tyrosine hydroxylase-expressing cells resulted
regulated by insulin excess [119]. IR in the brain in increased body weight, elevated fat mass, and
is present at all stages of development, although hyperphagia and seemed to be dependent of
its distribution and concentration vary between dopaminergic neurons in the Ventral Tegmental
embryonic and adult brain suggesting a role of Area (VTA) and Substantia Nigra (SN) [159].
insulin in neurogenesis [144]. In situ hybridiza- Experiments using intranasal insulin, which
tion of IR mRNA in the adult rat brain revealed allow the uptake of insulin from olfactory nerves
high gene expression in the hypothalamus, more [250], bypassing the BBB, were of significant
particularly in the anterior nuclei, in the hippo- importance to isolate insulin effects on the brain
campus, the olfactory bulb and the choroid plexus from the periphery. In Humans, intranasal insulin
[189], regions concerned with olfaction, appetite, has been shown to regulate responses to food
cognition and autonomic functions [311]. cues and smelling capacity [82], increase satiety
Although insulin receptors in the brain possess [114] and to decrease food intake [135]. In ani-
kinetic [326], pharmacological and signaling mal models, brain insulin has been reported to be
characteristics very similar to those present at the involved in satiety [18–20], reduction of lipolysis
periphery, a smaller size of the β subunits and a in adipose tissue [261], and regulation of insulin
more large glycosylation of insulin  receptors in sensitivity in skeletal muscle and liver (brain
periphery is however notable [119, 140, 180]. insulin is required for glucose production) [209].
In the rat brain, insulin regulates enzymes of
 rain Insulin Physiological Functions
B cerebral glucose metabolism such as hexokinase
The biological role of insulin in the CNS is diffi- and phosphofructokinase [128]. Like at the
cult to study and is thus not fully characterized. periphery, insulin is involved in glucose uptake in
However, receptor localization and in vivo and in the brain, although to a lesser extent since the
vitro studies have established several insulin most abundant glucose transporters in the brain,
brain functions. GLUT-1 (entire brain astrocytes and endothelial
One of the best-known roles of central IR acti- cells), GLUT-2 (hypothalamus), and GLUT-3,
vation is its involvement in the regulation of glu- (the major glucose transporter in the neurons of
cose metabolism and feeding behavior. For the cerebellum, striatum, cortex, and hippocam-
instance, study of the neuron-specific knockout pus and some glial and endothelial cells) are not
mouse model of the insulin receptor (NIRKO insulin dependent (reviewed in Blazquez et  al.
mice) showed that inactivation of IR in the CNS [33]). Indeed insulin dependent GLUT-4 is found
led to increased food intake and obesity, together in much lower amounts in selective areas of the
with insulin resistance, hyperinsulinemia and brain, including the olfactory bulb, dentate gyrus
hypertriglyceridemia [43]. Further studies using of the hippocampus, hypothalamus, and cortex
selective inactivation of IR in specific neuronal [76]. The neuron-specific glucose transporter
populations of arcuate nucleus in the hypothala- GLUT-8 (also known as GLUTX1), also insulin
mus showed that agouti-related peptide (AgRP)- dependent, is expressed in several areas of the
expressing neurons, but not pro-opiomelanocortin brain, in particular in the hippocampus where it is
(POMC)-expressing neurons were required for thought to contribute to glucose homeostasis in
central insulin control of hepatic glucose produc- neurons [225, 246].
tion [160]. In the same manner, inactivation of IR Insulin in the brain also seems to play a role in
in Neuropeptide Y (NPY) expressing neurons reproduction. The same mice previously
induced an obese phenotype, leading to the con- described as being deleted for insulin receptors
clusion that insulin signaling in NPY neurons only in neurons (NIRKO mice) have reduced fer-
controls food intake and energy expenditure tility, due to impaired spermatogenesis and matu-
[179]. In addition to the hypothalamus, insulin ration of ovarian follicles  following luteinizing
21  Tau, Diabetes and Insulin 263

hormone (LH) dysregulation [43]. The presence [26, 27], with no effect on word recall and non-­
of insulin receptors on the neurons producing declarative memory but improvement of declara-
Gonadotropin-Releasing Hormone (GnRH) in tive, hippocampus-dependent memory contents
the hypothalamus could explain this phenome- [26, 148]. Besides, epidemiological studies have
non. Indeed, GnRH is a hormone that controls the shown that patients with deregulation of insulin
development of ovarian follicle, ovulation and secretion and/or function (type 1 and type 2 dia-
maintenance of the luteal body in the menstrual betics)  have cognitive impairment and an
cycle in women, but also spermatogenesis in increased risk of AD [253].
men. Studies have shown that insulin stimulates Thus, insulin is a key hormone in the body,
the release of GnRH both in vitro and in vivo [45, whether in the CNS or at the periphery. It is
162, 289]. therefore not surprising that an alteration of the
Other insulin functions not related to metabo- functions of this hormone leads to pathological
lism and reproduction have been suggested to conditions, the best known being diabetes
occur in the brain: neuroprotective qualities, mellitus.
stimulation of neuritic growth and synaptic plas-
ticity, promotion of neuronal survival during
development, stimulation of protein synthesis as Pathophysiology of Insulin
well as a role in the neuronal activity (for review Functions: Diabetes Mellitus
Blazquez et al. [33]).
Finally, numerous studies suggest that insulin Diabetes mellitus, commonly referred to as dia-
may be involved in formation and storage of betes, is a chronic condition characterized by
memory. The hypothesis linking insulin to mem- hyperglycemia. Chronic hyperglycemia is caused
ory is born from the abundant presence of insulin by a lack or failure of insulin use. There are dif-
receptors in the hippocampus and the cortex, ferent forms of diabetes, the two major ones
highly involved in reward and recognition sys- being type 1 diabetes (T1D) and type 2 diabetes
tems as well as in memory. Insulin through (T2D) accounting for 5–10% and 90–95% of
GLUT-4 and GLUT-8 could allow that areas cases respectively. According to the WHO, 422
involved in cognition receive the necessary million people worldwide had diabetes in 2014.
amount of fuel to function. Several studies This pathology is in full expansion; by 2030, dia-
in  mice indicate that insulin contributes to betes will be the seventh leading cause of death in
changes in hippocampal synaptic plasticity by the world [191].
increasing the induction of long term potentiation
(LTP) [202] or long-term depression (LTD)
[300], two molecular mechanisms involved in Type 1 Diabetes
hippocampal-dependent learning and memory. In
accordance, impaired LTP and spatial learning T1D is insulin-dependent diabetes, characterized
have been reported in mice with a down-­ by autoimmune destruction of insulin-producing
regulation of IRs in the hippocampus [109]. One β-cells of pancreatic islets. This destruction leads
of the mechanisms proposed is the regulation of to very low or no insulin production resulting in
cell membrane expression of N-methyl-D-­ hyperglycemia. This type of diabetes, whose
aspartate (NMDA) receptors by insulin [269]. diagnosis is made early in life (children and
These changes are thought to involve insulin acti- young adults), requires treatment with insulin
vation of ERK1/2 [315] or PI3K signaling [257]. accompanied by tight blood glucose control to
Insulin signaling has been shown to contribute to maintain glycemic homeostasis in the body.
synaptogenesis and synaptic remodeling in the Although the exact causes of this disease are still
rat brain and cultured hippocampal neurons [1]. not well understood, it seems that a combination
In Humans, administration of intranasal insulin of genetic factors, mainly related to the major
improved memory functions in healthy subjects histocompatibility complex, and environmental
264 M. Gratuze et al.

contribution, probably viral or toxic, are neces- [262], Turban and Hajduch [298], and Roberts
sary for the development of the disease [13]. In et al. [251]).
terms of clinical symptoms, T1D results in weight There are several risk factors that promote the
loss despite an increased appetite, accompanied development of T2D, the three main ones being
by polyuria (abundant urine) and polydipsia genetic factors, age and obesity [49, 129, 196].
(feeling of intense thirst). Historically diagnosed in adults, T2D is a grow-
ing disease that is now increasingly present in
children and adolescents mainly because of the
Type 2 Diabetes raise of childhood obesity in recent decades.
According to the WHO, T2D cases have almost
T2D is a diabetes independent of insulin. In T2D, quadrupled since 1980  in the world, mainly
β-cells of pancreatic islets produce insulin cor- because of the growth of obesity and the aging
rectly, but when secreted because of hyperglyce- population. More than one in ten people will be
mia, cells are unable to use it to initiate insulin diabetic by 2025 if this trend persists. These
signaling cascade to restore physiological blood results are worrying as T2D is an important risk
glucose level: this is insulin resistance [6]. Insulin factor for many diseases such as vascular dis-
resistance results in decreased skeletal muscle eases that can lead to heart problems, blindness
uptake of glucose and increased production of or amputations [224, 316]. T2D is also a subse-
free fatty acids by adipose tissue, as adipose cells quent risk factor for the development of neurode-
are no longer able to use insulin to inhibit lipoly- generative diseases such as Alzheimer’s disease
sis. In most patients, persistent hyperglycemia [30, 31, 203].
due to insulin resistance in the cells causes the
body to overproduce insulin in an attempt to
restore circulating glucose homeostasis. This  iabetes Mellitus, Tau Pathology
D
positive feedback from insulin-producing cells and AD
leads to hyperinsulinemia that accompanies
hyperglycemia [80]. Ultimately, this overproduc- T1D, Tau Pathology and AD
tion of insulin can lead to depletion of β cells of
pancreatic islets [64, 65]. Patients become then, In Humans
as in T1D, hyperglycemic and hypoinsulinemic. There are currently no epidemiological data to
The exact causes of insulin resistance are not link T1D to AD. These data are difficult to col-
yet well understood. Rare cases have been lect given the lower life expectancy of T1D
explained by mutations in genes encoding insulin patients [130, 220]. However, many studies have
receptors or key players in its signaling pathway shown significant cognitive and intellectual
[291]. However, for most patients, no mutation impairments in patients with T1D, including
has been found. The strongest hypothesis to memory and mental plasticity, as in AD [38, 83,
explain insulin resistance is related to physical 121]. The early onset of T1D appears to be an
inactivity that could lead to chronic inflammation aggravating factor for cognitive impairment in
of the tissues and thus contribute to the develop- patients [90, 176]. In addition, higher phosphor-
ment of insulin resistance through overproduc- ylated Tau levels were observed in the CSF of
tion of lipids and metabolic damage. This patients with T1D compared to healthy subjects,
hypothesis suggests that the accumulation of as in patients with AD [212]. Interestingly, the
lipids in the tissues activates pro-inflammatory same study also revealed a rise of Aβ levels in
signaling pathways. This pro-inflammatory acti- the CSF of these patients, contrary to what is
vation impairs insulin signal transduction by seen in AD (Fig. 21.1).
altering phosphorylation events and key protein-­ Several studies have observed Alzheimer-type
protein interactions (for review Schmitz-Peiffer neuropathologies in animal models of T1D,
21  Tau, Diabetes and Insulin 265

Type 1 diabetes Type 2 diabetes

Patients
Memory loss
Cognitive impairments
Hyperphosphorylated Tau Patients
Brain atrophy
Animal ApoE dependent:
(Normal and STZ treated mice) Amyloid plaques
Worsening of Tau pathology NFTs
Memory impairments
Patients Animal
Increase in Tau csf (leptin protein or
receptor mutation)
Animal Tau hyperphosphorylation
(Normal and STZ treated mice) (HFD, HFHSD)
Decrease affinity Tau-MT Increased Tau pathology

Alzheimer’s disease

Fig. 21.1  Venn diagram of common symptoms between microtubules, ApoE apolipoprotein E, NFT neurofribril-
Alzheirmer’s disease and diabetes in patients and animal lary tangles, HFD high fat diet, HFHSD high fat high
models. STZ streptozotocin, CSF cerebrospinal fluid, MT sucrose diet

including the development of Tau pathology. ferent STZ injections in many studies [56, 138,
Although the link between T1D and AD is not 151, 228, 239].
clearly established in humans, these data provide Although Tau hyperphosphorylation seems
valuable information because they allow to eval- well characterized following STZ injections, the
uate the effects of hyperglycemia and hypoinsu- mechanisms proposed to explain this phenomenon
linemia on Tau pathology. differ from one study to another. Some suggest an
increase in kinase activity, and more specifically
In Animal Models GSK-3β, Tau major kinase, via a decrease of its
One of the most commonly used methods to inhibitory phosphorylation on serine 9 [138, 239].
mimic T1D is based on intraperitoneal injection Other studies attribute hyperphosphorylation
of streptozotocin or STZ (2-deoxy-2- (3- (methyl-­ of Tau following STZ injections to inhibition of
3-­nitrosoureido)-D-glucopyranose). This drug Tau phosphatases, mainly PP2A, Tau’s major
enters the β cells of pancreatic islets, by GLUT2 phosphatase [228, 239]. According to Planel et al.,
receptors, thanks to its analogy with glucose the decrease of phosphatase activity leading to
[283]. STZ then destroys these cells by methyla- Tau hyperphosphorylation is the result of animal
tion of the DNA, resulting in hypoinsulinemia hypothermia caused by profound metabolic altera-
accompanied by hyperglycemia, as in tions after STZ injections [228]. The latter study
T1D.  Hyperphosphorylation of Tau protein on demonstrated that hyperphosphorylation of Tau is
many epitopes, a decrease in its affinity for greatly reduced in STZ-­injected mice whose nor-
microtubules and impaired memory have been mothermia was restored. Indeed, the phosphoryla-
observed in non-transgenic rodents following dif- tion of Tau is very s­ ensitive to temperature changes;
266 M. Gratuze et al.

a degree below 37°C can induce an 80% increase T2D, Tau Pathology and AD
of Tau phosphorylation in mice [227]. Hypothermia
is a common consequence of diabetes in humans In Humans
[204, 265] and in animal models [149, 207, 267]. Although early epidemiological studies failed to
Unfortunately, very few studies that observed find a link between T2D and AD [120, 166], it is
hyperphosphorylation of Tau measured animal now well established that T2D is a major risk fac-
body temperature following STZ injection. tor for AD. One of the first studies to clearly dem-
Several authors have induced T1D with STZ onstrate this, is the Rotterdam study of over 6000
injections on transgenic models of AD.  These subjects [124, 210, 211]. The authors reported
studies evaluate whether metabolic disorders that T2D doubled the risk of developing AD, an
may exacerbate or interfere with the development even greater risk when diabetic patients were
of AD-like neuropathology in these models. In treated with insulin. Subsequently, many other
addition, the use of transgenic models makes it epidemiological studies have confirmed these
possible to observe markers that are very difficult findings [12, 42, 170, 182, 216, 318]. A 2010
to observe in non-transgenic animals such as Tau meta-analysis determined, based on available
aggregates. Ke et al. analyzed the effects of STZ studies at that time, a relative risk of 1.54 to
on Tau pathology in pR5 transgenic mice over-­ develop AD for a type 2 diabetic patient [235].
expressing mutated Tau protein (P301L) [146]. This study also notes that the presence of the
These mice, which spontaneously develop hyper- ApoE4 isoform significantly increases the risk of
phosphorylation of Tau and NFT, show an exac- AD in T2D patients compared to diabetics with
erbation of these 2 phenotypes following STZ ApoE3 or ApoE2 isoforms. Moreover, progres-
injection. In hTau mice (overexpressing non-­ sion to AD appears to be more rapid in diabetic
mutant human Tau), we did not detect any change patients, both symptomatically [5], and neuro-
in Tau solubility after STZ injection despite pathologically [174, 299].
marked hyperphosphorylation [106]. The differ- Finally, these data are reinforced by the 2014
ences between the two models  could be due to Alzheimer Society report, which announces a
the fact that P301L Tau mutant is more prone to comorbidity of 29% between AD and T2D [190].
aggregation than non-mutant Tau. Indeed, the Moreover, a study in a retirement home described
P301L mutation has been shown to accelerate the dementia as the most common comorbidity in
formation of paired helical filaments and pro- T2D patients [87]. These observations are not
mote β-sheet formation [21, 81, 308]. While hTau surprising, as one in ten cases of dementia may
mice do develop Tau aggregates, they do it slower be linked to T2D [32].
than pR5 mice. Interestingly, insulin injection Since the link between T2D and AD is well
30 min before sacrifice partially restored physio- established at the epidemiological level, it was
logical Tau phosphorylation levels in STZ-­ subsequently essential to understand the molecu-
injected hTau mice [106], confirming a link lar mechanisms underlying this link, to better
between insulin homeostasis and Tau phosphory- understand why and how T2D increases the risk
lation. Some animal models develop T1D sponta- of AD.
neously and therefore do not require drug Studies in humans have shown that T2D
injection. Phosphorylation of Tau was analyzed patients display  many neuropathological altera-
in two of these models, BB / Wor rats (for Bio-­ tions similar to those of AD. Indeed, significant
Breeding/Worcester) and NOD mice (for Non-­ cognitive impairments  including memory disor-
Obese Diabetic). While no change in Tau ders have been observed in diabetics [15, 278],
phosphorylation was observed in BB/Wor rats and mainly in the elderly [255]. These cognitive
[175], NOD mice showed hyperphosphorylation disorders could be caused by marked brain atro-
of Tau protein, accompanied by inhibition of phy in diabetic patients [187]. Brain atrophy
PP2A activity [213], as in some STZ-injected affects very similar regions to those of AD, such
models of T1D [228, 239]. as the hippocampus [32, 200, 252]. Moreover,
21  Tau, Diabetes and Insulin 267

brain loss, a normal process seen during aging, is results explain the presence of NFT at more
much more pronounced in T2D patients than in advanced Braak stages in the brains of these
healthy people of the same age [63, 78, 161, 301]. patients with T2D.
Several hypotheses exist to explain this brain However, post mortem data on humans are
atrophy. Among the most supported, the duration rare and have many limitations as to the mecha-
of diabetes, specifically insulin resistance corre- nistic study linking the two pathologies. That is
lating with atrophy, as well as cerebral infarc- why many studies have turned to animal models
tions related to vascular alterations in T2D seem of T2D, and combined models of T2D and AD.
quite plausible [186, 187]. The resulting vascular
disorders of T2D are also present in the micro- I n Animal Models
vasculature of the brain and could contribute to There are now many animal models that sponta-
the loss of brain matter and cognitive deficits. neously develop a T2D (for review King [154]).
Cerebrovascular abnormalities, also present in These models are derived from animals with one
AD, may explain, at least in part, the decline in or more genetic mutations transmitted from gen-
glucose uptake and hypometabolism in some eration to generation, or by selective cross-­
brain regions of T2D patients [16, 89], as in breeding of individuals leading to diabetic lines
AD.  A recent study using brain imaging with such as Otsuka Long-Evans Tokushima Fatty
PET and 18F-FDG biotracer showed an exacerba- (OLETF) in the rat. Studies have explored some
tion of reduced cerebral glucose metabolism in Alzheimer-type pathologies to better understand
patients with mild cognitive impairment when the links between these diseases in many of these
they were also diabetic compared to patients with models.
mild cognitive impairment without diabetes The db/db mice, with a mutation in the leptin
[327]. No statistical difference was observed receptor  gene,  leptin being a satiety hormone,
between healthy and diabetic patients despite a show an important hyperphosphorylation of Tau
strong tendency to reduced glucose metabolism protein in their brains according to several stud-
in the brain of diabetic patients. Nevertheless, ies [75, 151, 172, 243, 268]. However, the study
other studies on larger cohorts of patients (only by Jolivalt et al. observed no change in Tau phos-
31 diabetic patients in the study by Zhang et al. phorylation in this model [138]. The mechanistic
[327]) are needed to give a definitive answer on hypotheses of this hyperphosphorylation of Tau
the impact of T2D on brain glucose metabolism. protein in db/db mice vary from one study to
Finally, amyloid and Tau pathologies have another; activation of some Tau kinases like
been observed in diabetic patients [134, 178, GSK-3β [268] or JNK [172] has been suggested,
195, 216]. Aβ deposits and hyperphosphorylated but also an alteration of the BBB [243] or an
Tau proteins were detected in the pancreatic increase of Tau cleavage [151]. Our study sug-
islets of patients with T2D [195]. Amyloid gests that Tau hyperphosphorylation is mainly
plaques as well as NFT were also found in greater due to hypothermia caused by impaired thermo-
numbers in the hippocampus of T2D patients regulation in db/db since restoring normothermia
with the ApoE4 allele compared to healthy recovered physiological phosphorylation of Tau
patients [216]. On the other hand, another study similar to control mice [75]. It is interesting to
shows that there is no increased incidence of note that some studies have observed a rise in lev-
amyloid deposits in patients with T2D [134]. els of Aβ (40 and/or 42) in these mice [172, 272].
However, when cerebral amyloid plaques are However, these results have not been confirmed
present, the extent of accumulation correlates by other teams [206, 241].
with the duration of T2D [134]. Regarding the Ob/ob mice, which have a mutation in the
Tau protein, a study showed its hyperphosphory- gene encoding leptin, are also a common model
lation on many epitopes (Ser202, Thr217, Ser262 of T2D in which hyperphosphorylated Tau pro-
and Ser396) in the frontal cortex of T2D patients tein has been observed in the brain [104, 150,
compared with healthy patients [178]. These 256]. This hyperphosphorylation of Tau is also
268 M. Gratuze et al.

present in OLETF [142] and Bio-Breeding Studies done on transgenic mice reproduced
Zucker diabetic/ Worcester (BBZDR/Wor) [175] these inconsistent results. Indeed, while
rats. These observations are often associated with Leboucher et al. observe a hyperphosphorylation
increases in Aβ levels [175, 256]. Unfortunately, of soluble Tau protein without increasing its
the mechanisms of Tau hyperphosphorylation in aggregation in THY-Tau 22 mice (Tauopathy
these models are still poorly understood. model with two mutations on Tau (G272 V and
Other pathological changes in Tau have been P301S)) under a high-fat diet [167], Koga et al.
observed in these rodents, like an increase in Tau report a hyperphosphorylation of Tau only in the
cleavage in db /db and ob/ob mice [150, 151], an insoluble fraction of PS19 mice (model of
alteration of Tau exon 10 splicing in favor of the Tauopathy with P301S Tau mutation) under high
Tau 3R form in OLETF rats [142] or a decrease fat diet [158]. In 3xTg-AD mice under high fat
of Tau O-glycosylation in ob/ob mice [104]. diet, Ma et al. observe abnormal phosphorylation
Some studies have crossed these spontaneous of Tau on serine 422 [184] while others detect no
models of T2D with transgenic models of AD in change in Tau phosphorylation [23, 141, 157,
order to better evaluate the interaction between 304]. Ramos-Rodriguez reported an increase in
T2D and AD pathologies. In contrast to the previ- Tau phosphorylation, microglial activation and
ous models, these crosses make it possible to impaired memory in APPswe/PS1dE9 after high-­
determine if T2D can aggravate AD pathologies. fat diet [242]. Some research groups have com-
The ob/ob and db/db mice were crossed with bined these high-fat diets with high sugar and/or
murine models of amyloid pathologies: APP23 cholesterol diets to mimic a “western” diet, often
mice [286] and APP/PS1 mice [241] respectively. accused of the worrisome growth of obesity and
Crossing these models of amyloid pathologies T2D [7, 29, 47, 123, 214, 218, 285]. Most
with T2D models worsens cognitive disorders observed hyperphosphorylation of Tau protein
and increases Aβ levels (but not aggregation). The regardless of the animal model used, but signifi-
study by Ramos-Rodriguez et al. even shows an cant variability in the results persists.
exacerbation of Tau phosphorylation in db/APP/ Given the variability of all these results, we
PS1 mice [241]. One team also transduced db/db can hypothesize that some confounding parame-
mice with mutated P301L Tau protein using an ters vary from one study to another, such as the
adeno-associated virus, exacerbating Tau pathol- composition of diets (caloric source, percentage
ogy (both phosphorylation and NFT), but the of vitamins or fatty acids saturated vs. unsatu-
mechanisms involved remain unexplored [229]. rated, omega 6/3 ratio, unsuitable diet control,
Although all these models provide very valu- etc.), duration of diet, animal model used or tem-
able information on the impact of T2D on AD, perature of animals and use of anesthesia. In
they are quite far from human and physiological addition, many studies directly attribute the
conditions. To get close to a human situation, effects of diets observed in T2D while this type
studies have used hypercaloric diets in animal of diet can modulate certain factors such as obe-
models to mimic a western diet, a common cause sity [167], leptin signaling [108] or cerebrovas-
of T2D in humans. Most of these studies use a cularization and inflammation [203] that may
diet with a high calorie fat percentage. also modify Tau phosphorylation.
Unfortunately, despite a large number of studies, With the main goal of clarifying the impact of
the results obtained vary widely from one study hypercaloric diets on Tau pathogenesis, we pub-
to another. For example, high-fat diets (between lished a study closer to human conditions [105].
54% and 60% kcal from fat) in wild-type mice We therefore chose hTau mice that express non-­
induced hyperphosphorylation of Tau in two mutated human Tau protein as in AD and capable
studies [136, 152], but no change in phosphoryla- of developing NFT [8, 9]. We used diets high in
tion in two other studies [25, 243], and even a fat, sugar and cholesterol (alone or combined),
decrease of Tau phosphorylation in another study the three key components of western diet, at lev-
compared to mice under control diet [292]. els comparable to human diets (in pathological
21  Tau, Diabetes and Insulin 269

conditions). Surprisingly, we found no effect of notable increase of inhibitory IRS-1 phosphoryla-

fat, sugar and cholesterol overconsumption, even tion at Serine 616 and 636/639 has been particu-
when combined, on Tau phosphorylation, larly observed [287]. Tau pathology would likely
O-GlcNAcylation, splicing, cleavage and aggre- contribute to the establishment of such brain insu-
gation, suggesting that their overconsumption lin resistance. Indeed, a study by Yarchoan et al.
does not seem to aggravate Tau pathology in found increased phosphorylation of Serine
these mice [105]. While we have demonstrated IRS-1 in primary Tauopathies, including Pick dis-
that fat, cholesterol and/or sugar overconsump- ease (PiD), corticobasal degeneration (CBD) and
tion did not exacerbate Tau pathology in this PSP, but to a lesser degree than in AD [323].
model, our study calls for further high quality
investigations (controlling and reporting a maxi-
mum of parameters) in different mice models to  onsequences of Brain Insulin
C
have a definite answer on whether these diets can Resistance for Cognition, Longevity,
enhance Tau pathogenesis. and Metabolism
Together, the data summarized above indicate
that a decrease of insulin production or sensitiv- In accordance with the above-mentioned role of
ity may exacerbate Tau pathology suggesting that insulin towards the regulation of synaptic plastic-
brain insulin resistance could play a key role in ity and memory, correlation between increased
AD and Tauopathies. serine phosphorylation of IRS-1 and cognitive
score in AD patients (episodic and working mem-
ory) sounds logical [287]. However, while brain
 rain Insulin Resistance and Tau
B insulin signaling finely tunes energy homeostasis
Pathology: A Vicious Circle? and peripheral glucose homeostasis, much less is
known about the potential impact of brain insulin
 rain Insulin Resistance in AD
B resistance towards metabolic disturbances in AD.
Patients Metabolic alterations have been reported in
patients with AD with or without diabetes. Weight
Considering the prominent role exerted by insulin loss was already described by Aloïs Alzheimer
towards plasticity and cognition, it is not surpris- [96, 288] and it is now recognized as a clinical
ing that the AD brain has been described to exhibit feature of AD [194], impacting 20–45% of
an insulin resistance state, the so-called “Type 3 patients [94, 95, 111, 112, 305]. Weight loss in
Diabetes” [127, 274]. Indeed, CSF insulin levels patients with dementia is associated with acceler-
have been described to be lower in AD patients ated progression of AD, higher rate of institution-
than in healthy subjects [61, 93] although these alization [10] and increased mortality [79, 88,
data have not always been confirmed by others 312]. Although dementia-associated weight loss
[86, 198]. Some explained reduced CSF insulin often begins before the onset of the clinical syn-
levels by a decreased transport of insulin across drome and accelerates by the time of diagnosis, it
the BBB as suggested by Banks et  al. who is unclear if it is a cause or a consequence of AD
reported that chronic plasma hyperinsulinemia, pathology [131]. Metabolic deregulations in ani-
often reported in AD patients [44, 61, 86, 183, mal models of Tauopathies resemble those found
277], diminishes the passage of insulin to the in patients with Tau related dementia. Indeed,
brain [18–20]. However, whether all AD patients decreased body weight is found in mouse models
with reduced CSF insulin display hyperinsu- of Tau deposition such as THY-22 [167] or Tg4510
linemia remained to be established and other [40, 41], sometimes despite increased feeding
mechanisms could be instrumental. Insulin recep- behavior [139]. Additionally, genetic deletion of
tor signaling pathway and ability to respond to Tau resulted in increased body weight [201].
insulin have been described to be strongly Impaired satiation and increased feeding behavior
impaired in the brain of AD patients [199, 287]. A [2], as well as changes in body temperature [156]
270 M. Gratuze et al.

were also observed in a mouse model combining Despite early studies reporting co-localization
amyloid plaques and Tau pathology (3xTgAD). In and correlation between decrease of total IRS-1
Tg4510 mice, weight loss was specific to fat mass and IRS-2 along with increased phosphorylated
and co-occurred with deregulation of metabolic IRS-1 on Ser636/639 and Ser616in the brain of AD
rate [139] as well as disturbances in circadian patients and the NFT deposition [184, 199, 287,
rhythm [276]. In both cases Tau pathology was 323] as well as the link between increased phos-
found in the hypothalamus. phorylated IRS-1 and primary Tauopathies [323],
One could wonder if these metabolic altera- no study had any interest in the direct role of Tau
tions, appearing for most part before the detec- protein on central insulin signaling until a recent
tion of pathological deposits in the brain, could work published by our laboratory [188]. In the
really be the result of early deregulation of study of Marciniak et al., we have demonstrated
central insulin signaling. However, only rare that Tau deletion induces a disruption of hippo-
observations support that brain insulin resistance campal response to insulin and impairs hypotha-
may play a role in the metabolism of AD patients. lamic anorexigenic effect of insulin associated
First, paradoxical overeating concomitant with with energy metabolism alterations (enhanced
weight loss has been observed in patients with food intake and body weight, increased adipose
AD [314] as well as in frontotemporal dementia tissue mass, hyperinsulinemia and glucose intol-
(FTD) [36, 223, 271]. Second, an increased risk erance). These findings strongly suggested a
to develop T2D has been reported in AD patients putative function of Tau protein as a regulator of
(35% diabetics +46% with glucose intolerance) insulin signaling in the brain. The potential
[134]. Furthermore, some sparse studies report mechanism by which Tau regulates hippocampal
that AD patients can exhibit hyperinsulinemia response to insulin seems to act through IRS-1
or alterations in glucose metabolism [44, 61, 86, and PTEN. Indeed, we reported decreased activa-
183, 245, 277]. tion of IRS-1 and AKT after insulin exposure in
Tau KO mice, suggesting brain insulin resistance.
This is consistent with altered IRS-1 activity in
 hat Is the Trigger for Brain Insulin
W AD brains that correlates with NFT deposition
Resistance in AD? [184, 199, 287] whereas a direct interaction
between Tau and IRS-1 has never been reported.
AD brain insulin resistance has been originally On the other hand, we found that Tau interacts
ascribed to the detrimental impact of Aβ oligo- with PTEN, already known to inhibit insulin sig-
mers. Aβ peptide in oligomeric form can promote naling, and that human Tau is able to reduce
insulin resistance by competitively binding and PTEN activity [188]. Although at the current
internalizing IRs [330] and by increasing the stage it is still impossible to determine which of
phosphorylation of IRS-1 and JNK [34] in vitro. IRS-1 or PTEN is the instrument of Tau to regu-
In vivo, these results were confirmed in crabgrass late brain insulin signaling, our study identified a
macaques and APP/PS1 transgenic mice, which new function of Tau protein as a modulator of
received intracerebroventricular (ICV) injections brain insulin signaling and highlights potential
of Aβ oligomers [34]. Consistently, several rodent mechanistic explanation whereby alteration of
models of amyloid pathology develop metabolic insulin signaling would occur in AD via patho-
alterations such as glucose intolerance [55, 137, logical Tau loss of function. In addition, we have
197, 302]. Particularly interesting is the observa- highlighted a link between Tau haplotype and
tion of impaired glucose homeostasis following peripheral metabolism using genome-wide
ICV injection of Aβ oligomers [34]. Recent data association study (GWAS) data [236, 259].
­
also pointed out that ApoE4 could be strongly Indeed, we reported that patients with H1 haplo-
related to the development of brain insulin resis- type, previously shown with higher risk of
tance [329]. Notably, ApoE4 reduces insulin-IR Tauopathies [226], exhibited higher circulating
interaction and impairs IR trafficking. glucose levels and lower insulin levels during an
21  Tau, Diabetes and Insulin 271

oral glucose tolerance test, suggesting that Tau and PP1), known to play a pivotal role in the
impacts peripheral metabolism in humans. These development of Tau pathology [14, 50, 54, 68,
GWAS data contribute to recent evidences sug- 99, 102, 103, 132, 219, 290, 297, 307, 332].
gesting that Tau can regulate both brain insulin Therefore, chronic insulin signaling impairment
signaling and peripheral glucose metabolism. seen in the brains of patients with AD and
Overall, these data open the possibility that Tauopathies, is prone to favour the development
cognitive and metabolic deficits seen in AD of Tau pathology, through the disruption of the
patients are the consequence of a Tau loss-of-­ balance between Tau kinases and phosphatases.
function. For a long time, Tau aggregation into Moreover, altered brain insulin signaling would
NFT was thought to be the only culprit of neurode- also promote Tau pathology by favoring Tau
generation and cognitive deficits. But obviously, cleavage [150, 151] and deregulating Tau alterna-
previous observations supported this is not the tive splicing [285].
case [258, 282]. More recent data ascribed Tau
oligomers as detrimental species towards plastic-
ity. Interestingly, as Tau oligomers [237], Tau dele- Therapeutic Considerations
tion also impairs hippocampal plasticity and
spatial behaviour [3, 4, 153, 154]. This raises the Intranasal Insulin
possibility that both “toxic gain of Tau function”
and “loss of normal Tau function” play a role in the Insulin treatment is a promising therapeutic
development of cognitive deficits in Tauopathies. strategy for treating AD.  Intravenous insulin
Respective contribution of both phenomena will injection has already shown numerous positive
need to be further elucidated in the future. effects on cognition and, in particular, on mem-
ory in healthy subjects [148] and in patients with
AD [60]. Because of hypoglycaemic side effects
I mpact of Insulin Resistance on Tau [58], researchers turned to intranasal insulin
Lesions delivery, allowing rapid delivery of insulin to the
brain with no peripheral hypoglycaemic effect
Interestingly, although the interest for the impact [35, 147, 217]. A first study showed that admin-
of Tau protein on insulin signaling in the brain is istration of intranasal insulin (20 IU twice daily
very recent, it has for many years been known for 21 days) improved memory and attention in
that, conversely, insulin is capable of modulating patients with AD [248]. Another study using
Tau protein, mainly its phosphorylation state doses of 20 IU or 40 IU for 4 months confirmed
[74]. There are increasing data showing that insu- memory and cognitive improvements as well as
lin regulates Tau phosphorylation and could beneficial effects on quality of life [59]. This
exacerbate NFT development in AD [53, 84, 126, same study demonstrated that both doses
171, 260, 263]. In vivo, mice deficient for insulin improve glucose uptake in many brain areas,
receptors in neurons show inhibition of PI3K/Akt using brain imaging techniques with PET and
signalling pathway and an increase in phosphory- 18F-FDG biotracer compared to placebo [59].
lation of Tau [264]. Similarly, the deletion of Although optimistic, these results need to be
IRS-2 gene causes inhibition of the PI3K/Akt confirmed over longer periods and larger cohorts
pathway and therefore induces Tau hyperphos- of patients. This will allow for large-scale analy-
phorylation in IRS-2 knockout mice [263]. sis of the effect of intranasal insulin by consider-
Overall, the pathological effect of brain insulin ing ApoE ε4 allele, since one study reported
resistance on Tau has mostly been attributable to cognitive improvements 45  min after insulin
the modulation of several downstream pathways therapy only in patients without ε4 allele, and
[74], involving some Tau kinases (GSK-3β, JNK, negative effects of insulin in ε4 allele carriers
ERK, and AMPK) and Tau phosphatases (PP2A with the highest dose [247].
272 M. Gratuze et al.

In vivo, Chen et  al. demonstrated that 7-day Although metformin has been suggested as an
intranasal insulin does not alter Tau phosphoryla- interesting therapeutic issue for AD in the past, an
tion in 9-month 3xTg-AD mice, while observing in vivo study has recently reported that metformin
a reduction in amyloid pathology and neuroin- is able to promote Tau aggregation and abnormal
flammation [51]. Conversely, another study behaviour in a mouse model of Tauopathy [22].
observed in the same murine model a decrease in These results could explain, at least in part, recent
anesthesia-induced Tau hyperphosphorylation data reporting that the risk of developing AD was
when mice were pretreated with intranasal insu- 2.13 times as high in type 2 diabetic patients who
lin for 1  week [52]. More recently, intranasal took metformin as in those who did not [165].
insulin has been shown to reduce Tau hyperphos- Results of a 12-month pilot trial of metformin in
phorylation as well as neurodegeneration, neuro- MCI patients have shown no change in glucose
inflammation and cognitive impairments in ICV uptake or plasma amyloid levels but an increase
streptozotocin-injected rats [113] and diabetic in verbal memory scores [181].
mother offspring mice [244]. Although intranasal GLP-1 receptor agonists, commonly used in
insulin appears to be an interesting therapeutic diabetic patients, have shown encouraging
issue, further studies are needed to understand results. GLP-1 is an incretin that increases
how insulin affects memory and whether it is able glucose-­ dependent insulin secretion and
to decrease, or at least slow down, the evolution decreases glucagon secretion, contributing thus
of Tau pathology and neurodegeneration in AD to better insulin sensitivity. In animal models of
and other Tauopathies. AD, exendin-4 and liraglutide, two GLP-1 recep-
tor agonists, restored altered insulin signaling,
decreased amyloid burden, and improved cogni-
 iabetic Drugs: Peroxisome
D tive deficits [34, 192]. A pilot clinical trial test-
Proliferator-Activated Receptor-γ ing liraglutide in AD patients has demonstrated
Agonists, Metformin, GLP-1, Amylin that it was able to prevent decline of brain glu-
and Future Drugs cose metabolism [91]. Two clinical studies are
underway to evaluate the effects of these two
Several diabetic drugs that belong to thiazolidin- drugs in Alzheimer’s patients or with mild cog-
ediones have also been tested in AD including nitive decline.
rosiglitazone, which has reached phase 3 of clini- Another strategy recently tested in AD patients
cal trials. Thiazolidinediones are Peroxisome is pramlintide, an amylin analog [331]. While
Proliferator-Activated Receptor (PPARγ) ago- authors suggested that a single injection of pram-
nists used in the treatment of T2D by improving lintide into AD reduced the amyloid burden and
the sensitivity of tissues to insulin. Following lowered the concentrations of Aβ, this pilot study
encouraging results in preclinical studies on used a too small number of patients to draw
amyloid, Tau and inflammatory pathologies and definitive conclusions. Moreover, no effect on
on cognitive impairments [77, 208, 215, 293, Tau has been found.
317, 325], a dozen clinical studies have evaluated Although diabetic drugs provide a therapeutic
the effects of rosiglitazone in AD patients. alternative for Alzheimer’s disease and other
Although a pilot study has observed encouraging Tauopathies, an important consideration must be
results on memory and Aβ levels in the CSF given to the adverse effects that these treatments
[309], the subsequent phase 3 clinical studies may have on periphery by affecting circulating
showed no beneficial effect regardless of the dose concentrations of glucose and insulin. Moreover,
used or the ApoE genotype [98, 116]. other potential future diabetic drugs may repre-
Metformin, that is a first-line drug for the sent attractive therapeutic approaches to correct
treatment of diabetes, reduces blood glucose by aberrant insulin signaling pathways in AD like
improving glucose uptake by peripheral tissues PTP1B inhibitors [306]. Indeed, PTP1B inhibi-
and reducing gluconeogenesis by the liver. tion has been shown as an effective approach to
21  Tau, Diabetes and Insulin 273

rescue neuronal insulin signaling in non-AD References

models with neuronal insulin resistance [164,
221, 238, 328]. 1. Abbott MA, Wells DG, Fallon JR.  The insulin
receptor tyrosine kinase substrate p58/53 and the
insulin receptor are components of CNS synapses.
J Neurosci. 1999;19(17):7300–8.
 ould Anti-Tau Immunotherapy
C 2. Adebakin A, Bradley J, Gumusgoz S, Waters EJ,
Impact on Brain Insulin Sensitivity? Lawrence CB. Impaired satiation and increased feed-
ing behaviour in the triple-transgenic Alzheimer’s
disease mouse model. PLoS One. 2012;7(10):e45179.
In recent years, Tau immunotherapy has become 3. Ahmed T, Blum D, Burnouf S, Demeyer D, Buee-­
a rapidly growing area of drug development in Scherrer V, D’Hooge R, Buee L, Balschun D. Rescue
clinical trials. A recent study suggests that Tau of impaired late-phase long-term depression in a
tau transgenic mouse model. Neurobiol Aging.
passive immunization reduced not only total and 2015;36(2):730–9.
hyperphosphorylated Tau in the brain, while 4. Ahmed T, Van der Jeugd A, Blum D, Galas MC,
improving memory, but also nudged down lev- D’Hooge R, Buee L, Balschun D.  Cognition and
els of the APP and cleared Aβ plaques [62]. hippocampal synaptic plasticity in mice with
a homozygous tau deletion. Neurobiol Aging.
These data lead us to wonder whether anti-Tau 2014;35(11):2474–8.
immunotherapy impacts on brain insulin sensi- 5. Alagiakrishnan K, Sclater A.  Psychiatric disorders
tivity. According to Marciniak et  al. study, a presenting in the elderly with type 2 diabetes mel-
reduction of Tau protein itself, by immunother- litus. Am J Geriatr Psychiatry. 2012;20(8):645–52.
6. American Diabetes Association. Diagnosis and
apy in this case, could lead to exacerbate brain classification of diabetes mellitus. Diabetes Care.
insulin signaling impairments in AD and 2004;27(Suppl 1):S5–S10.
Tauopathies. However, C2N 8E12, an antibody 7. Anderson RA, Qin B, Canini F, Poulet L, Roussel
currently in phase 2 trial in AD patients might AM. Cinnamon counteracts the negative effects of a
high fat/high fructose diet on behavior, brain insulin
be interesting here since it recognizes an aggre- signaling and Alzheimer-associated changes. PLoS
gated, extracellular form of pathological Tau. In One. 2013;8(12):e83243.
P301S mice, it reduced brain neurofibrillary 8. Andorfer C, Kress Y, et al. Hyperphosphorylation and
pathology, insoluble Tau, microgliosis, seeding aggregation of tau in mice expressing normal human
tau isoforms. J Neurochem. 2003;86(3):582–90.
and brain atrophy without affecting soluble 9. Andorfer C, Acker CM, et al. Cell-cycle reen-
total Tau levels [320, 321]. This antibody could try and cell death in transgenic mice express-
therefore be an interesting alternative to elimi- ing nonmutant human tau isoforms. J Neurosci.
nate pathological forms of Tau (potentially less 2005;25(22):5446–54.
10. Andrieu S, Reynish W, Nourhashemi F, Ousset
able to regulate insulin signaling) while preserv- PJ, Grandjean H, Grand A, Albarede JL, Vellas
ing the soluble Tau protein and its physiological B. Nutritional risk factors for institutional placement
role on insulin signaling. in Alzheimer’s disease after one year follow-up.
J Nutr Health Aging. 2001;5(2):113–7.
11. Arendt T, Stieler JT, Holzer M. Tau and tauopathies.
Acknowledgments Our laboratories are supported by
Brain Res Bull. 2016;126(Pt 3):238–92.
grants from France Alzheimer/Fondation de France
12. Arvanitakis Z, Wilson RS, Bienias JL, Evans DA,
(InsTauBrain project), FHU VasCog research network
Bennett DA. Diabetes mellitus and risk of Alzheimer
(Lille, France), CoEN LiCEND and programs
disease and decline in cognitive function. Arch
d’investissements d’avenir LabEx (excellence laboratory)
Neurol. 2004;61(5):661–6.
DISTALZ (Development of Innovative Strategies for a
13. Atkinson MA, Eisenbarth GS, Michels AW. Type 1
Transdisciplinary approach to ALZheimer’s disease). Our
diabetes. Lancet. 2014;383(9911):69–82.
laboratories are also supported by ANR (ADORATAU to
14. Avila J, Leon-Espinosa G, Garcia E, Garcia-­
DB, SPREADTAU and GRAND to LB), Fondation pour
Escudero V, Hernandez F, Defelipe J.  Tau phos-
la Recherche Médicale, LECMA/Alzheimer Forschung
phorylation by GSK3  in different conditions. Int
Initiative, Fondation Plan Alzheimer as well as Inserm,
J Alzheimers Dis. 2012;2012:578373.
CNRS, Université Lille 2, Métropole Européenne de
15. Awad N, Gagnon M, Messier C.  The relationship
Lille, Région Nord/Pas-de-Calais, FEDER, DN2M.  We
between impaired glucose tolerance, type 2 diabetes,
also want to thank Alzheimer’s Association, the NIH and
and cognitive function. J  Clin Exp Neuropsychol.
Byrd Alzheimer’s Institute for their support.
2004;26(8):1044–80.
274 M. Gratuze et al.

16. Baker LD, Cross DJ, Minoshima S, Belongia 29. Bhat NR, Thirumangalakudi L. Increased tau phos-
D, Watson GS, Craft S.  Insulin resistance and phorylation and impaired brain insulin/IGF sig-
Alzheimer-like reductions in regional cerebral glu- naling in mice fed a high fat/high cholesterol diet.
cose metabolism for cognitively normal adults with J Alzheimers Dis. 2013;36(4):781–9.
prediabetes or early type 2 diabetes. Arch Neurol. 30. Biessels GJ, Deary IJ, Ryan CM.  Cognition and
2011;68(1):51–7. diabetes: a lifespan perspective. Lancet Neurol.
17. Baloyannis SJ, Mavroudis I, Mitilineos D, 2008;7(2):184–90.
Baloyannis IS, Costa VG.  The hypothalamus in 31. Biessels GJ, Kappelle LJ.  Increased risk of
Alzheimer’s disease: a Golgi and electron micro- Alzheimer’s disease in type II diabetes: insulin resis-
scope study. Am J  Alzheimers Dis Other Dement. tance of the brain or insulin-induced amyloid pathol-
2015;30(5):478–87. ogy? Biochem Soc Trans. 2005;33(Pt 5):1041–4.
18. Banks WA, Jaspan JB, Huang W, Kastin 32. Biessels GJ, Reijmer YD. Brain changes underlying
AJ. Transport of insulin across the blood-brain bar- cognitive dysfunction in diabetes: what can we learn
rier: saturability at euglycemic doses of insulin. from MRI? Diabetes. 2014;63(7):2244–52.
Peptides. 1997;18(9):1423–9. 33. Blazquez E, Velazquez E, Hurtado-Carneiro V, Ruiz-­
19. Banks WA, Jaspan JB, Kastin AJ. Effect of diabetes Albusac JM.  Insulin in the brain: its pathophysi-
mellitus on the permeability of the blood-brain bar- ological implications for states related with central
rier to insulin. Peptides. 1997;18(10):1577–84. insulin resistance, type 2 diabetes and Alzheimer’s
20. Banks WA, Jaspan JB, Kastin AJ.  Selective, disease. Front Endocrinol. 2014;5:161.
physiological transport of insulin across the 34. Bomfim TR, Forny-Germano L, Sathler LB, Brito-­
blood-brain barrier: novel demonstration by Moreira J, Houzel JC, Decker H, Silverman MA,
species-specific radioimmunoassays. Peptides. Kazi H, Melo HM, McClean PL, Holscher C, Arnold
1997;18(8):1257–62. SE, Talbot K, Klein WL, Munoz DP, Ferreira ST,
21. Barghorn S, Zheng-Fischhofer Q, Ackmann De Felice FG.  An anti-diabetes agent protects the
M, Biernat J, von Bergen M, Mandelkow EM, mouse brain from defective insulin signaling caused
Mandelkow E.  Structure, microtubule interactions, by Alzheimer’s disease- associated Abeta oligomers.
and paired helical filament aggregation by tau J Clin Invest. 2012;122(4):1339–53.
mutants of frontotemporal dementias. Biochemistry. 35. Born J, Lange T, Kern W, McGregor GP, Bickel
2000;39(38):11714–21. U, Fehm HL.  Sniffing neuropeptides: a transna-
22. Barini E, Antico O, Zhao Y, Asta F, Tucci V, Catelani sal approach to the human brain. Nat Neurosci.
T, Marotta R, Xu H, Gasparini L.  Metformin pro- 2002;5(6):514–6.
motes tau aggregation and exacerbates abnor- 36. Bozeat S, Gregory CA, Ralph MA, Hodges
mal behavior in a mouse model of tauopathy. Mol JR.  Which neuropsychiatric and behavioural fea-
Neurodegener. 2016;11:16. tures distinguish frontal and temporal variants of
23. Barron AM, Rosario ER, Elteriefi R, Pike CJ. Sex-­ frontotemporal dementia from Alzheimer’s disease?
specific effects of high fat diet on indices of meta- J Neurol Neurosurg Psychiatry. 2000;69(2):178–86.
bolic syndrome in 3xTg-AD mice: implications for 37. Braak H, Braak E.  Neuropathological stageing
Alzheimer’s disease. PLoS One. 2013;8(10):e78554. of Alzheimer-related changes. Acta Neuropathol.
24. Barthel A, Schmoll D.  Novel concepts in insulin 1991;82(4):239–59.
regulation of hepatic gluconeogenesis. Am J Physiol 38. Brands AM, Biessels GJ, de Haan EH, Kappelle LJ,
Endocrinol Metab. 2003;285(4):E685–92. Kessels RP.  The effects of type 1 diabetes on cog-
25. Becker K, Freude S, Zemva J, Stohr O, Krone W, nitive performance: a meta-analysis. Diabetes Care.
Schubert M.  Chronic peripheral hyperinsulinemia 2005;28(3):726–35.
has no substantial influence on tau phosphorylation 39. Brion JP. Immunological demonstration of tau pro-
in vivo. Neurosci Lett. 2012;516(2):306–10. tein in neurofibrillary tangles of Alzheimer’s dis-
26. Benedict C, Hallschmid M, Hatke A, Schultes B, ease. J Alzheimers Dis. 2006;9(3 Suppl):177–85.
Fehm HL, Born J, Kern W. Intranasal insulin improves 40. Brownlow ML, Benner L, D’Agostino D, Gordon
memory in humans. Psychoneuroendocrinology. MN, Morgan D.  Ketogenic diet improves motor
2004;29(10):1326–34. performance but not cognition in two mouse
27. Benedict C, Kern W, Schultes B, Born J, Hallschmid models of Alzheimer’s pathology. PLoS One.
M.  Differential sensitivity of men and women 2013;8(9):e75713.
to anorexigenic and memory-improving effects 41. Brownlow ML, Joly-Amado A, Azam S, Elza M,
of intranasal insulin. J  Clin Endocrinol Metab. Selenica ML, Pappas C, Small B, Engelman R,
2008;93(4):1339–44. Gordon MN, Morgan D.  Partial rescue of mem-
28. Berriman J, Serpell LC, Oberg KA, Fink AL, Goedert ory deficits induced by calorie restriction in a
M, Crowther RA.  Tau filaments from human brain mouse model of tau deposition. Behav Brain Res.
and from in  vitro assembly of recombinant protein 2014;271:79–88.
show cross-beta structure. Proc Natl Acad Sci U S 42. Bruce DG, Davis WA, Casey GP, Starkstein SE,
A. 2003;100(15):9034–8. Clarnette RM, Almeida OP, Davis TM. Predictors of
21  Tau, Diabetes and Insulin 275

cognitive decline in older individuals with diabetes. 56. Clodfelder-Miller BJ, Zmijewska AA, Johnson GV,
Diabetes Care. 2008;31(11):2103–7. Jope RS. Tau is hyperphosphorylated at multiple sites
43. Brüning JC, Gautam D, Burks DJ, Gillette J, in mouse brain in vivo after streptozotocin-induced
Schubert M, Orban PC, Klein R, Krone W, Muller-­ insulin deficiency. Diabetes. 2006;55(12):3320–5.
Wieland D, Kahn CR. Role of brain insulin receptor 57. Combettes-Souverain M, Issad T. Molecular basis of
in control of body weight and reproduction. Science. insulin action. Diabetes Metab. 1998;24(6):477–89.
2000;289(5487):2122–5. 58. Craft S, Asthana S, Newcomer JW, Wilkinson CW,
44. Bucht G, Adolfsson R, Lithner F, Winblad Matos IT, Baker LD, Cherrier M, Lofgreen C,
B.  Changes in blood glucose and insulin secretion Latendresse S, Petrova A, Plymate S, Raskind M,
in patients with senile dementia of Alzheimer type. Grimwood K, Veith RC.  Enhancement of mem-
Acta Med Scand. 1983;213(5):387–92. ory in Alzheimer disease with insulin and soma-
45. Burcelin R, Thorens B, Glauser M, Gaillard RC, tostatin, but not glucose. Arch Gen Psychiatry.
Pralong FP. Gonadotropin-releasing hormone secre- 1999;56(12):1135–40.
tion from hypothalamic neurons: stimulation by 59. Craft S, Baker LD, Montine TJ, Minoshima S,
insulin and potentiation by leptin. Endocrinology. Watson GS, Claxton A, Arbuckle M, Callaghan
2003;144(10):4484–91. M, Tsai E, Plymate SR, Green PS, Leverenz J,
46. Caillet-Boudin ML, Buee L, Sergeant N, Lefebvre Cross D, Gerton B. Intranasal insulin therapy for
B.  Regulation of human MAPT gene expression. Alzheimer disease and amnestic mild cognitive
Mol Neurodegener. 2015;10:28. impairment: a pilot clinical trial. Arch Neurol.
47. Calvo-Ochoa E, Hernandez-Ortega K, Ferrera P, 2012;69(1):29–38.
Morimoto S, Arias C.  Short-term high-fat-and-­ 60. Craft S, Newcomer J, Kanne S, Dagogo-Jack S,
fructose feeding produces insulin signaling altera- Cryer P, Sheline Y, Luby J, Dagogo-Jack A, Alderson
tions accompanied by neurite and synaptic reduction A. Memory improvement following induced hyper-
and astroglial activation in the rat hippocampus. insulinemia in Alzheimer’s disease. Neurobiol
J Cereb Blood Flow Metab. 2014;34(6):1001–8. Aging. 1996;17(1):123–30.
48. Cassimeris L, Spittle C.  Regulation of 61. Craft S, Peskind E, Schwartz MW, Schellenberg
microtubule-­ associated proteins. Int Rev Cytol. GD, Raskind M, Porte D Jr. Cerebrospinal fluid and
2001;210:163–226. plasma insulin levels in Alzheimer’s disease: rela-
49. Chang AM, Halter JB. Aging and insulin secretion. tionship to severity of dementia and apolipoprotein
Am J Phys Endocrinol Metab. 2003;284(1):E7–12. E genotype. Neurology. 1998;50(1):164–8.
50. Chen S, Li B, Grundke-Iqbal I, Iqbal K.  I1PP2A 62. Dai CL, Tung YC, Liu F, Gong CX, Iqbal K.  Tau
affects tau phosphorylation via association with the passive immunization inhibits not only tau but also
catalytic subunit of protein phosphatase 2A. J Biol Abeta pathology. Alzheimers Res Ther. 2017;9(1):1.
Chem. 2008;283(16):10513–21. 63. de Bresser J, Tiehuis AM, van den Berg E,
51. Chen Y, Run X, Liang Z, Zhao Y, Dai CL, Iqbal Reijmer YD, Jongen C, Kappelle LJ, Mali WP,
K, Liu F, Gong CX.  Intranasal insulin prevents Viergever MA, Biessels GJ.  Progression of cere-
anesthesia-­induced hyperphosphorylation of tau in bral atrophy and white matter hyperintensities
3xTg-AD mice. Front Aging Neurosci. 2014;6:100. in patients with type 2 diabetes. Diabetes Care.
52. Chen Y, Zhao Y, Dai CL, Liang Z, Run X, Iqbal K, 2010;33(6):1309–14.
Liu F, Gong CX. Intranasal insulin restores insulin 64. DeFronzo RA. Pathogenesis of type 2 diabetes mel-
signaling, increases synaptic proteins, and reduces litus. Med Clin North Am. 2004;88(4):787–835, ix.
Abeta level and microglia activation in the brains of 65. Defronzo RA.  Banting lecture. From the triumvi-
3xTg-AD mice. Exp Neurol. 2014;261:610–9. rate to the ominous octet: a new paradigm for the
53. Cheng CM, Tseng V, Wang J, Wang D, Matyakhina treatment of type 2 diabetes mellitus. Diabetes.
L, Bondy CA.  Tau is hyperphosphorylated 2009;58(4):773–95.
in the insulin-­ like growth factor-I null brain. 66. DeFronzo RA, Jacot E, et al. The effect of insu-
Endocrinology. 2005;146(12):5086–91. lin on the disposal of intravenous glucose.
54. Chung SH.  Aberrant phosphorylation in the Results from indirect calorimetry and hepatic
pathogenesis of Alzheimer’s disease. BMB Rep. and femoral venous catheterization. Diabetes.
2009;42(8):467–74. 1981;30(12):1000–7.
55. Clarke JR, Lyra ESNM, Figueiredo CP, Frozza RL, 67. Devaskar SU, Giddings SJ, Rajakumar PA, Carnaghi
Ledo JH, Beckman D, Katashima CK, Razolli D, LR, Menon RK, Zahm DS. Insulin gene expression
Carvalho BM, Frazao R, Silveira MA, Ribeiro FC, and insulin synthesis in mammalian neuronal cells.
Bomfim TR, Neves FS, Klein WL, Medeiros R, J Biol Chem. 1994;269(11):8445–54.
LaFerla FM, Carvalheira JB, Saad MJ, Munoz DP, 68. Domise M, Didier S, Marinangeli C, Zhao H,
Velloso LA, Ferreira ST, De Felice FG. Alzheimer-­ Chandakkar P, Buee L, Viollet B, Davies P,
associated Abeta oligomers impact the central Marambaud P, Vingtdeux V. AMP-activated ­protein
nervous system to induce peripheral metabolic kinase modulates tau phosphorylation and tau
deregulation. EMBO Mol Med. 2015;7(2):190–210. pathology in vivo. Sci Rep. 2016;6:26758.
276 M. Gratuze et al.

69. Dorn A, Bernstein HG, Rinne A, Ziegler M, Hahn in normal weight and overweight individuals:
HJ, Ansorge S. Insulin- and glucagonlike peptides in a meta-analysis of test meal studies. Br J  Nutr.
the brain. Anat Rec. 1983;207(1):69–77. 2007;98(1):17–25.
70. Doyle ME, Egan JM. Pharmacological agents that 83. Fox MA, Chen RS, Holmes CS.  Gender differ-
directly modulate insulin secretion. Pharmacol Rev. ences in memory and learning in children with
2003;55(1):105–31. insulin-dependent diabetes mellitus (IDDM) over
71. Duffy KR, Pardridge WM. Blood-brain barrier tran- a 4-year follow-up interval. J  Pediatr Psychol.
scytosis of insulin in developing rabbits. Brain Res. 2003;28(8):569–78.
1987;420(1):32–8. 84. Freude S, Plum L, Schnitker J, Leeser U, Udelhoven
72. Dujardin S, Colin M, Buee L.  Invited review: ani- M, Krone W, Bruning JC, Schubert M.  Peripheral
mal models of tauopathies and their implications hyperinsulinemia promotes tau phosphorylation
for research/translation into the clinic. Neuropathol in vivo. Diabetes. 2005;54(12):3343–8.
Appl Neurobiol. 2015;41(1):59–80. 85. Fu Z, Gilbert ER, Liu D.  Regulation of insu-
73. Duyckaerts C, Hauw JJ.  Diagnosis and staging of lin synthesis and secretion and pancreatic Beta-­
Alzheimer disease. Neurobiol Aging. 1997;18(4 cell dysfunction in diabetes. Curr Diabetes Rev.
Suppl):S33–42. 2013;9(1):25–53.
74. El Khoury NB, Gratuze M, Papon MA, Bretteville 86. Fujisawa Y, Sasaki K, Akiyama K. Increased insulin
A, Planel E. Insulin dysfunction and Tau pathology. levels after OGTT load in peripheral blood and cere-
Front Cell Neurosci. 2014;8:22. brospinal fluid of patients with dementia of Alzheimer
75. El Khoury NB, Gratuze M, Petry F, Papon MA, type. Biol Psychiatry. 1991;30(12):1219–28.
Julien C, Marcouiller F, Morin F, Nicholls SB, Calon 87. Gadsby R, Barker P, Sinclair A. People living with
F, Hebert SS, Marette A, Planel E.  Hypothermia diabetes resident in nursing homes – assessing lev-
mediates age-dependent increase of tau phosphory- els of disability and nursing needs. Diabet Med.
lation in db/db mice. Neurobiol Dis. 2016;88:55–65. 2011;28(7):778–80.
76. El Messari S, Leloup C, Quignon M, Brisorgueil 88. Gambassi G, Landi F, Lapane KL, Sgadari A, Mor
MJ, Penicaud L, Arluison M. Immunocytochemical V, Bernabei R.  Predictors of mortality in patients
localization of the insulin-responsive glucose trans- with Alzheimer’s disease living in nursing homes.
porter 4 (Glut4) in the rat central nervous system. J Neurol Neurosurg Psychiatry. 1999;67(1):59–65.
J Comp Neurol. 1998;399(4):492–512. 89. Garcia-Casares N, Berthier ML, Jorge RE, Gonzalez-­
77. Escribano L, Simon AM, Perez-Mediavilla Alegre P, Gutierrez Cardo A, Rioja Villodres J,
A, Salazar-Colocho P, Del Rio J, Frechilla Acion L, Ariza Corbo MJ, Nabrozidis A, Garcia-­
D. Rosiglitazone reverses memory decline and hip- Arnes JA, Gonzalez-Santos P.  Structural and func-
pocampal glucocorticoid receptor down-regulation tional brain changes in middle-aged type 2 diabetic
in an Alzheimer’s disease mouse model. Biochem patients: a cross-sectional study. J  Alzheimers Dis.
Biophys Res Commun. 2009;379(2):406–10. 2014;40(2):375–86.
78. Espeland MA, Bryan RN, Goveas JS, Robinson 90. Gaudieri PA, Chen R, Greer TF, Holmes
JG, Siddiqui MS, Liu S, Hogan PE, Casanova R, CS.  Cognitive function in children with type
Coker LH, Yaffe K, Masaki K, Rossom R, Resnick 1 diabetes: a meta-analysis. Diabetes Care.
SM. Influence of type 2 diabetes on brain volumes 2008;31(9):1892–7.
and changes in brain volumes: results from the 91. Gejl M, Gjedde A, Egefjord L, Moller A, Hansen
Women’s Health Initiative Magnetic Resonance SB, Vang K, Rodell A, Braendgaard H, Gottrup
Imaging studies. Diabetes Care. 2013;36(1):90–7. H, Schacht A, Moller N, Brock B, Rungby J.  In
79. Faxen-Irving G, Basun H, Cederholm T. Nutritional Alzheimer’s disease, 6-month treatment with GLP-1
and cognitive relationships and long-term mortal- analog prevents decline of brain glucose metabo-
ity in patients with various dementia disorders. Age lism: randomized, placebo-controlled, double-blind
Ageing. 2005;34(2):136–41. clinical trial. Front Aging Neurosci. 2016;8:108.
80. Festa A, Williams K, D’Agostino R Jr, Wagenknecht 92. Ghasemi R, Haeri A, Dargahi L, Mohamed Z,
LE, Haffner SM.  The natural course of beta-cell Ahmadiani A.  Insulin in the brain: sources,
function in nondiabetic and diabetic individuals: the localization and functions. Mol Neurobiol.
Insulin Resistance Atherosclerosis Study. Diabetes. 2013;47(1):145–71.
2006;55(4):1114–20. 93. Gil-Bea FJ, Solas M, Solomon A, Mugueta C,
81. Fischer D, Mukrasch MD, von Bergen M, Klos-­ Winblad B, Kivipelto M, Ramirez MJ, Cedazo-­
Witkowska A, Biernat J, Griesinger C, Mandelkow Minguez A. Insulin levels are decreased in the cere-
E, Zweckstetter M.  Structural and microtubule brospinal fluid of women with prodomal Alzheimer’s
binding properties of tau mutants of frontotemporal disease. J Alzheimers Dis. 2010;22(2):405–13.
dementias. Biochemistry. 2007;46(10):2574–82. 94. Gillette-Guyonnet S, Cortes F, Cantet C, Vellas
82. Flint A, Gregersen NT, Gluud LL, Moller BK, Raben B. Long-term cholinergic treatment is not associated
A, Tetens I, Verdich C, Astrup A.  Associations with greater risk of weight loss during Alzheimer’s
between postprandial insulin and blood glucose disease: data from the French REAL.FR cohort.
responses, appetite sensations and energy intake J Nutr Health Aging. 2005;9(2):69–73.
21  Tau, Diabetes and Insulin 277

95. Gillette-Guyonnet S, Nourhashemi F, Andrieu S, tau hyperphosphorylation in hTau mice, a model of

de Glisezinski I, Ousset PJ, Riviere D, Albarede Alzheimer’s disease-like tau pathology. Sci Rep.
JL, Vellas B. Weight loss in Alzheimer disease. Am 2017;7:46359.
J Clin Nutr. 2000;71(2):637s–42s. 107. Gray SM, Aylor KW, Barrett EJ.  Unravelling the
96. Gillette GS, Abellan VK, Alix E, Andrieu S, Belmin regulation of insulin transport across the brain endo-
J, Berrut G, Bonnefoy M, Brocker P, Constans T, thelial cell. Diabetologia. 2017;60(8):1512–21.
Ferry M, Ghisolfi-Marque A, Girard L, Gonthier 108. Greco SJ, Bryan KJ, Sarkar S, Zhu X, Smith MA,
R, Guerin O, Hervy MP, Jouanny P, Laurain MC, Ashford JW, Johnston JM, Tezapsidis N, Casadesus
Lechowski L, Nourhashemi F, Raynaud-Simon A, G.  Chronic leptin supplementation ameliorates
Ritz P, Roche J, Rolland Y, Salva T, Vellas B. IANA pathology and improves cognitive performance in
(International Academy on Nutrition and Aging) a transgenic mouse model of Alzheimer’s disease.
Expert Group: weight loss and Alzheimer’s disease. J  Alzheimers Dis. 2009; https://doi.org/10.3233/
J Nutr Health Aging. 2007;11(1):38–48. JAD-2009-1308.
97. Goedert M, Spillantini MG, Jakes R, Rutherford 109. Grillo CA, Piroli GG, Lawrence RC, Wrighten SA,
D, Crowther RA.  Multiple isoforms of human Green AJ, Wilson SP, Sakai RR, Kelly SJ, Wilson
microtubule-­associated protein tau: sequences and MA, Mott DD, Reagan LP.  Hippocampal insulin
localization in neurofibrillary tangles of Alzheimer’s resistance impairs spatial learning and synaptic plas-
disease. Neuron. 1989;3(4):519–26. ticity. Diabetes. 2015;64(11):3927–36.
98. Gold M, Alderton C, Zvartau-Hind M, Egginton 110. Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M,
S, Saunders AM, Irizarry M, Craft S, Landreth G, Wisniewski HM, Binder LI. Abnormal phosphoryla-
Linnamagi U, Sawchak S.  Rosiglitazone mono- tion of the microtubule-associated protein tau (tau)
therapy in mild-to-moderate Alzheimer’s disease: in Alzheimer cytoskeletal pathology. Proc Natl Acad
results from a randomized, double-blind, placebo-­ Sci U S A. 1986;83(13):4913–7.
controlled phase III study. Dement Geriatr Cogn 111. Guerin O, Andrieu S, Schneider SM, Milano M,
Disord. 2010;30(2):131–46. Boulahssass R, Brocker P, Vellas B.  Different
99. Gong CX, Iqbal K.  Hyperphosphorylation of modes of weight loss in Alzheimer disease: a pro-
microtubule-­ associated protein tau: a promising spective study of 395 patients. Am J  Clin Nutr.
therapeutic target for Alzheimer disease. Curr Med 2005;82(2):435–41.
Chem. 2008;15(23):2321–8. 112. Guerin O, Soto ME, Brocker P, Robert PH, Benoit
100. Gong CX, Lidsky T, Wegiel J, Zuck L, Grundke-­ M, Vellas B.  Nutritional status assessment dur-
Iqbal I, Iqbal K.  Phosphorylation of microtubule-­ ing Alzheimer’s disease: results after one year (the
associated protein tau is regulated by protein REAL French Study Group). J  Nutr Health Aging.
phosphatase 2A in mammalian brain. Implications 2005;9(2):81–4.
for neurofibrillary degeneration in Alzheimer’s dis- 113. Guo Z, Chen Y, Mao YF, Zheng T, Jiang Y, Yan Y,
ease. J Biol Chem. 2000;275(8):5535–44. Yin X, Zhang B.  Long-term treatment with intra-
101. Gong CX, Liu F, Wu G, Rossie S, Wegiel J, Li L, nasal insulin ameliorates cognitive impairment, tau
Grundke-Iqbal I, Iqbal K.  Dephosphorylation of hyperphosphorylation, and microglial activation in
microtubule-associated protein tau by protein phos- a streptozotocin-induced Alzheimer’s rat model. Sci
phatase 5. J Neurochem. 2004;88(2):298–310. Rep. 2017;7:45971.
102. Gong CX, Shaikh S, Wang JZ, Zaidi T, Grundke-­ 114. Hallschmid M, Higgs S, Thienel M, Ott V,

Iqbal I, Iqbal K. Phosphatase activity toward abnor- Lehnert H.  Postprandial administration of intra-
mally phosphorylated tau: decrease in Alzheimer nasal insulin intensifies satiety and reduces
disease brain. J Neurochem. 1995;65(2):732–8. intake of palatable snacks in women. Diabetes.
103. Gong CX, Singh TJ, Grundke-Iqbal I, Iqbal 2012;61(4):782–9.
K.  Phosphoprotein phosphatase activities 115. Harper DG, Stopa EG, Kuo-Leblanc V, McKee
in Alzheimer disease brain. J  Neurochem. AC, Asayama K, Volicer L, Kowall N, Satlin
1993;61(3):921–7. A. Dorsomedial SCN neuronal subpopulations sub-
104. Gratuze M, El Khoury NB, Turgeon A, Julien C, serve different functions in human dementia. Brain.
Marcouiller F, Morin F, Whittington RA, Marette A, 2008;131(Pt 6):1609–17.
Calon F, Planel E. Tau hyperphosphorylation in the 116. Harrington C, Sawchak S, Chiang C, Davies
brain of ob/ob mice is due to hypothermia: impor- J, Donovan C, Saunders AM, Irizarry M,
tance of thermoregulation in linking diabetes and Jeter B, Zvartau-Hind M, van Dyck CH, Gold
Alzheimer’s disease. Neurobiol Dis. 2017;98:1–8. M.  Rosiglitazone does not improve cognition or
105. Gratuze M, Julien J, Morin F, Calon F, Hebert SS, global function when used as adjunctive therapy to
Marette A, Planel E. High-fat, high-sugar, and high-­ AChE inhibitors in mild-to-moderate Alzheimer’s
cholesterol consumption does not impact tau patho- disease: two phase 3 studies. Curr Alzheimer Res.
genesis in a mouse model of Alzheimer’s disease-like 2011;8(5):592–606.
tau pathology. Neurobiol Aging. 2016;47:71–3. 117.
Havrankova J, Roth J, Brownstein MJ.
106. Gratuze M, Julien J, Petry FR, Morin F, Planel Concentrations of insulin and insulin receptors
E. Insulin deprivation induces PP2A inhibition and in the brain are independent of peripheral insulin
278 M. Gratuze et al.

levels. Studies of obese and streptozotocin-treated 133. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A,
rodents. J Clin Invest. 1979;64(2):636–42. van Eersel J, Wolfing H, Chieng BC, Christie MJ,
118. Havrankova J, Schmechel D, Roth J, Brownstein Napier IA, Eckert A, Staufenbiel M, Hardeman E,
M.  Identification of insulin in rat brain. Proc Natl Gotz J. Dendritic function of tau mediates amyloid-­
Acad Sci U S A. 1978;75(11):5737–41. beta toxicity in Alzheimer’s disease mouse models.
119. Heidenreich KA, Zahniser NR, Berhanu P, Cell. 2010;142(3):387–97.
Brandenburg D, Olefsky JM. Structural differences 134. Janson J, Laedtke T, Parisi JE, O’Brien P, Petersen
between insulin receptors in the brain and peripheral RC, Butler PC. Increased risk of type 2 diabetes in
target tissues. J Biol Chem. 1983;258(14):8527–30. Alzheimer disease. Diabetes. 2004;53(2):474–81.
120. Heitner J, Dickson D.  Diabetics do not have 135. Jauch-Chara K, Friedrich A, Rezmer M, Melchert
increased Alzheimer-type pathology compared with UH, Scholand-Engler HG, Hallschmid M, Oltmanns
age-matched control subjects. A retrospective post- KM.  Intranasal insulin suppresses food intake via
mortem immunocytochemical and histofluorescent enhancement of brain energy levels in humans.
study. Neurology. 1997;49(5):1306–11. Diabetes. 2012;61(9):2261–8.
121. Hershey T, Bhargava N, Sadler M, White NH, Craft 136. Jeon BT, Jeong EA, Shin HJ, Lee Y, Lee DH, Kim HJ,
S. Conventional versus intensive diabetes therapy in Kang SS, Cho GJ, Choi WS, Roh GS.  Resveratrol
children with type 1 diabetes: effects on memory and attenuates obesity-associated peripheral and central
motor speed. Diabetes Care. 1999;22(8):1318–24. inflammation and improves memory deficit in mice
122. Henquin JC. Regulation of insulin secretion: a mat- fed a high-fat diet. Diabetes. 2012;61(6):1444–54.
ter of phase control and amplitude modulation. 137. Jimenez-Palomares M, Ramos-Rodriguez JJ, Lopez-­
Diabetologia. 2009;52(5):739–51. Acosta JF, Pacheco-Herrero M, Lechuga-Sancho
123. Hiltunen M, Khandelwal VK, Yaluri N, Tiilikainen AM, Perdomo G, Garcia-Alloza M, Cozar-Castellano
T, Tusa M, Koivisto H, Krzisch M, Vepsalainen S, I.  Increased Abeta production prompts the onset
Makinen P, Kemppainen S, Miettinen P, Haapasalo of glucose intolerance and insulin resistance. Am
A, Soininen H, Laakso M, Tanila H. Contribution of J Phys Endocrinol Metab. 2012;302(11):E1373–80.
genetic and dietary insulin resistance to Alzheimer 138. Jolivalt CG, Lee CA, Beiswenger KK, Smith JL,
phenotype in APP/PS1 transgenic mice. J Cell Mol Orlov M, Torrance MA, Masliah E. Defective insu-
Med. 2012;16(6):1206–22. lin signaling pathway and increased glycogen syn-
124. Hofman A, Ott A, et al. Atherosclerosis, apo- thase kinase-3 activity in the brain of diabetic mice:
lipoprotein E, and prevalence of dementia and parallels with Alzheimer’s disease and correction by
Alzheimer’s disease in the Rotterdam Study. Lancet. insulin. J Neurosci Res. 2008;86(15):3265–74.
1997;349(9046):151–4. 139. Joly-Amado A, Serraneau KS, Brownlow M, Marin
125. Holgado-Madruga M, Emlet DR, Moscatello DK, de Evsikova C, Speakman JR, Gordon MN, Morgan
Godwin AK, Wong AJ.  A Grb2-associated dock- D. Metabolic changes over the course of aging in a
ing protein in EGF- and insulin-receptor signalling. mouse model of tau deposition. Neurobiol Aging.
Nature. 1996;379(6565):560–4. 2016;44:62–73.
126. Hong M, Lee VM. Insulin and insulin-like growth fac- 140. Joost HG. Structural and functional heterogeneity of
tor-­1 regulate tau phosphorylation in cultured human insulin receptors. Cell Signal. 1995;7(2):85–91.
neurons. J Biol Chem. 1997;272(31):19547–53. 141. Julien C, Tremblay C, Phivilay A, Berthiaume
127. Hoyer S.  Is sporadic Alzheimer disease the brain L, Emond V, Julien P, Calon F.  High-fat diet
type of non-insulin dependent diabetes melli- aggravates amyloid-beta and tau pathologies in
tus? A challenging hypothesis. J  Neural Transm. the 3xTg-AD mouse model. Neurobiol Aging.
1998;105(4–5):415–22. 2010;31(9):1516–31.
128. Hoyer S, Prem L, Sorbi S, Amaducci L. Stimulation 142. Jung HJ, Park SS, Mok JO, Lee TK, Park CS,
of glycolytic key enzymes in cerebral cortex by insu- Park SA.  Increased expression of three-repeat iso-
lin. Neuroreport. 1993;4(7):991–3. forms of tau contributes to tau pathology in a rat
129. Hu FB.  Globalization of diabetes: the role model of chronic type 2 diabetes. Exp Neurol.
of diet, lifestyle, and genes. Diabetes Care. 2011;228(2):232–41.
2011;34(6):1249–57. 143. Kang JM, Lee SY, Seo S, Jeong HJ, Woo SH, Lee
130. Huo L, Harding JL, Peeters A, Shaw JE, Magliano H, Lee YB, Yeon BK, Shin DH, Park KH, Kang H,
DJ. Life expectancy of type 1 diabetic patients dur- Okamura N, Furumoto S, Yanai K, Villemagne VL,
ing 1997–2010: a national Australian registry-based Seong JK, Na DL, Ido T, Cho J, Lee KM, Noh Y. Tau
cohort study. Diabetologia. 2016;59(6):1177–85. positron emission tomography using [18F]THK5351
131. Inelmen EM, Sergi G, Coin A, Girardi A, Manzato and cerebral glucose hypometabolism in Alzheimer’s
E. An open-ended question: Alzheimer’s disease and disease. Neurobiol Aging. 2017;59:210–9.
involuntary weight loss: which comes first? Aging 144. Kar S, Chabot JG, Quirion R. Quantitative autoradio-
Clin Exp Res. 2010;22(3):192–7. graphic localization of [125I]insulin-like growth fac-
132. Iqbal K, Liu F, Gong CX, Alonso Adel C, Grundke-­ tor I, [125I]insulin-like growth factor II, and [125I]
Iqbal I. Mechanisms of tau-induced neurodegenera- insulin receptor binding sites in developing and adult
tion. Acta Neuropathol. 2009;118(1):53–69. rat brain. J Comp Neurol. 1993;333(3):375–97.
21  Tau, Diabetes and Insulin 279

145. Kasus-Jacobi A, Perdereau D, Tartare-Deckert S, Van sity and voluntary exercise in a tauopathy mouse
Obberghen E, Girard J, Burnol AF.  Evidence for a model: implications of persistent hyperleptinemia
direct interaction between insulin receptor substrate-1 and enhanced astrocytic leptin receptor expression.
and Shc. J Biol Chem. 1997;272(27):17166–70. Neurobiol Dis. 2014;71:180–92.
146. Ke YD, Delerue F, Gladbach A, Gotz J, Ittner 159. Konner AC, Hess S, Tovar S, Mesaros A, Sanchez-­
LM.  Experimental diabetes mellitus exacerbates Lasheras C, Evers N, Verhagen LA, Bronneke HS,
tau pathology in a transgenic mouse model of Kleinridders A, Hampel B, Kloppenburg P, Bruning
Alzheimer’s disease. PLoS One. 2009;4(11):e7917. JC.  Role for insulin signaling in catecholaminer-
147. Kern W, Born J, Schreiber H, Fehm HL.  Central gic neurons in control of energy homeostasis. Cell
nervous system effects of intranasally adminis- Metab. 2011;13(6):720–8.
tered insulin during euglycemia in men. Diabetes. 160. Konner AC, Janoschek R, Plum L, Jordan SD,
1999;48(3):557–63. Rother E, Ma X, Xu C, Enriori P, Hampel B, Barsh
148. Kern W, Peters A, Fruehwald-Schultes B, GS, Kahn CR, Cowley MA, Ashcroft FM, Bruning
Deininger E, Born J, Fehm HL.  Improving influ- JC.  Insulin action in AgRP-expressing neurons is
ence of insulin on cognitive functions in humans. required for suppression of hepatic glucose produc-
Neuroendocrinology. 2001;74(4):270–80. tion. Cell Metab. 2007;5(6):438–49.
149. Kilgour RD, Williams PA.  Effects of diabetes and 161. Kooistra M, Geerlings MI, Mali WP, Vincken KL,
food deprivation on shivering activity during pro- van der Graaf Y, Biessels GJ.  Diabetes mellitus
gressive hypothermia in the rat. Comp Biochem and progression of vascular brain lesions and brain
Physiol A Physiol. 1996;114(2):159–65. atrophy in patients with symptomatic atheroscle-
150.
Kim B, Backus C, Oh S, Feldman EL. rotic disease. The SMART-MR study. J Neurol Sci.
Hyperglycemia-­ induced tau cleavage in  vitro 2013;332(1–2):69–74.
and in  vivo: a possible link between diabe- 162. Kovacs P, Parlow AF, Karkanias GB.  Effect of
tes and Alzheimer’s disease. J  Alzheimers Dis. centrally administered insulin on gonadotropin-­
2013;34(3):727–39. releasing hormone neuron activity and lutein-
151. Kim B, Backus C, Oh S, Hayes JM, Feldman izing hormone surge in the diabetic female rat.
EL.  Increased tau phosphorylation and cleavage Neuroendocrinology. 2002;76(6):357–65.
in mouse models of type 1 and type 2 diabetes. 163. Kraegen EW, Chisholm DJ, et al. The gastrointes-
Endocrinology. 2009;150(12):5294–301. tinal stimulus to insulin release. II. A dual action of
152. Kim B, Figueroa-Romero C, Pacut C, Backus secretin. J Clin Invest. 1970;49(3):524–9.
C, Feldman EL.  Insulin resistance prevents 164. Krishnan N, Krishnan K, Connors CR, Choy
AMPK-induced Tau dephosphorylation through MS, Page R, Peti W, Van Aelst L, Shea SD,
Akt-mediated increase in AMPKSer-485 phosphor- Tonks NK.  PTP1B inhibition suggests a thera-
ylation. J Biol Chem. 2015;290(31):19146–57. peutic strategy for Rett syndrome. J  Clin Invest.
153. Kimura T, Whitcomb DJ, Jo J, Regan P, Piers T, Heo 2015;125(8):3163–77.
S, Brown C, Hashikawa T, Murayama M, Seok H, 165. Kuan YC, Huang KW, Lin CL, Hu CJ, Kao
Sotiropoulos I, Kim E, Collingridge GL, Takashima CH. Effects of metformin exposure on neurodegen-
A, Cho K.  Microtubule-associated protein tau is erative diseases in elderly patients with type 2 dia-
essential for long-term depression in the hippo- betes mellitus. Prog Neuro-Psychopharmacol Biol
campus. Philos Trans R Soc Lond Ser B Biol Sci. Psychiatry. 2017;79(Pt B):77–83.
2014;369(1633):20130144. 166. Landin K, Blennow K, Wallin A, Gottfries CG. Low
154. King AJ.  The use of animal models in diabetes blood pressure and blood glucose levels in
research. Br J Pharmacol. 2012;166(3):877–94. Alzheimer’s disease. Evidence for a hypometabolic
155. King ME, Ahuja V, Binder LI, Kuret J.  Ligand-­ disorder? J Intern Med. 1993;233(4):357–63.
dependent tau filament formation: implications for 167.
Leboucher A, Laurent C, Fernandez-Gomez
Alzheimer’s disease progression. Biochemistry. FJ, Burnouf S, Troquier L, Eddarkaoui S,
1999;38(45):14851–9. Demeyer D, Caillierez R, Zommer N, Vallez E,
156. Knight EM, Brown TM, Gumusgoz S, Smith JC, Bantubungi K, Breton C, Pigny P, Buee-Scherrer
Waters EJ, Allan SM, Lawrence CB.  Age-related V, Staels B, Hamdane M, Tailleux A, Buee L,
changes in core body temperature and activity in Blum D.  Detrimental effects of diet-induced obe-
transgenic Alzheimer’s disease (3xTgAD)
triple-­ sity on tau pathology are independent of insu-
mice. Dis Model Mech. 2013;6(1):160–70. lin resistance in tau transgenic mice. Diabetes.
157. Knight EM, Martins IV, Gumusgoz S, Allan SM, 2013;62(5):1681–8.
Lawrence CB.  High-fat diet-induced memory 168. Lebouvier T, Pasquier F, Buee L. Update on tauopa-
impairment in triple-transgenic Alzheimer’s dis- thies. Curr Opin Neurol. 2017;30(6):589–98.
ease (3xTgAD) mice is independent of changes 169. Lee VM, Goedert M, Trojanowski JQ.
in amyloid and tau pathology. Neurobiol Aging. Neurodegenerative tauopathies. Annu Rev
2014;35(8):1821–32. Neurosci. 2001;24:1121–59.
158. Koga S, Kojima A, Ishikawa C, Kuwabara S, Arai 170. Leibson CL, Rocca WA, Hanson VA, Cha R,

K, Yoshiyama Y.  Effects of diet-induced obe- Kokmen E, O’Brien PC, Palumbo PJ.  Risk of
280 M. Gratuze et al.

dementia among persons with diabetes mellitus: replication study and meta-analysis. Curr Alzheimer
a population-based cohort study. Am J Epidemiol. Res. 2016;13(3):223–33.
1997;145(4):301–8. 184. Ma QL, Yang F, Rosario ER, Ubeda OJ, Beech
171. Lesort M, Johnson GV. Insulin-like growth factor-1 W, Gant DJ, Chen PP, Hudspeth B, Chen C, Zhao
and insulin mediate transient site-selective increases Y, Vinters HV, Frautschy SA, Cole GM.  Beta-­
in tau phosphorylation in primary cortical neurons. amyloid oligomers induce phosphorylation of tau
Neuroscience. 2000;99(2):305–16. and inactivation of insulin receptor substrate via
172. Li J, Deng J, Sheng W, Zuo Z. Metformin attenuates c-Jun N-terminal kinase signaling: suppression
Alzheimer’s disease-like neuropathology in obese, by omega-3 fatty acids and curcumin. J  Neurosci.
leptin-resistant mice. Pharmacol Biochem Behav. 2009;29(28):9078–89.
2012;101(4):564–74. 185. Ma QL, Zuo X, Yang F, Ubeda OJ, Gant DJ,
173. Li T, Braunstein KE, Zhang J, Lau A, Sibener Alaverdyan M, Kiosea NC, Nazari S, Chen PP,
L, Deeble C, Wong PC.  The neuritic plaque Nothias F, Chan P, Teng E, Frautschy SA, Cole
facilitates pathological conversion of tau in an GM.  Loss of MAP function leads to hippocampal
Alzheimer’s disease mouse model. Nat Commun. synapse loss and deficits in the Morris water maze
2016;7:12082. with aging. J Neurosci. 2014;34(21):7124–36.
174. Li W, Risacher SL, Huang E, Saykin AJ. Type 2 dia- 186. Manschot SM, Biessels GJ, de Valk H, Algra A,
betes mellitus is associated with brain atrophy and Rutten GE, van der Grond J, Kappelle LJ. Metabolic
hypometabolism in the ADNI cohort. Neurology. and vascular determinants of impaired cognitive
2016;87(6):595–600. performance and abnormalities on brain magnetic
175. Li ZG, Zhang W, Sima AA. Alzheimer-like changes resonance imaging in patients with type 2 diabetes.
in rat models of spontaneous diabetes. Diabetes. Diabetologia. 2007;50(11):2388–97.
2007;56(7):1817–24. 187. Manschot SM, Brands AM, van der Grond J, Kessels
176. Lin A, Northam EA, Rankins D, Werther GA, RP, Algra A, Kappelle LJ, Biessels GJ. Brain mag-
Cameron FJ. Neuropsychological profiles of young netic resonance imaging correlates of impaired cog-
people with type 1 diabetes 12 yr after disease onset. nition in patients with type 2 diabetes. Diabetes.
Pediatr Diabetes. 2010;11(4):235–43. 2006;55(4):1106–13.
1 77.
Liu F, Grundke-Iqbal I, Iqbal K, Gong 188. Marciniak E, Leboucher A, Caron E, Ahmed

CX.  Contributions of protein phosphatases T, Tailleux A, Dumont J, Issad T, Gerhardt E,
PP1, PP2A, PP2B and PP5 to the regula- Pagesy P, Vileno M, Bournonville C, Hamdane M,
tion of tau phosphorylation. Eur J  Neurosci. Bantubungi K, Lancel S, Demeyer D, Eddarkaoui
2005;22(8):1942–50. S, Vallez E, Vieau D, Humez S, Faivre E,
178.
Liu Y, Liu F, Grundke-Iqbal I, Iqbal K, Grenier-Boley B, Outeiro TF, Staels B, Amouyel
Gong CX.  Brain glucose transporters, P, Balschun D, Buee L, Blum D.  Tau deletion
O-GlcNAcylation and phosphorylation of tau in promotes brain insulin resistance. J  Exp Med.
diabetes and Alzheimer’s disease. J  Neurochem. 2017;214(8):2257–69.
2009;111(1):242–9. 189. Marks JL, Porte D Jr, Stahl WL, Baskin
179. Loh K, Zhang L, Brandon A, Wang Q, Begg D, Qi DG.  Localization of insulin receptor mRNA in
Y, Fu M, Kulkarni R, Teo J, Baldock P, Bruning JC, rat brain by in situ hybridization. Endocrinology.
Cooney G, Neely G, Herzog H. Insulin controls food 1990;127(6):3234–6.
intake and energy balance via NPY neurons. Mol 190. Masuda T, Wahlin K, Wan J, Hu J, Maruotti J, Yang
Metab. 2017;6(6):574–84. X, Iacovelli J, Wolkow N, Kist R, Dunaief JL, Qian
1 80. Lowe WL Jr, Boyd FT, Clarke DW, Raizada
J, Zack DJ, Esumi N.  Transcription factor SOX9
MK, Hart C, LeRoith D.  Development of brain plays a key role in the regulation of visual cycle gene
insulin receptors: structural and functional expression in the retinal pigment epithelium. J Biol
studies of insulin receptors from whole brain Chem. 2014;289(18):12908–21.
and primary cell cultures. Endocrinology. 191. Mathers CD, Loncar D. Projections of global mor-
1986;119(1):25–35. tality and burden of disease from 2002 to 2030.
181. Luchsinger JA, Perez T, Chang H, Mehta P, Steffener PLoS Med. 2006;3(11):e442.
J, Pradabhan G, Ichise M, Manly J, Devanand DP, 192. McClean PL, Parthsarathy V, Faivre E, Holscher
Bagiella E.  Metformin in amnestic mild cogni- C. The diabetes drug liraglutide prevents degenera-
tive impairment: results of a pilot randomized pla- tive processes in a mouse model of Alzheimer’s dis-
cebo controlled clinical trial. J  Alzheimers Dis. ease. J Neurosci. 2011;31(17):6587–94.
2016;51(2):501–14. 193. McDuff T, Sumi SM.  Subcortical degeneration in
182. Luchsinger JA, Tang MX, Shea S, Mayeux Alzheimer’s disease. Neurology. 1985;35(1):123–6.
R. Hyperinsulinemia and risk of Alzheimer disease. 194. McKhann G, Drachman D, Folstein M, Katzman
Neurology. 2004;63(7):1187–92. R, Price D, Stadlan EM.  Clinical diagnosis of
183. Ma J, Zhang W, Wang HF, Wang ZX, Jiang T, Tan Alzheimer’s disease: report of the NINCDS-­
MS, Yu JT, Tan L. Peripheral blood Adipokines and ADRDA Work Group under the auspices of
insulin levels in patients with Alzheimer’s disease: a Department of Health and Human Services
21  Tau, Diabetes and Insulin 281

Task Force on Alzheimer’s Disease. Neurology. 206. Niedowicz DM, Reeves VL, Platt TL, Kohler K,
1984;34(7):939–44. Beckett TL, Powell DK, Lee TL, Sexton TR, Song
195. Miklossy J, Qing H, Radenovic A, Kis A, Vileno B, ES, Brewer LD, Latimer CS, Kraner SD, Larson KL,
Laszlo F, Miller L, Martins RN, Waeber G, Mooser Ozcan S, Norris CM, Hersh LB, Porter NM, Wilcock
V, Bosman F, Khalili K, Darbinian N, McGeer DM, Murphy MP. Obesity and diabetes cause cog-
PL.  Beta amyloid and hyperphosphorylated tau nitive dysfunction in the absence of accelerated
deposits in the pancreas in type 2 diabetes. Neurobiol beta-amyloid deposition in a novel murine model
Aging. 2010;31(9):1503–15. of mixed or vascular dementia. Acta Neuropathol
196. Mills GW, Avery PJ, McCarthy MI, Hattersley Commun. 2014;2:64.
AT, Levy JC, Hitman GA, Sampson M, Walker 207. O’Donnell JM, Banyasz T, Kovacs T. Altered thermo-
M. Heritability estimates for beta cell function and regulatory responses to clonidine in streptozotocin-­
features of the insulin resistance syndrome in UK diabetic rats. Br J Pharmacol. 1996;117(5):938–42.
families with an increased susceptibility to type 2 208. O’Reilly JA, Lynch M. Rosiglitazone improves spa-
diabetes. Diabetologia. 2004;47(4):732–8. tial memory and decreases insoluble Abeta(1-­ 42)
197. Mody N, Agouni A, McIlroy GD, Platt B,
in APP/PS1 mice. J  Neuroimmune Pharmacol.
Delibegovic M. Susceptibility to diet-induced obe- 2012;7(1):140–4.
sity and glucose intolerance in the APP (SWE)/ 209. Obici S, Zhang BB, Karkanias G, Rossetti
PSEN1 (A246E) mouse model of Alzheimer’s L.  Hypothalamic insulin signaling is required
disease is associated with increased brain levels for inhibition of glucose production. Nat Med.
of protein tyrosine phosphatase 1B (PTP1B) and 2002;8(12):1376–82.
retinol-binding protein 4 (RBP4), and basal phos- 210. Ott A, Stolk RP, et al. Association of diabetes

phorylation of S6 ribosomal protein. Diabetologia. mellitus and dementia: the Rotterdam Study.
2011;54(8):2143–51. Diabetologia. 1996;39(11):1392–7.
198. Molina JA, Jimenez-Jimenez FJ, Vargas C, Gomez P, 211. Ott A, Stolk RP, van Harskamp F, Pols HA, Hofman
de Bustos F, Gomez-Escalonilla C, Zurdo M, Tallon A, Breteler MM.  Diabetes mellitus and the risk
A, Martinez-Salio A, Porta-Etessam J, Villanueva of dementia: The Rotterdam Study. Neurology.
C, Arenas J.  Cerebrospinal fluid levels of insulin 1999;53(9):1937–42.
in patients with Alzheimer’s disease. Acta Neurol 212. Ouwens DM, van Duinkerken E, Schoonenboom SN,
Scand. 2002;106(6):347–50. Herzfeld de Wiza D, Klein M, van Golen L, Pouwels
199. Moloney AM, Griffin RJ, Timmons S, O’Connor PJ, Barkhof F, Moll AC, Snoek FJ, Teunissen CE,
R, Ravid R, O’Neill C.  Defects in IGF-1 recep- Scheltens P, Diamant M.  Cerebrospinal fluid lev-
tor, insulin receptor and IRS-1/2  in Alzheimer’s els of Alzheimer’s disease biomarkers in middle-­
disease indicate possible resistance to IGF-1 aged patients with type 1 diabetes. Diabetologia.
and insulin signalling. Neurobiol Aging. 2014;57(10):2208–14.
2010;31(2):224–43. 213. Papon MA, El Khoury NB, et al. Deregulation of
200. Moran C, Phan TG, Chen J, Blizzard L, Beare R, protein phosphatase 2A and hyperphosphorylation
Venn A, Munch G, Wood AG, Forbes J, Greenaway of tau protein following onset of diabetes in NOD
TM, Pearson S, Srikanth V. Brain atrophy in type 2 mice. Diabetes. 2013;62(2):609–17.
diabetes: regional distribution and influence on cog- 214. Park SH, Kim JH, Choi KH, Jang YJ, Bae SS, Choi
nition. Diabetes Care. 2013;36(12):4036–42. BT, Shin HK.  Hypercholesterolemia accelerates
201. Morris M, Hamto P, Adame A, Devidze N, amyloid beta-induced cognitive deficits. Int J  Mol
Masliah E, Mucke L.  Age-appropriate cognition Med. 2013;31(3):577–82.
and subtle dopamine-independent motor defi- 215. Pedersen WA, McMillan PJ, Kulstad JJ, Leverenz
cits in aged tau knockout mice. Neurobiol Aging. JB, Craft S, Haynatzki GR. Rosiglitazone attenuates
2013;34(6):1523–9. learning and memory deficits in Tg2576 Alzheimer
202. Moult PR, Harvey J. Hormonal regulation of hippo- mice. Exp Neurol. 2006;199(2):265–73.
campal dendritic morphology and synaptic plastic- 216. Peila R, Rodriguez BL, Launer LJ. Type 2 diabetes,
ity. Cell Adhes Migr. 2008;2(4):269–75. APOE gene, and the risk for dementia and related
203. Mukaetova-Ladinska EB, Abdel-All Z, Mugica ES, pathologies: The Honolulu-Asia Aging Study.
Li M, Craggs LJ, Oakley AE, Honer WG, Kalaria Diabetes. 2002;51(4):1256–62.
RN. Tau proteins in the temporal and frontal cortices 217. Pelosi L, Blumhardt LD. Effects of age on working
in patients with vascular dementia. J  Neuropathol memory: an event-related potential study. Brain Res
Exp Neurol. 2015;74(2):148–57. Cogn Brain Res. 1999;7(3):321–34.
204. Neil HA, Dawson JA, Baker JE.  Risk of hypo- 218. Peng D, Pan X, Cui J, Ren Y, Zhang
thermia in elderly patients with diabetes. Br Med J.  Hyperphosphorylation of tau protein in hippo-
J. 1986;293(6544):416–8. campus of central insulin-resistant rats is associated
205. Newsholme EA, Dimitriadis G.  Integration of bio- with cognitive impairment. Cell Physiol Biochem.
chemical and physiologic effects of insulin on glu- 2013;32(5):1417–25.
cose metabolism. Exp Clin Endocrinol Diabetes. 219. Perry G, Roder H, Nunomura A, Takeda A, Friedlich
2001;109(Suppl 2):S122–34. AL, Zhu X, Raina AK, Holbrook N, Siedlak SL,
282 M. Gratuze et al.

Harris PL, Smith MA. Activation of neuronal extra- 231. Plymate SR, Breitner JC, DeGroodt W, Mehta P,
cellular receptor kinase (ERK) in Alzheimer disease Craft S.  Intranasal insulin improves cognition and
links oxidative stress to abnormal phosphorylation. modulates beta-amyloid in early AD.  Neurology.
Neuroreport. 1999;10(11):2411–5. 2008;70(6):440–8.
2 20. Petrie D, Lung TW, Rawshani A, Palmer AJ, 232. Pollock NJ, Mirra SS, Binder LI, Hansen LA,

Svensson AM, Eliasson B, Clarke P.  Recent Wood JG.  Filamentous aggregates in Pick’s
trends in life expectancy for people with disease, progressive supranuclear palsy, and
type 1 diabetes in Sweden. Diabetologia. Alzheimer’s disease share antigenic determinants
2016;59(6):1167–76. with microtubule-­associated protein, tau. Lancet.
221. Picardi PK, Calegari VC, Prada PO, Moraes JC, 1986;2(8517):1211.
Araujo E, Marcondes MC, Ueno M, Carvalheira 233. Pooler AM, Phillips EC, Lau DH, Noble W,
JB, Velloso LA, Saad MJ.  Reduction of hypotha- Hanger DP.  Physiological release of endogenous
lamic protein tyrosine phosphatase improves insu- tau is stimulated by neuronal activity. EMBO Rep.
lin and leptin resistance in diet-induced obese rats. 2013;14(4):389–94.
Endocrinology. 2008;149(8):3870–80. 234. Posner BI, Kelly PA, Shiu RP, Friesen HG. Studies
222. Piguet O.  Eating disturbance in behavioural-­ of insulin, growth hormone and prolactin binding:
variant frontotemporal dementia. J  Mol Neurosci. tissue distribution, species variation and character-
2011;45(3):589–93. ization. Endocrinology. 1974;95(2):521–31.
223. Piguet O, Petersen A, Lam BYK, Gabery S, Murphy 235. Profenno LA, Porsteinsson AP, Faraone SV.  Meta-­
K, Hodges JR, Halliday GM. Eating and hypothala- analysis of Alzheimer’s disease risk with obesity,
mus changes in behavioral-variant frontotemporal diabetes, and related disorders. Biol Psychiatry.
dementia. Ann Neurol. 2011;69(2):312–9. 2010;67(6):505–12.
224. Pirola L, Johnston AM, Van Obberghen 236. Prokopenko I, Poon W, Magi R, Prasad BR, Salehi
E.  Modulation of insulin action. Diabetologia. SA, Almgren P, Osmark P, Bouatia-Naji N, Wierup
2004;47(2):170–84. N, Fall T, Stancakova A, Barker A, Lagou V, Osmond
225. Piroli GG, Grillo CA, Hoskin EK, Znamensky V, C, Xie W, Lahti J, Jackson AU, Cheng YC, Liu J,
Katz EB, Milner TA, McEwen BS, Charron MJ, O’Connell JR, Blomstedt PA, Fadista J, Alkayyali S,
Reagan LP. Peripheral glucose administration stimu- Dayeh T, Ahlqvist E, Taneera J, Lecoeur C, Kumar
lates the translocation of GLUT8 glucose transporter A, Hansson O, Hansson K, Voight BF, Kang HM,
to the endoplasmic reticulum in the rat hippocam- Levy-Marchal C, Vatin V, Palotie A, Syvanen AC,
pus. J Comp Neurol. 2002;452(2):103–14. Mari A, Weedon MN, Loos RJ, Ong KK, Nilsson
226. Pittman AM, Myers AJ, Abou-Sleiman P, Fung P, Isomaa B, Tuomi T, Wareham NJ, Stumvoll M,
HC, Kaleem M, Marlowe L, Duckworth J, Leung Widen E, Lakka TA, Langenberg C, Tonjes A,
D, Williams D, Kilford L, Thomas N, Morris Rauramaa R, Kuusisto J, Frayling TM, Froguel P,
CM, Dickson D, Wood NW, Hardy J, Lees AJ, de Walker M, Eriksson JG, Ling C, Kovacs P, Ingelsson
Silva R.  Linkage disequilibrium fine mapping and E, McCarthy MI, Shuldiner AR, Silver KD,
haplotype association analysis of the tau gene in Laakso M, Groop L, Lyssenko V. A central role for
progressive supranuclear palsy and corticobasal GRB10 in regulation of islet function in man. PLoS
degeneration. J Med Genet. 2005;42(11):837–46. Genet. 2014;10(4):e1004235.
227. Planel E, Richter KE, Nolan CE, Finley JE, Liu 237. Puzzo D, Piacentini R, Fa M, Gulisano W, Li Puma
L, Wen Y, Krishnamurthy P, Herman M, Wang DD, Staniszewski A, Zhang H, Tropea MR, Cocco
L, Schachter JB, Nelson RB, Lau LF, Duff S, Palmeri A, Fraser P, D’Adamio L, Grassi C. LTP
KE.  Anesthesia leads to tau hyperphosphorylation and memory impairment caused by extracellular
through inhibition of phosphatase activity by hypo- Abeta and Tau oligomers is APP-dependent. elife.
thermia. J Neurosci. 2007;27(12):3090–7. 2017;6:e26991.
228. Planel E, Tatebayashi Y, Miyasaka T, Liu L, Wang 238. Qin Z, Pandey NR, Zhou X, Stewart CA, Hari
L, Herman M, Yu WH, Luchsinger JA, Wadzinski B, A, Huang H, Stewart AF, Brunel JM, Chen
Duff KE, Takashima A. Insulin dysfunction induces HH.  Functional properties of Claramine: a novel
in  vivo tau hyperphosphorylation through distinct PTP1B inhibitor and insulin-mimetic compound.
mechanisms. J Neurosci. 2007;27(50):13635–48. Biochem Biophys Res Commun. 2015;458(1):21–7.
229. Platt TL, Beckett TL, Kohler K, Niedowicz DM, 239. Qu Z, Jiao Z, Sun X, Zhao Y, Ren J, Xu

Murphy MP. Obesity, diabetes, and leptin resistance G.  Effects of streptozotocin-induced diabetes on
promote tau pathology in a mouse model of disease. tau phosphorylation in the rat brain. Brain Res.
Neuroscience. 2016;315:162–74. 2011;1383:300–6.
230. Ploia C, Antoniou X, Sclip A, Grande V, Cardinetti 240. Rahman A, Grundke-Iqbal I, Iqbal K. PP2B isolated
D, Colombo A, Canu N, Benussi L, Ghidoni R, from human brain preferentially dephosphorylates
Forloni G, Borsello T. JNK plays a key role in tau Ser-262 and Ser-396 of the Alzheimer disease abnor-
hyperphosphorylation in Alzheimer’s disease mod- mally hyperphosphorylated tau. J  Neural Transm.
els. J Alzheimers Dis. 2011;26(2):315–29. 2006;113(2):219–30.
21  Tau, Diabetes and Insulin 283

241. Ramos-Rodriguez JJ, Jimenez-Palomares M, WA, Petersen RC, Jack CR Jr. Association of type
Murillo-Carretero MI, Infante-Garcia C, Berrocoso 2 diabetes with brain atrophy and cognitive impair-
E, Hernandez-Pacho F, Lechuga-Sancho AM, ment. Neurology. 2014;82(13):1132–41.
Cozar-Castellano I, Garcia-Alloza M.  Central vas- 253. Roriz-Filho SJ, Sá-Roriz TM, Rosset I, Camozzato
cular disease and exacerbated pathology in a mixed AL, Santos AC, Chaves MLF, Moriguti JC, Roriz-­
model of type 2 diabetes and Alzheimer’s disease. Cruz M. (Pre)diabetes, brain aging, and cognition.
Psychoneuroendocrinology. 2015;62:69–79. Biochim Biophys Acta. 2009;1792(5):432–43.
242. Ramos-Rodriguez JJ, Ortiz-Barajas O, Gamero-­
254. Rosenfeld L.  Insulin: discovery and controversy.
Carrasco C, de la Rosa PR, Infante-Garcia C, Clin Chem. 2002;48(12):2270–88.
Zopeque-Garcia N, Lechuga-Sancho AM, Garcia-­ 255. Ryan CM, Geckle M. Why is learning and memory
Alloza M.  Prediabetes-induced vascular altera- dysfunction in type 2 diabetes limited to older adults?
tions exacerbate central pathology in APPswe/ Diabetes Metab Res Rev. 2000;16(5):308–15.
PS1dE9 mice. Psychoneuroendocrinology. 256. Sajan M, Hansen B, Ivey R 3rd, Sajan J, Ari C,
2014;48:123–35. Song S, Braun U, Leitges M, Farese-Higgs M,
243. Ramos-Rodriguez JJ, Ortiz O, Jimenez-­ Farese RV.  Brain insulin signaling is increased in
Palomares M, Kay KR, Berrocoso E, Murillo-­ insulin-resistant states and decreases in FOXOs
Carretero MI, Perdomo G, Spires-Jones T, and PGC-1alpha and increases in Abeta1-40/42 and
Cozar-Castellano I, Lechuga-Sancho AM, Garcia-­ Phospho-Tau May Abet Alzheimer development.
Alloza M. Differential central pathology and cogni- Diabetes. 2016;65(7):1892–903.
tive impairment in pre-diabetic and diabetic mice. 257. Sanna PP, Cammalleri M, Berton F, Simpson
Psychoneuroendocrinology. 2013;38(11):2462–75. C, Lutjens R, Bloom FE, Francesconi
244. Ramos-Rodriguez JJ, Sanchez-Sotano D, Doblas-­ W. Phosphatidylinositol 3-kinase is required for the
Marquez A, Infante-Garcia C, Lubian-Lopez S, expression but not for the induction or the mainte-
Garcia-Alloza M.  Intranasal insulin reverts central nance of long-term potentiation in the hippocampal
pathology and cognitive impairment in diabetic CA1 region. J Neurosci. 2002;22(9):3359–65.
mother offspring. Mol Neurodegener. 2017;12(1):57. 258. Santacruz K, Lewis J, Spires T, Paulson J, Kotilinek
245. Razay G, Wilcock GK.  Hyperinsulinaemia and L, Ingelsson M, Guimaraes A, DeTure M, Ramsden
Alzheimer’s disease. Age Ageing. 1994;23(5):396–9. M, McGowan E, Forster C, Yue M, Orne J, Janus
246. Reagan LP, Rosell DR, Alves SE, Hoskin EK, C, Mariash A, Kuskowski M, Hyman B, Hutton M,
McCall AL, Charron MJ, McEwen BS. GLUT8 glu- Ashe KH.  Tau suppression in a neurodegenerative
cose transporter is localized to excitatory and inhibi- mouse model improves memory function. Science.
tory neurons in the rat hippocampus. Brain Res. 2005;309(5733):476–81.
2002;932(1–2):129–34. 259. Saxena R, Hivert MF, Langenberg C, Tanaka T,
247. Reger MA, Watson GS, Frey WH 2nd, Baker LD, Pankow JS, Vollenweider P, Lyssenko V, Bouatia-­
Cholerton B, Keeling ML, Belongia DA, Fishel Naji N, Dupuis J, Jackson AU, Kao WH, Li M,
MA, Plymate SR, Schellenberg GD, Cherrier MM, Glazer NL, Manning AK, Luan J, Stringham HM,
Craft S. Effects of intranasal insulin on cognition in Prokopenko I, Johnson T, Grarup N, Boesgaard
memory-­impaired older adults: modulation by APOE TW, Lecoeur C, Shrader P, O’Connell J, Ingelsson
genotype. Neurobiol Aging. 2006;27(3):451–8. E, Couper DJ, Rice K, Song K, Andreasen CH,
248. Reger MA, Watson GS, Green PS, Wilkinson Dina C, Kottgen A, Le Bacquer O, Pattou F, Taneera
CW, Baker LD, Cholerton B, Fishel MA, Plymate J, Steinthorsdottir V, Rybin D, Ardlie K, Sampson
SR, Breitner JC, DeGroodt W, Mehta P, Craft M, Qi L, van Hoek M, Weedon MN, Aulchenko
S.  Intranasal insulin improves cognition and YS, Voight BF, Grallert H, Balkau B, Bergman
modulates beta-amyloid in early AD.  Neurology. RN, Bielinski SJ, Bonnefond A, Bonnycastle
2008;70(6):440–8. LL, Borch-Johnsen K, Bottcher Y, Brunner E,
249. Rehfeld JF, Stadil F. The effect of gastrin on basal- Buchanan TA, Bumpstead SJ, Cavalcanti-Proenca
and glucose-stimulated insulin secretion in man. J C, Charpentier G, Chen YD, Chines PS, Collins
Clin Invest. 1973;52(6):1415–26. FS, Cornelis M, Crawford JG, Delplanque J,
250. Renner DB, Svitak AL, Gallus NJ, Ericson ME, Frey Doney A, Egan JM, Erdos MR, Firmann M,
WH 2nd, Hanson LR. Intranasal delivery of insulin Forouhi NG, Fox CS, Goodarzi MO, Graessler
via the olfactory nerve pathway. J Pharm Pharmacol. J, Hingorani A, Isomaa B, Jorgensen T, Kivimaki
2012;64(12):1709–14. M, Kovacs P, Krohn K, Kumari M, Lauritzen T,
251. Roberts CK, Hevener AL, Barnard RJ.  Metabolic Levy-Marchal C, Mayor V, McAteer JB, Meyre D,
syndrome and insulin resistance: underlying causes Mitchell BD, Mohlke KL, Morken MA, Narisu N,
and modification by exercise training. Compr Palmer CN, Pakyz R, Pascoe L, Payne F, Pearson
Physiol. 2013;3(1):1–58. D, Rathmann W, Sandbaek A, Sayer AA, Scott
252. Roberts RO, Knopman DS, Przybelski SA, Mielke LJ, Sharp SJ, Sijbrands E, Singleton A, Siscovick
MM, Kantarci K, Preboske GM, Senjem ML, DS, Smith NL, Sparso T, Swift AJ, Syddall H,
Pankratz VS, Geda YE, Boeve BF, Ivnik RJ, Rocca Thorleifsson G, Tonjes A, Tuomi T, Tuomilehto J,
284 M. Gratuze et al.

Valle TT, Waeber G, Walley A, Waterworth DM, 269. Skeberdis VA, Lan J, Zheng X, Zukin RS,

Zeggini E, Zhao JH, GIANT Consortium, MAGIC Bennett MV.  Insulin promotes rapid delivery of
Investigators, Illig T, Wichmann HE, Wilson JF, N-methyl-D- aspartate receptors to the cell sur-
van Duijn C, Hu FB, Morris AD, Frayling TM, face by exocytosis. Proc Natl Acad Sci U S A.
Hattersley AT, Thorsteinsdottir U, Stefansson K, 2001;98(6):3561–6.
Nilsson P, Syvanen AC, Shuldiner AR, Walker M, 270. Smith U.  Impaired (‘diabetic’) insulin signaling
Bornstein SR, Schwarz P, Williams GH, Nathan and action occur in fat cells long before glucose
DM, Kuusisto J, Laakso M, Cooper C, Marmot intolerance  – is insulin resistance initiated in the
M, Ferrucci L, Mooser V, Stumvoll M, Loos RJ, adipose tissue? Int J  Obes Relat Metab Disord.
Altshuler D, Psaty BM, Rotter JI, Boerwinkle E, 2002;26(7):897–904.
Hansen T, Pedersen O, Florez JC, McCarthy MI, 271. Snowden JS, Bathgate D, Varma A, Blackshaw

Boehnke M, Barroso I, Sladek R, Froguel P, Meigs A, Gibbons ZC, Neary D.  Distinct behavioural
JB, Groop L, Wareham NJ, Watanabe RM. Genetic profiles in frontotemporal dementia and seman-
variation in GIPR influences the glucose and insu- tic dementia. J  Neurol Neurosurg Psychiatry.
lin responses to an oral glucose challenge. Nat 2001;70(3):323–32.
Genet. 2010;42(2):142–8. 272. Son SM, Song H, Byun J, Park KS, Jang HC,
260. Schechter R, Beju D, Miller KE. The effect of insu- Park YJ, Mook-Jung I.  Altered APP process-
lin deficiency on tau and neurofilament in the insulin ing in insulin-resistant conditions is mediated by
knockout mouse. Biochem Biophys Res Commun. autophagosome accumulation via the inhibition of
2005;334(4):979–86. mammalian target of rapamycin pathway. Diabetes.
261. Scherer T, O’Hare J, Diggs-Andrews K, Schweiger 2012;61(12):3126–38.
M, Cheng B, Lindtner C, Zielinski E, Vempati P, 273. Spillantini MG, Bird TD, Ghetti B. Frontotemporal
Su K, Dighe S, Milsom T, Puchowicz M, Scheja L, dementia and Parkinsonism linked to chromo-
Zechner R, Fisher SJ, Previs SF, Buettner C. Brain some 17: a new group of tauopathies. Brain Pathol.
insulin controls adipose tissue lipolysis and lipogen- 1998;8(2):387–402.
esis. Cell Metab. 2011;13(2):183–94. 274. Steen E, Terry BM, Rivera EJ, Cannon JL, Neely
262. Schmitz-Peiffer C.  Protein kinase C and lipid-­ TR, Tavares R, Xu XJ, Wands JR, de la Monte
induced insulin resistance in skeletal muscle. Ann N SM. Impaired insulin and insulin-like growth factor
Y Acad Sci. 2002;967:146–57. expression and signaling mechanisms in Alzheimer’s
263. Schubert M, Brazil DP, Burks DJ, Kushner JA, disease – is this type 3 diabetes? J Alzheimers Dis.
Ye J, Flint CL, Farhang-Fallah J, Dikkes P, Warot 2005;7(1):63–80.
XM, Rio C, Corfas G, White MF.  Insulin recep- 275. Steiner DF, Rouillé Y, Gong Q, Martin S, Carroll
tor substrate-2 deficiency impairs brain growth R, Chan SJ. The role of prohormone convertases in
and promotes tau phosphorylation. J  Neurosci. insulin biosynthesis: evidence for inherited defects
2003;23(18):7084–92. in their action in man and experimental animals.
264. Schubert M, Gautam D, Surjo D, Ueki K, Baudler Diabetes Metab. 1996;22(2):94–104.
S, Schubert D, Kondo T, Alber J, Galldiks N, 276. Stevanovic K, Yunus A, Joly-Amado A, Gordon M,
Kustermann E, Arndt S, Jacobs AH, Krone W, Kahn Morgan D, Gulick D, Gamsby J. Disruption of nor-
CR, Bruning JC. Role for neuronal insulin resistance mal circadian clock function in a mouse model of
in neurodegenerative diseases. Proc Natl Acad Sci U tauopathy. Exp Neurol. 2017;294:58–67.
S A. 2004;101(9):3100–5. 277. Stolk RP, Breteler MM, Ott A, Pols HA, Lamberts
265. Scott AR, Bennett T, Macdonald IA. Diabetes melli- SW, Grobbee DE, Hofman A. Insulin and cognitive
tus and thermoregulation. Can J Physiol Pharmacol. function in an elderly population. The Rotterdam
1987;65(6):1365–76. Study. Diabetes Care. 1997;20(5):792–5.
266. Sergeant N, Bretteville A, Hamdane M, Caillet-­ 278. Strachan MW, Deary IJ, Ewing FM, Frier BM.  Is
Boudin ML, Grognet P, Bombois S, Blum D, type II diabetes associated with an increased risk
Delacourte A, Pasquier F, Vanmechelen E, Schraen-­ of cognitive dysfunction? A critical review of pub-
Maschke S, Buee L.  Biochemistry of Tau in lished studies. Diabetes Care. 1997;20(3):438–45.
Alzheimer’s disease and related neurological disor- 279. Strubbe JH, Porte D Jr, Woods SC. Insulin responses
ders. Expert Rev Proteomics. 2008;5(2):207–24. and glucose levels in plasma and cerebrospinal
267. Shalaby TH, Yousef MK, Dupre RK. fluid during fasting and refeeding in the rat. Physiol
Thermoregulatory responses of diabetic rats. Behav. 1988;44(2):205–8.
Comp Biochem Physiol A Comp Physiol. 280. Sultan A, Nesslany F, Violet M, Begard S, Loyens
1989;94(1):153–7. A, Talahari S, Mansuroglu Z, Marzin D, Sergeant
268. Shuai H, Zhang J, Xie J, Zhang M, Ma J, Zhang N, Humez S, Colin M, Bonnefoy E, Buee L, Galas
L, Wang X.  Role of stereotaxically injected IgG MC. Nuclear tau, a key player in neuronal DNA pro-
from db/db mice in the phosphorylation of the tection. J Biol Chem. 2011;286(6):4566–75.
microtubule-­associated protein tau in hippocampus. 281. Swaab DF, Fliers E, Partiman TS.  The supra-
Brain Res. 2012;1486:14–26. chiasmatic nucleus of the human brain in rela-
21  Tau, Diabetes and Insulin 285

tion to sex, age and senile dementia. Brain Res. 293. Toledo EM, Inestrosa NC.  Activation of Wnt

1985;342(1):37–44. signaling by lithium and rosiglitazone reduced
282. Sydow A, Van der Jeugd A, Zheng F, Ahmed T, spatial memory impairment and neurodegenera-
Balschun D, Petrova O, Drexler D, Zhou L, Rune tion in brains of an APPswe/PSEN1DeltaE9 mouse
G, Mandelkow E, D’Hooge R, Alzheimer C, model of Alzheimer’s disease. Mol Psychiatry.
Mandelkow EM.  Tau-induced defects in synaptic 2010;15(3):272–85, 228.
plasticity, learning, and memory are reversible in 294. Torrent L, Ferrer I.  PP2A and Alzheimer disease.
transgenic mice after switching off the toxic Tau Curr Alzheimer Res. 2012;9(2):248–56.
mutant. J Neurosci. 2011;31(7):2511–25. 295. Tournissac M, Vandal M, Francois A, Planel E,
283. Szkudelski T. The mechanism of alloxan and strepto- Calon F.  Old age potentiates cold-induced tau
zotocin action in B cells of the rat pancreas. Physiol phosphorylation: linking thermoregulatory defi-
Res. 2001;50(6):537–46. cit with Alzheimer’s disease. Neurobiol Aging.
284. Taguchi A, White MF. Insulin-like signaling, nutri- 2017;50:25–9.
ent homeostasis, and life span. Annu Rev Physiol. 296. Towle HC, Kaytor EN, Shih HM. Regulation of the
2008;70:191–212. expression of lipogenic enzyme genes by carbohy-
285. Takalo M, Haapasalo A, Martiskainen H, Kurkinen drate. Annu Rev Nutr. 1997;17:405–33.
KM, Koivisto H, Miettinen P, Khandelwal VK, 297. Trojanowski JQ, Lee VM. Phosphorylation of paired
Kemppainen S, Kaminska D, Makinen P, Leinonen helical filament tau in Alzheimer’s disease neurofi-
V, Pihlajamaki J, Soininen H, Laakso M, Tanila H, brillary lesions: focusing on phosphatases. FASEB
Hiltunen M.  High-fat diet increases tau expression J. 1995;9(15):1570–6.
in the brain of T2DM and AD mice independently 298. Turban S, Hajduch E.  Protein kinase C isoforms:
of peripheral metabolic status. J  Nutr Biochem. mediators of reactive lipid metabolites in the
2014;25(6):634–41. development of insulin resistance. FEBS Lett.
286. Takeda S, Sato N, Uchio-Yamada K, Sawada K, 2011;585(2):269–74.
Kunieda T, Takeuchi D, Kurinami H, Shinohara 299. Valente T, Gella A, Fernandez-Busquets X, Unzeta
M, Rakugi H, Morishita R.  Diabetes-accelerated M, Durany N.  Immunohistochemical analysis of
memory dysfunction via cerebrovascular inflamma- human brain suggests pathological synergism of
tion and Abeta deposition in an Alzheimer mouse Alzheimer’s disease and diabetes mellitus. Neurobiol
model with diabetes. Proc Natl Acad Sci U S A. Dis. 2010;37(1):67–76.
2010;107(15):7036–41. 300. van der Heide LP, Kamal A, Artola A, Gispen
2 87. Talbot K, Wang HY, Kazi H, Han LY, Bakshi KP, WH, Ramakers GM.  Insulin modulates hippo-
Stucky A, Fuino RL, Kawaguchi KR, Samoyedny campal activity-dependent synaptic plasticity in a
AJ, Wilson RS, Arvanitakis Z, Schneider N-methyl-d-aspartate receptor and phosphatidyl-­
JA, Wolf BA, Bennett DA, Trojanowski JQ, inositol-­3-kinase-dependent manner. J  Neurochem.
Arnold SE.  Demonstrated brain insulin resis- 2005;94(4):1158–66.
tance in Alzheimer’s disease patients is 301. van Elderen SG, de Roos A, de Craen AJ,
associated with IGF-1 resistance, IRS-1 dys- Westendorp RG, Blauw GJ, Jukema JW, Bollen EL,
regulation, and cognitive decline. J  Clin Invest. Middelkoop HA, van Buchem MA, van der Grond
2012;122(4):1316–38. J. Progression of brain atrophy and cognitive decline
288. Tamura BK, Masaki KH, Blanchette P. Weight loss in diabetes mellitus: a 3-year follow-up. Neurology.
in patients with Alzheimer’s disease. J  Nutr Elder. 2010;75(11):997–1002.
2007;26(3–4):21–38. 302. Vandal M, White PJ, Chevrier G, Tremblay C,
289. Tanaka T, Nagatani S, Bucholtz DC, Ohkura S, St-Amour I, Planel E, Marette A, Calon F.  Age-­
Tsukamura H, Maeda K, Foster DL. Central action dependent impairment of glucose tolerance in the
of insulin regulates pulsatile luteinizing hormone 3xTg-AD mouse model of Alzheimer’s disease.
secretion in the diabetic sheep model. Biol Reprod. FASEB J. 2015;29(10):4273–84.
2000;62(5):1256–61. 303. Vandal M, White PJ, Tournissac M, Tremblay

290. Tanimukai H, Grundke-Iqbal I, Iqbal C, St-Amour I, Drouin-Ouellet J, Bousquet
K.  Up-regulation of inhibitors of protein phos- M, Traversy MT, Planel E, Marette A, Calon
phatase-­2A in Alzheimer’s disease. Am J  Pathol. F.  Impaired thermoregulation and beneficial
2005;166(6):1761–71. effects of thermoneutrality in the 3xTg-AD
291. Taylor SI, Arioglu E.  Syndromes associated with model of Alzheimer’s disease. Neurobiol Aging.
insulin resistance and acanthosis nigricans. J Basic 2016;43:47–57.
Clin Physiol Pharmacol. 1998;9(2–4):419–39. 304. Vandal M, White PJ, Tremblay C, St-Amour I,
292. To AW, Ribe EM, Chuang TT, Schroeder JE, Chevrier G, Emond V, Lefrancois D, Virgili J, Planel
Lovestone S.  The epsilon3 and epsilon4 alleles of E, Giguere Y, Marette A, Calon F.  Insulin reverses
human APOE differentially affect tau phosphoryla- the high-fat diet-induced increase in brain Abeta and
tion in hyperinsulinemic and pioglitazone treated improves memory in an animal model of Alzheimer
mice. PLoS One. 2011;6(2):e16991. disease. Diabetes. 2014;63(12):4291–301.
286 M. Gratuze et al.

305. Vellas B, Lauque S, Gillette-Guyonnet S, Andrieu 317. Xu S, Guan Q, Wang C, Wei X, Chen X, Zheng B,
S, Cortes F, Nourhashemi F, Cantet C, Ousset PJ, An P, Zhang J, Chang L, Zhou W, Mody I, Wang
Grandjean H. Impact of nutritional status on the evo- Q.  Rosiglitazone prevents the memory deficits
lution of Alzheimer’s disease and on response to ace- induced by amyloid-beta oligomers via inhibi-
tylcholinesterase inhibitor treatment. J  Nutr Health tion of inflammatory responses. Neurosci Lett.
Aging. 2005;9(2):75–80. 2014;578:7–11.
306. Vieira MN, Lyra ESNM, Ferreira ST, De Felice 318. Xu W, Qiu C, Winblad B, Fratiglioni L. The effect
FG.  Protein tyrosine phosphatase 1B (PTP1B): of borderline diabetes on the risk of dementia and
a potential target for Alzheimer’s therapy? Front Alzheimer’s disease. Diabetes. 2007;56(1):211–6.
Aging Neurosci. 2017;9:7. 319. Yamamoto H, Saitoh Y, Fukunaga K, Nishimura H,
307. Vintem AP, Henriques AG, da Cruz ESOA, da Cruz Miyamoto E.  Dephosphorylation of microtubule
ESEF. PP1 inhibition by Abeta peptide as a poten- proteins by brain protein phosphatases 1 and 2A, and
tial pathological mechanism in Alzheimer’s disease. its effect on microtubule assembly. J  Neurochem.
Neurotoxicol Teratol. 2009;31(2):85–8. 1988;50(5):1614–23.
308. von Bergen M, Barghorn S, Li L, Marx A, Biernat 320. Yanamandra K, Jiang H, Mahan TE, Maloney SE,
J, Mandelkow EM, Mandelkow E.  Mutations Wozniak DF, Diamond MI, Holtzman DM. Anti-tau
of tau protein in frontotemporal dementia pro- antibody reduces insoluble tau and decreases brain
mote aggregation of paired helical filaments by atrophy. Ann Clin Transl Neurol. 2015;2(3):278–88.
enhancing local beta-structure. J  Biol Chem. 321. Yanamandra K, Kfoury N, Jiang H, Mahan TE,
2001;276(51):48165–74. Ma S, Maloney SE, Wozniak DF, Diamond MI,
309. Watson GS, Cholerton BA, Reger MA, Baker Holtzman DM.  Anti-tau antibodies that block
LD, Plymate SR, Asthana S, Fishel MA, Kulstad tau aggregate seeding in  vitro markedly decrease
JJ, Green PS, Cook DG, Kahn SE, Keeling ML, pathology and improve cognition in  vivo. Neuron.
Craft S.  Preserved cognition in patients with early 2013;80(2):402–14.
Alzheimer disease and amnestic mild cognitive 322. Yang YJ, Hope ID, Ader M, Bergman RN.  Insulin
impairment during treatment with rosiglitazone: transport across capillaries is rate limiting for insulin
a preliminary study. Am J  Geriatr Psychiatry. action in dogs. J Clin Investig. 1989;84(5):1620–8.
2005;13(11):950–8. 323. Yarchoan M, Toledo JB, Lee EB, Arvanitakis

310. Weingarten MD, Lockwood AH, Hwo SY, Z, Kazi H, Han LY, Louneva N, Lee VM, Kim
Kirschner MW.  A protein factor essential for SF, Trojanowski JQ, Arnold SE.  Abnormal ser-
microtubule assembly. Proc Natl Acad Sci U S A. ine phosphorylation of insulin receptor substrate
1975;72(5):1858–62. 1 is associated with tau pathology in Alzheimer’s
311. Werther GA, Hogg A, Oldfield BJ, McKinley MJ, disease and tauopathies. Acta Neuropathol.
Figdor R, Allen AM, Mendelsohn FA. Localization 2014;128(5):679–89.
and characterization of insulin receptors in rat brain 324. Young WS 3rd. Periventricular hypothalamic cells in
and pituitary gland using in  vitro autoradiography the rat brain contain insulin mRNA. Neuropeptides.
and computerized densitometry. Endocrinology. 1986;8(2):93–7.
1987;121(4):1562–70. 325. Yu Y, Li X, Blanchard J, Li Y, Iqbal K, Liu F,
312. White H, Pieper C, Schmader K. The association of Gong CX. Insulin sensitizers improve learning and
weight change in Alzheimer’s disease with sever- attenuate tau hyperphosphorylation and neuroin-
ity of disease and mortality: a longitudinal analysis. flammation in 3xTg-AD mice. J  Neural Transm.
J Am Geriatr Soc. 1998;46(10):1223–7. 2015;122(4):593–606.
313. White MF, Yenush L.  The IRS-signaling system: 326. Zahniser NR, Goens MB, Hanaway PJ, Vinych

a network of docking proteins that mediate insulin JV.  Characterization and regulation of insu-
and cytokine action. Curr Top Microbiol Immunol. lin receptors in rat brain. J  Neurochem.
1998;228:179–208. 1984;42(5):1354–62.
314. Wolf-Klein GP, Silverstone FA, Levy AP. Nutritional 327. Zhang X, Ding L, Yang L, Qin W, Yuan J, Li S, Hu
patterns and weight change in Alzheimer patients. W.  Brain atrophy correlates with severe enlarged
Int Psychogeriatr. 1992;4(1):103–18. perivascular spaces in basal ganglia among lacunar
315. Wu SP, Lu KT, Chang WC, Gean PW. Involvement stroke patients. PLoS One. 2016;11(2):e0149593.
of mitogen-activated protein kinase in hippo- 328. Zhang ZY, Dodd GT, Tiganis T.  Protein tyro-

campal long-term potentiation. J  Biomed Sci. sine phosphatases in hypothalamic insulin
1999;6(6):409–17. and leptin signaling. Trends Pharmacol Sci.
316. Wu Y, Ding Y, Tanaka Y, Zhang W.  Risk factors 2015;36(10):661–74.
contributing to type 2 diabetes and recent advances 329. Zhao N, Liu CC, Van Ingelgom AJ, Martens YA,
in the treatment and prevention. Int J  Med Sci. Linares C, Knight JA, Painter MM, Sullivan PM,
2014;11(11):1185–200. Bu G.  Apolipoprotein E4 impairs neuronal insulin
21  Tau, Diabetes and Insulin 287

signaling by trapping insulin receptor in the endo- Qiu WQ. An amylin analog used as a challenge test
somes. Neuron. 2017;96(1):115–129. e115. for Alzheimer’s disease. Alzheimers Dement (N Y).
330. Zhao WQ, De Felice FG, Fernandez S, Chen 2017;3(1):33–43.
H, Lambert MP, Quon MJ, Krafft GA, Klein 332. Zhu X, Raina AK, Rottkamp CA, Aliev G,

WL.  Amyloid beta oligomers induce impair- Perry G, Boux H, Smith MA.  Activation and
ment of neuronal insulin receptors. FASEB redistribution of c-Jun N-terminal kinase/stress
J. 2008;22(1):246–60. activated protein kinase in degenerating neu-
331. Zhu H, Stern RA, Tao Q, Bourlas A, Essis MD, rons in Alzheimer’s disease. J  Neurochem.
Chivukula M, Rosenzweig J, Steenkamp D, Xia 2001;76(2):435–41.
W, Mercier GA, Tripodis Y, Farlow M, Kowall N,
 Part V
Tau Aggregation, and Propagation
 Top-Down Projections Direct
the Gradual Progression 22
of Alzheimer-Related Tau
Pathology Throughout
the Neocortex

Heiko Braak and Kelly Del Tredici

Introduction severity, systematically progressing throughout

the brain and spinal cord and leaving behind a
Sporadic Alzheimer’s disease (sAD) is a disorder characteristic regional distribution pattern
that is not known to occur in non-human primates (Fig.  22.1) [10, 19, 25, 45, 66]. Only the late
or non-primate vertebrates [9, 108]. Involved in phases of the pathological process cause readily
the disease-associated pathological process are recognizable symptoms and correlate with the
nerve cells of the central nervous system (CNS) clinical picture of sAD [4, 26, 82, 98, 102].
and, among the many different neuronal types, However, the frequency of clinically overt cases
only a few show a pronounced proclivity to increases considerably with age and imposes a
become involved. Nearly all of the vulnerable major socio-economic burden on societies with
types belong to the class of postnatally-maturing advancing life expectancies [31, 39, 50, 90].
projection neurons that generate in relation to the As pointed out above, one of the typical fea-
size of their cell body a long and sparsely myelin- tures of the sAD-associated process is that vul-
ated axon [25, 65]. nerable brain regions become involved according
The relentlessly progressive process associ- to a predictable topographical sequence
ated with the disease requires an extended period (Figs. 22.1 and 22.2, red arrow at left). Whereas
of time – nearly a lifetime – to reach its full extent the pace of inter-individual disease progression
[29]. The reasons for this extremely slow pro- varies considerably, the degree of inter-individual
gression are presently unknown, and it is likewise variability with respect to the sequence of the
unclear why sAD consistently fails to go into pathological tau changes is so minimal that the
remission [25]. Abnormal changes in the protein distribution pattern of the lesions can serve to
tau begin early in life [24, 29, 48] and, after their distinguish different neuropathological stages in
appearance, the tau lesions gradually increase in asymptomatic and symptomatic individuals [19,
25, 28, 47].
Crucial to the pathological process are alter-
In commemoration of Korbinian Brodmann (November ations of the protein tau [3, 8, 11, 54, 57, 67,
17, 1868 – August 22, 1918). 74]. In the cerebral cortex, the first abnormal
tau aggregates develop in the transentorhinal
H. Braak · K. Del Tredici (*)
Clinical Neuroanatomy Section/Department of region, a phylogenetically late-appearing tran-
Neurology, Center for Biomedical Research, sition cortex intercalated between allocortical
University of Ulm, Ulm, Germany regions of the temporal lobe (entorhinal region,
e-mail: heiko.braak@uni-ulm.de; hippocampal formation) and the neocortex
kelly.del-tredici@uni-ulm.de

© Springer Nature Singapore Pte Ltd. 2019 291

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_22
292 H. Braak and K. Del Tredici

Fig. 22.1  Overview of intraneuronal AT8-­region (arrowheads indicate its borders) into the laterally
immunopositive pathology (NFT stages I–IV) in fron- adjoining neocortex of the occipitotemporal gyrus,
tal sections (100 μm) through the temporal lobe. The whereas the remaining areas of the temporal lobe, par-
tau pathology is macroscopically visible in anteromedial ticularly the superior temporal gyrus, appear to be unin-
portions of the temporal lobe (brown chromogen DAB). volved. During NFT stage IV (female, 82 years of age),
The density of the lesions is highest in the transentorhinal neurofibrillary pathology advances more widely into high
and entorhinal regions, decreases from there, and gradu- order association areas of the temporal lobe up to (sel-
ally tapers off. In NFT stage I (female, 80 years of age) dom beyond) the medial temporal gyrus and into the
abnormal tau pathology is confined to the allocortical adjoining frontal fields. However, at NFT stages I–IV,
transentorhinal region. During NFT stage II (male, isolated AT8-­immunoreactive pyramidal cells are found
80 years of age), not only the transentorhinal region but in the superior temporal gyrus, possibly representing an
also the entorhinal region proper and a portion of the hip- initial ‘reaction’ to the neuron-to-neuron progression of
pocampal formation display clear involvement. In a case the pathological process originating in earlier-involved
assigned to NFT stage III (female, 90 years of age), neu- cortical regions (see [27, 70]). (Reproduced with permis-
rofibrillary changes extend beyond the transentorhinal sion from Ref. [25])

(Fig.  22.1, NFT stage I). The transentorhinal involvement is complemented by the develop-
region exists in higher primates and achieves its ment of tau aggregates within additional sen-
largest extent in the human brain [18]. From sory high order association fields and prefrontal
there, the pathological process enters both the areas (Fig.  22.1, NFT stage IV), followed by
entorhinal region and hippocampal formation sensory first order association fields and premo-
(Fig. 22.1, NFT stage II) and the temporal neo- tor areas (NFT stage V), and, finally, the pri-
cortex of the occipitotemporal gyrus accompa- mary areas of the neocortex (NFT stage VI).
nying the transentorhinal region as well as Notably, all of the regions that gradually
adjoining prefrontal fields (Fig.  22.1, NFT become affected during sAD-process are ana-
stage III). Subsequently, this pattern of tau tomically interconnected (Fig. 22.2) [2, 25, 52].
22  Top-Down Projections Direct the Gradual Progression of Alzheimer-Related Tau Pathology… 293

Fundamental Organization of data between the different cortical areas [40,

of the Neocortex and of Cortico-­ 83]. Notably, a considerable proportion of the
Cortical Connections neurons that generate such cortico-cortical fibers
is highly susceptible to the sAD-associated pro-
The various fields of the neocortex evolve, phylo- cess and becomes severely affected in the early
and ontogenetically, in an orderly sequence and phase of the disease [23, 25, 27, 34, 42, 62].
they also mature consecutively. This phenome-
non is traceable in the ontogenetic gradually pro-
gressive process of myelination as well as in the Cortico-Cortical Bottom-Up
functional maturation of the neocortex that is and Top-Down Connectivities
linked to myelination [58, 88, 109] (Fig.  22.2, in the Neocortex
blue arrow at center). In general, the neocortical
myelination commences late, progresses slowly Connectivities exist that establish links between
during childhood, and, after puberty, persists well the primary sensory and motor fields to the sen-
into adulthood. Notably, the myelination of the sory first order association areas and premotor
human neocortex is influenced by environmental fields (Fig. 22.2A, B), and, similarly, links from
and social interactions [15, 21, 43, 58, 86, 88, these fields to the sensory high order association
109, 112]. areas and prefrontal fields (Fig.  22.2B, C), and
The neocortical primary sensory and motor from these fields to the transentorhinal region
areas are the first to emerge (Fig. 22.2A). A girdle (Fig.  22.2C, D, thin black arrows between
of first order sensory association areas and pre- boxes). All of these pathways constitute bottom-
motor are as follows (Fig.  22.2B). These again ­up or ‘upstream’ connectivities, whereas other
are complemented by a profusion of high order pathways running in the opposite direction con-
sensory association areas and prefrontal fields stitute top-down or ‘downstream’ connectivities
(Fig.  22.2C). The last region is the transitional (Fig. 22.2 from D to C, from C to B, and from B
cortex between the temporal neo- and allocortex: to A, marked by thick red arrows between the
the transentorhinal region (Fig. 22.2D) [25, 27]. boxes) [14, 49, 52, 61, 95].
One of the essential tasks of the cerebral cor- The top-down connectivities are presently the
tex is the continuous reception and processing of focus of experimental studies seeking to explain
large units of input from the sensory organs and the mechanisms of impaired conscious percep-
the visceral organs together with the transforma- tion during anesthesia [33, 61, 78, 79]. Bottom-up
tion of these data into steering impulses that con- connectivities terminate preferentially on small
trol movements of skeletal muscles and functions projection neurons of the granular layer (IV),
of other organs as part of the whole organism’s namely, the spiny stellate cells [41]. These cells
interaction with the world outside [5, 83, 116] establish contacts via their short radially-oriented
(Fig. 22.2). In the context of the present chapter, axon to the proximal segments of the basal den-
it is important to emphasize that nerve cells of the drites of their target pyramidal cells in neocorti-
CNS that are primarily and directly occupied cal layers III and V [76, 77]. In this manner, the
with these tasks remain intact during sAD and, if spiny stellate cells establish a point-to-point
at all, only become involved at end-stage transsynaptic transmission and disseminate data
disease. in a highly ordered manner to select projection
A frequently underestimated requirement for neurons belonging to a single cortical column
the effective functioning of the neocortex is the [12, 13].
tight functional linkage on the part of the differ- In terms of functional maturation, bottom-up
ent fields (Fig.  22.2). This is accomplished and connectivities from A to B mature earlier than
ensured by means of a multitude of association those from B to C, and these connectivities also
fibers that guarantee a precise and rapid exchange mature earlier that those from C to D (Fig. 22.2).
294 H. Braak and K. Del Tredici

Fig. 22.2  Diagram depicting the top-down progres- (C) From the transentorhinal region, the lesions enter into
sion of sAD-related tau pathology along cortico-­ high order sensory association areas of the basal temporal
cortical connections. The degrees of shading from black neocortex, at first those covering the occipitotemporal
to aquamarine represent the growing severity and the pro- gyrus, as well as into adjoining prefrontal fields (NFT
posed spread of the lesions. (D) Cortical tau lesions begin stage III, dark blue). From neocortical fields that become
in the transentorhinal region, a transitional cortex between involved at NFT stage III, tau pathology develops in the
allocortical and neocortical regions of the temporal lobe remaining high order association fields of the temporal,
(Roman numerals indicate the first macroscopically visi- parietal, and occipital cortices as well as in additional pre-
ble lesions in NFT stages, here stage I, black). From the frontal areas (NFT stage IV, blue). (B) From there, tau
transentorhinal region, the tau pathology progresses into lesions progress into the first order sensory association
the entorhinal region proper and into portions of the hip- fields and into premotor areas (NFT stage V, light blue).
pocampal formation (e.g., CA1 sector) (NFT stage II, (A) Finally, the pathological process arrives at the primary
navy blue). The lesions then encroach upon additional fields of the neocortex (NFT stage VI, aquamarine). The
sectors of the Ammon’s horn (NFT stage III, dark blue) red arrow at the far left indicates the systematic
22  Top-Down Projections Direct the Gradual Progression of Alzheimer-Related Tau Pathology… 295

Top-down connectivities lag behind all of their of primate evolution and also mature late during
corresponding bottom-up connections. Those ontogenesis [6, 7, 9, 20, 91–93]. From the stand-
from B to A mature only after the connections point of evolutionary biology, this neuronal vul-
from A to B have attained sufficient maturation. nerability is probably inconsequential because it
Similarly, the top-down connectivities from C to does not threaten the preservation of the human
B and from D to C develop later than their bot- species [23].
tom-­up counterparts [27] (Fig. 22.2). After axons The uniform regional distribution pattern of
of spiny stellate cells belonging to the bottom-up the tau lesions is consistent with the concept put
pathway find their contacts to the proximal den- forth by many researchers that spreading of tau
drites of their target pyramidal neurons, the den- pathology within the human brain might proceed
drites continue to sprout and to produce distal along axons of involved nerve cells within a
tips with new sites for synapses of top-down given field to hitherto uninvolved nerve cells in
axons that grow into the cortex during this devel- other, including more distant, fields [22, 89, 99].
opmental phase. Immediately after birth, the Experiments show that pathogenic forms of tau,
axons of top-down projections from C to B and including human tau, are capable of seeding and
from D to C are unmyelinated and subsequently can be transmitted over distances axonally and
acquire their myelin sheathes during childhood transsynaptically [2, 32, 35, 37, 38, 46, 55, 59,
and adolescence in an activity-dependent manner 70, 73, 81].
organized by external stimuli [23]. Projections from subcortical nuclei probably
do not cause the early cortical lesions because the
only subcortical source involved during NFT
Involvement of Cortico-Cortical stage I that could possibly transmit abnormal tau
Projection Neurons and Their Target seeds to neocortical projection neurons is the
Cells in the sAD Process group of non-thalamic nuclei (locus coeruleus,
upper raphe nuclei, magnocellular nuclei of the
The sAD process progresses from D to C, from C basal forebrain) with diffuse projections to all
to B and, finally, from B to A (Fig.  22.2, red areas of the cortex [53, 84, 96]. However, highly
arrow at left margin)  – i.e., it advances in the ramified axonal projections from these subcorti-
reverse sequence and opposite direction to corti- cal nuclei exert their influence upon the expan-
cal myelination and the functional maturation of sive fields of the cortex chiefly via non-junctional
the neocortex (retrogenesis) [20, 93, 94, 97] varicosities and volume transmission rather than
(Fig. 22.2, blue arrow at center). An explanation via classical synapses with pre- and postsynaptic
for the pronounced order in the trajectory of the densities [1, 85, 87]. In addition, non-junctional
tau pathology is still lacking, but it remains clear varicosities (incomplete synapses) are not suited
that the vulnerable nerve cell types in sAD that for the seeding-induced transmission of tau
are subjected to early abnormal tau changes pathology [70]. It also means that the ascending
emerged, for the most part, during the later phases diffuse projections from the subcortical non-­

Fig. 22.2  (continued)  progression of the disease-related cortical CNS components into the regulation of cortical
tau pathology, whereas the blue arrow indicates the grad- output. Foremost among these is the ‘limbic’ circuit that
ual progression of myelination in the diverse cortical carries output data from the hippocampal CA1 sector and
regions and, at the same time, represents the increasing subiculum via the ventral striatum (v.striat.), ventral palli-
degree of the functional maturation of their neuronal con- dum (v.pall.), and mediodorsal thalamus to the prefrontal
stituents. The short red arrows between involved regions fields. The ‘striatal’ circuit follows, guiding data through
represent the top-down cortico-cortical connections that the dorsal striatum, pallidum, and thalamus to chiefly the
direct the pathological tau process. By contrast, the bot- premotor areas, and these, again, influence the group of
tom-up projections and their target neurons (thin black precerebellar nuclei/cerebellum that provide via the thala-
arrows) refrain from developing neurofibrillary pathology mus important input to the primary motor cortex.
in sAD. The efferent side of the diagram (at the right) also Abbreviations: CA1 first sector of the Ammon’s horn,
includes three important circuits. These incorporate sub- ento entorhinal, hippocamp form hippocampal formation
296 H. Braak and K. Del Tredici

thalamic nuclei would be unlikely to transport and it enters portions of the superior temporal
seed-containing tau via axons to distant cortical gyrus only in end-phase disease stages [23, 25].
sites in a highly selective manner because a gen- The reason for the delay is that the cortex of
eralized spreading pattern is inconsistent with the the superior temporal gyrus in the human adult
specific regional distribution pattern of abnormal brain is characterized, compared to other tempo-
tau in NFT stages I–VI. ral lobe regions, by a stronger degree of myelina-
Remarkably, the sAD-associated process from tion [64, 86]. When looking for potentially very
beginning to end leaves the spiny stellate cells early pathological tau lesions within the target
completely intact [25]. Whether projection neu- neurons of top-down connectivities, one could
rons receiving chiefly bottom-up connectivities begin with portions of Brodmann area 22 in the
become involved at all during sAD is presently superior temporal gyrus. In fact, although the first
unknown. The spiny stellate neurons are not spe- macroscopically detectable lesions routinely
cifically vulnerable and, thus, it is improbable develop in Brodmann area 22 in NFT stage V, a
that they would produce tau seeds or transport few isolated AT8-immunoreactive pyramidal
them via their axons (Fig. 22.2, thin black arrows cells mostly limited to layers III and V already
between boxes). The question then emerges as to occur there at lower NFT stages I–III.
how the parent cells of the top-down projections
and their target cells fare (Fig.  22.2, thick red
arrows between boxes). Our recent findings in Morphological and Abnormal Tau
the target cells are summarized in section Changes in Target Cells of Top-­
“Morphological and abnormal tau changes in tar- Down Connectivities
get cells of top-down connectivities” below and
indicate that it is the targets of the top-down pro- Abnormal tau in neocortical pyramidal cells is
jections which bear the brunt of the sAD neurofi- either confined to specific compartments of the
brillary pathology. involved neuron or fills it completely. Sections of
Early NFT stages (e.g., stages I–III) display 100  μm thickness facilitate recognition of the
obvious AT8-immunoreactive neurofibrillary entire dendritic tree of involved projection cells,
pathology limited to portions of the anteromedial and we recently found there four distinct groups
temporal lobe (Fig. 22.1). In thick tissue sections of sequential morphological and pathological
(100–300 μm), the existence of the lesions is vis- changes [27].
ible to the unaided eye. Heavily involved regions The first group displays swollen AT8-positive
can be easily distinguished from uninvolved ones distal segments of the basal dendrites and of the
and those with less pathology (Fig. 22.1). If one side branches of the apical dendrite belonging to
follows an imaginary line between the transento- involved pyramidal cells (Fig.  22.3a, b). These
rhinal region and the primary areas of the neocor- skirt-like distal tips point to a central immuno-
tex corresponding to that of the red arrows negative area (Fig. 22.3a, b). A second group dis-
connecting the boxes in Fig. 22.2, one only has to plays additional thread-like trailing structures
keep looking at the areas of the cortex, beginning (‘spokes’) surrounding the immunonegative area
with the transentorhinal region, in the direction (Fig. 22.3c, d). A third group reveals for the first
of the less involved or apparently uninvolved time the presence of abnormal tau in the cell
regions and check them for the presence or soma as well as in the stem of the apical dendrite
absence of abnormal tau. This would include the (Fig.  22.3e, f). In the last group, the pyramidal
regions in the occipitotemporal, inferior, medial, cells display pathological tau even within their
and superior temporal gyri, including the trans- axons (Fig. 22.3g, h). Cells of all four groups can
verse gyrus of Heschl (Fig. 22.1). The sAD pro- occur in one and the same individual, and the
cess proceeds uniformly in this direction with a four groups possibly represent phases in a devel-
delay upon reaching the border between the opmental sequence [27] (Fig.  22.4). The four
medial and superior gyri of the temporal lobe, phases may be part of an initial ‘reaction’ by a
22  Top-Down Projections Direct the Gradual Progression of Alzheimer-Related Tau Pathology… 297

Fig. 22.3  AT8-immunoreactive abnormal tau during higher magnification in (d). (e and f) Micrographs of two
early NFT stages develops in isolated pyramidal cells AT8-positive neurons assigned to group 3 (NFT stage II,
in neocortical areas that routinely become involved superior temporal gyrus, male, 77 years of age in (e), and
only in late NFT stages (100 μm sections). (a) Two AT8-­ NFT stage I, superior temporal gyrus, male 57  years of
immunoreactive neurons assigned to group 1 (NFT stage age in (f)). In addition to the dendritic changes, the patho-
III, superior temporal gyrus, male 72  years of age). logical material also occurs in the cell soma and in proxi-
Framed area indicates a nerve cell that is shown at higher mal portions of the apical dendrite. (g) A group 4 neuron
magnification in (b). Its distal dendritic segments are bent, displays in addition to the entirely AT8-immunopositive
thickened, and filled with abnormal tau. The skirt of basal dendritic tree an AT8-immunoreactive axon (NFT stage
dendrites surrounds a central immunonegative area. (c) A III, superior temporal gyrus, male, 57 years of age). The
group 2 neuron (NFT stage I, located in the medial tempo- framed area indicates the same neuron reproduced at
ral gyrus, male, 57  years of age), which, in addition to higher magnification in (h). There are no detectable signs
thickened distal segments of the basal dendrites, exhibits of severe axonal pathology nor are there varicosities or
thread-like extensions from there into the proximal seg- visible morphological changes at the junctions of the axon
ments that also head towards an AT8-immunonegative collaterals. (Reproduced, in part, with permission from
center. Framed area indicates the same cell shown at Ref. [27])

previously uninvolved neocortical field to a pro- activity at low NFT stages (I–III) anticipates tau
gressive neuron-to-neuron and top-down pathology in, e.g., the superior temporal gyrus or
­transaxonal spread of pathological tau originat- high order association fields of the occipital cor-
ing from cortical fields with a previous (and tex [70]. As such, it can be postulated that the
thereby greater) degree of involvement. All newly progression of the sAD-associated pathological
involved neocortical fields react similarly: The process may take place chiefly by means of an
same four phases develop regardless of age, gen- anterograde transport of tau seeds along top-­
der, or the presence/absence of amyloid-β down connectivities [27].
plaques. Taken together, the four phases repeat, Several implications emerge from the descrip-
in reverse order, the phylo- and ontogenetic tion and theory of the four phases. One is that
development of neocortical pyramidal cells [27, abnormal tau seeds would only be transmitted in
75] (Fig. 22.4). the first phase and solely by means of the distal
The existence of such an initial ‘reaction’ by dendritic segments, where the seeds trigger a
hitherto uninvolved regions of the neocortex is in renewed pathological response in the recipient
line with a recent study, in which tau seeding neuron (Fig. 22.3a, b). It is known that, following
298 H. Braak and K. Del Tredici

Fig. 22.4  Ontogenetic development of neocortical development of immunopositive thread-like structures

pyramidal cells (a–d) Generation of the axon (a) is fol- that emanate from the distal segments into the proximal
lowed by the development of a radially oriented apical dendrites leading towards an AT8- immunonegative soma
dendrite, including its tuft of terminal twigs (b). Next (c), (f). The soma of the involved cell together with proximal
short basal dendrites (later the proximal segments) portions of the apical dendrite become AT8-­
develop, which are complemented by a skirt (d) of later-­ immunopositive (g) Finally, the entire pyramidal cell,
maturing distal segments of the basal dendrites. Early including its axon, becomes AT8-immunoreactive (h) The
development of AT8 lesions in solitary pyramidal cells sequence of pathological tau changes shown in e–h repeat,
located in regions of the neocortex that typically in reverse order, the phylo- and ontogenetic development
become involved only in late NFT stages (e–h) Initially, of neocortical pyramidal cells in a–d. (Reproduced with
distal dendritic segments become filled with AT8-­ permission from Ref. [27])
immunoreactive tau aggregates (e) This is followed by the

its production in the soma, some of the native pro- dendrites are not essential for the disease propa-
tein tau undergoes resorting into the axon [67, gation. During the last two phases, the cell soma,
111], where it stabilizes microtubules [3, 44, 101, including the stem of the apical dendrite
106, 114]. It is not likely that part of the axonal (Fig. 22.3e, f), and the axon become filled with
tau protein, induced by signaling in the dendrites, abnormal tau (Fig.  22.3g, h). The axonal cyto-
peels away from the axonal microtubules and is skeleton remains without light microscopically
rapidly transferred without a trace from the axo- detectable injury, i.e., abnormal tau changes do
nal compartment into the dendritic tips [26, 27]. not have their origin in the axon. Initially, the
Moreover, none of the nerve cells in the first phase axonal cytoskeleton remains intact and thereby
displays the slightest traces of abnormal tau pro- would make the propagation of tau seeds possi-
tein in their axonal compartments (Fig. 22.3a, b). ble. The pathological “dendritic” tau is soluble
Abnormal tau first develops there during the and uniformly distributed within the axoplasm
fourth phase (Fig. 22.3g, h). Our findings are con- [72]. Gradually, presumably less soluble and
sonant with earlier reports, which show that nor- spindle-like tau aggregates develop in the axon
mal tau is not confined to axons but is also and these are interspersed by immunonegative
consistently present in dendrites [63, 68, 69, 71, segments [25, 110]. Therefore, the transmission
77, 104, 105]. This in turn implies that, following of tau seeds only appears possible for a brief
the potential transsynaptic transmission of tau- period, a constraint that could contribute to the
containing seeds, it is dendritic tau that perpetu- interminably protracted progression of
ates the pathological cascade in sAD [77, 104]. sAD. Most of the abnormal tau-containing nerve
During the second phase, abnormal tau aggre- cells survive for decades [17, 25, 80], and their
gates progress with thin protrusions from the dis- longevity may be attributable to the circumstance
tal tips into the proximal dendrites (Fig. 22.3c, d), that the soluble pathological tau material is pri-
thereby emphasizing the distinction between dis- marily generated within the dendritic compart-
tal and proximal dendritic segments [30, 56, 60, ment while leaving the axonal cytoskeleton
77]. Apparently, synaptic contacts on proximal virtually unimpaired.
22  Top-Down Projections Direct the Gradual Progression of Alzheimer-Related Tau Pathology… 299

It is seldom that one finds a lone AT8-­ (Fig. 22.2). Connections from D to C and C to B
immunopositive dendrite in the neuropil. Instead, (Fig. 22.2) achieve functional maturity along with
as a rule, all of the distal dendrites of an initially their developing myelin sheath only postnatally
involved dendritic tree belonging to an affected under the influence of factors that reach the human
projection neuron are AT8-immunoreactive organism from the outside world and enhance the
(Fig. 22.3a, b). This finding supports the assump- activity of the top-down projection cells [23].
tion that a single seed-bearing top-down axon, The sAD-related pathological process as a
including its numerous terminal collaterals, syn- whole clearly reflects the phylo- and ontogenetic
apses on all dendrites of a single target neuron and development of the human brain. As such, it is
refrains from contacting other nerve cells in the understandable why projection neurons that com-
immediate vicinity. The mechanisms of this one- mence axonal myelination early and develop a
to-one selectivity require more study; neverthe- sturdy myelin sheath prenatally are spared [25].
less, it is conceivable that axons of late-­maturing By contrast, projection cells that begin their axo-
pyramidal cells establish connections with late- nal myelination late (postnatally)  – such as the
emerging distal dendrites of neocortical pyramidal majority of the top-down cells and their target
cells in distant cortical regions in a highly ordered neurons – are highly susceptible. The molecular
rather than haphazard manner [113]. backgrounds of these relationships are not fully
This top-down one-to-one neuronal selectivity understood. It should be pointed out, however,
may underlie a capacity for a point-to-point that it is also the activity of a projection neuron
transmission of data. Traditionally, top-down that provides the physiological stimulus for oli-
neurons are thought to send their axons in a dif- godendroglia cells to produce and sustain its
fuse manner to numerous and different neuronal myelin sheath [16, 107, 109]. The greater the
constituents of their target areas, with the excep- degree of neuronal activity and the thicker the
tion of the layer IV spiny stellate cells [13, 79]. caliber of the myelin sheath during neuronal
Our findings, by contrast, point to the existence development, the better the nerve cell is protected
of a highly selective involvement limited only to against the sAD process.
the distal dendrites of a single target cell The question arises whether the ultimate mat-
(Figs. 22.3a, b and 22.4e). This selectivity could uration of the top-down connectivities (Fig. 22.2,
increase the toxic effects of transmitted abnormal from D to C and C to B) can be effectively pro-
tau seeds and might explain in part why the moted by placing greater functional demands on
abnormal tau protein spreads so effectively from them in childhood and early adulthood during the
neuron to neuron and why the progressive patho- postnatal myelination phase [16, 23, 36, 51, 100,
logical process associated with sAD fails to go 103, 107, 115]. When toddlers and young chil-
into remission. dren acquire their first language, the cytoarchi-
tectural areas of the superior temporal gyrus are
especially challenged functionally. These possi-
Late-Maturing Top-Down Neurons bly manifest themselves in the strikingly robust
and Their Target Cells Develop myelination of the superior temporal gyrus com-
the Earliest Tau Changes. Do They pared to that of the remaining regions of the tem-
Deliver Directives for Slowing poral lobe [64, 86]. As pointed out already above,
the Pathological Process? the predictable progression of the pathological
process is subject to a noticeable delay at the bor-
Top-down neurons and their target cells become der between the medial and superior temporal
involved sequentially during sAD, beginning with gyri. The more intensively the late-maturing neu-
those that establish connections from D to C and rons are used, the more successful their myelina-
followed by those from C to B and from B to A, tion and, perhaps, also the chances of delaying
i.e., neocortical pyramidal cells that mature latest the sAD-related process, so that the onset of the
are the first to develop tau pathology [27] clinical phase could be postponed.
300 H. Braak and K. Del Tredici

Acknowledgments This chapter was made possible by 13. Barbas H. Specialized elements of orbitofrontal cor-
support from the Hans and Ilse Breuer Foundation (Frankfurt tex in primates. Ann N Y Acad Sci. 2007;1121:10–32.
am Main, Germany) and the Braak Collection (Frankfurt am 14. Barbas H, Rempel-Clower N.  Cortical structure
Main, Germany). We thank Ms. Simone Feldengut (immu- predicts the pattern of corticocortical connections.
noreactions) and Mr. David Ewert (University of Ulm) for Cereb Cortex. 1997;7:635–46.
technical assistance with the graphics. 15. Bartzokis G. Age-related myelin breakdown: a devel-
opmental model of cognitive decline and Alzheimer’s
disease. Neurobiol Aging. 2004;25:5–18.
16. Bengtsson SL, Nagy Z, Skare S, Forsman L,
References Forssberg H, Ullén F. Extensive piano practicing has
regionally specific effects on white matter develop-
1. Agnati LF, Bjelke B, Fuxe K.  Volume versus wir- ment. Nat Neurosci. 2005;8:1140–50.
ing transmission in the brain: a new theoretical 17. Bobinski M, Wegiel J, Tarnawski M, de Leon MJ,
frame of neuropsychopharmacology. Med Res Rev. Reisberg B, Miller DC, Wisniewski HM. Duration of
1995;15:33–45. neurofibrillary changes in the hippocampal pyrami-
2. Ahmed Z, Cooper J, Murray TK, Garn K, dal neurons. Brain Res. 1998;799:156–8.
McNaughton E, Clark H, Parhizkar S, Ward MA, 18. Braak H, Braak E.  On areas of transition between
Cavallini A, Jackson S, Bose S, Clavaguera F, Tolnay entorhinal allocortex and temporal isocortex in the
M, Lavenir I, Goedert M, Hutton ML, O’Neill MJ. A human brain. Normal morphology and lamina-­
novel in  vivo model of tau propagation with rapid specific pathology in Alzheimer’s disease. Acta
and progressive neurofibrillary tangle pathology: the Neuropathol. 1985;68:325–32.
pattern of spread is determined by connectivity, not 19. Braak H, Braak E.  Neuropathological stageing
proximity. Acta Neuropathol. 2014;127:667–83. of Alzheimer-related changes. Acta Neuropathol.
3. Alonso AC, Li B, Grundke-Iqbal I, Iqbal 1991;82:239–59.
K.  Mechanism of tau-induced neurodegeneration 20. Braak H, Braak E.  Development of Alzheimer-­
in Alzheimer disease and related tauopathies. Curr related neurofibrillary changes in the neocortex
Alzheimer Res. 2008;5:375–84. inversely recapitulates cortical myelogenesis. Acta
4. Amieva H, Le Goff M, Millet X, Orgogozo JM, Neuropathol. 1996;92:97–101.
Pérès K, Barberger-Gateau A, Jacqmin-Gadda 21. Braak H, Del Tredici K. Poor and protracted myelin-
H, Dartigues JF.  Prodromal Alzheimer’s disease: ation as a contributory factor to neurodegenerative
successive emergence of clinical symptoms. Ann disorders. Neurobiol Aging. 2004;25:19–23.
Neurol. 2008;64:492–8. 22. Braak H, Del Tredici K.  Alzheimer’s pathogen-
5. Amunts K, Zilles K.  Architectonic mapping of esis: is there neuron-to-neuron propagation? Acta
the human brain beyond Brodmann. Neuron. Neuropathol. 2011;121:589–95.
2015;88:1086–107. 23. Braak H, Del Tredici K.  Alzheimer’s disease:
6. Arendt T. Alzheimer’s disease as a loss of differenti- pathogenesis and prevention. Alzheimers Dement.
ation control in a subset of neurons that retain imma- 2012;8:227–33.
ture features in the adult brain. Neurobiol Aging. 24. Braak H, Del Tredici K. Where, when, and in what
2000;21:783–96. form does sporadic Alzheimer’s disease begin? Curr
7. Arendt T, Brückner MK, Gertz HJ, Marcova Opin Neurol. 2012;25:708–14.
L. Cortical distribution of neurofibrillary tangles in 25. Braak H, Del Tredici K. Neuroanatomy and pathol-
Alzheimer’s disease matches the pattern of neurones ogy of sporadic Alzheimer’s disease. Adv Anat
that retain their capacity of plastic remodelling in the Embryol Cell Biol. 2015;215:1–162.
adult brain. Neuroscience. 1998;83:991–1002. 26. Braak H, Del Tredici K. The preclinical phase of the
8. Arendt T, Stieler JT, Holzer M. Tau and tauopathies. pathological process underlying Alzheimer’s dis-
Brain Res Bull. 2016;126:238–92. ease. Brain. 2015;138:2814–33.
9. Arendt T, Stieler J, Ueberham U.  Is sporadic 27. Braak H, Del Tredici K.  Spreading of tau pathol-
Alzheimer’s disease a developmental disorder? ogy in sporadic Alzheimer’s disease along cortico-­
J Neurochem. 2017;143:396–408. cortical top-down connections. Cereb Cortex.
10. Arnold SE, Hyman BT, Flory J, Damasio AR, Van 2018;28:3372–84.
Hoesen GW. The topographical and neuroanatomi- 28. Braak H, Alafuzoff I, Arzberger T, Kretzschmar
cal distribution of neurofibrillary tangles and neu- H, Del Tredici K.  Staging of Alzheimer disease-­
ritic plaques in the cerebral cortex of patients with associated neuropathology using paraffin sections
Alzheimer’s disease. Cereb Cortex. 1991;1:103–16. and immunocytochemistry. Acta Neuropathol.
11. Avila J.  Tau phosphorylation and aggregation 2006;112:3894–04.
in Alzheimer’s disease pathology. FEBS Lett. 29. Braak H, Thal DR, Ghebremedhin E, Del Tredici
2006;580:2922–7. K. Stages of the pathological process in Alzheimer’s
12. Bannister AP.  Inter- and intra-laminar connections disease: age categories from 1 to 100 years.
of pyramidal cells in the neocortex. Neurosci Res. J Neuropathol Exp Neurol. 2011;70:960–9.
2005;53:95–103.
22  Top-Down Projections Direct the Gradual Progression of Alzheimer-Related Tau Pathology… 301

30. Branco T, Häusser M. Synaptic integration gradients hippocampal neurons in the rat in situ and in cell
in single cortical pyramidal cell dendrites. Neuron. culture. Neuroscience. 1987;23:121–30.
2011;69:885–92. 45. Dugger BN, Hildalgo JA, Chiarolanza G, Mariner
31. Brookmeyer R, Johnson E, Ziegler-Graham JM, Henry-Watson J, Sue LI, Beach TG.  The dis-
K, Arrighi HM.  Forecasting the global burden tribution of phosphorylated tau in spinal cords of
of Alzheimer’s disease. Alzheimers Dement. Alzheimer's disease and non-demented individuals.
2007;3:186–91. J Alzheimers Dis. 2013;34:529–36.
32. Brundin P, Melki R, Kopito R. Prion-like transmis- 46. Dujardin S, Lécolle K, Caillierez R, Bégard S,
sion of protein aggregates in neurodegenerative dis- Zommer N, Lachaud C, Carrier S, Dufour N,
eases. Nat Rev Mol Cell Biol. 2010;11:301–7. Aurégan G, Winderickx J, Hantraye P, Déglon N,
33. Bullier J.  Feedback connections and conscious Colin M, Buée L.  Neuron-to-neuron wild-type tau
vision. Trends Cogn Sci. 2001;5:369–570. protein transfer through a trans-synaptic mechanism:
34. Bussiere T, Giannakopoulos P, Bouras C, Perl DP, relevance to sporadic tauopathies. Acta Neuropathol
Morrison JH, Hof PR.  Progressive degeneration of Commun. 2014;2:1–14.
nonphosphorylated neurofilament protein-enriched 47. Duyckaerts C, Braak H, Brion J-P, Buée L, Del
pyramidal neurons predicts cognitive impairment in Tredici K, Goedert M, Halliday G, Neumann M,
Alzheimer’s disease: stereologic analysis of prefron- Spillantini MG, Tolnay M, Uchihara T.  PART
tal cortex area 9. J Comp Neurol. 2003;463:281–302. is part of Alzheimer disease. Acta Neuropathol.
35. Calafate S, Buist A, Miskiewicz K, Vijayan V, 2015;129:749–56.
Daneels G, de Strooper B, de Wit J, Verstreken K, 48. Elobeid A, Soininen H, Alafuzoff
Moechars D. Synaptic contacts enhance cell-to-cell I.  Hyperphosphorylated tau in young and middle-­
tau pathology propagation. Cell Rep. 2015;11:1–8. aged subjects. Acta Neuropathol. 2011;123:97–104.
36. Cass SP. Alzheimer’s disease and exercise: a litera- 49. Felleman DJ, van Essen DC. Distributed hierarchi-
ture review. Curr Sports Med Rep. 2017;16:19–22. cal processing in the primate cerebral cortex. Cereb
37. Clavaguera F, Bolmont T, Crowther RA, Cortex. 1991;1:1–47.
Abramowski A, Frank S, Probst A, Fraser G, Stalder 50. Frahm-Falkenberg S, Ibsen R, Kjellberg J, Jennum
AK, Beibel M, Staufenbiel M, Jucker M, Goedert P.  Health, social and economic consequences of
M, Tolnay M.  Transmission and spreading of dementias: a comparative national cohort study. Eur
tauopathy in transgenic mouse brain. Nat Cell Biol. J Neurol. 2016;23:1400–7.
2009;11:909–13. 51. Freedman M, Alladi S, Chertkow H, Bialystok E,
38. De Calignon A, Polydoro M, Suárez-Calvet M, Craik FI, Phillips NA, Duggirala V, Raju SB, Bak
William C, Adamowicz DH, Kopeikina KJ, Pitstick TH. Delaying onset of dementia: are two languages
R, Sahara N, Ashe KH, Carlson GA, Spires-Jones enough? Behav Neurol. 2014;2014:808137.
TL, Hyman BT.  Propagation of tau pathology in 52. García-Cabezas MÁ, Zikopoulos B, Barbas H.
a model of early Alzheimer’s disease. Neuron. The structural model: a theory linking connec-
2012;73:685–97. tion, plasticity, pathology, development and evo-
39. Deb A, Thornton JD, Sambamoorthi U, Innes lution of the cerebral cortex. Brain Struct Funct.
K. Direct and indirect cost of managing Alzheimer’s 2019;224:985–1008.
disease and related dementias in the United 53. German DC, White CL, Sparkman DR. Alzheimer’s
States. Expert Rev Pharmaoecon Outcomes Res. disease: neurofibrillary tangles in nuclei that
2017;17:189–202. project to the cerebral cortex. Neuroscience.
40. DeFelipe J, Farinas I. The pyramidal neuron of the 1987;21:305–12.
cerebral cortex: morphological and chemical char- 54. Goedert M, Spillantini MG. A century of Alzheimer’s
acteristics of the synaptic inputs. Prog Neurobiol. disease. Science. 2006;314:777–81.
1992;39:563–607. 55. Goedert M, Clavaguera F, Tolnay M. The propaga-
41. DeFelipe J, Alonso-Nanclares L, Arellano tion of prion-like protein inclusions in neurodegen-
JI.  Microstructure of the neocortex: comparative erative diseases. Trends Neurosci. 2010;33:317–25.
aspects. J Neurocytol. 2002;31:299–316. 56. Gordon U, Polsky A, Schiller J. Plasticity compart-
42. Delatour B, Blanchard V, Pradier L, Duyckaerts ments in basal dendrites of neocortical pyramidal
C.  Alzheimer pathology disorganizes cortico-­ neurons. J Neurosci. 2006;26:12717–26.
cortical circuitry: direct evidence from a transgenic 57. Grundke-Iqbal I, Iqbal K, Tung Y, Quinlan M,
animal model. Neurobiol Dis. 2004;16:41–7. Wisniewski HM, Binder LI.  Abnormal phosphory-
43. Deoni SC, Dean DC, Remer J, Dirks H, lation of the microtubule-associated protein τ (tau)
O’Muircheartaigh J.  Cortical maturation and in Alzheimer cytoskeletal pathology. Proc Natl Acad
myelination in healthy toddlers and young children. Sci U S A. 1986;83:4913–7.
Neuroimage. 2015;115:147–61. 58. Grydeland H, Walhovd KB, Tamnes CK, Westlye
44. Dotti CG, Banker GA, Binder LI.  The expression LT, Fjell AM.  Intracortical myelin links with per-
and distribution of the microtubule-associated pro- formance variability across the human lifespan:
teins tau and microtubule-associated protein 2  in results from T1- and T2-weighter MRI myelin
302 H. Braak and K. Del Tredici

mapping and diffusion tensor imaging. J  Neurosci. 74. Mandelkow EM, Mandelkow E.  Biochemistry
2013;33:18618–30. and cell biology of tau protein in neurofibrillary
59. Guo JL, Lee VM. Seeding of normal tau by pathologi- degeneration. Cold Spring Harb Perspect Med.
cal tau conformers drives pathogenesis of Alzheimer-­ 2012;2:a006247.
like tangles. J Biol Chem. 2011;286:15317–31. 75. Marin-Padilla M.  Ontogenesis of the pyramidal
60. Häusser M, Mel B. Dendrites: bug or feature? Curr cell of the mammalian neocortex and developmen-
Opin Neurobiol. 2003;13:372–83. tal cytoarchitectonics: a unifying theory. J  Comp
61. Hochstein S, Ahissar M. View from the top: hierar- Neurol. 1992;321:223–40.
chies and reverse hierarchies in the visual system. 76. Markram H. A network of tufted layer 5 pyramidal
Neuron. 2002;36:791–804. neurons. Cereb Cortex. 1997;7:523–33.
62. Hof PR, Morrison JH.  The aging brain: morpho- 77. Merino-Serrais P, Benavides-Piccione R, Blazquez-­
molecular senescence of cortical circuits. Trends Llorca L, Kastanauskaite A, Rábano A, Avila J,
Neurosci. 2004;27:607–13. DeFelipe J.  The influence of phospho-tau on den-
63. Hoover BR, Reed MN, Su J, Penrod RD, Kotilinek dritic spines of cortical pyramidal neurons in patients
LA, Grant MK, Pitstick R, Carlson GA, Lanier with Alzheimer’s disease. Brain. 2013;136:1913–28.
LM, Yuan LL, Ashe KH, Liao D.  Tau mislocaliza- 78. Meyer K.  Primary sensory cortices, top-down pro-
tion to dendritic spines mediates synaptic dysfunc- jections and conscious experience. Prog Neurobiol.
tion independently of neurodegeneration. Neuron. 2011;94:408–17.
2010;68:1067–81. 79. Meyer K. The role of dendritic signaling in the anes-
64. Hopf A.  Photometric studies on the myeloarchi- thetic suppression of consciousness. Anesthesiology.
tecture of the human temporal lobe. J  Hirnforsch. 2015;122:1415–31.
1968;10:285–97. 80. Morsch R, Simon W, Coleman PD.  Neurons
65. Hyman BT, Goméz-Isla T.  Alzheimer’s disease is may live for decades with neurofibrillary tangles.
a laminar, regional, and neural system specific dis- J Neuropathol Exp Neurol. 1999;58:188–97.
ease, not a global brain disease. Neurobiol Aging. 81. Mudher A, Colin M, Dujardin S, Medina M,
1994;15:353–4. Dewachter I, Alavi Naini SM, Mandelkow EM,
66. Hyman BT, Phelps CH, Beach TG, Bigio EH, Mandelkow E, Buée L, Goedert M, Brion JP. What
Cairns NJ, Carrillo MC, Dickson DW, Duyckaerts is the evidence that tau pathology spreads through
C, Frosch MP, Masliah E, Mirra SS, Nelson PT, prion-like propagation? Acta Neuropathol Commun.
Schneider JA, Thal DR, Thies B, Trojanowski JQ, 2017;5:99–119.
Vinters HV, Montine TJ.  National Institute on 82. Nelson PT, Jicha GA, Schmitt FA, Liu H, Davis
Aging-Alzheimer’s association guidelines for the DG, Mendiondo MS, Abner EL, Markesbery
neuropathologic assessment of Alzheimer’s disease. WR.  Clinicopathologic correlations in a large
Alzheimers Dement. 2012;8:1–13. Alzheimer disease center autopsy cohort: neuritic
67. Iqbal K, Liu F, Gong CX, Alonso C, Grundke-Iqbal plaques and neurofibrillary tangles “do count” when
I.  Mechanisms of tau-induced neurodegeneration. staging disease severity. J Neuropathol Exp Neurol.
Acta Neuropathol. 2009;118:53–69. 2007;66:1136–46.
68. Ittner LM, Götz J. Amyloid-β and tau – a toxic pas 83. Nieuwenhuys R. The neocortex. An overview of its
de deux in Alzheimer’s disease. Nat Rev Neurosci. evolutionary development, structural organization
2011;12:65–72. and synaptology. Anat Embryol. 1994;190:307–37.
69. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A, 84. Nieuwenhuys R.  The greater limbic system, the
van Eersel J, Wölfing H, Chieng BC, Christie MJ, emotional motor system and the brain. Prog Brain
Napier IA, Eckert A, Staufenbiel M, Hardeman E, Res. 1996;107:551–80.
Götz J. Dendritic function of tau mediates amyloid-­ 85. Nieuwenhuys R.  Comparative aspects of vol-
beta toxicity in Alzheimer’s disease mouse models. ume transmission, with sidelight on other forms
Cell. 2010;142:387–97. of intercellular communication. Prog Brain Res.
70. Kaufman S, Del Tredici K, Thomas TL, Braak 2000;125:49–126.
H, Diamond MI.  Tau seeding activity anticipates 86. Nieuwenhuys R, Broere CA.  A map of the human
phospho-tau pathology in Alzheimer’s disease. Acta neocortex showing the estimated overall myelin
Neuropathol. 2018;136:57–67. content of the individuals architectonic areas based
71. Khatoon S, Grundke-Iqbal I, Iqbal K. Levels of nor- on the studies of Adolf Hopf. Brain Struct Funct.
mal and abnormally phosphorylated tau in different 2017;222:465–80.
cellular and regional compartments of Alzheimer dis- 87. O’Donnell J, Zeppenfeld D, McConnell E, Pena S,
ease and control brains. FEBS Lett. 1994;94:80–4. Nedergaard M. Norepinephrine: a neuromodulator that
72. Kopelkina KJ, Hyman BT, Spires-Jones TL. Soluble boosts the function of multiple cell types to optimize
forms of tau are toxic in Alzheimer’s disease. Transl CNS performance. Neurochem Res. 2012;37:2496–512.
Neurosci. 2012;3:223–33. 88. Paus T, Collins DL, Evans AC, Leonard G, Pike
73. Liu L, Drouet V, Wu JW, Witter MP, Smith SA, B, Zijdenbos A.  Maturation of white matter in the
Clelland C, Duff K.  Trans-synaptic spread of tau human brain: a review of magnetic resonance stud-
pathology in vivo. PLoS One. 2012;7:e31302. ies. Brain Res Bull. 2001;54:255–66.
22  Top-Down Projections Direct the Gradual Progression of Alzheimer-Related Tau Pathology… 303

89. Pearson RCA. Cortical connections and the pathol- tudinal, population-based sample of aging. Ann
ogy of Alzheimer's disease. Neurodegeneration. Neurol. 2007;62:406–13.
1996;5:429–34. 103. Swaab DF, Dubelaar EJ, Hofman MA, Scherder
90. Purohit DP, Batheja NO, Sano M, Jashnani KD, EJ, van Someren EJ, Verwer RW.  Brain aging and
Kalaria RN, Karunamurthy A, Kaur S, Shenoy Alzheimer’s disease; use it or lose it. Prog Brain Res.
AS, Van Dyk K, Schmeidler J, Perl DP.  Profiles 2002;138:343–73.
of Alzheimer’s disease-related pathology in an 104. Tai HC, Serrano-Pozo A, Hashimoto T, Frosch MP,
aging urban population sample. J  Alzheimers Dis. Spires-Jones TL, Hyman BT.  The synaptic accu-
2011;24:187–96. mulation of hyperphosphorylated tau oligomers in
91. Rapoport SI.  Hypothesis: Alzheimer’s disease Alzheimer disease is associated with dysfunction
is a phylogenetic disease. Med Hypotheses. of the ubiquitin-proteasome system. Am J  Pathol.
1989;29:147–50. 2012;181:14263–5.
92. Rapoport SI.  Integrated phylogeny of the primate 105. Tashiro K, Hasegawa M, Ihara Y, Iwatsubo
brain, with special reference to humans and their T.  Somatodendritic localization of phosphory-
diseases. Brain Res Rev. 1990;15:267–94. lated tau in neonatal and adult rat cerebral cortex.
93. Reisberg B, Franssen EH, Hasan SM, Monteiro I, Neuroreport. 1997;8:2797–801.
Boksay I, Souren LE, Kenowsky S, Auer SR, Elahi 106. Thies E, Mandelkow EM.  Missorting of tau in
S, Kruger A. Retrogenesis: clinical, physiologic, and neurons causes degeneration of synapses that can
pathologic mechanisms in brain aging, Alzheimer’s be rescued by kinase MARK2/Par-1. J  Neurosci.
and other dementing processes. Eur Arch Psychiatry 2007;27:2896–907.
Clin Neurosci. 1999;249(Suppl 3):28–36. 107. Ullén F. Is activity regulation of late myelination a
94. Reisberg B, Franssen EH, Souren LE, Auer SR, plastic mechanism in the human nervous system?
Akram I, Kenowsky S.  Evidence and mechanisms Neuron Glia Biol. 2009;5:29–34.
of retrogenesis in Alzheimer’s and other dementias: 108. Van Dam D, De Deyn PP.  Non human primate
management and treatment import. Am J Alzheimers models for Alzheimer’s disease-related research
Dis Other Demen. 2002;17:2021–2. and drug discovery. Expert Opin Drug Discovery.
95. Rockland KS, Pandya DN. Laminar origins and ter- 2017;12:187–200.
minations of cortical connections of the occipital lobe 109. van der Knaap MS, Valk J, Bakker CJ, Schooneveld
in the rhesus monkey. Brain Res. 1979;179:3–20. M, Faber JA, Willemse J, Gooskens RH. Myelination
96. Rüb U, Del Tredici K, Schultz C, Thal DR, as an expression of the functional maturity of the
Braak E, Braak H.  The evolution of Alzheimer’s brain. Dev Med Child Neurol. 1991;33:849–57.
disease-­related cytoskeletal pathology in the 110. Velasco ME, Smith MA, Siedlak SI, Nunomura
human raphe nuclei. Neuropathol Appl Neurobiol. A, Perry G.  Striation is the characteristic neu-
2000;26:553–67. ritic abnormality in Alzheimer disease. Brain Res.
97. Rubial-Álvarez S, de Sola S, Machado MC, Sintas 1998;813:329–33.
E, Böhm P, Sánchez-Benavides G, Langohr K, 111. von Bergen M, Barghorn S, Biernat J, Mandelkow
Muñiz R, Peña-Casanova J.  The comparison of EM, Mandelkow E. Tau aggregation is driven by a
cognitive and functional performance in children transition from random coil to beta sheet structure.
and Alzheimer’s disease supports the retrogenesis Biochim Biophys Acta. 2005;1739:158–66.
model. J Alzheimers Dis. 2013;33:191–203. 112. Wozniak JR, Lim KO.  Advances in white matter
98. Sabbagh MN, Cooper K, DeLange J, Stoehr JD, imaging: a review of in  vivo magnetic resonance
Thind K, Lahti T, Reisberg B, Sue L, Vedders L, methodologies and their applicability to the study
Fleming SR, Beach TG.  Functional, global and of development and aging. Neurosci Biobehav Rev.
cognitive decline correlates to accumulation of 2006;30:762–74.
Alzheimer’s pathology in MCI and AD.  Curr 113. Yogev S, Shen K.  Cellular and molecular mecha-
Alzheimer Res. 2010;7:280–6. nisms of synaptic specificity. Annu Rev Cell Dev
99. Saper CB, Wainer BH, German DC.  Axonal and Biol. 2014;30:417–37.
transneuronal transport in the transmission of neu- 114. Zempel H, Thies E, Mandelkow E, Mandelkow
rological disease: potential role in system degenera- EM. Aβ oligomers cause localized Ca2+ elevation,
tions, including Alzheimer’s disease. Neuroscience. missorting of endogeneous tau into dendrites, tau
1987;23:3893–8. phosphorylation, and destruction of microtubules
100. Schlaug G, Norton A, Overy K, Winner E.  Effects and spines. J Neurosci. 2010;30:11938–50.
of music training on the child's brain and cognitive 115. Zhu QB, Bao AM, Swaab D. Activation of the brain
development. Ann N Y Acad Sci. 2003;1060:219–30. to postpone dementia: a concept originating from
101. Serrano-Pozo A, Frosch MP, Masliah E, Hyman postmortem human brain studies. Neurosci Bull.
BT.  Neuropathological alterations in Alzheimer dis- 2019;35:253–66.
ease. Cold Spring Harb Perspect Med. 2011;1:a006189. 116. Zilles K, Amunts K.  Architecture of the cerebral
102. Sonnen JA, Larson EB, Crane PK, Haneuse S, Li cortex. In: Mai JK, Paxinos G, editors. The human
G, Schellenberg GD, Craft S, Leverenz JB, Montine nervous system. 3rd ed. New York: Elsevier; 2012.
TJ.  Pathological correlates of dementia in a longi- p. 836–95.
 Tau Prion-Like Propagation: State
of the Art and Current Challenges 23
Simon Dujardin and Bradley T. Hyman

 au Prion-Like Propagation, State

T prion-diseases. Therefore, it was postulated that
of the Art all these pathologies could be gathered and that
all these proteins could be considered as prion-­
In Alzheimer’s disease, tau deposition generally like protein or prionoids (for review: Aguzzi and
starts in the entorhinal cortex and then, following Rajendran [3], Clavaguera et al. [22], Frost and
neuronal connections, progresses through the Diamond [45], Goedert et  al. [53], Hall and
hippocampus, the limbic and association cortices Patuto [56], Jucker and Walker [72], Walker et al.
(Fig. 23.1a) [7, 12, 13, 26, 33]. This stereotypical [151], and Dujardin [29]).
“staging scheme” was recently confirmed by tau The designation “prion-like” is subject to an
PET scan studies [60, 71, 131, 132] and strongly intense semantic debate among researchers
correlates with cognitive decline [51]. because, although it is evident that these proteins
Interestingly, although affecting different cir- share similar pathological principles, major differ-
cuits, similar progression patterns through con- ences can also be pointed out [102]. Prion diseases
nected neuronal circuits have been identified in are rare neurodegenerative pathologies character-
other tauopathies including Progressive ized by the accumulation of misfolded prion pro-
Supranuclear Palsy (PSP) (Fig.  23.1b) [149, teins. In 1982, Prusiner and collaborators ventured
155], Argyrophilic grain disease (Fig.  23.1c) the hypothesis that the protein prion itself would
[126], or more recently Pick’s disease [67]. be an infectious agent, and the unique responsible
The similarities between these pathologies as of the pathology spreading in the body and
well as the clear neuron-to-neuron progression between individuals [119]. This hypothesis was
led to the hypothesis that cerebral proteinopa- rapidly confirmed, even if the term “infectious” is
thies share similar cellular and molecular path- still controversial [87], showing that abnormal
ways of pathological protein propagation. Other prion proteins are able to convert normal prion
types of well-characterized cerebral proteinopa- proteins into the pathological form. The latter
thies that spreads through neuronal networks are form soluble oligomers and consequently amyloid
fibers. The mechanism of formation of the fibers
by recruitment of monomers by an initial oligomer
S. Dujardin · B. T. Hyman (*) (or nucleus) is called seeding (Fig.  23.2) [72].
Department of Neurology, Massachusetts General
Hospital, Harvard Medical School, MassGeneral Interestingly, different strains of prion exist with
Institute for Neurodegenerative Disease, dissimilar, and transmissible, fibril conformations
Charlestown, MA, USA [145]. Moreover, the pathological prions can be
e-mail: sdujardin@mgh.harvard.edu;
bhyman@mgh.harvard.edu

© Springer Nature Singapore Pte Ltd. 2019 305

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_23
306 S. Dujardin and B. T. Hyman

Fig. 23.1  Staging in tauopathies. Progressive appear- (b) and Argyrophilic grain disease (c). Light purple, pur-
ance of tau pathology lesions in the brain of AD (a), PSP ple and red respectively represents a small, moderate or
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 307

Fig. 23.2  Tau seeding. Tau monomers are naturally monomers in a seeding mechanism, misfolded tau pro-
soluble, unfolded and flexible (orange). In pathophysi- teins transmit their folding properties to a large number
ology context, some transient-to-stable folding can of naïve monomers and recruit them into a higher order
appear (black). In tauopathies, misfolding leads to the stable fibril. The black structures represent the core of
recruitment of several monomers into an unstable multi- the fibril as shown by Fitzpatrick and collaborators [39].
meric forms or oligomers. Through the addition of Not to scale

transferred from c­ ell-to-­cell, spreading the pathol- tional and physicochemical properties seem to
ogy in several brain areas. co-exist [11, 109, 111, 161]. Lesions also appear
This was the first description of a protein with to actively propagate across neural circuits for
such pathological properties but recently, the these two pathological proteins [43, 58, 90, 150].
hypothesis that other proteins like the amyloid These studies led the way to the hypothesis
peptide (Aβ), alpha synuclein and Tau could that tau could actively propagate trans-­
share these properties has been raised. For the synaptically. Tau propagation (or spreading) is a
amyloid peptide and synuclein, the lesions are broad term including several cellular pathways
transmissible by a seeding mechanism [27, 63, including active neuron-to-neuron transfer and
70, 73, 89, 90, 93, 100, 101, 139, 150]. Similarly the contamination of secondary cells potentially
to prions, different strains with different func- through a seeding mechanism (Fig. 23.3).

Fig. 23.1  (continued) severe amount tau pathology in the ganglia and dentate nucleus are then affected followed by
affected brain region. (a) In Alzheimer’s disease, tau affection of cerebellum, putamen, caudate nucleus and all
deposition generally starts in the entorhinal cortex and neocortical regions at the exception of temporal areas
then, following neuronal connections, progresses through (sketch inspired by Williams et al. [155] and Braak and Del
the hippocampus, the limbic and association cortices Tredici [13]). (c) In Argyrophilic grain disease, tau deposi-
(sketch inspired by Braak and Del Tredici [13]). (b) In tion starts in the ambiens gyrus and CA1 region of the hip-
PSP, tau deposition progresses from the system pallidus – pocampus before affecting amygdala, transentorhinal
subthalamic nucleus – substantia nigra to the pedonculo- cortex, subiculum and temporal lobe. In stage III, frontal
pontine nucleus and premotor cortex. The whole basal lobe is affected (sketch inspired by Saito et al. [126])
308 S. Dujardin and B. T. Hyman

Fig. 23.3  Tau prion-like propagation model. (a) In nantly located into the axonal part of neurons. (b) Tau
physiology normal tau proteins (orange) are predomi- propagation starts with an early misfolded event that
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 309

 ransmissibility of Tau Pathology, Tau

T formation of tau aggregation in neuritic processes
Seeding? around plaques. We showed that in human AD
patients, tau seeding at the synapses seems to
Several studies have investigated the transmissi- precede tau pathology [28] and that tau seeding
bility and seeding of Tau pathology (Fig.  23.2) was markedly increased when amyloid pathology
confirming this concept in vitro [44], in cellulo is present in both mouse models and human brain
[46] or in vivo [20]. After these three studies, [9]. These studies could show that amyloid
many have confirmed and reproduced this data. pathology potentiates tau propagation likely at
Briefly, Clavaguera and coworkers showed that the synaptic terminals and actively triggers its
injecting Tau aggregates extracted from mice progression outside of the hippocampal forma-
overexpressing mutated Tau to mice overexpress- tion in the cortical regions where both patholo-
ing wild-type Tau was sufficient to induce Tau gies ultimately “meet” in Alzheimer’s disease.
pathology [20]. It is interesting to note that when In parallel with data on animal models, in cel-
a Tau-immunodepleted extract is injected, no lulo, several authors showed that after the incuba-
pathology can be detected showing that Tau is the tion of aggregates, these were internalized and
responsible factor. The same group has also were able to promote aggregation of overex-
shown similar results injecting human brain pressed Tau in cell lines [46, 55, 62, 107, 129,
lysates of several tauopathies and reproducing 141, 147, 156] but also in primary neurons or
the morphology of the lesions seen in the human human iPS cells overexpressing human tau [106,
disease [19]. It is also worth noting- even if not 124, 142]. These studies showed, using truncated
reproduced so far in the literature- that they Tau species, that the microtubule-binding domain
reported the formation of tau pathology in the is necessary to promote aggregation. There is,
brain of transgenic mice injected intraperitone- however, a debate on what is the needed species
ally with Tau aggregates [21]. Many following of tau that is needed to promote tau aggregation
studies have shown injections of cerebral lysates or seeding. In a 2015 study, Mirbaha and cowork-
or synthetic fibers in tau transgenic animals, ers observed that Tau trimer is the minimal unit
potentiating the transmissibility [4, 24, 52, 65, with which a seeding mechanism could be
66, 68, 77, 79, 110, 128, 135, 138]. Several of observed [97]. The same authors in 2018 have
these studies show the presence of murine Tau in identified a monomeric tau specie could be con-
the aggregates, suggesting that aggregates are verted to a seeding competent form [96].
able to recruit endogenous Tau proteins convert- Although the minimal unit for seed-competency
ing them to a pathological form [20, 25, 65, 104]. is still subject to debate, most of the studies agree
This is also supported by the observation that tau to the idea that soluble oligomeric tau are the
pathology is transmissible to wild type (WT) ani- most seeding prone species [68, 79, 82, 83, 142].
mals that don’t usually get any sign of tau pathol-
ogy [20, 104], and by early observations that
endogenous mouse tau was incorporated into the Neuron-to-Neuron Transfer of Tau
NFT that develop in transgenic P301L human tau
overexpressing mice [130]. An interesting recent I s Tau Transferred from One Cell
study show strong tau seeding in vivo when tau to the Other?
aggregates are injected into transgenic animals Most of the models cited above analyzed the
developing β-amyloid plaques [59]. Tau seeding propagation of Tau pathology looking for cell-to-­
in this study seems to be highly dependent on the cell transfer of Tau pathology. In cell models,

Fig. 23.3  (continued) spreads through a neuron and a neuron through the synapses and (d) in the recipient neu-
large number of tau proteins accumulate into the somato- ron, recruit endogenous tau, transmit their misfolding
dendritic as well as the synaptic terminals of the neurons. properties in a seeding mechanism and spread the pathol-
(c) Misfolded tau proteins are transferred from neuron-to- ogy. Not to scale
310 S. Dujardin and B. T. Hyman

both the (presumably misfolded or posttransla- to 3R and mutants [14, 30, 32]. Interestingly,
tionally modified) WT proteins [32] and aggre- we showed in two different models that the
gated truncated Tau proteins seem to transfer to presence of endogenous tau was not required
secondary cells [15, 46, 78, 160]. Models of for tau cell-to-cell transfer. However, mouse tau
microfluidic axonal isolations strongly suggest was needed to observe pathological changes in
that Tau is trans-synaptically transferred and that secondary neurons [154] again arguing for the
synapses are required for neuron-to-neuron trans- importance of endogenous tau in the pathologi-
fer [15, 32, 142, 152]. In other cell models, trans- cal recruitment of tau. Worth noting, Tau cell-
fer of tau-containing conditioned medium seems to-cell transfer was also observed in vivo in one
to be sufficient to transfer tau to secondary cells model of lamprey [80, 85]. All together these
suggesting that cell-to-cell contact or the close studies demonstrate that tau can be transferred
synaptic environment might not be a prerequisite neuron-­to-­neuron (and, more generally, in some
for tau uptake by cells [15, 78, 160]. We hypoth- instances cell-to-cell) which leads to experi-
esize that the synaptic uptake therefore is at least ments aimed at understanding the cellular path-
in part due to the relative proximity of the pre and ways implicated in this process (Fig. 23.4).
post synaptic elements as well as the potentially
high local concentrations of any tau released into Routes of Cell-to-Cell Transfer
the extremely small perisynaptic space. Different routes of cell-to-cell transfer of tau
Cell-to-cell transfer of tau was also suggested in have been proposed. In principle, cell to cell
in vivo models in which Tau pathology evolves transfer of an intracellular protein like Tau could
from the injection site to closely connected areas occur either through release into the extracellular
[4, 20, 65, 138]. However, it is difficult from these space, or via direct cell-to-cell contact. Regarding
particular models to firmly conclude and determine the extracellular space, it necessitates two dis-
the part of active propagation of aggregates versus tinct mechanisms: Tau secretion and tau uptake
diffusion area of the injection. In order to answer (Fig. 23.4).
this question, in two independent studies, a trans-
genic model using the neuropsin promoter allow- (a) Tau secretion
ing for the overexpression of mutated Tau
specifically in the entorhinal cortex has been devel- Tau protein is a cytosolic protein thus its secre-
oped and used to convincingly show the propaga- tion via the conventional exocytosis implicating
tion of tau pathology from the entorhinal cortex to the Golgi apparatus is hardly conceivable.
recipient neurons in the hippocampus [25, 88]. Therefore, several studies examined different
In addition, viral vector-mediated focal routes of secretion and notably via extracellular
overexpression of tau proteins either in the hip- vesicles. Two types of extracellular vesicles have
pocampus of rats or the entorhinal cortex of particularly been shown to be involved in tau
mice clearly and undoubtfully show that tau secretion: microvesicles (also called microparti-
proteins can be cell-to-cell transported across cles or ectosomes) and exosomes. The generation
long distances in accord with connectional neu- of microvesicles is still poorly understood but
roanatomical connections [30, 32, 153, 154]. they are large vesicles (between 100 nm to 1 μm
The advantage of viral vector-mediated systems diameter) directly shed from plasma membrane.
is that we can easily test different constructs The release of exosomes is better characterized:
and species of tau. Indeed, we could compare during endocytosis, invagination of the plasma
tau propagation of both mutant and wild-type membrane leads to the formation of the primary
tau as well as 3R versus 4R tau. We found that endosome. Other invaginations occur in the mem-
if the pathological conversion was faster for brane of this endosome leading to the formation
mutant tau, the long-distance propagation of of a multivesicular body. During this step, the
tau pathological species was more efficient intraluminal vesicles trap cytoplasmic material.
when 4R wild-type tau was used as compared When the multivesicular body fuses with the
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 311

Fig. 23.4  Mechanisms of cell-to-cell transfer. Different cytosis (brown) has also been a proposed mechanism but
pathways of neuron-to-neuron transfer have been pro- is largely questionable. Tau can also be transferred
posed and observed. Tau is secreted into the extracellular between cells via nanotubes (light blue). Uptake of tau
space via translocation through the plasma membrane largely happens through endocytosis (clathrin-mediated
(Purple), extracellular vesicles (Ectosomes – pink or exo- or clathrin-independent) and regulated by heparan sulfate
somes – red originating from the fusion of multivesicular proteoglycans (HSPGs) (orange). Tau aggregates are able
bodies (MVB) with the plasma membrane). Regular exo- to escape the endosome to reach the cytoplasm. (Sketch
inspired by Dujardin [29] and Mudher et al. [102])

plasma membrane, the intraluminal vesicles are ent models [37, 75, 116, 129, 143]. These discrep-
released and called exosomes. Exosomes are ancies probably rely on the models used but also
largely characterized (for review: Kowal et  al. on the intracellular concentration of tau as it
[81] and Rajendran et al. [120]) and were shown seems that tau is re-directed to exosomes when
to carry prion [38], Aβ peptide [121] and synu- there is an accumulation in the cytoplasm.
clein [5, 84]. Moreover, it was shown that exo- Contrariwise, tau seems to be present physiologi-
somes convey some lipids which potentiate cally in larger microvesicles [31] (Fig. 23.4).
fibrillogenesis and accelerates the formation of Regardless of whether tau is present in extra-
amyloid structures [162]. In the case of tauopa- cellular vesicles, several studies showed that
thies, the implication of extracellular vesicles extracellular tau is mostly found free in the
remains somewhat dependent on techniques and medium, not -associated with vesicles [31, 76].
model systems used. Indeed, we and other authors Several studies have proposed different secretory
describe the presence of Tau in both microvesicles pathways by which tau reaches the extracellular
[31] and exosomes purified from cells overex- space free. Katsinelos and coworkers found that
pressing Tau and from patient-derived cerebrospi- phosphorylated tau can potentially be translo-
nal fluid [8, 31, 127, 134, 152], whereas other cated across the plasma membrane via a type I
studies failed to reproduce these results in differ- unconventional protein secretion mechanism
312 S. Dujardin and B. T. Hyman

(Fig.  23.4) [76]. Interestingly, chaperone-­ endocytosis either clathrin-­independent, likely

mediated translocation through membranes through macropinocytosis, or clathrin-mediated
seems to lead to secretion as well in another type [16, 35, 61, 94]. The uptake of free tau aggre-
of unconventional secretion mechanism [41]. gates seems to be mediated by heparan sulfate
Another study by Tang and colleagues showed [61, 122, 140] even if it is likely that other
that in case of overexpression, Tau seems to relo- receptors are also implicated in these mecha-
calize, thanks to the mTor kinase, in autophagic nisms (Figs. 23.4 and 23.5).
vesicles and in the endoplasmic reticulum, which This uptake mechanism through the endocytic
could suggest also the existence of a mechanism pathway might actively participate in the patho-
of exocytosis [143]. Interestingly, in cellulo, Rab logical propagation mechanism (Fig.  23.5).
GTPase regulates tau secretion in the extracellu- Indeed, three independent studies demonstrated
lar space arguing for an active tau secretion that tau aggregates can induce a rupture of the
mechanism rather than a potential passive diffu- endomembrane [16, 35, 40] and reach the cyto-
sion through the plasma membrane [98, 125]. plasm. Calafate and coworkers report that BIN1,
Additionally, the presence of tau in the extracel- a genetic risk factor for late-onset Alzheimer’s
lular media is stimulated by neuronal activity disease, negatively regulates this endocytic path-
[116, 158, 159]. Interestingly, Tau proteins seem way and that a loss of function of BIN1 subse-
to be mainly dephosphorylated and truncated in quently results in more uptake and more
their carboxy-terminal part when secreted [74, underlying tau seeding of aggregates [16]. When
80, 99, 105, 113, 116, 136]. However, truncation the endosomal membrane is ruptured autophagy
of Tau is not retrieved in every model [18]. The pathways are activated through the galectin-8 and
mechanisms of secretion described above were NDP52 receptors [35]. Stress-mediated or phar-
mainly identified in vitro and it is likely that part macological inhibition of autophagy also potenti-
of them represent a physiological secretion rather ate tau aggregation arguing for a key role of
than pathological processes [18, 116]. In human autophagy in tau pathology propagation [35,
CSF, where tau is used as a biomarker of 133]. Michel and co-workers even suggest that
Alzheimer disease, both freely soluble and exo- tau aggregation from monomers can happen
some containing pools have been described [10, directly inside the endocytosis vesicles and
108, 127, 152]. thereby favor tau pathology spreading [94]
(Fig.  23.5). Worth noting, exosomes containing
(b) Tau uptake tau proteins are also endocytosed and hijacking
the endosomal pathway are transferred to several
Several studies have shown that tau placed in connected neurons [114] (Fig. 23.5).
the extracellular medium is internalized into the
soma of secondary cells arguing for the exis- (c) Nanotubes
tence of one or several mechanism(s) of uptake
[15, 46, 78, 106, 142, 157, 160]. Uptake of tau Although a secretory pathway is likely to take
has also been observed in vivo [17, 142]. Tau part in the propagation of Tau pathology, another
monomers and oligomers are also efficiently route has been identified: nanotubes [1, 144]
internalized by human iPSC-derived neurons (Fig.  23.4). Nanotubes are filamentous-actin-­
[146] and by astrocytes in culture [91, 112]. containing membranous structures with a diame-
Interestingly, Piacentini and coworkers see a ter of 50 to 800 nm forming bridges that connect
toxic effect of this uptake of oligomeric tau remote cells. They have been shown to carry
inside the astrocytes on the synaptic activity of prion proteins as well as viruses [54, 137].
neurons. Martini-Stoica and colleagues see that Tardivel and co-authors showed that tau could
an increased uptake in astrocyte can markedly spread through these tunneling nanotubes and
reduce tau propagation [91, 112]. Several stud- that interestingly extracellular tau can induce the
ies showed that uptake of tau happens through formation of these structures.
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 313

Fig. 23.5  Model of uptake and subsequent pathology into the cytoplasm (red). Once in the cytoplasm misfolded
spreading in the recipient neuron. Tau is released by tau induces seeding and spread of the pathology (red).
cells probably from synaptic terminals mostly freely, but However, autophagic mechanisms are activated to clear tau
also via extracellular vesicles and taken up by secondary out causing tau to be internalized inside autophagosomes
cells through a clathrin-independent mechanism of endo- (purple). Once there, the autophagosome fusion with the
cytosis or micropinocytosis (orange). Once in the endo- lysosome leads to degradation of its content (yellow) or
some, the endosomal membrane is ruptured, and tau gets secretion via multivesicular bodies (MVB) (light blue)

 xistence of “Strains” of Tau

E ferent species were described with distinct behav-
Aggregates iors in term of tau propagation phenotypes.
Differences in different tau preparations ability to
The question that actual strains of tau aggregates undergo cell to cell transfer and seeded misfold-
exist, parallel to the clinical strains observed in ing, differences in structure of tau aggregates in
prion diseases, has been highly investigated and different diseases, and differences in the patterns
the answers are not yet definitive. What “strain” of tau in varied neuropathological diseases like
means for prion diseases is, based on the same AD compared to Pick’s disease or PSP all point
protein template, there exist different but stable to the existence of something analogous to strains
conformations of misfolding that are transmissi- for tau. Firstly, comparing mutant and WT tau
ble to native non-misfolded proteins. In prion dis- gives some insight. In fact, mutated tau proteins
ease, these different conformers convey different seem to bear different fiber morphologies and
biological properties, including their clinical biochemical characteristics that influence the
course and aggressiveness of disease. ability of seeding [6, 34, 47]. It is interesting to
In support of this idea, there is substantial evi- observe that the specific fibers’ conformation of
dence for tau that different conformers exist and mutant tau can be transmitted and conserved by
have various behaviors in pathology. Indeed, dif- WT tau [34, 47]. In vivo, the propagation and
314 S. Dujardin and B. T. Hyman

conformations of mutant vs WT tau is also differ- seeding potential in AD and other tauopathies is
ent [30, 32]. Secondly, the different isoforms of therefore really intriguing but must be currently
tau – particularly the presence or absence of the viewed as uncertain.
exon 10 in the respectively called 4R or 3R iso- In conclusion, tau species share with prion
forms  – impacts tau behavior. It is of note that strains different behavior in tau pathology propa-
fibers made of various Tau isoforms bear variable gation (fibrils structures; lesions structures; cell
properties [2, 36, 49, 86, 107], and 3R and 4R tau types affected; brain regional patterns etc.).
proteins differ in behavior in in vivo propagation Structural evidence shows at least three stable
[30]. These properties could explain the different conformations (straight filaments, paired helical
morphologies of lesions observed in human filaments, and wide filaments in Pick bodies) but
pathologies and transmissible to animal or cell much more structural work in underway [36]. It
models [19, 128] and the differences in spatio- remains unknown if these stable conformational
temporal evolution of Tau pathology between so-­ changes can be reliably transmitted in a human
called 3R (Pick’s disease) and 4R (PSP, disease setting. Techniques used in the research
Corticobasal degeneration) tauopathies. and diagnosis for prion such as protein misfold-
Comparing tau fibers from different sporadic ing cyclic amplification are also not available for
tauopathies and their capacity to seed led to the use for tau proteins mostly because tau needs the
identification of potential strains [19, 77, 128]. addition of a polyanion such as heparin to aggre-
The Diamond laboratory used a stable cell gate in vitro [92]. This is one of the challenges for
reporter assay [48, 62] to model seeding using future studies of tau propagation.
brain lysate from several human cases with vari-
ous tauopathies. They show that the shapes of
resulting intracellular aggregates differ from case  au Prion-Like Propagation,
T
to case and that this could be transmitted to mice. Challenges and Future Directions
Different regional pattern of progression in the
brain are also observed consistent with the exis- Extensive studies have been conducted in the past
tence in sporadic tauopathies of different tau decade to understand tau propagation and the
strains [77, 128]. underlying cellular and molecular pathways,
Structural analysis of tau are difficult to carry researchers are still facing several challenges.
out due to the biophysical properties of tau.
Nevertheless, recent studies of Cryogenic
Electron Microscopy have clearly identified the  sing More Sporadic Tauopathies
U
structure of aggregated tau, either in Alzheimer’s Models
disease [39] or in Pick’s disease [36]. As the con-
tent in isoforms differs, the structures of the Tau pathological changes such as hyperphosphor-
fibrils are different as well explaining the shapes ylation or aggregation are fairly frequent in the
of fibers as previously observed with electron human brain of even non-demented people.
microscopy. Mirbaha and colleagues also However, for diverse and mostly unknown rea-
recently suggested that even a single monomer sons, in cell and animal models, tau pathology is
can bear a stable and transmissible conformation fairly hard to model. Therefore, most of the time,
that seed aggregation [96]. This result also researchers have to use artifices such as mutating
argues for the presence of specific strains but the tau gene (MAPT gene) and/or artificially
contradicts previous results showing that mono- cleaving tau protein in order to obtain markers of
meric tau in solution has only transient folding tau pathology. Mutations of the MAPT gene are
[103], raising the question of what stabilizes the rare events that only occur in rare genetic familial
monomeric tau into a stable conformation under frontotemporal dementia cases and don’t occur in
some circumstances but not others. Whether all the other sporadic tauopathies such as AD,
monomeric tau can have stable structures with a PSP, Pick’s Disease etc. [42, 50]. Tau propagation
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 315

studies are no exception to this limitation. Some  au Propagation, Physiological

T
examples: the main cell reporter assay for the Versus Pathological Mechanisms
study of tau seeding is a truncated and mutated
version of tau protein [62]; the main animal model As stated just above, one drawback of some mod-
used for the study of tau cell-to-cell transfer is a els of tau propagation is that the pathological
mutant version of tau [25, 88]; the main animal phenotypes such as tau phosphorylation, mis-
model used to study tau transmission and seeding folding and/or aggregation can be limited. This
is the PS19 mouse model that bears a P301S leads us to hypothesize that at least part of the
mutation [65]. Using these tau versions is indeed described mechanisms of tau propagation are
convenient but one could ask their relevance for happening physiologically. Indeed, for example,
sporadic tauopathies especially knowing the pre- regarding the trans-synaptic transfer of tau, in a
viously described structural differences between recent in vivo study we showed that this transfer
mutant and WT tau [6, 34, 47]. Some studies tried could happen in the absence of epitopes of tau
to use more “sporadic” models of tau propaga- hyperphosphorylation and misfolding [30]. This
tion. For example, Clavaguera and collaborators result suggests that the trans-synaptic transfer of
were able to induce transmission/seeding of tau tau can be a physiological mechanism. This idea
pathology into a mouse model overexpressing a is also supported by the fact that tau propagation
WT form of human tau protein [20] or even into still occurs in absence of any pathological con-
WT mice [20, 104]. We used WT rats and overex- version in a mouse lacking endogenous tau [154].
pressed WT 3R or 4R tau into their hippocampus However, several questions remain: is this physi-
using lentiviral vectors resulting in the long-dis- ological mechanism conserved in a diseased
tance transfer of tau proteins and also propagation brain? Are the routes of cell-to-cell transfer the
of pathological epitopes [30, 32]. These results same in physiology versus pathology? Is the
are promising but they have limitations as well kinetic of this transfer impacted? Some elements
because (1) of the very long kinetic needed to ana- of response are present in the literature. Firstly,
lyze them (analysis 8–24 months after pathology the routes seem to differ: tau secretion via exo-
transduction), (2) you hardly get the full tau somes, for instance, seem to happen only when
pathology phenotypes with only weak aggrega- tau accumulate intracellularly and not physiolog-
tion in neurofibrillary tangles, neurodegeneration, ically [31]. Secondly, the kinetic of nanotubes
or synaptic and behavioral deficits). These limita- formation is impacted by the extracellular pres-
tions of the models probably rely on fundamental ence of tau aggregates [144]. Finally, numerous
kinetic differences in the rodents brain (month of studies all show that secreted tau proteins are
pathological development at max) compared to mostly non-phosphorylated, monomeric, soluble
the human brains (years of pathological evolu- and truncated [18, 31, 41, 75, 76, 99, 113, 115,
tion). Metabolism, brain clearance and neuroin- 116, 127, 129, 134, 143, 159]. This goes in line
flammation also seem to occur faster in rodents with a physiological secretion of tau and could
than in humans. In addition to that, fundamental explain the presence of tau in the cerebrospinal
molecular differences cannot be ignored as mice fluid of every individuals independently of neu-
only express the 4R isoforms of tau when humans ronal death [57]. However, on the other side, tau
express both 3R and 4R. The use of more human- uptake seems to be more efficient for aggregated
ized models such as tau Knock-In rodents, human species than monomeric tau [97, 142].
iPS cells and/or organoids might give us more Understanding this discrepancy between secre-
insights in the future. In conclusion, it would be tion and uptake may be a key point for under-
interesting to invent efficient tools to study tau standing the exact mechanisms of tau propagation.
pathological propagation in a “sporadic” environ- Whether this is due to different kinetic of uptake,
ment to tackle the specific mechanisms of tau route of uptake or intracellular stability after
release uptake and instructive misfolding. uptake is difficult to understand at that time.
316 S. Dujardin and B. T. Hyman

 recisely Identify the Kinetics of Tau

P Identify Good Therapeutic Strategies
Life Cycle Events
Identifying tau propagation as a key mechanism
Understanding tau propagation also comes by of tauopathies and particularly AD has raised
deciphering the kinetic of events in the brain. some hope to find a disease-modifying treatment
This notion is largely unknown and very depen- tackling the progression of the disease. An obvi-
dent on the models used but the current literature ous target is the extracellular species of tau. This
can inform us about the slow versus fast points of led to intense research by academic and industry
this process. Firstly, it seems that the neuron-to-­ teams around the development of immunothera-
neuron transfer of tau is a slow point both in vitro pies. Most of the studies find a high potential for
and in vivo where we can observe it only after reducing tau propagation in diverse models.
respectively days or weeks/months [15, 25, 32, Examples include the work of Holtzman’s team
88]. It is interesting and counterintuitive to notice showing a reduction of tau uptake and seeding in
that both secretion and uptake are fast points as vitro in addition to reduction of tau pathology in
they can be observed in hours mostly in vitro [15, a transgenic mouse model of genetic
31, 46, 78, 106, 116, 142, 157–160]. The discrep- Frontotemporal Lobar Dementia [160].
ancy between transfer and secretion/uptake prob- Antibodies against oligomeric tau (TOMA anti-
ably rely on (1) the sensitivity of the histology bodies) showed reduction of seeding after injec-
techniques used to see tau transfer and (2) the tion of oligomers in the brain of hTau mouse
mostly non-physiologic amount of tau used to model [17]. We more recently showed diverse
observe fast secretion and fast uptake. Lastly, tau efficacies of different antibodies raised against
seeding seems to be a process regulated by a sig- total tau and also phosphorylated tau or truncated
moidal growth with a long lag phase, an expo- tau onto uptake and seeding of high molecular
nential growth phase and a plateau phase [62]. In weight tau [105, 106]. Recent studies confirm
vitro, when the right conditions are reunited (high these data in other models and trying to assess the
intracellular content of tau, presence of tau trun- potential of different epitopes [23, 24]. Beside
cation/mutation favoring misfolding etc.), the the use of immunotherapy, tau propagation
whole seeding process can happen in less than research could lead to additional efforts to
24 h [28] when in vivo the lag phase takes prob- develop therapies by targeting the pathway by
ably weeks and also relies on the amount and which tau propagates. This includes the media-
quality of tau in the recipient brain. Indeed, tors of tau propagation like targeting the mecha-
inducing seeding and aggregates formation in a nisms of secretion (chaperones, exosome
tau transgenic animal with a mutation (as soon as formations), the mechanisms of tau uptake (the
2 weeks post induction in animals such as PS19 endosomal pathway), stimulating the autophagy
mice (see Ahmed et al. [4] and Iba et al. [65, 66]) pathways to favor clearance, stimulation of
is faster than for a tau transgenic animal without deacetylase [95] etc. To our knowledge, this
a mutation (months post-injection in animals research is still preliminary but the understanding
such as alz17 or hTau mice (see Castillo-Carranza of mechanisms of tau propagation might lead to
et  al. [17] and Clavaguera et  al. [20] or in WT more potential therapies in the future.
animals [20])). The fast/exponential appearance
of neurofibrillary tangle after the lag phase in
vivo is supported by observations using 2-photon  nderstand the Regional Patterns,
U
microscopy [25]. Thus both transfer and seeding the Cell Vulnerabilities
have their fast and slow points. It is also evident
that factors such as the presence of amyloid beta It is striking to notice the differences between the
[9, 59, 69, 117, 148] or brain activity [116, 158, pattern of progression of the different tauopathies
159] can also accelerate these slow points. in the human brain. While they are all patterns of
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 317

tau aggregation, they are quite heterogenous. occurs via an endocytic mechanism. Once in the
Why are some pathologies very focal (Pick’s dis- endosome, tau aggregates cause the endosomal
ease or Argyrophilic Grain Disease) and some membrane to rupture and reach the cytoplasm. It
more global (PSP and AD)? Why tau progression has been demonstrated that when they reach the
leads to asymmetry in some diseases and global cytoplasm, tau aggregates induce recruit endog-
symmetry in others? Interestingly, the human enous naïve tau proteins in a seeding mechanism
data show for example that in AD the neurons (Fig. 23.5). The similarities between these mech-
receiving the neurofibrillary tangles are strongly anisms and the prion protein spreading mecha-
interconnected but dentate gyrus granule cells, nism has opened a discussion between researchers
which receive the bulk of the perforant pathway over the idea that we can reunite tau, prion and
projection from the early affected entorhinal cor- other proteins under the banner of prion-like pro-
tex, are relatively resistant [64]. On the other teins. This semantic debate is still open but we
hand, Pick bodies, another type of lesion made of clearly, here, brought arguments for the existence
tau aggregates are predominantly present in these of strong convergence points but as well pointed
particular granule cells in the dentate gyrus that out differences. Many other questions remain to
are resistant to the alternatively folded form of be answered. Why would different brain areas
tau found in neurofibrillary tangles [118]. These and neuronal subtypes be selectively vulnerable
differences are thought to be due to different cell for different forms of inclusions? What causes
populations vulnerabilities to tau pathology the initial misfolding events that are then propa-
[123]. This also suggests unique relationships gated across neural systems? Why distinct and
between specific misfolded tau species and host specific neural systems would respond differ-
cell characteristics. It would be really important ently to tau “strains”? What causes toxicity of
to understand why different brain areas and neu- neural networks? What is the impact of known
ronal subtypes are selectively vulnerable for dif- molecular differences (isoforms and posttransla-
ferent forms of inclusions? Beyond the tional modifications) onto these mechanisms?
observation that propagation is therefore depen- And, perhaps most importantly, can these insights
dent on both the anatomy and the properties of help in the design and application of tau as a ther-
the recipient cell, these concepts are largely apeutic target to slow progressive neurodegener-
unexplored so far. ation in AD and other tauopathies?

Conclusions References

Tau propagation, this sequence of events leading 1. Abounit S, Wu JW, Duff K, Victoria GS, Zurzolo
C. Tunneling nanotubes: a possible highway in the
to the progression of the disease in the brain of spreading of tau and other prion-like proteins in neu-
tauopathies (Fig. 23.2), has now been experimen- rodegenerative diseases. Prion. 2016;10(5):344–51.
tally described. We can distinguish several path- https://doi.org/10.1080/19336896.2016.1223003.
ways leading to tau propagation: (1) the transfer 2. Adams SJ, DeTure MA, McBride M, Dickson DW,
Petrucelli L.  Three repeat isoforms of tau inhibit
from cell-to-cell of both pathological and non-­ assembly of four repeat tau filaments. PLoS One.
pathological species of tau (Fig. 23.4) and (2) the 2010;5(5):e10810. https://doi.org/10.1371/journal.
transmission of tau pathology in a second neuron pone.0010810.
via a seeding mechanism (Fig.  23.1). The most 3. Aguzzi A, Rajendran L. The transcellular spread of
cytosolic amyloids, prions, and prionoids. Neuron.
plausible route of tau transfer is through the 2009;64(6):783–90. https://doi.org/10.1016/j.
extracellular space and particularly via its secre- neuron.2009.12.016.
tion (mostly freely in the medium but also 4. Ahmed Z, Cooper J, Murray TK, Garn K,
through extracellular vesicles) and uptake into McNaughton E, Clarke H, Parhizkar S, Ward MA,
Cavallini A, Jackson S, Bose S, Clavaguera F, Tolnay
recipient neurons (Fig. 23.4). This uptake likely M, Lavenir I, Goedert M, Hutton ML, O’Neill MJ. A
318 S. Dujardin and B. T. Hyman

novel in  vivo model of tau propagation with rapid P, Moechars D.  Synaptic contacts enhance cell-­
and progressive neurofibrillary tangle pathology: the cell Tau pathology propagation. Cell Rep.
to-­
pattern of spread is determined by connectivity, not 2015;11(8):1176–83. https://doi.org/10.1016/j.
proximity. Acta Neuropathol. 2014;127(5):667–83. celrep.2015.04.043.
https://doi.org/10.1007/s00401-014-1254-6. 16. Calafate S, Flavin W, Verstreken P, Moechars
5. Alvarez-Erviti L, Seow Y, Yin H, Betts C, Lakhal D.  Loss of Bin1 promotes the propagation of Tau
S, Wood MJ.  Delivery of siRNA to the mouse pathology. Cell Rep. 2016;17(4):931–40. https://doi.
brain by systemic injection of targeted exosomes. org/10.1016/j.celrep.2016.09.063.
Nat Biotechnol. 2011;29(4):341–5. https://doi. 17. Castillo-Carranza DL, Gerson JE, Sengupta U,
org/10.1038/nbt.1807. Guerrero-Munoz MJ, Lasagna-Reeves CA, Kayed
6. Aoyagi H, Hasegawa M, Tamaoka A. Fibrillogenic R. Specific targeting of tau oligomers in Htau mice
nuclei composed of P301L mutant tau induce prevents cognitive impairment and tau toxicity fol-
elongation of P301L tau but not wild-type tau. lowing injection with brain-derived tau oligomeric
J  Biol Chem. 2007;282(28):20309–18. https://doi. seeds. J  Alzheimers Dis. 2014;40(Suppl 1):S97–
org/10.1074/jbc.M611876200. S111. https://doi.org/10.3233/JAD-132477.
7. Arnold SE, Hyman BT, Flory J, Damasio AR, 18. Chai X, Dage JL, Citron M.  Constitutive secretion
Van Hoesen GW.  The topographical and neuro- of tau protein by an unconventional mechanism.
anatomical distribution of neurofibrillary tangles Neurobiol Dis. 2012;48(3):356–66. https://doi.
and neuritic plaques in the cerebral cortex of org/10.1016/j.nbd.2012.05.021.
patients with Alzheimer’s disease. Cereb Cortex. 19. Clavaguera F, Akatsu H, Fraser G, Crowther RA,
1991;1(1):103–16. Frank S, Hench J, Probst A, Winkler DT, Reichwald
8. Asai H, Ikezu S, Tsunoda S, Medalla M, Luebke J, J, Staufenbiel M, Ghetti B, Goedert M, Tolnay
Haydar T, Wolozin B, Butovsky O, Kugler S, Ikezu M.  Brain homogenates from human tauopathies
T.  Depletion of microglia and inhibition of exo- induce tau inclusions in mouse brain. Proc Natl
some synthesis halt tau propagation. Nat Neurosci. Acad Sci U S A. 2013;110(23):9535–40. https://doi.
2015;18(11):1584–93. https://doi.org/10.1038/ org/10.1073/pnas.1301175110.
nn.4132. 20. Clavaguera F, Bolmont T, Crowther RA,
9. Bennett RE, DeVos SL, Dujardin S, Corjuc B, Gor Abramowski D, Frank S, Probst A, Fraser G,
R, Gonzalez J, Roe AD, Frosch MP, Pitstick R, Stalder AK, Beibel M, Staufenbiel M, Jucker M,
Carlson GA, Hyman BT.  Enhanced Tau aggrega- Goedert M, Tolnay M. Transmission and spreading
tion in the presence of amyloid beta. Am J  Pathol. of tauopathy in transgenic mouse brain. Nat Cell
2017;187(7):1601–12. https://doi.org/10.1016/j. Biol. 2009;11(7):909–13. https://doi.org/10.1038/
ajpath.2017.03.011. ncb1901.
10. Blennow K.  A review of fluid biomarkers for 21. Clavaguera F, Hench J, Lavenir I, Schweighauser G,
Alzheimer’s disease: moving from CSF to blood. Frank S, Goedert M, Tolnay M.  Peripheral admin-
Neurol Ther. 2017;6(Suppl 1):15–24. https://doi. istration of tau aggregates triggers intracerebral
org/10.1007/s40120-017-0073-9. tauopathy in transgenic mice. Acta Neuropathol.
11. Bousset L, Pieri L, Ruiz-Arlandis G, Gath J, Jensen 2014;127(2):299–301. https://doi.org/10.1007/
PH, Habenstein B, Madiona K, Olieric V, Bockmann s00401-013-1231-5.
A, Meier BH, Melki R.  Structural and functional 22. Clavaguera F, Lavenir I, Falcon B, Frank S, Goedert
characterization of two alpha-synuclein strains. Nat M, Tolnay M. “Prion-like” templated misfolding in
Commun. 2013;4:2575. https://doi.org/10.1038/ tauopathies. Brain Pathol. 2013;23(3):342–9. https://
ncomms3575. doi.org/10.1111/bpa.12044.
12. Braak H, Braak E.  Neuropathological stageing 23. Courade JP, Angers R, Mairet-Coello G, Pacico N,
of Alzheimer-related changes. Acta Neuropathol. Tyson K, Lightwood D, Munro R, McMillan D,
1991;82(4):239–59. Griffin R, Baker T, Starkie D, Nan R, Westwood
13. Braak H, Del Tredici K.  The pathological process M, Mushikiwabo ML, Jung S, Odede G, Sweeney
underlying Alzheimer’s disease in individuals under B, Popplewell A, Burgess G, Downey P, Citron
thirty. Acta Neuropathol. 2011;121(2):171–81. M. Epitope determines efficacy of therapeutic anti-­
https://doi.org/10.1007/s00401-010-0789-4. Tau antibodies in a functional assay with human
14. Caillierez R, Begard S, Lecolle K, Deramecourt Alzheimer Tau. Acta Neuropathol. 2018;136(5):729–
V, Zommer N, Dujardin S, Loyens A, Dufour N, 45. https://doi.org/10.1007/s00401-018-1911-2.
Auregan G, Winderickx J, Hantraye P, Deglon N, 24. Dai CL, Hu W, Tung YC, Liu F, Gong CX, Iqbal
Buee L, Colin M. Lentiviral delivery of the human K.  Tau passive immunization blocks seeding and
wild-type tau protein mediates a slow and progres- spread of Alzheimer hyperphosphorylated Tau-­
sive neurodegenerative tau pathology in the rat induced pathology in 3 x Tg-AD mice. Alzheimers
brain. Mol Ther. 2013;21(7):1358–68. https://doi. Res Ther. 2018;10(1):13. https://doi.org/10.1186/
org/10.1038/mt.2013.66. s13195-018-0341-7.
15. Calafate S, Buist A, Miskiewicz K, Vijayan V, 25. de Calignon A, Polydoro M, Suarez-Calvet M,
Daneels G, de Strooper B, de Wit J, Verstreken William C, Adamowicz DH, Kopeikina KJ, Pitstick
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 319

R, Sahara N, Ashe KH, Carlson GA, Spires-Jones 2015;290(2):1049–65. https://doi.org/10.1074/jbc.

TL, Hyman BT.  Propagation of tau pathology in M114.589309.
a model of early Alzheimer’s disease. Neuron. 35. Falcon B, Noad J, McMahon H, Randow F, Goedert
2012;73(4):685–97. https://doi.org/10.1016/j. M.  Galectin-8-mediated selective autophagy pro-
neuron.2011.11.033. tects against seeded tau aggregation. J  Biol Chem.
26. Delacourte A, David JP, Sergeant N, Buee L, 2018;293(7):2438–51. https://doi.org/10.1074/jbc.
Wattez A, Vermersch P, Ghozali F, Fallet-Bianco M117.809293.
C, Pasquier F, Lebert F, Petit H, Di Menza C. The 36. Falcon B, Zhang W, Murzin AG, Murshudov G,
biochemical pathway of neurofibrillary degenera- Garringer HJ, Vidal R, Crowther RA, Ghetti B,
tion in aging and Alzheimer’s disease. Neurology. Scheres SHW, Goedert M.  Structures of filaments
1999;52(6):1158–65. from Pick’s disease reveal a novel tau protein
27. Desplats P, Lee HJ, Bae EJ, Patrick C, Rockenstein fold. Nature. 2018;561(7721):137–40. https://doi.
E, Crews L, Spencer B, Masliah E, Lee SJ. Inclusion org/10.1038/s41586-018-0454-y.
formation and neuronal cell death through neuron-­ 37. Faure J, Lachenal G, Court M, Hirrlinger J,
to-­neuron transmission of alpha-synuclein. Proc Natl Chatellard-Causse C, Blot B, Grange J, Schoehn G,
Acad Sci U S A. 2009;106(31):13010–5. https://doi. Goldberg Y, Boyer V, Kirchhoff F, Raposo G, Garin
org/10.1073/pnas.0903691106. J, Sadoul R. Exosomes are released by cultured cor-
28. DeVos SL, Corjuc BT, Commins C, Dujardin S, tical neurones. Mol Cell Neurosci. 2006;31(4):642–
Bannon RN, Corjuc D, Moore BD, Bennett RE, Jorfi 8. https://doi.org/10.1016/j.mcn.2005.12.003.
M, Gonzales JA, Dooley PM, Roe AD, Pitstick R, 38. Fevrier B, Vilette D, Archer F, Loew D, Faigle W,
Irimia D, Frosch MP, Carlson GA, Hyman BT. Tau Vidal M, Laude H, Raposo G.  Cells release prions
reduction in the presence of amyloid-beta prevents in association with exosomes. Proc Natl Acad Sci U
tau pathology and neuronal death in  vivo. Brain. S A. 2004;101(26):9683–8. https://doi.org/10.1073/
2018;141(7):2194–212. https://doi.org/10.1093/ pnas.0308413101.
brain/awy117. 39. Fitzpatrick AWP, Falcon B, He S, Murzin AG,
29. Dujardin S. De la cellule au primate, propagation Murshudov G, Garringer HJ, Crowther RA, Ghetti
physiopathologique de la protéine Tau (From cells B, Goedert M, Scheres SHW.  Cryo-EM structures
to primates, pathophysiological propagation of tau of tau filaments from Alzheimer’s disease. Nature.
protein). Catalogue SCD Lille. 2015;2:1–334. 2017;547(7662):185–90. https://doi.org/10.1038/
30. Dujardin S, Begard S, Caillierez R, Lachaud C, nature23002.
Carrier S, Lieger S, Gonzalez JA, Deramecourt 40. Flavin WP, Bousset L, Green ZC, Chu Y,
V, Deglon N, Maurage CA, Frosch MP, Hyman Skarpathiotis S, Chaney MJ, Kordower JH, Melki R,
BT, Colin M, Buee L. Different tau species lead Campbell EM.  Endocytic vesicle rupture is a con-
to heterogeneous tau pathology propagation served mechanism of cellular invasion by amyloid
and misfolding. Acta Neuropathol Commun. proteins. Acta Neuropathol. 2017;134(4):629–53.
2018;6(1):132. https://doi.org/10.1186/ https://doi.org/10.1007/s00401-017-1722-x.
s40478-018-0637-7. 41. Fontaine SN, Zheng D, Sabbagh JJ, Martin MD,
31. Dujardin S, Begard S, Caillierez R, Lachaud C, Chaput D, Darling A, Trotter JH, Stothert AR,
Delattre L, Carrier S, Loyens A, Galas MC, Bousset Nordhues BA, Lussier A, Baker J, Shelton L, Kahn
L, Melki R, Auregan G, Hantraye P, Brouillet E, M, Blair LJ, Stevens SM Jr, Dickey CA.  DnaJ/
Buee L, Colin M. Ectosomes: a new mechanism for Hsc70 chaperone complexes control the extracel-
non-exosomal secretion of tau protein. PLoS One. lular release of neurodegenerative-associated pro-
2014;9(6):e100760. https://doi.org/10.1371/journal. teins. EMBO J.  2016;35(14):1537–49. https://doi.
pone.0100760. org/10.15252/embj.201593489.
32. Dujardin S, Lecolle K, Caillierez R, Begard S, 42. Forrest SL, Kril JJ, Halliday GM. Reply: will FTLD-­
Zommer N, Lachaud C, Carrier S, Dufour N, tau work for all when FTDP-17 retires? Brain. 2018;
Auregan G, Winderickx J, Hantraye P, Deglon N, https://doi.org/10.1093/brain/awy179.
Colin M, Buee L.  Neuron-to-neuron wild-type Tau 43. Freundt EC, Maynard N, Clancy EK, Roy S, Bousset
protein transfer through a trans-synaptic mecha- L, Sourigues Y, Covert M, Melki R, Kirkegaard K,
nism: relevance to sporadic tauopathies. Acta Brahic M. Neuron-to-neuron transmission of alpha-­
Neuropathol Commun. 2014;2:14. https://doi. synuclein fibrils through axonal transport. Ann
org/10.1186/2051-5960-2-14. Neurol. 2012;72(4):517–24. https://doi.org/10.1002/
33. Duyckaerts C, Hauw JJ.  Diagnosis and staging of ana.23747.
Alzheimer disease. Neurobiol Aging. 1997;18(4 44. Friedhoff P, von Bergen M, Mandelkow EM, Davies
Suppl):S33–42. P, Mandelkow E. A nucleated assembly mechanism
34. Falcon B, Cavallini A, Angers R, Glover S, Murray of Alzheimer paired helical filaments. Proc Natl
TK, Barnham L, Jackson S, O’Neill MJ, Isaacs Acad Sci U S A. 1998;95(26):15712–7.
AM, Hutton ML, Szekeres PG, Goedert M, Bose 45. Frost B, Diamond MI.  Prion-like mechanisms in
S. Conformation determines the seeding potencies of neurodegenerative diseases. Nat Rev Neurosci.
native and recombinant Tau aggregates. J Biol Chem. 2010;11(3):155–9. https://doi.org/10.1038/nrn2786.
320 S. Dujardin and B. T. Hyman

46. Frost B, Jacks RL, Diamond MI. Propagation of tau disease? Commun Integr Biol. 2012;5(6):623–6.
misfolding from the outside to the inside of a cell. https://doi.org/10.4161/cib.21437.
J  Biol Chem. 2009;284(19):12845–52. https://doi. 58. Hamaguchi T, Eisele YS, Varvel NH, Lamb BT,
org/10.1074/jbc.M808759200. Walker LC, Jucker M. The presence of Abeta seeds,
47. Frost B, Ollesch J, Wille H, Diamond and not age per se, is critical to the initiation of
MI.  Conformational diversity of wild-type Tau Abeta deposition in the brain. Acta Neuropathol.
fibrils specified by templated conformation change. 2012;123(1):31–7. https://doi.org/10.1007/
J  Biol Chem. 2009;284(6):3546–51. https://doi. s00401-011-0912-1.
org/10.1074/jbc.M805627200. 59. He Z, Guo JL, McBride JD, Narasimhan S, Kim
48. Furman JL, Holmes BB, Diamond MI.  Sensitive H, Changolkar L, Zhang B, Gathagan RJ, Yue C,
detection of proteopathic seeding activity with Dengler C, Stieber A, Nitla M, Coulter DA, Abel T,
FRET flow cytometry. J Vis Exp. 2015;106:e53205. Brunden KR, Trojanowski JQ, Lee VM.  Amyloid-­
https://doi.org/10.3791/53205. beta plaques enhance Alzheimer’s brain tau-seeded
49. Furukawa Y, Kaneko K, Nukina N.  Tau protein pathologies by facilitating neuritic plaque tau aggre-
assembles into isoform- and disulfide-dependent gation. Nat Med. 2018;24(1):29–38. https://doi.
polymorphic fibrils with distinct structural proper- org/10.1038/nm.4443.
ties. J  Biol Chem. 2011;286(31):27236–46. https:// 60. Hoenig MC, Bischof GN, Seemiller J, Hammes
doi.org/10.1074/jbc.M111.248963. J, Kukolja J, Onur OA, Jessen F, Fliessbach K,
50. Ghetti B, Oblak AL, Boeve BF, Johnson KA, Neumaier B, Fink GR, van Eimeren T, Drzezga
Dickerson BC, Goedert M.  Invited review: fron- A.  Networks of tau distribution in Alzheimer’s
totemporal dementia caused by microtubule-­ disease. Brain. 2018;141(2):568–81. https://doi.
associated protein tau gene (MAPT) mutations: a org/10.1093/brain/awx353.
chameleon for neuropathology and neuroimaging. 61. Holmes BB, DeVos SL, Kfoury N, Li M, Jacks R,
Neuropathol Appl Neurobiol. 2015;41(1):24–46. Yanamandra K, Ouidja MO, Brodsky FM, Marasa J,
https://doi.org/10.1111/nan.12213. Bagchi DP, Kotzbauer PT, Miller TM, Papy-Garcia
51. Giannakopoulos P, Herrmann FR, Bussiere T, D, Diamond MI.  Heparan sulfate proteoglycans
Bouras C, Kovari E, Perl DP, Morrison JH, Gold G, mediate internalization and propagation of spe-
Hof PR. Tangle and neuron numbers, but not amy- cific proteopathic seeds. Proc Natl Acad Sci U S A.
loid load, predict cognitive status in Alzheimer’s dis- 2013;110(33):E3138–47. https://doi.org/10.1073/
ease. Neurology. 2003;60(9):1495–500. pnas.1301440110.
52. Gibbons GS, Banks RA, Kim B, Xu H, Changolkar 62. Holmes BB, Furman JL, Mahan TE, Yamasaki
L, Leight SN, Riddle DM, Li C, Gathagan RJ, Brown TR, Mirbaha H, Eades WC, Belaygorod L, Cairns
HJ, Zhang B, Trojanowski JQ, Lee VM. GFP-mutant NJ, Holtzman DM, Diamond MI.  Proteopathic tau
human Tau transgenic mice develop tauopathy fol- seeding predicts tauopathy in vivo. Proc Natl Acad
lowing CNS injections of Alzheimer’s brain-derived Sci U S A. 2014;111(41):E4376–85. https://doi.
pathological Tau or synthetic mutant human Tau org/10.1073/pnas.1411649111.
fibrils. J  Neurosci. 2017;37(47):11485–94. https:// 63. Hu X, Crick SL, Bu G, Frieden C, Pappu RV,
doi.org/10.1523/JNEUROSCI.2393-17.2017. Lee JM.  Amyloid seeds formed by cellular
53. Goedert M, Clavaguera F, Tolnay M. The propaga- uptake, concentration, and aggregation of the
tion of prion-like protein inclusions in neurodegener- amyloid-beta peptide. Proc Natl Acad Sci U S A.
ative diseases. Trends Neurosci. 2010;33(7):317–25. 2009;106(48):20324–9. https://doi.org/10.1073/
https://doi.org/10.1016/j.tins.2010.04.003. pnas.0911281106.
54. Gousset K, Schiff E, Langevin C, Marijanovic 64. Hyman BT, Van Hoesen GW, Damasio AR, Barnes
Z, Caputo A, Browman DT, Chenouard N, de CL.  Alzheimer’s disease: cell-specific pathol-
Chaumont F, Martino A, Enninga J, Olivo-Marin ogy isolates the hippocampal formation. Science.
JC, Mannel D, Zurzolo C.  Prions hijack tunnel- 1984;225(4667):1168–70.
ling nanotubes for intercellular spread. Nat Cell 65. Iba M, Guo JL, McBride JD, Zhang B, Trojanowski
Biol. 2009;11(3):328–36. https://doi.org/10.1038/ JQ, Lee VM. Synthetic tau fibrils mediate transmis-
ncb1841. sion of neurofibrillary tangles in a transgenic mouse
55. Guo JL, Lee VM.  Seeding of normal Tau by model of Alzheimer’s-like tauopathy. J  Neurosci.
pathological Tau conformers drives pathogen- 2013;33(3):1024–37. https://doi.org/10.1523/
esis of Alzheimer-like tangles. J  Biol Chem. JNEUROSCI.2642-12.2013.
2011;286(17):15317–31. https://doi.org/10.1074/ 66. Iba M, McBride JD, Guo JL, Zhang B, Trojanowski
jbc.M110.209296. JQ, Lee VM.  Tau pathology spread in PS19 tau
56. Hall GF, Patuto BA.  Is tau ready for admission to transgenic mice following locus coeruleus (LC)
the prion club? Prion. 2012;6(3):223–33. https://doi. injections of synthetic tau fibrils is determined by
org/10.4161/pri.19912. the LC’s afferent and efferent connections. Acta
57. Hall GF, Saman S. Death or secretion? The demise of Neuropathol. 2015;130(3):349–62. https://doi.
a plausible assumption about CSF-tau in Alzheimer org/10.1007/s00401-015-1458-4.
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 321

67. Irwin DJ, Brettschneider J, McMillan CT, Cooper Tau. Cell Rep. 2018;23(7):2039–55. https://doi.
F, Olm C, Arnold SE, Van Deerlin VM, Seeley org/10.1016/j.celrep.2018.04.056.
WW, Miller BL, Lee EB, Lee VM, Grossman M, 77. Kaufman SK, Sanders DW, Thomas TL, Ruchinskas
Trojanowski JQ.  Deep clinical and neuropatho- AJ, Vaquer-Alicea J, Sharma AM, Miller TM,
logical phenotyping of Pick disease. Ann Neurol. Diamond MI.  Tau prion strains dictate patterns of
2016;79(2):272–87. https://doi.org/10.1002/ cell pathology, progression rate, and regional vulner-
ana.24559. ability in vivo. Neuron. 2016;92(4):796–812. https://
68. Jackson SJ, Kerridge C, Cooper J, Cavallini A, doi.org/10.1016/j.neuron.2016.09.055.
Falcon B, Cella CV, Landi A, Szekeres PG, Murray 78. Kfoury N, Holmes BB, Jiang H, Holtzman DM,
TK, Ahmed Z, Goedert M, Hutton M, O’Neill Diamond MI.  Trans-cellular propagation of Tau
MJ, Bose S. Short fibrils constitute the major spe- aggregation by fibrillar species. J  Biol Chem.
cies of seed-­competent Tau in the brains of mice 2012;287(23):19440–51. https://doi.org/10.1074/
transgenic for human P301S Tau. J  Neurosci. jbc.M112.346072.
2016;36(3):762–72. https://doi.org/10.1523/ 79. Kim D, Lim S, Haque MM, Ryoo N, Hong HS,
JNEUROSCI.3542-15.2016. Rhim H, Lee DE, Chang YT, Lee JS, Cheong E, Kim
69. Jacobs HIL, Hedden T, Schultz AP, Sepulcre J, Perea DJ, Kim YK. Identification of disulfide cross-linked
RD, Amariglio RE, Papp KV, Rentz DM, Sperling tau dimer responsible for tau propagation. Sci Rep.
RA, Johnson KA. Structural tract alterations predict 2015;5:15231. https://doi.org/10.1038/srep15231.
downstream tau accumulation in amyloid-positive 80. Kim W, Lee S, Jung C, Ahmed A, Lee G, Hall
older individuals. Nat Neurosci. 2018;21(3):424–31. GF.  Interneuronal transfer of human tau between
https://doi.org/10.1038/s41593-018-0070-z. lamprey central neurons in situ. J  Alzheimers
70. Jarrett JT, Lansbury PT Jr. Seeding “one-­dimensional Dis. 2010;19(2):647–64. https://doi.org/10.3233/
crystallization” of amyloid: a pathogenic mecha- JAD-2010-1273.
nism in Alzheimer’s disease and scrapie? Cell. 81. Kowal J, Tkach M, Thery C. Biogenesis and secre-
1993;73(6):1055–8. tion of exosomes. Curr Opin Cell Biol. 2014;29:116–
71. Johnson KA, Schultz A, Betensky RA, Becker 25. https://doi.org/10.1016/j.ceb.2014.05.004.
JA, Sepulcre J, Rentz D, Mormino E, Chhatwal 82. Lasagna-Reeves CA, Castillo-Carranza DL,
J, Amariglio R, Papp K, Marshall G, Albers M, Sengupta U, Guerrero-Munoz MJ, Kiritoshi T,
Mauro S, Pepin L, Alverio J, Judge K, Philiossaint Neugebauer V, Jackson GR, Kayed R.  Alzheimer
M, Shoup T, Yokell D, Dickerson B, Gomez-Isla brain-derived tau oligomers propagate pathology
T, Hyman B, Vasdev N, Sperling R.  Tau positron from endogenous tau. Sci Rep. 2012;2:700. https://
emission tomographic imaging in aging and early doi.org/10.1038/srep00700.
Alzheimer disease. Ann Neurol. 2016;79(1):110–9. 83. Lasagna-Reeves CA, Castillo-Carranza DL,
https://doi.org/10.1002/ana.24546. Sengupta U, Sarmiento J, Troncoso J, Jackson GR,
72. Jucker M, Walker LC.  Self-propagation of patho- Kayed R. Identification of oligomers at early stages
genic protein aggregates in neurodegenerative dis- of tau aggregation in Alzheimer’s disease. FASEB
eases. Nature. 2013;501(7465):45–51. https://doi. J.  2012;26(5):1946–59. https://doi.org/10.1096/
org/10.1038/nature12481. fj.11-199851.
73. Kane MD, Lipinski WJ, Callahan MJ, Bian F, 84. Lee HJ, Patel S, Lee SJ. Intravesicular localization
Durham RA, Schwarz RD, Roher AE, Walker and exocytosis of alpha-synuclein and its aggre-
LC. Evidence for seeding of beta -amyloid by intra- gates. J Neurosci. 2005;25(25):6016–24. https://doi.
cerebral infusion of Alzheimer brain extracts in org/10.1523/JNEUROSCI.0692-05.2005.
beta -amyloid precursor protein-transgenic mice. 85. Lee S, Kim W, Li Z, Hall GF.  Accumulation
J Neurosci. 2000;20(10):3606–11. of vesicle-associated human tau in distal den-
74. Kanmert D, Cantlon A, Muratore CR, Jin M, drites drives degeneration and tau secre-
O’Malley TT, Lee G, Young-Pearse TL, Selkoe DJ, tion in an in situ cellular tauopathy model. Int
Walsh DM. C-terminally truncated forms of Tau, but J  Alzheimers Dis. 2012;2012:172837. https://doi.
not full-length Tau or its C-terminal fragments, are org/10.1155/2012/172837.
released from neurons independently of cell death. 86. Levarska L, Zilka N, Jadhav S, Neradil P, Novak
J  Neurosci. 2015;35(30):10851–65. https://doi. M. Of rodents and men: the mysterious interneuronal
org/10.1523/JNEUROSCI.0387-15.2015. pilgrimage of misfolded protein tau in Alzheimer’s
75. Karch CM, Jeng AT, Goate AM.  Extracellular disease. J  Alzheimers Dis. 2013;37(3):569–77.
Tau levels are influenced by variability in Tau https://doi.org/10.3233/JAD-131106.
that is associated with tauopathies. J  Biol Chem. 87. Liberski PP. Prion, prionoids and infectious amyloid.
2012;287(51):42751–62. https://doi.org/10.1074/ Parkinsonism Relat Disord. 2014;20(Suppl 1):S80–
jbc.M112.380642. 4. https://doi.org/10.1016/S1353-8020(13)70021-X.
76. Katsinelos T, Zeitler M, Dimou E, Karakatsani 88. Liu L, Drouet V, Wu JW, Witter MP, Small SA,
A, Muller HM, Nachman E, Steringer JP, Ruiz de Clelland C, Duff K.  Trans-synaptic spread of tau
Almodovar C, Nickel W, Jahn TR. Unconventional pathology in  vivo. PLoS One. 2012;7(2):e31302.
secretion mediates the trans-cellular spreading of https://doi.org/10.1371/journal.pone.0031302.
322 S. Dujardin and B. T. Hyman

89. Luk KC, Kehm V, Carroll J, Zhang B, O’Brien P, ics. PLoS One. 2017;12(5):e0178288. https://doi.
Trojanowski JQ, Lee VM.  Pathological alpha-­ org/10.1371/journal.pone.0178288.
synuclein transmission initiates Parkinson-like 99. Mohamed NV, Plouffe V, Remillard-Labrosse G,
neurodegeneration in nontransgenic mice. Science. Planel E, Leclerc N.  Starvation and inhibition of
2012;338(6109):949–53. https://doi.org/10.1126/ lysosomal function increased tau secretion by pri-
science.1227157. mary cortical neurons. Sci Rep. 2014;4:5715. https://
90. Luk KC, Kehm VM, Zhang B, O’Brien P, Trojanowski doi.org/10.1038/srep05715.
JQ, Lee VM.  Intracerebral inoculation of patho- 100. Morales R, Duran-Aniotz C, Castilla J, Estrada LD,
logical alpha-synuclein initiates a rapidly progres- Soto C. De novo induction of amyloid-beta deposi-
sive neurodegenerative alpha-­ synucleinopathy in tion in vivo. Mol Psychiatry. 2012;17(12):1347–53.
mice. J  Exp Med. 2012;209(5):975–86. https://doi. https://doi.org/10.1038/mp.2011.120.
org/10.1084/jem.20112457. 101. Mougenot AL, Nicot S, Bencsik A, Morignat
91. Martini-Stoica H, Cole AL, Swartzlander DB, E, Verchere J, Lakhdar L, Legastelois S, Baron
Chen F, Wan YW, Bajaj L, Bader DA, Lee VMY, T.  Prion-like acceleration of a synucleinopathy
Trojanowski JQ, Liu Z, Sardiello M, Zheng in a transgenic mouse model. Neurobiol Aging.
H. TFEB enhances astroglial uptake of extracellular 2012;33(9):2225–8. https://doi.org/10.1016/j.
tau species and reduces tau spreading. J  Exp Med. neurobiolaging.2011.06.022.
2018;215(9):2355–77. https://doi.org/10.1084/ 102. Mudher A, Colin M, Dujardin S, Medina M,
jem.20172158. Dewachter I, Alavi Naini SM, Mandelkow
92. Meyer V, Dinkel PD, Rickman Hager E, Margittai EM, Mandelkow E, Buee L, Goedert M, Brion
M.  Amplification of Tau fibrils from minute quan- JP. What is the evidence that tau pathology spreads
tities of seeds. Biochemistry. 2014;53(36):5804–9. through prion-like propagation? Acta Neuropathol
https://doi.org/10.1021/bi501050g. Commun. 2017;5(1):99. https://doi.org/10.1186/
93. Meyer-Luehmann M, Coomaraswamy J, Bolmont s40478-017-0488-7.
T, Kaeser S, Schaefer C, Kilger E, Neuenschwander 103. Mukrasch MD, Bibow S, Korukottu J, Jeganathan S,
A, Abramowski D, Frey P, Jaton AL, Vigouret JM, Biernat J, Griesinger C, Mandelkow E, Zweckstetter
Paganetti P, Walsh DM, Mathews PM, Ghiso J, M.  Structural polymorphism of 441-residue tau at
Staufenbiel M, Walker LC, Jucker M.  Exogenous single residue resolution. PLoS Biol. 2009;7(2):e34.
induction of cerebral beta-­ amyloidogenesis https://doi.org/10.1371/journal.pbio.1000034.
is governed by agent and host. Science. 104. Narasimhan S, Guo JL, Changolkar L, Stieber A,
2006;313(5794):1781–4. https://doi.org/10.1126/ McBride JD, Silva LV, He Z, Zhang B, Gathagan
science.1131864. RJ, Trojanowski JQ, Lee VMY.  Pathological Tau
94. Michel CH, Kumar S, Pinotsi D, Tunnacliffe A, St strains from human brains recapitulate the diver-
George-Hyslop P, Mandelkow E, Mandelkow EM, sity of tauopathies in nontransgenic mouse brain.
Kaminski CF, Kaminski Schierle GS.  Extracellular J  Neurosci. 2017;37(47):11406–23. https://doi.
monomeric tau protein is sufficient to initiate the org/10.1523/JNEUROSCI.1230-17.2017.
spread of tau protein pathology. J  Biol Chem. 105. Nicholls SB, DeVos SL, Commins C, Nobuhara C,
2014;289(2):956–67. https://doi.org/10.1074/jbc. Bennett RE, Corjuc DL, Maury E, Eftekharzadeh B,
M113.515445. Akingbade O, Fan Z, Roe AD, Takeda S, Wegmann
95. Min SW, Sohn PD, Li Y, Devidze N, Johnson JR, S, Hyman BT.  Characterization of TauC3 antibody
Krogan NJ, Masliah E, Mok SA, Gestwicki JE, and demonstration of its potential to block tau prop-
Gan L. SIRT1 Deacetylates Tau and reduces patho- agation. PLoS One. 2017;12(5):e0177914. https://
genic Tau spread in a mouse model of tauopa- doi.org/10.1371/journal.pone.0177914.
thy. J  Neurosci. 2018;38(15):3680–8. https://doi. 106. Nobuhara CK, DeVos SL, Commins C, Wegmann
org/10.1523/JNEUROSCI.2369-17.2018. S, Moore BD, Roe AD, Costantino I, Frosch MP,
96. Mirbaha H, Chen D, Morazova OA, Ruff KM, Pitstick R, Carlson GA, Hock C, Nitsch RM,
Sharma AM, Liu X, Goodarzi M, Pappu RV, Colby Montrasio F, Grimm J, Cheung AE, Dunah AW,
DW, Mirzaei H, Joachimiak LA, Diamond MI. Inert Wittmann M, Bussiere T, Weinreb PH, Hyman BT,
and seed-competent tau monomers suggest struc- Takeda S. Tau antibody targeting pathological species
tural origins of aggregation. elife. 2018;7 https://doi. blocks neuronal uptake and interneuron propagation
org/10.7554/eLife.36584. of Tau in vitro. Am J Pathol. 2017;187(6):1399–412.
97. Mirbaha H, Holmes BB, Sanders DW, Bieschke https://doi.org/10.1016/j.ajpath.2017.01.022.
J, Diamond MI.  Tau trimers are the minimal 107. Nonaka T, Watanabe ST, Iwatsubo T, Hasegawa
propagation unit spontaneously internalized to M.  Seeded aggregation and toxicity of {alpha}-
seed intracellular aggregation. J  Biol Chem. synuclein and tau: cellular models of neurodegenera-
2015;290(24):14893–903. https://doi.org/10.1074/ tive diseases. J Biol Chem. 2010;285(45):34885–98.
jbc.M115.652693. https://doi.org/10.1074/jbc.M110.148460.
98. Mohamed NV, Desjardins A, Leclerc N. Tau secre- 108. Olsson B, Lautner R, Andreasson U, Ohrfelt A,
tion is correlated to an increase of Golgi dynam- Portelius E, Bjerke M, Holtta M, Rosen C, Olsson C,
Strobel G, Wu E, Dakin K, Petzold M, Blennow K,
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 323

Zetterberg H. CSF and blood biomarkers for the diag- 118. Probst A, Tolnay M, Langui D, Goedert M,
nosis of Alzheimer’s disease: a systematic review and Spillantini MG. Pick’s disease: hyperphosphorylated
meta-analysis. Lancet Neurol. 2016;15(7):673–84. tau protein segregates to the somatoaxonal compart-
https://doi.org/10.1016/S1474-4422(16)00070-3. ment. Acta Neuropathol. 1996;92(6):588–96.
109. Peelaerts W, Bousset L, Van der Perren A, Moskalyuk 119. Prusiner SB.  Novel proteinaceous infectious parti-
A, Pulizzi R, Giugliano M, Van den Haute C, Melki cles cause scrapie. Science. 1982;216(4542):136–44.
R, Baekelandt V.  Alpha-synuclein strains cause 120. Rajendran L, Bali J, Barr MM, Court FA, Kramer-­
distinct synucleinopathies after local and systemic Albers EM, Picou F, Raposo G, van der Vos KE,
administration. Nature. 2015;522(7556):340–4. van Niel G, Wang J, Breakefield XO.  Emerging
https://doi.org/10.1038/nature14547. roles of extracellular vesicles in the nervous sys-
110. Peeraer E, Bottelbergs A, Van Kolen K, Stancu IC, tem. J  Neurosci. 2014;34(46):15482–9. https://doi.
Vasconcelos B, Mahieu M, Duytschaever H, Ver org/10.1523/JNEUROSCI.3258-14.2014.
Donck L, Torremans A, Sluydts E, Van Acker N, 121. Rajendran L, Honsho M, Zahn TR, Keller P,
Kemp JA, Mercken M, Brunden KR, Trojanowski Geiger KD, Verkade P, Simons K.  Alzheimer’s
JQ, Dewachter I, Lee VM, Moechars D. Intracerebral disease beta-amyloid peptides are released in asso-
injection of preformed synthetic tau fibrils initiates ciation with exosomes. Proc Natl Acad Sci U S A.
widespread tauopathy and neuronal loss in the brains 2006;103(30):11172–7. https://doi.org/10.1073/
of tau transgenic mice. Neurobiol Dis. 2015;73:83– pnas.0603838103.
95. https://doi.org/10.1016/j.nbd.2014.08.032. 122. Rauch JN, Chen JJ, Sorum AW, Miller GM, Sharf
111. Petkova AT, Leapman RD, Guo Z, Yau WM, T, See SK, Hsieh-Wilson LC, Kampmann M, Kosik
Mattson MP, Tycko R. Self-propagating, molecular-­ KS.  Tau internalization is regulated by 6-O sulfa-
level polymorphism in Alzheimer’s beta-amyloid tion on heparan sulfate proteoglycans (HSPGs).
fibrils. Science. 2005;307(5707):262–5. https://doi. Sci Rep. 2018;8(1):6382. https://doi.org/10.1038/
org/10.1126/science.1105850. s41598-018-24904-z.
112. Piacentini R, Li Puma DD, Mainardi M, Lazzarino 123. Rebeck GW, Hyman BT.  Neuroanatomical
G, Tavazzi B, Arancio O, Grassi C.  Reduced glio- connections and specific regional vulner-
transmitter release from astrocytes mediates ability in Alzheimer’s disease. Neurobiol Aging.
tau-induced synaptic dysfunction in cultured hippo- 1993;14(1):45–7; discussion 55-46
campal neurons. Glia. 2017;65(8):1302–16. https:// 124. Reilly P, Winston CN, Baron KR, Trejo M,
doi.org/10.1002/glia.23163. Rockenstein EM, Akers JC, Kfoury N, Diamond M,
113. Plouffe V, Mohamed NV, Rivest-McGraw J, Bertrand Masliah E, Rissman RA, Yuan SH.  Novel human
J, Lauzon M, Leclerc N. Hyperphosphorylation and neuronal tau model exhibiting neurofibrillary
cleavage at D421 enhance tau secretion. PLoS One. tangles and transcellular propagation. Neurobiol
2012;7(5):e36873. https://doi.org/10.1371/journal. Dis. 2017;106:222–34. https://doi.org/10.1016/j.
pone.0036873. nbd.2017.06.005.
114. Polanco JC, Li C, Bodea LG, Martinez-Marmol R, 125. Rodriguez L, Mohamed NV, Desjardins A, Lippe R,
Meunier FA, Gotz J.  Amyloid-beta and tau com- Fon EA, Leclerc N. Rab7A regulates tau secretion.
plexity  – towards improved biomarkers and tar- J  Neurochem. 2017;141(4):592–605. https://doi.
geted therapies. Nat Rev Neurol. 2018;14(1):22–39. org/10.1111/jnc.13994.
https://doi.org/10.1038/nrneurol.2017.162. 126. Saito Y, Ruberu NN, Sawabe M, Arai T, Tanaka N,
115. Polydoro M, Dzhala VI, Pooler AM, Nicholls SB, Kakuta Y, Yamanouchi H, Murayama S. Staging of
McKinney AP, Sanchez L, Pitstick R, Carlson GA, argyrophilic grains: an age-associated tauopathy.
Staley KJ, Spires-Jones TL, Hyman BT.  Soluble J Neuropathol Exp Neurol. 2004;63(9):911–8.
pathological tau in the entorhinal cortex leads to 127. Saman S, Kim W, Raya M, Visnick Y, Miro S, Saman
presynaptic deficits in an early Alzheimer’s disease S, Jackson B, McKee AC, Alvarez VE, Lee NC, Hall
model. Acta Neuropathol. 2014;127(2):257–70. GF. Exosome-associated tau is secreted in tauopathy
https://doi.org/10.1007/s00401-013-1215-5. models and is selectively phosphorylated in cerebro-
116. Pooler AM, Phillips EC, Lau DH, Noble W, spinal fluid in early Alzheimer disease. J Biol Chem.
Hanger DP.  Physiological release of endogenous 2012;287(6):3842–9. https://doi.org/10.1074/jbc.
tau is stimulated by neuronal activity. EMBO M111.277061.
Rep. 2013;14(4):389–94. https://doi.org/10.1038/ 128. Sanders DW, Kaufman SK, DeVos SL, Sharma AM,
embor.2013.15. Mirbaha H, Li A, Barker SJ, Foley AC, Thorpe JR,
117. Pooler AM, Polydoro M, Maury EA, Nicholls SB, Serpell LC, Miller TM, Grinberg LT, Seeley WW,
Reddy SM, Wegmann S, William C, Saqran L, Diamond MI.  Distinct tau prion strains propa-
Cagsal-Getkin O, Pitstick R, Beier DR, Carlson gate in cells and mice and define different tauopa-
GA, Spires-Jones TL, Hyman BT.  Amyloid accel- thies. Neuron. 2014;82(6):1271–88. https://doi.
erates tau propagation and toxicity in a model org/10.1016/j.neuron.2014.04.047.
of early Alzheimer’s disease. Acta Neuropathol 129. Santa-Maria I, Varghese M, Ksiezak-Reding H,
Commun. 2015;3:14. https://doi.org/10.1186/ Dzhun A, Wang J, Pasinetti GM. Paired helical fila-
s40478-015-0199-x. ments from Alzheimer disease brain induce intracel-
324 S. Dujardin and B. T. Hyman

lular accumulation of Tau protein in aggresomes. genic mice. Acta Neuropathol. 2015;129(6):875–94.
J  Biol Chem. 2012;287(24):20522–33. https://doi. https://doi.org/10.1007/s00401-015-1413-4.
org/10.1074/jbc.M111.323279. 139. Stohr J, Watts JC, Mensinger ZL, Oehler A, Grillo
130. Santacruz K, Lewis J, Spires T, Paulson J, Kotilinek SK, DeArmond SJ, Prusiner SB, Giles K.  Purified
L, Ingelsson M, Guimaraes A, DeTure M, Ramsden and synthetic Alzheimer’s amyloid beta (Abeta) pri-
M, McGowan E, Forster C, Yue M, Orne J, Janus ons. Proc Natl Acad Sci U S A. 2012;109(27):11025–
C, Mariash A, Kuskowski M, Hyman B, Hutton M, 30. https://doi.org/10.1073/pnas.1206555109.
Ashe KH.  Tau suppression in a neurodegenerative 140. Stopschinski BE, Holmes BB, Miller GM, Manon
mouse model improves memory function. Science. VA, Vaquer-Alicea J, Prueitt WL, Hsieh-Wilson
2005;309(5733):476–81. https://doi.org/10.1126/ LC, Diamond MI.  Specific glycosaminoglycan
science.1113694. chain length and sulfation patterns are required
131. Schwarz AJ, Yu P, Miller BB, Shcherbinin S, for cell uptake of tau versus alpha-synuclein
Dickson J, Navitsky M, Joshi AD, Devous MD Sr, and beta-­ amyloid aggregates. J  Biol Chem.
Mintun MS.  Regional profiles of the candidate tau 2018;293(27):10826–40. https://doi.org/10.1074/
PET ligand 18F-AV-1451 recapitulate key features of jbc.RA117.000378.
Braak histopathological stages. Brain. 2016;139(Pt 141. Takahashi M, Miyata H, Kametani F, Nonaka T,
5):1539–50. https://doi.org/10.1093/brain/aww023. Akiyama H, Hisanaga S, Hasegawa M. Extracellular
132. Sepulcre J, Grothe MJ, d’Oleire Uquillas F, Ortiz-­ association of APP and tau fibrils induces intracel-
Teran L, Diez I, Yang HS, Jacobs HIL, Hanseeuw lular aggregate formation of tau. Acta Neuropathol.
BJ, Li Q, El-Fakhri G, Sperling RA, Johnson 2015;129(6):895–907. https://doi.org/10.1007/
KA. Neurogenetic contributions to amyloid beta and s00401-015-1415-2.
tau spreading in the human cortex. Nat Med. 2018; 142. Takeda S, Wegmann S, Cho H, DeVos SL, Commins
https://doi.org/10.1038/s41591-018-0206-4. C, Roe AD, Nicholls SB, Carlson GA, Pitstick R,
133. Silva JM, Rodrigues S, Sampaio-Marques B, Gomes Nobuhara CK, Costantino I, Frosch MP, Muller DJ,
P, Neves-Carvalho A, Dioli C, Soares-Cunha C, Irimia D, Hyman BT.  Neuronal uptake and propa-
Mazuik BF, Takashima A, Ludovico P, Wolozin B, gation of a rare phosphorylated high-molecular-­
Sousa N, Sotiropoulos I. Dysregulation of autophagy weight tau derived from Alzheimer’s disease brain.
and stress granule-related proteins in stress-driven Nat Commun. 2015;6:8490. https://doi.org/10.1038/
Tau pathology. Cell Death Differ. 2018; https://doi. ncomms9490.
org/10.1038/s41418-018-0217-1. 143. Tang Z, Ioja E, Bereczki E, Hultenby K, Li C, Guan
134. Simon D, Garcia-Garcia E, Royo F, Falcon-Perez Z, Winblad B, Pei JJ. mTor mediates tau localization
JM, Avila J. Proteostasis of tau. Tau overexpression and secretion: implication for Alzheimer’s disease.
results in its secretion via membrane vesicles. FEBS Biochim Biophys Acta. 2015;1853(7):1646–57.
Lett. 2012;586(1):47–54. https://doi.org/10.1016/j. https://doi.org/10.1016/j.bbamcr.2015.03.003.
febslet.2011.11.022. 144. Tardivel M, Begard S, Bousset L, Dujardin S, Coens
135. Smolek T, Jadhav S, Brezovakova V, Cubinkova A, Melki R, Buee L, Colin M. Tunneling nanotube
V, Valachova B, Novak P, Zilka N.  First-in-rat (TNT)-mediated neuron-to neuron transfer of patho-
study of human Alzheimer’s disease Tau propaga- logical Tau protein assemblies. Acta Neuropathol
tion. Mol Neurobiol. 2018; https://doi.org/10.1007/ Commun. 2016;4(1):117. https://doi.org/10.1186/
s12035-018-1102-0. s40478-016-0386-4.
136. Sokolow S, Henkins KM, Bilousova T, Gonzalez 145. Telling GC, Parchi P, DeArmond SJ, Cortelli P,
B, Vinters HV, Miller CA, Cornwell L, Poon WW, Montagna P, Gabizon R, Mastrianni J, Lugaresi E,
Gylys KH. Pre-synaptic C-terminal truncated tau is Gambetti P, Prusiner SB.  Evidence for the confor-
released from cortical synapses in Alzheimer’s dis- mation of the pathologic isoform of the prion pro-
ease. J Neurochem. 2015;133(3):368–79. https://doi. tein enciphering and propagating prion diversity.
org/10.1111/jnc.12991. Science. 1996;274(5295):2079–82.
137. Sowinski S, Jolly C, Berninghausen O, Purbhoo 146. Usenovic M, Niroomand S, Drolet RE, Yao L,
MA, Chauveau A, Kohler K, Oddos S, Eissmann Gaspar RC, Hatcher NG, Schachter J, Renger JJ,
P, Brodsky FM, Hopkins C, Onfelt B, Sattentau Parmentier-Batteur S.  Internalized Tau oligomers
Q, Davis DM.  Membrane nanotubes physically cause neurodegeneration by inducing accumula-
connect T cells over long distances presenting a tion of pathogenic Tau in human neurons derived
novel route for HIV-1 transmission. Nat Cell Biol. from induced pluripotent stem cells. J  Neurosci.
2008;10(2):211–9. https://doi.org/10.1038/ncb1682. 2015;35(42):14234–50. https://doi.org/10.1523/
138. Stancu IC, Vasconcelos B, Ris L, Wang P, Villers A, JNEUROSCI.1523-15.2015.
Peeraer E, Buist A, Terwel D, Baatsen P, Oyelami T, 147. Varghese M, Santa-Maria I, Ho L, Ward L,
Pierrot N, Casteels C, Bormans G, Kienlen-Campard Yemul S, Dubner L, Ksiezak-Reding H, Pasinetti
P, Octave JN, Moechars D, Dewachter I. Templated GM.  Extracellular Tau paired helical filaments
misfolding of Tau by prion-like seeding along neuro- differentially affect Tau pathogenic mechanisms
nal connections impairs neuronal network function in mitotic and post-mitotic cells: implications
and associated behavioral outcomes in Tau trans- for mechanisms of Tau propagation in the brain.
23  Tau Prion-Like Propagation: State of the Art and Current Challenges 325

J  Alzheimers Dis. 2016;54(2):477–96. https://doi. drome. Brain. 2007;130(Pt 6):1566–76. https://doi.

org/10.3233/JAD-160166. org/10.1093/brain/awm104.
148. Vasconcelos B, Stancu IC, Buist A, Bird M, Wang P, 156. Woerman AL, Aoyagi A, Patel S, Kazmi SA, Lobach
Vanoosthuyse A, Van Kolen K, Verheyen A, Kienlen-­ I, Grinberg LT, McKee AC, Seeley WW, Olson SH,
Campard P, Octave JN, Baatsen P, Moechars D, Prusiner SB.  Tau prions from Alzheimer’s disease
Dewachter I. Heterotypic seeding of Tau fibrillization and chronic traumatic encephalopathy patients prop-
by pre-aggregated Abeta provides potent seeds for agate in cultured cells. Proc Natl Acad Sci U S A.
prion-like seeding and propagation of Tau-pathology 2016;113(50):E8187–96. https://doi.org/10.1073/
in  vivo. Acta Neuropathol. 2016;131(4):549–69. pnas.1616344113.
https://doi.org/10.1007/s00401-015-1525-x. 157. Wu JW, Herman M, Liu L, Simoes S, Acker CM,
149. Verny M, Jellinger KA, Hauw JJ, Bancher C, Litvan Figueroa H, Steinberg JI, Margittai M, Kayed R,
I, Agid Y. Progressive supranuclear palsy: a clinico- Zurzolo C, Di Paolo G, Duff KE. Small misfolded
pathological study of 21 cases. Acta Neuropathol. Tau species are internalized via bulk endocytosis and
1996;91(4):427–31. anterogradely and retrogradely transported in neu-
150. Volpicelli-Daley LA, Luk KC, Patel TP, Tanik SA, rons. J  Biol Chem. 2013;288(3):1856–70. https://
Riddle DM, Stieber A, Meaney DF, Trojanowski JQ, doi.org/10.1074/jbc.M112.394528.
Lee VM.  Exogenous alpha-synuclein fibrils induce 158. Wu JW, Hussaini SA, Bastille IM, Rodriguez GA,
Lewy body pathology leading to synaptic dysfunc- Mrejeru A, Rilett K, Sanders DW, Cook C, Fu H,
tion and neuron death. Neuron. 2011;72(1):57–71. Boonen RA, Herman M, Nahmani E, Emrani S,
https://doi.org/10.1016/j.neuron.2011.08.033. Figueroa YH, Diamond MI, Clelland CL, Wray S,
151. Walker LC, Levine H 3rd, Mattson MP, Jucker Duff KE.  Neuronal activity enhances tau propa-
M.  Inducible proteopathies. Trends Neurosci. gation and tau pathology in  vivo. Nat Neurosci.
2006;29(8):438–43. https://doi.org/10.1016/j. 2016;19(8):1085–92. https://doi.org/10.1038/
tins.2006.06.010. nn.4328.
152. Wang Y, Balaji V, Kaniyappan S, Kruger L, Irsen S, 159. Yamada K, Holth JK, Liao F, Stewart FR, Mahan TE,
Tepper K, Chandupatla R, Maetzler W, Schneider Jiang H, Cirrito JR, Patel TK, Hochgrafe K, Mandelkow
A, Mandelkow E, Mandelkow EM.  The release EM, Holtzman DM. Neuronal activity regulates extra-
and trans-synaptic transmission of Tau via exo- cellular tau in vivo. J Exp Med. 2014;211(3):387–93.
somes. Mol Neurodegener. 2017;12(1):5. https://doi. https://doi.org/10.1084/jem.20131685.
org/10.1186/s13024-016-0143-y. 160. Yanamandra K, Kfoury N, Jiang H, Mahan TE, Ma S,
153. Wegmann S, Bennett RE, Amaral AS, Hyman Maloney SE, Wozniak DF, Diamond MI, Holtzman
BT. Studying tau protein propagation and pathology DM.  Anti-tau antibodies that block tau aggregate
in the mouse brain using adeno-associated viruses. seeding in  vitro markedly decrease pathology and
Methods Cell Biol. 2017;141:307–22. https://doi. improve cognition in vivo. Neuron. 2013;80(2):402–
org/10.1016/bs.mcb.2017.06.014. 14. https://doi.org/10.1016/j.neuron.2013.07.046.
154. Wegmann S, Maury EA, Kirk MJ, Saqran L, 161. Yonetani M, Nonaka T, Masuda M, Inukai Y,
Roe A, DeVos SL, Nicholls S, Fan Z, Takeda Oikawa T, Hisanaga S, Hasegawa M. Conversion of
S, Cagsal-Getkin O, William CM, Spires-Jones wild-type alpha-synuclein into mutant-type fibrils
TL, Pitstick R, Carlson GA, Pooler AM, Hyman and its propagation in the presence of A30P mutant.
BT. Removing endogenous tau does not prevent tau J  Biol Chem. 2009;284(12):7940–50. https://doi.
propagation yet reduces its neurotoxicity. EMBO org/10.1074/jbc.M807482200.
J.  2015;34(24):3028–41. https://doi.org/10.15252/ 162. Yuyama K, Yamamoto N, Yanagisawa K. Accelerated
embj.201592748. release of exosome-associated GM1 ganglioside
155. Williams DR, Holton JL, Strand C, Pittman A, de (GM1) by endocytic pathway abnormality: another
Silva R, Lees AJ, Revesz T. Pathological tau burden putative pathway for GM1-induced amyloid fibril
and distribution distinguishes progressive supra- formation. J  Neurochem. 2008;105(1):217–24.
nuclear palsy-parkinsonism from Richardson’s syn- https://doi.org/10.1111/j.1471-4159.2007.05128.x.
 Tau Condensates
24
Kenneth S. Kosik and Songi Han

Introduction of membraneless organelles and biomolecular

condensates by cell biologists and (2) discoveries
The tau protein embodies a deep paradox. Under by chemists of coacervation, which usually refers
normal conditions it is one of the most soluble to an electrostatically driven liquid-liquid phase
proteins known, but once it aggregates in a vari- separation resulting from weak association of
ety of different diseases, it becomes one of the oppositely charged, but typically fully hydrated,
most insoluble proteins known. How this transi- macro-ions in the form of polyelectrolytes.
tion occurs has remained baffling. Recently we Membraneless organelles are a set of cellular
and other groups have identified a novel state of entities recognizable microscopically by immu-
tau in which the protein becomes phase separated nostaining of specific marker proteins.
within the cytoplasm [1–5]. The core unresolved Biomolecular condensates imply a more general
question addressed in this chapter is whether biophysical structure that draws upon principles
LLPS (liquid liquid phase separation) of tau in of polymer chemistry and soft matter physics to
brain cells is on-pathway to pathological tau describe membraneless entities with an emphasis
aggregates. Two streams of prior work have on the condensation of proteins and nucleic acids
brought the field to this question: (1) discoveries as a “crucible” for localized molecular reactions
[6]. We will discuss these membraneless organ-
elles and coacervation separately and then con-
K. S. Kosik (*)
Department of Molecular, Cellular and join them toward the problem of tau aggregation
Developmental Biology, University of California, in the tauopathies. The question is whether
Santa Barbara, CA, USA physico-chemical knowledge of coacervation
Neuroscience Research Institute, University of processes can inform the cell biological problem:
California, Santa Barbara, CA, USA when do physico-chemical principles apply and
Biomolecular Science and Engineering, University of when do the specific biological interactions take
California, Santa Barbara, CA, USA over and dictate the relevant mechanisms.
e-mail: kosik@lifesci.ucsb.edu Discoveries in this domain related to other
S. Han proteins involved in neurodegenerative condi-
Biomolecular Science and Engineering, University of tions have greatly informed the work on tau.
California, Santa Barbara, CA, USA
Among the most well-studied and informative of
Department of Chemistry and Biochemistry, these proteins are FUS [7–10], hnRNPA2B1 and
University of California, Santa Barbara, CA, USA
hnRNPA1 [11], TDP-43 [11, 12], C9ORF72 [13–
Department of Chemical Engineering, University of 15], Ddx4 [16] all of which are genetically
California, Santa Barbara, CA, USA

© Springer Nature Singapore Pte Ltd. 2019 327

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_24
328 K. S. Kosik and S. Han

a­ssociated with amyotrophic lateral sclerosis (a)–(c) should underlie all biomolecular conden-
(ALS) and frontotemporal lobar degeneration. As sates that can be deemed droplets, (d) weak inter-
noted above, the tau protein now joins this list of actions should be foundational for biomolecular
amyloidogenic proteins undergoing LLPS [1–5]. condensates originating from and constituting
The results of studies on FUS (fused in sarcoma/ droplets, while (e) droplet fusion is a stringent
translocated in liposarcoma), a FET family mem- criteria only met by rapid dynamics, and possibly
ber, exemplify a number of principles that have apply only to nascent droplets upon formation,
guided recent thinking about aggregation prone before solidification processes have occurred.
proteins in neurodegeneration. The protein has a
prion-like low complexity domain near its car-
boxy terminus, which is RGG-rich (Arg-Gly-­ Membraneless Organelles
Gly-rich) and where most of the disease mutations
reside. Interestingly, an arginine motif also shows To contextualize tau in this context, some consid-
up in C9ORF72. This region can polymerize into eration of the recognized organelles that derive
fibrous amyloid-like assemblies [11, 17, 18], from liquid liquid phase separation is useful.
which exhibit the macroscopic behavior of Cells contain a class of organelles in which cel-
hydrogels. Extrapolating these observations to lular material is condensed and moves cohesively
living cells and assessing whether liquid conden- through the cell, but are devoid of any limiting
sates are on pathway to fibers represents a con- membrane. The condensed constituents are
siderable challenge. In the case of FUS, retained within the volume of the organelle, pre-
liquid-like droplets can form in living cells, FUS sumably held together by electrostatic interac-
droplets have the appearance of stress granules, tions, while water can be freely exchanged.
the liquid-like state can convert into a solid state Liquid-like behavior is observed by FRAP of
and disease-associated mutations enhance this their protein components as well as flowing, fus-
conversion [10]. These observations and others ing, and rapid reorganization of internal compo-
have led to the speculation that phase transitions nents [20, 21]. The liquid-liquid phase separation
are a central pathobiological phenomenon in (LLPS) often arises from binding sites on RNA-­
some neurodegenerative diseases. binding proteins that interact with themselves or
One can ask how well the observations con- other RNA-binding proteins to create a larger
cerning FUS generalize to other proteins that multivalent assembly [10, 16, 22–26]. Organelles,
undergo aggregation. This remains an open ques- which meet these biophysical criteria are diverse
tion that is asked by many in the field. It starts and include RNA granules, germline P granules
with the very basic challenge of knowing what [20], P-bodies, stress granules that mostly tend to
criteria are necessary to conclude that a conden- reside in the cytoplasm and nucleoli, Cajal bod-
sate in cells has droplet properties. Among the ies, nuclear speckles, paraspeckles, histone-locus
properties assigned to droplets are (a) dynamic bodies, and promyelocytic leukemia bodies,
protein exchange as measured by fluorescence which are all present in the nucleus [27–30].
recovery after photobleaching (FRAP); (b) con- Synapsin can form a distinct liquid phase in an
densation of fully hydrated proteins in high con- aqueous environment and capture small lipid
centration within a droplet that can have a vesicles [31]. The presence of RNA is one unify-
viscosity of 10- to 100-fold that of water; (c) a ing, but not absolute feature of these structures.
spherical shape that balances interfacial tension The phase separated states are often rapidly dis-
to its medium; (d) the interference with assembled upon phosphorylation. In the case of
1,6-­hexanediol of weak hydrophobic protein-­ synapsin, by calcium/calmodulin-dependent pro-
protein or protein-RNA interactions that are tein kinase II, which could serve as a mechanism
responsible for the formation of dynamic, liquid-­ for clustering synaptic vesicles at synapses.
like assemblies; and/or (e) droplet fusion into a Ribonucleoprotein (RNP) granules such as the
single droplet [19]. It is our belief that criteria nucleolus and centrosomes are also phase
24  Tau Condensates 329

s­ eparated entities. Furthermore, macromolecular nal RNA granules was their translational silence
complexes such as super enhancer-enriched tran- despite a dense collection of ribosomes and co-­
scriptional coactivators have properties of phase localization with poly(A+) mRNA.  Consistent
separation [19, 32, 33]. Proteins that contribute to with their translational incompetence is their fail-
the formation of these organelles often have low ure to incorporate radioactive amino acids and
complexity domains [34] or more generally have the absence of eIF4E, 4G, and tRNAs. Silencing
regions of intrinsic disorder, a common feature may be due to the inadequacy of translation ini-
found in RNA binding proteins. In fact, tau is an tiation factors, incomplete ribosomal subunit
intrinsically disordered protein that binds RNA, composition, inadequate tRNA populations, the
but is not classified as an RNA binding protein. presence of FMRP or steric hindrance due to
Clearly, a taxonomy based upon organizational their dense packing. When visualized with
principles of their diverse elements is sorely SYTO14  in living neurons they demonstrated
needed. motility—often small bidirectional displace-
When turning to the pathobiology of proteins ments—around synapses. These features sug-
known to form inclusions in neurodegeneration, gested that RNA granules transported mRNAs to
such as FUS, TDP-43, c9orf72, hnRNPA2B1 and dendritic locales in a translationally silent state
hnRNPA1, TIA-1 and tau several biophysical that could then be de-repressed by synaptic depo-
facets begin to converge. These proteins can larization. With depolarization, many mRNAs,
undergo LLPS in vitro, appear to form biomo- including those involved in plasticity, rapidly
lecular condensates in living cells, have low com- shift from the RNA granule fraction to polysomes
plexity domains (or more generally intrinsically where they can function as the building blocks
disordered regions) with a resemblance to yeast for the translational needs of a synapse. Among
prion proteins, often bind to RNA and can be the mRNAs detected are CaMKIIα, MAP 2, and
associated with membraneless organelles. Yeast β-actin, and the proteins Staufen, RNA helicases
prion proteins have low complexity domains, and of the DDX family and other RNA binding pro-
can be driven to interconvert from a liquid state to teins [37, 42]. A key component is RNG105/cap-
fibers [35]. This, however, does not mean that the rin1, an RNA-binding protein [43]. Knockdown
liquid state is an intermediate state towards, or is and knockout of RNG105  in cultured neurons
facilitating, fibril formation. Rather, the current reduces the synaptic connections on dendrites
state of observation merely tells us that there is a and the density of neural network [44, 45]. By
pathway for proteins in the LLPS state to trans- establishing an exposed charge surface by virtue
form to fibrils. For some proteins and environ- of the intrinsic disorder of RNA binding proteins,
ments, this path may be a highly likely pathway this class of proteins is crucial to the partitioning
under pathological conditions, and for other pro- of many phase-separated structures.
teins this might not be the case. Depolarization reorganizes granules and induces
a less compact organization of their ribosomes.
Neuronal RNA Granules  Among the earliest Thus, RNA granules serve as a local storage
sightings of a membraneless organelle in neurons compartment for mRNAs under translational
were reports beginning in 1996 of “RNA gran- arrest, but are poised for release to actively trans-
ules” [36–38] that appear to have a role in synap- lated pools. The local release of mRNAs and
tic plasticity. Historical precedents for “RNA ribosomes from granules may link RNA localiza-
granules” are rooted in developmental biology tion to translation and synaptic plasticity.
beginning with Metschnikoff [39], who observed
dark staining granules at one pole in the Miastor
metraloas (fly) larvae and subsequent research Germinal Granules  Among various mem-
identified “polar granules” involved in primordial braneless organelles, RNA granules in neurons
germ cell differentiation from a variety of insect bear the closest resemblance to germinal gran-
species [40, 41]. A remarkable feature of neuro- ules as described in Xenopus laevis, polar
330 K. S. Kosik and S. Han

g­ ranules in Drosophila melanogaster, and P gran- mRNAs rapidly move from the granules to poly-
ules in Caenorhabditis elegans. This category of ribosomes and begin translating. Heat shock
granule is collectively referred to as germ cell mRNAs that are selectively translated during
granules [46, 47]. In Caenorhabditis elegans, stress are excluded from the granules [58]. Thus,
when the first germ cell is established, P granules stress moved many “housekeeping” mRNAs to a
localize to the posterior of the one-cell embryo translationally incompetent state, while the trans-
and undergo rapid dissolution and condensation. lation of mRNAs encoding molecular chaperones
When condensation was biased to the posterior, and enzymes involved in damage repair was
localization occurred as a classic phase transi- enhanced. Anderson [60] showed that the assem-
tion, in which polarity proteins vary the conden- bly of TIA-1 positive stress granules is initiated
sation point across the cell and potentially by the phosphorylation of eIF-2α. A TIA-1
structure the cytoplasm [20]. By silencing and mutant lacking its RNA-binding domains func-
then releasing mRNA translation, they direct the tioned as a transdominant inhibitor of stress gran-
timing of maternal translation to promote germ ule assembly and therefore first suggested the
cell specification in the early embryo. role of RNA-protein complex sequestration in the
assembly of stress granules [60]. An interesting
link to neurodegeneration is the finding that stress
P Bodies  One year after the description of RNA granules are cleared by autophagy [61]. Most
granules, P bodies (processing bodies) were dis- recently Roy Parker showed that RNA homopol-
covered as small discrete cytoplasmic foci that ymers as well as RNA binding protein–protein
functioned in RNA turnover via the 5'-3'exoribo- interactions contribute to stress granule forma-
nuclease, mXRN1p [48]. At this time they were tion [62]. The importance of this finding is its
not recognized as membraneless organelles, nor implication of a role for the neurotoxic arginine-­
had they been named. Many more proteins were containing dipeptides, which are among the
discovered in these bodies that mediated decap- translation products that derive from the
ping of mRNAs, nonsense-mediated mRNA C9ORF72 expansion associated with ALS (amy-
decay, adenylate-uridylate-rich element mediated otrophic lateral sclerosis) and FTD (fronto-­
mRNA decay, RNA storage and translational temporal dementia). These dipeptides stabilize
repression by miRNAs (microRNAs) [48–51]. the RNA-RNA interactions and thereby reveal a
These diverse functions generated different possible step in the longstanding question: What
names for the structures, but P bodies have initiates pathological aggregation?
become a consensus term for the localization of
most of these RNA processing functions [52]
especially after their extensive characterization in Coacervation
yeast [52, 53].
The physico-chemical basis for the many biolog-
ical phenomena enumerated above is complex or
Stress Granules  Stress granules serve as triage simple coacervation. The term, coacervation, and
centers that sort, remodel, and export specific the concept were first introduced by Bungenberg
mRNA transcripts for reinitiation, decay, or stor- de Jong in an under cited publication [63]. A
age [54]. The response of the cell to stress by the complex coacervate refers to a phase-separated
formation of cytoplasmic foci or granules was fluid, typically upon weak association of two
reported in 1983  in tomato cells [55], in 1988 oppositely charged polyelectrolytes in solution.
using chicken embryo fibroblasts [56] and HeLa When chemists apply the term “complex” the
cells [57]. Cultured Peruvian tomato cells dem- complexity pales in comparison to the analogous
onstrated that untranslated mRNAs accumulate chemistry in a living biological system. The term
within these discrete foci with heat stress [58, “complex” in complex coacervation refers to
59]; once returned to ambient temperature, the multiple polymer types involved in the
24  Tau Condensates 331

c­omplexation; often it is just two oppositely engulf macromolecular surfaces under aqueous
charged proteins or components compared to the conditions, provided a weakly attractive interac-
thousands of proteins associated with membrane- tion between the polyelectrolyte constituents and
less organelles. the surface. The engulfment properties may
Historically, “simple” coacervation refers to a account for recent observations concerning pos-
related process resulting in LLPS, however, only sible A-beta engulfment and fibrilization path-
involving a single biopolymer type. The mecha- ways as an antimicrobial response [67]. Surface
nism behind condensation for simple coacerva- wetting of solid structures in the biological con-
tion can be similar to that of complex coacervation text and in cells by coacervate-like liquid con-
if an intrinsically disordered biopolymer has spa- densates is a relatively unexplored area in cell
tially separated charges, and when modulation of biology, and may present yet another mechanism
pH and salt concentration can achieve effective facilitated by LLPS of how biomolecules interact
local charge neutralization for polymer chain and cellular constituents associate. Taken
condensation to LLPS. An adhesive foot mussel together, the process of complex coacervation
protein from Mytilus californianus provides one spans an enormous breadth of biological systems
illustrative example [64]. The formation of a and has even been attributed a role in the origin of
bifluidic sponge-like nanostructured network life [68].
undergoing water exchange is thought to be
driven by tuning the electrostatic interactions
between the polyelectrolyte constituents or their Tau Droplets
complexes. However, simple coacervation can
instead also be driven by other forces, such as With this foundational framework we now pose
hydrophobic or π-cation interactions. Broadly, the question whether tau can undergo LLPS in
the field of coacervation has recently been invig- vitro and in vivo, whether cofactors are required
orated by the realization that LLPS occurs in and whether in a condensed state, tau can transi-
vivo, in particular involving proteins involved in tion to fibers. This thinking led to consideration
neurodegenerative diseases, suggesting a possi- of LLPS as an intermediate regulatory state,
ble physiological or pathophysiologic role for which could redissolve into a soluble state or
these assemblies. Specifically in the case of tau, transition to irreversible aggregation/amyloid
both complex coacervation between tau and RNA fibrils. Intracellular aggregates consisting of the
or simple coacervation of hyperphosphorylated tau protein occur in many neurological condi-
tau have been reported on in the literature. tions with Alzheimer’s disease the most promi-
Importantly, in vitro studies can validate the nent among them. A variety of mutations in the
isolated properties of the components within a tau gene lead to another neurodegenerative con-
biological system to undergo LLPS.  The liquid dition labeled frontotemporal dementia. Tau has
phase consists of a dense polyelectrolyte-rich been regarded as in a dynamic equilibrium
phase, which form droplets and retain liquid-like between a microtubule-bound state and the solu-
mobility, as shown by NMR measurements [65, tion state; however phase separation studies of
66] and a dilute polyelectrolyte-depleted phase. tau suggest this two compartment model is sim-
The concentrated liquid phase is termed the coac- plistic. Tau exists in multiple phase states, as a
ervate phase and the dilute phase the equilibrium solid when it polymerizes into fibrils and in liq-
phase or solution phase, within which the coacer- uid liquid phase separated state. Transitions
vate characteristically appears as droplets of sub-­ across these states are central to understanding
micron to several micron dimensions that may how tau contributes to disease. Under disease
further coalesce to larger droplets, even macro- conditions tau self-assembles into solid fibrils
scopic phase separation. Complex coacervate flu- that eventually lead to highly insoluble polymeric
ids critically rely on their low interfacial tension inclusions known as neurofibrillary tangles. The
with respect to water, and thereby can wet or underlying biophysical process by which tau
332 K. S. Kosik and S. Han

transitions to an insoluble fibril is unknown. An with different tau:tRNA ratios coexist. The higher
early clue about the nature of this transition came order multimeric structures are likely held
from the observation that polyanions, such as together by weak interactions between multiple
heparin, promote tau fibrillization [69]. Although tau proteins and tRNA.
less effectively, RNA can also induce tau fibril- Along with the recognition of tau as intrinsi-
lization [70, 71], and unlike heparin, RNA is cally disordered, tau shares many properties with
present intracellularly, making it accessible to other RNA-binding proteins involved in neurode-
interact with tau. These early observations generation. These proteins include FUS [10, 65,
showed that certain types of tau interactions drive 73] TDP-43 [11], C9ORF72 [14, 74], TIA1 [75],
fibril formation and the fact that these interac- hnRNPA2B1 and hnRNPA1 [7, 76, 77], all of
tions involve polyanions suggested the possibil- which can undergo liquid-liquid phase-­separation
ity of a phase separation due to their tendency to (LLPS) from the surrounding aqueous medium to
decrease solubility and favor entropy driven form droplets in vitro.
effects [72]. Coupling of assembly and phase That tau can bind to RNA in living cells
separation of multivalent macromolecules is an prompted a more detailed quantitative analysis of
important organizing principle for biomolecular the in vitro conditions under which tau phase
condensates. separates. We showed that tau-RNA complex-
Our experiments extended the in vitro data of ation can lead to complex coacervation [1], and
tau-polyanion binding to prove that tau can bind we also know from literature and our own experi-
RNA in living cells. These experiments utilized a ments that tau can undergo simple coacervation,
protein-RNA crosslinking method followed by either when tau is phosphorylated or under high
antibody pull down of tau crosslinked to RNAs salt conditions. In the complex coacervation pro-
followed by RNAseq. Interestingly, tau-RNA cess, multiple tau molecules weakly bind RNA,
binding showed some enhancement for tRNAs and when overall charge matching is achieved
even after normalizing for their relative abun- between the polycation, tau, and polyanion,
dance. Among the tRNAs, tRNAArg topped the RNA, tau undergoes reversible condensation and
list as differentially selected. Full-length recom- liquid-liquid phase separation into micrometer-­
binant human tau (4R2N) induced a gel shift in sized droplets detected as a turbid solution. The
unacetylated tRNALys to yield a dissociation con- spontaneous and reversible droplet formation
stant (Kd) of 460 ± 47 nM. The derived Hill coef- suggests that tau is held in a fluid state with low
ficient was 2.8, implying cooperative binding of energy-barrier between the dilute solution and
multiple tau proteins to tRNA.  Isothermal titra- the complex coacervate condensate, as well as
tion calorimetry experiments independently con- with a modest free energy difference between
firmed the affinity of tau binding to tRNA them, toggled by interactions mediated by ions
(Kd = 735 ± 217 nM). The dissociation constant and hydration water. Key experimentally acces-
for 4R2N tau binding to a random 43 nucleotide sible parameters that tune LLPS were found to be
RNA sequence was similar, suggesting that under the biopolymer concentration, salt concentration,
in vitro conditions, tau effectively and non-­ pH and temperature, crowding pressure, post-­
specifically binds RNA.  The gel shift assay translational charge modification of the protein
showed multiple bands corresponding to high and the presence of biopolymer constituents that
molecular weight protein-RNA complexes pres- engage in multivalent electrostatic interactions,
ent at different tau:tRNA stoichiometries. The including RNA, single-stranded DNA and intrin-
fraction of bound tRNA to 4R2N tau from the sically disordered proteins (IDPs). Droplet for-
low and high molecular weight bands, plotted as mation followed spontaneously from the mixing
a function of tau:tRNA molar ratios represent a of two oppositely charged biopolymers, tau and
range of stoichiometries from 1:1 to 6:1. Thus, RNA, within a given range of pH, salt and protein
multiple tau molecules bind tRNA with increas- concentration, as well as temperature. This pro-
ing tau concentrations and multiple populations cess is fully reversible and reproducible,
24  Tau Condensates 333

c­haracteristic of equilibrium states. Complex solution state. In fact, the result of this experi-
coacervation of tau as a function of temperature ment was quite remarkable, in that the EPR line-
verified the process to be entropy-driven and shape of labeled Δtau187 within the droplet
likely toggled by the release of hydration water— phase was indistinguishable from that of
not counterions, as commonly assumed in the Δtau187 in the dilute solution state when coacer-
literature—that reduces the net excluded volume vated with poly(A) RNA or tRNA [1]. In con-
of the hydrated biopolymer constituents. [78]. trast, the cw EPR line shape of spin label at site
Under physiological temperature conditions 322 dramatically broadened within minutes of
(37 °C), droplet formation was observed at pro- adding heparin—a highly effective aggregation
tein concentrations as low as 2.5–5  μM, in the inducer of tau. Therefore, the condensation of
presence of salt concentrations as high as 100 mM Δtau187 to high protein concentration and for-
when glycerol was added as a crowding reagent mation of long-range associations with RNA
to mimic the intracellular environment. All of the within droplets, does not induce dehydration and
relevant parameters interact to maintain the elec- aggregation of tau as indicated by the retention of
trostatic charge ratio at approximately 1:1. the solution state dynamical properties. These
Importantly, tau coacervation can occur under results were further supported by doubly spin
physiologic conditions. Tau droplets also demon- labeling at tau sites 272 and 285 to access the dis-
strate a low interfacial tension that promotes tances across the PHF6 and PHF6∗ hexapeptide
fusion and coating. This phenomenon is associ- region by double electron-electron resonance
ated with low cohesive energy between hydrated (DEER) spectroscopy [82]. This PHF6(∗) region
polyelectrolyte complexes and weakly bound of the tau sequence packs into the β-sheet core
water constituents, consistent with high internal once fibrils are formed of tau [81], and hence
fluid dynamics [79, 80]. Tau droplets were clearly undergoes a dramatic opening from a compact to
visible with a highly spherical boundary, and two a fully extended local conformation, well before
droplets frequently merged into one with no vis- fibril formation is observed, and remains extended
ible boundary at the fusion interface left, indicat- as they pack into β-sheet structures as part of
ing that the droplets are fluidic with a relatively insoluble fibrils [82]. Dipolar EPR spectroscopy,
low interfacial tension. Confocal microscopy widely known as dipolar electron electron reso-
images of fluorescence-labeled tau verified that nance (DEER), showed that the mean distance
tau was predominantly contained within the flanking the PHF6∗ region of Δtau187 remained
droplet. unchanged from dilute solution state to when tau
Within a condensed complex coacervate fluid, was condensed into a concentrated complex
held together by non-specific and weak electro- coacervate phase in association with poly(A)
static interactions, the polyelectrolyte constitu- RNA or tRNA, unless aggregation-inducing fac-
ents must maintain their hydration layer and tors were added or present. This is in contrast to
remain dynamic which we proved by labeling a the conformational change that Δtau187 under-
tau truncation known as Δtau187 (4R tau 255– goes, within minutes upon the addition of hepa-
441) with fluorescence or spin labels at cysteine rin, of dramatic extension of the mean distance
site 322. Confocal fluorescence imaging showed between the labels flanking the PHF6(∗) seg-
that fluorescence labeled Δtau187 was predomi- ment, and hence signifying the opening of the
nantly localized within the droplet. Continuous PHF6(∗) region of Δtau187 followed by neat
wave electron paramagnetic resonance (cw EPR) stacking into β-sheets [82].
spectral line shape analysis of spin labeled These studies revealed that tau in the phase
Δtau187 revealed that the protein side chain separated state can maintain the same conforma-
dynamics and degrees of freedom of the tethered tion as in its solution state, depending on the
spin label of Δtau187, associated with poly(A) cofactor and tau variant used, despite the con-
RNA or tRNA, in the dense liquid coacervate straints of very high tau concentration and strong
phase was like that of spin labeled tau in dilute interaction and electrostatic association with
334 K. S. Kosik and S. Han

RNA.  The same studies raise the question a map of the free energy balance between the
whether tau in its droplet state is on pathway to condensed LLPS and dilute phase, constituting
fibril formation. We observed a low-level the phase diagram. By constructing the phase
Thioflavin T (ThT) fluorescence under tau-­ diagram from cloud-point measurements at the
poly(U) RNA droplet forming conditions that onset of complex coacervation under varying
gradually increased over 15 h [1]. When the salt conditions of temperature, salt, and polymer con-
concentration was increased to conditions under centrations we established the features of phase
which tau-RNA associations are weakened and coexistence boundaries which were modeled
droplets disperse, the ThT fluorescence was using theory and simulation. The resulting model
nearly eliminated. However, even after 15  h of supported the experimental observations that the
incubation, the ThT fluorescence intensity from equilibrium window for the complex coacerva-
the tau-RNA droplet samples was significantly tion of tau and RNA under cellular conditions is
less than occurred in the presence of heparin, narrow. Extrapolating from the empirically deter-
under similar charge ratio and mass concentra- mined phase diagram computed from in vitro
tion and following a brief incubation time of less experiments and validated by simulation, we
than 20 min. ThT fluorescence is commonly used conclude that the narrow window of tau-RNA
as an assay to detect β-sheet content, suggesting LLPS likely resides within or near physiologic
that droplet formation of tau increases its aggre- conditions of 37 °C, 150 mM salt concentration
gation propensity. Prolonged residence in this and cellular crowding pressure. The narrow win-
phase state may induce β-sheet formation, sug- dow for LLPS may account for pathological
gesting that the highly condensed phase state of potential of the tau splice site mutations which
tau can be a precursor to fibril formation. This subtly increase the 4r to 3R ratio, and therefore
observation is consistent with experiments that alter one of the finely tunable elements that con-
show liquid compartments of FUS convert with trol phase separation. Likewise, the extended
time into an aberrant aggregated state [10]. In the projection of MAP 2 may prevent its microtubule-­
cited study, FUS droplets carrying FUS muta- binding domain, which is highly homologous to
tions or aged droplets increased their fusion time tau from forming condensates vulnerable to
when single droplets were manipulated by laser fibrillization.
tweezers. Over time some of the droplets had Post translational modifications involve
short fibers and others had longer fibers that changes in the charge state of tau, again altering
extended beyond the boundary of the droplet the propensity for LLPS, while cellular processes
eventually taking on the appearance of a that alter the local pH would similarly affect the
“starburst.” LLPS propensity of tau. Fragmentation of tau, or
Having observed several tunable conditions in changes in the local concentration of RNA spe-
which tau/RNA transitions towards a homoge- cies will also shift the LLPS equilibrium. Thus,
neous phase or an LLPS state, we characterized our finding that LLPS formation occurs within a
the phase diagram of tau/RNA LLPS using narrow range of conditions close to or within cel-
experiment and simulation [78]. Using an lular conditions imply that there can be a myriad
N-terminus truncated version of the longest iso- of possible biological conditions that can regu-
form of human 4R tau in vitro, we showed that late the formation, dissolution, or in fact solidifi-
tau/RNA complexation is reversible, and tau cation and transformation, of the tau droplet
remains dynamic without a persistent structure state.
within the dense phase. The phase coexistence
curve separating a supernatant phase from a con-
densate phase is determined by the concentration,
temperature, salt, and the nature of the interac-
tion strength between the various solution con-
stituents, including the solvent. This work offers
24  Tau Condensates 335

 he Necessity for Cofactors in Tau

T factors available, not merely its dissolution from
Droplet Formation tau fibrils. Furthermore the high seeding capacity
of transgenic tauopathy mouse brain extracted
In vitro tau fibrillation of tau is typically triggered tau fibrils is exhausted after one generation, while
with cofactors, most commonly heparin [69], but supplementation with RNA cofactors resulted in
also other cofactors such as RNA [70] or arachi- sustained seeding over multiple generations. We
donic acid [83]. Seeding tau aggregation utilizes suggest that a polyelectrolyte cofactor is essential
premade fibrils added to fresh monomers, but the for the stability of fibrils prepared in  vitro.
seeding process is improved by the presence of Targeting of assembly-inducing cofactors repre-
cofactors (RNA or heparin) in solution [84, 85]. sents a unique approach to therapeutics.
Heparin is a limiting factor in fibril formation Remarkably, it is the same type of cofactor
[86, 87], but whether heparin is part of [88] or not that appears to stabilize tau fibrils and promote
part of mature fibrils [87, 89] is unclear. In our the seeding activity of tau fibrils by overcoming
work we addressed whether cofactors are crucial the seeding barrier, as well as induce LLPS of
constituents of mature fibrils and contribute to tau. We have shown that in the presence of aggre-
their stability, or whether they only catalyze gation promoting factors or aggregation-prone
aggregation toward a self-sustained protein tau variants can transform tau from a droplet state
assembly [90]. Because seeding is presumed to to a fibril state. This implies that there is a path-
convert naïve tau monomers into aggregates [91, way from tau in solution to the fibril state via the
92], the hypothesis emerged that in vivo, mono- droplet state. However, whether the droplet is an
meric tau can spontaneously polymerize into fila- obligatory intermediate state, or even whether the
ments when an appropriate seed template is droplet path is a biologically relevant avenue is
provided. Therefore, in searching for the basis of an open and important question.
aggregation, the focus has been on the properties
of tau itself, such as hyperphosphorylation, cleav- Acknowledgements Many individuals from the Kosik
age, high local concentrations, and alternative and Han labs contributed to the work described here
through their laboratory experiments, their ideas and the
splicing, but only marginally in abnormal inter- lively discussions at a weekly meetings held jointly
actions with cofactors. between the two labs. Contributing lab members are
We have recently shown that mature fibrils Xuemei Zhang, Yanxian Lin, Neil A. Eschmann, Hongjun
made of 4R tau variants, prepared with heparin or Zhou, Jennifer Rauch, Israel Hernandez, Elmer Guzman,
Dasol Han, Andrew Longhini, Yann Fichou, Michael
RNA, spontaneously depolymerize and release Vigers, Zhikai Zeng, Madhur Srivastava, Timothy
monomers when their cofactors are removed J. Keller. Additional contributions from Joan Shea, Glen
[90]. We next asked if the tau monomers released Fredrickson and James McCarty regarding the tau/RNA
from the digested fibrils retained the conforma- phase diagram were invaluable.
Work described here was funded by the Tau
tion assumed in the fibril. If so, one would expect Consortium, the National Institutes of Health
a much higher seeding capacity of the tau mono- 1U54NS100717, R01AG056058, Cohen Veterans
mers; however the seeding propensity of heparin Bioscience, The Edward N. & Della L. Thome Memorial
and RNA fibrils did not significantly differ Foundation and the Leo and Ann Albert Charitable Trust.
between digested and nondigested fibrils, sug-
gesting that the monomers released from fibrils
are not seeding-active. Therefore, the loss of References
seeding capacity is due to the loss of fibril con-
1. Zhang X, Lin Y, Eschmann NA, Zhou H, Rauch JN,
tent after cofactor digestion and the monomers Hernandez I, Guzman E, Kosik KS, Han S.  RNA
derived from fibrils have no memory of the fibril stores tau reversibly in complex coacervates. PLoS
state, neither in conformation nor in seeding Biol. 2017;15:e2002183.
capacity. Tau may adopt seeding competent, or 2. Ambadipudi S, Biernat J, Riedel D, Mandelkow E,
Zweckstetter M. Liquid-liquid phase separation of the
even seeding active, conformers in solution or
oligomer state, but there must be other promoting
336 K. S. Kosik and S. Han

microtubule-binding repeats of the Alzheimer-related 13. Kwon I, Xiang S, Kato M, Wu L, Theodoropoulos

protein Tau. Nat Commun. 2017;8:275. P, Wang T, Kim J, Yun J, Xie Y, Mcknight SL. Poly-­
3. Hernandez-vega A, Braun M, Scharrel L, Jahnel M, dipeptides encoded by the C9orf72 repeats bind
Wegmann S, Hyman BT, Alberti S, Diez S, Hyman nucleoli, impede RNA biogenesis, and kill cells.
AA. Local nucleation of microtubule bundles through Science. 2014;345:1139–45.
tubulin concentration into a condensed tau phase. Cell 14. Lee KH, Zhang P, Kim HJ, Mitrea DM, Sarkar M,
Rep. 2017;20:2304–12. Freibaum BD, Cika J, Coughlin M, Messing J,
4. Ferreon JC, Jain A, Choi KJ, Tsoi PS, Mackenzie KR, Molliex A, Maxwell BA, Kim NC, Temirov J, Moore
Jung SY, Ferreon AC. Acetylation disfavors tau phase J, Kolaitis RM, Shaw TI, Bai B, Peng J, Kriwacki
separation. Int J Mol Sci. 2018;19:E1360. RW, Taylor JP. C9orf72 dipeptide repeats impair the
5. Wegmann S, Eftekharzadeh B, Tepper K, Zoltowska assembly, dynamics, and function of membrane-less
KM, Bennett RE, Dujardin S, Laskowski PR, organelles. Cell. 2016;167:774–788 e17.
Mackenzie D, Kamath T, Commins C, Vanderburg 15. Boeynaems S, Bogaert E, Kovacs D, Konijnenberg A,
C, Roe AD, Fan Z, Molliex AM, Hernandez-Vega Timmerman E, Volkov A, Guharoy M, De Decker M,
A, Muller D, Hyman AA, Mandelkow E, Taylor Jaspers T, Ryan VH, Janke AM, Baatsen P, Vercruysse
JP, Hyman BT.  Tau protein liquid-liquid phase T, Kolaitis RM, Daelemans D, Taylor JP, Kedersha
separation can initiate tau aggregation. EMBO N, Anderson P, Impens F, Sobott F, Schymkowitz J,
J. 2018;37:e98049. Rousseau F, Fawzi NL, Robberecht W, Van Damme
6. Banani SF, Lee HO, Hyman AA, Rosen P, Tompa P, Van Den Bosch L.  Phase separation of
MK.  Biomolecular condensates: organizers of C9orf72 dipeptide repeats perturbs stress granule
cellular biochemistry. Nat Rev Mol Cell Biol. dynamics. Mol Cell. 2017;65:1044–1055 e5.
2017;18:285–98. 16. Nott TJ, Petsalaki E, Farber P, Jervis D, Fussner E,
7. Molliex A, Temirov J, Lee J, Coughlin M, Kanagaraj Plochowietz A, Craggs TD, Bazett-Jones DP, Pawson
AP, Kim HJ, Mittag T, Taylor JP.  Phase separation T, Forman-Kay JD, Baldwin AJ.  Phase transition of
by low complexity domains promotes stress granule a disordered nuage protein generates environmen-
assembly and drives pathological fibrillization. Cell. tally responsive membraneless organelles. Mol Cell.
2015;163:123–33. 2015;57:936–47.
8. Li YR, King OD, Shorter J, Gitler AD.  Stress gran- 17. Han TW, Kato M, Xie S, Wu LC, Mirzaei H, Pei J,
ules as crucibles of ALS pathogenesis. J  Cell Biol. Chen M, Xie Y, Allen J, Xiao G, Mcknight SL. Cell-­
2013;201:361–72. free formation of RNA granules: bound RNAs iden-
9. Murakami T, Qamar S, Lin JQ, Schierle GS, Rees E, tify features and components of cellular assemblies.
Miyashita A, Costa AR, Dodd RB, Chan FT, Michel Cell. 2012;149:768–79.
CH, Kronenberg-Versteeg D, Li Y, Yang SP, Wakutani 18. Kwon I, Kato M, Xiang S, Wu L, Theodoropoulos
Y, Meadows W, Ferry RR, Dong L, Tartaglia GG, P, Mirzaei H, Han T, Xie S, Corden JL, Mcknight
Favrin G, Lin WL, Dickson DW, Zhen M, Ron D, SL. Phosphorylation-regulated binding of RNA poly-
Schmitt-Ulms G, Fraser PE, Shneider NA, Holt C, merase II to fibrous polymers of low-complexity
Vendruscolo M, Kaminski CF, St George-Hyslop domains. Cell. 2013;155:1049–60.
P.  ALS/FTD mutation-induced phase transition of 19. Sabari BR, Dall'agnese A, Boija A, Klein IA,

FUS liquid droplets and reversible hydrogels into Coffey EL, Shrinivas K, Abraham BJ, Hannett NM,
irreversible hydrogels impairs RNP granule function. Zamudio AV, Manteiga JC, Li CH, Guo YE, Day DS,
Neuron. 2015;88:678–90. Schuijers J, Vasile E, Malik S, Hnisz D, Lee TI, Cisse
10. Patel A, Lee HO, Jawerth L, Maharana S, Jahnel M, II, Roeder RG, Sharp PA, Chakraborty AK, Young
Hein MY, Stoynov S, Mahamid J, Saha S, Franzmann RA.  Coactivator condensation at super-enhancers
TM, Pozniakovski A, Poser I, Maghelli N, Royer LA, links phase separation and gene control. Science.
Weigert M, Myers EW, Grill S, Drechsel D, Hyman 2018;361(6400):pii: eaar3958.
AA, Alberti S.  A liquid-to-solid phase transition of 20.
Brangwynne CP, Eckmann CR, Courson DS,
the ALS protein FUS accelerated by disease mutation. Rybarska A, Hoege C, Gharakhani J, Julicher F,
Cell. 2015;162:1066–77. Hyman AA. Germline P granules are liquid droplets
11. Kato M, Han TW, Xie S, Shi K, Du X, Wu LC, that localize by controlled dissolution/condensation.
Mirzaei H, Goldsmith EJ, Longgood J, Pei J, Grishin Science. 2009;324:1729–32.
NV, Frantz DE, Schneider JW, Chen S, Li L, Sawaya 21. Brangwynne CP.  Phase transitions and size scal-

MR, Eisenberg D, Tycko R, Mcknight SL. Cell-free ing of membrane-less organelles. J  Cell Biol.
formation of RNA granules: low complexity sequence 2013;203:875–81.
domains form dynamic fibers within hydrogels. Cell. 22. Elbaum-Garfinkle S, Kim Y, Szczepaniak K, Chen
2012;149:753–67. CC, Eckmann CR, Myong S, Brangwynne CP.  The
12. Li HR, Chiang WC, Chou PC, Wang WJ, Huang
disordered P granule protein LAF-1 drives phase sepa-
JR.  TAR DNA-binding protein 43 (TDP-43) liquid-­ ration into droplets with tunable viscosity and dynam-
liquid phase separation is mediated by just a few aro- ics. Proc Natl Acad Sci U S A. 2015;112:7189–94.
matic residues. J Biol Chem. 2018;293:6090–8. 23. Feric M, Vaidya N, Harmon TS, Mitrea DM, Zhu
L, Richardson TM, Kriwacki RW, Pappu RV,
24  Tau Condensates 337

Brangwynne CP.  Coexisting liquid phases underlie 39.

Metschnikoff E.  Veber die Entwicklung der
nucleolar subcompartments. Cell. 2016;165:1686–97. Cecidomyienlarven aus dem Pseudovum. Arch furr
24. Mitrea DM, Cika JA, Guy CS, Ban D, Banerjee
Naturg Bd 1. 1865;1(1):304–10.
PR, Stanley CB, Nourse A, Deniz AA, Kriwacki 40. Ritter R. Die Entwicklung dur Geschlechtsorgane und
RW.  Nucleophosmin integrates within the nucleolus des Darmes bei Chiromomus. Zeit furr Wiss Zool Bd.
via multi-modal interactions with proteins displaying 1890;50:408–27.
R-rich linear motifs and rRNA. Elife. 2016;5:e13571. 41. Hegner RW.  Studies on germ cells. I The history of
25. Pak CW, Kosno M, Holehouse AS, Padrick SB,
the germ cells in insects with special reference to the
Mittal A, Ali R, Yunus AA, Liu DR, Pappu RV, Rosen Keimbahn-determinants II The origin and significance
MK.  Sequence determinants of intracellular phase of the Keimbahn-determinants in animals. J Morphol.
separation by complex coacervation of a disordered 1914;25:375–509.
protein. Mol Cell. 2016;63:72–85. 42. Kanai Y, Dohmae N, Hirokawa N. Kinesin transports
26. Zhang H, Elbaum-Garfinkle S, Langdon EM, Taylor RNA: isolation and characterization of an RNA-­
N, Occhipinti P, Bridges AA, Brangwynne CP, transporting granule. Neuron. 2004;43:513–25.
Gladfelter AS.  RNA controls polyQ protein phase 43. Shiina N, Shinkura K, Tokunaga M.  A novel RNA-­
transitions. Mol Cell. 2015;60:220–30. binding protein in neuronal RNA granules: regula-
27.
Mao YS, Zhang B, Spector DL.  Biogenesis tory machinery for local translation. J  Neurosci.
and function of nuclear bodies. Trends Genet. 2005;25:4420–34.
2011;27:295–306. 44. SHIINA N, TOKUNAGA M.  RNA granule protein
28. Mitrea DM, Kriwacki RW. Phase separation in biol- 140 (RNG140), a paralog of RNG105 localized to dis-
ogy; functional organization of a higher order. Cell tinct RNA granules in neuronal dendrites in the adult
Commun Signal. 2016;14:1. vertebrate brain. J Biol Chem. 2010;285:24260–9.
29. Handwerger KE, Gall JG. Subnuclear organelles: new 45. Shiina N, Yamaguchi K, Tokunaga M. RNG105 defi-
insights into form and function. Trends Cell Biol. ciency impairs the dendritic localization of mRNAs
2006;16:19–26. for Na+/K+ ATPase subunit isoforms and leads to
30. Decker CJ, Parker R.  P-bodies and stress gran-
the degeneration of neuronal networks. J  Neurosci.
ules: possible roles in the control of translation and 2010;30:12816–30.
mRNA degradation. Cold Spring Harb Perspect Biol. 46. Schisa JA, Pitt JN, Priess JR. Analysis of RNA asso-
2012;4:a012286. ciated with P granules in germ cells of C. elegans
31. Milovanovic D, Wu Y, Bian X, De Camilli P. A liq- adults. Development. 2001;128:1287–98.
uid phase of synapsin and lipid vesicles. Science. 47. Leatherman JL, Jongens TA. Transcriptional silencing
2018;361:604–7. and translational control: key features of early germ-
32. Chong S, Dugast-Darzacq C, Liu Z, Dong P, Dailey line development. Bioessays. 2003;25:326–35.
GM, Cattoglio C, Heckert A, Banala S, Lavis L, 48. Bashkirov VI, Scherthan H, Solinger JA, Buerstedde
Darzacq X, Tjian R. Imaging dynamic and selective JM, Heyer WD. A mouse cytoplasmic exoribonucle-
low-complexity domain interactions that control gene ase (mXRN1p) with preference for G4 tetraplex sub-
transcription. Science. 2018;361:eaar2555. strates. J Cell Biol. 1997;136:761–73.
33. Cho WK, Spille JH, Hecht M, Lee C, Li C, Grube 49. Eystathioy T, Chan EK, Tenenbaum SA, Keene JD,
V, Cisse II.  Mediator and RNA polymerase II clus- Griffith K, Fritzler MJ. A phosphorylated cytoplasmic
ters associate in transcription-dependent condensates. autoantigen, GW182, associates with a unique popu-
Science. 2018;361:412–5. lation of human mRNAs within novel cytoplasmic
34. Gilks N, Kedersha N, Ayodele M, Shen L, Stoecklin speckles. Mol Biol Cell. 2002;13:1338–51.
G, Dember LM, Anderson P. Stress granule assembly 50. Ingelfinger D, Arndt-Jovin DJ, Luhrmann R, Achsel
is mediated by prion-like aggregation of TIA-1. Mol T.  The human LSm1-7 proteins colocalize with the
Biol Cell. 2004;15:5383–98. mRNA-degrading enzymes Dcp1/2 and Xrnl in dis-
35. Alberti S, Halfmann R, King O, Kapila A, Lindquist tinct cytoplasmic foci. RNA. 2002;8:1489–501.
S.  A systematic survey identifies prions and illumi- 51. Van Dijk E, Cougot N, MEYER S, Babajko S, Wahle
nates sequence features of prionogenic proteins. Cell. E, Seraphin B.  Human Dcp2: a catalytically active
2009;137:146–58. mRNA decapping enzyme located in specific cyto-
36. Knowles RB, Sabry JH, Martone ME, Deerinck TJ, plasmic structures. EMBO J. 2002;21:6915–24.
Ellisman MH, Bassell GJ, Kosik KS.  Translocation 52. Sheth U, Parker R.  Decapping and decay of mes-
of RNA granules in living neurons. J  Neurosci. senger RNA occur in cytoplasmic processing bodies.
1996;16:7812–20. Science. 2003;300:805–8.
37. Krichevsky AM, Kosik KS. Neuronal RNA granules: 53. Teixeira D, Sheth U, Valencia-Sanchez MA, Brengues
a link between RNA localization and stimulation-­ M, Parker R.  Processing bodies require RNA for
dependent translation. Neuron. 2001;32:683–96. assembly and contain nontranslating mRNAs. RNA.
38. Rook MS, Lu M, Kosik KS. CaMKIIalpha 3' untrans- 2005;11:371–82.
lated region-directed mRNA translocation in living 54. Anderson P, Kedersha N. RNA granules. J Cell Biol.
neurons: visualization by GFP linkage. J  Neurosci. 2006;172:803–8.
2000;20:6385–93.
338 K. S. Kosik and S. Han

55. Nover L, Scharf KD, Neumann D. Formation of cyto- induced by sulphated glycosaminoglycans. Nature.
plasmic heat shock granules in tomato cell cultures 1996;383:550–3.
and leaves. Mol Cell Biol. 1983;3:1648–55. 70. Kampers T, Friedhoff P, Biernat J, Mandelkow

56. Collier NC, Heuser J, Levy MA, Schlesinger
EM, Mandelkow E.  RNA stimulates aggrega-
MJ.  Ultrastructural and biochemical analysis of the tion of microtubule-­ associated protein tau into
stress granule in chicken embryo fibroblasts. J  Cell Alzheimer-like paired helical filaments. FEBS Lett.
Biol. 1988;106:1131–9. 1996;399:344–9.
57. Arrigo AP, Suhan JP, Welch WJ.  Dynamic changes 71. Wang X, Wang D, Zhao J, Qu M, Zhou X, He H,
in the structure and intracellular locale of the mam- He R.  The proline-rich domain and the microtubule
malian low-molecular-weight heat shock protein. Mol binding domain of protein tau acting as RNA binding
Cell Biol. 1988;8:5059–71. domains. Protein Pept Lett. 2006;13:679–85.
58. Nover L, Scharf KD, Neumann D. Cytoplasmic heat 72. Flory PJ.  Principles of polymer chemistry. Ithaca/

shock granules are formed from precursor particles New York: Cornell Univ. Press; 1953.
and are associated with a specific set of mRNAs. Mol 73. Shin Y, Berry J, Pannucci N, Haataja MP, Toettcher
Cell Biol. 1989;9:1298–308. JE, Brangwynne CP. Spatiotemporal control of intra-
59. Scharf KD, Heider H, Hohfeld I, Lyck R, Schmidt E, cellular phase transitions using light-activated opto-
Nover L. The tomato Hsf system: HsfA2 needs inter- Droplets. Cell. 2017;168:159–171 e14.
action with HsfA1 for efficient nuclear import and 74. Lin Y, Mori E, Kato M, Xiang S, Wu L, Kwon I,
may be localized in cytoplasmic heat stress granules. Mcknight SL. Toxic PR poly-dipeptides encoded by
Mol Cell Biol. 1998;18:2240–51. the C9orf72 repeat expansion target LC domain poly-
60. Kedersha NL, Gupta M, Li W, Miller I, Anderson mers. Cell. 2016;167:789–802 e12.
P.  RNA-binding proteins TIA-1 and TIAR link 75. Mackenzie IR, Nicholson AM, Sarkar M, Messing J,
the phosphorylation of eIF-2 alpha to the assem- Purice MD, Pottier C, Annu K, Baker M, Perkerson
bly of mammalian stress granules. J  Cell Biol. RB, Kurti A, Matchett BJ, Mittag T, Temirov J,
1999;147:1431–42. Hsiung GR, Krieger C, Murray ME, Kato M, Fryer
61.
Buchan JR, Kolaitis RM, Taylor JP, Parker JD, Petrucelli L, Zinman L, Weintraub S, Mesulam
R. Eukaryotic stress granules are cleared by autophagy M, Keith J, Zivkovic SA, Hirsch-Reinshagen V, Roos
and Cdc48/VCP function. Cell. 2013;153:1461–74. RP, Zuchner S, Graff-Radford NR, Petersen RC,
62. Van Treeck B, Protter DSW, Matheny T, Khong A, Caselli RJ, Wszolek ZK, Finger E, Lippa C, Lacomis
Link CD, Parker R.  RNA self-assembly contributes D, Stewart H, Dickson DW, Kim HJ, Rogaeva E,
to stress granule formation and defining the stress Bigio E, Boylan KB, Taylor JP, Rademakers R. TIA1
granule transcriptome. Proc Natl Acad Sci U S A. mutations in amyotrophic lateral sclerosis and fronto-
2018;115:2734–9. temporal dementia promote phase separation and alter
63. Bungenberg DJH.  Crystallisation-coacervation-­ stress granule dynamics. Neuron. 2017;95:808–816
flocculation. Colloid Sci. 1949;2:232–58. e9.
64. Wei W, Tan Y, Martinez Rodriguez NR, Yu J,
76. Kim HJ, Kim NC, Wang YD, Scarborough EA,

Israelachvili JN, Waite JH.  A mussel-derived one Moore J, Diaz Z, Maclea KS, Freibaum B, Li S,
component adhesive coacervate. Acta Biomater. Molliex A, Kanagaraj AP, Carter R, Boylan KB,
2014;10:1663–70. Wojtas AM, Rademakers R, Pinkus JL, Greenberg
65. Burke KA, Janke AM, Rhine CL, Fawzi NL. Residue-­ SA, Trojanowski JQ, Traynor BJ, Smith BN, Topp S,
by-­residue view of in vitro FUS granules that bind the Gkazi AS, Miller J, Shaw CE, Kottlors M, Kirschner
C-terminal domain of RNA polymerase II. Mol Cell. J, Pestronk A, Li YR, Ford AF, Gitler AD, Benatar M,
2015;60:231–41. King OD, Kimonis VE, Ross ED, Weihl CC, Shorter
66. Brady JP, Farber PJ, Sekhar A, Lin YH, Huang R, J, Taylor JP.  Mutations in prion-like domains in
Bah A, Nott TJ, Chan HS, Baldwin AJ, Forman-Kay hnRNPA2B1 and hnRNPA1 cause multisystem pro-
JD, Kay LE. Structural and hydrodynamic properties teinopathy and ALS. Nature. 2013;495:467–73.
of an intrinsically disordered region of a germ cell-­ 77. Lin Y, Protter DS, Rosen MK, Parker R.  Formation
specific protein on phase separation. Proc Natl Acad and maturation of phase-separated liquid droplets by
Sci U S A. 2017;114:E8194–203. RNA-binding proteins. Mol Cell. 2015;60:208–19.
67. Eimer WA, Vijaya Kumar DK, Navalpur Shanmugam 78. Lin Y, Mccarty J, Rauch JN, Delaney KT, Kosik KS,
NK, Rodriguez AS, Mitchell T, Washicosky Fredrickson GH, Shea JE, Han S. Narrow equilibrium
KJ, Gyorgy B, Breakefield XO, Tanzi RE, Moir window for complex coacervation of tau and RNA
RD.  Alzheimer’s disease-associated beta-amyloid under cellular conditions. bioRxiv. 2018; https://doi.
is rapidly seeded by herpesviridae to protect against org/10.1101/424358.
brain infection. Neuron. 2018;99(56–63):e3. 79. Jho Y, Yoo HY, Lin Y, Han S, Hwang DS. Molecular
68. Oparin A.  Origin of life. Moscow: Moscow Worker and structural basis of low interfacial energy of com-
Publisher; 1924. plex coacervates in water. Adv Colloid Interface Sci.
69. Goedert M, Jakes R, Spillantini MG, Hasegawa M, 2017;239:61–73.
Smith MJ, Crowther RA.  Assembly of microtubule-­ 80. Huang KY, Yoo HY, Jho Y, Han S, Hwang

associated protein tau into Alzheimer-like filaments DS.  Bicontinuous fluid structure with low cohesive
24  Tau Condensates 339

energy: molecular basis for exceptionally low interfa- inducer mediated tau polymerization. Biochemistry.
cial tension of complex coacervate fluids. ACS Nano. 2007;46:8838–49.
2016;10:5051–62. 88. Sibille N, Sillen A, Leroy A, Wieruszeski JM, Mulloy
81. Von Bergen M, Barghorn S, Biernat J, Mandelkow B, Landrieu I, Lippens G. Structural impact of heparin
EM, Mandelkow E.  Tau aggregation is driven by a binding to full-length Tau as studied by NMR spec-
transition from random coil to beta sheet structure. troscopy. Biochemistry. 2006;45:12560–72.
Biochim Biophys Acta. 2005;1739:158–66. 89. Von Bergen M, Barghorn S, Muller SA, Pickhardt
82. Eschmann NA, Georgieva ER, Ganguly P, Borbat PP, M, Biernat J, Mandelkow EM, Davies P, Aebi U,
Rappaport MD, Akdogan Y, Freed JH, Shea JE, Han Mandelkow E.  The core of tau-paired helical fila-
S. Signature of an aggregation-prone conformation of ments studied by scanning transmission electron
tau. Sci Rep. 2017;7:44739. microscopy and limited proteolysis. Biochemistry.
83. Wilson DM, Binder LI.  Free fatty acids stimulate 2006;45:6446–57.
the polymerization of tau and amyloid beta pep- 90. Fichou Y, Lin Y, Rauch JN, Vigers M, Zeng Z,

tides. In vitro evidence for a common effector of Srivastava M, Keller TJ, Freed JH, Kosik KS, Han
pathogenesis in Alzheimer’s disease. Am J  Pathol. S.  Cofactors are essential constituents of stable and
1997;150:2181–95. seeding-active tau fibrils. Proc Natl Acad Sci U S A.
84. Dinkel PD, Holden MR, Matin N, Margittai M. RNA 2018;115(52):13234–9.
binds to Tau fibrils and sustains template-assisted 91. Falcon B, Cavallini A, Angers R, Glover S, Murray
growth. Biochemistry. 2015;54:4731–40. TK, Barnham L, Jackson S, O'Neill MJ, Isaacs
85. Meyer V, Dinkel PD, Rickman Hager E, Margittai AM, Hutton ML, Szekeres PG, Goedert M, Bose
M.  Amplification of Tau fibrils from minute quanti- S. Conformation determines the seeding potencies of
ties of seeds. Biochemistry. 2014;53:5804–9. native and recombinant Tau aggregates. J Biol Chem.
86. Ramachandran G, Udgaonkar JB. Understanding the 2015;290:1049–65.
kinetic roles of the inducer heparin and of rod-like 92. Woerman AL, Aoyagi A, Patel S, Kazmi SA, Lobach
protofibrils during amyloid fibril formation by Tau I, Grinberg LT, Mckee AC, Seeley WW, Olson SH,
protein. J Biol Chem. 2011;286:38948–59. Prusiner SB.  Tau prions from Alzheimer’s disease
87. Carlson SW, Branden M, Voss K, Sun Q, Rankin and chronic traumatic encephalopathy patients propa-
CA, Gamblin TC.  A complex mechanism for gate in cultured cells. Proc Natl Acad Sci U S A.
2016;113:E8187–96.
 Liquid-Liquid Phase Separation
of Tau Protein in Neurobiology 25
and Pathology

Susanne Wegmann

States of Tau in the Cell aggregates that often have a highly ordered

structure.
The microtubule associated protein tau (MAPT) Only recently, a new class of dynamic macro-
is expressed in neuron of the central nervous sys- molecular tau assemblies has been identified:
tem, where is involved in microtubule regulation Liquid condensed phases of tau, which form
in the axon. In the adult human brain, 6 isoforms spontaneously in aqueous solution and have
of tau coexist and are generated by alternative liquid-­like fluid properties. These liquid tau clus-
splicing of one precursor mRNA. Tau is an intrin- ters contain high concentrations of tau, and they
sically disordered protein (IDP) with very few differ from monomeric and from aggregated tau
transient secondary structural elements, mainly in their biochemical and biophysical properties,
in its microtubule binding region (MTBR) and therefore provide a new mode of interaction
(Fig. 25.1a–c). Interestingly, although tau is one with other biomolecules and organelles
of the most abundant proteins in the CNS, murine (Fig. 25.1d).
knock-out models suggest only minor effects on
brain function upon tau ablation. In contrast, the
removal of tau appears to have a large positive Monomeric (or Dimeric) Soluble Tau
effect on neuronal health and survival in different
diseases associated stress situations. Soluble monomeric tau is intrinsically disor-
In a number of neurodegenerative diseases, dered, which reflects in a floppy molecular struc-
such as Alzheimer’s disease (AD), Frontotemporal ture that is guided by transient intramolecular
dementia (FTD), Pick’s disease, and supranu- interactions [1, 2]. The intrinsic disorder impli-
clear palsy, Tau owns a very negative prominence cates that tau is likely involved in a plethora of
as a mediator of neurotoxicity (likely in its oligo- interactions with other proteins and organelles
meric state), and the intracellular aggregation of [3]. Most prominently, tau binds microtubules
tau is a hallmark of these diseases. Until now it is [4–6], but also actin binding [7, 8], membrane
not clear, how and why the highly soluble intrin- association [9–11], kinase interactions through
sically disordered tau protein transitions into SH3 and SH2 domains [12, 13], and nuclear pore
interactions [14] have been reported.

S. Wegmann (*)
Deutsches Zentrum für Neurodegenerative
Erkrankungen (DZNE), Berlin, Germany
e-mail: susanne.wegmann@dzne.de

© Springer Nature Singapore Pte Ltd. 2019 341

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_25
342 S. Wegmann

Fig. 25.1  Human tau domains, disorder, and conden- dered (green regions in middle horizontal line), whereas
sation states. (a) Sequence and its domain structure of the some secondary structure is predicted for the repeat
longest brain isoform of human tau having two inserts (N1 domain (R1 to R4 and R′; pink regions in middle line). The
and N2) in the N-terminal projection domain and four predicted low complexity regions (grey bars; http://glob-
pseudo-repeats (R1-4) in the C-terminal half (2N4R; 441 plot.embl.de/) result mostly from high proline content. (c)
amino acids). Repeats R1-4, and small parts of the proline-­ Tau can exist in different assembly states that have differ-
rich region P2 and the region R’, are involved in the micro- ent physicochemical and interaction properties: soluble tau
tubule binding of tau, whereas the unstructured N-terminal monomers (or dimers) can reversibly condensate into liq-
half projects from the microtubule surface. (b) A protein uid dense phases, from which small soluble oligomeric tau
disorder prediction for full-length tau (2N4R; http://www. assemblies can form; oligomers could also  assemble
pondr.com/) shows that the projection domain, the proline- directly from monomers/dimers, and can evolve further
rich domain (P1+P2), and the C-terminus are mostly disor- into larger fibrillary aggregates
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 343

 ligomeric and Aggregated States

O uid droplets out of monomeric tau solutions can
of Tau initiate the aggregation of tau, and (2) condensed
tau phases (liquid or gel-like) can establish both
In AD and primary tauopathies, the aggregation molecular interactions with single molecules in
of tau into neurofibrillary tangles (NFTs) repre- the inner body and the outer surface of the con-
sents a pathological hallmark. Tau aggregation densates, and macromolecular interactions with
also occurs in other neurodegenerative diseases other polymer phases and cellular structures
such as dementia with Lewy bodies [15], chronic (organelles, filaments, etc.) at their surfaces.
traumatic brain encephalopathy [16], and prion In the following, we will describe the forma-
diseases [17]. NFTs contain highly ordered, tion and interactions of these condensed tau
amyloid-­like twisted or straight filamentous tau phases – as far as is known to date – with a focus
aggregates, whose structure has been described on these two aspects. The mechanism and bio-
by electron [18] and atomic force [19] micros- logical implication of tau phase separation by
copy, and recently was solved by Cryo-EM [20, coacervation with RNA molecules is covered in
21]. The core of these fibrils is assembled from another chapter of this book, authored by Kenneth
stacks of tau microtubule binding domain Kosik and colleagues.
(Fig. 25.1a; more details on tau domains in sec-
tion “Tau protein domains in LLPS”). The in
vitro assembly of tau aggregates seems to involve  rivers of Tau Liquid-Liquid Phase
D
oligomeric intermediates [22], and tau oligomers Separation
have also been suggested as the toxic species in
neurons and in vivo [23–26]. As of now, it is not I ntroduction to Protein Phase
clear if toxic oligomers are on-path with larger Separation
ordered aggregates of tau in vivo. In any case, the
mechanism, by which monomeric tau converts The adhoc liquid phase of proteins together with
into oligomers and/or amyloid-like aggregates is RNA in living cells has been first described for
unclear as well; the liquid condensates of tau p-granules in germline cells of Caenorhabditis
described in this chapter may provide a possible elegans [31]. In this system, the condensation of
mechanism for this transition. PGL-1 protein together with RNA enables the
establishment of a transcriptional gradient in the
cell, which is essential to break the symmetry
Liquid and Gel-Like Phases of Tau during development; similar mechanism exists
also for other proteins involved in development
In addition to the monomeric, oligomeric and of the nematode C. elegans [32]. The biophysical
aggregated phases of tau, another assembly state process, in which aqueous solutes in a solution
of tau in the cell has very recently been discov- spontaneous condensate into separate aqueous
ered [27–30]: dense liquid-like tau protein phases, is also called liquid-liquid phase separa-
phases, formed by liquid-liquid phase separation tion (LLPS). In the cell, the condensation of pro-
(LLPS) of tau from the surrounding aqueous teins (and RNA) into liquid dense phases is used
solution, can form spontaneously in a tau concen- to efficiently and reversibly generate membrane-­
tration dependent manner, or upon interaction less organelles with a high inner protein concen-
with polyanions such as heparin, the microtubule tration that can function, for example, as little
surface, or RNA (references). nanoreactors to catalyze translation, transcrip-
These phases can be seen as an intermediate tion, or other reactions [33]. A large number of
between monomeric/dimeric and oligomeric/ the proteins that are known to undergo LLPS are
aggregated tau species for two reasons: (1) they RNA binding proteins that co-condensate with
seem to facilitate the transition between the two RNA into liquid droplets under stress conditions
mentioned states in that the formation of tau liq- to form reversible intracellular stress granules
344 S. Wegmann

[34, 35]. The aberrant phase separation and tran- that contain aggregation prone sequences and
sition into aggregates have been observed for dif- have the ability to self-propagate certain protein
ferent RNA binding proteins, such as FUS [36, conformations and misfolding, similar to yeast
37], hnRNPA1 [38], and TDP-43 [39] in the con- prion proteins [46]; for example, PrLDs facilitate
text of disease-associated mutations and condi- both the LLPS and the misfolding of the RNA
tions in amyotrophic lateral sclerosis (ALS). binding proteins FUS, TDP43 and hnRNPA1/2.
Another ALS condition, the nuclear accumula- Note that lateral infectivity between individuals
tion of di-peptide repeat proteins derived from and species, a hallmark of prion diseases, does
C9orf72 hexanucleotide repeat expansions, has usually not occur in proteins with PrLDs.
been shown to be driven by LLPS and at the same The amino acid sequence in LCRs of LLPS
time also to interfere with LLPS of RNA binding proteins is enriched in polar amino acids such as
proteins [40, 41]. serine, glutamine, asparagine, and glycine, with
But also proteins with other functions have interspersed charged (e.g. lysine, arginine) or
been shown to undergo LLPS under physiologi- aromatic residues (e.g. tyrosine, phenylalanine)
cal model conditions in vitro and in cells. This [38, 50]. The residue composition determines the
includes the synaptic structural proteins synGAP phase behavior of LCRs, and with that regulates
together with PSD-95 [42] and synapsin-1 [43], their function. Accordingly, changes in the appar-
and the tumor suppressor SPOP [44]. And the ent sequence or the biochemical character of an
assembly of nucleoli, paraspeckles, and hetero- LDR, e.g. by mutations [30, 36, 39, 44, 51] or
chromatin, which organize RNA and chromatin post-translational modification [30, 37, 52], can
in the nucleus and thereby regulate gene tran- largely alter the phase behavior of LLPS proteins
scription, is likely based on LLPS of proteins as and either promote or inhibit the formation and
well [45–47]. Interestingly, LLPS of another stability of the liquid dense phases. External fac-
microtubule binding protein, BugZ, has been tors that change the interactions involved in the
shown to be important for spindle assembly dur- formation of liquid dense protein phases are for
ing mitosis [48]. example protein concentration, molecular crowd-
The molecular base of protein condensation ing (which changes the apparent protein concen-
into liquid droplets seems to be encoded in the tration through excluded volume effects),
primary amino acid sequence of LLPS proteins: electrolyte concentration, pH, and temperature.
intrinsically unfolded protein stretches of low
amino acid diversity, so called low complexity
regions (LCRs), can facilitate the condensation Tau Protein Domains in LLPS
of proteins into liquid dense phases [38]. The
basic principle behind the phase separation is Tau is an axonal neuronal microtubule binding
governed by the preference of the LCRs to inter- protein that is very conserved across mammalian
act with other proteins of its own kind, rather species, and has analogues in C. elegans [53] and
than with the surrounding solvent and biomole- birds [54]. In the human brain, six isoforms of tau
cules; thereby, certain protein conformations, or co-exist [55] with varying abundance depending
complex formation with other molecules like on developmental stage [56, 57] and brain region
RNA, may be necessary to expose the LCD in a [58]; an exceptional large variant of tau (often
protein to enable LLPS.  The molecular interac- called “big tau”) exists in the peripheral nervous
tions established within the liquid dense phase system. The isoforms in the central nervous sys-
are thought to be weak (to ensure reversibility tem are generated by alternative splicing of
and liquid behavior of the phase) and can rely for mRNA from a single gene, MAPT, on chromo-
example on electrostatic or pi-pi, or weak hydro- some 17, and vary in the presence of pseudo-­
phobic interactions [49]. repeats (each ~30aa long) in the N-terminal end
In many RNA binding proteins, LCRs coin- and the C-terminal half of the protein (Fig. 25.1a):
cide with prion-like domains (PrLDs or PDLs) two N-terminal insets (N1+2) encoded by Exon 2
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 345

and Exon 3, and four central pseudo-repeats (R1-­ domains. It appears that different regions of tau
4), of which R2 (Exon 10) can be alternatively can contribute to the process of liquid-liquid
spliced-out to generate the 3-repeat isoforms of phase separation in an individual as well as in a
tau. The longest human tau variant (2N4R) con- combined fashion, making the phase separation
sists of 441 amino acids, large parts of which of tau a complex process.
have no fixed secondary or tertiary structure, In addition, the LLPS conditions and charac-
assigning tau to the group of intrinsically disor- teristics for the different tau domains seem to be
dered proteins (IDPs) [1, 59] (Fig. 25.1a). very different. For example, the condensation of
Functionally, the sequence of tau can be phosphorylated full-length tau is inhibited by ali-
divided into protein domains according to their phatic alcohols like hexanediol but largely resis-
function in MT binding: the pseudo-repeats R1-4 tant to high ionic strength of the buffer, whereas
build the microtubule binding region (MTBR; aa LLPS of the phosphorylated N-terminal half is
244–368); this domain harbors some transient inhibited under such conditions [30]. These data
beta-structure with two hydrophobic hexapep- suggest that hydrophobic interactions – presum-
tides that mediate tau’s aggregation into fibrillar ably by two short beta-strands in repeats R2 and
aggregates [60, 61]. The N-terminal projections R3 of the tau repeat region  – may stabilize the
domain of tau (aa 1–150) and the proline-rich liquid phase of phospho-tau, whereby charge
domain (P1+P2; aa 151–243) are unstructured interactions in the N-terminal half may help to
and protrude from the surface of the microtu- initiate condensation or have an unknown addi-
bules. In fibrillar aggregates assembled from full-­ tional function in the process of
length tau, the N-terminal half builds a polymer LLPS.  Un-phosphorylated tau (from E. coli)
brush around the fibril core [62]. appears generally more sensitive to high salt con-
In most LLPS proteins, low complexity or ditions. In summary, these different LLPS char-
prion-like domains drive their condensation into acteristics indicate a distinct and complementary
liquid phases. Tau does not have a clear LCR that role of the protein domains in tau LLPS behavior
is particularly enriched in polar and charged resi- and function, which need to be further explored.
dues [38, 63, 64]  - low complexity predictors
mostly point out the proline-rich regions
(P1+P2)). The aggregation prone region of tau,  rotein Concentration Critical for Tau
P
which could maybe be compared to PrLDs in LLPS
RNA binding proteins because of its aggregation
propensity, is located in the transiently structured Tau is one of the most abundant proteins in the
MBTR.  In fact, tau is generally very rich in brain and is physiological intracellular concen-
charged residues, with regions of positive and tration was suggested to be ~2 μM [30, 65]. This,
negative net charges in the protein that create however, does not take into account local differ-
multivalence along the primary sequence of tau ences in tau concentration that occur in the neu-
(Fig.  25.1a). Low amino acid complexity (high ron: tau is almost exclusively found in the axon,
proline content) exists only in the proline-rich where its local concentration is likely much
regions. higher, but it has been suggested that at least 50%
In the absence of a LCR or PrLD in the tau of tau molecules are bound to microtubules [66].
sequence, predicting which part of tau can Due to tau’s high solubility, the tau concentra-
undergo LLPS becomes difficult. In fact, differ- tions in cells can reach up to ~100 μM (for exam-
ent tau protein domains have now been shown to ple during recombinant protein expression in
individually be able to condensate into liquid insect cells) while remaining in a free soluble
droplets. As of now, it has been shown that full-­ state. Immunohistological labeling of neurons in
length tau [29, 30], but also the N-terminal half culture and in vitro microtubule binding experi-
(aa1–256; [28, 30]) and the MTBR [27, 28]) can ments with fluorescently labeled recombinant tau
undergo LLPS in absence of the other protein show that there are even “hot spots” of locally
346 S. Wegmann

enriched tau in the axon, which recently have phosphates per tau molecule) from insect cells
been suggested to consist of liquid condensed tau increases the LLPS rate and decreases the critical
phases as well (more information below). tau protein concentration even more [30]. On the
Different independently performed studies other hand, the addition of 17 negative charges in
reported the condensation of tau into liquid a phospho-mimetic of tau, which mimics the
­droplets in vitro at concentrations between charge changes introduced by phosphorylation of
1–25 μM in the presence of crowding agents [29, tau, can also facilitate tau LLPS [30]. This shows
30], which were added to mimic the molecular that changes in the molecular conformation, as
crowding (~100–200 mg/ml proteins) present in well as charge compensation or addition, are
the intracellular environment. Typical crowding favoring the condensation of tau into liquid
agents used in in vitro protein condensation stud- phases. Charge compensation through phosphor-
ies are for example PEG-8000, Ficoll-400, or ylation followed by conformational changes/mis-
dextrans in concentration of 5–15% (w/vol); they folding have previously been implicated in the
increase the apparent protein concentration oligomerization of tau [69]. Liquid tau conden-
through excluded volume effects. In the absence sates may thus be resemble a macromolecular
of such molecular crowders, the critical LLPS state between soluble and aggregated forms of
concentration of tau is about one magnitude tau.
higher (10–100 μM; [28, 30]). It remains unclear, In healthy cells, phosphorylation by multiple
however, what is the critical tau concentration for kinases (e.g. MARK2, GSK3b, Fyn, CDK5,
intracellular tau condensation, which may in fact PKA, and JNK) is the main regulator of tau’s
be highly dependent on the local subcellular functions such as microtubule binding. In fact,
environment (electrolytes, pH, biomolecules, tau in neurons is mostly phosphorylated.
binding partners etc.). In addition, one has to Accordingly, liquid-like tau condensates can also
keep in mind that the concentration of tau (and be found in primary neurons expressing GFP-tau
other molecules) in highly differentiated cell like [30], and several conditions that are known to
neurons is strongly dependent on the neuronal trigger tau phosphorylation may also cause tau
compartment (e.g. soma, dendrites, axon, condensation. For example, several neuronal
nucleus, synapse) (Fig. 25.2a). stress situations associated with neurodegenera-
tive diseases, such as mechanical injury [70],
chronical stress enhancing glucocorticoids and
Phosphorylation of Tau ROS [71], ER stress [72], and the exposure to
Amyloid-beta [73], acutely enhance tau phos-
An important finding is that tau phosphorylation phorylation. Interestingly, tau LLPS – at least in
reduces the critical LLPS concentration of tau by vitro – seems to be inhibited at low temperatures
an order of magnitude. Phosphorylation (and [27, 28], whereas tau phosphorylation has been
other PTMs) may thus be an important regulator shown to increase in hibernating animals during
not only of tau microtubule binding and aggrega- torpor [74]. These data suggest a complex rela-
tion, but also of tau condensation. For example, tionship between tau phosphorylation and con-
minor phosphorylation by the kinase MARK2 at densation, in which functional and dysfunctional
4 sites in the MTBR, which has been shown to LLPS might be differentially regulated. The reg-
catalyzes the detachment of tau from microtu- ulation of tau LLPS by phosphorylation (and
bules [67, 68], is sufficient to induce condensa- other PTMs) could thus be an important switch
tion at tau concentrations of 1–2 μM [28, 30] (Fig. between function and disease.
25.2b). Interestingly, the phosphorylation of tau At the moment, however, in vivo evidence for
by MARK2 was shown to increase the local beta-­ tau condensates in rodent models or human post-­
structure content in the tau repeat domain; these mortem tissue is still very sparse, mainly because
data support the idea that beta-strand interactions of technical difficulties in characterizing the liq-
may enhance tau condensation. Higher levels of uid state in living animals or fixed tissue. As a
phosphorylation at multiple sites in tau (~12 first approach to this challenge, we recently were
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 347

Fig. 25.2  Conditions enhancing tau liquid-liquid LLPS at a physiological concentration of 2 μM [30]. (c)
phase transition. (a) Schematic phase diagram of tau FTD mutations, which are known to increase the aggrega-
LLPS: the phase separation of tau, which leads to the tion and oligomerization propensity of tau, significantly
formation of liquid “droplets”, depends on the concen- decrease the critical concentration necessary for tau LLPS
tration of tau protein and other soluents (e.g. salts, sol- to physiological concentrations (~2 μM). The small white
vents, biomolecules, crowding agent) in the solution. rectangles represent two hexapeptide motifs that mediate
Phosphorylation, FTD mutations, RNA, and molecular the aggregation of tau; their beta-structure propensity and
crowding lower the critical tau LLPS concentration [c0] hydrophobicity is enhanced by FTD mutations in this
and enhance tau condensation into liquid droplets (pink region [60]. (d) Phase transition of tau in vitro, from
line). (b) Phosphorylation sites are distributed along the monomeric to liquid condensates to aggregates, can
sequence of tau but enriched in the proline-rich region and occur spontaneously or can be initiated at ~50-fold lower,
in the C-terminus. In contrast to most kinases (e.g. GSK- physiological tau concentrations by using polymeric
3beta, CDK5, Fyn), kinase MARK2, which catalyzes the crowding agents. Over time, liquid droplets of phospho-
detachment of tau from microtubules [68], phosphorylates tau form small aggregates on their surface that can be
tau at four sites in the repeat region (red sites). This minor labeled with Thioflavine-S, suggesting beta-sheet content
phosphorylation by MARK2 is sufficient to enable tau of the aggregates
348 S. Wegmann

able to detect first signs of droplet-shaped tau molecular crowding agents. RNA and heparin
accumulations in the mouse cortex by in vivo are commonly used as polyanionic cofactors to
two-photon imaging, and could initiate induce the aggregation of recombinant tau, con-
­condensation of tau isolated from AD brains in necting the concepts of tau liquid phase separa-
vitro [30]. tion with tau aggregation. Notably, the binding
of RNA to tau and the RNA-induced tau LLPS
may have functions in the regulation of RNA
Frontotemporal Dementia Mutations transcription that are not fully understood yet
(see Chap. 26 by Benjamin Wolozin and Chap.
In Alzheimer’s disease, tau aggregation is associ- 24 by Kenneth Kosik in this book for more
ated with hyperphosphorylation and likely occurs details).
due to changes in the intramolecular interactions
and protein conformation caused by aberrant
post-translational modification. In contrast, in  hase Transition from Tau Liquid
P
frontotemporal dementias (FTDs), mutations in Condensates into Aggregates
the tau gene (MAPT) drive the aggregation of tau,
and several point mutations have been identified Several RNA-binding proteins that transiently
in patients that trigger the aggregation in tau in condensate into liquid phases under certain con-
vitro [60, 75]. The molecular mechanism for ditions have been suggested to also be involved in
some of these mutations has been identified: the pathobiology of motor neurons in ALS. In the
Aggregation-prone mutations in the tau repeat presence of disease-associated mutations or aber-
domain – such as P301L, P301S, and DeltaK280 – rant PTMs, liquid-like protein condensates of
seem to increase the beta-structure propensity of FUS [36], TDP-43 [39] and hnRNP [38] can tran-
a hexapeptide motif in R2 (PHF6*), which sition into aggregates. Such phase transition from
increases the local hydrophobicity and favors the a liquid into an aggregated state has also been
assembly of tau into amyloid-like fibrils [76]. described for tau [28, 30].
Interestingly, the same mutations also enhance
the condensation of tau into liquid phases. More
precisely, they substantially lower the critical  aturation of Condensed Liquid Tau
M
concentration of non-phosphorylated tau to Phases
undergo LLPS in vitro (down to <2  μM; [30]).
Another FTD mutation located in the N-terminal The spontaneous formation of liquid-like con-
projection domain of human tau, A152T, which densates of recombinant tau in the presence of
has been shown to enhance oligomerization but crowding agents is an instant process. “Young”
not aggregation of tau in vitro [77] and cause droplets (~0–15 min old) show several liquid-like
functional deficits in vivo [78–80], also facilitates characteristics, for example they readily fuse into
tau LLPS at low concentrations. These findings larger ones and show surface wetting behavior.
provide further intriguing evidence that tau LLPS However, already after ~15  min, the fusion of
may be “on-path” with tau oligomerization; tau individual tau condensates seems to be hindered,
oligomers are currently considered most respon- and after ~1 h, the condensates start to lose their
sible for tau induced neurotoxicity (Fig. 25.2c). perfect spherical symmetry and spots of deformed
solid-like material start to appear on the surface
of the condensates [30].
RNA and Polyanions Fluorescence recovery after photo bleaching
(FRAP) shows that this “maturation” of liquid
A class of very potent inducers of tau LLPS are tau condensates is accompanied by a rapid (in
polyanionic molecules: tau condensates form minutes) decrease in the exchange of tau mole-
spontaneously in the presence of RNA [27, 30] cules between the droplet phase and the sur-
and heparin [28, 30], without the need of rounding buffer, as well as with a decrease of tau
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 349

diffusion rate in the droplets [30]. In fact, after a  otential Roles for Tau LLPS
P
few hours, the condensed phase of tau can be in Neurodegenerative Diseases
enriched by centrifugation. Stiffness and visco-
elasticity of the tau droplets during maturation I gnition of Tau Aggregation
could for example be quantified using atomic by Liquid-Liquid Phase Separation
force microscopy and spectroscopy; from such
first measurements (unpublished data) it seems Tau aggregation is one of the main pathological
that, with time, the liquid state of tau condensates changes in AD and tauopathy brains. Until now,
autonomously evolves into a viscoelastic the role of tau liquid dense phases for tau neuro-
hydrogel-­ like state, similar to what has been pathological changes observed in AD and tauop-
described for the polymerization process of gels. athies is uncertain. However, there is strong
Investigating the dynamic molecular rearrange- evidence that tau LLPS – at least in vitro – can be
ments of tau condensates during maturation is on-path with tau aggregation (Fig.  25.1d), sug-
one of the current aims in the field and will be gesting a possible connection of these two pro-
needed to understand the molecular mechanisms cesses. In fact, multiple observations related to
of phase evolution from liquid-like tau phases tau in neurodegenerative diseases could be
into aggregates. explained by dysfunctional phase separation in
these diseases. For example, intracellular and/or
physiological differences between brain areas
Emerging of Tau Aggregates and neuronal subpopulations could lead to differ-
from Condensates ential tau condensation, and thereby explain the
observation that tau aggregation occurs in dis-
The condensation into liquid droplets leads to a tinct brain areas in AD. In other tauopathies with
high local concentration of tau in the conden- a less “organized” tau aggregation pathology,
sates. Calibrated fluorescence imaging of fluores- such as FTD, the condensation of tau, which is
cently labeled recombinant phospho-tau in the enhanced by tau mutations that lower the thresh-
presence of crowding agents and the measure- old concentration for LLPS [30], could occurs
ment of tau protein concentration in the con- more spontaneously and independently of the
densed phase (after centrifugation) suggest a tau neuronal population. This could explain the dif-
concentration of >30  μM in the liquid dense ferent pattern in tau aggregation between AD and
phases [30]. This concentration is similar to con- other tauopathies.
centrations commonly used in in vitro tau aggre- On the cellular level, intraneuronal tau con-
gation assays (20–50 μM). It is not surprising that densation may, for example, occur as a conse-
protein aggregates emerge from the condensates quence of tau mislocalization into the
already 1 h after their formation, and that more somato-dendritic compartment, where it encoun-
and more spherical condensates transition into ters a different local environment that may favor
amorphous aggregates over time. Interestingly, its LLPS.  Another possibility is that aberrant
although these tau aggregates do not show the post-translational modifications of tau (hyper-
well-described ordered structure of tau paired-­ phosphorylation, acetylation, ubiquitination,
helical filaments (PHFs), they harbor the poten- SUMO-ylation, truncation etc.) increases its ten-
tial to seed tau aggregation in cells [30]. The dency for tau LLPS in the axon.
formation of seeding-competent tau aggregates Tau has also been shown to be associated with
from tau condensates implicates an intriguing neuronal stress granules, which are reversibly
connection between tau phase separation and assembled, membrane-less organelles formed by
pathological aggregation. However, at the LLPS of RNA-binding proteins and RNA mole-
moment it remains unclear to what extend tau cules under stress conditions in neurons [81]. A
LLPS plays a role in neurodegenerative diseases co-partitioning of tau in these granules has been
with pathological tau aggregation (Fig. 25.2d). suggested to be involved in tau-mediated toxicity
350 S. Wegmann

in transgenic mice [82], however, independent of folded tau [94–97], and for soluble oligomers of
tau aggregation. tau [98, 99]. An unexplored possibility is that
matured liquid- or gel-like tau condensates,
which seem to be stable outside of cells at least in
Tau Protein Propagation vitro [30] and in in neuronal cultures, could be
Through Condensation in the Brain? released and internalized by neurons.
For the uptake of tau, multiple mechanisms
The progressive deposition of aggregated tau in have been suggested to play a role depending on
the brain is a hallmark of tauopathies and AD. In the state of tau: whereas neurons in vitro internal-
sporadic AD (sAD), the tau neurofibrillary tan- ize monomeric soluble tau by Clathrin-mediated
gles (NFTs) pathology follows a very stereotypi- endocytosis [100], the uptake of pre-aggregated
cal pattern: NFTs start to form in the locus tau seems to involve heparan sulfates on the cell
coeruleus and the entorhinal cortex, then appear surface and to be facilitated by pinocytosis [101]
in the hippocampal formation from where they or bulk endocytosis [99, 100] of tau. Largely
“spread” through limbic areas to the forebrain unexplored is the process by which the uptake of
and other cortical areas [83, 84]. Neuronal loss tau (not associated with synaptic vesicles) could
and cognitive decline in patients with AD corre- happen at the pre-synapse.
lates with the progression of NFT pathology. One In context of tau liquid condensates [30] and/
main working hypothesis for the mechanism of or tau associated with stress granules [81], one
tau progression in sAD is based on the trans-­ can imagine different mechanisms for the propa-
synaptic propagation of tau proteins between gation of tau aggregation pathology in the brain,
neurons [85]: misfolded tau travels from a neuron which do not rely on the physical travel of tau
to downstream connected neurons, where it proteins between neurons. One could speculate
“seeds” the misfolding and aggregation of tau, that the progressive accumulation and aggrega-
which then leads to NFT formation in the recipi- tion of tau may be a consequence of neuron-­
ent neuron, and continues to propagate further to intrinsic changes manifesting in certain neurons
other synaptically connected neurons; this during the disease and favoring an aberrant con-
hypothesis nicely explains the occurrence of tau densation and subsequent aggregation of tau in
aggregation in sAD in brain regions that are con- these neurons. However, the intrinsic neuronal
nected by major neuronal projections. However, changes involved in aberrant LLPS or stress
the cellular mechanism of tau protein propaga- granule association in this model are speculative
tion between neurons is still unclear, and differ- as well, and the signal that causes the changes
ent possibilities are discussed. Tau could may as well be tau protein that was received by
propagate either by general release and uptake in neuron-to-neuron protein propagation.
by neighboring cells, or across synapses by pre-­
synaptic release followed by uptake at the post-­
synapse. The release at the synapse has been Condensed Tau and Microtubules
shown for tau associated with synaptic vesicles
[86], tau associated with exosomes [87, 88] or The most prominent and best studies role of tau is
ectosomes [89], or by unusual protein secretion the stabilization of microtubules in the axon. In
depending on lipids in the synaptic membrane fact, MAPT was discovered during co-­purification
[90]. Furthermore, tau propagation appears to be of tubulin from the brain, as a co-factor enhancing
enhanced by the presence of amyloid-beta [91] tubulin polymerization into microtubules [102].
and by neuronal activity in mice in vivo [92]. The In the context of tau protein liquid phase separa-
species of tau released from neurons are dis- tion, it is obviously important to examine if tau
cussed as well, and propagation has been reported condensation processes are involved in the bind-
for non-misfolded full-length tau [93], for mis- ing, bundling, and nucleation of microtubules.
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 351

 ubulin Co-condensation with Tau

T not other motor proteins along microtubules, sug-
Nucleates Microtubules gesting a mechanism for how tau may regulate
the microtubule transport of cargo.
Tau phosphorylation (by certain kinases at certain
sites) regulates the microtubule binding of tau, is
relevant for the regulation of microtubule dynam-  ethods Used to Study Tau Protein
M
ics [103], and is also discussed for the regulation Condensation
of transport along on microtubules [104].
Accordingly, disease associated hyperphosphory- To characterize the transition of tau between its
lation [105] or mutations [75] decreases the different condensation states  – soluble mono-
microtubule binding of tau, thereby causing trans- meric, liquid dense condensates, gel-like dense
port deficits and neurotoxicity. Tau can also nucle- condensates, small soluble oligomers, and larger
ate microtubules in vitro, and only recently it was aggregates  – one is confronted with various
shown that the nucleation involves liquid phases molecular sizes, interactions and dynamics. For
of tau [29]. Non-polymerized tubulin co-parti- example, highly soluble intrinsically unfolded
tions into preformed tau liquid droplets, where it tau monomers have minor amounts of transient
reaches the critical concentration for microtubule structure and a large amount of interaction part-
nucleation, and minutes after the addition of GTP, ners, whereas tau PHFs are larger by at least 2–3
microtubules start growing out bidirectionally magnitudes, have a highly ordered rigid amyloid-­
from of the tau condensates, whereby the growing like structure (in their fibrils core), and presum-
microtubules remain “coated” by a liquid phase of ably a limited amount of interaction partners. Tau
tau. This mechanism of a priori nucleation of liquid condensates can, to some extent, be seen as
microtubules could be important for local non- an intermediate between these two states: tau
centrosomal microtubule growth and organization molecules remain somewhat mobile but with a
in the axon at sites of need, for example after reduced diffusion coefficient compared to free
injury, or during axon growth and branching. monomeric tau, the conformational sampling of
Furthermore, the liquid phase of tau seemed to tau in the condensed phase is more restricted than
reversibly bundle microtubules, which could be in its monomeric freely moving form, and tau
an important mechanism for microtubule stability molecules establish many transient interactions
in cells. Liquid-like Tau phases on microtubules. with themselves and with other constituents of
the liquid phase.
To describe the characteristics of all the states
Liquid Tau Phases on Microtubules that tau can adopt  – monomers, liquids, oligo-
mers, aggregates -, many different protein bio-
A more close view on the association of liquid chemical and biophysical methods have to be
tau phases with preformed microtubules in vitro employed. The following techniques were used to
showed that actually very little tau (~0.5 nM) is describe the evolution of tau condensates, their
needed to form small local tau condensates on liquid state, and their maturation into gel-like
microtubules [106]. A similar inhomogeneous phases and aggregates: Light microscopy can be
“coating” of microtubules by tau can also be used to identify droplet-like tau condensates in
observe in the brain of mice [107]. Interestingly, solution and to describe their fusion/fission and
the local dense condensates seem to form prefer- size distribution; it is limited though to droplets
entially at areas of high microtubule curvature, with diameters larger than the diffraction limit
and one can imagine a mechanism for repair or [29, 30]. Optical tweezers can be used to measure
stabilization of microtubules after forced bending the forces occurring during tau droplet fusion
or during axon branching. Furthermore, in the [29]. Bulk measurements of dynamic light scat-
same study, the tau condensates appeared to tering and turbidity can detect the formation of
selectively decrease the migration of some but tau condensates in a solution [27, 28]. Nuclear
352 S. Wegmann

magnetic resonance and circular dichroism mea- some of the long-standing paradigms in the con-
surements were used to determine changes in the text of LLPS.  For example, enhanced LLPS in
interactions and molecular structure of tau in the neuronal stress situations (elevated phosphoryla-
free soluble versus the condensed phase [28]. tion) or over-stabilized droplets of FTD-mutant
Electron microscopy and AFM could produce tau could result in tau aggregation and thereby
images of tau liquid condensates and aggregates, explain the enigmatic transition from highly solu-
and in vitro aggregation assays could proof the ble to ordered aggregated states of tau in the brain
formation of tau aggregates [27, 28, 30] from (Fig. 25.3). In another example, the aberrant regu-
recombinant tau and in cells [30]. In future, the lation of tau LLPS on microtubules could lead to
use of fluorescence life-time imaging and novel failures in motor protein processivity and, hence,
time-resolved light-scattering and x-ray tech- mitochondria or other cargo transport observed in
niques promises to give more insides into the neurodegenerative diseases. Additionally, altered
dynamic structural changes during the maturation co-phase separation of other (yet unknown) bio-
of tau condensates into aggregates, both in vitro molecules into tau liquid phases may result either
and in cells. For the examination of tau condensa- in local microtubule destabilization, or in too high
tion processes and their consequences in the brain, (or low) local concentrations of certain essential
two-photon imaging in combination with calcium molecules; this, in turn, can deregulate molecular
imaging and electrophysiology in different ani- pathways by locally changing the reaction equi-
mal models, for example in C. elegans but also in librium in the axoplasm. Furthermore, tau LLPS
rodent models, appear to be promising approaches characteristics likely depend on local distinct con-
to study the role of tau LLPS in vivo. ditions in the cytoplasm, and thus may differ in
the different cellular compartments (soma, axon,
dendrites, synapses). While tau LLPS may be
Conclusion and Outlook needed for functional microtubule bundling and
nucleation in the axon, the condensation of tau in
The wide-spread biological relevance of protein the soma could result in aggregation or other mis-
condensation into dense phases with liquid- or functions of tau. For example, we recently discov-
gel-like biophysical properties has recently been ered that phosphorylated tau, when in the soma
discovered. In the case of tau, the roles of LLPS and perinuclear space as in AD neurons, can inter-
(and the resulting liquid dense condensates) for act with nucleoporins in the pore of the nuclear
“normal” and pathobiological tau functions are pore complex  – another physiological liquid/gel
still quite unknown. However, the concept of tau phase in the cell – and that this interaction causes
liquid phases opens up another level of complex- an impairment of nucleocytoplasmic protein
ity in tau biology, in which interactions, reaction transport [14]; this interaction may involve co-­
kinetics, and molecular behavior differ from pro- condensation of tau with FG-nucleoporins.
cesses involving mono-molecular tau. All the dif- To conclude, it is now on us and others to
ferent physiochemical states in that tau can exist experimentally test the hypotheses around func-
in  – monomers/dimers, small oligomers, larger tional and dysfunctional tau LLPS, and to evalu-
aggregates, and now liquid phases – enable tau to ate the role of tau condensation in the healthy and
establish different sets of interactions at individ- the diseased brain.
ual time-scales, depending on their molecular size
and their intrinsic conformational sampling rate. Acknowledgements  A deep and grateful thanks goes to
This may enable us to explain observations that Daniel J Mueller, Eckhard & Eva Mandelkow, and
Bradley T Hyman for their outstanding and passionate
are not yet understood. The liquid phase of tau mentoring through the last years, and for their uplifting
offers us additional possibilities to explain the support that shaped my scientific and personal
mis-regulations of tau in diseases like AD and ­development. I also cordially thank Gabriela Ruano for
tauopathies, and it should encourage us to rethink her unlimited support on my side.
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 353

Fig. 25.3  Possible roles of neuronal tau condensation. the type of RNA [27]; tau-RNA condensates may, for
Tau liquid-liquid phase separation could have multiple example, be involved in RNA transport, transcription, and
roles for the neuronal cell biology. There is evidence that, stability. More generally, co-condensation or partitioning
in the pathological setting of AD and tauopathies (i), tau of tau into other liquid-like condensates could lay the
phase separation is enhanced by hyperphosphorylation or foundation for “unusual” interactions of tau, such as the
FTD mutations and can initiate tau aggregation [30]. In interaction of phospho-tau with the hydrogel formed by
the healthy neuron, tau liquid condensates can form on FG-domain nucleoporins in the pore of the nuclear pore
microtubules and facilitate efficient bundling and nucle- complexes [14]. Lastly, tau condensates could also be
ation at sites needed (ii) [29]. Partitioning of tau into involved in the trans-neuronal propagation of tau between
stress granules, which assemble through LLPS of RNA-­ neurons in the brain. In this context, the travel of seeding-­
binding proteins with RNA, could change the stress gran- competent tau condensates could, for example, play a role
ule function and/or trigger tau aggregation [81]. The for the spread of pathological tau aggregation in neurode-
efficiency of coacervation of tau with RNA depends on generative diseases such as AD

and generates a pathological (MC-1) conformation.

References J Biol Chem. 2008;283(46):32066–76.
3. Uversky VN. Intrinsic disorder-based protein inter-
1. Mukrasch MD, Bibow S, Korukottu J, Jeganathan actions and their modulators. Curr Pharm Des.
S, Biernat J, Griesinger C, et al. Structural polymor- 2013;19(23):4191–213.
phism of 441-residue Tau at single residue resolu- 4. Gustke N, Trinczek B, Biernat J, Mandelkow
tion. PLoS Biol. 2009;7(2):399–414. EM, Mandelkow E.  Domains of tau protein and
2. Jeganathan S, Hascher A, Chinnathambi S, Biernat interactions with microtubules. Biochemistry.
J, Mandelkow EM, Mandelkow E. Proline-directed 1994;33(32):9511–22.
pseudo-phosphorylation at AT8 and PHF1 epitopes 5. Janning D, Igaev M, Sundermann F, Bruhmann J,
induces a compaction of the paperclip folding of tau Beutel O, Heinisch JJ, et al. Single-molecule track-
ing of tau reveals fast kiss-and-hop interaction with
354 S. Wegmann

microtubules in living neurons. Mol Biol Cell. polymorphic structure and stability. J  Biol Chem.
2014;25(22):3541–51. 2010;285(35):27302–13.
6. Kadavath H, Hofele RV, Biernat J, Kumar S, Tepper 20. Fitzpatrick AWP, Falcon B, He S, Murzin AG,
K, Urlaub H, et  al. Tau stabilizes microtubules by Murshudov G, Garringer HJ, et al. Cryo-EM struc-
binding at the interface between tubulin heterodi- tures of tau filaments from Alzheimer’s disease.
mers. Proc Natl Acad Sci. 2015;112(24):7501–6. Nature. 2017;547(7662):185–90.
7. Elie A, Prezel E, Guérin C, Denarier E, Ramirez-Rios 21. Falcon B, Zhang W, Schweighauser M, Murzin AG,
S, Serre L, et al. Tau co-organizes dynamic microtu- Vidal R, Garringer HJ, et  al. Tau filaments from
bule and actin networks. Sci Rep. 2015;5:9964. multiple cases of sporadic and inherited Alzheimer’s
8. Cabrales Fontela Y, Kadavath H, Biernat J, Riedel disease adopt a common fold. Acta Neuropathol.
D, Mandelkow E, Zweckstetter M.  Multivalent 2018;136(5):699–708.
cross-linking of actin filaments and microtubules 22. Lasagna-Reeves CA, Castillo-Carranza DL,
through the microtubule-associated protein Tau. Nat Sengupta U, Sarmiento J, Troncoso J, Jackson GR,
Commun. 2017;8(1):1981. et  al. Identification of oligomers at early stages of
9. Elbaum-Garfinkle S, Ramlall T, Rhoades E.  The tau aggregation in Alzheimer’s disease. FASEB
role of the lipid bilayer in tau aggregation. Biophys J. 2012;26(5):1946–59.
J. 2010;98(11):2722–30. 23. Ward SM, Himmelstein DS, Lancia JK, Binder
10. Mari SA, Wegmann S, Tepper K, Hyman BT, LI. Tau oligomers and tau toxicity in neurodegenera-
Mandelkow EM, Mandelkow E, et  al. Reversible tive disease. Biochem Soc Trans. 2012;40(4):667–71.
cation-selective attachment and self-assembly of 24. Kaniyappan S, Chandupatla RR, Mandelkow
human tau on supported brain lipid membranes. EM, Mandelkow E.  Extracellular low-n oligo-
Nano Lett. 2018;18(5):3271–81. mers of tau cause selective synaptotoxicity with-
11. Brandt R, Léger J, Lee G.  Interaction of tau with out affecting cell viability. Alzheimers Dement.
the neural plasma membrane mediated by tau’s 2017;13(11):1270–91.
amino-terminal projection domain. J  Cell Biol. 25. Lasagna-Reeves CA, Castillo-Carranza DL,
1995;131(5):1327–40. Sengupta U, Clos AL, Jackson GR, Kayed R.  Tau
12. Usardi A, Pooler AM, Seereeram A, Reynolds oligomers impair memory and induce synaptic and
CH, Derkinderen P, Anderton B, et  al. Tyrosine mitochondrial dysfunction in wild-type mice. Mol
phosphorylation of tau regulates its interactions Neurodegener. 2011;6:39.
with Fyn SH2 domains, but not SH3 domains, 26. Tepper K, Biernat J, Kumar S, Wegmann S, Timm
altering the cellular localization of tau. FEBS T, Hübschmann S, et al. Oligomer formation of tau
J. 2011;278(16):2927–37. protein hyperphosphorylated in cells. J Biol Chem.
13. Reynolds CH, Garwood CJ, Wray S, Price C, 2014;289(49):34389–407.
Kellie S, Perera T, et  al. Phosphorylation regulates 27. Zhang X, Lin Y, Eschmann NA, Zhou H, Rauch JN,
tau interactions with Src homology 3 domains Hernandez I, et al. RNA stores tau reversibly in com-
of phosphatidylinositol 3-kinase, phospholipase plex coacervates. PLoS Biol. 2017;15(7):e2002183.
Cγ1, Grb2, and Src family kinases. J  Biol Chem. 28. Ambadipudi S, Biernat J, Riedel D, Mandelkow E,
2008;283(26):18177–86. Zweckstetter M.  Liquid-liquid phase separation of
14. Eftekharzadeh B, Daigle JG, Kapinos LE, Dujardin the microtubule-binding repeats of the Alzheimer-­
S, Coyne A, Miller SJ, et  al. Tau protein disrupts related protein Tau. Nat Commun. 2017;8(1):275.
nucleocytoplasmic transport in Alzheimer’s disease. 29. Hernández-Vega A, Braun M, Scharrel L, Jahnel
Neuron. 2018;99(5):925–940.e7. M, Wegmann S, Hyman BT, et  al. Local nucle-
15. Kovacs GG, Botond G, Budka H. Protein coding of ation of microtubule bundles through tubulin con-
neurodegenerative dementias: the neuropathological centration into a condensed tau phase. Cell Rep.
basis of biomarker diagnostics. Acta Neuropathol. 2017;20(10):2304–12.
2010;119(4):389–408. 30. Wegmann S, Eftekharzadeh B, Tepper K, Zoltowska
16. McKee AC, Stein TD, Nowinski CJ, Stern RA, KM, Bennett RE, Dujardin S, et al. Tau protein liq-
Daneshvar DH, Alvarez VE, et al. The spectrum of uid–liquid phase separation can initiate tau aggrega-
disease in chronic traumatic encephalopathy. Brain. tion. EMBO J. 2018;37(7):pii: e98049.
2013;136(Pt 1):43–64. 31. Brangwynne CP, Eckmann CR, Courson DS,
17. Kovacs GG, Budka H. The spectrum of tau pathol- Rybarska A, Hoege C, Gharakhani J, et al. Germline
ogy in human prion disease. In: Zou WQ, Gambetti P granules are liquid droplets that localize by con-
P, editors. Prions and diseases. 1st ed. New  York: trolled dissolution/condensation. 2009;Science ­(80-),
Springer; 2012. 324(5935):1729–32. Available from: http://www.
18. Huseby CJ, Kuret J.  Analyzing Tau aggregation sciencemag.org/cgi/doi/10.1126/science.1172046
with electron microscopy. Methods Mol Biol. 32. Nott TJ, Petsalaki E, Farber P, Jervis D, Fussner
2015;1345:101–12. E, Plochowietz A, et  al. Phase transition of a dis-
19. Wegmann S, Yu JJ, Chinnathambi S, Mandelkow ordered nuage protein generates environmentally
EM, Mandelkow E, Muller DJ.  Human tau iso- responsive membraneless organelles. Mol Cell.
forms assemble into ribbon-like fibrils that display 2015;57(5):936–47.
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 355

33. Hyman AA, Weber CA, Jülicher F.  Liquid-liquid 47. Larson AG, Elnatan D, Keenen MM, Trnka
phase separation in biology. Annu Rev Cell MJ, Johnston JB, Burlingame AL, et  al. Liquid
Dev Biol. 2014;30(1):39–58. Available from: droplet formation by HP1α suggests a role for
http://www.annualreviews.org/doi/abs/10.1146/ phase separation in heterochromatin. Nature.
annurev-cellbio-100913-013325 2017;547(7662):236–40.
34. Protter DSW, Parker R. Principles and properties of 48. Jiang H, Wang S, Huang Y, He X, Cui H, Zhu X,
stress granules. Trends Cell Biol. 2016;26(9):668–79. et  al. Phase transition of spindle-associated pro-
35. Wheeler JR, Matheny T, Jain S, Abrisch R, Parker tein regulate spindle apparatus assembly. Cell.
R. Distinct stages in stress granule assembly and dis- 2015;163(1):108–22.
assembly. Elife. 2016;5:pii:e18413. 49. Brangwynne CP, Tompa P, Pappu RV.  Polymer
36. Patel A, Lee HO, Jawerth L, Maharana S, Jahnel M, physics of intracellular phase transitions. Nat Phys.
Hein MY, et al. A liquid-to-solid phase transition of 2015;11(11):899–904.
the ALS protein FUS accelerated by disease muta- 50. Wang J, Choi JM, Holehouse AS, Lee HO, Zhang
tion. Cell. 2015;162(5):1066–77. X, Jahnel M, et al. A molecular grammar governing
37. Hofweber M, Hutten S, Bourgeois B, Spreitzer the driving forces for phase separation of prion-like
E, Niedner-Boblenz A, Schifferer M, et  al. Phase RNA binding proteins. Cell. 2018;174(3):688–699.
separation of FUS is suppressed by its nuclear e16.
import receptor and arginine methylation. Cell. 51. Murakami T, Qamar S, Lin JQ, Schierle GSK, Rees
2018;173(3):706–719.e13. E, Miyashita A, et  al. ALS/FTD mutation-induced
38. Molliex A, Temirov J, Lee J, Coughlin M, Kanagaraj phase transition of fus liquid droplets and reversible
AP, Kim HJ, et  al. Phase separation by low com- hydrogels into irreversible hydrogels impairs RNP
plexity domains promotes stress granule assem- granule function. Neuron. 2015;88(4):678–90.
bly and drives pathological fibrillization. Cell. 52. Monahan Z, Ryan VH, Janke AM, Burke KA,
2015;163(1):123–33. Rhoads SN, Zerze GH, et  al. Phosphorylation of
39. Conicella AE, Zerze GH, Mittal J, Fawzi NL. ALS the FUS low-complexity domain disrupts phase
mutations disrupt phase separation mediated by separation, aggregation, and toxicity. EMBO
α-helical structure in the TDP-43 low-complexity J.  2017;36(20):2951–67. Available from: http://
C-terminal domain. Structure. 2016;24(9):1537–49. emboj.embopress.org/lookup/doi/10.15252/
40. Lee KH, Zhang P, Kim HJ, Mitrea DM, Sarkar embj.201696394
M, Freibaum BD, et  al. C9orf72 dipeptide repeats 53. Chew YL, Fan X, Gotz J, Nicholas HR. PTL-1 regu-
impair the assembly, dynamics, and function of lates neuronal integrity and lifespan in C. elegans.
membrane-less organelles. Cell. 2016;167(3):774– J Cell Sci. 2013;126(Pt 9):2079–91.
788.e17. Available from: http://linkinghub.elsevier. 54. Farah G, Siwek D, Cummings P.  Tau accumula-
com/retrieve/pii/S0092867416313836 tions in the brains of woodpeckers. PLoS One.
41. Boeynaems S, Bogaert E, Kovacs D, Konijnenberg 2018;13(2):e0191526.
A, Timmerman E, Volkov A, et al. Phase separation 55. Goedert M, Spillantini MG, Jakes R, Rutherford
of C9orf72 dipeptide repeats perturbs stress granule D, Crowther RA.  Multiple isoforms of human
dynamics. Mol Cell. 2017;65(6):1044–1055.e5. microtubule-­associated protein tau: sequences and
42. Zeng M, Shang Y, Araki Y, Guo T, Huganir RL, localization in neurofibrillary tangles of Alzheimer’s
Zhang M. Phase transition in postsynaptic densities disease. Neuron. 1989;3(4):519–26.
underlies formation of synaptic complexes and syn- 56. Takuma H, Arawaka S, Mori H. Isoforms changes of
aptic plasticity. Cell. 2016;166(5):1163–1175.e12. tau protein during development in various species.
43. Milovanovic D, Wu Y, Bian X, De Camilli P. A liq- Dev Brain Res. 2003;142(2):121–7.
uid phase of synapsin and lipid vesicles. Science 57. Kosik KS, Orecchio LD, Bakalis S, Neve
(80-). 2018;361(6402):604–7. RL.  Developmentally regulated expression of spe-
44. Bouchard JJ, Otero JH, Scott DC, Szulc E, Martin cific tau sequences. Neuron. 1989;2(4):1389–97.
EW, Sabri N, et  al. Cancer mutations of the 58. Majounie E, Cross W, Newsway V, Dillman A,
tumor suppressor SPOP disrupt the formation of Vandrovcova J, Morris CM, et  al. Tau expression
active, phase-separated compartments. Mol Cell. varies in different brain regions and disease state.
2018;72(1):19–36.e8. Neurobiol Aging. 2013;34(7):1922.e7–1922.e12.
45. Berry J, Weber SC, Vaidya N, Haataja M, 59. Jeganathan S, Von Bergen M, Mandelkow EM,
Brangwynne CP. RNA transcription modulates phase Mandelkow E.  The natively unfolded character of
transition-driven nuclear body assembly. Proc Natl Tau and its aggregation to Alzheimer-like paired heli-
Acad Sci U S A. 2015;112(38):E5237–45. Available cal filaments. Biochemistry. 2008;47(40):10526–39.
from: http://www.pnas.org/content/112/38/E5237. 60. von Bergen M, Friedhoff P, Biernat J, Heberle
full J, Mandelkow EM, Mandelkow E.  Assembly of
46. Hennig S, Kong G, Mannen T, Sadowska A, Kobelke tau protein into Alzheimer paired helical fila-
S, Blythe A, et al. Prion-like domains in RNA bind- ments depends on a local sequence motif ((306)
ing proteins are essential for building subnuclear VQIVYK(311)) forming beta structure. Proc Natl
paraspeckles. J Cell Biol. 2015;210(4):529–39. Acad Sci U S A. 2000;97(10):5129–34.
356 S. Wegmann

61. Von Bergen M, Barghorn S, Biernat J, Mandelkow process associated with neuronal plasticity in hiber-
EM, Mandelkow E. Tau aggregation is driven by a nating animals. J Neurosci. 2003;23(18):6972–81.
transition from random coil to beta sheet structure. 75. Fischer D, Mukrasch MD, Von Bergen M, Klos-­
Biochim Biophys Acta. 2005;1739(2–3):158–66. Witkowska A, Biemat J, Griesinger C, et  al.
62. Wegmann S, Medalsy ID, Mandelkow E, Müller Structural and microtubule binding properties of tau
DJ. The fuzzy coat of pathological human Tau fibrils mutants of frontotemporal dementias. Biochemistry.
is a two-layered polyelectrolyte brush. Proc Natl 2007;46(10):2574–82.
Acad Sci U S A. 2013;110(4):E313–21. Available 76. Von Bergen M, Barghorn S, Li L, Marx A, Biernat
from: http://www.pubmedcentral.nih.gov/articleren- J, Mandelkow EM, et  al. Mutations of tau protein
der.fcgi?artid=3557036&tool=pmcentrez&renderty in frontotemporal dementia promote aggregation
pe=abstract of paired helical filaments by enhancing local beta-­
63. Martin EW, Mittag T. Relationship of sequence and structure. J Biol Chem. 2001;276(51):48165–74.
phase separation in protein low-complexity regions. 77. Coppola G, Chinnathambi S, Lee JJY, Dombroski
Biochemistry. 2018;57(17):2478–87. BA, Baker MC, Soto-ortolaza AI, et al. Evidence for
64. Kato M, Han TW, Xie S, Shi K, Du X, Wu LC, et al. a role of the rare p.A152T variant in mapt in increas-
Cell-free formation of RNA granules: low complex- ing the risk for FTD-spectrum and Alzheimer’s dis-
ity sequence domains form dynamic fibers within eases. Hum Mol Genet. 2012;21(15):3500–12.
hydrogels. Cell. 2012;149(4):753–67. 78. Pir GJ, Choudhary B, Mandelkow E, Mandelkow
65. Drubin DG, Feinstein SC, Shooter EM, Kirschner E-M. Tau mutant A152T, a risk factor for FTD/PSP,
MW. Nerve growth factor-induced neurite outgrowth induces neuronal dysfunction and reduced lifes-
in PC12 cells involves the coordinate induction of pan independently of aggregation in a C. elegans
microtubule assembly and assembly-promoting fac- Tauopathy model. Mol Neurodegener. 2016;11:33.
tors. J Cell Biol. 1985;101(5):1799–807. Available from: http://www.ncbi.nlm.nih.gov/
66. Ackmann M, Wiech H, Mandelkow E. Nonsaturable pubmed/27118310%5Cn. http://www.pubmedcen-
binding indicates clustering of Tau on the microtu- tral.nih.gov/articlerender.fcgi?artid=PMC4847334
bule surface in a paired helical filament-like confor- 79. Maeda S, Djukic B, Taneja P, Yu G, Lo I, Davis A,
mation. J Biol Chem. 2000;275(39):30335–43. et  al. Expression of A152T human tau causes age-­
67. Matenia D, Mandelkow EM.  The tau of MARK: a dependent neuronal dysfunction and loss in trans-
polarized view of the cytoskeleton. Trends Biochem genic mice. EMBO Rep. 2016;17(4):530–51.
Sci. 2009;34(7):332–42. 80. Decker JM, Krüger L, Sydow A, Dennissen FJ,
68. Schwalbe M, Biernat J, Bibow S, Ozenne V, Jensen Siskova Z, Mandelkow E, et  al. The Tau/A152T
MR, Kadavath H, et  al. Phosphorylation of human mutation, a risk factor for frontotemporal-spectrum
tau protein by microtubule affinity-regulating kinase disorders, leads to NR2B receptor-mediated excito-
2. Biochemistry. 2013;52(50):9068–79. toxicity. EMBO Rep. 2016;17(4):552–69.
69. Lippens G, Sillen A, Landrieu I, Amniai L, Sibille 81. Vanderweyde T, Apicco DJ, Youmans-Kidder K,
N, Barbier P, et al. Tau aggregation in Alzheimer’s Ash PEA, Cook C, Lummertz da Rocha E, et  al.
disease: what role for phosphorylation? Prion. Interaction of tau with the RNA-binding protein
2007;1(1):21–5. TIA1 regulates tau pathophysiology and toxicity.
70. Rubenstein R, Chang B, Grinkina N, Drummond Cell Rep. 2016;15(7):1455–66.
E, Davies P, Ruditzky M, et  al. Tau phosphoryla- 82. Apicco DJ, Ash PEA, Maziuk B, Leblang C, Medalla
tion induced by severe closed head traumatic brain M, Al Abdullatif A, et al. Reducing the RNA binding
injury is linked to the cellular prion protein. Acta protein TIA1 protects against tau-mediated neurode-
Neuropathol Commun. 2017;5(1):30. generation in vivo. Nat Neurosci. 2018;21(1):72–80.
71. Lopes S, Vaz-Silva J, Pinto V, Dalla C, Kokras 83. Braak H, Braak E. Staging of Alzheimer’s disease-­
N, Bedenk B, et  al. Tau protein is essential related neurofibrillary changes. Neurobiol Aging.
for stress-induced brain pathology. Proc Natl 1995;16(3):271–8.
Acad Sci. 2016;113(26):E3755–63. Available 84. Montine TJ, Phelps CH, Beach TG, Bigio EH,
from: http://www.pnas.org/lookup/doi/10.1073/ Cairns NJ, Dickson DW, et  al. National institute
pnas.1600953113 on aging-Alzheimer’s association guidelines for
72. Salminen A, Kauppinen A, Suuronen T, Kaarniranta the neuropathologic assessment of Alzheimer’s
K, Ojala J. ER stress in Alzheimer’s disease: a novel disease: a practical approach. Acta Neuropathol.
neuronal trigger for inflammation and Alzheimer’s 2012;123(1):1–11.
pathology. J Neuroinflammation. 2009;6:41. 85. Pooler AM, Polydoro M, Wegmann S, Nicholls
73. Busciglio J, Lorenzo A, Yeh J, Yankner BA. SB, Spires-Jones TL, Hyman BT.  Propagation of
β-Amyloid fibrils induce tau phosphoryla- tau pathology in Alzheimer’s disease: identification
tion and loss of microtubule binding. Neuron. of novel therapeutic targets. Alzheimers Res Ther.
1995;14(4):879–88. 2013;5(5):49. Available from: http://alzres.com/
74. Arendt T, Stieler J, Strijkstra AM, Hut RA, Rüdiger content/5/5/49
J, Van der Zee EA, et  al. Reversible paired helical 86. Sokolow S, Henkins KM, Bilousova T, Gonzalez
filament-like phosphorylation of tau is an adaptive B, Vinters HV, Miller CA, et  al. Pre-synaptic
25  Liquid-Liquid Phase Separation of Tau Protein in Neurobiology and Pathology 357

C-terminal truncated tau is released from corti- 98. Takeda S, Wegmann S, Cho H, DeVos SL,
cal synapses in Alzheimer’s disease. J  Neurochem. Commins C, Roe AD, et al. Neuronal uptake and
2015;133(3):368–79. propagation of a rare phosphorylated high-molec-
87. Wang Y, Balaji V, Kaniyappan S, Krüger L, Irsen S, ular-weight tau derived from Alzheimer’s disease
Tepper K, et al. The release and trans-synaptic trans- brain. Nat Commun. 2015;6:8490. Available from:
mission of Tau via exosomes. Mol Neurodegener. http://www.pubmedcentral.nih.gov/articlerender.
2017;12(1):5. fcgi?artid=4608380&tool=pmcentrez&rendertype
88. Saman S, Kim WH, Raya M, Visnick Y, Miro S, =abstract
Saman S, et al. Exosome-associated tau is secreted 99. Wu JW, Herman M, Liu L, Simoes S, Acker CM,
in tauopathy models and is selectively phosphory- Figueroa H, et  al. Small misfolded tau species are
lated in cerebrospinal fluid in early Alzheimer dis- internalized via bulk endocytosis and anterogradely
ease. J Biol Chem. 2012;287(6):3842–9. and retrogradely transported in neurons. J  Biol
89. Dujardin S, Bégard S, Caillierez R, Lachaud C, Chem. 2013;288(3):1856–70.
Delattre L, Carrier S, et al. Ectosomes: a new mech- 100. Evans LD, Wassmer T, Fraser G, Smith J, Perkinton
anism for non-exosomal secretion of Tau protein. M, Billinton A, et al. Extracellular monomeric and
PLoS One. 2014;9(6):e100760. aggregated tau efficiently enter human neurons
90. Katsinelos T, Zeitler M, Dimou E, Karakatsani through overlapping but distinct pathways. Cell Rep.
A, Müller HM, Nachman E, et  al. Unconventional 2018;22(13):3612–24.
secretion mediates the trans-cellular spreading of 101. Holmes BB, DeVos SL, Kfoury N, Li M, Jacks R,
tau. Cell Rep. 2018;23(7):2039–55. Yanamandra K, et  al. Heparan sulfate proteogly-
91. Pooler AM, Polydoro M, Maury EA, Nicholls SB, cans mediate internalization and propagation of
Reddy SM, Wegmann S, et al. Amyloid accelerates specific proteopathic seeds. Proc Natl Acad Sci.
tau propagation and toxicity in a model of early 2013;110(33):E3138–47.
Alzheimer’s disease. Acta Neuropathol Commun. 102. Weingarten MD, Lockwood AH, Hwo SY,
2015;3:14. Available from: http://www.pubmedcen- Kirschner MW.  A protein factor essential for
tral.nih.gov/articlerender.fcgi?artid=4371800&tool microtubule assembly. Proc Natl Acad Sci U S A.
=pmcentrez&rendertype=abstract 1975;72(5):1858–62. Available from: http://www.
92. Wu JW, Hussaini SA, Bastille IM, Rodriguez pubmedcentral.nih.gov/articlerender.fcgi?artid=432
GA, Mrejeru A, Rilett K, et  al. Neuronal activity 646&tool=pmcentrez&rendertype=abstract
enhances tau propagation and tau pathology in vivo. 103. Bunker JM, Wilson L, Jordan MA, Feinstein

Nat Neurosci. 2016;19(8):1085–92. SC.  Modulation of microtubule dynamics by
93. Dujardin S, Lecolle K, Caillierez R, Begard S, tau in living cells: implications for develop-
Zommer N, Lachaud C, et al. Neuron-to-neuron wild-­ ment and neurodegeneration. Mol Biol Cell.
type Tau protein transfer through a trans-synaptic 2004;15(6):2720–8.
mechanism: relevance to sporadic tauopathies. Acta 104. Dixit R, Ross JL, Goldman YE, Holzbaur
Neuropathol Commun. 2014;2:14. Available from: ELF.  Differential regulation of dynein and
http://www.ncbi.nlm.nih.gov/pubmed/24479894 kinesin motor proteins by tau. Science (80-).
94. De Calignon A, Polydoro M, Suárez-Calvet M, 2008;319(5866):1086–9.
William C, Adamowicz DH, Kopeikina KJ, et  al. 105. Gong CX, Lidsky T, Wegiel J, Zuck L, Grundke-­
Propagation of tau pathology in a model of early Iqbal I, Iqbal K.  Phosphorylation of microtubule-­
Alzheimer’s disease. Neuron. 2012;73(4):685–97. associated protein tau is regulated by protein
95. Liu L, Drouet V, Wu JW, Witter MP, Small SA, phosphatase 2A in mammalian brain. Implications
Clelland C, et  al. Trans-synaptic spread of tau for neurofibrillary degeneration in Alzheimer’s dis-
pathology in vivo. PLoS One. 2012;7(2):e31302. ease. J Biol Chem. 2000;275(8):5535–44.
96. Wegmann S, Maury EA, Kirk MJ, Saqran L, Roe A, 106. Tan R, Lam AJ, Tan T, Han J, Nowakowski DW,
DeVos SL, et al. Removing endogenous tau does not Vershinin M, et  al. Microtubules gate tau conden-
prevent tau propagation yet reduces its neurotoxicity. sation to spatially regulate microtubule functions.
EMBO J. 2015;34(24):1–14. Available from: http:// bioRxiv. 2018; https://doi.org/10.1101/423376.
www.ncbi.nlm.nih.gov/pubmed/26538322 107. Dennissen FJA, Anglada-Huguet M, Sydow A,
97. Kfoury N, Holmes BB, Jiang H, Holtzman DM, Mandelkow E, Mandelkow E-M.  Adenosine A1
Diamond MI.  Trans-cellular propagation of Tau receptor antagonist rolofylline alleviates axonopathy
aggregation by fibrillar species. J  Biol Chem. caused by human Tau ΔK280. Proc Natl Acad Sci.
2012;287(23):19440–51. 2016;113(41):11597–602.
 The Pathophysiology of Tau
and Stress Granules in Disease 26
Anna Cruz, Mamta Verma, and Benjamin Wolozin

Introduction cal concept that particular elements of the stress

response mediated by RNA binding proteins
Stress and the stress response are fundamentally actually accelerate tau oligomerization and
central features of any disease. Stress and the thereby accelerate disease progression in tauopa-
stress response are also clearly central features in thy. Conversely, we will also show how RNA
the pathophysiology of tauopathy and the biol- binding proteins are therefore key targets for
ogy of AD.  Chapters 16, 23, 28 and 29 in this therapeutic intervention in tauopathies, including
book address how stress modifies the phosphory- Alzheimer’s disease.
lation of tau, the role of stress in the biology of
the endoplasmic reticulum and the translational
stress response, and the effects of behavioral  ysfunction of RNA Binding
D
stress and glucocorticoids on tau. The stress Proteins in Neurodegenerative
response is undoubtedly necessary to cope with Disorders
the harm arising from chronic exposure to
β-amyloid (Aβ), oligomeric tau and age or dis- Reviewing the basic biology of RBPs, stress
ease related reductions in cerebral blood flow. granules and the translational stress response is
However, this chapter will address the paradoxi- necessary in order to understand how and why
tau would interact with RBPs. An increasing
Authors Anna Cruz and Mamta Verma have been equally number of genetic studies show that RNA bind-
contributed to this chapter. ing proteins (RBPs) are central to the pathophysi-
ology of multiple neurological disorders.
A. Cruz · M. Verma Mutations in genes that encode RBPs cause
Department of Pharmacology and Experimental
Therapeutics, Boston University School of Medicine, motor neuron diseases such as amyotrophic lat-
Boston, MA, USA eral sclerosis (ALS), spinal muscular atrophy
B. Wolozin (*) (SMA), multisystem proteinopathy (MSP), and
Department of Pharmacology and Experimental frontotemporal lobar degeneration (FTLD) [40].
Therapeutics, Boston University School of Medicine, ALS is the most common motor neuron disorder
Boston, MA, USA that leads to progressive loss of upper and lower
Department of Neurology, Boston University School motor neurons, muscle weakness, atrophy and
of Medicine, Boston, MA, USA ultimately death [78].
Program in Neuroscience, Boston University School TAR DNA-binding proteins 43 (TDP-43) is
of Medicine, Boston, MA, USA the major constituent of pathological
e-mail: bwolozin@bu.edu

© Springer Nature Singapore Pte Ltd. 2019 359

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_26
360 A. Cruz et al.

u­ biquitinated inclusions present in ALS and fron- nucleus to the cytoplasm where they associate
totemporal dementias defined by progranulin with SGs (Fig. 26.1). These granules serve as a
haploinsufficiency (FTD-TDP-43 or FTD-U) cytoprotective mechanism against stress as it
[69]. Mutations in TDP-43 also cause rare forms temporarily inhibits the translation of non-­
of familial ALS, which demonstrates that dys- essential mRNA and promote the translation of
function of TDP-43 is sufficient to cause disease transcripts necessary for cell survival [41, 49].
[83]. TDP-43 is not the only RBP to exhibit However, mutations in TDP-43 and FUS increase
mutations linked to ALS, though. Multiple RBPs the propensity to aggregate, leading to the accu-
exhibit mutations that are genetically linked to mulation of persistent cytoplasmic SGs, and the
ALS: heterogeneous nuclear ribonucleoprotein formation of pathological inclusions of these pro-
A1 (hnRNP A1), heterogeneous nuclear ribonu- teins in the human brain [9, 16, 48]. Support for
cleoprotein A2/B1 (hnRNP A2/B1), fused in sar- the hypothesis that the pathological inclusions in
coma/translocates in liposarcoma (FUS/TLS), brain derive from SGs comes from observed
Ewing’s sarcoma breakpoint region 1 (EWSR1), colocalization with SG markers such as eIF3,
TATA-box binding protein associated factor 15 eIF4G, TIA1 and PABP [48, 16].
(TAF15), Matrin3 (MATR3), and TIA1 cytotoxic This chapter will focus on SGs because the
associated granule binding protein (TIA1) [40, relationship to disease is clear, SGs provide a lin-
53] (Table  26.1). In addition, polyglutamine ear pathway between chronic stress and disease,
(polyQ) expansions (27–33 Qs) in Ataxin-2 and the disease-associated RBPs examined to
(ATXN-2) are an important genetic risk for ALS date have been shown to co-localize with SGs.
[19]. RBPs are generally defined by the presence However, RBPs form many types of RNA gran-
of a homologous RNA binding domain, the RNA ules to mediate many different functions, includ-
recognition motif (RRM); many RBPs also share ing splicing, transcription, ribosome genesis,
homologous low complexity Gly-rich domains RNA transport, RNA degradation, RNA transla-
(LCDs) [31, 53]. FUS, EWSR1, and TAF15, tion, viral defense and many other functions. The
which comprises the FET family, also share com- pathophysiological principles linking SGs to dis-
mon zinc finger domains [31]. The RRMs allow ease presented below might also apply to these
the binding of these proteins to DNA and RNA in other types of RBPs, particularly for familial
a sequence-specific manner, while the LCD is a RBPs that exhibit mutations rendering the RBPs
prion-like domain that plays a critical role in the more aggregation prone. However, differing
formation and the dynamic biophysical state of types of RNA granules likely differ in their pro-
stress granules (SGs) [30]. These RBPs are mul- pensity to precipitate the irreversible protein
tifunctional RNA processing proteins that pre- aggregation that causes disease because of differ-
dominantly reside in the nucleus and are generally ences in the types of proteins associated with
expressed in many different types of cells and tis- each RNA granule type and/or differences in
sues [40]. post-translational modifications, as discussed
below. SGs currently appear to represent the
RNA granule exhibiting the greatest propensity
 NA Binding Proteins Mediate
R to elicit the irreversible protein aggregation asso-
Disease Through Stress Granules ciated with neurodegenerative diseases.
Many of the mutations in FUS are thought to
The identification of disease-linked mutations in increase cytoplasmic localization by disrupting
the genes that encode these RBPs, particularly the nuclear localization signal, which prevents
TDP-43 and FUS, introduced a paradigm shift in the transportation of these proteins to the nucleus
the study of ribostasis and proteostasis in disease. [16, 71, 94]. The disease associated mutations in
Under physiological conditions, TDP-43 and TDP-43 occur predominantly in the LCD domain,
FUS localize in the nucleus. But, in the presence increase the tendency to aggregate, and forma-
of a cellular stress, they redistribute from the tion of cytoplasmic SGs [48, 38]. In most cases,
26  The Pathophysiology of Tau and Stress Granules in Disease 361

Table 26.1  Physiological and pathophysiological functions of RBPs and non-RBPs implicated in neurodegenerative
diseases
362 A. Cruz et al.

Fig. 26.1  The stress granule cycle. (a) Under basal con- SGs mature over time, consolidating and beginning to
ditions most RNA binding proteins are nuclear, while incorporate multiple secondary proteins that include RNA
some are predominantly cytoplasmic. (b) With stress, binding proteins as well as proteins with functions inde-
nuclear RNA binding proteins exit the nucleus through the pendent of RNA (e.g., autophagy, apoptosis). (e) Upon
nuclear pore (arrow) and enter the cytoplasm. (c) The SGs resolution of the stress, the SGs dissolve and nuclear RNA
begin to nucleate. Proteins such as TIA1, TIAR, TTP, binding proteins return to the nucleus
G3BP and FMRP are primary nucleators of SGs. (d) The

the increased tendency to aggregate shifts the DNA damage or proteostatic dysfunction [14, 25,
biophysical state of these granules, that is thought 41, 42, 56, 57, 64]. SGs are classically made up
to extend their persistence, resulting in the for- of mRNAs, RNA-binding proteins, small 40S
mation of persistent pathological SGs which can ribosome subunits, translation initiation factors,
then consolidate to form the classic pathologic and broadly non-RNA binding proteins [12, 64,
structures that are associated with disease [55]. 76]. SG formation occurs in stages, with core
Finally, it is important to note that cleavage of nucleating RBPs initiating SG formation, fol-
TDP-43 can produce fragments with a strong ten- lowed by secondary association of a wide variety
dency to aggregate through a mechanism that is of proteins. The complexity of SGs varies with
not strongly linked to the translational stress the type of stress, the type of cell and the duration
response [71, 94]. of the stress [54].
The initial changes in RNA metabolism
induced by stress result in polysome stalling and
 he Biology of Stress Granules
T nucleation of SGs by a set of core nucleating
and the Translational Stress RBPs, which include TIA1, G3BP1 and 2, FMRP,
Response TTP and TIAR; proteomic experiments point to a
comprehensive list of core RBPs (Fig. 26.1) [5,
SGs are cytoplasmic complexes that form in con- 54]. Nucleation of these RBPs is controlled by
cert with inhibition of RNA translation; stimulate PTMs, which are described below, and by loca-
of SG include a wide variety of stresses that tion. TIA1 for instance, translocates from the
include nutritional stress, heat or osmotic shock, nucleus to the cytoplasm during stress [41].
26  The Pathophysiology of Tau and Stress Granules in Disease 363

These core nucleating components associate with aggregates that form in neurodegenerative dis-
the mRNA transcripts and protein components of eases [65, 74]; mutations that are associated with
the stalled initiation complexes including eIF3, familial disease frequently increase the rate of
eIF4F (consist of eIF4E, eIF4A and eIF4G), amyloidogenic transition [65, 74]. Secondary
eIF4B, small ribosomal subunits and Poly A nucleation also allows the association of SGs
Binding protein 1 (Table 26.1) [41, 42]. A wide with other proteins, such as tau, that exhibit a
variety of secondary proteins associate with SGs. high propensity to aggregate into stable amyloids
These proteins include many different RBPs, [88].
including those associated with ALS, such as Three different biophysical considerations
TDP-43, FUS and hnRNPA0 [41, 42]. However, explain the biology of LLPS. (1) The physical
secondary proteins that associate with SGs also chemistry of liquids creates the conditions for
include scaffolding proteins, such as caprin, sig- LLPS.  Chemicals that are in liquid form and
naling proteins such as HDAC6 and SirT6, exhibit strong physical differences, such as oil
nuclear pore proteins such as nup98 [17], disag- and water, will phase separate to minimize the
gregases such as karyopherin-b2 (Transportin-1, free energy of the mixture, reduce unfavorable
Kapnb2) (Table 26.1) [16, 27, 28, 32], and pro- interactions and promote weak bonding. The
teins linked to cell death pathways, such as TRAF aqueous nature of proteins thus provides the con-
and FAST (Table 26.1) [11, 23, 44]. ditions that allow for phase separation. The phase
separation is promoted by weak bonding of low
complexity protein sequences that consist pri-
Phase Transition and the Role marily of alanine, glycine, glutamine and proline,
of Protein Aggregation with some extra complexity arising from inter-
in the Biology of SGs spersed arginine and asparagine [34, 75]. (Note
that the phase separation of proteins does not pro-
The formation and consolidation of SGs (and duce the extreme concentrations (e.g., 55  M)
other RNA granules) appears to depend on the occurring when oil and water phase separate
biophysical processes of liquid-liquid phase sep- because the weak interactions present in proteins
aration (LLPS) and protein aggregation. The pro- are only moderately favored over interactions
cess of LLPS is described in detail in Chaps. 24 between proteins and water in the aqueous solu-
and 25 (Kosik & Han, Wegmann), but a brief tion) (Fig.  26.1a). The low affinity binding of
overview will be given in this chapter because of multiple short regions of low complexity domains
its fundamental importance for understanding the produces the dynamic phase separation that char-
biology of SGs. The roles of LLPS and protein acterizes RNA granules [91, 92]. (2) The low
aggregation are the essential features that make complexity regions that promote LLPS occur in
SGs so important for the pathophysiology of neu- intrinsically disordered regions (referred to as
rodegenerative diseases generally, and tauopa- IDRs in the literature). The lack of order enables
thies specifically. Under transient stress the “sticky” sequences in these regions to move
conditions SG components assemble and disas- in a dynamic manner forming the multiple weak
semble quickly, forming the highly dynamic associations that drive the LLPS. (3) The final
structures that are governed by the biophysics of consideration is polymer chemistry. RNA greatly
phase separation [5, 46]. The dynamic nature of facilitates protein based LLPS by forming a scaf-
these phase separated proteins enables transitions fold that helps to stabilize the phase separating
between multiple protein conformations. A fun- proteins, keeping them generally in the same
damental weakness of this biology gives rise to region [20]. Thus, RBPs bound to RNA phase
neurodegenerative diseases. With extended time, separate at a lower concentration than is required
such as might occur with chronic stress, some SG in absence of RNA [20]. The tendency of RBPs
proteins transit into highly stable amyloid like to cluster around RNA combines with the pres-
states, similar to that observed in the protein ence of intrinsically disordered regions that can
364 A. Cruz et al.

self-associate in a low affinity manner to render that hyperphosphorylated tau accumulates in the
LLPS a prominent feature of RBP biology, lead- somatodendritic arbor during stress to adapt pro-
ing to formation of many types of RNA granules, tein synthesis to address the stress [88].
including SGs.

Tau Regulates Stress Granules

Tau and Stress Granules
The role of tau in the translational stress response
 au Is Sorted to the Somatodendritic
T is evident from studies of RBPs in neurons during
Domain in Stress stress. The relationship of tau to SGs is readily
apparent when examining TIA1, an RBP that is
The information presented above provide a clear one of the core nucleating SG proteins. In cell
mechanistic pathway through which RBPs, SGs lines TIA1 is completely nuclear under basal con-
and RNA metabolism contribute to pathological ditions, and translocates into the cytoplasm under
aggregation and the pathophysiology of moto- stress. Translocation of TIA1 has been demon-
neuron diseases. The section below will explain strated in response to many different stresses as
how this mechanism involves tau, and the pro- well as viral infections (for a general review of
found manner in which RBPs, SGs and the trans- TIA1, see Anderson et al. [1]), but also includes
lational stress response contribute to the stresses that are very relevant to disease, such as
pathophysiology of tauopathy. arsenite, glucocorticoids, Aβ and tau oligomers
Tau is normally most abundant in the axons of [37, 82, 87, 88]. In each case, the resulting TIA1
neurons, where its primary function is to stabilize SGs co-localize with hyperphosphorylated tau.
microtubules [4]. In AD, as well as in stress, tau However, comparison of SGs associated with
becomes phosphorylated by proline directed TIA1 and G3BP1 demonstrate that SGs are not
kinases, such as glycogen synthase kinase β uniform species [54]. The neuronal SY5Y cell line
(GSK3β), cyclin dependent kinase 5 (CDK5) and exhibits a strong stress response after glucocorti-
microtubules affinity-regulating kinases coid treatment, exhibiting both TIA1 and G3BP1
(MARKs); this type of phosphorylation will be positive SGs [82]. Interestingly, hyperphosphory-
referred to as “hyperphosphorylation” [89]; this lated tau inclusions strongly colocalize with TIA1-
hyperphosphorylated tau accumulates in the positive SGs, but show little colocalization with
somatodendritic arbor where it eventually forms G3BP1-positive SGs [82]. The relevance of this
neurofibrillary tangles. Originally hyperphos- point becomes clearer when considering tau
phorylated tau was through to translocate from pathology in vivo (discussed below), where one
the axon to the cytoplasm, but the translocation sees induction of both TIA1 and G3BP1 pathology
model didn’t fit the pathology, which does not with disease, but only TIA1 colocalizes with
show extensive axonal tau phosphorylation. More hyperphosphorylated tau [87]. In neurons under
recent data indicate that accumulation occurs as basal conditions TIA1 is also abundant in the
newly synthesized tau becomes phosphorylated nucleus, but also has some presence in the cyto-
and is prevented from entering the axon [33, 84, plasm [88]. However, in tau knockout neurons,
96]. In the presence stress, tau is phosphorylated TIA1 is completely nuclear [88]. In addition,
and localizes in the soma and dendrite where it under conditions of stress (e.g., Aβ toxicity), TIA1
can interact with RBPs associated with SGs [33, exhibits reduced translocation to the cytoplasm in
84]. The accumulating phospho-­tau arises from tau KO neurons (Wolozin, personal communica-
phosphorylation of de novo synthesized tau rather tion). Conversely, over-­expressing tau increases
than translocation of phospho-tau from the axon SG formation and the associated repression of pro-
[96]. The physiological logic for which the neu- tein synthesis [60, 87, 88].
ron would change the distribution of tau repre- Independent approaches support the intimate
sents a fundamental question for tau research, and link between tau, SGs and translational control.
is one that has never been explained. We propose Tau is known to exist in dendrites near dendritic
26  The Pathophysiology of Tau and Stress Granules in Disease 365

boutons [36]. These small tau granules appear to be ies show colocalization of pathological tau
linked to translational control/protein synthesis (hyperphosphorylated or misfolded) with multi-
because chemicals that modulate protein synthesis ple RBPs [2, 3, 55, 82, 87, 88]. The degree of
affect the tau granule distribution. Cycloheximide, co-localization detected in human tissues,
which inhibits protein synthesis and inhibits SG though, likely under-represents reality because
formation, also prevents clustering of tau into gran- detection of RBPs in situ decreases dramatically
ules in the dendritic arbor. Conversely, puromycin, with fixation time [55]. RBPs are abundantly
which also inhibits protein synthesis but stimulates detectable at <2 h of fixation and remain readily
SG formation, enhances clustering of tau into gran- detectable at <24 h of fixation, but become diffi-
ules in the dendritic arbor [88]. Other SG inhibi- cult to detect at >48 h of fixation [55]; this sensi-
tors, such as the protein kinase R inhibitor C16 or tivity to fixation duration impacts greatly on
the PERK inhibitor GSK2606414 also prevent staining of human tissues because most human
coalescence of tau with SGs [88]. cases have been fixed for much more than 48 h.
Immunoprecipitating TIA1 identifies tau in the The pattern of SG pathology and co-­
protein interactome network, as well as many other localization with tau differs dramatically based
RBPs for which binding to TIA1 requires tau, and on the RNA binding protein examined. TIA1, co-­
immunoprecipitating tau identifies multiple co- localizes strongly with pathological tau in human
associating RBPs [26, 35, 55, 88, 90]. These data tissues [87]. In contrast, rasGAP-binding protein
demonstrate that the biology of tau is tightly con- (G3BP) only shows weak colocalization with
nected with that of SGs and translational control. phosphorylated tau, despite exhibiting increased
Tau exhibits a tendency to phase separate in accumulation in neurons with increasing disease
the presence of RNA in vitro much like RBPs, as severity [87]. In animal tissues, where shorter
discussed in the Chaps. 24 and 25 by Kosik and fixation times are possible, tau is observed to co-­
by Wegmann [18, 97]. The propensity of tau to localize with multiple other RBPs, including
phase separate might facilitate its interaction DDX6, eIF2α, hnRNPA0, and PABP [55, 82].
with RBPs and the formation of SGs, although Interestingly, the pattern of reactivity appears to
this point has yet to be empirically demonstrated. differ with the type of pathology. Co-localization
Hyperphosphorylation of tau increases its pro- of tau with RBPs is strongest with smaller inclu-
pensity to form droplets in vitro, which suggests sions; mature neurofibrillary tangles exhibit
that tau hyperphosphorylation might function to accumulation of RBPs adjacent to the pathologi-
promote the formation of phase separated com- cal tau tangles, suggesting the hypothesis that the
plexes of tau, RBPs and RNA [18]. However, tau RPBs become excluded as tau consolidates to
also has a strong tendency to fibrillize, and stud- form the mature tangle [55].
ies of human tau in neurons show that hyperphos- The putative dysfunction of RBPs and RNA
phorylation renders tau prone to irreversible metabolism in tauopathy can be tested by exam-
aggregation [18]. Thus, the consolidation of tau ining RNA splicing. If RBPs become seques-
into droplets and SGs might increase the ten- tered as protein aggregates in persistent
dency of tau to aggregate, thereby enhancing a pathological SGs, then one might expect to
pathway that leads to neurodegeneration. observe effects on RNA metabolism when exam-
ined through the lens of RNAseq. Indeed, multi-
ple transcriptome studies show dramatic changes
 au Colocalizes with RNA Binding
T in RNA transcriptomes in tauopathies. Studies
Proteins in Disease from several laboratories, including our own,
show that splicing is dramatically altered in ani-
These cell culture studies complement cogent mal models of tauopathy as well as in cases of
pathological data. The connection between tau AD [2, 3, 6, 27, 28, 63, 81]. The changes in
and RBPs is evident in pathological tissues from splicing are far greater than the comparatively
human cases of AD, FTD-tau as well as animal modest changes in transcript levels. Since the
models of tauopathy. Molecular pathology stud- splicesome is made up of RBPs, the large
366 A. Cruz et al.

changes in splicing that occur with disease are The selective interaction between tau oligo-
consistent with a model in which RBPs become mers and TIA1 is supported by independent stud-
sequestered away from splicesomes in the ies of tau propagation. Tau oligomers and fibrils
nucleus leading to dysfunctional splicing. were isolated from 9-month P301S tau mice, and
propagated in both WT and P301S tau mice; the
results were similar for both but more striking in
 au Oligomers Mediate Interactions
T the mice over-expressing human tau [37]. Both
with RNA Binding Proteins oligomeric and fibrillar tau exhibited robust prop-
agation, which is consistent with Chaps. 30 and
Animal models provide insight into the mecha- 31 and multiple reports in the literature [15, 47,
nisms underlying the interaction of tau with RBPs, 79]. The experimental design also allowed side by
SGs and the translational stress response. Our side comparison of toxicity, with the results show-
laboratory recently crossed PS19 P301S tau mice ing that oligomeric tau produced abundant tau
with TIA1−/− mice, and demonstrated that reduc- pathology and abundant neurodegeneration while
ing TIA1 in vivo provides striking rescue of the the fibrillar tau produced abundant tau pathology
degenerative phenotype associated with the PS19 but no degeneration evident after 3  months of
P301S tau mice [2, 3]. Reducing TIA1 expression propagation [37]. This confirms prior studies sug-
by 50% greatly decreased the number and size of gesting that oligomeric tau is much more toxic
cytoplasmic pathological TIA1 granules (which than fibrillar tau. [80, 86]. These studies suggest
colocalize with SG markers). The TIA1 reduction that oligomeric or misfolded forms of tau are
also yielded striking rescue of behavior, neuronal toxic, drive cognitive decline, and act through a
and synaptic degeneration, cortical thickness, as mechanism that occurs before or independently of
well as a 26% increase in survival despite the con- the development of NFTs [73, 80, 95].
tinued five-fold over-­expression of tau [2, 3]. TIA1 The propagation studies also demonstrated the
reduction also decreased the amount of hyper- strong link between oligomeric tau and TIA1-­
phosphorylated tau evident at 3 months of age [2, positive SGs. The oligomeric tau propagated tau
3]. This acute reduction in tau pathology is consis- pathology that co-localized with cytoplasmic
tent with cell culture studies showing that TIA1 TIA1-positive SGs, as shown by co-localization
knockdown also provides neuroprotection and with TIA1, PABP and eIF3η; this was true in both
reduces tau pathology [88]. the ipsilateral and contralateral cortex providing
Insights into the mechanism of tau/TIA1 inter- clear demonstration of tau propagation [37]. In
actions arise from our studies examining the mice contrast, the fibrillar tau propagated nicely, but
at later ages, as well as from a subsequent study showed no colocalization with TIA1-positive
comparing the propagation of oligomeric and granules. An additional mechanistic link between
fibrillar tau. Aging of the P301S tau::TIA1+/+ TIA1 and tau was evident from similar studies
and +/− mice showed dramatic changes in the performed in P301S tau::TIA1+/−. These mice
aggregation of tau. The P301S tau::TIA1+/− showed abundant propagation of fibrillar tau, but
exhibited striking (90%) reduction in the accu- very little (if any) propagation of oligomeric tau,
mulation of oligomeric tau at 9  months, and an which provides support for the hypothesis that
equally striking increase (>10-fold) in fibrillar TIA1 selectively interacts with oligomeric tau.
tau at 9  months of age. Analysis by immuno-­
electron microscopy demonstrated that TIA1
binds to phosphorylated tau oligomers but not tau  Model for the Interactions of Tau
A
fibrils; the ability of TIA1 to increase tau oligo- in Stress and with RNA Binding
merization (assessed by ELISA) confirmed this Proteins
observation [88]. These data suggest the hypoth-
esis that TIA1 interacts selectively with phos- The accumulating data presented above suggest a
phorylated tau oligomers. model in which phosphorylated tau accumulates
26  The Pathophysiology of Tau and Stress Granules in Disease 367

in the somatodendritic arbor where it oligomer- reducing eIF2α phosphorylation (by expressing
izes and then interacts with TIA1 and possibly the phosphatase adapter protein, GADD34)
other RNA binding proteins and/or ribosomal relieves the translational repression caused by
proteins. Binding of tau to these proteins appears eIF2α-P and delays neurodegeneration [66].
to promotes the translational stress response, Studies in drosophila and in rat primary cortical
which reduces synthesis of specialized proteins neurons expressing TDP-43 show that inhibiting
(such as those related to synaptic function) and the phosphorylation of eIF2α alleviates the toxic-
increases synthesis of proteins needed for the ity induced by TDP-43 [43]. Similar approaches
stress response. The proline directed phosphory- have now also been shown to apply to inhibition
lation that is characteristic of the stress response of toxicity associated with tau, as well as with Aβ
also inhibits binding of tau to microtubules, per- [51, 52, 59, 88]. These studies point to inhibition
haps allowing for more tau oligomerization and of eIF2α phosphorylation as a potential therapeu-
interaction with the translational machinery. tic intervention to neurodegenerative diseases.
Chronic stress, though, leads to the accumulation However, the clinical value of each of these
of oligomeric tau which is toxic, although the approaches is limited by toxic liabilities. For
mechanism of toxicity is not currently known. example, PERK inhibition leads to pancreatic
toxicity, although partial inhibition might be clin-
ically tolerated [29], and inhibition of PKR
 herapeutic Approaches Based
T enables reemergence of retroviruses [68].
on Modulating SGs Other teams are developing innovative small
and the Translational Stress molecule therapeutics to inhibit the accumulation
Response of persistent pathological SGs. Our group used
neuronal PC12 cells inducibly over-expressing
One of the most important aspects of studying the TDP-43 to screen a library of brain penetrant
relationship between tau, RBPs, SGs and the small molecules [10]. This screen lead to the
translational stress response is the possibility of identification of 16 hits that reduced the accumu-
innovative disease therapies. The biology of SGs lation of TDP-43 inclusions [10]. Some of these
involves multiple biochemical pathways, some of compounds were able to improve survival of neu-
which not been considered previously in the con- rons in a C. elegans model of over-expressing
text of neurodegenerative diseases. TDP-43, suggesting the potential for in vivo effi-
The classic SG/translational stress response is cacy [10]. Cell lines expressing TDP-43 have
regulated by phosphorylation of eukaryotic initi- also been used to screen a variety of other puta-
ation factor 2 (eIF2α-P); this pathway has been tive therapeutics, with promising results although
studied by a number of different laboratories. In these approaches have yet to be tested in vivo [13,
presence of stress, stress-related kinases, includ- 67].
ing PERK, PKR, HRI and GCN2, phosphorylate An alternative approach for disease modifying
eIF2α trigger the assembly of SGs that inhibit therapy has been direct therapeutic targeting of
global cellular protein synthesis [93]. Chronic RBPs for therapy. Studies in this area have
diseases produces a sustained stress response, focused on delaying neurodegeneration in mod-
persistent SGs and continued global translational els of tauopathy, based on tau over-expression,
repression [8, 45, 77, 85]. These observations and models of ALS based on TDP-43 over-­
suggest the hypothesis that the translational stress expression [2, 3, 7]. The discussion of tau,
response is too active, and inhibiting the stress described above, demonstrates how reducing
response might be beneficial in neurodegenera- TIA1  in the P301S tau mouse model rescues
tive disease. One of the first studies demonstrat- memory loss, reduces neurodegeneration and
ing the value of inhibiting the translational stress improves survival [2, 3]. Protection is also be
response was performed in a mouse prion model, achieved by TIA1 knockdown with shRNA
where Mallucci and colleagues demonstrated that directed against TIA1 [88]. TIA1 reduction also
368 A. Cruz et al.

protects against tauopathy in a model of tau prop- affecting the total TDP-43 expression levels and
agation, which indicates that reducing TIA1 pro- increases survival percentage of flies expressing
vides generalized protection against tau pathology TDP-43  in the brain. Inhibiting Tankyrase pre-
produced within neurons as well as propagated vents the stress-induced formation of cytoplas-
among neurons [37]. These results suggest that mic TDP-43 foci without altering the dynamics
reducing TIA1 might provide broad-based neuro- and assembly of SGs. TDP-43 also becomes
protection in AD and other tauopathies. phosphorylated upon prolonged exposure to
The therapeutic potential of reducing RNA stress, but the phosphorylation appears to stimu-
binding proteins has also been examined in a late aggregation of TDP-43 through a pathway
model of ALS based on over-expressing TDP-43 that does not co-localize with SG markers [58].
[7]. These studies focused on ATXN2, which par- Since phospho-TDP-43 is associated with dis-
ticipates in RNA metabolism, contributing to ease pathology, this TDP-43 phosphorylation
RNA splicing, and degradation [39, 72]. ATXN2 pathway might identify a disease-relevant path-
contains a domain with a small number of CAG way [70]. Since tankyrase downregulation
trinucleotide repeats (producing glutamines) increases nuclear TDP-43 and decreases cyto-
whose expansion is associated with disease. plasmic TDP-43, inhibiting tankyrase might
Disease linked mutations in ATXN2 that expand reduce the amount of cytoplasmic TDP-43 that is
the CAG domain to 34 or more repeats cause the available to be phosphorylated and provide a
neurodegenerative disorder spinocerebellar potential pharmacological intervention for dis-
ataxia type 2 (SCA2) [24]. However, disease eases associated with pathological TDP-43 [58].
linked mutations that expand the CAG domain to The most striking success in delay of neurode-
27–33 repeats increase the risk of ALS, with generative disease based on targeting RNA metab-
associated TDP-43 pathology [19]. The link olism comes from the field of spinal muscular
between ATXN2 and TDP-43 was strengthened atrophy (SMA). This disease is caused by a mis-
with the observation that reducing ATXN2 sig- sense mutation that causes exon skipping that pro-
nificantly extends survival in an animal model of duces an inactive form of the gene survival of
ALS based on TDP-43 overexpression [7]. motor neuron 1 (SMN1) [50]. Teams of investiga-
Knockout and antisense-mediated knockdown of tors from Ionis Pharmaceuticals and Biogen devel-
ATXN2  in TDP-43 transgenic mice decreased oped antisense oligomers capable of correcting the
SGs containing TDP-43, reduced the accumula- exon skipping, which increases levels of SMN1.
tion of phosphorylated TDP-43 spinal cord inclu- Multiple clinical trials now demonstrate that appli-
sions, improved the motor performance and cation of these antisense oligomers to children with
increased the median lifespan by 35% [7]. These SMA produces striking rescue from disease, and
findings indicate that ATXN2 plays a crucial role prolonged delay of disease progression [21, 22, 61,
in the development of pathological SGs and aug- 62]. The striking ability of these antisense oligo-
mentation of TDP-43 toxicity. mers to delay disease progression and actually
Targeting other regulators of RNA metabo- improve clinical conditions in patients with SMA
lism has also been shown to ameliorate disease. now serves as a bench post for future therapies.
McGurk et  al. [58] found that downregulating
Tankyrase 1 and 2, a poly (ADB-ribose) poly-
merase, reduced the formation of cytoplasmic Conclusion
TDP-43 foci without affecting the SG assembly.
PAR binds to the PAR-binding motif in the The work covered in this chapter presents a cogent
N-terminal region of TDP-43 and is necessary for paradigm for understanding the pathophysiology of
its sequestration in SGs in mammalian cells and tauopathy. Accumulating data suggest that tau
neurons upon exposure to stress [58]. Inhibiting becomes hyperphosphorylated and oligomerize as
Tankyrase significantly increases nuclear TDP-­ part of an endogenous mechanism to promote the
43 and decreases cytoplasmic TDP-43 without translational stress response. Tauopathies, though,
26  The Pathophysiology of Tau and Stress Granules in Disease 369

are chronic diseases. The prolonged stress leads to a 4. Ballatore C, Lee VMY, Trojanowski JQ. Tau-mediated
neurodegeneration in Alzheimer’s disease and related
persistent stress response, which provides time for disorders. Nat Rev Neurosci. 2007;8(9):663–72.
oligomeric tau to accumulate, cause toxicity and https://doi.org/10.1038/nrn2194.
neurodegeneration. We show that tau-mediated 5. Banani SF, et  al. Compositional control of phase-­
neurodegeneration occurs through a mechanism separated cellular bodies. Cell. 2016;166(3):651–63.
https://doi.org/10.1016/j.cell.2016.06.010.
that is mediated by RNA binding proteins and the 6. Barbash S, et al. Alzheimer’s brains show inter-related
translational stress response. Discovery of the role changes in RNA and lipid metabolism. Neurobiol
of RNA metabolism in tauopathy opens a wide vari- Dis. 2017;106:1–13. https://doi.org/10.1016/j.
ety of novel therapeutic approaches. The abundant nbd.2017.06.008.
7. Becker LA, et  al. Therapeutic reduction of ataxin-2
studies targeting TDP-43-mediated disease demon- extends lifespan and reduces pathology in TDP-43
strate the large variety of approaches designed to mice. Nature. 2017;544(7650):367–71. https://doi.
therapeutically modulate RNA metabolism, and org/10.1038/nature22038.
indicate strong potential for success in disease mod- 8. Bentmann E, Haass C, Dormann D.  Stress granules
in neurodegeneration  – lessons learnt from TAR
ification in models of ALS based on abrogating DNA binding protein of 43 kDa and fused in sar-
TDP-43 mediated disease. These studies comple- coma. FEBS J.  2013;280(18):4348–70. https://doi.
ment the documented success of targeting RNA org/10.1111/febs.12287.
metabolism modify disease progression in animal 9. Bosco DA, et al. Mutant FUS proteins that cause amy-
otrophic lateral sclerosis incorporate into stress gran-
models of tauopathy as well as β-amyloidosis. The ules. Hum Mol Genet. 2010; https://doi.org/10.1093/
success in treating SMA point to a new era for ther- hmg/ddq335.
apy of neurodegenerative diseases, and suggest that 10. Boyd JD, et  al. A high-content screen identi-

the right type of therapy can actually go beyond fies novel compounds that inhibit stress-induced
TDP-43 cellular aggregation and associated cyto-
delay of disease progression to actually halt disease toxicity. J  Biomol Screen. 2014; https://doi.
progression and improve clinical outcomes for org/10.1177/1087057113501553.
some neurodegenerative diseases. Taken together, 11. Brehm MA, et al. Intracellular localization of human
these approaches point to a future in which novel Ins(1,3,4,5,6)P5 2-kinase. Biochem J.  2007; https://
doi.org/10.1042/BJ20070382.
therapeutic approaches might be able to signifi- 12. Buchan JR, Parker R. Eukaryotic stress granules: the
cantly delay disease progression in tauopathy. ins and outs of translation. Mol Cell. 2009;36(6):932–
41. https://doi.org/10.1016/j.molcel.2009.11.020
Acknowledgements  This work was supported by fund- 13. Chang CF, et  al. Therapeutic effect of berberine on
ing to BW from the NIH (AG050471, NS089544, TDP-43-related pathogenesis in FTLD and ALS.  J
AG056318), BrightFocus Foundation, Alzheimer Biomed Sci. 2016;23(1):72. https://doi.org/10.1186/
Association, Cure Alzheimer’s Fund and the Thome s12929-016-0290-z.
Medical Foundation. 14. Dang Y, et  al. Eukaryotic initiation factor 2alpha-­
independent pathway of stress granule induction
by the natural product pateamine A.  J Biol Chem.
Conflict of Interest  BW is co-founder and Chief
2006;281(43):32870–8. https://doi.org/10.1074/jbc.
Scientific Officer of Aquinnah Pharmaceuticals Inc., www.
M606149200
aquinnahpharma.com.
15. de Calignon A, et  al. Propagation of tau pathology
in a model of early Alzheimer’s disease. Neuron.
2012;73(4):685–97. https://doi.org/10.1016/j.
neuron.2011.11.033.
16. Dormann D, et  al. ALS-associated fused in sar-

References coma (FUS) mutations disrupt Transportin-mediated
nuclear import. EMBO J.  2010;29(16):2841–57.
1. Anderson P, Kedersha N. Stress granules: the Tao of https://doi.org/10.1038/emboj.2010.143.
RNA triage. Trends Biochem Sci. 2008;33(3):141– 17. Eftekharzadeh B, Daigle JG, et  al. Tau protein dis-
50. https://doi.org/10.1016/j.tibs.2007.12.003. rupts nucleocytoplasmic transport in Alzheimer’s dis-
2. Apicco DJ et  al. Dysregulation of RNA splicing in ease. Neuron. 2018a;99(5):925–940.e7. https://doi.
tauopathies, Submitted. 2018 org/10.1016/j.neuron.2018.07.039.
3. Apicco DJ, et  al. Reducing the RNA binding pro- 18. Eftekharzadeh B, Tepper K, et al. Tau protein liquid–
tein TIA1 protects against tau-mediated neurode- liquid phase separation can initiate tau aggregation.
generation in  vivo. Nat Neurosci. 2018; https://doi. EMBO J. 2018b;37:e98049. https://doi.org/10.15252/
org/10.1038/s41593-017-0022-z. embj.201798049.
370 A. Cruz et al.

19. Elden AC, et al. Ataxin-2 intermediate-length polyglu- of neurodegeneration. Neuron. 2010;68(6):1067–81.
tamine expansions are associated with increased risk https://doi.org/10.1016/j.neuron.2010.11.030.
for ALS.  Nature. 2010;466(7310):1069–75. https:// 34. Hughes MP, et al. Atomic structures of low-­complexity
doi.org/10.1038/nature09320 protein segments reveal kinked beta sheets that
20. Feric M, et  al. Coexisting liquid phases underlie
assemble networks. Science. 2018;359(6376):698–
nucleolar subcompartments. Cell. 2016;165(7):1686– 701. https://doi.org/10.1126/science.aan6398.
97. https://doi.org/10.1016/j.cell.2016.04.047. 35. Ikezu T, et al. Tau phosphorylation is impacted by rare
21. Finkel RS, et  al. Treatment of infantile-onset spi-
AKAP9 mutations associated with Alzheimer disease
nal muscular atrophy with nusinersen: a phase in African Americans. J  Neuroimmune Pharmacol.
2, open-label, dose-escalation study. Lancet. 2018; https://doi.org/10.1007/s11481-018-9781-x.
2016;388(10063):3017–26. https://doi.org/10.1016/ 36. Ittner LM, et  al. Dendritic function of tau mediates
S0140-6736(16)31408-8. amyloid-beta toxicity in Alzheimer’s disease mouse
22. Finkel RS, et  al. Nusinersen versus sham control in models. Cell. 2010;142(3):387–97. https://doi.
infantile-onset spinal muscular atrophy. N Engl J Med. org/10.1016/j.cell.2010.06.036.
2017;377(18):1723–32. https://doi.org/10.1056/ 37. Jiang L, et  al. TIA1 regulates the generation and

NEJMoa1702752. response to toxic tau oligomers. Acta Neuropathol.
23. Gennarelli M, et  al. Survival motor neuron gene
2018; https://doi.org/10.1007/s00401-018-1937-5.
transcript analysis in muscles from spinal muscular 38. Johnson BS, et al. TDP-43 is intrinsically aggregation-­
atrophy patients. Biochem Biophys Res Commun. prone, and amyotrophic lateral sclerosis-linked muta-
1995;213(1):342–8. https://doi.org/10.1006/bbrc. tions accelerate aggregation and increase toxicity.
1995.2135 J  Biol Chem. 2009;284(30):20329–39. https://doi.
24. Geschwind DH, et al. The prevalence and wide clinical org/10.1074/jbc.M109.010264
spectrum of the spinocerebellar ataxia type 2 trinucle- 39. Kaehler C, et al. Ataxin-2-like is a regulator of stress
otide repeat in patients with autosomal dominant cer- granules and processing bodies. PLoS One. 2012;
ebellar ataxia. Am J Hum Genet. 1997;60(4):842–50. https://doi.org/10.1371/journal.pone.0050134.
25. Gilks N, et  al. Stress granule assembly is mediated 40. Kapeli K, Martinez FJ, Yeo GW.  Genetic muta-

by prion-like aggregation of TIA-1. Mol Biol Cell. tions in RNA-binding proteins and their roles in
2004;15(12):5383–98. https://doi.org/10.1091/mbc. ALS. Hum Genet. 2017;136(9):1193–214. https://doi.
e04-08-0715 org/10.1007/s00439-017-1830-7.
26. Gunawardana CG, et al. The human tau interactome: 41. Kedersha N, et al. Dynamic shuttling of TIA-1 accom-
binding to the ribonucleoproteome, and impaired panies the recruitment of mRNA to mammalian stress
binding of the proline-to-leucine mutant at position granules. J  Cell Biol. 2000;151(6):1257–68. https://
301 (P301L) to chaperones and the proteasome. Mol doi.org/10.1083/jcb.151.6.1257
Cell Proteomics. 2015;14(11):3000–14. https://doi. 42. Kedersha N, et  al. Evidence that ternary complex

org/10.1074/mcp.M115.050724. (eIF2-GTP-tRNA(i)(Met))-deficient preinitiation
27. Guo C, et  al. Tau activates transposable elements in complexes are core constituents of mammalian stress
Alzheimer’s disease. Cell Rep. 2018a;23(10):2874– granules. Mol Biol Cell. 2002;13(1):195–210. https://
80. https://doi.org/10.1016/j.celrep.2018.05.004. doi.org/10.1091/mbc.01-05-0221.
28. Guo L, et al. Nuclear-import receptors reverse aber- 43. Kim HJ, et  al. Therapeutic modulation of eIF2α

rant phase transitions of RNA-binding proteins with phosphorylation rescues TDP-43 toxicity in amyo-
prion-like domains. Cell. 2018b;173(3):677–692.e20. trophic lateral sclerosis disease models. Nat Genet.
https://doi.org/10.1016/j.cell.2018.03.002. 2014;46(2):152–60. https://doi.org/10.1038/ng.2853.
29. Halliday M, et al. Partial restoration of protein synthesis 44. Li W, et  al. Cell proteins TIA-1 and TIAR interact
rates by the small molecule ISRIB prevents neurode- with the 3 ′ stem-loop of the West Nile virus comple-
generation without pancreatic toxicity. Cell Death Dis. mentary minus-strand RNA and facilitate virus repli-
2015;6:e1672. https://doi.org/10.1038/cddis.2015.49. cation. J Virol. 2002;76(23):11989–2000. https://doi.
30. Han TW, et al. Cell-free formation of RNA granules: org/10.1128/JVI.76.23.11989.
bound RNAs identify features and components of cel- 45. Li YR, et  al. Stress granules as crucibles of ALS
lular assemblies. Cell. 2012;149(4):768–79. https:// pathogenesis. J  Cell Biol. 2013;201(3):361–72.
doi.org/10.1016/j.cell.2012.04.016. https://doi.org/10.1083/jcb.201302044.
31. He Y, Smith R.  Nuclear functions of heterogeneous 46. Lin Y, et  al. Formation and maturation of phase-­

nuclear ribonucleoproteins A/B.  Cell Mol Life separated liquid droplets by RNA-binding pro-
Sci. 2009;66(7):1239–56. https://doi.org/10.1007/ teins. Mol Cell. 2015;60(2):208–19. https://doi.
s00018-008-8532-1. org/10.1016/j.molcel.2015.08.018.
32. Hofweber M, et al. Phase separation of FUS is sup- 47. Liu L, et  al. Trans-synaptic spread of tau pathology
pressed by its nuclear import receptor and arginine in  vivo. PLoS One. 2012;7(2):e31302. https://doi.
methylation. Cell. 2018;173(3):706–719 e13. https:// org/10.1371/journal.pone.0031302.
doi.org/10.1016/j.cell.2018.03.004. 48. Liu-Yesucevitz L, et al. Tar DNA binding protein-43
33. Hoover BR, et  al. Tau mislocalization to dendritic (TDP-43) associates with stress granules: analysis
spines mediates synaptic dysfunction independently of cultured cells and pathological brain tissue. PLoS
26  The Pathophysiology of Tau and Stress Granules in Disease 371

One. 2010;5(10):e13250. https://doi.org/10.1371/ 62. Mercuri E, et  al. Nusinersen versus sham con-

journal.pone.0013250. trol in later-onset spinal muscular atrophy. N Engl
49.
Liu-Yesucevitz L, et  al. Local RNA transla- J Med. 2018;378(7):625–35. https://doi.org/10.1056/
tion at the synapse and in disease. J  Neurosci. NEJMoa1710504.
2011;31(45):16086–93. https://doi.org/10.1523/ 63. Mills JD, et  al. RNA-Seq analysis of the parietal

JNEUROSCI.4105-11.2011. cortex in Alzheimer’s disease reveals alternatively
50. Lorson CL, et al. A single nucleotide in the SMN gene spliced isoforms related to lipid metabolism. Neurosci
regulates splicing and is responsible for spinal mus- Lett. 2013;536:90–5. https://doi.org/10.1016/j.
cular atrophy. Proc Natl Acad Sci. 1999; https://doi. neulet.2012.12.042.
org/10.1073/pnas.96.11.6307. 64. Mokas S, et  al. Uncoupling stress granule assem-
51. Lourenco MV, et  al. TNF-alpha mediates PKR-­
bly and translation initiation inhibition. Mol Biol
dependent memory impairment and brain IRS-1 inhi- Cell. 2009;20(11):2673–83. doi: E08-10-1061
bition induced by Alzheimer’s beta-amyloid oligomers [pii]10.1091/mbc.E08-10-1061.
in mice and monkeys. Cell Metab. 2013;18(6):831– 65. Molliex A, et al. Phase separation by low complexity
43. https://doi.org/10.1016/j.cmet.2013.11.002. domains promotes stress granule assembly and drives
52. Ma T, et al. Suppression of eIF2alpha kinases allevi- pathological fibrillization. Cell. 2015;163(1):123–33.
ates Alzheimer’s disease-related plasticity and mem- https://doi.org/10.1016/j.cell.2015.09.015.
ory deficits. Nat Neurosci. 2013;16(9):1299–305. 66. Moreno JA, et  al. Sustained translational repres-

https://doi.org/10.1038/nn.3486. sion by eIF2α-P mediates prion neurodegeneration.
53. Mackenzie IR, et al. TIA1 mutations in amyotrophic Nature. 2012; https://doi.org/10.1038/nature11058.
lateral sclerosis and frontotemporal dementia promote 67. Moujalled D, et al. Kinase inhibitor screening identi-
phase separation and Alter stress granule dynam- fies cyclin-dependent kinases and glycogen ­synthase
ics. Neuron. 2017;95(4):808–816.e9. https://doi. kinase 3 as potential modulators of TDP-43 cyto-
org/10.1016/j.neuron.2017.07.025. solic accumulation during cell stress. PLoS One.
54. Markmiller S, et al. Context-dependent and disease-­ 2013;8(6):e67433. https://doi.org/10.1371/journal.
specific diversity in protein interactions within stress pone.0067433.
granules. Cell. 2018;172(3):590–604.e13. https://doi. 68. Muto NF, et al. Inhibition of replication of reactivated
org/10.1016/j.cell.2017.12.032. human immunodeficiency virus type 1 (HIV-1) in
55. Maziuk BF, et  al. RNA binding proteins co-­localize latently infected U1 cells transduced with an HIV-1
with small tau inclusions in tauopathy. Acta long terminal repeat-driven PKR cDNA construct.
Neuropathol Commun. 2018;6(1):71. https://doi. J  Virol. 1999;73(11):9021–8. Available at: https://
org/10.1186/s40478-018-0574-5. www.ncbi.nlm.nih.gov/pubmed/10516008
56. Mazroui R, et al. Inhibition of ribosome recruitment 69. Neumann M, et al. Ubiquitinated TDP-43 in fronto-
induces stress granule formation independently of temporal lobar degeneration and amyotrophic lateral
eukaryotic initiation factor 2alpha phosphorylation. sclerosis. Science. 2006;314(5796):130–3. https://
Mol Biol Cell. 2006;17(10):4212–9. https://doi. doi.org/10.1126/science.1134108
org/10.1091/mbc.e06-04-0318 70. Neumann M, et  al. Phosphorylation of S409/410

57. McEwen E, et  al. Heme-regulated inhibitor kinase-­ of TDP-43 is a consistent feature in all sporadic
mediated phosphorylation of eukaryotic translation and familial forms of TDP-43 proteinopathies.
initiation factor 2 inhibits translation, induces stress Acta Neuropathol. 2009; https://doi.org/10.1007/
granule formation, and mediates survival upon arse- s00401-008-0477-9
nite exposure. J Biol Chem. 2005;280(17):16925–33. 71. Nonaka T, et  al. Truncation and pathogenic muta-
https://doi.org/10.1074/jbc.M412882200 tions facilitate the formation of intracellular aggre-
58. McGurk L, et  al. Poly(ADP-ribose) prevents patho- gates of TDP-43. Hum Mol Genet. 2009; https://doi.
logical phase separation of TDP-43 by promoting liq- org/10.1093/hmg/ddp275.
uid demixing and stress granule localization. Mol Cell. 72. Nonhoff U, et al. Ataxin-2 interacts with the DEAD/H-­
2018;71(5):703–717.e9. https://doi.org/10.1016/j. box RNA helicase DDX6 and interferes with P-bodies
molcel.2018.07.002. and stress granules. Mol Biol Cell. 2007;18(4):1385–
59. Meier S, et al. Identification of novel tau interactions 96. https://doi.org/10.1091/mbc.e06-12-1120.
with endoplasmic reticulum proteins in Alzheimer’s 73. Oddo S, et  al. Reduction of soluble Aβ and tau,

disease brain. J Alzheimers Dis. 2015;48(3):687–702. but not soluble Aβ alone, ameliorates cognitive
https://doi.org/10.3233/JAD-150298. decline in transgenic mice with plaques and tangles.
60. Meier S, et  al. Pathological tau promotes neuronal J  Biol Chem. 2006;281(51):39413–23. https://doi.
damage by impairing ribosomal function and decreas- org/10.1074/jbc.M608485200.
ing protein synthesis. J  Neurosci. 2016; https://doi. 74. Patel A, et al. A liquid-to-solid phase transition of the
org/10.1523/JNEUROSCI.3029-15.2016. ALS protein FUS accelerated by disease mutation.
61. Mendell JR, et  al. Single-dose gene-replacement
Cell. 2015;162(5):1066–77. https://doi.org/10.1016/j.
therapy for spinal muscular atrophy. N Engl J  Med. cell.2015.07.047.
2017;377(18):1713–22. https://doi.org/10.1056/ 75. Prilusky J, et  al. FoldIndex: a simple tool to predict
NEJMoa1706198. whether a given protein sequence is intrinsically
372 A. Cruz et al.

unfolded. Bioinformatics. 2005;21(16):3435–8. 2012;123(6):787–805. https://doi.org/10.1007/

https://doi.org/10.1093/bioinformatics/bti537. s00401-012-0987-3.
76. Protter DSW, Parker R.  Principles and properties of 87. Vanderweyde T, et  al. Contrasting pathology of

stress granules. Trends Cell Biol. 2016;26(9):668–79. the stress granule proteins TIA-1 and G3BP in
https://doi.org/10.1016/j.tcb.2016.05.004. Tauopathies. J  Neurosci. 2012;32(24):8270–83.
77. Ramaswami M, Taylor JP, Parker R. Altered ribosta- https://doi.org/10.1523/JNEUROSCI.1592-12.2012.
sis: RNA-protein granules in degenerative disorders. 88. Vanderweyde T, et  al. Interaction of tau with the

Cell. 2013;154(4):727–36. https://doi.org/10.1016/j. RNA-binding protein TIA1 regulates tau patho-
cell.2013.07.038. physiology and toxicity. Cell Rep. 2016; https://doi.
78. Robberecht W, Eykens C.  The genetic basis of
org/10.1016/j.celrep.2016.04.045.
amyotrophic lateral sclerosis: recent breakthroughs. 89. Wang Y, Mandelkow E. Tau in physiology and pathol-
Adv Genomics Genet. 2015;5:327–45. https://doi. ogy. Nat Rev Neurosci. 2016;17(1):5–21. https://doi.
org/10.2147/AGG.S57397. org/10.1038/nrn.2015.1.
79. Sanders DW, et  al. Distinct tau prion strains propa- 90. Wang P, et al. Tau interactome mapping based iden-
gate in cells and mice and define different tauopa- tification of Otub1 as Tau deubiquitinase involved
thies. Neuron. 2014;82(6):1271–88. https://doi. in accumulation of pathological Tau forms in  vitro
org/10.1016/j.neuron.2014.04.047. and in vivo. Acta Neuropathol. 2017;133(5):731–49.
80. SantaCruz K. Tau suppression in a neurodegenerative https://doi.org/10.1007/s00401-016-1663-9.
mouse model improves memory function. Science. 91. Wang J, et  al. A molecular grammar governing the
2005;309(5733):476–81. https://doi.org/10.1126/ driving forces for phase separation of prion-like RNA
science.1113694. binding proteins. Cell. 2018;174(3):688–699.e16.
81. Scheckel C, et  al. Regulatory consequences of neu- https://doi.org/10.1016/j.cell.2018.06.006.
ronal ELAV-like protein binding to coding and non-­ 92. Wei MT, et  al. Phase behaviour of disordered pro-
coding RNAs in human brain. Elife. 2016;5 https:// teins underlying low density and high permeability
doi.org/10.7554/eLife.10421. of liquid organelles. Nat Chem. 2017;9(11):1118–25.
82. Silva JM, et al. Dysregulation of autophagy and stress https://doi.org/10.1038/nchem.2803.
granule-related proteins in stress-driven Tau pathol- 93. Wek RC, Jiang H-Y, Anthony TG. Coping with stress:
ogy. Cell Death Differ. 2018;26(8):1411–27. https:// eIF2 kinases and translational control. Biochem Soc
doi.org/10.1038/s41418-018-0217-1. Trans. 2006; https://doi.org/10.1042/BST20060007.
83. Sreedharan J, et  al. TDP-43 mutations in familial
94. Winton MJ, et  al. Disturbance of nuclear and cyto-
and sporadic amyotrophic lateral sclerosis. Science. plasmic TAR DNA-binding protein (TDP-43) induces
2008;319(5870):1668–72. https://doi.org/10.1126/ disease-like redistribution, sequestration, and aggre-
science.1154584 gate formation. J  Biol Chem. 2008; https://doi.
84. Stamer K, et al. Tau blocks traffic of organelles, neuro- org/10.1074/jbc.M800342200.
filaments, and APP vesicles in neurons and enhances 95. Zempel H, Mandelkow E. Lost after translation: mis-
oxidative stress. J  Cell Biol. 2002;156(6):1051–63. sorting of Tau protein and consequences for Alzheimer
https://10.1083/jcb.200108057jcb.200108057 [pii] disease. Trends Neurosci. 2014;37(12):721–32.
85. Thomas M, Alegre-Abarrategui J, Wade-Martins
https://doi.org/10.1016/j.tins.2014.08.004.
R.  RNA dysfunction and aggrephagy at the centre 96. Zempel H, et al. Amyloid-beta oligomers induce syn-
of an amyotrophic lateral sclerosis/frontotemporal aptic damage via Tau-dependent microtubule severing
dementia disease continuum. Brain. 2013;136(Pt by TTLL6 and spastin. EMBO J. 2013;32(22):2920–
5):1345–60. https://doi.org/10.1093/brain/awt030. 37. https://doi.org/10.1038/emboj.2013.207.
86. Van der Jeugd A, et  al. Cognitive defects are
97. Zhang X, et  al. RNA stores tau reversibly in com-
reversible in inducible mice expressing pro-aggre- plex coacervates. PLoS Biol. 2017;15(7):e2002183.
gant full-length human Tau. Acta Neuropathol. https://doi.org/10.1371/journal.pbio.2002183.
 Tau Oligomers
27
Sumihiro Maeda and Akihiko Takashima

Prediction of Non-fibrillar loss without NFT formation [6]. Therefore, over-

Aggregates of Tau Protein expression of tau, at least, induces neuronal tox-
icity in an animal model, but the toxic species of
Neurofibrillary tangles (NFTs), a pathological tau was not the tau filament itself, but something
hallmark of Alzheimer’s disease (AD), are com- produced in the process of tau aggregation. In
posed of filamentous polymers of tau protein [1]. that sense, granular tau oligomers, the intermedi-
NFTs are insoluble and resistant to proteases. ate form of tau filament, meet this criterion for
Thus, even after neurons are lost, NFTs remain as the toxic tau species.
“tombstones” of the NFT-bearing neurons and
are called ghost tangles. In AD, the number of
neurons lost and the number of NFT/ghost tan-  au Oligomer as an Intermediate
T
gles should be the same if all of the neurons were Species of Tau Filament
lost after NFT formation. But they aren’t. More
neurons are lost than ghost tangles remain [2, 3]. Tau is highly hydrophilic and does not aggregate
Although the number of NFTs, neurons lost, and by itself. The core of the tau filament, the micro-
the severity of the disease all correlate well with tubule binding domain (MBD), is positively
each other [3], several findings point to a missing charged, and that charge prevents intermolecular
element. NFT formation and neuronal loss have interactions of tau. Under physiological condi-
been reported to be distinct events. Suppression tions, the positively charged residues of tau inter-
of human tau (hTau) expression in an hTau trans- act with negatively charged residues of tubulin
genic mouse model did not block tau filament [1]. Under pathological conditions, tau is highly
formation but reduced neuronal loss [4], and phosphorylated, and the negative charge of phos-
NFT-bearing neurons were functionally intact in phor residues is thought to neutralize the positive
hTau transgenic mice [5]. In an hTau Drosophila change of the MBD, induce detachment of tau
model, hTau over-expression induced neuronal from tubulin, and allow tau-tau interactions [1].
Polyanionic compounds, such as RNA or hep-
S. Maeda (*) arin, induce aggregation of recombinant tau
Department of Physiology, Keio University School of probably by neutralizing the positive charges of
Medicine, Tokyo, Japan tau [7–9]. Lipids, such as arachidonic acid, also
e-mail: sumihiro.maeda@keio.jp induce tau aggregation above the critical micelle
A. Takashima concentration because the surfaces of lipid
Department of Life Science, Gakushuin University, micelles are negatively charged [10]. In an
Tokyo, Japan

© Springer Nature Singapore Pte Ltd. 2019 373

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_27
374 S. Maeda and A. Takashima

in  vitro tau aggregation assay, non-filamentous and nuclear magnetic resonance analysis [16,
tau aggregates called granular tau oligomers were 17]. In human brains, the number of tau oligo-
found [11]. The non-filamentous tau aggregates mers was increased in the frontal cortex of Braak
could be separated from tau monomers and fila- stage I patients, when NFTs have yet to form.
ments by sucrose gradient centrifugation Thus, tau oligomers form before NFTs [12–14].
(Fig.  27.1). Tau filaments do not form in  vitro
without aggregation inducers. However, simply
concentrating tau oligomers induces tau filament Various Types of Non-filamentous
formation without additional heparin, and under Tau Aggregates
atomic force microscopy (AFM), filaments
formed in vitro or purified and AD brain degraded Definition and preparation of tau oligomers vary
to granular tau structures [11] (Fig.  27.2). This among researchers. For example, Maeda et  al.
suggested that the tau oligomer is an intermediate induced tau aggregation by mixing recombinant
species of the tau filament. In addition, tau oligo- tau with heparin and defined tau oligomers as
mers were found in animal models and human granular structures under AFM [11]. The Davies
brains [12–14] (Fig.  27.3). The amino acid group reported non-filamentous tau species that
sequences critical for tau filament formation, were extracted by a conformation-dependent
PHF6 and PHF6∗ [15], are also reported to be key antibody, MC1 [18]. The non-filamentous tau
for tau oligomerization by western blots (WBs) species formed filaments after concentration,

Fig. 27.1  The purification of granular tau oligomers (c) are shown. Granular tau oligomers could be recovered
using sucrose step gradient centrifugation. (a) Tau pro- in fraction 3 as indicated (b, c). In fraction 1, tau mono-
teins were aggregated in vitro and layered onto 20–50% mers and multimers were collected (b), but no aggregated
sucrose step gradients and centrifuged at 200,000 × G for structure were observed under AFM (c). The amount of
2  h to separate non-aggregated tau, granular tau oligo- tau in fraction 3 was less than that in fraction 1 (b), but
mers, and tau filaments. (b, c) Staining with Coomassie granular structures were detected under AFM (c). Longer
brilliant blue (CBB) (b) and AFM images of all fractions filaments were collected in fractions 4–6 (c) [11]
27  Tau Oligomers 375

Fig. 27.2  The formation and degradation of tau fila- age by AFM tips. Continuous AFM observation revealed
ments. (a) Granular tau oligomers in fraction 3 and non-­ that mechanical damage degraded tau filaments formed
aggregated tau in fraction 1 were concentrated without in vitro (b) or purified from an AD brain (c) to granular tau
adding heparin. Filaments were observed by AFM in con- structures. The numbers at the top right indicate the order
centrated fraction 3 (C3) but not in the concentrated frac- of the images. It took about 9 min to take an image [11].
tion 1 (C1). Blue arrows indicate the generated filaments Blue arrows indicate the breaking points of the tau
in C3. (b, c) Structures were exposed to mechanical dam- filament

indicating that the oligomers extracted here are Sf9 cells) because highly phosphorylated tau pro-
identical with the oligomers described above. teins can be extracted, whereas recombinant tau
Berger et al. reported a correlation of tau multim- protein produced in E. coli had no post-­
ers and behavioral scores in an hTau transgenic translational modifications [20]. They confirmed
mouse model [19]. Multimers were defined by that phosphorylation enhanced tau aggregation
the molecular mass by WB (140- and 170-kDa [20, 21]. The Kayed group generated tau oligo-
bands). The multimer can be a component of mers by cross-seeding with amyloid beta (Aβ)
higher order aggregates, such as globular oligo- oligomers [22]. Those oligomers induced cell
mers or tau filaments, but it degrades to a multi- death when added exogenously to cells [22, 23].
meric form under WB denaturing conditions. Anti-tau oligomer antibodies were generated
Thus, one component of tau oligomers may be using the cross-seeded tau oligomers, and the
the 140-/170-kDa multimers. Multimers might antibodies blocked tau-induced neurodegenera-
also be complexes of tau and other molecules, but tion in hTau transgenic mouse models that lack
another group reported that similar multimers Aβ aggregates and could not form Aβ seeds [24,
may form only from recombinant tau [16]. 25]. Thus, tau aggregates, like tau oligomers
Mandelkow’s group induced tau oligomers using seeded by Aβ aggregates, can be induced in the
recombinant protein produced in insect cells (i.e., absence of Aβ aggregates. Binder’s group [26]
376 S. Maeda and A. Takashima

Fig. 27.3  Purification

of granular tau
oligomers from human
brains. (a) Granular tau
oligomers in human
brains were purified in a
combination of
immune-affinity column
chromatography with a
pan-tau antibody and
sucrose step gradient
centrifugation [12]. (b)
AD brains (Braak stage
V) contained more
granular tau oligomers
than control brains
(Braak stage 0) [12].
Inserts show the
magnified images of
granular tau oligomers
purified from AD or
control brains

stabilized tau aggregates with a chemical cross generation conformation-dependent antibody

linker and found that, by WB, the tau dimer is a that was raised against tau aggregates purified
component of tau aggregates and that cross-­ with an Alz50 immunoaffinity column. MC1
linked tau dimers form short filaments and oligo- reacts with aggregated tau species more specifi-
meric globular structures but not long filaments. cally than Alz50 [27, 29] and detects both tau
filaments and oligomers [2]. With a combination
of sucrose step gradients and the MC1 antibody,
Tau Oligomer Detection Methods oligomeric species can be detected even in mouse
EM and Oligomer Antibodies tissue lysates [14]. However, the requirement for
sucrose step gradient centrifugation limits the
Granular tau oligomers can be detected by AFM ability to do mechanistic studies of tau oligomer-­
and electron microscopy (EM) [11, 22, 26], but dependent pathogenesis.
not by other methods. To expand their ability to Researchers next sought to distinguish tau
study tau, researchers generated antibodies spe- oligomers from tau filaments without gradient
cific for tau oligomers. centrifugation. They generated more specific
The Davies group generated two conformation-­ antibodies against tau oligomers than Alz50 and
dependent antibodies, Alz50 and MC1 [27], that MC1 antibodies. Binder’s group found the 180-­
react with tau filaments purified from AD brains. kDa tau species in AD brain lysates [26]. To
Alz50 recognizes discontinuous sequences at the examine the pathogenesis of this tau species, they
N-terminus and MBD. During abnormal confor- cross-linked tau dimers to obtain stable lower-­
mation changes, inter- or intramolecular attach- order tau aggregates. When incubated with ara-
ment of the N-terminus to the MBD may generate chidonic acid, cross-linked tau dimers formed
the Alz50 epitope [28]. MC1 is a second-­ granular tau oligomer but not long fibrils. The
27  Tau Oligomers 377

researchers immunized tau-knockout mice with  au Oligomerization Enhancers,

T
the cross-linked tau dimers and obtained the Tau Blockers, and the Toxic Mechanism
Oligomeric Component-1 (TOC1) antibody [26].
TOC1 preferentially labels granular tau oligo- Not all FTDP-17 mutations increase tau filament
mers and the end of tau filaments, supporting the formation [36, 37]. However, all FTDP-17 tau
idea that tau oligomers are an intermediate spe- mutations examined so far for tau oligomeriza-
cies of tau filament and the attachment of tau tion enhanced tau oligomerization. Notably, the
oligomer to tau filament will elongate the fila- P301L mutation decreases the size and increases
ment [11]. Unlike another tau-conformation the number of tau oligomers [14].
dependent antibody, Alz50 that recognizes dis- Tau aggregates (mainly filaments) have been
continuous tau sequences, amino acids 2–10 and suggested to spread from one brain region to
312–342, TOC1 recognizes amino acids 209–224 another trans-synaptically [38]. The spreading
of tau. may be mediated by the tau aggregates them-
TNT1 is another conformation-dependent selves, which are called prion-like tau species,
antibody generated by the Binder group [30]. Its but not by the dysfunction of projecting neurons
epitope is the tau phosphatase-activating domain [38]. Although it is not clear that the soluble frac-
(PAD) in the N-terminal domain. The N-terminus tion has prion-like activity [39, 40], tau oligo-
attaches to the MBD under physiological condi- mers may mediate the propagation [41]. They
tions. It is detached by aggregation or phosphor- may function as a template for newly formed tau
ylation and impairs axonal transport via oligomers that can be amplified by adding non-­
phosphatase activation, GSK-3β activation, and aggregated tau [22]. Thus, the tau oligomer itself
kinesin light chain phosphorylation [30, 31]. As may increase tau oligomer numbers.
with all other tau oligomer antibodies, reactivity Heat shock proteins (HSPs) are involved in
depends on conformation. Notably, the PAD tau oligomerization [42]. HSPs, including Hsp90
sequence lacks the endogenous mouse tau and 27 that block tau aggregation, are inversely
sequence, and thus, TNT1 cannot detect mouse correlated with tau oligomer levels in human
tau even though mouse tau aggregates like brains [16]. Thus, HSPs are an intriguing target
human tau [32]. Both TOC1 and TNT1 are pan- for therapies to prevent tau pathogenesis [43, 44].
tau antibodies in WB because their epitopes are Other small molecules also block tau oligomer-
freely accessible even under denatured condi- ization. A screen of a library of natural compound
tions [33]. derivatives for tau-binding compounds revealed
The Kayed group raised antibodies against the several positives. For example,
cross-seeding oligomeric species of tau and 1,2-­dihydroxybenzene blocked tau oligomeriza-
obtained the T22 (rabbit polyclonal) and TOMA tion, and DL-isoproterenol reduced tau aggrega-
(mouse monoclonal) antibodies [25, 34]. For tion and neuronal cell loss in human tau transgenic
unknown reasons, the epitopes of these antibod- mice [45]. These findings also support the idea
ies are preserved even in WB, whereas other that tau oligomer is a culprit of tau-mediated
conformation-­dependent antibodies function as pathogenesis.
pan-tau antibodies in WB because all sequences Tau aggregates accumulate inside of neurons.
of tau protein are accessible to the antibodies. Thus, to assess their toxicity, tau oligomers must
Normal tau immunization induced neurologic be introduced into neurons. However, tau oligo-
deficits in normal mice [35]. However, immuni- mers are thought to be incorporated into cells [46]
zation with these antibodies blocked tau-induced and to induce synaptotoxicity and Ca dysregula-
neuronal dysfunction in hTau transgenic mice as tion [20, 47]. Surprisingly, the synaptotoxicity
mentioned above, suggesting that granular tau may be separate from cell viability, even though
oligomers, not tau monomers, should be explored neuronal cell loss correlates well with the NFT
as therapeutic targets for tauopathies [24, 25]. formation as mentioned above. Ca ­dysregulation
378 S. Maeda and A. Takashima

by tau oligomers may be mediated by M1 and M3 References

muscarinic receptors [48], even though aggre-
gated tau species showed less toxicity than mono- 1. Morris M, Maeda S, Vossel K, Mucke L. The many
faces of tau. Neuron. 2011;70:410–26.
meric non-aggregated tau in that system. 2. Ihara Y.  PHF and PHF-like fibrils--cause or conse-
The mechanism of tau toxicity is unknown, quence? Neurobiol Aging [Internet]. 2001;22(1):123–
but a hint might be provided by its conformation. 6. Available from: http://www.ncbi.nlm.nih.gov/
Tau is reported to have a paper-clip conforma- entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt
=Citation&list_uids=11164285
tion: the N-terminus is bent over to MBD in 3. Gómez-Isla T, Hollister R, West H, Mui S, Growdon
native state [49]. Tau aggregation disrupt that JH, Petersen RC, et al. Neuronal loss correlates with
conformation and exposes the N-terminus. The but exceeds neurofibrillary tangles in Alzheimer’s dis-
exposure allows the PAD domain to induce the ease. Ann Neurol. 1997;41(1):17–24.
4. Santacruz K, Lewis J, Spires T, Paulson J, Kotilinek
impairment of axonal trafficking [30]. L, Ingelsson M, et  al. Tau suppression in a neuro-
degenerative mouse model improves memory func-
tion. Science [Internet]. 2005;309(5733):476–81.
Conclusion Available from: http://www.ncbi.nlm.nih.gov/
pubmed/16020737%5Cn, http://www.pubmedcentral.
nih.gov/articlerender.fcgi?artid=PMC1574647
Tau aggregation is toxic to cells. Thus, many 5. Kuchibhotla KV, Wegmann S, Kopeikina KJ, Hawkes
researchers have visualized tau aggregation in J, Rudinskiy N, Andermann ML, et al. Neurofibrillary
their experimental models to search for a com- tangle-bearing neurons are functionally integrated
in cortical circuits in  vivo. Proc Natl Acad Sci.
mon feature between AD patients and the mod- 2014;111:510–4.
els. The exact mechanism of tau aggregation and 6. Wittmann CW, Wszolek MF, Shulman JM, Salvaterra
toxicity and the targets of tau aggregates are still PM, Lewis J, Hutton M, et  al. Tauopathy in
open issues. The discrepancy of tau filaments Drosophila: neurodegeneration without neurofibril-
lary tangles. Science. 2001;293(5530):711–4.
from neuronal cell loss implicates tau oligomers 7. Kampers T, Friedhoff P, Biernat J, Mandelkow
as the culprit of tau-mediated pathogenesis. EM, Mandelkow E.  RNA stimulates aggregation of
Researchers have characterized tau oligomers by microtubule-associated protein tau into Alzheimer-­
various methods. However, to directly examine like paired helical filaments. FEBS Lett [Internet].
1996;399(3):344–9. Available from: http://www.ncbi.
their toxicity mechanisms, we will need methods nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=Pu
to enhance or block tau oligomerization inside bMed&dopt=Citation&list_uids=8985176
neurons. Also, to develop drugs that target tau 8. Goedert M, Jakes R, Spillantini MG, Hasegawa M,
oligomers, we will need tau oligomer probes that Smith MJ, Crowther RA.  Assembly of microtubule-­
associated protein tau into Alzheimer-like filaments
can be used for the target engagement in human induced by sulphated glycosaminoglycans. Nature
live imaging. Antibodies specific for tau oligo- [Internet]. 1996;383(6600):550–3. Available from:
mers were used to show that tau oligomers have a http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?
structure distinct from other tau species. Thus, it cmd=Retrieve&db=PubMed&dopt=Citation&l
ist_uids=8849730
might be possible to develop tau oligomer–spe- 9. Perez M, Valpuesta JM, Medina M, Montejo de Garcini
cific probes for positron emission tomography. E, Avila J. Polymerization of tau into filaments in the
Those probes would be a powerful tool in combi- presence of heparin: the minimal sequence required
nation with similar probes for tau filaments [50]. for tau-tau interaction. J  Neurochem [Internet].
1996;67(3):1183–90. Available from: http://www.
ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&d
Acknowledgement  The authors thank Dr. Gary Howard b=PubMed&dopt=Citation&list_uids=8752125
for editorial review. This research was partially supported 10. Chirita CN, Necula M, Kuret J. Anionic micelles and
by grants from the Cyclic Innovation for Clinical vesicles induce tau fibrillization in vitro. J Biol Chem.
Empowerment (CiCLE) and the Acceleration Program for 2003;278(28):25644–50. Available from: http://www.
Intractable Diseases Research Utilizing Disease-specific ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&d
iPS Cells from the Japan Agency for Medical Research b=PubMed&dopt=Citation&list_uids=12730214
and Development (AMED) and JSPS to SM. 11. Maeda S, Sahara N, Saito Y, Murayama M, Yoshiike Y,
Kim H, et al. Granular tau oligomers as intermediates
of tau filaments. Biochemistry. 2007;46(12):3856–61.
27  Tau Oligomers 379

12. Maeda S, Sahara N, Saito Y, Murayama S, Ikai A, 24. Castillo-Carranza DL, Sengupta U, Guerrero-Muñoz
Takashima A. Increased levels of granular tau oligo- MJ, Lasagna-Reeves CA, Gerson JE, Singh G, et al.
mers: an early sign of brain aging and Alzheimer’s Passive immunization with Tau oligomer monoclo-
disease. Neurosci Res. 2006;54(3):197–201. nal antibody reverses tauopathy phenotypes with-
13. Lasagna-Reeves CA, Castillo-Carranza DL, Sengupta out affecting hyperphosphorylated neurofibrillary
U, Guerrero-Munoz MJ, Kiritoshi T, Neugebauer V, tangles. J  Neurosci [Internet]. 2014;34(12):4260–
et al. Alzheimer brain-derived tau oligomers propagate 72. Available from: http://www.ncbi.nlm.nih.gov/
pathology from endogenous tau. Sci Rep [Internet]. pubmed/24647946
2012;2:700. Available from: http://www.nature.com/ 25. Castillo-Carranza DL, Gerson JE, Sengupta U,

srep/2012/121003/srep00700/full/srep00700.html Guerrero-Muñoz MJ, Lasagna-Reeves CA, Kayed
14. Maeda S, Sato Y, Takashima A.  Frontotemporal
R.  Specific targeting of tau oligomers in Htau mice
dementia with Parkinsonism linked to chromosome- prevents cognitive impairment and tau toxicity follow-
17 mutations enhance tau oligomer formation.
­ ing injection with brain-derived tau oligomeric seeds.
Neurobiol Aging. 2018;69:26–32. J Alzheimers Dis. 2014;40(Suppl 1):S97–S111.
15. Von Bergen M, Barghorn S, Li L, Marx A, Biernat 26. Patterson KR, Remmers C, Fu Y, Brooker S, Kanaan
J, Mandelkow EM, et  al. Mutations of tau protein NM, Vana L, et al. Characterization of prefibrillar tau
in frontotemporal dementia promote aggregation oligomers in  vitro and in Alzheimer disease. J  Biol
of paired helical filaments by enhancing local beta-­ Chem. 2011;286(26):23063–76.
structure. J Biol Chem. 2001;276(51):48165–74. 27. Jicha GA, Bowser R, Kazam IG, Davies P. Alz-50 and
16. Sahara N, Maeda S, Murayama M, Suzuki T, Dohmae MC-1, a new monoclonal antibody raised to paired
N, Yen SH, et  al. Assembly of two distinct dimers helical filaments, recognize conformational epit-
and higher-order oligomers from full-length tau. Eur opes on recombinant tau. J  Neurosci Res [Internet].
J Neurosci. 2007;25(10):3020–9. 1997;48(2):128–32. Available from: http://www.ncbi.
17. Peterson DW, Zhou H, Dahlquist FW, Lew J.  A
nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=Pu
soluble oligomer of tau associated with fiber forma- bMed&dopt=Citation&list_uids=9130141
tion analyzed by NMR.  Biochemistry [Internet]. 28. Carmel G, Mager EM, Binder LI, Kuret J.  The

2008;47(28):7393–404. Available from: https://www. structural basis of monoclonal antibody Alz50’s
ncbi.nlm.nih.gov/pubmed/18558718 selectivity for Alzheimer’s disease pathology. J  Biol
18.
Weaver CL, Espinoza M, Kress Y, Davies Chem [Internet]. 1996;271(51):32789–95. Available
P.  Conformational change as one of the earliest from: http://www.ncbi.nlm.nih.gov/entrez/query.fcg
alterations of tau in Alzheimer’s disease. Neurobiol i?cmd=Retrieve&db=PubMed&dopt=Citation&l
Aging [Internet]. 2000;21(5):719–27. Available ist_uids=8955115
from: http://www.ncbi.nlm.nih.gov/entrez/query.fcg 29. Jeganathan S, Hascher A, Chinnathambi S, Biernat
i?cmd=Retrieve&db=PubMed&dopt=Citation&l J, Mandelkow EM, Mandelkow E.  Proline-directed
ist_uids=11016541 pseudo-phosphorylation at AT8 and PHF1 epitopes
19. Berger Z, Roder H, Hanna A, Carlson A, Rangachari induces a compaction of the paperclip folding of tau
V, Yue M, et  al. Accumulation of pathological tau and generates a pathological (MC-1) conformation.
species and memory loss in a conditional model of J Biol Chem. 2008;283:32066–76.
tauopathy. J Neurosci. 2007;27(14):3650–62. 30. Kanaan NM, Morfini GA, LaPointe NE, Pigino GF,
20. Tepper K, Biernat J, Kumar S, Wegmann S, Timm Patterson KR, Song Y, et al. Pathogenic forms of tau
T, Hübschmann S, et  al. Oligomer formation of tau inhibit kinesin-dependent axonal transport through
protein hyperphosphorylated in cells. J  Biol Chem. a mechanism involving activation of axonal phos-
2014;289(49):34389–407. photransferases. J Neurosci. 2011;31(27):9858–68.
21. Alonso A, Zaidi T, Novak M, Grundke-Iqbal I, Iqbal 31. LaPointe NE, Morfini G, Pigino G, Gaisina IN,

K.  Hyperphosphorylation induces self-assembly of Kozikowski AP, Binder LI, et al. The amino terminus
tau into tangles of paired helical filaments/straight of tau inhibits kinesin-dependent axonal transport:
filaments. Proc Natl Acad Sci U S A [Internet]. implications for filament toxicity. J  Neurosci Res.
2001;98(12):6923–8. Available from: http://www. 2009;87:440–51.
ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&d 32. Kampers T, Pangalos M, Geerts H, Wiech H,

b=PubMed&dopt=Citation&list_uids=11381127 Mandelkow E.  Assembly of paired helical filaments
22. Lasagna-Reeves CA, Castillo-Carranza DL, Guerrero-­ from mouse tau: implications for the neurofibrillary
Muñoz MJ, Jackson GR, Kayed R.  Preparation pathology in transgenic mouse models for Alzheimer’s
and characterization of neurotoxic tau oligomers. disease. FEBS Lett [Internet]. 1999;451(1):39–44.
Biochemistry. 2010;49(47):10039–41. Available from: http://www.ncbi.nlm.nih.gov/entrez/
23. Lasagna-Reeves CA, Castillo-Carranza DL, Sengupta query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citati
U, Clos AL, Jackson GR, Kayed R.  Tau oligomers on&list_uids=10356980
impair memory and induce synaptic and mitochondrial 33. Cox K, Combs B, Abdelmesih B, Morfini G, Brady
dysfunction in wild-type mice. Mol Neurodegener. ST, Kanaan NM.  Analysis of isoform-specific
2011;6(1):39. tau ­aggregates suggests a common toxic mecha-
nism involving similar pathological conformations
380 S. Maeda and A. Takashima

and axonal transport inhibition. Neurobiol Aging. ity control for treatment of adult onset neurode-
2016;47:113–26. generative diseases. Annu Rev Pharmacol Toxicol.
34.
Lasagna-Reeves CA, Castillo-Carranza DL, 2015;55:353–71.
Sengupta U, Sarmiento J, Troncoso J, Jackson GR, 43. Fontaine SN, Martin MD, Akoury E, Assimon VA,
et  al. Identification of oligomers at early stages of Borysov S, Nordhues BA, et  al. The active Hsc70/
tau aggregation in Alzheimer’s disease. FASEB tau complex can be exploited to enhance tau turnover
J. 2012;26:1946–59. without damaging microtubule dynamics. Hum Mol
35. Rosenmann H, Grigoriadis N, Karussis D, Boimel Genet. 2015;24:3971–81.
M, Touloumi O, Ovadia H, et  al. Tauopathy-like 44. Abisambra J, Jinwal UK, Miyata Y, Rogers J, Blair L,
abnormalities and neurologic deficits in mice Li X, et al. Allosteric heat shock protein 70 inhibitors
immunized with neuronal tau protein. Arch Neurol. rapidly rescue synaptic plasticity deficits by reducing
2006;63(10):1459–67. aberrant tau. Biol Psychiatry. 2013;74:367–74.
36. Nacharaju P, Lewis J, Easson C, Yen S, Hackett J, 45. Soeda Y, Yoshikawa M, Almeida OFX, Sumioka A,
Hutton M, Yen S-H. Accelerated filament formation Maeda S, Osada H, et al. Toxic tau oligomer formation
from tau protein with specific FTDP-17 missense blocked by capping of cysteine residues with 1,2-dihy-
mutations. FEBS Lett. 1999;447 (2–3):195–9. droxybenzene groups. Nat Commun [Internet].
37. Arrasate M, Mitra S, Schweitzer ES, Segal MR, 2015;6:1–12. https://doi.org/10.1038/ncomms10216.
Finkbeiner S. Inclusion body formation reduces levels 46. Shafiei SS, Guerrero-Muñoz MJ, Castillo-Carranza
of mutant huntingtin and the risk of neuronal death. DL. Tau oligomers: cytotoxicity, propagation, and mito-
Nature. 2004;431(7010):805–10. chondrial damage. Front Aging Neurosci. 2017;9:83.
38. Mudher A, Colin M, Dujardin S, Medina M,
47.
Kaniyappan S, Chandupatla RR, Mandelkow
Dewachter I, Alavi Naini SM, et  al. What is the EM, Mandelkow E.  Extracellular low-n oligo-
evidence that tau pathology spreads through prion- mers of tau cause selective synaptotoxicity with-
like propagation? Acta Neuropathol Commun. out affecting cell viability. Alzheimers Dement.
2017;5(1):99. 2017;13(11):1270–91.
39. Clavaguera F, Bolmont T, Crowther RA, Abramowski 48. Gómez-Ramos A, Díaz-Hernández M, Rubio A,

D, Frank S, Probst A, et al. Transmission and spread- Miras-Portugal MT, Avila J.  Extracellular tau pro-
ing of tauopathy in transgenic mouse brain. Nat Cell motes intracellular calcium increase through M1 and
Biol. 2009;11(7):909–13. M3 muscarinic receptors in neuronal cells. Mol Cell
40. Sanders DW, Kaufman SK, DeVos SL, Sharma AM, Neurosci. 2008;37:673–81.
Mirbaha H, Li A, et al. Distinct tau prion strains prop- 49. Jeganathan S, Von Bergen M, Brutlach H, Steinhoff
agate in cells and mice and define different tauopa- HJ, Mandelkow E.  Global hairpin folding of tau in
thies. Neuron. 2014;82(6):1271–88. solution. Biochemistry. 2006;45(7):2283–93.
41. Gerson JE, Kayed R.  Formation and propagation of 50. Sahara N, Shimojo M, Ono M, Takuwa H, Febo M,
tau oligomeric seeds. Front Neurol. 2013;4:93. Higuchi M, et al. In vivo tau imaging for a diagnostic
42. Pratt WB, Gestwicki JE, Osawa Y, Lieberman
platform of tauopathy using the rTg4510 mouse line.
AP.  Targeting Hsp90/Hsp70-based protein qual- Front Neurol. 2017;8:663.
 Experimental Models
of Tauopathy – From Mechanisms 28
to Therapies

Julika J. Götz and Jürgen Götz

Introduction – The Concept the somatodendritic accumulation of tau and its

of ‘Tauopathy’ hyperphosphorylation at selected epitopes [3].
The focus of this book chapter is on animal
One of the routinely used ‘labels’ for neurode- models for tauopathies that have been established
generative diseases that provides a clue about the by manipulating the germline (i.e. viral
underlying pathomechanisms is the term ‘pro- approaches or the intracerebral injection of brain
teinopathy’. Proteinopathies come in many varia- extracts will not be discussed). Although AD is a
tions, including ‘tauopathy’, a term coined by tauopathy which pathologically involves not only
Drs Bernardino Ghetti and Michel Goedert to tau but also Aβ and even TDP-43 deposition [4],
categorize a group of human neurological disor- we will, in this chapter, restrict our discussion to
ders that have tau inclusions in neurons or glia as tau models, their validity, the insight they have
their common, defining denominator. Previously, provided into pathomechanisms, as well as the
when Alzheimer’s disease (AD) research was therapeutic interventions that have been tested
dominated by studies into the processing and tox- and validated in these models. We will then finish
icity of amyloid-β (Aβ, in part owing to the fact the chapter with an outlook into what the future
that disease-causing mutations had been identi- may hold as far as animal models for tauopathy
fied in the amyloid precursor protein-encoding (or brain diseases more generally) are
gene APP, but not in the tau-encoding gene concerned.
MAPT) [1], it was helpful to stress the point that
tau pathology is a defining feature of more than
two dozen so-called primary tauopathies, whereas Approaches to Model Tauopathy
in the secondary tauopathy AD, tau pathology in Animals
exists on par with Aβ pathology [2]. It was at this
time that Götz et  al. generated the first human The mouse (Mus musculus) is the most frequently
wild-type tau transgenic mouse model that repro- used species for modeling tauopathy because of
duced aspects of the human pathology, including its ease of breeding and genetic manipulation, as
well as the relatively low cost of its husbandry.
As discussed recently, there are a few principal
J. J. Götz · J. Götz (*) approaches in generating animal models, and
Clem Jones Centre for Ageing Dementia Research these strategies have been applied to the study of
(CJCADR) Queensland Brain Institute (QBI), The
University of Queensland, Brisbane, Australia tau under both physiological and pathological
e-mail: j.goetz@uq.edu.au conditions [5].

© Springer Nature Singapore Pte Ltd. 2019 381

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_28
382 J. J. Götz and J. Götz

For the generation of classical transgenic tion of the transgene, which often forms multiple
models, the gene of interest and a particular pro- tandem repeats, and interestingly, for most of the
moter are combined and delivered into the zygote strains the analysis is incomplete and more often
by pronuclear injection [6]. The injected zygotes than not, motor or cognitive deficits have not
are then transplanted into pseudopregnant been assessed [13].
females to obtain pups that are screened for the The first mouse strain with significant tau fila-
expression of the transgene and then used to ment formation in neurons was reported in 2000;
establish a new animal strain [5]. During the cre- this mouse expressed the prevalent P301L muta-
ation of the first human wild-type tau transgenic tion of FTD tau in neurons [14]. Whereas the
mouse strain [3], the longest human tau isoform majority of vectors that have been used target tau
(2N4R) in its wild-type form was expressed by expression to neurons, a few have also achieved
cloning the tau-encoding cDNA into a human pathology in glia [15–17]. Surprisingly, glial tau
Thy1 (hThy1) expression vector. At the time, the pathology has remained an underexplored
authors also tested the human rhombotin 1, the research area, despite the fact that in tauopathies
rat neuron-specific enolase and the murine neuro- such as corticobasal degeneration (CBD) or pro-
filament light chain promoter; however, neither gressive supranuclear palsy (PSP), glial pathol-
of these expression vectors yielded significantly ogy often outnumbers neuronal pathology. A
high expression levels [3]. Subsequently, stron- transactivator-based strategy was chosen to gen-
ger promoters were chosen for transgene expres- erate P301L transgenic mice that are character-
sion. These included the murine Thy1.2 ized by massive neuronal cell loss and early-onset
(mThy1.2) promoter, which has been success- tau pathology [18]. Other regulatable systems
fully used to overexpress different frontotempo- expressed tau fragments which were either prone
ral dementia (FTD) mutant forms of tau, to or resistant to aggregation [19]. Whereas tau
achieving tau filament formation [7, 8], the transgenic mice are generally maintained on a
murine prion protein (PrP) promoter [9, 10], or C57Bl/6 wild-type background, breeding a
the Ca2+/calmodulin-dependent protein kinase II P301L tau transgene onto a senescence-­
(CaMKII) promoter [11, 12]. In addition to pro- accelerated mouse-prone 8 (SAMP8) background
moter choice, these strains differ in terms of the over ten generations achieved an accelerated tau
type of tau isoform they express, and whether or phenotype, indicating that age (the major risk
not they carry pathogenic mutations of FTD. An factor for AD) accentuated tau pathology [20].
excellent recent review by Dr. Simon Dujardin Finally, the entire tau-encoding MAPT locus has
and colleagues provides a comprehensive, been introduced into the mouse genome, yielding
although as they write, ‘non-exhaustive’ over- a more pronounced pathology when endogenous
view of 50 tau transgenic mouse strains (some in tau was removed [21].
combination with other genetic modifications), a For models based on homology-directed
small number of rat and C. elegans strains, and recombination (HDR) mechanisms, the targeting
around 25 Drosophila strains [13]. When one vector containing the sequence to be introduced
assesses the additional information provided in into the host genome is flanked by homology
this review, the heterogeneity of the models arms corresponding to the recipient genomic site
becomes evident: there are differences in the tau of interest and also contains a selection marker.
isoform being expressed, the presence or type of In the first step, the targeting vector is introduced
FTD mutation(s), the promoter used to drive into embryonic stem cells that are initially
transgene expression, the distribution pattern of selected for an HDR event rather than random
tau, the degree of overexpression compared to integration. These are then microinjected into
endogenous tau levels, and the type and time-­ blastocysts, followed by transfer into pseudo-
course of tau pathology, including its phosphory- pregnant females to establish a new line [5]. To
lation pattern. The way these models were obtain models based on gene editing (such as
generated did not allow for a controlled integra- transcription activator-like effector nucleases
28  Experimental Models of Tauopathy – From Mechanisms to Therapies 383

(TALEN) or CRISPR), a donor template contain- reported parkinsonism [29] and another reported
ing homology arms is injected into the pronu- subtle dopamine-independent motor deficits with
cleus of zygotes together with the method-specific age [30] (for a detailed discussion see [31]).
components (either TALENs or the Cas9 enzyme With the aim of reproducing tau pathology
and guiding RNA molecules, respectively) to without relying on an overexpression approach, a
introduce breaks in the genomic DNA at specific P301L tau knock-in mouse was generated in
sites. 2012 [32]. However, no overt tau pathology
Embryonic stem cell-based approaches have developed, suggesting that endogenous levels of
also been used to generate tau knock-out mice. even pathogenic tau were not sufficient to achieve
The first strain was established by Harada et al. in tau pathology, including tau filament formation,
1994 but, surprisingly, did not display any patho- within the relatively short lifespan of a mouse –
logical phenotype [22]. In a compensatory mech- this has only been achieved through the genera-
anism, the tau homolog MAP 1A was upregulated tion of overexpression models. In 2016, we used
and the axon caliber was altered to one typically the TALEN gene editing tool to generate knock-
found in dendrites. Subsequently, in 2001, two ­in mice in which a photoconvertible mEOS2 tag
additional strains became available. The was introduced in-frame into the carboxy-­
‘Dawson’ strain was established by Dawson and terminus of the MAPT gene [33]. The mice
colleagues, who used the same strategy as Harada expressed all three major brain isoforms of tau,
and colleagues by replacing the first coding exon with the isoform ratio being similar to that
of tau with a neomycin selection cassette, thereby reported for wild-type mice. Interestingly how-
abrogating tau expression [23]. The second strain ever, we found that the level of tau expression
was a ‘knock-in’ generated by Tucker and col- was approximately 15-fold lower than that of
leagues, who inserted a GFP cassette in frame, wild-type animals. By establishing primary cul-
resulting in a fusion protein that contained the tures from these mice, the distribution of tau in
first 31 amino acids of tau [24]. Neither of these different subcellular compartments could be ana-
two strains revealed any obvious impairment (at lyzed using live-cell imaging, photoconversion
the time, the Tucker strain was only used as a and FRAP (fluorescence recovery after photo-
‘tool’ to study neurotrophins, although it has bleaching) assays. It was found that Tau-mEOS2
since been used for behavioural studies: [25]). followed a proximo-distal gradient in axons and a
Interestingly, in the Dawson strain, MAP 1A lev- subcellular distribution similar to that of endog-
els were increased twofold at birth, whereas they enous Tau in neurons obtained from wild-type
returned back to normal levels as the mice mice. These features were abolished, when either
became older, suggesting that MAP 1A may human wild-type or P301L mutant tau were over-­
compensate for the loss of tau during early brain expressed. These results highlight the potential
development, but not in the mature brain [23]. confound of overexpression systems. They simul-
Potential compensatory functions attributable to taneously highlight the value of gene-edited
closely related MAPs can be excluded in C. ele- models in gaining a deeper understanding of the
gans, where protein with Tau-like repeats (PTL-­ mobility and dynamics of tau in response to both
1) is the sole homolog of TAU/MAP 2/MAP 4 physiological and pathological stimuli.
[26, 27], meaning that shared physiological func- To address the issue of a potential overexpres-
tions of the different family members can be sion artifact in the context of AD and FTD patho-
addressed. Knock-out studies in C. elegans genesis, attempts have been made to humanize
revealed that PTL-1 has an essential role in main- the relevant murine genes and then introduce
taining neuronal health with age and in regulating pathogenic mutations found in familial cases of
whole organism lifespan [28]. Several follow-up AD and FTD.  This approach has not remained
studies using the above tau knock-out mouse without challenges. Most progress has been made
strains mostly failed to report significant memory towards modeling Aβ pathology. When the
and motor impairments; however, one group sequence in the murine APP locus was ­humanized
384 J. J. Götz and J. Götz

by replacing three amino acids, predictably, these pathology, but also with the accompanying syn-
mice did not develop Aβ plaques [34]. However, aptic and neuronal degeneration. These aspects
the simultaneous knock-in of APP harboring have been achieved in a subset of models and
both the APPswe and the Beyreuther/Iberian generally required the overexpression of FTD
(I716F) mutations has led to Aβ deposition by mutant tau [5, 13]. A major challenge lies in com-
6 months of age, with the introduction of an addi- paring the different models and assessing their
tional APP mutation, the Arctic (E693G) muta- validity, due to either incomplete analysis or the
tion, leading to the formation of Aβ deposits as lack of standardized analytic techniques.
early as 2  months of age in homozygous mice A further challenge that arises is modeling the
[34]. These strains presented age-dependent role of tau pathology in impairing various
memory impairments corresponding to the num- domains of cognition and motor function. Such
ber of mutations that had been introduced [34], impairments occur due to the accumulation of
together with an Aβ pathology that was more pro- pathological tau in particular brain areas, which
nounced in female mice [35]. Obviously, this is determined by several factors including the
strategy combines mutations and generates a integration site of the transgene (which deter-
hybrid protein that does not exist in nature and mines the expression pattern) and the copy num-
may lead to an altered interactome. Humanizing ber (which is somewhat linked to expression
tau is even more challenging due to the large size levels). The pattern of expression and pathology
of the MAPT locus and the coding sequence, and inevitably differs from that encountered in for
the difference in isoform composition in adult example AD brains, where the pathology follows
mice and humans [36]. There is also the question a stereotypical pattern with disease progression
of which sequences should be humanized. The as formulated in the Braak staging scheme [41],
challenge that these issues present is reflected by which is believed to be due to a stereotypical pro-
the absence of published outcomes despite cess of tau spreading [42, 43], in part mediated
attempts by several laboratories. by exosomes [44–46], and an area discussed else-
As reviewed previously [37], other species, where in this book in more detail Chaps. 21, 34
such as the sea lamprey (Petromyzon marinus) and 35).
and zebrafish (Danio rerio), have also been Finally, a recent review by Ahmed et  al. of
explored as a means to dissect specific aspects of mouse models of FTD (which goes beyond tau
tau pathology. Large animal species in which and also includes C9ORF72, progranulin, TDP-­
neurodegeneration has been successfully mod- 43 and VCP) [47] discusses them in the context
eled include the pig and sheep [38, 39], with fea- of the functional impairments observed in human
tures of AD pathology being modeled in tauopathy. These include behavioral, socioemo-
transgenic pigs [40]. tional, motor and memory functions, as well as
language, eating and metabolism. Several of
these impairments have indeed been modeled in
 omparative Analysis of Tauopathy
C mice, adding to the validity of such models for
Models the study of tauopathy. For example, it has been
reported that impaired ultrasonic vocalizations in
With advances in research technologies, there is aged mutant tau mice correlate with tau pathol-
often a change in focus and therefore also in the ogy in the mid-brain and brainstem nuclei con-
expectations of what an animal model needs to trolling vocalization and respiration [48]. This
deliver. As discussed above, an initial goal has may partially resemble the language disorders
been to reproduce aspects of the human pathol- observed in FTD patients, although the finding
ogy, at least at a histological and biochemical awaits confirmation in other lines. Another FTD
level. Considering that AD and FTD are neurode- tau mutant mouse model displayed age-­dependent
generative diseases, an authentic model would signs of impulsivity, decreased social explora-
need to present not only with a tau (and Aβ) tion, and executive dysfunction. The deficit in
28  Experimental Models of Tauopathy – From Mechanisms to Therapies 385

executive function was initially limited to levels were massively induced by Fyn [55]. In
decreased spatial working memory, but with addition, tau was found to be strongly phosphor-
aging extended to impaired instrumental short-­ ylated. This boosting was not seen when the den-
term memory. Importantly, tau pathology was dritic tau homolog MAP 2 or GFP was
prominent in brain regions underlying these co-expressed with Fyn, indicating the specificity
behaviours [49]. One of the propositions of the of this effect [55]. Importantly, the boosting
review by Ahmed and colleagues [47] was to (re) effect of Fyn on tau was suppressed with cyclo-
examine existing FTD mouse models based on heximide and not with actinomycin D, suggest-
the functional impairment in human patients. ing (and subsequently confirmed) that Fyn
They further suggested that defining FTD mouse induces tau translation rather than transcription.
models, in isolation or in combination, based on We further showed that Fyn activates the ERK/S6
functional deficits driven by clinical observations pathway in an mTOR-independent manner. We
may improve drug development and testing. then incubated primary neurons with Aβ and
showed that the oligomeric form of the peptide
induced the de novo protein synthesis of tau via a
Insight into Pathomechanisms Fyn/ERK/S6 pathway. We further revealed that
Aβ employed this pathway to achieve de novo
It is impossible to provide an exhaustive over- protein synthesis specifically in the somatoden-
view of the pathomechanistic insight that has dritic compartment. Activation of the same path-
been gained in animal models and pay tribute to way was demonstrated in a range of cellular
the wealth of data that have been obtained. Here, systems, and in vivo in brains from Aβ-depositing
we would like to focus on two findings made in APP23 mice [56], Aβ-injected wild-type mice,
our laboratory. Both involved studies in which and Fyn-overexpressing FynCA mice [57] with
cell culture systems were complemented by ani- tau accumulation [55]. Our findings even
mal models. extended to human patients, where tau levels in
The first study focused on why tau, which is the cerebrospinal fluid of AD patients, but not
mainly perceived as an axonal protein, accumu- patients with FTD, were shown to rise steeply,
lates in the somatodendritic domain. Early in suggesting that Aβ is a driver of the tau increases
development, tau is distributed throughout the that occur in AD [58, 59]. We believe that tau is
neuron; with maturation, however, it becomes not only a substrate of the Aβ/Fyn/EEK/S6 cas-
enriched in the axon [50]. In AD and related cade, but may also have a role in the translational
tauopathies, tau accumulates in both the soma machinery itself. Indeed, tau has been shown to
and the dendrites in a hyperphosphorylated form interact with RNA-binding proteins such as
[51]. It is generally assumed that hyperphosphor- TIA1, which has been reported to facilitate the
ylated tau in the axon detaches from the microtu- formation of stress granules in tauopathy [60].
bules and passes through the axon initial segment, These stress granules are believed to sequester
which serves as a diffusion barrier for physiolog- specific mRNAs, altering the synthesis of partic-
ically phosphorylated tau, before accumulating ular sets of proteins and therefore potentially
in the cell body and dendrites, a process that is having a role in tau pathology [60, 61]. Together,
partly mediated by Aβ [52, 53]. We asked whether these findings indicate that there are distinct dif-
Aβ employs a mechanism other than relocaliza- ferences in the pathomechanisms of primary and
tion of tau to account for the massive accumula- secondary tauopathies.
tion of this protein in the somatodendritic The second example has to do with the fact
compartment. Because tau is known to interact that tau is best perceived as a scaffolding protein.
with the Src kinase Fyn [25, 54], we investigated This means that, with its pathological accumula-
this crosstalk in more detail, initially by co-­ tion in the somatodendritic domain, it may bind
transfecting tagged tau and Fyn expression vec- to additional proteins. We showed this to be the
tors into HEK293 cells. This revealed that tau case, when we investigated the parkinsonism and
386 J. J. Götz and J. Götz

axonal transport impairment phenotype in K369I tle phenotype of tau knock-out mice [69], the
tau transgenic K3 mice [62, 63]. We found that view has emerged that, rather than targeting dis-
the pathological tau which accumulated in the tinct post-translational modifications, it may suf-
soma trapped the kinesin adaptor JIP1, thereby fice to reduce overall tau levels. In this context,
preventing it from loading distinct cargoes, pan-tau-specific antibodies represent a potential
including mitochondria, onto the kinesin 5 motor, treatment strategy.
resulting in their impaired axonal transport [62, Many phospho-epitopes have already  been
63]. Interestingly, in very recent work, we identi- tested in mice. The active immunization
fied a similar trapping effect as an underlying approaches used a peptide containing important
pathomechanism when we tried to understand tau phospho-epitopes (pS396/pS404 (known as
why pathological tau (this time wild-type and the PHF1 epitope), pS202/pT205 (AT8), pT212/
P301L human tau) impairs mitophagy [64, 65]. pS214 (AT100), pT231/pS235 (AT180) and
Here, we complemented a HEK293 cell system pS422), proving efficacious in preventing or
with a C. elegans model of tauopathy to reveal reducing pathology in tau transgenic models, in
that pathological tau traps parkin in the soma, the absence of overt side effects [68].
thereby preventing it from localizing onto mito- Complementing these active approaches, passive
chondria [65]. These findings add a new dimen- strategies have also been pursued using the PHF1
sion to our understanding of tau toxicity, antibody and the conformation-dependent anti-­
illustrating the existence of a vicious cycle tau antibody MC1, resulting in behavioral and
whereby tau not only contributes to mitochon- cognitive improvements in mouse models [68].
drial dysfunction, but also inhibits the clearance Neuronal uptake of therapeutic anti-tau antibod-
of the resultant damaged mitochondria [66, 67]. ies is believed to be crucial for vaccine efficacy,
although such uptake has not been reported in all
preclinical studies. Because tau is also present in
 rom Mechanisms to Therapeutic
F the extracellular space, however, antibodies that
Interventions have not been shown to be taken up into neurons
are predicted to block both the uptake and the
Strategies that directly target tau and have seeding activity of extracellular tau aggregates,
reached clinical development involve blocking its and have been demonstrated to ameliorate the
aggregation and vaccination, whereas indirect cognitive deficits in tau transgenic mice [70]. A
strategies involve stabilizing microtubules, as recent study has shown that different antibodies
well as manipulating kinases and phosphatases targeting extracellular tau might have different
that govern the post-translational modification of modes of action, and some antibodies have been
tau [68]. Other important approaches that have shown to facilitate the clearance of extracellu-
been tested in animal studies include reducing larly added tau aggregates by microglia-like BV2
tau oligomerization, facilitating autophagy and cells, although others do not [71]. Overall, sev-
protein clearance, supporting mitochondrial eral tau-targeted active and passive vaccines are
function and mitigating oxidative stress, as well currently in phase I and II clinical trials in AD
as reducing tau levels, hyperphosphorylation and and PSP, with PSP being an attractive testing
pathological tau propagation, either directly or ground because of the absence of Aβ pathology
through vaccination [68]. As far as the latter is as a comorbidity factor [68].
concerned, guidance has been provided by immu- More recently, we have begun to investigate
nization strategies targeting Aβ, although tau is a ultrasound as a new treatment strategy for AD
more challenging target because it is mainly and tauopathies more generally. Ultrasound is a
deposited intracellularly and because we still do mechanical pressure (sound) wave with a fre-
not fully understood which post-translational quency above 20  kHz, the highest frequency
modifications and tau species need to be targeted that can be detected by humans [72]. Ultrasound
for therapeutic intervention. Considering the sub- is used in a wide range of diagnostic imaging
28  Experimental Models of Tauopathy – From Mechanisms to Therapies 387

a­ pplications, primarily in the fields of obstetrics reduced  effectively, plaques were partially
and cardiology, but also for examining the abdo- cleared, and memory functions were restored in
men and musculoskeletal system. In this con- three complementary behavioral tests [78]. Our
text, ultrasound waves are transmitted from the results also revealed that the underlying clear-
transducer into the patient [72]. The wave is par- ance mechanism involved the internalization of
tially reflected at tissue interfaces and the Aβ by brain-resident dormant microglial cells
‘echoes’ are detected by the same transducer. that became activated by unidentified blood-
Bone tissue highly attenuates the propagation of borne factors which entered the brain as a conse-
ultrasound, so current neurological ultrasound quence of the transient opening of the
imaging must be performed at ‘acoustic win- BBB. Importantly, we found that tau pathology
dows’ in the skull, where the bone is thin enough (being mostly intraneuronal) could also be par-
for adequate signal transmission. To date, this tially ameliorated with this approach, as shown
approach has only proven suitable for imaging in the pR5 transgenic mouse model of tauopathy
structures in limited areas within the brain, such which accumulates hyperphosphorylated tau in
as blood vessels that are basal to the skull. neurons [81]. In this particular study, we per-
Therapeutic ultrasound uses similar instrumen- formed four treatments and also tested the syner-
tation at a higher power, and can be broadly cat- gistic effects of scanning ultrasound using a
egorized as either thermal or non-thermal. In the single chain variant (ScFv) antibody fragment
latter case, ultrasound is combined with intrave- targeting the 2 N isoform of tau [81]. However,
nously injected microbubbles to achieve tran- the mechanism by which ultrasound treatment
sient opening of the blood-­brain barrier (BBB). clears tau remains to be determined. Long-term
Microbubbles are biologically inert and have safety in wild-type mice has also been demon-
either a lipid or polymer shell and a stabilized strated, based on a wide range of behavioral,
gas core. They can be systemically administered electrophysiological and imaging modalities
and subsequently exposed to non-invasively [82, 83], together with related studies in sheep
delivered focused ultrasound pulses [72, 73]. [84]. Considering the safety and efficacy profile
Microbubbles within the target volume become of ultrasound, which is highly tunable, these
‘acoustically activated’ by what is known as studies suggest that therapeutic ultrasound may
acoustic cavitation. In this process, the micro- offer a potential treatment strategy for tauopa-
bubbles (with a diameter of up to 10 μm, match- thies, either on its own or in combination with a
ing that of small capillaries) expand and contract therapeutic agent [85].
over several cycles. These dilations and contrac-
tions displace the vessel wall [74, 75]. More
specifically, the mechanical interactions I s There a Future for Animal Models
between ultrasound, microbubbles and the vas- of Tauopathies?
culature transiently open tight junctions and
facilitate size-dependent transport across the There is an ongoing need for animal models of
BBB [76]. These microbubbles are clinically tauopathies. We believe that there is value in the
approved and routinely used as contrast agents existing models, which can be further probed to
for diagnostic ultrasound. reflect clinical features, to dissect additional
We and others have shown that microbubble-­ pathomechanisms and to validate therapeutic
mediated BBB opening in the absence of any interventions. To increase their utility, a stan-
therapeutic agent reduces the Aβ pathology in dardized approach of analysis is highly recom-
APP mutant APP23 mice [77–80]. Following mended. It can also be anticipated that gene
5–8 weekly treatments of 12 month-old APP23 editing tools will be increasingly employed to
mice with ultrasound in a scanning mode, Aβ generate improved animal models and to study
species ranging from monomers to oligomers to key molecules such as tau and the enzymes with
high molecular weight species could be which it interacts, such as Fyn, Pyk2 and STEP,
388 J. J. Götz and J. Götz

at a subcellular level, using tools such as super-­ Acknowledgements We acknowledge support by the
Estate of Dr. Clem Jones AO, the Australian Research
resolution microscopy [86]. Mice are likely to Council [DP160103812] and the National Health and
remain a  major species for modeling and ana- Medical Research Council of Australia [GNT1037746,
lyzing tauopathy. However, although used less GNT1127999], and thank Rowan Tweedale for critically
frequently than mice, rats present with several reading this manuscript.
advantages for the study of human diseases,
including their relatively large brain size, which
facilitates surgery. Rats are also much easier to References
handle than mice and are less easily stressed by
1. Hardy J.  The discovery of Alzheimer-causing
human contact. In addition, several of the con- mutations in the APP gene and the formula-
founds in the behavioral paradigms found in the tion of the “amyloid cascade hypothesis”. FEBS
mouse studies do not apply to rats [87]. It is J. 2017;284(7):1040–4.
anticipated that the increased availability of 2. Lee VM, Goedert M, Trojanowski
JQ.  Neurodegenerative tauopathies. Annu Rev
gene-­editing tools is likely to pave the way for a Neurosci. 2001;24:1121–59.
multitude of new genetic models based on rats 3. Götz J, Probst A, Spillantini MG, Schafer T, Jakes
[87]. There is also a continued need for inverte- R, Burki K, et  al. Somatodendritic localization and
brate models such as flies and worms. These hyperphosphorylation of tau protein in transgenic
mice expressing the longest human brain tau isoform.
have short generation times, are cheap in their EMBO J. 1995;14(7):1304–13.
handling, do not require animal ethics approval 4. Amador-Ortiz C, Lin WL, Ahmed Z, Personett D,
and are ideal for fast genetic screening. In an Davies P, Duara R, et al. TDP-43 immunoreactivity in
ideal world, pathological findings would be hippocampal sclerosis and Alzheimer’s disease. Ann
Neurol. 2007;61(5):435–45.
demonstrated in multiple models, either in two 5. Götz J, Bodea LG, Goedert M.  Rodent mod-
or more strains of the same species (such as two els for Alzheimer disease. Nat Rev Neurosci.
complementary tau transgenic mouse models) 2018;19(10):583–98.
or in a cellular system complemented by an 6. Ittner LM, Götz J.  Pronuclear injection for
the generation of transgenic mice. Nat Protoc.
invertebrate and a vertebrate model. There is 2007;2(5):1206–15.
also value in large animal models such as mini- 7. Götz J, Chen F, Barmettler R, Nitsch RM.  Tau fila-
pigs, sheep or non-human primates; however, ment formation in transgenic mice expressing P301L
their availability, the time required to establish tau. J Biol Chem. 2001;276(1):529–34.
8. Allen B, Ingram E, Takao M, Smith MJ, Jakes R,
an aged colony, the hurdles posed by ethics Virdee K, et  al. Abundant tau filaments and non-
approval, and the cost and effort required for apoptotic neurodegeneration in transgenic mice
their analysis presents practical impediments expressing human P301S tau protein. J  Neurosci.
for their widespread use. Cell culture systems, 2002;22(21):9340–51.
9. Brion JP, Tremp G, Octave JN.  Transgenic expres-
even when derived from human induced plurip- sion of the shortest human tau affects its compart-
otent stem cells [88], cannot model brain or mentalization and its phosphorylation as in the
immune interactions or the BBB, and aging can pretangle stage of Alzheimer’s disease. Am J Pathol.
only be modeled to some extent. Whether organ- 1999;154(1):255–70.
10. Yoshiyama Y, Higuchi M, Zhang B, Huang SM, Iwata
oids provide a viable solution remains to be N, Saido TC, et al. Synapse loss and microglial acti-
determined [89]. There is no doubt that this vation precede tangles in a P301S tauopathy mouse
technology opens up numerous possibilities for model. Neuron. 2007;53(3):337–51.
scientific discovery in developmental biology, 11. Tatebayashi Y, Miyasaka T, Chui DH, Akagi T,

Mishima K, Iwasaki K, et al. Tau filament formation
as well as in translational research, but whether and associative memory deficit in aged mice express-
organoids can truly live up to this challenge is, ing mutant (R406W) human tau. Proc Natl Acad Sci
for some, still an open question [90]. U S A. 2002;99(21):13896–901.
Transplanting human cell-derived organoids 12. Kimura T, Yamashita S, Fukuda T, Park JM,

Murayama M, Mizoroki T, et al. Hyperphosphorylated
into the brains of animals such as mice would be tau in parahippocampal cortex impairs place learning
a compromise and may allow the brain to func- in aged mice expressing wild-type human tau. EMBO
tion as a Petri dish for human tissue [91]. J. 2007;26(24):5143–52.
28  Experimental Models of Tauopathy – From Mechanisms to Therapies 389

13. Dujardin S, Colin M, Buee L.  Invited review: ani- mechanosensation and development in Caenorhabditis
mal models of tauopathies and their implications for elegans. Dev Genes Evol. 2008;218(10):541–51.
research/translation into the clinic. Neuropathol Appl 28. Chew YL, Fan X, Götz J, Nicholas HR. Protein with
Neurobiol. 2015;41(1):59–80. tau-like repeats regulates neuronal integrity and lifes-
14. Lewis J, McGowan E, Rockwood J, Melrose H,
pan in C. elegans. J Cell Sci. 2013;126:2079–91.
Nacharaju P, Van Slegtenhorst M, et al. Neurofibrillary 29. Lei P, Ayton S, Finkelstein DI, Spoerri L, Ciccotosto
tangles, amyotrophy and progressive motor distur- GD, Wright DK, et al. Tau deficiency induces parkin-
bance in mice expressing mutant (P301L) tau protein. sonism with dementia by impairing APP-mediated
Nat Genet. 2000;25(4):402–5. iron export. Nat Med. 2012;18(2):291–5.
15. Ishihara T, Hong M, Zhang B, Nakagawa Y, Lee MK, 30. Morris M, Hamto P, Adame A, Devidze N, Masliah
Trojanowski JQ, et al. Age-dependent emergence and E, Mucke L.  Age-appropriate cognition and subtle
progression of a tauopathy in transgenic mice over- dopamine-­ independent motor deficits in aged Tau
expressing the shortest human tau isoform. Neuron. knockout mice. Neurobiol Aging. 2013;34(6):1523–9.
1999;24(3):751–62. 31. Götz J, Xia D, Leinenga G, Chew YL, Nicholas

16. Ishihara T, Zhang B, Higuchi M, Yoshiyama Y,
H. What renders TAU toxic. Front Neurol. 2013;4:72.
Trojanowski JQ, Lee VM.  Age-dependent induction 32. Gilley J, Seereeram A, Ando K, Mosely S, Andrews
of congophilic neurofibrillary tau inclusions in tau S, Kerschensteiner M, et  al. Age-dependent axo-
transgenic mice. Am J Pathol. 2001;158(2):555–62. nal transport and locomotor changes and tau hypo-
17. Götz J, Tolnay M, Barmettler R, Chen F, Probst A, phosphorylation in a “P301L” tau knockin mouse.
Nitsch RM.  Oligodendroglial tau filament forma- Neurobiol Aging. 2012;33(3):621 e1–e15.
tion in transgenic mice expressing G272V tau. Eur 33. Xia D, Gutmann JM, Götz J. Mobility and subcellu-
J Neurosci. 2001;13(11):2131–40. lar localization of endogenous, gene-edited Tau dif-
18. Santacruz K, Lewis J, Spires T, Paulson J, Kotilinek fers from that of over-expressed human wild-type and
L, Ingelsson M, et al. Tau suppression in a neurode- P301L mutant Tau. Sci Rep. 2016;6:29074.
generative mouse model improves memory function. 34. Saito T, Matsuba Y, Mihira N, Takano J, Nilsson P,
Science. 2005;309(5733):476–81. Itohara S, et al. Single App knock-in mouse models of
19. Hochgrafe K, Sydow A, Mandelkow EM. Regulatable Alzheimer’s disease. Nat Neurosci. 2014;17(5):661–3.
transgenic mouse models of Alzheimer disease: onset, 35. Masuda A, Kobayashi Y, Kogo N, Saito T, Saido

reversibility and spreading of Tau pathology. FEBS TC, Itohara S.  Cognitive deficits in single App
J. 2013;280:4371–81. knock-in mouse models. Neurobiol Learn Mem.
20. Bodea LG, Evans HT, Van der Jeugd A, Ittner LM, 2016;135:73–82.
Delerue F, Kril J, et al. Accelerated aging exacerbates 36. Goedert M. NEURODEGENERATION. Alzheimer’s
a pre-existing pathology in a tau transgenic mouse and Parkinson’s diseases: the prion concept in rela-
model. Aging Cell. 2017;16(2):377–86. tion to assembled Abeta, tau, and alpha-synuclein.
21. Andorfer C, Kress Y, Espinoza M, de Silva R, Tucker Science. 2015;349(6248):1255555.
KL, Barde YA, et  al. Hyperphosphorylation and 37. Götz J, Ittner LM. Animal models of Alzheimer’s dis-
aggregation of tau in mice expressing normal human ease and frontotemporal dementia. Nat Rev Neurosci.
tau isoforms. J Neurochem. 2003;86(3):582–90. 2008;9(7):532–44.
22. Harada A, Oguchi K, Okabe S, Kuno J, Terada S, 38. Handley RR, Reid SJ, Patassini S, Rudiger SR,

Ohshima T, et  al. Altered microtubule organization Obolonkin V, McLaughlan CJ, et  al. Metabolic dis-
in small-calibre axons of mice lacking tau protein. ruption identified in the Huntington’s disease trans-
Nature. 1994;369(6480):488–91. genic sheep model. Sci Rep. 2016;6:20681.
23. Dawson HN, Ferreira A, Eyster MV, Ghoshal N,
39. Yan S, Tu Z, Liu Z, Fan N, Yang H, Yang S, et al. A
Binder LI, Vitek MP. Inhibition of neuronal matura- huntingtin knockin pig model recapitulates features of
tion in primary hippocampal neurons from tau defi- selective neurodegeneration in Huntington’s disease.
cient mice. J Cell Sci. 2001;114(Pt 6):1179–87. Cell. 2018;173(4):989–1002.. e13
24. Tucker KL, Meyer M, Barde YA. Neurotrophins are 40. Lee SE, Hyun H, Park MR, Choi Y, Son YJ, Park YG,
required for nerve growth during development. Nat et al. Production of transgenic pig as an Alzheimer’s
Neurosci. 2001;4(1):29–37. disease model using a multi-cistronic vector system.
25. Ittner LM, Ke YD, Delerue F, Bi M, Gladbach A, PLoS One. 2017;12(6):e0177933.
van Eersel J, et al. Dendritic function of tau mediates 41. Braak H, Braak E.  Staging of Alzheimer’s disease-­
amyloid-beta toxicity in Alzheimer’s disease mouse related neurofibrillary changes. Neurobiol Aging.
models. Cell. 2010;142(3):387–97. 1995;16(3):271–8; discussion 8-84
26. McDermott JB, Aamodt S, Aamodt E. ptl-1, a
42. Clavaguera F, Bolmont T, Crowther RA, Abramowski
Caenorhabditis elegans gene whose products are D, Frank S, Probst A, et al. Transmission and spread-
homologous to the tau microtubule-associated pro- ing of tauopathy in transgenic mouse brain. Nat Cell
teins. Biochemistry. 1996;35(29):9415–23. Biol. 2009;11(7):909–13.
27. Gordon P, Hingula L, Krasny ML, Swienckowski JL, 43. Khan UA, Liu L, Provenzano FA, Berman DE,

Pokrywka NJ, Raley-Susman KM.  The invertebrate Profaci CP, Sloan R, et  al. Molecular drivers and
microtubule-associated protein PTL-1 functions in cortical spread of lateral entorhinal cortex dysfunc-
390 J. J. Götz and J. Götz

tion in preclinical Alzheimer’s disease. Nat Neurosci. with Alzheimer disease-like pathology. Proc Natl
2014;17(2):304–11. Acad Sci U S A. 1997;94(24):13287–92.
44. Polanco JC, Scicluna BJ, Hill AF, Götz J. Extracellular 57. Xia D, Götz J. Premature lethality, hyperactivity, and
vesicles isolated from the brains of rTg4510 mice aberrant phosphorylation in transgenic mice express-
seed tau protein aggregation in a threshold-dependent ing a constitutively active form of Fyn. Front Mol
manner. J Biol Chem. 2016;291(24):12445–66. Neurosci. 2014;7:40.
45. Baker S, Polanco JC, Götz J.  Extracellular vesicles 58. Hampel H, Buerger K, Zinkowski R, Teipel SJ,

containing P301L mutant tau accelerate pathologi- Goernitz A, Andreasen N, et  al. Measurement of
cal tau phosphorylation and oligomer formation but phosphorylated tau epitopes in the differential diagno-
do not seed mature neurofibrillary tangles in ALZ17 sis of Alzheimer disease: a comparative cerebrospinal
mice. J Alzheimers Dis. 2016;54(3):1207–17. fluid study. Arch Gen Psychiatry. 2004;61(1):95–102.
46. Polanco JC, Li C, Durisic N, Sullivan R, Gotz
59. Olsson A, Vanderstichele H, Andreasen N, De Meyer
J.  Exosomes taken up by neurons hijack the endo- G, Wallin A, Holmberg B, et  al. Simultaneous mea-
somal pathway to spread to interconnected neurons. surement of beta-amyloid(1-42), total tau, and phos-
Acta Neuropathol Commun. 2018;6(1):10. phorylated tau (Thr181) in cerebrospinal fluid by the
47. Ahmed RM, Irish M, van Eersel J, Ittner A, Ke YD, xMAP technology. Clin Chem. 2005;51(2):336–45.
Volkerling A, et al. Mouse models of frontotemporal 60. Vanderweyde T, Apicco DJ, Youmans-Kidder K,

dementia: A comparison of phenotypes with clinical Ash PEA, Cook C, Lummertz da Rocha E, et  al.
symptomatology. Neurosci Biobehav Rev. 2017;74(Pt Interaction of tau with the RNA-binding protein TIA1
A):126–38. regulates tau pathophysiology and toxicity. Cell Rep.
48. Menuet C, Cazals Y, Gestreau C, Borghgraef P, Gielis 2016;15(7):1455–66.
L, Dutschmann M, et  al. Age-related impairment of 61. Apicco DJ, Ash PEA, Maziuk B, LeBlang C, Medalla
ultrasonic vocalization in Tau.P301L mice: possible M, Al Abdullatif A, et al. Reducing the RNA binding
implication for progressive language disorders. PLoS protein TIA1 protects against tau-mediated neurode-
One. 2011;6(10):e25770. generation in vivo. Nat Neurosci. 2018;21(1):72–80.
49. Van der Jeugd A, Vermaercke B, Halliday GM,
62. Ittner LM, Fath T, Ke YD, Bi M, van Eersel J, Li KM,
Staufenbiel M, Gotz J. Impulsivity, decreased social et al. Parkinsonism and impaired axonal transport in
exploration, and executive dysfunction in a mouse a mouse model of frontotemporal dementia. Proc Natl
model of frontotemporal dementia. Neurobiol Learn Acad Sci U S A. 2008;105(41):15997–6002.
Mem. 2016;130:34–43. 63. Ittner LM, Ke YD, Götz J.  Phosphorylated Tau

50. Kosik KS, Finch EA.  MAP2 and tau segregate into interacts with c-Jun N-terminal kinase-interacting
dendritic and axonal domains after the elaboration of protein 1 (JIP1) in Alzheimer disease. J  Biol Chem.
morphologically distinct neurites: an immunocyto- 2009;284(31):20909–16.
chemical study of cultured rat cerebrum. J Neurosci. 64. Cummins N, Götz J.  Shedding light on mitoph-

1987;7(10):3142–53. agy in neurons: what is the evidence for PINK1/
51. Wang JZ, Xia YY, Grundke-Iqbal I, Iqbal K. Abnormal Parkin mitophagy in  vivo? Cell Mol Life Sci.
hyperphosphorylation of tau: sites, regulation, and 2018;75(7):1151–62.
molecular mechanism of neurofibrillary degeneration. 65. Cummins N, Tweedie A, Zuryn S, Bertran-Gonzalez
J Alzheimers Dis. 2013;33(Suppl 1):S123–39. J, Götz J.  Disease-associated tau impairs mitophagy
52. Li X, Kumar Y, Zempel H, Mandelkow EM, Biernat by inhibiting Parkin translocation to mitochondria.
J, Mandelkow E.  Novel diffusion barrier for axonal EMBO J. 2018;38(3):e99360.
retention of Tau in neurons and its failure in neurode- 66. David DC, Hauptmann S, Scherping I, Schuessel

generation. EMBO J. 2011;30:4825–37. K, Keil U, Rizzu P, et  al. Proteomic and func-
53. Sohn PD, Tracy TE, Son HI, Zhou Y, Leite RE, Miller tional analysis reveal a mitochondrial dysfunc-
BL, et al. Acetylated tau destabilizes the cytoskeleton tion in P301L tau transgenic mice. J  Biol Chem.
in the axon initial segment and is mislocalized to the 2005;280(25):23802–14.
somatodendritic compartment. Mol Neurodegener. 67. Duboff B, Götz J, Feany MB. Tau promotes neurode-
2016;11(1):47. generation via DRP1 mislocalization in vivo. Neuron.
54. Lee G, Thangavel R, Sharma VM, Litersky JM,
2012;75(4):618–32.
Bhaskar K, Fang SM, et al. Phosphorylation of tau by 68. Li C, Götz J. Tau-based therapies in neurodegenera-
fyn: implications for Alzheimer’s disease. J Neurosci. tion  – opportunities and challenges. Nat Rev Drug
2004;24(9):2304–12. Discov. 2017;16(12):863–83.
55.
Li C, Götz J.  Somatodendritic accumulation 69. Morris M, Maeda S, Vossel K, Mucke L. The many
of Tau in Alzheimer’s disease is promoted by faces of tau. Neuron. 2011;70(3):410–26.
Fyn-mediated local protein translation. EMBO 70. Yanamandra K, Kfoury N, Jiang H, Mahan TE, Ma
J. 2017;36(21):3120–38. S, Maloney SE, et  al. Anti-tau antibodies that block
56.
Sturchler-Pierrat C, Abramowski D, Duke M, tau aggregate seeding in  vitro markedly decrease
Wiederhold KH, Mistl C, Rothacher S, et  al. Two pathology and improve cognition in  vivo. Neuron.
amyloid precursor protein transgenic mouse models 2013;80(2):402–14.
28  Experimental Models of Tauopathy – From Mechanisms to Therapies 391

71. Funk KE, Mirbaha H, Jiang H, Holtzman DM,

81. Nisbet RM, van der Jeugd A, Leinenga G, Evans HT,
Diamond MI.  Distinct therapeutic mechanisms Janowicz PW, Götz J.  Combined effects of scanning
of tau antibodies: promoting microglial clearance ultrasound and a tau-specific single chain antibody in a tau
versus blocking neuronal uptake. J  Biol Chem. transgenic mouse model. Brain. 2017;140(5):1220–30.
2015;290(35):21652–62. 82. Hatch RJ, Leinenga G, Götz J.  Scanning ultrasound
72. Leinenga G, Langton C, Nisbet R, Götz J. Ultrasound (SUS) causes no changes to neuronal excitability and
treatment of neurological diseases – current and emerg- prevents age-related reductions in hippocampal CA1
ing applications. Nat Rev Neurol. 2016;12(3):161–74. dendritic structure in wild-type mice. PLoS One.
73. Choi JJ, Selert K, Vlachos F, Wong A, Konofagou 2016;11(10):e0164278.
EE.  Noninvasive and localized neuronal delivery 83. Blackmore DG, Turpin F, Mohamed AZ, Zong F,

using short ultrasonic pulses and microbubbles. Proc Pandit R, Pelekanos M, et  al. Multimodal analysis
Natl Acad Sci U S A. 2011;108(40):16539–44. of aged wild-type mice exposed to repeated scanning
74.
Raymond SB, Skoch J, Hynynen K, Bacskai ultrasound treatments demonstrates long-term safety.
BJ.  Multiphoton imaging of ultrasound/Optison Theranostics. 2018;8(22):6233–47.
mediated cerebrovascular effects in  vivo. J  Cereb 84. Pelekanos M, Leinenga G, Odabaee M, Odabaee M,
Blood Flow Metab. 2007;27(2):393–403. Saifzadeh S, Steck R, et al. Establishing sheep as an
75. Caskey CF, Stieger SM, Qin S, Dayton PA, Ferrara experimental species to validate ultrasound-mediated
KW.  Direct observations of ultrasound microbubble blood-brain barrier opening for potential therapeutic
contrast agent interaction with the microvessel wall. interventions. Theranostics. 2018;8(9):2583–602.
J Acoust Soc Am. 2007;122(2):1191–200. 85. Hynynen K. Focused ultrasound for blood-brain dis-
76. McDannold N, Arvanitis CD, Vykhodtseva N,
ruption and delivery of therapeutic molecules into the
Livingstone MS. Temporary disruption of the blood-­ brain. Expert Opin Drug Deliv. 2007;4(1):27–35.
brain barrier by use of ultrasound and microbubbles: 86. Bademosi AT, Lauwers E, Amor R, Verstreken P, van
safety and efficacy evaluation in rhesus macaques. Swinderen B, Meunier FA.  In vivo single-molecule
Cancer Res. 2012;72(14):3652–63. tracking at the Drosophila presynaptic motor nerve
77. Jordao JF, Thevenot E, Markham-Coultes K, Scarcelli terminal. J Vis Exp. 2018;131:e56952.
T, Weng YQ, Xhima K, et  al. Amyloid-beta plaque 87. Ellenbroek B, Youn J.  Rodent models in neurosci-
reduction, endogenous antibody delivery and glial ence research: is it a rat race? Dis Model Mech.
activation by brain-targeted, transcranial focused 2016;9(10):1079–87.
ultrasound. Exp Neurol. 2013;248:16–29. 88. Abud EM, Ramirez RN, Martinez ES, Healy LM,
78.
Leinenga G, Götz J.  Scanning ultrasound Nguyen CHH, Newman SA, et  al. iPSC-derived
removes amyloid-beta and restores memory in an human microglia-like cells to study neurological dis-
Alzheimer’s disease mouse model. Sci Transl Med. eases. Neuron. 2017;94(2):278–93.. e9
2015;7(278):278ra33. 89. Bian S, Repic M, Guo Z, Kavirayani A, Burkard T,
79. Leinenga G, Götz J. Safety and efficacy of scanning Bagley JA, et  al. Genetically engineered cerebral
ultrasound treatment of aged APP23 mice. Front organoids model brain tumor formation. Nat Methods.
Neurosci. 2018;12:55. 2018;15(8):631–9.
80. Jordao JF, Ayala-Grosso CA, Markham K, Huang
90. Huch M, Knoblich JA, Lutolf MP, Martinez-Arias
Y, Chopra R, McLaurin J, et  al. Antibodies targeted A.  The hope and the hype of organoid research.
to the brain with image-guided focused ultrasound Development. 2017;144(6):938–41.
reduces amyloid-beta plaque load in the TgCRND8 91. Pham MT, Pollock KM, Rose MD, Cary WA, Stewart
mouse model of Alzheimer’s disease. PLoS One. HR, Zhou P, et al. Generation of human vascularized
2010;5(5):e10549. brain organoids. Neuroreport. 2018;29(7):588–93.
 Cerebrospinal Fluid and Plasma
Tau as a Biomarker for Brain 29
Tauopathy

Mikio Shoji

Introduction and correlated with severity of dementia. In this

assay, AT120 was adopted for captured antibody
Presence of tau in cerebrospinal fluid (CSF) was and combination with HT7 and BT2 was used for
first discovered by Vandermeeren in 1993 [1]. reporter antibodies [6]. Blennow first reported
They developed specific monoclonal antibodies the presence of phosphorylated tau at threonine
against human tau, and phosphorylated tau (ptau) 181 (p181tau) and threonine 231 (ptau231) by
[2, 3], and set up specific ELISA consisted of ELISA using AT180 for ptau231 and AT270 for
AT120 antibody and rabbit anti-human tau anti- ptau181 as captured antibodies and combination
serum. Quantitation of total 190 CSF samples of HT7 and AT120 for reporter antibodies.
from Alzheimer’s disease, controls and other Although tau and ptau levels were significantly
neurological disease showed assay detection increased compared with controls, overlap of
limit of CSF tau was less than 5  pg/ml and assay values were observed also between AD and
increased tau levels in AD compared to those of other neurological diseases [7]. Presently, this
controls. However, marked overlap value between p181tau assay system is altered to use HT7 for
AD and other neurological diseases was observed capture and AT270 for detection antibodies [8].
in this initial report. In 1995, 5 assay results of Arai showed significant increase of CSF tau and
CSF tau were published at once. Vigo-Pelfrey presence 50~65 kd tau bands in CSF samples by
group measured CSF tau from 181 patients using western blot suggesting that CSF tau might
ELISA using 16G7 and 16B5 antibody to tau [4]. reflect the progressive accumulation of altered
We also developed different ELISA system using tau due to progressive death of neurons in the AD
anti-Ht-2 and F-F11 antibodies and confirmed brain [9]. This is short history of developing of
increased CSF tau in AD patients [5]. Data using CSF tau and ptau assay. Afterward these basic
present common ELISA system for tau and ptau findings, huge number of studies of CSF tau and
provided Innogenetics N.V. in Belgium (now ptau open up doors to the definite biomarkers for
Fujirebio Europe N.V.) published from this year. diagnosis and prediction of AD, and the clarifica-
Hock reported that CSF tau levels are increased tion of tauopathy mechanisms, establishment for
in AD, preclinical stage of AD, familial AD cases, global standardization, and finally valuable tools
for development of essential therapy for AD.
Major papers about CSF tau studies are
M. Shoji (*) reviewed and summarized here, and the vison of
Department of Neurology, Geriatrics Research future is commented.
Institute and Hospital, Maebashi, Japan
e-mail: m-shoji@ronenbyo.or.jp

© Springer Nature Singapore Pte Ltd. 2019 393

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_29
394 M. Shoji

Definite Biomarker for AD demented neurological disease and 181 normal

controls showed the cut-off value of CSF tau was
In the same 1995, another way of AD biomarker 375  pg/ml, 59% sensitivity and 90% specificity
had initiated. Motter first reported that CSF Aß42 for diagnosis AD compared to other groups.
levels were decreased in spite of increased levels Simultaneously, elevation of CSF tau level was
of CSF tau in AD patient [10]. CSF samples from observed in other chronic and acute brain damage
37 AD, 32 neurological disease and 20 nonde- disease suggesting required attention for clinical
mented controls were evaluated by ELISA using practice [16]. Andreasen also showed sensitivity,
266/277-2 antibodies for Aß42 and 16B5/16G7 specificity and stability of CSF tau in AD in a
antibodies for tau. In 1998, Kanai published community-based patient samples showed the
the  first longitudinal and multicenter study of cut-off value of CSF tau was 302  pg/ml, 93%
CSF Aß40, Aß42 and tau. The study consisted of sensitivity and 86% specificity for diagnosis AD
93 AD, 33 non-AD dementia, 56 other neurologi- compared with controls and suggested some neu-
cal disease CSF samples were evaluated by rological conditions (e.g., stroke) increases CSF
Innogenetics tau assay and common Aß ELISAs tau [17]. Finally, these findings were validated by
of BAN-50 for capture and BA-27/BC-05 for prospective comparisons between antemortem
detecting Aß40/Aß42, respectively. A significant CSF tau and Aß and autopsy-confirmed dementia
elevation of tau and correlation between the tau diagnosis [18].
levels and the clinical progression were observed
in AD. A significant decrease of the Aß42 levels
and a significant increase of Aß40/42 ratio were Systemic Review and Meta-Analysis
observed in AD suggesting that CSF tau increases
with clinical progression of dementia and altera- From these basic findings era, CSF biomarker
tion of Aß42 and Aß40/42 ratio starts at early study expanded to clarify the diagnostic efficacy
stages in AD.  Efficient diagnostic sensitivity of CSF tau and ptau in huge number of cohort
(71%) and specificity (83%) were  revealed  by studies including AD dementia, mild cognitive
using combination of tau and Aß40/42 ratio val- impairment (MCI) due to AD, cognitively unim-
ues. Improvement in sensitivity up to 91% was paired (CU)  state in AD and other neurological
obtained in longitudinal evaluation [11]. diseases. According to recent systemic review by
Additional reports by different assay systems Olsson, 231 articles comprising 15,699 AD and
supported these findings [12, 13]. 13,018 controls were analyzed and showed CSF
Different assay methods of ptau also reported. tau, ptau and Aß42 strongly differentiated AD
Increased level of ptau at serine 199 using ELISA form controls, and MCI due to AD from stable
with HT-7 and Anti-PS199 was published by Ito MCI. Total 164 cohorts with AD and 153 control
in a  large scale multicenter study consisted of cohorts representing 11,341 AD and 7,086 con-
570 CSF samples, in 2001 [14]. However, in trols showed that average tau concentration ratio
global standardization study in comparative CSF between AD and controls are 2.54. In CSF ptau,
study among ptau231, ptau181, and ptau199 98 studies comprising 7,498 AD from 96 cohorts
showed that ptau181 assay reached specificity and 5,126 controls from 91 cohorts showed the
levels greater than 75% when sensitivity was set ptau  concentration ratio is 1.88. Comparison
at 85% or greater [15]. between 12 cohorts with 307 MCI and 570 stable
For clinical application in differential diagno- MCI showed that average concentration ratio of
sis of AD, another large-scale multicenter study CSF tau is 1.76 and those analysis with 9 cohorts
by Shoji analyzing total 1,031 samples from 366 comprising 251 MCI due to AD and 501 stable
AD and 168 non-Alzheimer dementia, 316 non-­ MCI of CSF ptau is 1.72 [19].
29  Cerebrospinal Fluid and Plasma Tau as a Biomarker for Brain Tauopathy 395

Differential Diagnosis 738.4  ±  290.6  pg/ml in AD dementia/MCI

and Prediction for Onset of AD (p < 0.0001), 1,337 ± 1554 pg/ml in encephalop-
athy (ENC) (p = 0.036), and 415.7 ± 158.2 pg/ml
During these 20 years, based on development of in multiple system atrophy (MSA) (p = 0.0164)
basic research and neuroimaging tools, clinical than in CU. One patient with CJD had a CSF tau
classification and diagnosis criteria of neurogen- level of 1,554 pg/ml (Fig. 29.1a). P181tau levels
erative dementia have been refined and subdi- were 41.69  pg/ml in CU and significantly
vided in detail. These developments facilitate increased to 88.62 ± 29.69 pg/ml in AD demen-
rigorous evaluation of CSF biomarkers in newly tia/MCI (p  <  0.0001). No significant changes
identified neurogenerative diseases. Toledo were observed in other diseases. The CSF tau/
examined pathology confirmed neurodegenera- p181 tau ratio was 6.0 ± 1.8 in CU, and increased
tive dementia patients and showed 26.8% of AD to 8.2 ± 1.2 in AD dementia/MCI (p = 0.0038),
combines another pathology and 14~17% under- 38.8 ± 55.6 in ENC (p = 0.013), and 10.2 ± 3.1 in
estimation of biomarker accuracy. CSF tau and MSA (P < 0.0001). In CJD, this ratio was 25.9.
ptau are increased in AD and AD-Dementia with Specific changes due to AD processes were rec-
Lewy bodies (DLB) [20]. Both tau and ptau val- ognized in P181tau levels. This study corre-
ues themselves did not discriminate behavior sponded of previous 1,031 subjects multicenter
variant frontotemporal lobar degeneration study showing overlap values in the tauopathy
(bvFTLD) and frontotemporal dementia (FTD) and other neurological disease groups, with mod-
without combination of Aß42 values [21, 22]. erate measurement sensitivity and specificity as a
The systemic review and meta-analysis of idio- biomarker using mean ± 2 SD as a cutoff value
pathic normal-pressure hydrocephalus indicated [16]. Total tau was increased in some diseases
significantly reduced levels of tau, ptau and Aß42 because of different pathological processes,
compared to healthy normal state [23]. In including tauopathy due to AD, acute brain injury
Creutzfeldt-Jakob disease (CJD), highly elevated by ENC and CJD, and axonal degeneration in
CSF levels of tau and 14-3-3 protein are estab- MSA.  No significant changes were detected in
lished biomarkers. Rumeileh showed tau yielded total tau, p181tau, or their ratio in CBD, PSP, or
80.6% sensitivity and 75.3% specificity in distin- FTD. The present results on the CSF tau/p181tau
guishing AD from CJD.  However, ptau alone ratio suggest that it was 6:1 and that the
showed no eligible significance. Cut-off value phosphorylated-­ tau tangle pathology increased
was proposed to be >1200 mg in tau and >16.4 in p181tau and secondarily induced brain injury due
tau/ptau ratio [24]. Ewers studied diagnostic to increased total tau levels (8:1), even in AD.
value of CSF Aß42, tau and ptau in 675 CSF sam-
ples from controls, AD dementia, subjective
memory impairment, vascular dementia, LBD,  DNI and DIAN Study
A
and FTD, depression, and other neurological dis- Demonstrated Signature of AD
ease. As the results, Aß42 showed the best diag-
nostic accuracy among them. At a sensitivity of Then, epochal 2 global studies initiated; One is
85%, the specificity to differentiate AD dementia Alzheimer’s Disease Neuroimaging Initiative
against other diagnosis ranged from 42% for (ADNI) from 2003, which demonstrated natural
DLB, 77% for FTD. However, significant overlap course of cognitive function, neuroimaging and
with other non-AD dementia, possibly reflected CSF biomarkers in cognitively unimpaired (CU),
the underlying mixed pathology [25]. MCI and dementia stage of pure sporadic
Recently, we have independently reexamined AD. Another is further convincing study of sig-
CSF tau and ptau in a total of 213 CSF samples nature of biomarkers in dominantly inherited AD,
from various neurological diseases and CU sub- Dominantly Inherited Alzheimer’s Disease
jects [26]. Tau levels were 259.3 ± 162.8 pg/ml in (DIAN) from 2008.
CU, and were significantly higher at
396 M. Shoji

Fig. 29.1  Total tau, phosphorylated-tau in CSF from 231 Parkinson’s disease, MSA multiple system atrophy, CBD
neurological diseases corticobasal degeneration, PSP progressive supranuclear
∗: p < 0.05; ∗∗: p < 0.005; ∗∗∗: p < 0.0001 palsy, SCD spinocerebellar degeneration, ALS amyo-
Abbreviations: ADD Alzheimer dementia, MCI mild cog- trophic lateral sclerosis, PN polyneuropathy, CU cogni-
nitive impairment, FTD frontotemporal dementia, NPH tively unimpaired control subjects. Units of CSF total tau
normal pressure hydrocephalus, ENC meningoencephali- and phosphorylated-tau were pg/mL
tis, MS multiple sclerosis, NMO neuromyelitis optica, PD

Shaw showed that CSF Aß42 was the most center CV% was 17.9% in Aß42, 13.1% in tau
sensitive biomarker for AD in ADNI cohort and and 14.6% in ptau181 [28]. Follow-up during
autopsy-confirmed subjects. Cut-offs, sensitivity 48 months in ADNI cohort showed that low Aß42
and specificity discriminating between AD and values were associated longitudinal increase in
CU subjects were 93 pg/ml, 69.6% and 92.3% in ptau181, and high baseline ptau181 values were
tau, 23  pg/ml, 67.9%, and 73.1% in ptau 181, conversely not associated win changes of Aß42
192 pg/ml, 96.4% and 76.9% in Aß42, and 0.39, levels [29]. Recent report from ADNI consisted
85.7% and 84.6% in tau/Aß42 [27]. Multicenter of 56 CU, 73 MCI and 17 AD over 1~7  years
quality control study of ADNI samples using follow-up divided by Aß+ and Aß− groups
INNO-BIA AlzBIo3, xMAP technology showed depend on Aß42 cut-offs 192 pg/ml, showed sig-
intra center assay CV% was 5.3% in Aß42, 6.7% nificantly increased baseline levels of CSF tau
in tau and 10.8% in ptau181. Those of inter-­ and ptau in Aß + CU, Aß + MCI and Aß + AD
29  Cerebrospinal Fluid and Plasma Tau as a Biomarker for Brain Tauopathy 397

dementia. Longitudinally, tau levels increased in specificity 83% in combination with tau and
both Aß + CU and Aß + MCI, but, decreased in Aß42, and those of 95% and 87% in combination
Aß  +  AD dementia. Longitudinally, ptau levels ptau181 and Aß42 for detection AD dementia
increased in Aß + CU and significantly decline in converter [33]. DESCRIPIA prospective study
Aß + AD dementia. Both follow-up study showed from 20 memory clinic across Europe during
increase of tau and ptau mainly increase MCI 2003–2005 showed CSF AD profile (low Aß42/
stage, and conversely decline in dementia stage high tau value) was common in 52% of subjec-
[30]. tive cognitive impairment (SCI), 68% of non-­
DIAN study is the unique study of dominantly amnestic MCI (naMCI), 79% of amnestic MCI
inherited AD (DIAD). DIAD is caused by muta- (aMCI) and 31% CU and associated with cogni-
tions of APP, APP duplication and PSEN-1/-2 tive decline in naMCI and aMCI [34]. In subjects
mutations. Although the onset age is variable with MCI and abnormal CSF Aß42 profile, CSF
depend on each gene mutation type, penetrance is tau and ptau and hippocampal atrophy can pre-
100% and the onset age and prognosis is essen- dict further cognitive decline [35]. Prospective
tially identical with carrier’s parent with 9-year study of 44 CU showed that 6 of 12 with
DIAD.  These findings indicate that survey of low baseline CSF Aß42 developed AD, but other
mutation carrier can reveal definite preclinical 6 subjects with low Aß42 and 32 with normal
alteration and natural course of biomarkers and Aß42 did not develop AD. CSF tau and ptau did
cognition before onset. In 2012, Bateman clearly not predict development AD/DLB over 9  years
showed the order and magnitude of pathologic [36].
processes in AD.  CSF Aß42 levels appeared to Adult Children Study of 169 middle-aged CU
decline 25  years before the symptom onset. Aß during 6  years prospectively also revealed
deposition in the brain detected by PiB amyloid that longitudinal reduction in Aß42 were observed
PET, increased CSF tau and brain atrophy initi- in some individuals as early middle age and low
ated before 15 years of onset. Cerebral hypome- Aß42 levels were associated with the develop-
tabolism and episodic memory disturbance ment of cortical amyloid deposition, especially in
appeared before 10 years. Finally, global cogni- mid middle age during 55–64 years. CSF tau and
tive impairment was detected 5  years before ptau as neuronal injury markers dramatically
onset. Then, 3  years after onset, patients met increased in some individuals in mid and late
diagnostic criteria for dementia [31, 32]. Thus, middle aged during 55–74 years [37]. CSF Aß42
DIAN study conclusively established big data of was correlated only with PiB binding, but, CSF
all alterations of biomarkers and cognitive func- tau, ptau and hippocampal volume were corre-
tions which gradually progress during 25  years lated with the longitudinal alteration in global
and confirmed these orders speculated by ADNI cognition [38, 39]. Toledo recently reported a
study. Both results by ADNI and DIAN study global large multicenter study of CSF samples
provide us extremely useful tool to open inter- from 1,233 healthy cohort subjects, 40–84 years,
ventions for prevention of AD. from 15 cohorts from 12 different centers by
Luminex® assay of Aß42, tau and ptau in
Gothenburg Laboratory. At 40 years of age, 76%
 rospective Study for Prediction
P of subjects were classified normal Aß42, tau and
of MCI and Dementia Due to AD ptau and their frequency decreased to 32% at
85  years. Normal Aß42 and increased tau/ptau
Natural course of CSF biomarkers of incipient group frequency increased slowly from 1% at
AD from CU to MCI was also studied. Hansson 44 years to 16% at 85 years. Low Aß42 with high
examined 137 MCI during 5.2  years follow up tau/ptau frequency increased from 1% at 53 years
and showed 42% developed AD dementia and to 28% at 85  years. Abnormal low Aß42 were
15% develop other form of dementia. CSF bio- already frequent in middle-life and APOE geno-
markers at baseline yielded sensitivity 95% and type strongly affects the Aß42, tau and ptau in
398 M. Shoji

Swedish BioFINDER (n  =  277) and ADNI effects were less than between-laboratory effects
(n = 646) [40]. [46]. Temperature at stored, non-frozen time,
contamination such as detergent and blood, cen-
trifugation and tube materials have a significant
Standardization and Newly effect on assay variability [47].
Developed Assay Technology

To assay AD biomarkers simultaneously and Origin of CSF Tau and pTau

automatically, the flow cytometric-based
Luminex xMAP® technology involves coupling During these 25 years, no one believed tau is nor-
of specific monoclonal antibody sets to the sur- mally secreted into CSF.  Everyone also tried to
face of microbeads uniquely identified with a clarify the missing link between Aß amyloidosis
combination of fluorescence dyes in a single and tauopathy in AD pathological processes. Sato
sample assuming no cross-reactivity of particular illuminated these issues using kinetic study of
antibodies. The standard assay system named stable isotope labelling tau and mass spectrome-
The INNO-BIA AlzBio3 is commonly used to try in the human central nervous system and
measure Aß42, tau and ptau181 consisted of cor- iPSC-derived neurons. Brain full length tau is
responding capturing antibodies (tau: AT120, C-terminally truncated at residues 210–230 and
ptau181: AT270, Aß42:4D7A3) and biotinylated released from human neurons in 3  days and
detection antibodies (HT7 and 3D6). Intra-and ~14 days into CSF. Average half-life of CSF tau
interassay CVs were less than 10% for all ana- is 23  days and its production rates are
lytes [41]. Multiplexed quantification correlated 26.3 ± 9.2 pg/ml/day. Increase in CSF tau in AD
with results by usual ELISA assay and improved is due to an increase in synthesis and release and
sample management and quality control of assay positively correlated with amyloidosis. There
[42]. Recently, Roche Diagnostics developed was no correlation between tau fraction turnover
Elecsys assays that utilize the automated cobas rate and tau PET imaging. Increased tau produc-
601 analyzer exhibited further precision accu- tion and soluble tau secretion are initiated by
racy, reliability, reproducibility with between-­ amyloid toxicity in very early MCI.  Then,
laboratory CV of approximately 4% [43, 44]. increased aggregated tau and decreasing elevated
Large concordant study between cut-offs for CSF tau appear and induce trans-synaptic spread-
Elecsys assay and amyloid PET using Swedish ing of aggregated tau, causing cortical cognitive
BioFINDER (n = 277), ADNI (n = 646) and clin- function deficits in early to mild AD dementia
ical progression in MCI (n  =  619) showed tau/ stage [48]. He reported that Aß plaque facilitates
Aß42 and ptau/Aß42 ratios were highly concor- the rapid amplification of aggregated tau seeds
dant with PET classification in BioFINDER into large tau aggregates in dystrophic neurites of
(overall agreement: 90%) and ADNI classifica- senile plaques, induces formation and spread of
tion (overall agreement: 89–90%) and predicted neurofibrillary tangles and neuropil threads as
greater 2-year clinical decline in MCI [44]. The secondary seeding events [49].
Alzheimer’s Association quality control program
participated by 40 laboratories in 2011 showed
total CVs among centers were 16–28% for Association Between CSF
ELISA,13–36% for xMAP, and 16–36% for Biomarkers and Newly Developing
Meso Scale Discovery [45]. Extended quality Tau Neuroimaging
control program participated 84 laboratories
reported that CVs between laboratories were Tau PET ligand [18F] flortaucipir have been
around 20–30%; within-run CVs, less than developed to detect in vivo tau accumulation in
5–10%; and longitudinal with-in laboratory CVs AD. Brain tau accumulation initiated from tem-
were 5–19%. For tau and ptau, between-kit lot poral lobe at early MCI stage, progressively
29  Cerebrospinal Fluid and Plasma Tau as a Biomarker for Brain Tauopathy 399

extended to parietofrontal lobes and closely asso- Scale Discovery are very small but totally differ-
ciated with cortical cognitive functions and ent. In former report, plasma ptau181 levels are
severity of dementia. Comparison between post- 0.171~0.045 pg/ml. In the later report, those are
mortem brain pathology and regional in  vivo 6.4~11.6 pg/ml. There are 100 times differences
uptake of [18F] flortaucipir showed close correla- between recent reports. Basic studies such as
tion with density of tau-positive neurites, intra- plasma Aß kinetics from brain to CSF, plasma
somal neurofibrillary tangles and total tau burden. and mass spectrographic identification studies
No correlations between [18F] flortaucipir and Aß have not yet performed. Presence of big tau, a
pathology were found [50]. CSF tau and ptau close homologues with brain tau presented in
increase from preclinical AD, despite normal human body organs and peripheral nerve sys-
[18F] flortaucipir retention, suggesting that tems, has not been clarified yet [58, 59]. Based on
appearance of positive tau PET findings initiates these issues, further developing basic study and
later stage of MCI than those of CSF tau and ptau large scale confirmation studies are expected.
[51, 52].

Recommendation in the Diagnostic
Plasma Phosphorylated Tau Evaluation of MCI and Dementia
as Possible Biomarker for AD
In 2017, evidence-based guidelines in the diag-
Recent studies have clarified that the plasma nostic evaluation of MCI and dementia due to AD
Aß42/40 ratio is inversely correlated with corti- were proposed based on systematic reviews using
cal amyloid burden in AD, which can be con- Grading of Recommendations, Assessment,
verted to MCI, and that the plasma Aß42/40 ratio Development, and Evaluation (GRADE) meth-
is a useful screening marker for brain Aß amyloi- ods by working group comprised 28 international
dosis in normal individuals [53, 54]. In a similar members [60, 61]. The former report recom-
way, quantitation of plasma tau and ptau as a mends the use of CSF markers in predicting the
screening biomarkers for brain tauopathy are functional or cognitive decline or conversion to
developing. Mattsson studied of plasma tau lev- AD dementia within 3 years and counseling both
els using total of 1,284 participants from ADNI before and after the biomarker evaluation. The
and BioFINDER cohorts. Plasma tau partially later report recommends the use of CSF AD bio-
reflects AD pathology, but the overlap between markers as a supplement to clinical evaluation, to
normal aging and AD is large, especially in identify or exclude AD as the cause of dementia,
patients without dementia [55]. Tatebe tried to for prognostic evaluation, and for guiding man-
quantitate plasma ptau181 using modified agement of patients, particularly in atypical and
SimoaTM Tau 2.0 kit on Simoa HD-1 analyzer uncertain cases. As summarized here, huge num-
(Quantrex). Plasma ptau181 levels were signifi- bers of basic and clinical dedications during these
cantly increased in AD and Down syndrome 25  years for CSF biomarkers revealed the total
patients compared to controls [56]. Mielke mea- figures occurred in person due to AD. However,
sured plasma ptau 182 using the Meso Scale essential aim of CSF biomarkers for contribution
Discovery platform with antibody AT270 for cap- of developing disease modifying therapy and
ture ptau181 and antibody SULFO-TAG-LRL for intervention in AD pathological processes are
detection tau from 172 CU, 57 MCI, 40 AD still ongoing. In 2018, National Institute on
dementia with concurrent Aß and tau PET. Plasma Aging -Alzheimer’s Association (NIA-AA) has
tau and ptau181 levels were higher in AD demen- proposed research framework using ATN classifi-
tia than those in CU. Plasma ptau181 was more cation system of biomarkers toward a biological
strongly associated with Aß and tau PET [57]. definition of AD.  In this criteria, AD is consid-
The values of ptau181 measured Simoa or Meso ered as a continuum, and cognitive staging is
400 M. Shoji

AD continuum: cognitive stage

Cognitively Unimpaired (CU) MCI Dementia

CSF Aß42 Positive tau PET

Biomarker Profile

CSF tau and ptau

Aß amyloidosis

normal AD biomarkers,
A-T- (N) - normal AD biomarkers with MCI normal AD biomarkers with dementia
cognitively unimpaired

A+T- (N) - Preclinical AD pathologic change AD pathological change with MCI AD pathologic change with dementia
A+T+(N)-
Preclinical AD AD with MCI AD dementia
A+T+(N)+

Fig. 29.2  NIA-AA Research Framework and natural course of CSF biomarkers

classified into cognitively unimpaired (CU), MCI

References
and dementia. Biomarker profile is classified in
to 3 groups; A: Aggregated Aß or associated 1. Vandermeeren M, Mercken M, Vanmechelen E, Six
pathologic state including CSF Aß42, or Aß42/ J, van de Voorde A, Martin JJ, Cras P.  Detection
Aß42 ratio, amyloid PET; T: Aggregated tau of tau proteins in normal and Alzheimer’s dis-
(neurofibrillary tangles) or associated pathologic ease cerebrospinal fluid with a sensitive sandwich
enzyme-linked immunosorbent assay. J  Neurochem.
state including CSF ptau and tau PET; (N): neu- 1993;61(5):1828–34. PubMed PMID: 8228996
rodegeneration or neuronal injury including 2. Mercken M, Vandermeeren M, Lübke U, Six J,
Anatomical MRI, FDG PET and CSF tau Boons J, Vanmechelen E, Van de Voorde A, Gheuens
(Fig. 29.2). Based on this novel NIA-AA research J.  Affinity purification of human tau proteins and
the construction of a sensitive sandwich enzyme-­
framework criteria, AD process in human brain linked immunosorbent assay for human tau detection.
will be biologically defined further and essential J  Neurochem. 1992;58(2):548–53. PubMed PMID:
therapy of next generation will be evaluated 1729400
hopefully [62]. 3. Goedert M, Jakes R, Crowther RA, Cohen P,
Vanmechelen E, Vandermeeren M, Cras P.  Epitope
mapping of monoclonal antibodies to the paired
Acknowledgments This study was supported by the helical filaments of Alzheimer’s disease: identifica-
Amyloidosis Research Committee Surveys and Research tion of phosphorylation sites in tau protein. Biochem
on Special Diseases, the Longevity Science Committee of J 1994;301 (Pt 3):871–877. PubMed PMID: 7519852;
the Ministry of Health and Welfare of Japan; Scientific PubMed Central PMCID: PMC1137067.
Research (C) (18K07385 MS) from the Ministry of 4. Vigo-Pelfrey C, Seubert P, Barbour R, Blomquist C,
Education, Science, and Culture of Japan; the Hirosaki Lee M, Lee D, Coria F, Chang L, Miller B, Lieberburg
University Institutional Research Grant, and the Center of I, et  al. Elevation of microtubule-associated pro-
Innovation Science and Technology-based Radical tein tau in the cerebrospinal fluid of patients with
Innovation and Entrepreneurship Program from the Japan Alzheimer’s disease. Neurology. 1995;45(4):788–93.
Science and Technology Agency. PubMed PMID: 7723971
5. Mori H, Hosoda K, Matsubara E, Nakamoto T, Furiya
Potential Conflicts of Interest  The authors have no con- Y, Endoh R, Usami M, Shoji M, Maruyama S, Hirai
flict of interest to report. S.  Tau in cerebrospinal fluids: establishment of the
sandwich ELISA with antibody specific to the repeat
29  Cerebrospinal Fluid and Plasma Tau as a Biomarker for Brain Tauopathy 401

sequence in tau. Neurosci Lett. 1995;186(2–3):181–3. K.  Large-scale, multicenter study of cerebrospinal
PubMed PMID: 7777192 fluid tau protein phosphorylated at serine 199 for the
6. Hock C, Golombowski S, Naser W, Müller-Spahn antemortem diagnosis of Alzheimer’s disease. Ann
F. Increased levels of tau protein in cerebrospinal fluid Neurol. 2001;50(2):150–6. PubMed PMID: 11506396
of patients with Alzheimer’s disease—correlation 15. Hampel H, Buerger K, Zinkowski R, Teipel SJ,

with degree of cognitive impairment. Ann Neurol. Goernitz A, Andreasen N, Sjoegren M, DeBernardis
1995;37(3):414–5. PubMed PMID: 7695246 J, Kerkman D, Ishiguro K, Ohno H, Vanmechelen E,
7. Blennow K, Wallin A, Agren H, Spenger C, Siegfried Vanderstichele H, McCulloch C, Moller HJ, Davies
J, Vanmechelen E.  Tau protein in cerebrospinal P, Blennow K.  Measurement of phosphorylated tau
fluid: a biochemical marker for axonal degenera- epitopes in the differential diagnosis of Alzheimer
tion in Alzheimer disease? Mol Chem Neuropathol. disease: a comparative cerebrospinal fluid study. Arch
1995;26(3):231–45. PubMed PMID: 8748926 Gen Psychiatry. 2004;61(1):95–102. PubMed PMID:
8. Vanmechelen E, Vanderstichele H, Davidsson P, Van 14706948
Kerschaver E, Van Der Perre B, Sjögren M, Andreasen 16. Shoji M, Matsubara E, Murakami T, Manabe Y, Abe
N, Blennow K. Quantification of tau phosphorylated K, Kanai M, Ikeda M, Tomidokoro Y, Shizuka M,
at threonine 181 in human cerebrospinal fluid: a sand- Watanabe M, Amari M, Ishiguro K, Kawarabayashi T,
wich ELISA with a synthetic phosphopeptide for Harigaya Y, Okamoto K, Nishimura T, Nakamura Y,
standardization. Neurosci Lett. 2000;285(1):49–52. Takeda M, Urakami K, Adachi Y, Nakashima K, Arai
PubMed PMID: 1078870 H, Sasaki H, Kanemaru K, Yamanouchi H, Yoshida
9. Arai H, Terajima M, Miura M, Higuchi S, Muramatsu Y, Ichise K, Tanaka K, Hamamoto M, Yamamoto
T, Machida N, Seiki H, Takase S, Clark CM, Lee H, Matsubayashi T, Yoshida H, Toji H, Nakamura
VM, et al. Tau in cerebrospinal fluid: a potential diag- S, Hirai S.  Cerebrospinal fluid tau in dementia dis-
nostic marker in Alzheimer’s disease. Ann Neurol. orders: a large scale multicenter study by a Japanese
1995;38(4):649–52. PubMed PMID: 7574462 study group. Neurobiol Aging. 2002;23(3):363–70.
10. Motter R, Vigo-Pelfrey C, Kholodenko D, Barbour PubMed PMID: 11959397
R, Johnson-Wood K, Galasko D, Chang L, Miller B, 17. Andreasen N, Minthon L, Clarberg A, Davidsson

Clark C, Green R, et  al. Reduction of beta-amyloid P, Gottfries J, Vanmechelen E, Vanderstichele H,
peptide42  in the cerebrospinal fluid of patients with Winblad B, Blennow K.  Sensitivity, specificity, and
Alzheimer’s disease. Ann Neurol. 1995;38(4):643–8. stability of CSF-tau in AD in a community-based
PubMed PMID: 7574461 patient sample. Neurology 1999;53(7):1488–1494.
11. Kanai M, Matsubara E, Isoe K, Urakami K,
PubMed PMID: 10534256.
Nakashima K, Arai H, Sasaki H, Abe K, Iwatsubo 18. Clark CM, Xie S, Chittams J, Ewbank D, Peskind E,
T, Kosaka T, Watanabe M, Tomidokoro Y, Shizuka Galasko D, Morris JC, McKeel DW Jr, Farlow M,
M, Mizushima K, Nakamura T, Igeta Y, Ikeda Y, Weitlauf SL, Quinn J, Kaye J, Knopman D, Arai H,
Amari M, Kawarabayashi T, Ishiguro K, Harigaya Doody RS, DeCarli C, Leight S, Lee VM, Trojanowski
Y, Wakabayashi K, Okamoto K, Hirai S, Shoji JQ.  Cerebrospinal fluid tau and beta-amyloid: how
M. Longitudinal study of cerebrospinal fluid levels of well do these biomarkers reflect autopsy-confirmed
tau, A beta1-40, and A beta1-42(43) in Alzheimer’s dementia diagnoses? Arch Neurol 2003;60(12):1696–
disease: a study in Japan. Ann Neurol. 1998;44(1):17– 1702. PubMed PMID: 14676043.
26. PubMed PMID: 9667589 19. Olsson B, Lautner R, Andreasson U, Öhrfelt A,

12. Galasko D, Chang L, Motter R, Clark CM, Kaye J, Portelius E, Bjerke M, Hölttä M, Rosén C, Olsson
Knopman D, Thomas R, Kholodenko D, Schenk D, C, Strobel G, Wu E, Dakin K, Petzold M, Blennow
Lieberburg I, Miller B, Green R, Basherad R, Kertiles K, Zetterberg H.  CSF and blood biomarkers for the
L, Boss MA, Seubert P. High cerebrospinal fluid tau diagnosis of Alzheimer’s disease: a systematic review
and low amyloid beta42 levels in the clinical diagno- and meta-analysis. Lancet Neurol. 2016;15(7):673–
sis of Alzheimer disease and relation to apolipopro- 84. https://doi.org/10.1016/S1474-4422(16)00070-3.
tein E genotype. Arch Neurol. 1998;55(7):937–45. Epub 2016 Apr 8. Review. PubMed PMID: 27068280
PubMed PMID: 9678311 20. Toledo JB, Brettschneider J, Grossman M, Arnold SE,
13. Hulstaert F, Blennow K, Ivanoiu A, Schoonderwaldt Hu WT, Xie SX, Lee VM, Shaw LM, Trojanowski
HC, Riemenschneider M, De Deyn PP, Bancher C, JQ. CSF biomarkers cutoffs: the importance of coin-
Cras P, Wiltfang J, Mehta PD, Iqbal K, Pottel H, cident neuropathological diseases. Acta Neuropathol.
Vanmechelen E, Vanderstichele H. Improved discrim- 2012;124(1):23–35. https://doi.org/10.1007/s00401-
ination of AD patients using beta-amyloid(1-42) and 012-0983-7. Epub 201e2 Apr 22. PubMed PMID:
tau levels in CSF.  Neurology. 1999;52(8):1555–62. 22526019; PubMed Central PMCID: PMC3551449
PubMed PMID: 10331678 21. Marelli C, Gutierrez LA, Menjot de Champfleur N,
14. Itoh N, Arai H, Urakami K, Ishiguro K, Ohno
Charroud C, De Verbizier D, Touchon J, Douillet
H, Hampel H, Buerger K, Wiltfang J, Otto M, P, Berr C, Lehmann S, Gabelle A.  Late-onset
Kretzschmar H, Moeller HJ, Imagawa M, Kohno behavioral variant of frontotemporal lobar degen-
H, Nakashima K, Kuzuhara S, Sasaki H, Imahori eration versus Alzheimer’s disease: interest of cere-
402 M. Shoji

brospinal fluid biomarker ratios. Alzheimers Dement 2011 Feb 11. PubMed PMID: 21311900; PubMed
(Amst). 2015;1(3):371–9. https://doi.org/10.1016/j. Central PMCID: PMC3175107
dadm.2015.06.004. eCollection 2015 Sep. PubMed 29.
Toledo JB, Xie SX, Trojanowski JQ, Shaw
PMID: 27239517; PubMed Central PMCID: LM. Longitudinal change in CSF Tau and Aβ biomark-
PMC4878372 ers for up to 48 months in ADNI. Acta Neuropathol.
22. Vergallo A, Carlesi C, Pagni C, Giorgi FS, Baldacci 2013;126(5):659–70. https://doi.org/10.1007/s00401-
F, Petrozzi L, Ceravolo R, Tognoni G, Siciliano G, 013-1151-4. Epub 2013 Jun 29. PubMed PMID:
Bonuccelli U. A single center study: Aβ42/p-Tau(181) 23812320; PubMed Central PMCID: PMC3875373
CSF ratio to discriminate AD from FTD in clinical 30. Sutphen CL, McCue L, Herries EM, Xiong C,

setting. Neurol Sci. 2017;38(10):1791–7. https://doi. Ladenson JH, Holtzman DM, Fagan AM. ADNI.
org/10.1007/s10072-017-3053-z. Epub 2017 July 19. Longitudinal decreases in multiple cerebrospinal
PubMed PMID: 28726050 fluid biomarkers of neuronal injury in symptomatic
23. Chen Z, Liu C, Zhang J, Relkin N, Xing Y, Li
late onset Alzheimer's disease. Alzheimers Dement.
Y.  Cerebrospinal fluid Aβ42, t-tau, and p-tau levels 2018;14(7):869–79. https://doi.org/10.1016/j.
in the differential diagnosis of idiopathic normal-­ jalz.2018.01.012. Epub 2018 Mar 23. PubMed PMID:
pressure hydrocephalus: a systematic review and 29580670; PubMed Central PMCID: PMC6110083
meta-analysis. Fluids Barriers CNS. 2017;14(1):13. 31. Bateman RJ, Xiong C, Benzinger TL, Fagan AM,
https://doi.org/10.1186/s12987-017-0062-5. Review. Goate A, Fox NC, Marcus DS, Cairns NJ, Xie X,
PubMed PMID: 28486988; PubMed Central PMCID: Blazey TM, Holtzman DM, Santacruz A, Buckles
PMC5424383 V, Oliver A, Moulder K, Aisen PS, Ghetti B, Klunk
24. Abu Rumeileh S, Lattanzio F, Stanzani Maserati M, WE, McDade E, Martins RN, Masters CL, Mayeux
Rizzi R, Capellari S, Parchi P.  Diagnostic accuracy R, Ringman JM, Rossor MN, Schofield PR, Sperling
of a combined analysis of cerebrospinal fluid t-PrP, RA, Salloway S, Morris JC, Dominantly Inherited
t-tau, p-tau, and Aβ42  in the differential diagno- Alzheimer Network. Clinical and biomarker changes
sis of Creutzfeldt-Jakob disease from Alzheimer’s in dominantly inherited Alzheimer’s disease. N
disease with emphasis on atypical disease variants. Engl J  Med. 2012;367(9):795–804. https://doi.
J  Alzheimers Dis. 2017;55(4):1471–80. https:// org/10.1056/NEJMoa1202753. Epub 2012 July 11.
doi.org/10.3233/JAD-160740. PubMed PMID: Erratum in: N Engl J Med. 2012 Aug 23;367(8):780.
27886009; PubMed Central PMCID: PMC5181677 PubMed PMID: 22784036; PubMed Central PMCID:
25. Ewers M, Mattsson N, Minthon L, Molinuevo JL, PMC3474597
Antonell A, Popp J, Jessen F, Herukka SK, Soininen 32. Fagan AM, Xiong C, Jasielec MS, Bateman RJ,

H, Maetzler W, Leyhe T, Bürger K, Taniguchi M, Goate AM, Benzinger TL, Ghetti B, Martins RN,
Urakami K, Lista S, Dubois B, Blennow K, Hampel Masters CL, Mayeux R, Ringman JM, Rossor MN,
H.  CSF biomarkers for the differential diagnosis of Salloway S, Schofield PR, Sperling RA, Marcus D,
Alzheimer’s disease: a large-scale international multi- Cairns NJ, Buckles VD, Ladenson JH, Morris JC,
center study. Alzheimers Dement. 2015;11(11):1306– Holtzman DM, Dominantly Inherited Alzheimer
15. https://doi.org/10.1016/j.jalz.2014.12.006. Epub Network. Longitudinal change in CSF biomarkers in
2015 Mar 21. PubMed PMID: 25804998 autosomal-dominant Alzheimer’s disease. Sci Transl
26. Seino Y, Nakamura T, Kawarabayashi T, Hirohata
Med. 2014;6(226):226ra30. https://doi.org/10.1126/
M, Narita S, Wakasaya Y, Kaito K, Ueda T, scitranslmed.3007901. PubMed PMID: 24598588;
Harigaya Y, Shoji M.  CSF and Plasma biomarkers PubMed Central PMCID: PMC4038930
in neurodegenerative diseases. J  Alzheimers Dis. 33. Hansson O, Zetterberg H, Buchhave P, Londos E,
2019;68(1):395–404. Blennow K, Minthon L.  Association between CSF
27. Shaw LM, Vanderstichele H, Knapik-Czajka M, Clark biomarkers and incipient Alzheimer’s disease in
CM, Aisen PS, Petersen RC, Blennow K, Soares H, patients with mild cognitive impairment: a follow-up
Simon A, Lewczuk P, Dean R, Siemers E, Potter study. Lancet Neurol. 2006;5(3):228–34. Erratum in:
W, Lee VM, Trojanowski JQ, Alzheimer’s Disease Lancet Neurol. 2006 Apr;5(4):293. PubMed PMID:
Neuroimaging Initiative. Cerebrospinal fluid bio- 16488378
marker signature in Alzheimer’s disease neuroimaging 34. Visser PJ, Verhey F, Knol DL, Scheltens P, Wahlund
initiative subjects. Ann Neurol. 2009;65(4):403–13. LO, Freund-Levi Y, Tsolaki M, Minthon L, Wallin AK,
https://doi.org/10.1002/ana.21610. PubMed PMID: Hampel H, Bürger K, Pirttila T, Soininen H, Rikkert
19296504; PubMed Central PMCID: PMC2696350 MO, Verbeek MM, Spiru L, Blennow K. Prevalence
28. Shaw LM, Vanderstichele H, Knapik-Czajka M,
and prognostic value of CSF markers of Alzheimer’s
Figurski M, Coart E, Blennow K, Soares H, Simon AJ, disease pathology in patients with subjective cogni-
Lewczuk P, Dean RA, Siemers E, Potter W, Lee VM, tive impairment or mild cognitive impairment in the
Trojanowski JQ, Alzheimer’s Disease Neuroimaging DESCRIPA study: a prospective cohort study. Lancet
Initiative. Qualification of the analytical and clini- Neurol. 2009;8(7):619–27. https://doi.org/10.1016/
cal performance of CSF biomarker analyses in S1474-4422(09)70139-5. Epub 2009 June 10.
ADNI.  Acta Neuropathol. 2011;121(5):597–609. PubMed PMID: 19523877
https://doi.org/10.1007/s00401-011-0808-0. Epub
29  Cerebrospinal Fluid and Plasma Tau as a Biomarker for Brain Tauopathy 403

35. van Rossum IA, Vos SJ, Burns L, Knol DL, Scheltens PubMed PMID: 26220940; PubMed Central PMCID:
P, Soininen H, Wahlund LO, Hampel H, Tsolaki M, PMC4643624.
Minthon L, L'italien G, van der Flier WM, Teunissen 41. Olsson A, Vanderstichele H, Andreasen N, De Meyer
CE, Blennow K, Barkhof F, Rueckert D, Wolz R, G, Wallin A, Holmberg B, Rosengren L, Vanmechelen
Verhey F, Visser PJ.  Injury markers predict time to E, Blennow K.  Simultaneous measurement of beta-­
dementia in subjects with MCI and amyloid pathol- amyloid (1-42), total tau, and phosphorylated tau
ogy. Neurology. 2012;79(17):1809–16. https://doi. (Thr181) in cerebrospinal fluid by the xMAP technol-
org/10.1212/WNL.0b013e3182704056. Epub 2012 ogy. Clin Chem. 2005;51(2):336–45. Epub 2004 Nov
Sep 26. PubMed PMID: 23019259; PubMed Central 24. PubMed PMID: 15563479
PMCID: PMC3475623 42.
Lewczuk P, Kornhuber J, Vanderstichele H,
36. Stomrud E, Minthon L, Zetterberg H, Blennow
Vanmechelen E, Esselmann H, Bibl M, Wolf S,
K, Hansson O.  Longitudinal cerebrospinal fluid Otto M, Reulbach U, Kölsch H, Jessen F, Schröder
biomarker measurements in preclinical sporadic J, Schönknecht P, Hampel H, Peters O, Weimer E,
Alzheimer’s disease: a prospective 9-year study. Perneczky R, Jahn H, Luckhaus C, Lamla U, Supprian
Alzheimers Dement (Amst). 2015;1(4):403–11. T, Maler JM, Wiltfang J.  Multiplexed quantifica-
https://doi.org/10.1016/j.dadm.2015.09.002. eCollec- tion of dementia biomarkers in the CSF of patients
tion 2015 DecPubMed PMID: 27239521; PubMed with early dementias and MCI: a multicenter study.
Central PMCID: PMC4879483 Neurobiol Aging. 2008;29(6):812–8. Epub 2007 Jan
37. Sutphen CL, Jasielec MS, Shah AR, Macy EM,
19. PubMed PMID: 17239996
Xiong C, Vlassenko AG, Benzinger TL, Stoops EE, 43. Schindler SE, Gray JD, Gordon BA, Xiong C, Batrla-­
Vanderstichele HM, Brix B, Darby HD, Vandijck Utermann R, Quan M, Wahl S, Benzinger TLS,
ML, Ladenson JH, Morris JC, Holtzman DM, Fagan Holtzman DM, Morris JC, Fagan AM. Cerebrospinal
AM.  Longitudinal cerebrospinal fluid biomarker fluid biomarkers measured by Elecsys assays com-
changes in preclinical Alzheimer disease during pared to amyloid imaging. Alzheimers Dement.
middle age. JAMA Neurol. 2015;72(9):1029–42. 2018;14(11):1460–9. https://doi.org/10.1016/j.
https://doi.org/10.1001/jamaneurol.2015.1285. jalz.2018.01.013. Epub 2018 Mar 2. PubMed PMID:
PubMed PMID: 26147946; PubMed Central PMCID: 29501462; PubMed Central PMCID: PMC6119652
PMC4570860 44. Hansson O, Seibyl J, Stomrud E, Zetterberg H,

38. Xiong C, Jasielec MS, Weng H, Fagan AM, Benzinger Trojanowski JQ, Bittner T, Lifke V, Corradini V,
TL, Head D, Hassenstab J, Grant E, Sutphen CL, Eichenlaub U, Batrla R, Buck K, Zink K, Rabe C,
Buckles V, Moulder KL, Morris JC.  Longitudinal Blennow K, Shaw LM, Swedish BioFINDER study
relationships among biomarkers for Alzheimer group; Alzheimer’s Disease Neuroimaging Initiative.
disease in the adult children study. Neurology. CSF biomarkers of Alzheimer’s disease concord
2016;86(16):1499–506. https://doi.org/10.1212/ with amyloid-β PET and predict clinical progres-
WNL.0000000000002593. Epub 2016 Mar 23. sion: a study of fully automated immunoassays in
PubMed PMID: 27009258; PubMed Central PMCID: BioFINDER and ADNI cohorts. Alzheimers Dement.
PMC4836885 2018;14(11):1470–81. https://doi.org/10.1016/j.
39. Roe CM, Fagan AM, Grant EA, Hassenstab J,
jalz.2018.01.010. Epub 2018 Mar 1. PubMed PMID:
Moulder KL, Maue Dreyfus D, Sutphen CL, 29499171; PubMed Central PMCID: PMC6119541
Benzinger TL, Mintun MA, Holtzman DM, Morris 45. Mattsson N, Andreasson U, Persson S, Arai H, Batish
JC.  Amyloid imaging and CSF biomarkers in pre- SD, Bernardini S, Bocchio-Chiavetto L, Blankenstein
dicting cognitive impairment up to 7.5 years later. MA, Carrillo MC, Chalbot S, Coart E, Chiasserini D,
Neurology. 2013;80(19):1784–91. https://doi. Cutler N, Dahlfors G, Duller S, Fagan AM, Forlenza
org/10.1212/WNL.0b013e3182918ca6. Epub 2013 O, Frisoni GB, Galasko D, Galimberti D, Hampel H,
Apr 10. PubMed PMID: 23576620; PubMed Central Handberg A, Heneka MT, Herskovits AZ, Herukka
PMCID: PMC3719431 SK, Holtzman DM, Humpel C, Hyman BT, Iqbal K,
40. Toledo JB, Zetterberg H, van Harten AC, Glodzik Jucker M, Kaeser SA, Kaiser E, Kapaki E, Kidd D,
L, Martinez-Lage P, Bocchio-Chiavetto L, Rami L, Klivenyi P, Knudsen CS, Kummer MP, Lui J, Lladó A,
Hansson O, Sperling R, Engelborghs S, Osorio RS, Lewczuk P, Li QX, Martins R, Masters C, McAuliffe
Vanderstichele H, Vandijck M, Hampel H, Teipl S, J, Mercken M, Moghekar A, Molinuevo JL, Montine
Moghekar A, Albert M, Hu WT, Monge Argilés JA, TJ, Nowatzke W, O'Brien R, Otto M, Paraskevas
Gorostidi A, Teunissen CE, De Deyn PP, Hyman GP, Parnetti L, Petersen RC, Prvulovic D, de Reus
BT, Molinuevo JL, Frisoni GB, Linazasoro G, de HP, Rissman RA, Scarpini E, Stefani A, Soininen H,
Leon MJ, van der Flier WM, Scheltens P, Blennow Schröder J, Shaw LM, Skinningsrud A, Skrogstad
K, Shaw LM, Trojanowski JQ; Alzheimer’s Disease B, Spreer A, Talib L, Teunissen C, Trojanowski JQ,
Neuroimaging Initiative Alzheimer’s disease cere- Tumani H, Umek RM, Van Broeck B, Vanderstichele
brospinal fluid biomarker in cognitively normal H, Vecsei L, Verbeek MM, Windisch M, Zhang
subjects. Brain 2015;138(Pt 9):2701-2715. https:// J, Zetterberg H, Blennow K.  The Alzheimer’s
doi.org/10.1093/brain/awv199. Epub 2015 July 27. Association external quality control program for
cerebrospinal fluid biomarkers. Alzheimers Dement.
404 M. Shoji

2011;7(4):386–395.e6. https://doi.org/10.1016/j. 2017;9(9):1212–23. https://doi.org/10.15252/

jalz.2011.05.2243. Erratum in: Alzheimers Dement. emmm.201707809. PubMed PMID: 28743782;
2011 Sep;7(5):556. PubMed PMID: 21784349; PubMed Central PMCID: PMC5582410
PubMed Central PMCID: PMC3710290 52. Mattsson N, Smith R, Strandberg O, Palmqvist

46. Mattsson N, Andreasson U, Persson S, Carrillo MC, S, Schöll M, Insel PS, Hägerström D, Ohlsson
Collins S, Chalbot S, Cutler N, Dufour-Rainfray D, T, Zetterberg H, Blennow K, Jögi J, Hansson
Fagan AM, Heegaard NH, Robin Hsiung GY, Hyman O.  Comparing (18)F-AV-1451 with CSF t-tau and
B, Iqbal K, Kaeser SA, Lachno DR, Lleó A, Lewczuk p-tau for diagnosis of Alzheimer disease. Neurology.
P, Molinuevo JL, Parchi P, Regeniter A, Rissman RA, 2018;90(5):e388–95. https://doi.org/10.1212/
Rosenmann H, Sancesario G, Schröder J, Shaw LM, WNL.0000000000004887. Epub 2018 Jan 10.
Teunissen CE, Trojanowski JQ, Vanderstichele H, PubMed PMID: 29321235; PubMed Central PMCID:
Vandijck M, Verbeek MM, Zetterberg H, Blennow PMC5791788
K.  Alzheimer’s Association QC Program Work 53. Ovod V, Ramsey KN, Mawuenyega KG, Bollinger
Group. CSF biomarker variability in the Alzheimer's JG, Hicks T, Schneider T, Sullivan M, Paumier K,
Association quality control program. Alzheimers Holtzman DM, Morris JC, Benzinger T, Fagan AM,
Dement. 2013;9(3):251–61. https://doi.org/10.1016/j. Patterson BW, Bateman RJ.  Amyloid β concentra-
jalz.2013.01.010. Erratum in: Alzheimers Dement. tions and stable isotope labeling kinetics of human
2015 Feb;11(2):237. Käser, Stephan A [corrected plasma specific to central nervous system amy-
to Kaeser, Stephan A]; Rissman, Robert [corrected loidosis. Alzheimers Dement. 2017;13(8):841–9.
to Rissman, Robert A]. PubMed PMID: 23622690; https://doi.org/10.1016/j.jalz.2017.06.2266. Epub
PubMed Central PMCID: PMC3707386 2017 July 19. Erratum in: Alzheimers Dement.
47. Hansson O, Mikulskis A, Fagan AM, Teunissen C, 2017 Oct;13(10):1185. PubMed PMID: 28734653;
Zetterberg H, Vanderstichele H, Molinuevo JL, Shaw PubMed Central PMCID: PMC5567785
LM, Vandijck M, Verbeek MM, Savage M, Mattsson 54. Nakamura A, Kaneko N, Villemagne VL, Kato

N, Lewczuk P, Batrla R, Rutz S, Dean RA, Blennow T, Doecke J, Doré V, Fowler C, Li QX, Martins R,
K. The impact of preanalytical variables on measur- Rowe C, Tomita T, Matsuzaki K, Ishii K, Ishii K,
ing cerebrospinal fluid biomarkers for Alzheimer’s Arahata Y, Iwamoto S, Ito K, Tanaka K, Masters CL,
disease diagnosis: a review. Alzheimers Dement. Yanagisawa K. High performance plasma amyloid-β
2018;14(10):1313–33. https://doi.org/10.1016/j. biomarkers for Alzheimer’s disease. Nature.
jalz.2018.05.008. Epub 2018 Jun 23. Review. PubMed 2018;554(7691):249–54. https://doi.org/10.1038/
PMID: 29940161 nature25456. Epub 2018 Jan 31. PubMed PMID:
48. Sato C, Barthélemy NR, Mawuenyega KG, Patterson 29420472
BW, Gordon BA, Jockel-Balsarotti J, Sullivan M, 55. Mattsson N, Zetterberg H, Janelidze S, Insel PS,

Crisp MJ, Kasten T, Kirmess KM, Kanaan NM, Andreasson U, Stomrud E, Palmqvist S, Baker D, Tan
Yarasheski KE, Baker-Nigh A, Benzinger TLS, Hehir CA, Jeromin A, Hanlon D, Song L, Shaw LM,
Miller TM, Karch CM, Bateman RJ.  Tau kinetics Trojanowski JQ, Weiner MW, Hansson O, Blennow
in neurons and the human central nervous system. K, ADNI Investigators. Plasma tau in Alzheimer dis-
Neuron. 2018;98(4):861–4. https://doi.org/10.1016/j. ease. Neurology. 2016;87(17):1827–35. Epub 2016
neuron.2018.04.035. PubMed PMID: 29772204; Sep 30. PubMed PMID: 27694257; PubMed Central
PubMed Central PMCID: PMC6192252 PMCID: PMC5089525
49. He Z, Guo JL, McBride JD, Narasimhan S, Kim H, 56. Tatebe H, Kasai T, Ohmichi T, Kishi Y, Kakeya

Changolkar L, Zhang B, Gathagan RJ, Yue C, Dengler T, Waragai M, Kondo M, Allsop D, Tokuda
C, Stieber A, Nitla M, Coulter DA, Abel T, Brunden T.  Quantification of plasma phosphorylated tau
KR, Trojanowski JQ, Lee VM.  Amyloid-β plaques to use as a biomarker for brain Alzheimer pathol-
enhance Alzheimer's brain tau-seeded pathologies by ogy: pilot case-control studies including patients
facilitating neuritic plaque tau aggregation. Nat Med. with Alzheimer's disease and down syndrome.
2018;24(1):29–38. https://doi.org/10.1038/nm.4443. Mol Neurodegener. 2017;12(1):63. https://doi.
Epub 2017 Dec 4. PubMed PMID: 29200205; org/10.1186/s13024-017-0206-8. PubMed PMID:
PubMed Central PMCID: PMC5760353 28866979; PubMed Central PMCID: PMC5582385
50. Smith R, Wibom M, Pawlik D, Englund E, Hansson 57. Mielke MM, Hagen CE, Xu J, Chai X, Vemuri

O. Correlation of in vivo [18F]flortaucipir with post- P, Lowe VJ, Airey DC, Knopman DS, Roberts
mortem Alzheimer disease tau pathology. JAMA RO, Machulda MM, Jack CR Jr, Petersen RC,
Neurol. 2018;76:310. https://doi.org/10.1001/jama- Dage JL.  Plasma phospho-tau181 increases with
neurol.2018.3692. [Epub ahead of print] PubMed Alzheimer’s disease clinical severity and is associated
PMID: 30508025 with tau- and amyloid-positron emission tomography.
51. Mattsson N, Schöll M, Strandberg O, Smith R,
Alzheimers Dement. 2018;14(8):989–97. https://doi.
Palmqvist S, Insel PS, Hägerström D, Ohlsson T, org/10.1016/j.jalz.2018.02.013. Epub 2018 Apr 5.
Zetterberg H, Jögi J, Blennow K, Hansson O. (18) PubMed PMID: 29626426; PubMed Central PMCID:
F-AV-1451 and CSF T-tau and P-tau as biomark- PMC6097897
ers in Alzheimer’s disease. EMBO Mol Med.
29  Cerebrospinal Fluid and Plasma Tau as a Biomarker for Brain Tauopathy 405

58. Goedert M, Spillantini MG, Crowther RA. Cloning of org/10.1016/j.jalz.2016.09.009. Epub 2016 Oct 27.
a big tau microtubule-associated protein characteristic Review. PubMed PMID: 28341066
of the peripheral nervous system. Proc Natl Acad Sci 61. Simonsen AH, Herukka SK, Andreasen N, Baldeiras
U S A. 1992;89(5):1983–7. PubMed PMID: 1542696; I, Bjerke M, Blennow K, Engelborghs S, Frisoni
PubMed Central PMCID: PMC48578 GB, Gabryelewicz T, Galluzzi S, Handels R,
59. Couchie D, Mavilia C, Georgieff IS, Liem RK,
Kramberger MG, Kulczyńska A, Molinuevo JL,
Shelanski ML, Nunez J.  Primary structure of high Mroczko B, Nordberg A, Oliveira CR, Otto M,
molecular weight tau present in the peripheral nervous Rinne JO, Rot U, Saka E, Soininen H, Struyfs
system. Proc Natl Acad Sci U S A. 1992;89(10):4378– H, Suardi S, Visser PJ, Winblad B, Zetterberg H,
81. PubMed PMID: 1374898; PubMed Central Waldemar G.  Recommendations for CSF AD bio-
PMCID: PMC49085 markers in the diagnostic evaluation of dementia.
60. Herukka SK, Simonsen AH, Andreasen N, Baldeiras Alzheimers Dement. 2017;13(3):274–84. https://doi.
I, Bjerke M, Blennow K, Engelborghs S, Frisoni org/10.1016/j.jalz.2016.09.008.
GB, Gabryelewicz T, Galluzzi S, Handels R, 62. Jack CR Jr, Bennett DA, Blennow K, Carrillo MC,
Kramberger MG, Kulczyńska A, Molinuevo JL, Dunn B, Haeberlein SB, Holtzman DM, Jagust W,
Mroczko B, Nordberg A, Oliveira CR, Otto M, Jessen F, Karlawish J, Liu E, Molinuevo JL, Montine
Rinne JO, Rot U, Saka E, Soininen H, Struyfs T, Phelps C, Rankin KP, Rowe CC, Scheltens P,
H, Suardi S, Visser PJ, Winblad B, Zetterberg H, Siemers E, Snyder HM, Sperling R, Contributors.
Waldemar G.  Recommendations for cerebrospinal NIA-AA research framework: toward a biological def-
fluid Alzheimer’s disease biomarkers in the diag- inition of Alzheimer’s disease. Alzheimers Dement.
nostic evaluation of mild cognitive impairment. 2018;14(4):535–62. https://doi.org/10.1016/j.
Alzheimers Dement. 2017;13(3):285–95. https://doi. jalz.2018.02.018. Review. PubMed PMID: 29653606;
PubMed Central PMCID: PMC5958625
 Dementia Therapy Targeting Tau
30
Luc Buee

This final chapter addresses some of the various In the following sections, we investigate tau
researches on a therapeutic potential around the silencing, tau alternative splicing, post-­
tau protein and its gene, MAPT. Some have fallen translational tau modifications, tau metabolism,
into oblivion and others have appeared. This microtubular tau function, tau aggregation, tau
research reflects the different streams of ideas immunotherapy and brain homeostasis.
about tau protein and neurofibrillary
degeneration.
Currently, MAPT is considered to encode a Tau Silencing
pleiotropic protein but this was not always the
case. Tau is primarily a microtubule-associated To stop tau toxicity, one of the most obvious
protein regulated by its phosphorylation state but approaches is probably to modulate the expres-
has functions in chromatin structure, signal trans- sion of tau protein. Indeed, there is some evi-
duction, and nucleic acid protection [1]. dence that decreasing tau expression is beneficial
Alternative splicing and tau expression are also in reducing the electrophysiological and/or
new aspects of its biology. Finally, its new func- behavioral disturbances found in models of
tions, its other post-translational modifications Alzheimer’s disease [3–5]. This positive side
that phosphorylation, its half-life, its secretion should not obscure the fact that tau protein is
and its degradation recently aroused new inter- essential for synaptic plasticity, signal transduc-
ests, as you could discover it in the various chap- tion and nucleic acid protection [6–8].
ters of this work. To silence tau, the antisense oligonucleotide
This chapter is not exhaustive and it covers (ASO) and siRNA approaches are the most com-
only a few representative examples. Other recent mon ones. MicroRNAs may also be of interest
reviews on tau therapy may also be of interest [2]. since they have been involved in many tau (dys-)
Keep in mind that tau is a recent therapeutic tar- functions [9, 10]. ASO are single-strand oligo-
get. The hypothesis of the amyloid cascade made nucleotides (8–50 nucleotides) designed to com-
of tau a secondary actor. However, there have plement pre-mRNA/mRNA with unique
been some pre-clinical and clinical therapeutic specificity. Such approach is already in clinical
approaches. trials (ClinicalTrials identifier NCT03186989)
following promising results in rodents and mon-
L. Buee (*) keys [11].
University of Lille, INSERM, CHU-Lille, Alzheimer Similar approaches are also investigated to
& Tauopathies, LabEx DISTALZ, Lille, France modulate tau alternative splicing.
e-mail: luc.buee@inserm.fr

© Springer Nature Singapore Pte Ltd. 2019 407

A. Takashima et al. (eds.), Tau Biology, Advances in Experimental Medicine and Biology 1184,
https://doi.org/10.1007/978-981-32-9358-8_30
408 L. Buee

Tau Alternative Splicing tion [19–23]. In addition, some works have sug-
gested that hyperphosphorylated tau proteins
Alternative splicing of tau has quickly emerged as may serve as nucleation agents for tau aggrega-
a therapeutic target. Indeed, some mutations on the tion [24, 25]. Thus, inhibition of tau phosphory-
MAPT gene in frontotemporal lobar degeneration lation is clearly considered as a therapeutic
(FTLD) are responsible for misplicing. Thus, the approach. Other post translational modifications
intron mutations around exon 10 favor its default such as acetylation may also be of interest. For
insertion within tau transcripts leading to overex- instance, acetylation on Lys residues 274 and
pression of isoforms with four microtubule bind- 281, as found in AD, can impair AMPAR traf-
ing domains. There are also some pathologies such ficking and LTP [26].
as type 1 myotonic dystrophy where CTG expan- Acetylation is the second most important post-
sions in the 3’ UTR of the DMPK gene are respon- translational modification after phosphorylation.
sible for misplicing or deregulation of alternative Tau protein contains 44 lysine residues that can be
splicing. The resulting CUG expansions of the acetylated. Acetylated tau protein was observed
transcripts would lead to the sequestration of cer- linked to neurofibrillary degeneration in the brains
tain splicing factors like muscleblind (MBNL). of patients with AD. In addition, tau acetylation
This loss of MBNL function would result in a would reduce its own degradation when it is phos-
decrease or absence of insertion of exons 2, 3 and phorylated, suggesting a link between hyperphos-
10 in the tau transcripts and lead to neurofibrillary phorylation, accumulation and aggregation of Tau
degeneration (for reviews, [12, 13]). in tauopathies [27]. Acetylated tau protein no lon-
Correcting the alternative splicing of tau thus ger polymerizes tubulin into microtubules and
seems to be an excellent strategy for some tauopa- aggregates more rapidly in vitro [28]. Nevertheless,
thies like FTLDs and myotonic dystrophies. On the the real role of acetylation on tau pathology
one hand, it will be a question of modulating the remains poorly defined. Other studies have shown
insertion of the exon 10 of the other to correct the the opposite effect where Tau protein nonacety-
loss of function of MBNL.  The modulation of lated aggregates more rapidly in vitro than acety-
splicing of exon 10 in FTLDs was not only tested lated Tau protein [29, 30]. Salsalate and salicylate
by the use of ASO [5] [14] for its exclusion but also have been shown to reduce tau acetylation, tau
by spliceosome-mediated RNA trans-­splicing for amounts and decrease cognitive deficits in a
its inclusion [15]. In any case, it is possible to mod- mouse model of tauopathy [31].
ulate the alternative splicing of tau. Concerning In fact, modulating phosphorylation has been
myotonic dystrophies, the most advanced results a major approach: inhibitors of different kinases
concern MBNL [16] A gene therapy approach with such as glycogen synthase kinases (GSK) and
a truncated form of MBNL would correct the alter- cyclin-dependent kinases and MAPK (erk, JNK,
native splicing of tau in the presence of CUG p38 …) are used.
expansions (Patent WO2015158740A1). Lithium chloride and GSK3ß inhibitors have
been widely studied [32]. Lithium has been very
promising in preclinical models [33]. It has been
Acetylation and Phosphorylation used for many years in mood disorders [32].
Unfortunately, clinical trial with Lithium did not
In the 1980s, tau was considered the constituent reach its endpoints [34]. Similarly, GS3ß inhibi-
of neurofibrillary degeneration but the notion of tors were used in preclinical models of tauopathy
phosphorylation was not clear. In the mid-1990s, [35–37]. In PSP, the GSK3ß inhibitor, Tideglusib,
the notion of pathological phosphorylation is did not allow for any clinical improvement [38].
established because even if this phosphorylation More recently, the PERK activation is also
is physiological and disappears in healthy sub- considered as a therapeutic target in tauopathies
jects due to post-mortem dephosphorylation [17, [39] with CCT020312, a drug developed for can-
18], there is also a “pathological” phosphoryla- cer [40] but there is no conclusive data yet.
30  Dementia Therapy Targeting Tau 409

Finally, kinases involved in AMPK and mTOR as chaperone modulators also modulate tau metab-
pathways have also been considered as therapeu- olism. For instance, trehalose, an enhancer of
tic targets [41–43]. autophagy, has been shown to lead to tau degrada-
Regarding tau phosphorylation, there are 85 tion in preclinical model [58]. Another approach is
phosphorylation sites. Among them, five are Tyr proteasome modulation by PROTACs. Protac
residues. Thus, we can also cite tyr kinase inhibi- (Proteolysis Targeting Chimeric Molecule) acts as
tors even if these molecules do not target tau a bridge, bringing together E3 ubiquitin ligase with
phosphorylation per se but mostly signal trans- a protein target, resulting in its ubiquitination and
duction pathways such as src/fyn and abl path- degradation [55]. Tau could be degraded through a
ways. Regarding the Fyn inhibition, the most Keap1-dependent peptide PROTAC [59]. Finally,
advanced molecule is saracatinib, which is in sGC-1061 is a NO mimetic which activates the
clinical trial [44]. Regarding Abl tyr kinase, its cGMP signaling [60]. It restores cognition and
inhibition by nilotinib, a drug cancer, may turn on lowers Aß and tau amounts in mouse models [61].
autophagy [45] and allow for the degradation of More recently, tau clearance via the lysosome has
misfolded proteins including tau proteins [46]. also been explored. Activation of this pathway by
Another way to modulate phosphorylation is inhibiting the enzyme farnesyltransferase blocks
dephosphorylation by phosphatases [47]. The the attachment of a neuronal protein, Rhes, to the
activation of an enzyme is always tricky and the cell membrane. Farnesyltransferase inhibitors are
broad spectrum of phosphatase activities is very already in use in human patients for treating cancer.
challenging to find a therapeutic approach. The authors treated a mouse model of tauopathy
Sodium selenate has been shown to activate PP2A with one such drug, lonafarnib, and were able to
and restore synaptic plasticity [37, 48]. Finally, prevent the formation of tau inclusions, decrease
indirect approaches have also been suggested microgliosis and brain atrophy, and attenuate
such as Pin1 modulators [49]. Pin1 is a peptidyl behavioral abnormalities [62].
cis/trans isomerase and its isomerization catalysis In addition, it has been suggested that neurons
may allow for regulating the (de)phosphorylation may suffer from toxicity of persistent stress gran-
of specific pThr/pSer-Pro motifs [50–53, 47]. ules. These latter are cytoplasmic RNA-protein
Other isomerases such FKBPs may also modulate macro-complexes which appear as a normal cell
tau phosphorylation and degradation [41, 54, 55]. response to stress. They inhibit translation of
Finally, N-acetyl glucosamine linked to Ser/ transcripts not required for the stress response,
Thr residues (O-GlcNAc) is also a post-­ sequestering these non-essential transcripts.
translational modification that regulates tau phos- However, in neurodegenerative disorders such as
phorylation. O-GlcNac is a reciprocal to ALS and tauopathies, some RNA binding pro-
phosphorylation. Thus, therapeutic approaches to teins such as TIA1 may allow for the persistence
inhibit deglycosylation on Ser/Thr residues have of stress granules facilitating both formation of
been also used. For instance, Thiamet G is a tau oligomers, tau toxicity and tau spreading
O-GlcNacase inhibitor and has shown promising [63]. Thus, the modulation of such binding pro-
results in preclinical models [56, 57]. Some simi- teins may provide insights into possible thera-
lar compounds such as MK8719 and ASN120920 peutic approaches to reduce the spread of
are currently in clinical trials. neurodegeneration in tauopathy [64].
Finally, some drugs have pleiotropic effects
and may act on both amyloid and tau degradation
Tau Metabolism [65, 66]. Some of those such as AZP2006 are in
clinical trials (http://www.alzprotect.com/en/rd/
As indicated earlier, some kinase modulators may azp2006-clinical-trials/).
also play on tau metabolism by facilitating tau pro- All of these approaches target proteolytic
teolysis such as nilotinib, rapamycin and pathways and modulate tau metabolism in a non-­
CCT020312 [39, 43, 46]. Other compounds such specific way.
410 L. Buee

Microtubule Stabilizers tial work of Dale Schenk on amyloid vaccine,

tau vaccine has initiated by the group of Einar
Tau is a microtubule-associated protein and its Sigurdsson in 2007 [76] and followed by tau
binding to microtubules is regulated by post-­ immunotherapy [77]. The main concern about
translational modifications such as acetylation and tau immunotherapy is the antigen location. Tau
phosphorylation as described earlier. Some com- is a cytosolic protein and it is not clear how anti-
pounds are also able to stabilize microtubules. For bodies can “clear” intraneuronal aggregates even
instance, taxol and its derivates have been widely if some the endosome-lysosome [78] or the cyto-
in cancer to stop proliferation of cancer cells by solic TRIM21 Fc receptor [79] pathways have
interfering with tubulin [67]. However, these com- been suggested. With the prion-like propagation
pounds do not cross the blood-brain barrier. Other hypothesis, tau becomes an extracellular target
drugs such as Epothilone D [68] and TPI-287 and thus may be easy to eliminate [80]. However,
(abeotaxane) [69, 70] have been developed and there is no consensus on the seeding and spread-
shown positive outputs in experimental models. ing tau species which may be truncated, mono-
Nevertheless, Epothilone D, a small molecule mers, oligomers or high molecular weight
microtubule stabilizer, has been tested in phase 1 species [81–83]. More research is needed to
and discontinued. TPI-287 has finished phase 1 specify which tau species have to be targeted.
clinical trials in AD and SSP in 2018. As indicated Thus, we should be all conscient that current tri-
earlier, tau is more than a microtubule-associated als jeopardize tau pathology as target since tau
protein [1] and thus, it may be more complicated research is not advanced enough. However,
than a simple microtubule stabilization. numerous pre-­ clinical and clinical trials are
ongoing.
ACI35 and AADvac-1 are the most advanced
 nti-Aggregating Agents
A clinical trials in vaccination. ACI35 is a liposome-­
based antigen. It allows for the targeting of pS396
Tau aggregation is considered as deleterious even if and pS404 and has shown promising preclinical
it is likely more complicated. In fact, there is first effect [84]. AADvac-1 is microtubule-binding
post-translational modifications which may lead to peptide around the PGGG repeat sequence. Such
gain of toxic functions. Such tau species may be peptide belongs to the core of PHF and may act
aggregated as inert species. However, with time, as seeds for aggregation [85].
such aggregates grow and also become toxic for the Different passive immunotherapy approaches
cell. Such postulate is based in different observa- are also ongoing and positive preclinical studies
tions in other proteinopathies and experimental have allowed for phase 1 and phase 2 clinical tri-
models of tauopathies [71, 72]. However, different als in AD and PSP. For instance, phosphorylation
ß-breakers and anti-­aggregating agents were tested at Ser422 is considered as pathological [21] and
in preclinical models [73]. Among them, methy- its targeting has shown improvement in behavior
lene blue and its derivate Rember™ have been in and decrease in tau aggregation [86]. An ­antibody
clinical trials but shown some bias due to the blue anti-phosphoSer422, has also shown promising
color lacking in placebo [74]. TRx 0237 (LMTX™) data in preclinical studies even if some lysosomal
is the new generation of anti-aggregating agent defects were reported [78]. Unfortunately, clini-
[75]. It is currently in phase 2/3 trial for AD. cal trial with passive immunotherapy RG7345,
has been stopped in phase 1. Other antibodies are
currently in clinical trials with different isotypes
I mmunotherapy (see Table  30.1). Human isotype IgG4 may be
preferred in a number of studies since it has low
With the success of immunotherapy in cancer antibody-dependent cellular cytotoxicity and low
and multiple sclerosis, this approach has been binding to Fc receptors [87] which may allow for
widely tested in other pathologies. With the ini- a better targeting of extracellular tau. For
30  Dementia Therapy Targeting Tau 411

Table 30.1  Examples of current clinical trials for tauopathies

Mode of action Drug Target Status
Tau RNA BIIB080 Tau mRNA Phase 1
Tau PTMs Tideglusib/Lithium GSK3ß inhibition Phase 1 discont.
Nicotinamide Desacetylase inhibitor Phase 2
Saracatinib AZD0530 Tyr kinase inhibition Phase 2
Nilotinib
VEL015 Selenate PP2A activation Phase 2
ASN120290 O-GlcNacase inhibitor Phase 1
Tau metabolism AZP2006 N.A. Phase 1
Tau microtubule Epothilone D Stabilization microtubule Discontinued
TPI-287 Stabilization microtubule Phase 1
Anti-aggregating Rember ß-breakers Discontinued
agents LMTX ß-breakers Phase 3
Active AADvac1 [KDNIKHVPGGGS]4 Phase 2
immunotherapy ACI-35 VYKpSPVVSGDTpSPRHL Phase 1
Passive BIIB076 Full tau Phase 1
immunotherapy BIIB092 EVMEDHAGTY Phase 2
C2N-8E12 DQGGYT Phase 2
JNJ-63733657 Tau mid-region Phase 1
LY33035690 N-terminal conformational Phase 2
NPT088 (not real Fused Ig-general amyloid Phase 1
immunoglobulin) interaction motif
RG7345 pS422 Phase 1 discontinued

RO7105705 pS409 Phase 2

UCB0107 SPSSAKSRLQTA Phase 1
Others CERE-110 NGF Phase 2
GSK3ß glycogen synthase kinase 3ß, PP2A protein phosphatase 2A, NGF Nerve Growth Factor

instance, both BIIB092 and UCB0107 are IgG4, sis: key actors are linked to tau pathology such as
which have been tested in phase 1 clinical trials adenosine and its A2A receptor which can be
and shown convincing preclinical studies [88– modulated by a non-selective antagonist caf-
91]. They are both in phase 2 clinical trials. Some feine) [94, 95] and AMPK which is hyperacti-
antibodies are also fused immunoglobulin with vated in neurofibrillary tangles [42, 96].
general amyloid interaction motif and target Aß, NGF and growth factors may also represent a
tau, synuclein amyloids [92]. nice therapeutic strategy linking tau pathology
Other immunotherapies are going and sum- and loss of cholinergic neurons [97, 98].
marized in Table 30.1. Modulation of cholesterol metabolism by acting
on its esterification as well as insulin may also be
of interest [99–101].
Brain/Body as a Whole Finally, recent findings suggest that peripheral
signaling such as microbiote and different infec-
Finally, many modulators of brain homeostasis tions may also participate to tau pathology and
have also been tested in experimental models of thus represent new therapeutic targets [102–104].
tauopathy and in clinical trials. They are not In conclusion, there are many (maybe too
directly linked to tau pathology but are clearly of many) therapeutic targets for tau pathology.
interest. Physical exercise has shown a beneficial Unfortunately, tau pathology is ill-defined and
effect in tau transgenic mice [93]. Energy metab- thus, it is difficult to identify those which are the
olism is also very important for brain homeosta- most promising ones.
412 L. Buee

Acknowledgements  I would like to thank all members of tau deposition and seeding in mice with tauopa-
the Alzheimer & Tauopathies team for their support and thy. Sci Transl Med. American Association for the
discussion on the present work. Advancement of Science. 2017;9(374):eaag0481.
12. Caillet-Boudin M-L, Buee L, Sergeant N, Lefebvre
B.  Regulation of human MAPT gene expression.
Mol Neurodegener. BioMed Central. 2015;10(1):28.
References 13. Caillet-Boudin M-L, Fernandez-Gomez F-J, Tran H,
Dhaenens C-M, Buee L, Sergeant N. Brain pathol-
1. Sotiropoulos I, Galas M-C, Silva JM, Skoulakis ogy in myotonic dystrophy: when tauopathy meets
E, Wegmann S, Maina MB, et  al. Atypical, non-­ spliceopathy and RNAopathy. Front Mol Neurosci.
standard functions of the microtubule associated Frontiers. 2014;6:57.
tau protein. Acta Neuropathol Commun. BioMed 14. Kalbfuss B, Mabon SA, Misteli T.  Correction of
Central. 2017;5(1):91. alternative splicing of tau in frontotemporal demen-
2. Jadhav S, Avila J, Schöll M, Kovacs GG, Kovari tia and parkinsonism linked to chromosome 17.
E, Skrabana R, et al. A walk through tau therapeu- J  Biol Chem. American Society for Biochemistry
tic strategies. Acta Neuropathol Commun. BioMed and Molecular Biology. 2001;276(46):42986–93.
Central. 2019;7(1):22. 15. Avale ME, Rodriguez-Martin T, Gallo J-M.  Trans-­
3. Busche MA, Wegmann S, Dujardin S, Commins splicing correction of tau isoform imbalance in a
C, Schiantarelli J, Klickstein N, et  al. Tau impairs mouse model of tau mis-splicing. Human Mol Gen.
neural circuits, dominating amyloid-β effects, in 2013;22(13):2603–11.
Alzheimer models in  vivo. Nat Neurosci.. Nature 16. Goodwin M, Mohan A, Batra R, Lee K-Y,
Publishing Group. 2019;22(1):57–64. Charizanis K, Fernandez-Gomez F-J, et  al. MBNL
4. Roberson ED, Scearce-Levie K, Palop JJ, Yan sequestration by toxic RNAs and RNA misprocess-
F, Cheng IH, Wu T, et  al. Reducing endogenous ing in the Myotonic dystrophy brain. Cell Rep.
tau ameliorates amyloid beta-induced deficits in 2015;12(7):1159–68.
an Alzheimer’s disease mouse model. Science. 17. Matsuo ES, Shin RW, Billingsley ML, Van de Voorde
American Association for the Advancement of A, O’Connor M, Trojanowski JQ, et  al. Biopsy-­
Science. 2007;316(5825):750–4. derived adult human brain tau is phosphorylated at
5. DeVos SL, Goncharoff DK, Chen G, Kebodeaux many of the same sites as Alzheimer’s disease paired
CS, Yamada K, Stewart FR, et al. Antisense reduc- helical filament tau. Neuron. 1994;13(4):989–1002.
tion of tau in adult mice protects against sei- 18. Buee-Scherrer V, Condamines O, Mourton-Gilles
zures. J  Neurosci.. Society for Neuroscience. C, Jakes R, Goedert M, Pau B, et  al. AD2, a
2013;33(31):12887–97. phosphorylation-­ dependent monoclonal antibody
6. Marciniak E, Leboucher A, Caron E, Ahmed T, directed against tau proteins found in Alzheimer’s dis-
Tailleux A, Dumont J, et al. Tau deletion promotes ease. Brain Res Mol Brain Res. 1996;39(1–2):79–88.
brain insulin resistance. J  Exp Med.. Rockefeller 19. Sergeant N, Bussiere T, Vermersch P, Lejeune JP,
University Press. 2017;214(8):2257–69. Delacourte A.  Isoelectric point differentiates PHF-­
7. Violet M, Chauderlier A, Delattre L, Tardivel M, tau from biopsy-derived human brain tau proteins.
Chouala MS, Sultan A, et al. Prefibrillar tau oligo- Neuroreport. 1995;6(16):2217–20.
mers alter the nucleic acid protective function of 20. Hoffmann R, Lee VM, Leight S, Varga I, Otvos
tau in hippocampal neurons in vivo. Neurobiol Dis. L.  Unique Alzheimer’s disease paired helical fila-
2015;82:540–51. ment specific epitopes involve double phosphory-
8. Ahmed T, Van der Jeugd A, Blum D, Galas M-C, lation at specific sites. Biochemistry.. American
D’Hooge R, Buee L, et  al. Cognition and hippo- Chemical Society. 1997;36(26):8114–24.
campal synaptic plasticity in mice with a homo- 21. Bussiere T, Hof PR, Mailliot C, Brown CD, Caillet-­
zygous tau deletion. Neurobiol Aging. Elsevier. Boudin ML, Perl DP, et al. Phosphorylated serine422
2014;35(11):2474–8. on tau proteins is a pathological epitope found in
9. Hebert SS, Papadopoulou AS, Smith P, Galas MC, several diseases with neurofibrillary degeneration.
Planel E, Silahtaroglu AN, et al. Genetic ablation of Acta Neuropathol. 1999;97(3):221–30.
dicer in adult forebrain neurons results in abnormal 22. Delobel P, Flament S, Hamdane M, Mailliot C,
tau hyperphosphorylation and neurodegeneration. Sambo AV, Begard S, et al. Abnormal tau phosphor-
Human Mol Gen. 2010;19(20):3959–69. ylation of the Alzheimer-type also occurs during
10. Smith PY, Hernandez-Rapp J, Jolivette F, Lecours mitosis. J Neurochem. 2002;83(2):412–20.
C, Bisht K, Goupil C, et  al. miR-132/212 defi- 23. Malia TJ, Teplyakov A, Ernst R, Wu S-J, Lacy ER,
ciency impairs tau metabolism and promotes path- Liu X, et al. Epitope mapping and structural basis for
ological aggregation in  vivo. Human Mol Gen. the recognition of phosphorylated tau by the anti-tau
2015;24(23):6721–35. antibody AT8. Proteins.. John Wiley & Sons, Ltd.
11. DeVos SL, Miller RL, Schoch KM, Holmes BB, 2016;84(4):427–34.
Kebodeaux CS, Wegener AJ, et  al. Tau reduction 24. Iqbal K, Liu F, Gong C-X, Alonso ADC, Grundke-­
prevents neuronal loss and reverses pathological Iqbal I.  Mechanisms of tau-induced neurode-
30  Dementia Therapy Targeting Tau 413

generation. Acta Neuropathol.. Springer-Verlag. 37. Ahmed T, Blum D, Burnouf S, Demeyer D, Buée-­
2009;118(1):53–69. Scherrer V, D’Hooge R, et  al. Rescue of impaired
25. Despres C, Byrne C, Qi H, Cantrelle F-X, Huvent I, late-phase long-term depression in a tau transgenic
Chambraud B, et al. Identification of the tau phos- mouse model. Neurobiol Aging. 2015;36(2):730–9.
phorylation pattern that drives its aggregation. Proc 38. Tolosa E, Litvan I, Höglinger GU, Burn D, Lees A,
Natl Acad Sci USA. 2017;114(34):9080–5. Andrés MV, et al. A phase 2 trial of the GSK-3 inhib-
26. Tracy TE, Sohn PD, Minami SS, Wang C, Min itor tideglusib in progressive supranuclear palsy.
S-W, Li Y, et  al. Acetylated tau obstructs KIBRA-­ Mov Disord. 2014;29(4):470–8.
mediated signaling in synaptic plasticity and pro- 39. Bruch J, Xu H, Rösler TW, De Andrade A, Kuhn
motes Tauopathy-related memory loss. Neuron. P-H, Lichtenthaler SF, et al. PERK activation miti-
2016;90(2):245–60. gates tau pathology in vitro and in vivo. EMBO Mol
27. Min S-W, Cho S-H, Zhou Y, Schroeder S, Med. 2017;9(3):371–84.
Haroutunian V, Seeley WW, et al. Acetylation of tau 40. Stockwell SR, Platt G, Barrie SE, Zoumpoulidou G,
inhibits its degradation and contributes to Tauopathy. Poele Te RH, Aherne GW, et al. Mechanism-based
Neuron.. Elsevier Inc. 2010;67(6):953–66. screen for G1/S checkpoint activators identifies a
28. Cohen TJ, Guo JL, Hurtado DE, Kwong LK, Mills selective activator of EIF2AK3/PERK signalling.
IP, Trojanowski JQ, et  al. The acetylation of tau PLoS ONE.  Bernhard EJ, editor. Public Library of
inhibits its function and promotes pathological tau Science. 2012;7(1):e28568.
aggregation. Nat Commun.. Nature Publishing 41. Mueed Z, Tandon P, Maurya SK, Deval R, Kamal
Group. 2011;2(1):252. MA, Poddar NK.  Tau and mTOR: the hotspots for
29. Cook C, Carlomagno Y, Gendron TF, Dunmore J, multifarious diseases in Alzheimer’s development.
Scheffel K, Stetler C, et al. Acetylation of the KXGS Front Neurosci. Frontiers. 2018;12:1017.
motifs in tau is a critical determinant in modulation 42. Domise M, Didier S, Marinangeli C, Zhao H,
of tau aggregation and clearance. Human Mol Gen. Chandakkar P, Buee L, et  al. AMP-activated pro-
2014;23(1):104–16. tein kinase modulates tau phosphorylation and
30. Kamah A, Huvent I, Cantrelle F-X, Qi H, Lippens G, tau pathology in  vivo. Sci Rep. Nature Publishing
Landrieu I, et al. Nuclear magnetic resonance analy- Group. 2016;6(1):26758.
sis of the acetylation pattern of the neuronal tau pro- 43. Frédérick C, Ando K, Leroy K, Heraud C, Suain
tein. Biochemistry. 2014;53(18):3020–32. V, Buee L, et al. Rapamycin ester analog CCI-779/
31. Min S-W, Chen X, Tracy TE, Li Y, Zhou Y, Wang Temsirolimus alleviates tau pathology and improves
C, et al. Critical role of acetylation in tau-mediated motor deficit in mutant tau transgenic mice.
neurodegeneration and cognitive deficits. Nat Med.. J Alzheimers Dis. IOS Press. 2015;44(4):1145–56.
Nature Publishing Group. 2015;21(10):1154–62. 44. Nygaard HB, Wagner AF, Bowen GS, Good SP,
32. Forlenza OV, De-Paula VJR, Diniz MacAvoy MG, Strittmatter KA, et al. A phase Ib mul-
BSO.  Neuroprotective effects of lithium: implica- tiple ascending dose study of the safety, tolerability,
tions for the treatment of Alzheimer’s disease and and central nervous system availability of AZD0530
related neurodegenerative disorders. ACS Chem (saracatinib) in Alzheimer’s disease. Alzheimers Res
Neurosci. 2014;5(6):443–50. Ther. BioMed Central. 2015;7(1):35.
33. Leroy K, Ando K, Heraud C, Yilmaz Z, Authelet 45. Hebron ML, Lonskaya I, Moussa CE-H.  Tyrosine
M, Boeynaems JM, et al. Lithium treatment arrests kinase inhibition facilitates autophagic SNCA/α-­-
the development of neurofibrillary tangles in mutant synuclein clearance. Autophagy.. Taylor & Francis.
tau transgenic mice with advanced neurofibrillary 2013;9(8):1249–50.
pathology. J Alzheimers Dis. 2010;19(2):705–19. 46. Hebron ML, Javidnia M, Moussa CE-H.  Tau
34. Forlenza OV, Diniz BS, Radanovic M, Santos FS, clearance improves astrocytic function and
Talib LL, Gattaz WF.  Disease-modifying proper- brain glutamate-­ glutamine cycle. J  Neurol Sci.
ties of long-term lithium treatment for amnestic 2018;391:90–9.
mild cognitive impairment: randomised controlled 47. Taleski G, Sontag E.  Protein phosphatase 2A and
trial. Br J Psychiatry.. Cambridge University Press. tau: an orchestrated ‘pas de Deux’. FEBS Lett.. John
2011;198(5):351–6. Wiley & Sons, Ltd. 2018;592(7):1079–95.
35. Serenó L, Coma M, Rodríguez M, Sánchez-Ferrer 48. van Eersel J, Ke YD, Liu X, Delerue F, Kril JJ,
P, Sánchez MB, Gich I, et  al. A novel GSK-3beta Götz J, et al. Sodium selenate mitigates tau pathol-
inhibitor reduces Alzheimer’s pathology and res- ogy, neurodegeneration, and functional deficits in
cues neuronal loss in  vivo. Neurobiol Diseas. Alzheimer’s disease models. Proc Natl Acad Sci
2009;35(3):359–67. USA. 2010;107(31):13888–93.
36. Georgievska B, Sandin J, Doherty J, Mörtberg A, 49. Nakamura K, Zhou XZ, Lu KP.  Distinct functions
Neelissen J, Andersson A, et al. AZD1080, a novel of cis and trans phosphorylated tau in Alzheimer’s
GSK3 inhibitor, rescues synaptic plasticity deficits disease and their therapeutic implications. Curr Mol
in rodent brain and exhibits peripheral target engage- Med. 2013;13(7):1098–109.
ment in humans. J Neurochem. John Wiley & Sons, 50. Galas MC, Dourlen P, Begard S, Ando K, Blum
Ltd (10.1111). 2013;125(3):446–56. D, Hamdane M, et  al. The peptidylprolyl cis/
414 L. Buee

trans-­
isomerase Pin1 modulates stress-induced generation and response to toxic tau oligomers. Acta
dephosphorylation of tau in neurons. Implication in Neuropathol. 2nd ed. Springer Berlin Heidelberg.
a pathological mechanism related to Alzheimer dis- 2019;137(2):259–77.
ease. J Biol Chem. 2006;281(28):19296–304. 64. Apicco DJ, Ash PEA, Maziuk B, LeBlang C, Medalla
51. Hamdane M, Dourlen P, Bretteville A, Sambo AV, M, Abdullatif Al A, et al. Reducing the RNA binding
Ferreira S, Ando K, et al. Pin1 allows for differential protein TIA1 protects against tau-mediated neurode-
tau dephosphorylation in neuronal cells. Mol Cell generation in vivo. Nat Neurosci.. Nature Publishing
Neurosci. 2006;32(1–2):155–60. Group. 2018;21(1):72–80.
52. Smet C, Sambo AV, Wieruszeski JM, Leroy A, 65. Sergeant N, Vingtdeux V, Eddarkaoui S, Gay M,
Landrieu I, Buee L, et  al. The peptidyl prolyl cis/ Evrard C, Le Fur N, et  al. New piperazine multi-­
trans-isomerase Pin1 recognizes the phospho-­Thr212-­ effect drugs prevent neurofibrillary degeneration and
Pro213 site on tau. Biochemistry. 2004;43(7):2032–40. amyloid deposition, and preserve memory in ani-
53. Smet C, Wieruszeski JM, Buee L, Landrieu I, mal models of Alzheimer’s disease. Neurobiol Dis.
Lippens G.  Regulation of Pin1 peptidyl-prolyl cis/ 2019;129:217–233.
trans isomerase activity by its WW binding module 66. Melnyk P, Vingtdeux V, Burlet S, Eddarkaoui S,
on a multi-phosphorylated peptide of tau protein. Grosjean M-E, Larchanché P-E, et al. Chloroquine
FEBS Lett. 2005;579(19):4159–64. and chloroquinoline derivatives as models for the
54. Giustiniani J, Guillemeau K, Dounane O, Sardin E, design of modulators of amyloid peptide precur-
Huvent I, Schmitt A, et al. The FK506-binding pro- sor metabolism. ACS Chem Neurosci.. American
tein FKBP52 in  vitro induces aggregation of trun- Chemical Society. 2015;6(4):559–69.
cated tau forms with prion-like behavior. FASEB 67. Kingston DG.  Taxol: the chemistry and structure-­
J. Federation of American Societies for Experimental activity relationships of a novel anticancer agent.
Biology. 2015;29(8):3171–81. Trends Biotechnol. 1994;12(6):222–7.
55. Gorantla NV, Chinnathambi S.  Tau protein squired 68. Brunden KR, Zhang B, Carroll J, Yao Y, Potuzak JS,
by molecular chaperones during Alzheimer’s disease. Hogan A-ML, et al. Epothilone D improves micro-
J Mol Neurosci.. Springer US. 2018;66(3):356–68. tubule density, axonal integrity, and cognition in a
56. Yu Y, Zhang L, Li X, Run X, Liang Z, Li Y, et  al. transgenic mouse model of tauopathy. J  Neurosci..
Differential effects of an O-GlcNAcase inhibitor Society for Neuroscience. 2010;30(41):13861–6.
on tau phosphorylation. PLoS ONE. Götz J, editor. 69. Fitzgerald DP, Emerson DL, Qian Y, Anwar T,
Public Library of Science. 2012;7(4):e35277. Liewehr DJ, Steinberg SM, et  al. TPI-287, a new
57. Wang X, Smith K, Pearson M, Hughes A, Cosden ML, taxane family member, reduces the brain metastatic
Marcus J, et al. Early intervention of tau pathology colonization of breast cancer cells. Mol Cancer
prevents behavioral changes in the rTg4510 mouse Therap. American Association for Cancer Research.
model of tauopathy. PLoS ONE. Ginsberg SD, edi- 2012;11(9):1959–67.
tor. Public Library of Science. 2018;13(4):e0195486. 70. Zumbar CT, Usubalieva A, King PD, Li X, Mifsud
58. Schaeffer V, Lavenir I, Ozcelik S, Tolnay M, Winkler CS, Dalton HM, et  al. The CNS penetrating tax-
DT, Goedert M.  Stimulation of autophagy reduces ane TPI 287 and the AURKA inhibitor alisertib
neurodegeneration in a mouse model of human induce synergistic apoptosis in glioblastoma cells.
tauopathy. Brain. 2012;135(Pt 7):2169–77. J Neurooncol.. Springer US. 2018;137(3):481–92.
59. Lu M, Liu T, Jiao Q, Ji J, Tao M, Liu Y, et  al. 71. Sun C-S, Lee C-C, Li Y-N, Yao-Chen Yang S, Lin
Discovery of a Keap1-dependent peptide PROTAC to C-H, Chang Y-C, et  al. Conformational switch of
knockdown tau by ubiquitination-proteasome degra- polyglutamine-expanded huntingtin into benign
dation pathway. Eur J Med Chem. 2018;146:251–9. aggregates leads to neuroprotective effect. Sci Rep..
60. Thatcher GRJ, Bennett BM, Reynolds JN. NO chi- Nature Publishing Group. 2015;5(1):14992.
meras as therapeutic agents in Alzheimer’s disease. 72. d’Orange M, Auregan G, Cheramy D, Gaudin-Guérif
Curr Alzheimer Res. 2006;3(3):237–45. M, Lieger S, Guillermier M, et al. Potentiating tan-
61. Thatcher GRJ, Luo J, Vandevrede L, Qin Z, Lee S, gle formation reduces acute toxicity of soluble tau
Abdelhamid R, et al. NO/cGMP/pCREB reactivation species in the rat. Brain. 2018;141(2):535–49.
reverses cognition deficits and attenuates amyloid ß 73. Bulic B, Pickhardt M, Mandelkow E.  Progress
neuropathology in transgenic models of Alzheimer’s and developments in tau aggregation inhibi-
disease. BMC Pharmacol Toxicol. 2013;14(Suppl tors for Alzheimer disease. J  Med Chem.
1):P72. 2013;56(11):4135–55.
62. Hernandez I, Luna G, Rauch JN, Reis SA, Giroux 74. Baddeley TC, McCaffrey J, Storey JMD, Cheung
M, Karch CM, et  al. A farnesyltransferase inhibi- JKS, Melis V, Horsley D, et  al. Complex dispo-
tor activates lysosomes and reduces tau pathology sition of methylthioninium redox forms deter-
in mice with tauopathy. Sci Transl Med. American mines efficacy in tau aggregation inhibitor therapy
Association for the Advancement of Science. for Alzheimer’s disease. J  Pharmacol Exp Ther.
2019;11(485):eaat3005. 2015;352(1):110–8.
63. Jiang L, Ash PEA, Maziuk BF, Ballance HI, 75. Wischik CM, Staff RT, Wischik DJ, Bentham P,
Boudeau S, Abdullatif AA, et al. TIA1 regulates the Murray AD, Storey JMD, et  al. Tau aggregation
30  Dementia Therapy Targeting Tau 415

inhibitor therapy: an exploratory phase 2 study in 87. Bournazos S, Ravetch JV.  Fcγ receptor function
mild or moderate Alzheimer’s disease. J Alzheimers and the design of vaccination strategies. Immunity.
Dis.. IOS Press. 2015;44(2):705–20. 2017;47(2):224–33.
76. Asuni AA, Boutajangout A, Quartermain D, 88. Albert M, Mairet-Coello G, Danis C, Lieger S,
Sigurdsson EM. Immunotherapy targeting patholog- Caillierez R, Carrier S, et al. Prevention of tau seed-
ical tau conformers in a tangle mouse model reduces ing and propagation by immunotherapy with a central
brain pathology with associated functional improve- tau epitope antibody. Brain. 2019;142(6):1736–1750
ments. J  Neurosci.. Society for Neuroscience. 89. Bright J, Hussain S, Dang V, Wright S, Cooper B, Byun
2007;27(34):9115–29. T, et  al. Human secreted tau increases amyloid-beta
77. Congdon EE, Sigurdsson EM.  Tau-targeting production. Neurobiol Aging. 2015;36(2):693–709.
therapies for Alzheimer disease. Nat Rev Neurol. 90. Qureshi IA, Tirucherai G, Ahlijanian MK, Kolaitis
2018;14(7):399–415. G, Bechtold C, Grundman M. A randomized, single
78. Collin L, Bohrmann B, Göpfert U, Oroszlan-Szovik ascending dose study of intravenous BIIB092  in
K, Ozmen L, Grüninger F. Neuronal uptake of tau/ healthy participants. Alzheimers Dement (NY).
pS422 antibody and reduced progression of tau 2018;4:746–55.
pathology in a mouse model of Alzheimer’s disease. 91. Courade J-P, Angers R, Mairet-Coello G, Pacico
Brain. 2014;137(Pt 10):2834–46. N, Tyson K, Lightwood D, et  al. Epitope deter-
79. McEwan WA, Falcon B, Vaysburd M, Clift D, Oblak mines efficacy of therapeutic anti-tau antibodies
AL, Ghetti B, et  al. Cytosolic fc receptor TRIM21 in a functional assay with human Alzheimer tau.
inhibits seeded tau aggregation. Proc Natl Acad Sci Acta Neuropathol. Springer Berlin Heidelberg.
USA. 2017;114(3):574–9. 2018;136(5):729–45.
80. Mudher A, Colin M, Dujardin S, Medina M, 92. Levenson JM, Schroeter S, Carroll JC, Cullen V,
Dewachter I, Alavi Naini SM, et al. What is the evi- Asp E, Proschitsky M, et al. NPT088 reduces both
dence that tau pathology spreads through prion-like amyloid-β and tau pathologies in transgenic mice.
propagation? Acta Neuropathol Commun. BioMed Alzheimers Dement (NY). 2016;2(3):141–55.
Central. 2017;5(1):99. 93. Belarbi K, Burnouf S, Fernandez-Gomez FJ,
81. Sharma AM, Thomas TL, Woodard DR, Kashmer Laurent C, Lestavel S, Figeac M, et  al. Beneficial
OM, Diamond MI.  Tau monomer encodes strains. effects of exercise in a transgenic mouse model of
Elife. eLife Sciences Publications Limited. Alzheimer’s disease-like tau pathology. Neurobiol
2018;7:909. Dis. 2011;43(2):486–94.
82. Usenovic M, Niroomand S, Drolet RE, Yao L, Gaspar 94. Laurent C, Burnouf S, Ferry B, Batalha VL,
RC, Hatcher NG, et  al. Internalized tau oligomers Coelho JE, Baqi Y, et  al. A2A adenosine recep-
cause Neurodegeneration by inducing accumulation tor deletion is protective in a mouse model of
of pathogenic tau in human neurons derived from Tauopathy. Mol Psychiat.. Nature Publishing Group.
induced pluripotent stem cells. J Neurosci.. Society 2016;21(1):97–107.
for Neuroscience. 2015;35(42):14234–50. 95. Laurent C, Eddarkaoui S, Derisbourg M, Leboucher
83. Audouard E, Houben S, Masaracchia C, Yilmaz Z, A, Demeyer D, Carrier S, et  al. Beneficial effects
Suain V, Authelet M, et  al. High-molecular-weight of caffeine in a transgenic model of Alzheimer’s
paired helical filaments from Alzheimer brain disease-like tau pathology. Neurobiol Aging.
induces seeding of wild-type mouse tau into an 2014;35(9):2079–90.
Argyrophilic 4R tau pathology in vivo. Am J Pathol. 96. Galasso A, Cameron CS, Frenguelli BG, Moffat
2016;186(10):2709–22. KG.  An AMPK-dependent regulatory pathway in
84. Theunis C, Crespo-Biel N, Gafner V, Pihlgren M, tau-mediated toxicity. Biol Open. The Company of
López-Deber MP, Reis P, et al. Efficacy and safety Biologists Ltd. 2017;6(10):1434–44.
of a liposome-based vaccine against protein tau, 97. Mitra S, Behbahani H, Eriksdotter M.  Innovative
assessed in tau.P301L mice that model tauopathy. therapy for Alzheimer’s disease-with focus on bio-
PLoS ONE.  Iijima KM, editor. Public Library of delivery of NGF. Front Neurosci. 2019;13:38.
Science. 2013;8(8):e72301. 98. Belarbi K, Burnouf S, Fernandez-Gomez FJ,
85. Kontsekova E, Zilka N, Kovacech B, Novak P, Desmercieres J, Troquier L, Brouillette J, et al. Loss
Novak M. First-in-man tau vaccine targeting struc- of medial septum cholinergic neurons in THY-Tau22
tural determinants essential for pathological tau-tau mouse model: what links with tau pathology? Curr
interaction reduces tau oligomerisation and neuro- Alzheimer Res. 2011;8(6):633–8.
fibrillary degeneration in an Alzheimer’s disease 99. Burlot M-A, Braudeau J, Michaelsen-Preusse K,
model. Alzheimers Res Ther. BioMed Central. Potier B, Ayciriex S, Varin J, et  al. Cholesterol
2014;6(4):44. 24-hydroxylase defect is implicated in memory
86. Troquier L, Caillierez M, Burnouf S, Fernandez-­ impairments associated with Alzheimer-like tau
Gomez FJ, Grosjean MJ, Zommer N, et al. Targeting pathology. Human Mol Gene. 2015;24(21):5965–76.
phospho-Ser422 by active tau immunotherapy in the 100. Braudeau J, Burlot M-A, Ayciriex S, Varin J, Gautier
THY-Tau22 mouse model: a suitable therapeutic B, Djelti F, et  al. Cholesterol and tau pathology:
approach. Curr Alzheimer Res. 2012;9(4):397–405. consequences of AAV-mediated CHOLESTEROL-­
416 L. Buee

24-­hydroxylase overexpression in the thy-TAU22 J  Alzheimers Dis. Albensi B, editor. IOS Press.
mouse. Alzheimer’s & Dementia. Elsevier; 2018;66(4):1657–82.
2014;10(4):P237. 103. Dominy SS, Lynch C, Ermini F, Benedyk M,

101. Craft S, Claxton A, Baker LD, Hanson AJ, Cholerton Marczyk A, Konradi A, et  al. Porphyromonas gin-
B, Trittschuh EH, et al. Effects of regular and long-­ givalis in Alzheimer’s disease brains: evidence for
acting insulin on cognition and Alzheimer’s disease disease causation and treatment with small-molecule
biomarkers: a pilot clinical trial. J Alzheimers Dis. la inhibitors. Sci Adv. American Association for the
Monte de S, editor. IOS Press. 2017;57(4):1325–34. Advancement of Science. 2019;5(1):eaau3333.
102. Syeda T, Sanchez-Tapia M, Pinedo-Vargas L,
104. Ball MJ, Lukiw WJ, Kammerman EM, Hill
Granados O, Cuervo-Zanatta D, Rojas-Santiago JM. Intracerebral propagation of Alzheimer’s disease:
E, et  al. Bioactive food abates metabolic and syn- strengthening evidence of a herpes simplex virus eti-
aptic alterations by modulation of gut micro- ology. Alzheimers Dement. 2013;9(2):169–75.
biota in a mouse model of alzheimer’s disease.