REVIEwS

Immunoneuropsychiatry — novel
perspectives on brain disorders
Katrin Pape1, Ryad Tamouza2,3,4, Marion Leboyer2,3,4,5 and Frauke Zipp   1,5*
Abstract | Immune processes have a vital role in CNS homeostasis, resilience and brain reserve.
Our cognitive and social abilities rely on a highly sensitive and fine-tuned equilibrium of immune
responses that involve both innate and adaptive immunity. Autoimmunity , chronic inflammation,
infection and psychosocial stress can tip the scales towards disruption of higher-order networks.
However, not only classical neuroinflammatory diseases, such as multiple sclerosis and
autoimmune encephalitis, are caused by immune dysregulation that affects CNS function. Recent
insight indicates that similar processes are involved in psychiatric diseases such as schizophrenia,
autism spectrum disorder, bipolar disorder and depression. Pathways that are common to these
disorders include microglial activation, pro-inflammatory cytokines, molecular mimicry ,
anti-neuronal autoantibodies, self-reactive T cells and disturbance of the blood–brain barrier.
These discoveries challenge our traditional classification of neurological and psychiatric diseases.
New clinical paths are required to identify subgroups of neuropsychiatric disorders that are
phenotypically distinct but pathogenically related and to pave the way for mechanism-based
immune treatments. Combined expertise from neurologists and psychiatrists will foster
translation of these paths into clinical practice. The aim of this Review is to highlight outstanding
findings that have transformed our understanding of neuropsychiatric diseases and to suggest
new diagnostic and therapeutic criteria for the emerging field of immunoneuropsychiatry.

Crosstalk between the immune and nervous systems processes and brain diseases. On the basis of the find-
is receiving increasing attention in a wide spectrum of ings and studies discussed, we suggest a new clini-
neurological and psychiatric diseases. As pathways cal approach to neuropsychiatric disorders from an
emerge that are common to disorders from both fields, immunological perspective.
the traditional boundaries between neurological and
psychiatric disorders are becoming blurred. Novel dis- Neuroimmune interplay in brain health
1
Department of Neurology, coveries about the roles of the immune system in CNS The CNS has traditionally been regarded as a site of
University Medical Center of function and in disease together with tremendous devel- immune privilege. However, despite being protected by
the Johannes Gutenberg opments in immune therapies make this topic of great specialized physical barriers, the brain is neither inert
University Mainz,
interest. This rapidly developing research is providing nor immunologically separated from the peripheral
Mainz, Germany.
new perspectives not only on disease and therapeu- immune system. Instead, an interplay between neurons,
2
Inserm, U955, Institut
Mondor de la Recherche
tic targets but also on brain reserve and resilience in glial cells and the immune system contributes to func-
Biomédicale, Créteil, France. neuropsychiatric disorders. tional properties, such as cognition, social behaviour and
3
Fondation FondaMental, In this Review, we first give an overview of neuro- learning performance in the healthy brain.
Créteil, France. immune interplay and inflammatory influences in Increasing evidence indicates that specific commu-
4
AP-HP, Department of the healthy brain and in disease. Accumulating data nication occurs between neurons and microglia, the
Psychiatry of Mondor suggest immune and autoimmune contributions to a brain parenchyma-resident macrophages that account
University Hospital, DHU wide variety of neurological and psychiatric disorders for ~10% of CNS cells. Microglia permanently survey
PePsy, University of Paris-Est-
(Box 1). These findings are reflected in a growing num- their local environment with fine, motile processes1
Créteil, Créteil, France.
ber of therapeutic studies of mechanism-based immune and, during brain development, they have an impor-
5
These authors contributed
equally: Marion Leboyer,
treatments in subgroups of patients with neurological tant role in synaptic pruning mediated by the comple-
Frauke Zipp and psychiatric disorders. In addition, we discuss recent ment proteins C1q and C3, as they phagocytose the
*e-mail: zipp@uni-mainz.de insights into the role of psychosocial stress and infectious complement-tagged synapses2,3. In addition to micro-
https://doi.org/10.1038/ events in the CNS that provide a mechanistic corner­ glia, soluble factors such as chemokines and chemokine
s41582-019-0174-4 stone for our understanding of neuroinflammatory receptors also contribute to physiological developmental

Nature Reviews | Neurology

Reviews

Key points a T cell receptor for a non-CNS-specific antigen17. The

advantageous effects of self-reactive T cells on the main-
• At the interface of neurological and psychiatric disorders, immune processes are tenance of neuronal function have led to the concept of
major factors in CNS health and disease; the immune system contributes to CNS protective autoimmunity, in which adaptive immune cell
homeostasis, resilience and brain reserve. function maintains tissue function. Further beneficial
• Although a certain level of neuroimmune interplay is required for optimal brain roles of the immune response in the CNS — again going
functioning, chronic inflammation and latent infections can cause higher-order against the anticipated pathogenic activity of inflamma-
network disturbances, resulting in cognitive and behavioural impairment.
tion — include protective roles for T helper 2 (TH2) cells
• Psychosocial stress correlates with inflammatory processes in the CNS. in CNS injury and for IL-4 signalling in neurons in mul-
• Immune dysregulation plays a role not only in classical autoimmune brain diseases such tiple sclerosis (MS) models18,19. The complex and diverse
as multiple sclerosis and autoimmune encephalitis but also in psychiatric disorders roles of adaptive immune cells are further underlined by
such as schizophrenia, autism spectrum disorder, bipolar disorder and depression.
the finding that a lack of B cells does not impair learning
• Immune treatments are emerging as therapeutic options for subgroups of patients behaviour in mice20.
with brain disorders that are associated with an inflammatory phenotype.
The interplay between the CNS and the periphery is
• New diagnostic and therapeutic criteria are required to translate immunopathogenic mediated by two drainage systems. The first is the glym-
findings into individualized treatment options for patients with neuropsychiatric
phatic system, a glial-dependent perivascular network that
disorders.
ensures provision of nutrients for neurons and glia and
clearance of extracellular metabolites, such as lactate21.
processes in the CNS; for example, CXC-chemokine Activation of the glymphatic system is higher during
receptor 4 (CXCR4) and CXCR7 are involved in the sleep than during wakefulness, underlining the impor-
migration of cortical interneurons4. In the healthy adult tance of sleep for removal of potentially neurotoxic waste
brain, microglia have a role in the homeostasis of synap- products22. The second is the meningeal lymphatic system
tic circuits in the CNS5. Their expression of receptors for that lines the dural sinuses and allows drainage from the
purines (such as ATP) and common neurotransmitters CNS to deep cervical lymph nodes, the discovery of which
(such as glutamate) enables them to sense local neuronal established the missing link to the peripheral immune
activity6. In response, microglia can directly contact system23,24. Via this system, molecules and immune cells
neurons via outgrowth of processes5 or can indirectly from the CNS can be transported to lymphoid organs
modulate neuronal firing rate via release of extracellular and evoke an immune response that involves subsequent
vesicles7 or signalling molecules, such as tumour necro- migration of immune cells to the brain25.
sis factor (TNF)8. In this way, microglia contribute to
activity-induced synaptic plasticity, for example, during Cognitive performance in inflammation
motor learning and memory. During systemic inflam- As described above, innate and adaptive immunity are
mation, microglia can become activated and produce essential for CNS homeostasis. Consequently, distur-
pro-inflammatory mediators and induce phagocytosis9. bances in the equilibrium of immune cells, neurons and
Microglia comprise the innate CNS immune com- glial cells needed for healthy CNS function are likely to
partment; the existence of adaptive immune cells in modify cognitive performance (Fig. 1a). In this section,
the brain has long been considered a sign of disease. we consider the evidence that low-grade inflammation
However, in healthy individuals, antigen-presenting and inflammation in response to infection can alter
cells and T cells patrol the brain’s borders, residing in cognitive performance.
the meninges, the choroid plexus and the cerebro­spinal Recognition of pathogen-associated or damage-
fluid (CSF)10. These patrol cells provide additional pro- associated molecular patterns and subsequent phago-
tection to that provided by the blood–brain barrier cytosis or cytokine production by microglia or invad-
(BBB) — a physical boundary that consists of the basal ing macrophages, representing the innate immune
lamina of endothelial cells, tight junctions between them response in the CNS, has to be distinguished from the
and astrocyte end-feet processes — and the blood–CSF antigen-specific responsiveness of adaptive immune cells.
barrier at the choroid plexus11, which is composed of
epithelial cells12. Notably, evidence suggests that the Low-grade inflammation. A compelling example of the
choroid plexus has a role not only in transmigration but pathogenic influence of the immune system on brain
also in stimulation of T cells in response to peripheral function is the change in mood, social behaviour and
inflammatory signals13. cognitive abilities — known as sickness behaviour —
Although T cells do not generally penetrate the paren- upon infection and systemic inflammation. The release
chyma in non-inflammatory conditions, they can release of pro-inflammatory cytokines, such as IL-1β, IL-6
soluble cytokines that affect CNS function14,15. Studies and TNF, outside the CNS affects the brain via neu-
in mice have demonstrated that adaptive immunity is ral (mainly vagal) pathways, interaction with cytokine
necessary for cognitive performance: mice with severe receptors on cerebral endothelial cells and/or microglial
combined immune deficiency exhibited impaired spa- activation26. The behavioural effects, such as social with-
tial learning and memory, but symptoms were reversed drawal and fatigue, are assumed to be adaptive responses
by injection of exogenous T cells16. Similar observations that increase survival of the host27. Similarly, treatment
have been made for social behaviour14. From a neuro- with the cytokine IFNβ, for MS or chronic viral hepa-
biological perspective, levels of neurogenesis in adult titis, for example, has been associated with depressive
transgenic mice that overexpress a CNS-specific T cell symptoms as an adverse effect28. Sickness behaviour can
receptor are higher than those in mice that overexpress be diminished after repeated subthreshold exposure to

www.nature.com/nrneurol
 Reviews

pathogens, a mechanism that is referred to as euflam- Infection-related inflammation. Infection can initiate
mation and that evidence suggests is the result of chronic inflammation that alters cognitive function.
tolerogenic processes29. Data from a Danish longitudinal register showed that
In contrast to this transient low-grade inflammation, prior hospitalization for an autoimmune disease or
chronic low-grade inflammation can occur and cause infection increased the risk of a major mood disorder
neurotoxicity and neurodegeneration30. For example, by 45% and 62%, respectively37. In rodents, even infec-
cytokines can have neurotoxic effects by increasing pro- tion and systemic inflammation in the fetus during the
duction of reactive oxygen species, reducing monoamine prenatal or perinatal period can cause long-term cogni-
transmission and potentiating glutamatergic transmis- tive damage, including learning, memory and attention
sion. Chronic, sterile, low-grade inflammation occurs abnormalities, a model that explains why early infection
during human ageing and can contribute to age-related increases the risk of psychosis in young adulthood38,39.
diseases; this process is referred to as inflammaging31. Similarly, a viral infection in the mother during the first
Chronic inflammation is implicated in schizophrenia and trimester of pregnancy and bacterial infection during the
other psychiatric disorders32, and systemic inflammation, second trimester of pregnancy were associated with
as well as acute infections, have been associated with an the development of autism spectrum disorder (ASD),
increased rate of cognitive decline and exacerbation of a pervasive neurodevelopmental disorder defined by
symptoms in patients with Alzheimer disease (AD)33. impairments in social skills and stereotypical behaviour,
Chronic low-grade inflammation is also present in people among the children in a large Danish cohort40.
who are obese34, and this inflammation influences cog- In children, throat infection with Streptococcus pyo-
nitive performance by damaging neuronal circuits and genes has repeatedly been associated with subsequent
the BBB and by activating pro-inflammatory immune neuropsychiatric disorders41. In particular, these dis-
cells35. These obesity-related mechanisms were initially orders include Sydenham chorea, a movement disor-
observed in the hypothalamus, but some evidence sug- der related to rheumatic fever that occurs in temporal
gests that they can also occur in the hippocampus, cortex, relationship to group A streptococcal infection42, and
brainstem and amygdala36. conditions encompassed by the umbrella term of pae-
Overall, in the interplay between the brain and the diatric autoimmune neuropsychiatric disorders asso-
immune system, inflammation is a double-edged sword. ciated with S. pyogenes (PANDAS) and concomitant
Beneficial effects such as euflammation and evolution- with obsessive–compulsive disorder, tic disorder or
arily advantageous sickness behaviour are on one side, choreiform motoric hyperactivity43. The suspected
but chronic inflammation that leads to neuronal dam- pathophysiology of PANDAS is cross reactivity of
age, cognitive decline and possibly dementia is on the anti-streptococcus A antibodies with brain tissue
other side. The right balance of inflammation is needed due to molecular mimicry44 that initiates an adaptive
for optimal CNS function. immune cell response. Nevertheless, in a prospective
study of children with post-streptococcal neuropsy-
Box 1 | Neuropsychiatric disorders with inflammatory disturbance chiatric symptoms, no correlation was seen between
clinical symptoms and a change in autoimmune mark-
Involvement of immune dysfunction in pathogenesis has been studied in a
ers, such as anti-neuronal antibodies or inflam­matory
broad range of neuropsychiatric diseases. This review focuses on the following
disorders:
cytokines45. Therefore, despite confirmation on an
epide­m iological level, the aetiological association
• Multiple sclerosis
between streptococcal infection and psychiatric symp-
• Neuromyelitis optica spectrum disorder
toms is still controversial and the subject of an ongoing,
• Autoimmune encephalitis (paraneoplastic, idiopathic or triggered by infection) large observational study46.
• Schizophrenia Severe infections, such as herpes simplex virus (HSV)
• Autism spectrum disorders encephalitis at any time in the lifespan, have been asso-
• Depression ciated with long-term impairment of memory and brain
• Bipolar disorder atrophy47. Herpes family viruses persist in the host and
• Dementia, especially Alzheimer disease require constant immune surveillance in order to pre-
vent reactivation. For example, latent cytomegalovirus
In addition, immune mechanisms are likely to contribute to many other
(CMV) infection leads to accumulation of functionally
disorders. These disorders include the following:
exhausted effector T cells while the naive T cell pool is
• Myelin oligodendrocyte glycoprotein antibody spectrum disorder
diminished48. In the prospective Northern Manhattan
• Chronic inflammatory optic neuritis Study, a high infectious burden (assessed with serolog-
• Rasmussen encephalitis ical markers for Chlamydia pneumoniae, Helicobacter
• Susac syndrome pylori, CMV, HSV-1 and HSV-2) was associated with
• Stroke cognitive decline independently of cardiovascular risk
• Parkinson disease among a cohort of 1,625 patients49. Evidence suggests
• Amyotrophic lateral sclerosis that among patients with schizophrenia or bipolar dis-
• Huntington disease order, infections with Toxoplasma gondii, HSV and CMV
• Obsessive–compulsive disorder affect the cognitive dysfunction already present in these
disorders, in particular working memory50. Another
• Anxiety disorders
example of the intricate relationship between cognition
• Eating disorders
and chronic infection is HIV-associated neurocognitive

Nature Reviews | Neurology

Reviews

a Inflammation b
(levels of cytokines
and myeloid cells) Non-microglia
Lymphatic vessel myeloid cells

T cells Oligodendrocyte
Astrocyte
Activated
myeloid cell

Infectious Neuron
burden

Brain
parenchyma Pro-inflammatory T cell
Regulatory
Neuronal T cell
network Microglia
Non-microglia
Blood–brain barrier myeloid cell

Cognitive
performance
Blood vessel Lymphocyte

Non-self-reactive Self-reactive

Fig. 1 | The interplay between the immune system and the CNS. a | A certain level of inflammation and autoimmunity
is necessary for optimal function of the CNS, but an overwhelming immune reaction leads to neuronal loss and impaired
cognition. b | Immune cell infiltration into the CNS in disease. Glial cells, including microglia and astrocytes, are resident
in the CNS. Lymphocytes and non-microglia myeloid cells are both thought to be present in the lymphatic vessels in health
but might transmigrate through the blood–brain barrier in inflammatory diseases.

disorder, which emerges in patients with HIV infection amyloid plaques are thought to alter the neuroimmune
despite highly active antiretroviral therapy51. crosstalk in various ways. For example, HSV-1 proteins
Our understanding of the mechanistic links between debilitate neuronal autophagy and, consequently, anti-
infectious burden and neuropsychiatric diseases is still gen presentation58. In mice, recurrent asymptomatic
in its infancy. In addition to molecular mimicry that activation of HSV-1 leads to upregulation of markers
triggers autoimmune reactions, a possible mecha­ of neuroinflammation (for example, Toll-like recep-
nism is stress-induced potentiation of microglial tor-4, IFNα and IFNβ) and early neurodegeneration59.
inflammasome activation, which causes an increase in In addition, extracellular deposits of Aβ activate rest-
pro-inflammatory mediators52,53. In this way, both innate ing microglia and trigger them to attack neurons via
(microglial activation with subsequent phagocytosis or various pathways, such as NADPH oxidase activa-
cytokine production) and adaptive immune responses tion, inducible nitric oxide synthase expression and
(antigen-specific responsiveness and production of anti- phagocytosis60.
bodies) contribute to pathology. In addition, infectious In summary, owing to the sensitive equilibrium of the
agents might act via epigenetic pathways to modulate immune system and brain function, infectious agents
the innate immune cell repertoire, thereby influencing the might tip the scales towards pathology. Furthermore,
risk profile for neurodegenerative disorders54. infections can trigger autoimmune phenomena that
In addition to the cumulative effect of infectious might contribute to a broad spectrum of neuropsychiatric
burden, pathogens contribute directly to neurodegen- diseases, discussed in more detail below.
eration, and therefore cognitive decline, in elderly peo-
ple. One example that has been investigated extensively Autoimmunity
but is still incompletely understood is the association Nonspecific inflammatory processes underlie the
between HSV infection and AD. On a population level, low-grade inflammation described above, but neuro­
data from a Swedish cohort suggest that reactivation inflammation can also be driven by autoimmunity.
of HSV-1, indicated by high serum levels of anti-HSV Autoimmune responses can be antibody driven or cell
immunoglobulin M (IgM) antibodies, increases the risk driven and, depending on the autoantigen involved, can
of developing AD55. In vitro experiments have shown involve antigen presentation by major histocompatibil-
that HSV-1 induces increases in intracellular amyloid-β ity complex (MHC) class II. Autoimmune responses can
(Aβ) levels and tau phosphorylation, both of which are follow infection, and molecular mimicry can under-
markers of AD56. In addition, HSV-1 DNA has been lie these responses41,61. Innate and adaptive immune
found in amyloid plaques in patients with AD, further mecha­nisms have roles in autoimmunity. What we know
underlining a link between the virus and pathological about autoimmune responses and neuro­psychiatric dis-
aggregations57. Virus-mediated cytotoxicity does not orders is largely based on specific disorders, discussed
explain all clinical developments in AD, but viruses and in detail below.

www.nature.com/nrneurol
 Reviews

Lessons from multiple sclerosis. Neurological auto- MS brains75. However, direct recognition of neurons by
immune diseases are mainly considered to be medi- patho­genic T cells has also been proposed76,77, and neuro­
ated by autoaggressive lymphocytes or anti-neuronal nal loss could occur as a result of Wallerian degene­ration
or anti-glial autoantibodies. These conditions include and mitochondrial dysfunction78. Grey matter damage,
neuro­myelitis optica and potentially disabling, or some- measured as the cortical lesion load or cortical thinning,
times life-threatening, autoimmune encephalomyelitis. is a key predictor of progressive disease and cognitive
MS is the classical and most common chronic autoim- decline79. Even in clinically stable patients, the pres-
mune encephalomyelitis; this disease leads to demyeli- ence of gadolinium-enhancing lesions on brain MRI
nation and progressive neurodegeneration. In early MS, is associated with impaired performance in the Paced
most patients experience transient physical deficits, such Auditory Serial Addition Test (PASAT), a screening tool
as optic neuritis, hemiparesis or sensory disturbances, for cognitive dysfunction80. These findings underline
in a relapsing–remitting presentation. Although these the fact that active neuroinflammation detectable with
physical deficits are the most obvious consequences, an gadoli­nium enhancement affects cognitive function and
increasing body of evidence illustrates that cognitive might cause detrimental effects on neuronal networks
deficits are abundant and extensive in MS62. and connectivity.
Patients with MS can experience neuropsychological Although rare, psychotic symptoms can also occur
symptoms, such as mild cognitive alterations or depres- in MS. A genetic overlap between MS and schizophre-
sion, related not only to relapses but also to phases of nia has been identified, and the genes that are common
remission63. The most common cognitive deficits, which to both conditions are immune related81. Despite being
occur in 40–60% of patients with MS, are reduced pro- separate disease entities, MS and classical psychiatric
cessing speed and impaired memory and/or executive disorders are accompanied by higher-order network
function, and these symptoms can improve to some disturbances82–85 (Fig.  2a). The study of MS has pro-
extent with immunomodulatory treatment64. Behavioural vided fundamental insights into the crosstalk between
signs, such as anxiety, can also occur and are currently the immune system and the nervous system, and these
seen as comorbidities but might be a consequence of the insights will be highly valuable for the study of other
pathology65. Notably, cognitive impairment at the time diseases in which underlying immune pathology has
of MS diagnosis predicts disability progression, transi- recently been identified, discussed below.
tion to secondary progressive MS and cortical thinning66.
In line with these findings, the risk of cognitive impair- Lessons from autoimmune psychosis. A compelling
ment increases with progression of disease and is high- example of the overlap between psychiatric and neuro­
est in secondary progressive MS67. Nevertheless, 30–40% logical pathologies is the occurrence of psychosis as a
of patients with clinically isolated syndrome, which is result of autoantibodies against neuropil, which can
considered to be a precursor of MS, have cognitive defi- develop spontaneously in neoplastic diseases or after
cits68,69, and in a study of so-called benign MS (defined viral infections, particularly HSV encephalitis86. Some
by an Expanded Disability Status Scale (EDSS) score of these autoantibodies target intracellular antigens,
≤3.0 despite a disease duration ≥15 years), up to 45% of such as onconeural proteins or 65 kDa glutamic acid
patients had cognitive impairment70. decarboxylase (GAD65). The presence of these anti-
Inflammation in MS, the local effects of which bodies is diagnostically useful but considered to be an
include microglial activation and myelin and neuro­ epiphenomenon, as cytotoxic T cells are responsible
nal damage, is thought to be primarily mediated by for neuronal damage that is associated with neuropsy-
auto­reactive T cells. Pro-inflammatory TH17 cells are chiatric symptoms87. However, a broad spectrum of
thought to migrate into the CNS and subsequently extracellular antigens involved in synaptic transmission
increase BBB permeability, enabling invasion of other and plasticity are located on the cell surface; therefore,
immune cells71 (Fig. 1b). Other important players include autoantibodies against these antigens that are present
regulatory T cells, which are functionally impaired in in patients with psychosis seem highly likely to have a
people with MS72, and B cells, the role of which is high- direct pathogenic role. One example of such an anti-
lighted by the effects of B cell-depleting therapies73. body is the autoantibody against the GABAA receptor, a
Relapses in patients with MS are influenced by stressful postsynaptic chloride channel that mediates fast inhib-
life events and by infections: a systematic meta-analysis itory neurotransmission in the mammalian brain; this
has shown that upper respiratory tract infections have autoantibody causes encephalitis with therapy-resistant
the most pronounced effect on relapse rates74. These epileptic seizures88. Moreover, antibodies against the
observations underline the role of inflammation in α-amino-3-hydroxy-5-methyl-4-isoxazole propionic
relapse activity, as both systemic infection and stress acid (AMPA) receptor can have a pathological role:
induce a pro-inflammatory status, as discussed in more administration of human pathogenic antibodies against
detail below. the AMPA receptor subunit GluA2 impairs long-term
A specific autoantigen has not yet been identified synaptic plasticity in vitro and affects learning and
in MS, but autoreactive T cells are widely assumed to memory in mice in vivo89.
target proteins of the myelin sheath, causing demyelina- More controversial is the importance of antibodies
tion and, thereby, white matter damage. Demyelination against the N-methyl-d-aspartate (NMDA) glutamate
deprives neurons of protective factors and has been receptor, one of the most commonly observed antigens
described to decrease axonal transport and synaptic in autoimmune encephalitis. Patients with this condi-
density in demyelinated hippocampi from post-mortem tion often present with anxiety, sleep disorders, mania,

Nature Reviews | Neurology

Reviews

a Upregulation of lupus erythematosus, and these antibodies cause neuro­

pro-inflammatory genes nal damage via activated microglia and complement
Activation of Altered microglial component C1q when administered to mice93,94.
adaptive immune cells function Despite this evidence, the pathological role of neuro­
transmitter receptor autoantibodies came into ques­
Autoantibodies Infectious burden
tion when anti-NMDA receptor autoantibodies were
detected in serum from ~10% of healthy controls as
Soluble inflammatory Maternal immune well as in patients with pure psychotic symptoms95.
markers (IL-6 and CRP) activation
Similar results have been seen for a variety of brain
antigens, challenging the pathological relevance of
CNS-specific autoantibodies96; BBB integrity has been
proposed as a pivotal factor in the clinical outcome95.
In another study, anti-NMDA receptor autoantibodies
from psychotic patients altered synaptic transmission
and long-term potentiation in cultured neurons and in
? mouse brain whereas those from healthy controls did
Stereotypical
Paranoia
behaviour not97. Independent of immune pathogenesis, glutamate
receptor hypofunction has been proposed as a key factor
Anhedonia Hallucinations in the development of schizophrenia98.
Epileptic Impaired
In combination, these results provide evidence that
seizures Cognitive social skills specific and different mechanisms underlie encephalitis
decline and autoimmune psychosis, in which autoantibodies
against NMDA receptors have different effects on the
b organization of the glutamate synapse through receptor
internalization or abnormal NMDA receptor dynamics.
Given that patients with schizophrenia who are posi-
tive for anti-NMDA receptor antibodies do not exhibit
classical signs of encephalitis97, we conclude that differ-
Neurological Psychiatric ent immune alterations mediated by similar antibodies
diseases diseases
result in different neuropsychiatric entities.

Immunopsychiatry — an emerging field

On the basis of the data discussed above, the concept of
autoimmune psychosis has become a compelling exam-
ple of the interface between neurological (autoimmune
Fig. 2 | overlap of neuropsychiatric disorders. a | Higher-order network disturbances encephalitis) and psychiatric (psychosis) disorders. In
are caused by common pathophysiological mechanisms of immune dysregulation (top). this section, we discuss accumulating evidence that sug-
How clinically distinct phenotypes emerge from these pathways is an unresolved gests that immune dysregulation is involved in a broad
question (bottom). b | As pathways emerge that are common to neurological and spectrum of psychiatric diseases. Owing to a lack of speci­
psychiatric autoimmune brain diseases, the traditional boundaries between these fic markers, studies in the emerging field of immuno­
disorders become blurred. Identification of subgroups of patients who are likely to psychiatry currently focus on systemic measures of
benefit from immunotherapeutic approaches is crucial. CRP, C-reactive protein. inflammation. A growing number of studies have iden-
tified nonspecific inflammatory disturbances in sub-
paranoia, memory impairment and disintegration of groups of patients, but knowledge of specific pathogenic
language, followed by a phase in which agitation and pathways remains scarce.
catatonia alternate, accompanied by abnormal move- One overarching concept is that immunopsychiat-
ments and autonomic instability90. In humans, antibody ric diseases involve a generally overactivated immune
titres in the CSF correlate with the clinical course of dis- system. In general, overactivation of the immune sys-
ease90. Experiments in rats indicate a pathophysiological tem is thought to increase brain vulnerability, increas-
role for anti-NMDA receptor antibodies. For example, ing the risk of psychiatric symptoms upon a so-called
in rat hippocampal neurons that have been incubated second hit later in life99. On an epidemiological level,
with anti-NMDA receptor antibodies, NMDA receptor this hypothesis is supported by a link between psychi-
internalization and reduced NMDA receptor-mediated atric diseases, such as schizophrenia, depression and
synaptic currents were observed91. In addition, treat- anxiety, and systemic autoimmune disorders100,101. Case
ment of rats with NMDA receptor blockers evoked reports similarly support this hypothesis; one striking
stereotyped behaviour and cataleptic freezing that were example is the development of severe psychosis in an
comparable to symptoms of autoimmune encephalitis individual who received a stem cell transplantation
in humans92. NMDA receptor encephalitis can occur from their brother with schizophrenia, indicating that
after HSV encephalitis in up to 27% of patients, indicat- adoptive immune transfer had a role102. Furthermore,
ing a role for molecular mimicry61. Furthermore, anti- extensive studies of cytokine profiles in disorders such
bodies that cross-react with the NMDA receptor have as schizophrenia, depression, suicidal ideation and
been found in patients with neuropsychiatric systemic post-traumatic stress disorder reveal wide-ranging

www.nature.com/nrneurol
 Reviews

dysregulation of mainly pro-inflammatory markers, tissue were detected in the serum of mothers of chil-
in particular IL-6, IL-2 receptor, IL-1β, IL-17 A and dren with ASD but not in mothers of healthy children121.
C-reactive protein (CRP)103–107. Several studies iden- Another study produced the same observation in up to
tified a positive correlation of serum levels of inflam- 23% of mothers of children with ASD122. In animal stud-
matory markers with disease severity and a negative ies, intravenous administration of 73 kDa and 37 kDa
correlation with cognitive performance103,108. IgG from mothers of children with ASD to pregnant
Changes in the innate CNS immune compartment rhesus macaques led to abnormal social behaviour in the
have also been associated with psychiatric disorders. macaque offspring123. Targets for these antibodies include
Either excessive upregulation or downregulation of proteins with functions in neurodevelopment, such as
microglia function evokes detrimental effects. For exam- contactin-associated protein-like 2 (CASPR2)122,124.
ple, genetic defects in microglial signalling pathways Although the exact mecha­nism by which these maternal
cause abnormal development of brain circuits and neuro­ antibodies affect the fetus is not yet understood, the find-
psychiatric symptoms, as in hereditary diffuse leuko­ ings in humans and the adoptive transfer study suggest
encephalopathy with spheroids (HDLS)109. Disruption that placental transfer of maternal antibodies contributes
of microglial function in later life can be initiated by to disease development, at least in a subgroup of patients.
psycho­social and environmental factors. For example, in Aberrant genetic regulation underlies most of the
a mouse model, chronic stress with depressive behaviour immune dysfunction discussed above, and a polygenic
has been associated with microglial loss in the hippo­ contribution to the risk of several severe psychiatric
campus110. By contrast, depression and schizophrenia disorders has been confirmed in a meta-analysis of
have been associated with increased microglial activity genome-wide association studies in which the MHC was
measured with PET, and this increased activity could be the most relevant shared risk locus125. Genes in the MHC
used to identify patients at high risk of disease exacer- are involved in physiological processes, including CNS
bations111,112. Although microglial activation in humans development and homeostasis, but are common among
has been studied only indirectly, early life challenges are neuropsychiatric disease-associated loci126,127. For exam-
assumed to ‘prime’ microglia and increase their response ple, the MHC-resident complement C4 cluster has been
to subsequent inflammatory stimuli113. implicated in the risk of schizophrenia128,129. Therefore,
A crucial discovery that many immune factors are fine dissection of the genetic contribution of the MHC
influential prenatally led to use of the so-called mater- is a promising approach to increasing our understanding
nal immune activation model in order to study ASD of brain disorders.
and schizophrenia in rodents. In this model, IL-6 was Although we are only beginning to develop a deeper,
identified as a key mediator of inflammatory effects mechanistic understanding of immunological pathways
on fetal brain development114. In mice, autism-like involved in the pathogenesis of psychiatric disorders, the
behaviour in offspring requires the presence of mater- emerging field of immunopsychiatry provides new per-
nal RORγt-positive TH17 cells and IL-17a downstream spectives on the disorders discussed. As even early evo-
of IL-6 (ref.115). In subsequent studies, a mechanism of lutionary processes such as the integration of sequences
ASD pathogenesis has been proposed in which TH17- of human endogenous retroviruses into the genome
dependent loss of inhibitory interneuron networks presumably through repeated infections were found to
leads to increased cortical activation in the primary be associated with neurological and psychiatric disor-
somatosensory cortex116. ders130,131, it becomes clear that dysfunctional processes
A study published in 2017 provides evidence that occur on an ancestral, maternal and individual level.
maternal gut commensal bacteria have a role in IL-17a
production in mothers117. These findings support the Immunomodulatory treatment
idea of a gut–immune–brain axis that has been impli- Classical neurological autoimmune diseases have long
cated in classical autoimmunity (such as modulation of been treated with immunomodulatory drugs. Routinely
the balance between pro-inflammatory and regulatory used treatments for acute exacerbations include steroids,
T cells by gut bacteria in MS118) and are in line with pre- plasmapheresis, intravenous immunoglobulin, cyclo-
vious reports of an altered intestinal barrier in people phosphamide and B cell-depleting monoclonal anti-
with ASD and their close relatives119. In rodents, the role bodies; a broadening range of disease-modifying drugs
of maternal immune activation in ASD was confirmed is emerging from the development of therapies for MS
at the transcriptome level, in which immune activation (Box 2). Immunosuppression is also an established treat-
in the mother caused dysregulation in the fetal brain ment strategy for autoimmune encephalitis, but clinical
expression of genes involved in developmental pro- studies of such treatments for psychiatric diseases are
cesses120. Although animal studies provide the proof of still in progress.
concept and agree with findings in diseases of neuro­ In a pilot study published in 2018, treatment with
logical immune dysfunction, such as MS, the mecha- high-dose intravenous immunoglobulin had positive
nisms described are plausible in humans but remain effects on scores in several cognitive and behavioural
speculative. tests for children with ASD and evidence of a dysregu­
Cytokine signalling is not the only aspect of mater- lated immune system132. In another study, intravenous
nal immunity that affects the developing fetal brain — immunoglobulin had beneficial effects for subgroups of
B cell-mediated immune responses and production children with ASD and inflammation133. A series of case
of antibodies can also have effects. For example, in an reports of immune modulatory treatment in PANDAS
epidemio­logical study, antibodies against fetal brain disorders demonstrates how an understanding that

Nature Reviews | Neurology

Reviews

Box 2 | Immunomodulatory drugs available for the treatment of CNS autoimmune diseases
Immunomodulatory drugs and treatments applied in • Glatiramer acetate
a broad range of autoimmune diseases • Teriflunomide
• Steroids • Dimethyl fumarate
• Plasmapheresis
• Fingolimod
• Intravenous immunoglobulin
• Natalizumab
• Cyclophosphamide
• Alemtuzumab
• Methotrexate
• Cladribine
• Mycophenolate mofetil
• Rituximab
• Azathioprin
• Mitoxantrone • Ocrelizumab
• Rituximab other drugs with partial immunomodulatory
• Ocrelizumab effects
• Infliximab • Nonsteroidal anti-inflammatory drugs
• Selective serotonin reuptake inhibitors
Disease-modifying drugs from the range of
therapeutics for multiple sclerosis • Minocycline
• IFNβ • Tryptophan metabolites

auto-inflammatory processes are involved in the patho- been done in humans, but a meta-analysis of existing
genesis can lead to successful treatment134. However, clinical data indicates beneficial effects of mino­cycline
syste­matic treatment approaches are still lacking and are in patients with depression, thereby providing a proof
currently merely performed with antibiotics. of concept144. Minocycline also had positive effects as
Currently, monoclonal antibody treatments are an adjunctive treatment to risperidone in children with
being trialled in schizophrenia. In one study, use of the ASD145, although not as a single medication in a small
mono­clonal antibody natalizumab, which targets the cell pilot study146. In addition to its inhibitory effects on
adhesion molecule α4 integrin and is currently used for microglia, minocycline has antibiotic properties that
treatment of MS, is being tested135. The recombinant might modify the gut flora, which could contri­bute to
humanized anti-human IL-6 receptor monoclonal anti- beneficial treatment outcomes. Furthermore, in depres-
body tocilizumab is being tested in two136,137 ongoing sion, anti-inflammatory treatment with the TNF antago­
studies, one of which136 focuses on patients with signs nist infliximab had beneficial effects in patients with
of peripheral inflammation. In one case report, remis- treatment-resistant depression and high baseline levels
sion of treatment-resistant schizophrenia was seen in a of inflammatory biomarkers147.
patient who underwent bone marrow transplantation In summary, immunomodulatory therapies are effec-
for acute myeloid leukaemia, providing evidence that tive options for a range of neurological and psychiatric
this treatment could be an effective, or even curative, diseases. To develop individualized therapeutic strate-
option for severe schizophrenia138. Similarly, infusion of gies, subgroups of patients with psychiatric disorders
autologous umbilical cord blood into children with ASD and signs of immune dysregulation must be identified
led to clinical improvement of core autistic symptoms139. (Fig. 2b). Clinical studies are more likely to reveal con-
Evidence that immune dysfunction is involved in sistent positive effects if they include patients with an
depression is supported by accumulating data that sug- inflammatory phenotype, indicated by MRI or high
gest that established antidepressant drugs such as selec- levels of inflammation markers in the CSF or serum.
tive serotonin reuptake inhibitors (SSRIs) exert their Furthermore, we assume that there is a narrow thera-
effects, at least partly, via anti-inflammatory effects. peutic time window for maximal therapeutic efficacy
For example, a meta-analysis showed that SSRIs reduce and avoidance of irreversible CNS injury. For example,
levels of IL-1β and IL-6 in patients with depression140. timely immunomodulatory treatment in autoimmune
In line with these observations, synthetic metabolites encephalitis probably prevents neuronal damage and
of tryptophan, a precursor of serotonin, suppressed permanent neurological deficits148, and this concept
pro-inflammatory T cells and autoimmune neuro­ might hold true for other neuropsychiatric disorders.
inflammation in experimental autoimmune encephalitis,
a mouse model of MS141. Additional support for the con- Resilience and reserve
cept that immune pathology underlies depression comes A well-recognized risk factor for exacerbation in MS
from a meta-analysis of randomized clinical trials that and many psychiatric disorders is psychosocial stress.
showed that, despite heterogeneous study results, non- Some evidence suggests that the association between
steroidal anti-inflammatory drugs had an overall positive psychiatric disorders and stress is mediated, at least
effect on depression142. In rats, the microglia inhibitor in part, by neuroinflammatory processes, including
minocycline improved symptoms of depression in the microglial activation. For example, the ‘two-hit hypo­
forced swimming test143, an observation that is consistent thesis’ proposes that stress in early life increases base-
with reports that microglia are overactivated in depres- line microglial activity, thereby increasing the risk of
sion112. A large randomized trial of minocycline has not psychiatric disorders upon subsequent challenges later

www.nature.com/nrneurol
 Reviews

in life149,150. Accordingly, a meta-analysis of clinical stud- findings for neurogenesis discussed above, self-reactive,
ies demonstrated a pro-inflammatory status (high levels brain-specific T cells are needed for resilience in mice,
of CRP, IL-6 and TNF) in adults with a history of child- leading to the idea that vaccination with CNS-related
hood trauma151. Similarly, in a prospective cohort study, peptides could stimulate coping behaviour157.
patients with a history of childhood maltreatment and From studies of chronic autoimmune inflammation
depression had higher levels of CRP than participants of the CNS in MS, we know that the brain has reserve
with depression only152. capacities30. Our recent studies indicate that the cortical
The question remains, however, of why some indi- network responds to an inflammatory attack with a
viduals are more resilient to stressful events than others. TNF-dependent increase in cortical synaptic strength65.
In a mouse model of chronic social defeat stress, BBB This upregulation of cortical neuronal activity presum-
integrity was altered in the nucleus accumbens owing to ably reflects repair mechanisms initially but eventually
downregulation of the tight junction protein claudin 5, causes maladaptive development of anxiety behaviour
thereby allowing influx of IL-6 into the brain paren- despite otherwise complete remission in disability.
chyma153. Importantly, however, loss of claudin 5 was Analysis of higher-order networks in patients who
observed only in stress-susceptible mice — susceptibil- progress rapidly and those who do not could increase
ity to stress was observed as social avoidance behaviour our understanding of brain reserve and identify future
after several days of exposure to social defeat by a larger, therapeutic approaches.
physically aggressive mouse; by contrast, this behaviour
was not seen in mice classified as resilient. Epigenetic New clinical approaches
modulation of IL-6 seems to further promote resilience The idea that the CNS is a site of immune privilege has
in mice, as an inhibitor of IL-6 DNA methylation that shifted with a growing understanding that immune cells
decreased IL-6 production reduced depressive symptoms have multifaceted roles in CNS homeostasis, even in
in a mouse model154. Furthermore, adaptive immune the healthy brain. Mechanistic aspects of the neuroim-
cells are needed for coping with stress. For example, mune interplay have been revealed, including roles for
exposure of mice to acute psychological stress in the peripherally induced adaptive immunity and for cross-
form of predator odour increases T cell trafficking to the talk between neurons and CNS-resident microglia. The
brain155, and T cell-deficient mice displayed more signs involvement of astrocytes has been studied to a lesser
of maladaptation after stress156. In addition, analogous to extent but is becoming a focus of research.

New-onset neuropsychiatric Is there reason for a high

clinical symptoms (cognitive clinical suspicion of an
decline, psychosis, epileptic Do MRI and lumbar NO inflammatory cause (e.g. high NO No evidence of
seizures or transient focal punctures reveal signs serum levels of CRP, IL-6, or immune-mediated
neurological deficit) and of inflammation? IL-2 receptors, and/or high pathology
exclusion of parallel ESR)?
extracerebral infection YES
YES
Is there evidence of an infectious
YES cause (e.g. bacterial meningitis,
Infection
neuroborreliosis, HSV, VZV, CMV,
HIV or opportunistic infections)? Does FDG–PET reveal signs NO
of inflammation and/or
NO does the condition respond
to trial therapy for immune-
Is there evidence of systemic inflammation? mediated disorders?
YES
YES NO
NO
Systemic autoimmune disease Autoimmune
with CNS involvement disease of the CNS Are anti-neuropil-
(e.g. rheumatologic disease, reactive YES Autoimmune encephalitis
sarcoidosis and vasculitis) (paraneoplastic, triggered
autoantibodies by infection or idiopathic)
present?
Is there evidence of YES
autoantibodies?
NO Are anti-aquaporin 4 YES NMOSD or
or anti-MOG MOG-associated
antibodies present? disease
Unilateral cerebral Dissemination in
atrophy and epilepsia time and space and
partialis continua positive OCBs in CSF Autoimmune subtypes
of psychiatric diseases,
(e.g. schizophrenia,
autism, depression or
Rasmussen encephalitis Multiple sclerosis bipolar disorder)

Fig. 3 | A proposed clinical pathway for patients with new-onset neuropsychiatric symptoms. Blue boxes indicate
investigations required and orange boxes indicate diagnoses. CMV, cytomegalovirus; CRP, C-reactive protein; CSF,
cerebrospinal fluid; ESR , erythrocyte sedimentation rate; HSV, herpes simplex virus; MOG, myelin oligodendrocyte
glycoprotein; NMOSD, neuromyelitis optica spectrum disorder ; OCB, oligoclonal bands; VZV, varicella zoster virus.

Nature Reviews | Neurology

Reviews

Box 3 | unresolved questions to direct future research be screened for blood, CSF and imaging markers of
inflammation. Psychiatric symptoms in patients with
• Which pathways mediate the interplay between levels classical neuroinflammatory disorders must be con-
of inflammation and cognitive performance? sidered when making treatment decisions. If there are
• Why do apparently similar immunological mechanisms any signs of active inflammation, immunomodulatory
lead to clinically distinct disorders? treatment should be discussed.
• What are the precise mechanisms of interaction
between immune cells and higher-order networks in Conclusion
neuropsychiatric diseases? How can this crosstalk be For decades, or even centuries, of medical history, neuro­
addressed therapeutically?
psychiatric diseases have been diagnosed and classified
• Are immunological disturbances the cause or on the basis of phenomenological criteria. The findings
consequence of neuropsychiatric diseases?
discussed in this Review provide novel perspectives and
• How can we better identify subgroups of patients who indicate the need for a future classification of neuropsy-
are likely to benefit from immunotherapies?
chiatric disorders on the basis of pathogenic mechanisms,
at least for subgroups of patients. In this context, diag-
Associations of the immune system with brain dis- nostic algorithms should classify patients into inflamma-
eases have been reported for decades, initially in case tory and non-inflammatory subgroups to help reach an
reports but later in epidemiological studies. Studies of individualized treatment decision. For example, future
the immune system in psychiatric disorders have thus understanding of neuropsychiatric diseases might no
far focused on classical markers of inflammation, such longer divide them into descriptive entities, such as MS,
as CRP and erythrocyte sedimentation rate (ESR), TNF neuromyelitis optica spectrum disorder, auto­immune
or IL-6 in serum, as more sensitive or specific markers encephalitis, schizophrenia, ASD and bipolar dis­order,
are often lacking. A level of inflammation that strikes but into pathogenic entities, such as autoimmune
a highly sensitive equilibrium seems to be required antibody-mediated or T cell-mediated brain disorders.
for health. Consequently, any disturbance can cause Although the knowledge of immune involvement
detrimental effects not only on focal neurological in neuropsychiatric disorders has emerged over several
symptoms but also, more importantly, on higher-order years, necessary changes to the clinical management of
network function. patients with these disorders have scarcely been imple-
As common immunological aetiologies of neurologi- mented. Furthermore, many questions remain unsolved
cal and psychiatric diseases are discovered, the traditional (Box 3), such as why apparently similar mechanisms lead
classification of these diseases is increasingly challenged. to clinically distinct disorders and whether immunolog-
In order to develop promising immune-based thera- ical disturbances are always a cause of certain disorders
peutic approaches for psychiatric diseases, we propose or sometimes a consequence. The next steps will require
new diagnostic pathways (Fig. 3). In this proposed algo- combined experience from neurologists and psychiatrists
rithm, individuals who present with new neuropsy- in a new field of immunoneuropsychiatry.
chiatric symptoms, such as cognitive and behavioural
disturbances, epileptic seizures and psychosis, must Published online xx xx xxxx

1. Nimmerjahn, A., Kirchhoff, F. & Helmchen, F. Resting 12. Schwartz, M. & Deczkowska, A. Neurological disease autoimmune encephalomyelitis, thereby
microglial cells are highly dynamic surveillants of as a failure of brain-immune crosstalk: the multiple highlighting a new pathway of neuroimmune
brain parenchyma in vivo. Science 308, 1314–1318 faces of neuroinflammation. Trends Immunol. 37, interplay as a potential therapeutic target.
(2005). 668–679 (2016). 20. Radjavi, A., Smirnov, I. & Kipnis, J. Brain
2. Stevens, B. et al. The classical complement cascade 13. Strominger, I. et al. The choroid plexus functions as a antigen-reactive CD4+ T cells are sufficient to support
mediates CNS synapse elimination. Cell 131, niche for T-cell stimulation within the central nervous learning behavior in mice with limited T cell repertoire.
1164–1178 (2007). system. Front. Immunol. 9, 1066 (2018). Brain Behav. Immun. 35, 58–63 (2014).
3. Paolicelli, R. C. et al. Synaptic pruning by microglia is 14. Filiano, A. J. et al. Unexpected role of interferon- 21. Iliff, J. J. et al. A paravascular pathway facilitates CSF
necessary for normal brain development. Science gamma in regulating neuronal connectivity and social flow through the brain parenchyma and the clearance
333, 1456–1458 (2011). behaviour. Nature 535, 425–429 (2016). of interstitial solutes, including amyloid beta.
4. Sanchez-Alcaniz, J. A. et al. Cxcr7 controls neuronal 15. Derecki, N. C. et al. Regulation of learning and Sci. Transl Med. 4, 147ra111 (2012).
migration by regulating chemokine responsiveness. memory by meningeal immunity: a key role for IL-4. 22. Xie, L. et al. Sleep drives metabolite clearance from
Neuron 69, 77–90 (2011). J. Exp. Med. 207, 1067–1080 (2010). the adult brain. Science 342, 373–377 (2013).
5. Wake, H., Moorhouse, A. J., Jinno, S., Kohsaka, S. 16. Kipnis, J., Cohen, H., Cardon, M., Ziv, Y. & 23. Louveau, A. et al. Structural and functional features
& Nabekura, J. Resting microglia directly monitor the Schwartz, M. T cell deficiency leads to cognitive of central nervous system lymphatic vessels. Nature
functional state of synapses in vivo and determine dysfunction: implications for therapeutic vaccination 523, 337–341 (2015).
the fate of ischemic terminals. J. Neurosci. 29, for schizophrenia and other psychiatric conditions. This article presents the first description of the
3974–3980 (2009). Proc. Natl Acad. Sci. USA 101, 8180–8185 (2004). re-discovery of the meningeal lymphatic system.
6. Pocock, J. M. & Kettenmann, H. Neurotransmitter This study shows that a lack of mature T cells 24. Absinta, M. et al. Human and nonhuman primate
receptors on microglia. Trends Neurosci. 30, causes cognitive and behavioural impairment in meninges harbor lymphatic vessels that can be visualized
527–535 (2007). mice, underlining the role of adaptive immunity for noninvasively by MRI. eLife 6, e29738 (2017).
7. Antonucci, F. et al. Microvesicles released from microglia normal CNS function. 25. Traka, M., Podojil, J. R., McCarthy, D. P., Miller, S. D.
stimulate synaptic activity via enhanced sphingolipid 17. Ziv, Y. et al. Immune cells contribute to the & Popko, B. Oligodendrocyte death results in
metabolism. EMBO J. 31, 1231–1240 (2012). maintenance of neurogenesis and spatial learning immune-mediated CNS demyelination. Nat. Neurosci.
8. Beattie, E. C. et al. Control of synaptic strength by glial abilities in adulthood. Nat. Neurosci. 9, 268–275 19, 65–74 (2016).
TNFα. Science 295, 2282–2285 (2002). (2006). 26. D’Mello, C. & Swain, M. G. Immune-to-brain
9. Gertig, U. & Hanisch, U. K. Microglial diversity by 18. Walsh, J. T. et al. MHCII-independent CD4+ T cells communication pathways in inflammation-associated
responses and responders. Front. Cell Neurosci. 8, protect injured CNS neurons via IL-4. J. Clin. Invest. sickness and depression. Curr. Top. Behav. Neurosci.
101 (2014). 125, 699–714 (2015). 31, 73–94 (2017).
10. Ellwardt, E., Walsh, J. T., Kipnis, J. & Zipp, F. 19. Vogelaar, C. F. et al. Fast direct neuronal signaling 27. Kelley, K. W. et al. Cytokine-induced sickness behavior.
Understanding the role of T cells in CNS homeostasis. via the IL-4 receptor as therapeutic target in Brain Behav. Immun. 17 (Suppl. 1), 112–118 (2003).
Trends Immunol. 37, 154–165 (2016). neuroinflammation. Sci. Transl Med. 10, eaao2304 28. Neilley, L. K., Goodin, D. S., Goodkin, D. E. &
11. Engelhardt, B. & Ransohoff, R. M. Capture, crawl, (2018). Hauser, S. L. Side effect profile of interferon beta-1b
cross: the T cell code to breach the blood-brain This study reveals a beneficial effect of intrathecal in MS: results of an open label trial. Neurology 46,
barriers. Trends Immunol. 33, 579–589 (2012). IL-4 treatment on progression in experimental 552–554 (1996).

www.nature.com/nrneurol
 Reviews

29. Tarr, A. J., Liu, X., Reed, N. S. & Quan, N. Kinetic 52. Zhang, C. J. et al. TLR-stimulated IRAKM activates progression in multiple sclerosis: a systematic review.
characteristics of euflammation: the induction of caspase-8 inflammasome in microglia and promotes Neurotoxicology 61, 189–212 (2017).
controlled inflammation without overt sickness neuroinflammation. J. Clin. Invest. 128, 5399–5412 75. Dutta, R. et al. Demyelination causes synaptic
behavior. Brain Behav. Immun. 42, 96–108 (2014). (2018). alterations in hippocampi from multiple sclerosis
30. Larochelle, C., Uphaus, T., Prat, A. & Zipp, F. 53. Heneka, M. T., McManus, R. M. & Latz, E. Inflammasome patients. Ann. Neurol. 69, 445–454 (2011).
Secondary progression in multiple sclerosis: neuronal signalling in brain function and neurodegenerative 76. Liblau, R. S., Gonzalez-Dunia, D., Wiendl, H. & Zipp, F.
exhaustion or distinct pathology? Trends Neurosci. disease. Nat. Rev. Neurosci. 19, 610–621 (2018). Neurons as targets for T cells in the nervous system.
39, 325–339 (2016). 54. Wendeln, A. C. et al. Innate immune memory in the Trends Neurosci. 36, 315–324 (2013).
31. Franceschi, C. et al. Inflamm-aging. An evolutionary brain shapes neurological disease hallmarks. Nature 77. Kebir, H. et al. Human TH17 lymphocytes promote
perspective on immunosenescence. Ann. NY Acad. Sci. 556, 332–338 (2018). blood-brain barrier disruption and central nervous
908, 244–254 (2000). This study shows that peripheral inflammatory system inflammation. Nat. Med. 13, 1173–1175 (2007).
32. Marrie, R. A. et al. Increased incidence of psychiatric stimuli induce differential epigenetic modulation of 78. Campbell, G. R. et al. Mitochondrial DNA deletions
disorders in immune-mediated inflammatory disease. brain-resident microglia, influencing symptoms in a and neurodegeneration in multiple sclerosis.
J. Psychosom. Res. 101, 17–23 (2017). mouse model of AD; these findings provide a Ann. Neurol. 69, 481–492 (2011).
33. Perry, V. H., Nicoll, J. A. & Holmes, C. Microglia in mechanistic link between inflammation, innate 79. Calabrese, M. et al. Cortical lesion load associates
neurodegenerative disease. Nat. Rev. Neurol. 6, immunity and neuropsychiatric disease. with progression of disability in multiple sclerosis.
193–201 (2010). 55. Lovheim, H., Gilthorpe, J., Adolfsson, R., Nilsson, L. G. Brain 135, 2952–2961 (2012).
34. Lumeng, C. N. & Saltiel, A. R. Inflammatory links & Elgh, F. Reactivated herpes simplex infection 80. Bellmann-Strobl, J. et al. Poor PASAT performance
between obesity and metabolic disease. J. Clin. Invest. increases the risk of Alzheimer’s disease. Alzheimers correlates with MRI contrast enhancement in multiple
121, 2111–2117 (2011). Dement. 11, 593–599 (2015). sclerosis. Neurology 73, 1624–1627 (2009).
35. Stranahan, A. M., Hao, S., Dey, A., Yu, X. & Baban, B. 56. Fulop, T., Itzhaki, R. F., Balin, B. J., Miklossy, J. & 81. Andreassen, O. A. et al. Genetic pleiotropy between
Blood-brain barrier breakdown promotes macrophage Barron, A. E. Role of microbes in the development of multiple sclerosis and schizophrenia but not bipolar
infiltration and cognitive impairment in leptin Alzheimer’s disease: state of the art - an International disorder: differential involvement of immune-related
receptor-deficient mice. J. Cereb. Blood Flow Metab. Symposium presented at the 2017 IAGG Congress in gene loci. Mol. Psychiatry 20, 207–214 (2015).
36, 2108–2121 (2016). San Francisco. Front. Genet. 9, 362 (2018). 82. Charalambous, T. et al. Structural network disruption
36. Guillemot-Legris, O. & Muccioli, G. G. Obesity-induced 57. Wozniak, M. A., Mee, A. P. & Itzhaki, R. F. Herpes simplex markers explain disability in multiple sclerosis.
neuroinflammation: beyond the hypothalamus. Trends virus type 1 DNA is located within Alzheimer’s disease J. Neurol. Neurosurg. Psychiatry 90, 219–226 (2019).
Neurosci. 40, 237–253 (2017). amyloid plaques. J. Pathol. 217, 131–138 (2009). 83. Fleischer, V. et al. Graph theoretical framework of
37. Benros, M. E. et al. Autoimmune diseases and severe 58. O’Connell, D. & Liang, C. Autophagy interaction with brain networks in multiple sclerosis: a review of
infections as risk factors for schizophrenia: a 30-year herpes simplex virus type-1 infection. Autophagy 12, concepts. Neuroscience https://doi.org/10.1016/
population-based register study. Am. J. Psychiatry 451–459 (2016). j.neuroscience.2017.10.033 (2017).
168, 1303–1310 (2011). 59. Martin, C. et al. Inflammatory and neurodegeneration 84. Dong, D., Wang, Y., Chang, X., Luo, C. & Yao, D.
This large Danish longitudinal register shows that markers during asymptomatic HSV-1 reactivation. Dysfunction of large-scale brain networks in
both autoimmune disease and prior hospitalization J. Alzheimers Dis. 39, 849–859 (2014). schizophrenia: a meta-analysis of resting-state
for infection increase the risk of schizophrenia on 60. Brown, G. C. & Neher, J. J. Inflammatory functional connectivity. Schizophr. Bull. 44, 168–181
an epidemiological level. neurodegeneration and mechanisms of microglial killing (2018).
38. Khandaker, G. M. et al. Inflammation and immunity in of neurons. Mol. Neurobiol. 41, 242–247 (2010). 85. Kaiser, R. H., Andrews-Hanna, J. R., Wager, T. D. &
schizophrenia: implications for pathophysiology and 61. Armangue, T. et al. Frequency, symptoms, risk Pizzagalli, D. A. Large-scale network dysfunction in
treatment. Lancet Psychiatry 2, 258–270 (2015). factors, and outcomes of autoimmune encephalitis major depressive disorder: a meta-analysis of
39. Canetta, S. et al. Elevated maternal C-reactive protein after herpes simplex encephalitis: a prospective resting-state functional connectivity. JAMA Psychiatry
and increased risk of schizophrenia in a national birth observational study and retrospective analysis. 72, 603–611 (2015).
cohort. Am. J. Psychiatry 171, 960–968 (2014). Lancet Neurol. 17, 760–772 (2018). 86. Armangue, T. et al. Autoimmune post-herpes simplex
40. Atladottir, H. O. et al. Maternal infection requiring This combined prospective observational study and encephalitis of adults and teenagers. Neurology 85,
hospitalization during pregnancy and autism spectrum retrospective analysis characterizes the frequency 1736–1743 (2015).
disorders. J. Autism Dev. Disord. 40, 1423–1430 and the natural course of autoimmune encephalitis 87. Bien, C. G. et al. Immunopathology of
(2010). after herpes simplex encephalitis. autoantibody-associated encephalitides: clues for
41. Orlovska, S. et al. Association of streptococcal throat 62. Carta, M. G. et al. The risk of bipolar disorders in multiple pathogenesis. Brain 135, 1622–1638 (2012).
infection with mental disorders: testing key aspects of sclerosis. J. Affect. Disord. 155, 255–260 (2014). 88. Petit-Pedrol, M. et al. Encephalitis with refractory
the PANDAS hypothesis in a nationwide study. JAMA 63. Feinstein, A., Magalhaes, S., Richard, J. F., Audet, B. seizures, status epilepticus, and antibodies to the
Psychiatry 74, 740–746 (2017). & Moore, C. The link between multiple sclerosis and GABAA receptor: a case series, characterisation of
42. Zomorrodi, A. & Wald, E. R. Sydenham’s chorea in depression. Nat. Rev. Neurol. 10, 507–517 (2014). the antigen, and analysis of the effects of antibodies.
western Pennsylvania. Pediatrics 117, e675–e679 64. Patti, F. Treatment of cognitive impairment in patients Lancet Neurol. 13, 276–286 (2014).
(2006). with multiple sclerosis. Expert Opin. Investig. Drugs 89. Haselmann, H. et al. Human autoantibodies against
43. Swedo, S. E. et al. Pediatric autoimmune neuropsychiatric 21, 1679–1699 (2012). the AMPA receptor subunit GluA2 induce receptor
disorders associated with streptococcal infections: clinical 65. Ellwardt, E. et al. Maladaptive cortical hyperactivity reorganization and memory dysfunction. Neuron 100,
description of the first 50 cases. Am. J. Psychiatry 155, upon recovery from experimental autoimmune 91–105 (2018).
264–271 (1998). encephalomyelitis. Nat. Neurosci. 21, 1392–1403 90. Dalmau, J., Lancaster, E., Martinez-Hernandez, E.,
44. Bronze, M. S. & Dale, J. B. Epitopes of streptococcal (2018). Rosenfeld, M. R. & Balice-Gordon, R. Clinical
M proteins that evoke antibodies that cross-react with This study identifies the emergence of cortical experience and laboratory investigations in patients
human brain. J. Immunol. 151, 2820–2828 (1993). network hyperexcitability and elevated anxiety in with anti-NMDAR encephalitis. Lancet Neurol. 10,
45. Singer, H. S., Gause, C., Morris, C. & Lopez, P. Serial remission after neuroinflammatory attack to the 63–74 (2011).
immune markers do not correlate with clinical brain; this network instability with anxiety 91. Hughes, E. G. et al. Cellular and synaptic mechanisms
exacerbations in pediatric autoimmune neuropsychiatric behaviour, also known in patients with MS, of anti-NMDA receptor encephalitis. J. Neurosci. 30,
disorders associated with streptococcal infections. represents the early stages of neurodegeneration. 5866–5875 (2010).
Pediatrics 121, 1198–1205 (2008). 66. Pitteri, M., Romualdi, C., Magliozzi, R., Monaco, S. & 92. Haggerty, G. C., Forney, R. B. & Johnson, J. M.
46. US National Library of Medicine. ClinicalTrials.gov Calabrese, M. Cognitive impairment predicts disability The effect of a single administration of phencyclidine
https://clinicaltrials.gov/ct2/show/NCT02889016 (2016). progression and cortical thinning in MS: an 8-year on behavior in the rat over a 21-day period.
47. Kapur, N. et al. Herpes simplex encephalitis: long term study. Mult. Scler. 23, 848–854 (2017). Toxicol. Appl. Pharmacol. 75, 444–453 (1984).
magnetic resonance imaging and neuropsychological 67. Potagas, C. et al. Cognitive impairment in different MS 93. DeGiorgio, L. A. et al. A subset of lupus anti-DNA
profile. J. Neurol. Neurosurg. Psychiatry 57, subtypes and clinically isolated syndromes. J. Neurol. antibodies cross-reacts with the NR2 glutamate
1334–1342 (1994). Sci. 267, 100–106 (2008). receptor in systemic lupus erythematosus. Nat. Med.
48. Almanzar, G. et al. Long-term cytomegalovirus infection 68. Ruano, L. et al. Age and disability drive cognitive 7, 1189–1193 (2001).
leads to significant changes in the composition of the impairment in multiple sclerosis across disease 94. Nestor, J. et al. Lupus antibodies induce behavioral
CD8+ T cell repertoire, which may be the basis for an subtypes. Mult. Scler. 23, 1258–1267 (2017). changes mediated by microglia and blocked by ACE
imbalance in the cytokine production profile in elderly 69. Di Filippo, M., Portaccio, E., Mancini, A. & Calabresi, P. inhibitors. J. Exp. Med. 215, 2554–2566 (2018).
persons. J. Virol. 79, 3675–3683 (2005). Multiple sclerosis and cognition: synaptic failure and 95. Hammer, C. et al. Neuropsychiatric disease relevance of
49. Katan, M. et al. Infectious burden and cognitive network dysfunction. Nat. Rev. Neurosci. 19, circulating anti-NMDA receptor autoantibodies depends
function: the Northern Manhattan Study. Neurology 599–609 (2018). on blood-brain barrier integrity. Mol. Psychiatry 19,
80, 1209–1215 (2013). 70. Amato, M. P. et al. Benign multiple sclerosis: cognitive, 1143–1149 (2014).
This cohort study shows that infectious burden psychological and social aspects in a clinical cohort. 96. Dahm, L. et al. Seroprevalence of autoantibodies
measured as serological exposure to common J. Neurol. 253, 1054–1059 (2006). against brain antigens in health and disease.
pathogens was associated with cognitive impairment 71. Nylander, A. & Hafler, D. A. Multiple sclerosis. J. Clin. Ann. Neurol. 76, 82–94 (2014).
independent of cardiovascular risk profile. Invest. 122, 1180–1188 (2012). 97. Jezequel, J. et al. Dynamic disorganization of synaptic
50. Hamdani, N. et al. Effects of cumulative herpesviridae 72. Viglietta, V., Baecher-Allan, C., Weiner, H. L. & NMDA receptors triggered by autoantibodies from
and Toxoplasma gondii infections on cognitive function Hafler, D. A. Loss of functional suppression by psychotic patients. Nat. Commun. 8, 1791 (2017).
in healthy, bipolar, and schizophrenia subjects. J. Clin. CD4+CD25+ regulatory T cells in patients with multiple This study reveals differential effects of
Psychiatry 78, e18–e27 (2017). sclerosis. J. Exp. Med. 199, 971–979 (2004). autoantibodies against the glutamate NMDA
51. Kamminga, J., Cysique, L. A., Lu, G., Batchelor, J. 73. Hauser, S. L. et al. Ocrelizumab versus interferon receptor on synaptic transmission and long-term
& Brew, B. J. Validity of cognitive screens for beta-1a in relapsing multiple sclerosis. N. Engl. J. potentiation in patients with psychosis versus
HIV-associated neurocognitive disorder: a systematic Med. 376, 221–234 (2017). healthy controls, providing a mechanistic
review and an informed screen selection guide. 74. McKay, K. A., Jahanfar, S., Duggan, T., Tkachuk, S. & framework for different clinical outcomes despite
Curr. HIV/AIDS Rep. 10, 342–355 (2013). Tremlett, H. Factors associated with onset, relapses or similar autoantibodies.

Nature Reviews | Neurology

Reviews

98. Kantrowitz, J. & Javitt, D. C. Glutamatergic transmission in their first-degree relatives. J. Pediatr. Gastroenterol. of minocycline combined with several glutamate
in schizophrenia: from basic research to clinical practice. Nutr. 51, 418–424 (2010). antagonists. Prog. Neuropsychopharmacol.
Curr. Opin. Psychiatry 25, 96–102 (2012). 120. Lombardo, M. V. et al. Maternal immune activation Biol. Psychiatry 32, 380–386 (2008).
99. Bergink, V., Gibney, S. M. & Drexhage, H. A. dysregulation of the fetal brain transcriptome and 144. Rosenblat, J. D. & McIntyre, R. S. Efficacy and
Autoimmunity, inflammation, and psychosis: a search relevance to the pathophysiology of autism spectrum tolerability of minocycline for depression: a systematic
for peripheral markers. Biol. Psychiatry 75, 324–331 disorder. Mol. Psychiatry 23, 1001–1013 (2018). review and meta-analysis of clinical trials. J. Affect.
(2014). 121. Zimmerman, A. W. et al. Maternal antibrain antibodies Disord. 227, 219–225 (2018).
100. Cullen, A. E. et al. Associations between non- in autism. Brain Behav. Immun. 21, 351–357 (2007). This systematic review and meta-analysis provides
neurological autoimmune disorders and psychosis: 122. Braunschweig, D. et al. Autism-specific maternal a proof of concept for the antidepressant effects of
a meta-analysis. Biol. Psychiatry 85, 35–48 (2018). autoantibodies recognize critical proteins in minocycline.
101. Siegmann, E. M. et al. Association of depression developing brain. Transl Psychiatry 3, e277 (2013). 145. Ghaleiha, A. et al. Minocycline as adjunctive treatment
and anxiety disorders with autoimmune thyroiditis: 123. Bauman, M. D. et al. Maternal antibodies from to risperidone in children with autistic disorder:
a systematic review and meta-analysis. JAMA Psychiatry mothers of children with autism alter brain growth and a randomized, double-blind placebo-controlled trial.
75, 577–584 (2018). social behavior development in the rhesus monkey. J. Child Adolesc. Psychopharmacol. 26, 784–791
102. Sommer, I. E. et al. Severe chronic psychosis after Transl Psychiatry 3, e278 (2013). (2016).
allogeneic SCT from a schizophrenic sibling. Bone 124. Brimberg, L. et al. Caspr2-reactive antibody cloned 146. Pardo, C. A. et al. A pilot open-label trial of minocycline
Marrow Transplant. 50, 153–154 (2015). from a mother of an ASD child mediates an ASD-like in patients with autism and regressive features.
103. Dahan, S. et al. The relationship between serum phenotype in mice. Mol. Psychiatry 21, 1663–1671 J. Neurodev. Disord. 5, 9 (2013).
cytokine levels and degree of psychosis in patients with (2016). 147. Raison, C. L. et al. A randomized controlled trial of the
schizophrenia. Psychiatry Res. 268, 467–472 (2018). This study describes the isolation and tumor necrosis factor antagonist infliximab for
104. Haapakoski, R., Mathieu, J., Ebmeier, K. P., Alenius, H. characterization of monoclonal brain-reactive treatment-resistant depression: the role of baseline
& Kivimaki, M. Cumulative meta-analysis of antibodies from a mother of a child with ASD. inflammatory biomarkers. JAMA Psychiatry 70,
interleukins 6 and 1β, tumour necrosis factor α and 125. Cross-Disorder Group of the Psychiatric Genomics 31–41 (2013).
C-reactive protein in patients with major depressive Consortium. Identification of risk loci with shared effects 148. Kalman, B. Autoimmune encephalitides: a broadening
disorder. Brain Behav. Immun. 49, 206–215 (2015). on five major psychiatric disorders: a genome-wide field of treatable conditions. Neurologist 22, 1–13
105. Black, C. & Miller, B. J. Meta-analysis of cytokines and analysis. Lancet 381, 1371–1379 (2013). (2017).
chemokines in suicidality: distinguishing suicidal 126. Trowsdale, J. & Knight, J. C. Major histocompatibility 149. Howes, O. D. & McCutcheon, R. Inflammation and the
versus nonsuicidal patients. Biol. Psychiatry 78, complex genomics and human disease. Annu. Rev. neural diathesis-stress hypothesis of schizophrenia:
28–37 (2015). Genomics Hum. Genet. 14, 301–323 (2013). a reconceptualization. Transl Psychiatry 7, e1024 (2017).
106. Al-Ayadhi, L. Y. & Mostafa, G. A. Elevated serum levels 127. Dendrou, C. A., Petersen, J., Rossjohn, J. & Fugger, L. 150. Fonken, L. K., Frank, M. G., Gaudet, A. D. & Maier, S. F.
of interleukin-17A in children with autism. HLA variation and disease. Nat. Rev. Immunol. 18, Stress and aging act through common mechanisms to
J. Neuroinflamm. 9, 158 (2012). 325–339 (2018). elicit neuroinflammatory priming. Brain Behav. Immun.
107. Passos, I. C. et al. Inflammatory markers in 128. Sekar, A. et al. Schizophrenia risk from complex 73, 133–148 (2018).
post-traumatic stress disorder: a systematic review, variation of complement component 4. Nature 530, 151. Baumeister, D., Akhtar, R., Ciufolini, S., Pariante, C. M.
meta-analysis, and meta-regression. Lancet Psychiatry 177–183 (2016). & Mondelli, V. Childhood trauma and adulthood
2, 1002–1012 (2015). 129. Prasad, K. M. et al. Neuropil contraction in relation to inflammation: a meta-analysis of peripheral C-reactive
108. Bulzacka, E. et al. Chronic peripheral inflammation is Complement C4 gene copy numbers in independent protein, interleukin-6 and tumour necrosis factor-α.
associated with cognitive impairment in schizophrenia: cohorts of adolescent-onset and young adult-onset Mol. Psychiatry 21, 642–649 (2016).
results from the multicentric FACE-SZ dataset. schizophrenia patients-a pilot study. Transl Psychiatry 152. Danese, A. et al. Elevated inflammation levels in
Schizophr. Bull. 42, 1290–1302 (2016). 8, 134 (2018). depressed adults with a history of childhood
109. Rademakers, R. et al. Mutations in the colony 130. Kury, P. et al. Human endogenous retroviruses in maltreatment. Arch. Gen. Psychiatry 65, 409–415
stimulating factor 1 receptor (CSF1R) gene cause neurological diseases. Trends Mol. Med. 24, 379–394 (2008).
hereditary diffuse leukoencephalopathy with (2018). 153. Menard, C. et al. Social stress induces neurovascular
spheroids. Nat. Genet. 44, 200–205 (2011). 131. Perron, H. et al. Molecular characteristics of human pathology promoting depression. Nat. Neurosci. 20,
110. Tong, L. et al. Microglia loss contributes to the endogenous retrovirus type-W in schizophrenia and 1752–1760 (2017).
development of major depression induced by different bipolar disorder. Transl Psychiatry 2, e201 (2012). This study reveals mechanistic links between
types of chronic stresses. Neurochem. Res. 42, 132. Melamed, I. R., Heffron, M., Testori, A. & Lipe, K. psychosocial stress and depression via disruption
2698–2711 (2017). A pilot study of high-dose intravenous immunoglobulin of BBB integrity by downregulation of claudin 5
111. Bloomfield, P. S. et al. Microglial activity in people 5% for autism: impact on autism spectrum and in mice.
at ultra high risk of psychosis and in schizophrenia: markers of neuroinflammation. Autism Res. 11, 154. Wang, J. et al. Epigenetic modulation of inflammation
an [11C]PBR28 PET brain imaging study. Am. J. 421–433 (2018). and synaptic plasticity promotes resilience against
Psychiatry 173, 44–52 (2016). 133. Connery, K. et al. Intravenous immunoglobulin for the stress in mice. Nat. Commun. 9, 477 (2018).
This PET imaging study shows that patients with treatment of autoimmune encephalopathy in children 155. Lewitus, G. M., Cohen, H. & Schwartz, M. Reducing
schizophrenia and individuals with a high risk of with autism. Transl Psychiatry 8, 148 (2018). post-traumatic anxiety by immunization. Brain Behav.
psychosis exhibit increased microglial activity, This article and that by Melamed et al. (2018) Immun. 22, 1108–1114 (2008).
indicating a connection between neuroinflammation describe pilot studies that provided the first 156. Cohen, H. et al. Maladaptation to mental stress
and the risk of psychosis. evidence for a beneficial effect of intravenous mitigated by the adaptive immune system via
112. Wachholz, S. et al. Microglia activation is associated immunoglobulin for the treatment of children with depletion of naturally occurring regulatory CD4+CD25+
with IFN-alpha induced depressive-like behavior. ASD and signs of inflammation. cells. J. Neurobiol. 66, 552–563 (2006).
Brain Behav. Immun. 55, 105–113 (2016). 134. Swedo, S. E., Frankovich, J. & Murphy, T. K. Overview 157. Lewitus, G. M. et al. Vaccination as a novel approach
113. Norden, D. M., Muccigrosso, M. M. & Godbout, J. P. of treatment of pediatric acute-onset neuropsychiatric for treating depressive behavior. Biol. Psychiatry 65,
Microglial priming and enhanced reactivity to syndrome. J. Child Adolesc. Psychopharmacol. 27, 283–288 (2009).
secondary insult in aging, and traumatic CNS injury, 562–565 (2017).
and neurodegenerative disease. Neuropharmacology 135. US National Library of Medicine. ClinicalTrials.gov Acknowledgements
96, 29–41 (2015). https://clinicaltrials.gov/ct2/show/NCT03093064 (2017). The authors thank C. Ernest, R. Gilchrist and S. Jennrich (all
114. Smith, S. E., Li, J., Garbett, K., Mirnics, K. & 136. US National Library of Medicine. ClinicalTrials.gov at the Department of Neurology, University Medical Center
Patterson, P. H. Maternal immune activation alters https://clinicaltrials.gov/ct2/show/NCT02874573 (2018). of the Johannes Gutenberg University Mainz, Mainz,
fetal brain development through interleukin-6. 137. US National Library of Medicine. ClinicalTrials.gov Germany) for proofreading the manuscript. The authors
J. Neurosci. 27, 10695–10702 (2007). https://clinicaltrials.gov/ct2/show/NCT02034474 (2018). acknowledge financial support for the underlying original
115. Choi, G. B. et al. The maternal interleukin-17a 138. Miyaoka, T. et al. Remission of psychosis in treatment- work from the German Research Foundation (DFG; CRC-TR
pathway in mice promotes autism-like phenotypes in resistant schizophrenia following bone marrow 128, CRC 1080 and CRC 1292 to F.Z.) and from the Agence
offspring. Science 351, 933–939 (2016). transplantation: a case report. Front. Psychiatry 8, Nationale de la Recherche (ANR; I-GIVE Samenta 2013;
The results of this study indicate that the 174 (2017). 13-SAMA-0004-01), the Institut National de la Recherche
development of ASD-like phenotypes in offspring in 139. Dawson, G. et al. Autologous cord blood infusions are Médicale and Fondation FondaMental (to M.L.). The views
the murine model of maternal immune activation is safe and feasible in young children with autism spectrum expressed are those of the authors.
mediated by TH17 cells. disorder: results of a single-center phase I open-label
116. Shin Yim, Y. et al. Reversing behavioural abnormalities trial. Stem Cells Transl Med. 6, 1332–1339 (2017). Author contributions
in mice exposed to maternal inflammation. Nature 140. Hannestad, J., DellaGioia, N. & Bloch, M. The effect of All authors conceived the article, performed literature
549, 482–487 (2017). antidepressant medication treatment on serum levels searches, wrote the article and reviewed and edited the
117. Kim, S. et al. Maternal gut bacteria promote of inflammatory cytokines: a meta-analysis. manuscript. K.P. designed the display items.
neurodevelopmental abnormalities in mouse offspring. Neuropsychopharmacology 36, 2452–2459 (2011).
Nature 549, 528–532 (2017). 141. Platten, M. et al. Treatment of autoimmune Competing interests
This study shows that the production of IL-17 in the neuroinflammation with a synthetic tryptophan The authors declare no competing interests.
maternal immune activation model in mice depends metabolite. Science 310, 850–855 (2005).
on the composition of maternal intestinal bacteria, 142. Kohler, O. et al. Effect of anti-inflammatory treatment Publisher’s note
underlining the role of the gut–immune–brain axis. on depression, depressive symptoms, and adverse Springer Nature remains neutral with regard to jurisdictional
118. Cekanaviciute, E. et al. Gut bacteria from multiple effects: a systematic review and meta-analysis of claims in published maps and institutional affiliations.
sclerosis patients modulate human T cells and randomized clinical trials. JAMA Psychiatry 71,
exacerbate symptoms in mouse models. Proc. Natl 1381–1391 (2014). Reviewer information
Acad. Sci. USA 114, 10713–10718 (2017). 143. Molina-Hernandez, M., Tellez-Alcantara, N. P., Nature Reviews Neurology thanks D. Agalliu and other anon-
119. de Magistris, L. et al. Alterations of the intestinal Perez-Garcia, J., Olivera-Lopez, J. I. & ymous reviewer(s) for their contribution to the peer review of
barrier in patients with autism spectrum disorders and Jaramillo-Jaimes, M. T. Antidepressant-like actions this work.

www.nature.com/nrneurol