Modelling Viral Encephalitis Caused by Herpes Simplex Virus 1 Infection in Cerebral Organoids

nature microbiology
Article https://doi.org/10.1038/s41564-023-01405-y
Modelling viral encephalitis caused

by herpes simplex virus 1 infection in
cerebral organoids
Received: 22 July 2022 Agnieszka Rybak-Wolf 1,17 , Emanuel Wyler 2,17,

Tancredi Massimo Pentimalli3,4,17, Ivano Legnini 3,13,17, Anna Oliveras Martinez5,
Accepted: 10 May 2023
Petar Glažar3, Anna Loewa 1,14, Seung Joon Kim 3,15, Benedikt B. Kaufer 6,
Published online: 22 June 2023 Andrew Woehler 5,16, Markus Landthaler 2,7 & Nikolaus Rajewsky 3,8,9,10,11,12
Check for updates

Herpes simplex encephalitis is a life-threatening disease of the central nervous
system caused by herpes simplex viruses (HSVs). Following standard of care
with antiviral acyclovir treatment, most patients still experience various
neurological sequelae. Here we characterize HSV-1 infection of human brain
organoids by combining single-cell RNA sequencing, electrophysiology and
immunostaining. We observed strong perturbations of tissue integrity,
neuronal function and cellular transcriptomes. Under acyclovir treatment
viral replication was stopped, but did not prevent HSV-1-driven defects such
as damage of neuronal processes and neuroepithelium. Unbiased analysis
of pathways deregulated upon infection revealed tumour necrosis factor
activation as a potential causal factor. Combination of anti-inflammatory drugs
such as necrostatin-1 or bardoxolone methyl with antiviral treatment prevented
the damages caused by infection, indicating that tuning the inflammatory
response in acute infection may improve current therapeutic strategies.
Herpes simplex virus type 1 (HSV-1) is a common pathogen However, stimuli such as stress signals and weakened immunity can
affecting a large part of the human population worldwide1. Follow- cause the re-activation of HSV-1 from sensory neurons at any time1
ing primary replication in epithelial cells, HSV-1 establishes latency in Herpes simplex encephalitis (HSE), caused by re-activation of viral
the trigeminal ganglia and latent infections can persist over decades. replication or by new lytic infection in the brain, is life threatening in
1
Organoid Platform, Berlin Institute for Medical Systems Biology (BIMSB), Max Delbrück Center for Molecular Medicine in the Helmholtz Association
(MDC), Berlin, Germany. 2Laboratory for RNA Biology, Berlin Institute for Medical Systems Biology (BIMSB), Max Delbrück Center for Molecular Medicine
in the Helmholtz Association (MDC), Berlin, Germany. 3Laboratory for Systems Biology of Gene Regulatory Elements, Berlin Institute for Medical Systems
Biology (BIMSB), Max Delbrück Center for Molecular Medicine in the Helmholtz Association (MDC), Berlin, Germany. 4Charité—Universitätsmedizin
Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin School of Integrative Oncology (BSIO), Berlin, Germany.
5
System Biology Imaging Platform, Berlin Institute for Medical Systems Biology (BIMSB), Max Delbrück Center for Molecular Medicine in the Helmholtz
Association (MDC), Berlin, Germany. 6Institut für Virologie, Freie Universität Berlin, Berlin, Germany. 7Institut für Biologie, Humboldt Universität zu Berlin,
Berlin, Germany. 8Charité—Universitätsmedizin, Berlin, Germany. 9German Center for Cardiovascular Research (DZHK), Site Berlin, Berlin, Germany.
10
NeuroCure Cluster of Excellence, Berlin, Germany. 11German Cancer Consortium (DKTK), Berlin, Germany. 12National Center for Tumor Diseases (NCT),
Site Berlin, Berlin, Germany. 13Present address: Centre for Genomics, Functional Genomics Programme, Human Technopole, Milan, Italy. 14Present
address: Department of Infectious Diseases and Respiratory Medicine, Charité—Universitätsmedizin Berlin, corporate member of Freie Universität Berlin
and Humboldt Universität zu Berlin, Berlin, Germany. 15Present address: Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridge,
UK. 16Present address: Howard Hughes Medical Institute, Janelia Research Campus, Ashburn, VA, USA. 17These authors contributed equally: Agnieszka
Rybak-Wolf, Emanuel Wyler, Tancredi Massimo Pentimalli, Ivano Legnini. e-mail: agnieszka.rybak@mdc-berlin.de; rajewsky@mdc-berlin.de
Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1252

absence of treatment2,3. HSV-1 is the most common cause of HSE, which For an unbiased and comprehensive cell type characterization,
accounts for 5–15% of infectious encephalitis in children and adults4. we performed scRNA-seq analysis of 60-day-old organoids (Fig. 1d and
Antiviral therapies strongly reduce the mortality rate of patients Extended Data Fig. 1c,d). Combined highly reproducible single-cell
with HSE3. However, large proportions of recovered patients show transcriptomes from two biological replicates from two lines (shown
moderate to severe neurological sequelae2,3. One possible explanation in Extended Data Fig. 1c) showed 16 transcriptionally distinct clusters
of brain injury is neuroinflammation3. (Fig. 1d). These clusters were defined by cell-type-specific marker
The current understanding of the mechanisms of HSV-1 infection, genes (examples in Extended Data Fig. 1d) in accordance with previ-
latency and re-activation in the central nervous system are based either ous annotations21,22 and published atlases (Supplementary Fig. 1a21).
on animal models or on in vitro differentiated neurons. These models Subsequently, we grouped them into progenitors, intermediate pro-
cannot fully account for human host–pathogen interaction specificity genitors, neurons, astroglia, choroid plexus/cortical hem, retinal and
or cellular and functional diversity in neural tissues, respectively5,6. mural cells (Fig. 1d).
In recent years, human stem cell-derived brain organoids emerged
as models that capture several important aspects of human brain HSV-1 disrupts neuronal integrity and functionality
development, tissue architecture and physiology7–9. Brain organoids To balance between cellular complexity and increasing necrosis
have been used to study Zika virus neurotropism10–13, Japanese encepha- upon prolonged culturing23, we used 60-day-old organoids for all the
litis virus14 and severe acute respiratory syndrome coronavirus 2 infec- infection experiments. We infected them with the green fluorescent
tion15. In the context of HSV-1, acute and latent-like infection states have protein (GFP)-expressing HSV-1 strain 17syn+ (ref. 24) and investigated
been established in 3D neuronal culture, which demonstrated molecular, cellular and physiological consequences of the infection
that HSV-1 can cause dramatic changes in neuronal morphology, pro- (Fig. 2a). We analysed GFP mRNA expression using spatial transcrip-
cesses and syncytium formation5,16–18. Furthermore, HSV-1 impairs tomics data (10x Genomics Visium platform) and GFP protein using
identity of neuroepithelial and neuronal progenitor cells during immunofluorescence and could show that the virus spreads into the
early organoid development16. Given the complexity of these models, organoid outer layers at 1 day post infection (dpi) and expands into
unbiased and high-throughput methods become necessary to get a deeper inner layers at 3 dpi (Fig. 2b and Extended Data Fig. 2a). Trans
complete picture of pathological changes and identify better thera- criptome changes of infected organoids were assessed using bulk and
peutic interventions. scRNA-seq at 1 and 3 dpi, and neuronal activity by calcium imaging
In this Article, we applied single-cell RNA sequencing (scRNA-seq), after HSV-1 infection. Immunostainings and western blotting served
neuronal activity measurements and imaging to capture the mole as validation tools.
cular and functional consequences of HSV-1 infection in human brain Previous reports showed that lytic HSV-1 neuronal infection
organoids. Real-time calcium imaging showed that HSV-1 infection caused synaptic dysfunction and disassembly of dendritic spines
massively decreases neuronal activity. An unbiased, exploratory gene and upregulation of activity-regulated cytoskeleton-associated
set enrichment analysis (GSEA) identified tumour necrosis factor protein25,26. In our system, we observed massive downregulation of
(TNF) signalling as the most highly and broadly induced pathway upon synaptic transmission genes (Extended Data Fig. 2b). Expression
infection. Its activation was not stopped by antiviral acyclovir (ACV) changes for synapsin-1 (SYN1) and postsynaptic density scaffolding
treatment, although this significantly reduced viral load in all cell protein 1 (HOMER1) were also confirmed at the p rotein level using
types. Only co-treatment with ACV and anti-inflammatory drugs such immunostaining (Extended Data Fig. 2c) and western blot analysis
as necrostatin-1 (NEC-1) or bardoxolone methyl (CDDO-Me) reduced (Extended Data Fig. 2d).
TNF pathway activation, prevented damages of neuronal processes, To analyse functional consequences of the infection, we measured
and preserved neuroepithelial integrity. Together, our data indicate intracellular calcium dynamics of spontaneous action potentials27.
that brain organoids serve as a powerful tool to study neurotropic/ After plating the organoids, neurons (TUJ1+ cells) showed active
neuroinvasive viruses and to model future therapeutic strategies as migration and neurite extension towards the glass dish bottom
exemplified here by identification of the synergistic effect displayed (Extended Data Fig. 2e). Calcium imaging analysis indicated a high
by a combinatorial antiviral/anti-inflammatory treatment. abundance of spontaneously active neurons in all organoids (day 0).
Uninfected organoids had a high number of active neurites (Fig. 2c
Results and Supplementary Video 1), which showed coordinated activity
Cerebral organoids reflect complex 3D neuronal tissue (Fig. 2c, right). In contrast, the robust initial spontaneous calcium
We generated cerebral brain organoids from two genetically distinct activity decreased after 24 h upon infection and was almost completely
induced pluripotent stem (iPS) cells lines, according to an optimized abolished after 48 h (Fig. 2d, Extended Data Fig. 2f and Supplemen-
version of the Lancaster et al. protocol7 (Fig. 1a). These organoids had tary Video 2). The number of neurites was clearly reduced (Extended
predominantly dorsal forebrain region specification and contained Data Fig. 2e), and the remaining ones were silent or showed limited
ventricle-like structures formed by SOX2+PAX6+ neural progenitors and uncoordinated activity (Fig. 2d). Synaptic dysfunction is prob-
(apical radial glia), around ZO-1+ (tight junction) ring structures. With ably related to the strong downregulation of stathmin-2 (STMN2) a
ongoing differentiation, EOMES+ intermediate progenitors and outer microtubule regulator important for axonal outgrowth and regenera-
radial glia (oRG, HOPX, PEA15 and MOXD1) emerge radially therefrom. tion, and synaptic signalling28,29. Notably, STMN2 expression strongly
Finally, an early cortical-like structure appears at later timepoints, declined after HSV-1 infection on protein level and RNA level (Fig. 2e and
marked by MAP2+TUJ1+ neurons. At 60 days of differentiation, three Supplementary Data Fig. 2a).
main neuronal subtypes are observable, namely immature deep-layer HSV-1 infection has been shown to impair neuroepithelial identity
projection neurons (NEUROD2, NEUROD6 and TBR1), deep-layer cortico during the early stage of organoid development16. We tested for this
fugal projection neurons (BCL11B), callosal projection neurons (SATB2) by immunostaining of tight junction protein ZO-1, which is typically
(Fig. 1b,c and Extended Data Fig. 1a,b), and GFAP astrocytes (Fig. 1c localized at the apical side of progenitors in neural rosettes16,30 (Figs. 1c
and Extended Data Fig. 1b,d). Our cerebral organoids thus contain the and 2f). Notably, the ring structures of tight junctions were gradually
major cell types marking early human brain development19. Although lost with the progress of HSV-1-infection and ZO-1 protein was markedly
organoids are not a direct equivalent of the adult temporal and frontal overexpressed and mislocalized (Fig. 2f).
lobes, which are commonly affected in HSE, they represent the most Lastly, we observed an induction of antisense transcription, as
complex human brain-like tissue system for studying viral infections previously reported in primary fibroblasts31 (Supplementary Data
of central nervous system in vitro20. Fig. 2b, two exemplary genes shown in Supplementary Data Fig. 2c).

a b EOMES TBR1
Merge
BCL11B SOX2
TUBB3 PAX6
iPS cell Day 0 Day 5 Day 11 Day 30
c d
Retinal
Immature cortical neurons 1

TBR1 SATB2 TBR1/SATB2 DAPI Immature cortical neurons 2
Choroid/hem
Mature cortical neurons
Hindbrain neurons 1
Neurons Intermediate Thalamic neurons
progenitor Hindbrain neurons 2
SOX2 TUJ1 Cajal retzius neurons
SOX2/TUJ1/DAPI
Radial glia proliferating
Progenitors
Radial glia
Intermediate progenitors
Retinal progenitors
GFAP MAP2 MAP2/GFAP/DAPI Astroglia
Retinal pigmented cells
Progenitors
Mural cells
Cortical hem/choroid plexus
Astroglia
UMAP2
Mural
SOX2 ZO-1 ZO-1/SOX2/DAPI

UMAP1
Fig. 1 | Cerebral organoids development at single-cell and spatial resolution. cerebral brain organoids showing expression of TBR1 (deeper-layer cortical
a, Schematic overview of cerebral organoids generation using iPS cells. neurons), SATB2 (upper-layer cortical neurons), SOX2 (progenitors), TUJ1/
Bottom: representative images of organoids at different developmental MAP2 (neurons), GFAP (astroglia) and ZO-1 (neuroepithelial junctions). Scale
stages. b, Representative images from the Resolve Biosciences’ Molecular bar, 500 µm. Right: merged images co-stained with nuclear marker DAPI (blue).
Cartography analysis of 60-day-old organoid from iPS cell line 1 for markers Representative image of three experiments from two independent iPS cell lines
EOMES (intermediate progenitors), SOX2/PAX2 (progenitors), TBR1 (deeper- (n = 3 independent organoids/iPS cell line/experiment). d, UMAP plot of scRNA-
layer cortical neurons), BCL11B (upper-layer cortical neurons) and TUBB3 seq data. Colours are mapped to the clusters (annotated on the right); broader
(neurons). Scale bar, 100 µm. Representative image of one experiment with two cell-type definitions are overlayed on the plot. Data from two independent iPS
biological replicates (n = 2). c, Exemplary immunohistochemistry of 60-day-old cell lines, n = 2 biological replicates each.
HSV-1 affects cellular composition in cerebral organoids and cluster composition across replicates and cell lines were overall
To dissect cell-type-specific effects of HSV-1 infection in cerebral similar (Extended Data Fig. 3c,d).
organoids, we performed scRNA-seq of 60-day-old organoids at To study cell-type-specific effects of viral infection, we hypothe
early (1 dpi) and late (3 dpi) stages of infection (Fig. 3a). Current sized that cell types as defined at the transcriptomic level would not
therapeutic strategies for HSE rely on the administration of viral replica- be changed upon infection. We therefore excluded viral transcripts for
tion inhibitors, such as ACV32. Accordingly, we also analysed organoids cell type labelling and performed data integration, low-dimensional
at 3 dpi, which were treated with ACV starting at 6–8 h after infection. embedding and clustering of control and infected samples (Fig. 3a,b).
Immunostaining showed efficient reduction of GFP protein express This analysis yielded 19 clusters, which we annotated on the basis
ion throughout the entire organoid tissue by this treatment (Extended of data from uninfected organoids (Methods). Notably, numerous
Data Fig. 3a). In addition, we performed scRNA-seq analysis (13.9–67.5%) cells from the infected samples formed four
of ACV-treated organoids in the absence of viral infection to control infection-specific clusters (Fig. 3a,b). The control cells detected in
for non-specific effects of the treatment alone, and found no these clusters (1,760 cells representing 15.9% of these clusters and 14.8%
evidence of substantial effects on cellular transcriptomes (Extended of the control cells), exhibited about half the transcript counts (737
Data Fig. 3b). median unique molecular identifiers (UMIs) compared with 1,184 for
After quality filtering, we collected 36,392 single-cell transcrip- the other control cells and 1,472 for cells from other timepoints in the
tomes across 16 samples from two independent cell lines (control, 1 dpi, same clusters) and were therefore excluded from downstream analyses.
3 dpi, and 3 dpi ACV-treated, two replicates/organoids each), capturing In line with previous studies16,33, a prominent decrease in the num-
host and viral polyadenylated transcripts. Numbers of recovered cells ber of proliferating cells (radial glia clusters) was observed (Fig. 3b

a b CTRL HSV-1 (1 dpi) HSV-1 (3 dpi)

iPS cell lines: 1 2
Uninfected control
(CTRL)
scRNA-seq
HSV-1 (GFP protein)

HSV-1 HSV-1 HSV-1
Electrophysiological
+ analysis
HSV-1 infection Validations

(1 dpi, 3 dpi)
HSV-1/DAPI
c CTRL d HSV-1
0.25 10 0.25 5
0.6 0.6
0.20
ROI number
ROI number 30 0.20 15
0.4 0.4
0.15 60 0.15 25
0.10 80 0.2 0.10 35 0.2
0 100 0 45 0
0
0 50 100 150 200 0 50 100 150 200
Times (s) Times (s)
e STMN2 NEUN HSV-1 Merge f ZO-1 HSV-1–GFP Merge
CTRL
CTRL
STMN2/NEUN/DAPI/HSV-1 ZO-1/HSV-1/DAPI
HSV-1 (1 dpi)
HSV-1 (1 dpi)
HSV-1 (3 dpi)
HSV-1 (3 dpi)
Fig. 2 | HSV-1 impairs synaptic activity and alters neuroepithelium identity. 48 hpi HSV-1-infected (organoids. Scale bar, 100 μm. Images representative of one
a, Schematic overview of experimental set-up. CTRL, uninfected control. experiment, n = 3 biological replicates per condition. e, Immunohistochemistry
b, HSV-1–GFP protein (black) expression in uninfected and 1 and 3 dpi HSV-1-infected for neuronal proteins STMN2 and NEUN, and HSV-1–GFP protein expression in
60-day-old organoids. Bottom: merged images showing HSV-1–GFP protein uninfected and 1 and 3 dpi HSV-1-infected 60-day-old organoids. Right: merged
(white) and nuclear marker DAPI (blue). Scale bars, 500 µm. Images representative images (HSV-1–GFP, white; STMN2, red, NEUN, green) co-stained with nuclear
of three experiments. c, CalBryte590 AM-loaded (green) organoids (left); spike marker DAPI (blue). Scale bars, 100 µm. f, Immunohistochemistry for tight
frequency illustrated by colour-coded image (middle); and spike detection junction protein ZO-1 and HSV-1–GFP protein expression in uninfected and 1 and
plot for all regions of interest (ROIs) (right) in uninfected organoids. Scale bar, 3 dpi HSV-1-infected 60-day-old organoids. Right: merged images (HSV-1–GFP,
100 μm. d, CalBryte590 AM-loaded (green) organoids showing HSV-1–GFP white; ZO-1, red) co-stained with nuclear marker DAPI (blue). Scale bars, 100 µm.
expression (grey) at 48 h post infection (hpi) (left); spike frequency illustrated Images representative of three experiments, with two independent iPS cell lines
by colour-coded image (middle); and spike detection plot for all ROIs (right) in (n = 3 independent organoids/condition/iPS cell line/experiment).
and Extended Data Fig. 4a), which we also validated by immuno (for example, DCX, NEUROD2 and BCL11B) and highly infected cluster 4
histochemistry using cell mitosis marker phospho-vimentin (Extended expressed progenitor genes (for example, VIM and GFAP), while highly
Data Fig. 4b). infected clusters 1 and 3 did not exhibit clear expression signatures and
The infection-specific cell clusters exhibited both extremely high probably represent a mixture of infected cells (Extended Data Fig. 4e). Of
viral loads (that is, the percentage of captured viral transcripts in each note, cells in highly infected cluster 3 showed an extremely high viral load
cell) and expression of numerous viral genes (Fig. 3c and Extended Data (on average, almost 75% of recovered RNA; Extended Data Fig. 4f). Interest-
Fig. 4c). The loss of cellular identity (that is, the lack of co-clustering ingly, despite considerably lower amounts of viral RNA in the ACV-treated
with the cell types in the control organoids) is therefore probably due organoids, the amount of highly infected cells was not markedly reduced.
to both high amounts of viral transcripts and perturbation of the host This could indicate that the strong perturbation of the cellular transcrip-
cell transcriptome. For simplicity, we annotated these cells as ‘highly tomes is a consequence of the infection itself and less of viral replication.
infected’ (Fig. 3a–d). Highly infected cluster 1 was equally composed In summary, we observed a drastic change in the organoid
of cells from each condition, while highly infected clusters 2, 3 and 4 cellular composition, which may be partially attributed to infected
were mainly composed of cells from the HSV-1 3 dpi and 3 dpi ACV cells becoming no longer recognizable from scRNA-seq data alone.
conditions (Extended Data Fig. 4d). In terms of canonical marker With the appearance of infection-specific clusters, the fraction of
gene expression, highly infected cluster 2 expressed neuronal genes almost all cell types was reduced (for example, proliferating radial glia

a b 100 c
Mixed neurons 0.0 2.5 27.8 12.1
Thalamic neurons 2 0.0 0.3 3.5 0.8
Highly infected Cortical neurons 0.0 1.6 1.1 0.7
Hindbrain neurons 1 0.0 1.5 2.3 1.2

UMIs (%)
Thalamic neurons 1 75 0.0 8.0 8.8 4.4
Hindbrain neurons 2 0.0 0.4 1.8 0.7 60

Neurons Radial glia G2M phase 0.0 0.3 1.4 0.6
50
Choroid/hem Radial glia S phase 0.0 0.7 1.8 0.7
Cluster (%)
Radial glia G1 phase 0.0 0.8 1.7 0.9
40
Astroglia 50
Intermediate progenitors 0.0 0.6 1.3 0.8
Mural 30
Retinal progenitors 0.0 0.7 6.0 1.0
Retinal Retinal pigmented cells 0.0 0.7 1.3 0.4

20
Mural cells 0.0 0.3 14.3 0.3
10
Hem/choroid plexus 25
0.0 0.9 4.3 1.2
Astroglia 0.0 0.3 1.9 0.6

0
Intermediate
progenitor Highly infected 1 NA 23.0 20.4 8.5
UMAP2
Highly infected 2 NA 46.2 5.2 3.2
Progenitors
Highly infected 3 NA 68.8 67.0 55.6
Highly infected 4 0 NA 58.7 23.6 4.4
UMAP1
3 i
pi
AC i
V
V
TR
TR
p
dp
dp
AC
1d
1d
C
C
3
HSV-1 HSV-1
d CTRL HSV-1 (1 dpi) HSV-1 (3 dpi) HSV-1 (3 dpi) +ACV

Density
0.03
0.02
0.01
0
UMAP2
UMAP2
UMAP2
UMAP2
UMAP1 UMAP1 UMAP1 UMAP1
Fig. 3 | Cell-type-specific responses in HSV-1-infected and ACV-treated uninfected organoids. d, UMAP plots of cells coloured by the density of cells from
organoids. a, UMAP plot of cells from uninfected (CTRL), HSV-1-infected (1 and each condition in the specific UMAP area shown for uninfected, HSV-1-infected
3 dpi) and 3 dpi HSV-1-infected ACV-treated organoids. b, Proportion of cell types (1 and 3 dpi) and 3 dpi HSV-1-infected ACV-treated organoids, in which yellow
in uninfected, HSV-1-infected (1 and 3 dpi) and 3 dpi HSV-1-infected ACV-treated and blue indicate increased and decreased density, respectively. Data from two
organoids. c, Mean viral load (percentage of viral UMI) in every cluster for each independent iPS cell lines, n = 2 biological replicates/condition/iPS cell line.
condition. Missing values (NA) are indicated for ‘highly infected’ clusters in
from 17.8% to 3.1%, Fig. 3b). Notably, highly infected cells already appear pathway activation was particularly strong in, but not restricted to,
at 1 dpi, increase at 3 dpi reaching 53.3% and only partially decrease highly infected cells (Fig. 4b and Extended Data Fig. 6b).
after 2 days of ACV treatment (Fig. 3b,c). The TNF cytokine is a key mediator of inflammation. The TNF
and receptor superfamilies (TNFSF and TNFRSF) induce an intricate
TNF signalling activated by HSV-1 persists antiviral treatment network of cellular signalling pathways, including activation of nuclear
To identify cell-type-specific molecular alterations induced by HSV-1 factor-kappa B (NF-κB) and its target genes35.
infection, we performed differential gene expression analysis compar- We quantified the expression of all TNF(R)SF members in bulk
ing cells from each infected condition (1 dpi, 3 dpi, and 3 dpi ACV-treated) RNA-seq data from 3-day-infected organoids. TNFSF9 and TNF were
to uninfected cells in the same cluster (Methods). To increase the robust- highly induced upon infection (Extended Data Fig. 7a). Using scRNA
ness of our analysis34, we aggregated randomly selected 50 cells per seq data, we then confirmed that most of the detected ligands are
organoid sample and cluster (‘pseudobulk’) and focused on 4 cell types upregulated upon infection both with and without ACV treatment
(‘mixed neurons’, ‘cortical neurons’, ‘thalamic neurons 1’, ‘radial glia G1 (Extended Data Fig. 7b).
phase’) with at least 50 cells across most organoids (Methods; at least We next examined whether HSV-1 infection leads to canonical
80% or 13 out of 16 samples). Then, we performed GSEA leveraging the NF-κB–RelA pathway activation, detectable by Ser536 phosphoryla-
Hallmark gene sets (Methods). Notably, the term ‘TNF signalling via tion in p65 (also known as RelA) in western blots. As expected, HSV-1
NF-κB’ was the only pathway being consistently upregulated in all ana- infection led to an increase of phosphorylated p65 (p-p65), which was
lysed clusters. Interestingly, significant upregulation was detected only only partially reduced with ACV treatment (Fig. 4c,d). Concomitantly,
at 3 dpi and was not stopped by ACV treatment (Fig. 4a and Extended the impaired tissue integrity at 3 dpi HSV-1 infection was not stopped
Data Fig. 5a). No effect was observed on the TNF signalling pathway in by ACV treatment, despite considerably lower amounts of GFP origina
ACV-treated uninfected organoids (Extended Data Fig. 6a). ting from the virus (Fig. 4e). We did not observe any increase in cleaved
To extend our analysis to all clusters, we ranked genes by their caspase-3 expression in infected organoids and treated organoids
relative expression in every cell and evaluated the enrichment of genes (3 dpi), and therefore excluded apoptosis as the cause of observed
in the ‘TNF signalling via NF-κB’ gene set (Methods). Of note, TNF tissue damage (Extended Data Fig. 6c).

Furthermore, we specifically looked at the expression of NF-κB ACV treatment (Fig. 6a,b), as previously shown in Fig. 4. Co-treatment
target genes, that is, having a NF-κB responsive site in the promoter with either CDDO-Me or NEC-1 significantly strengthened this effect.
or whose expression has been shown to be associated with NF-κB Treatment with CDDO-Me or NEC-1 alone did not reduce NF-κB–RelA
(ref. 36). While some genes were either constitutively expressed pathway activation, indicating again that inhibition of viral replica-
or mildly responding to the infection (for example, HSP90AA1 and tion is a prerequisite for effective treatment (Fig. 6a,b and Extended
YY1), many were broadly activated upon infection and/or enriched Data Fig. 10a). Also, no effect was observed on p-p65 phosporylation
in the ‘highly infected clusters’ (for example, GADD45B, BCL2L11, with irradiated (inactive) HSV-1–GFP virus (Extended Data Fig. 10b).
NFΚBIA, CDKN1A and BNIP3), whereas others were subject to These results show that anti-inflammatory treatment in combination
more cell-type-specific changes (for example, CD83, CDKN1A, PLK3 with antiviral therapies can ameliorate the cellular and tissue damage
and E2F3) (Supplementary Data Fig. 3a). induced by HSV-1 infection.
These results indicate that antiviral treatment, although very To determine whether the HSV-1-driven inflammation is directly
effective when it comes to inhibition of viral replication, cannot stop dependent on HSV-1 replication, we compared wild-type HSV-1 with
the global inflammatory processes triggered by HSV-1 and possibly a mutant lacking ICP27, an essential protein for the expression of late
mediated by TNFSF members. genes and viral DNA synthesis. Because of its replication defect38, we
compared p65 phosphorylation in organoids infected with wild-type
Combinatorial treatment reduces defects in organoids virus for 3 days with those infected with the mutant strain for 6 days. Six
To test whether anti-inflammatory drugs can prevent the observed days of infection with the mutant strain were sufficient to observe viral
tissue and cellular damage observed upon infection, we used two dif- protein levels similar to those at 3 days with the wild-type virus (Extended
ferent compounds targeting the TNF pathway, alone and in combina- Data Fig. 10c). However, already at 3 dpi the p65 phosphorylation level
tion with ACV: first, necrostatin-1 (NEC-1), an RIP1-targeted inhibitor was increased in the mutant strain, despite the viral protein levels being
of TNF-induced necroptosis, and second, CDDO-Me, an NRF2 agonist lower than in the wild-type strain (Fig. 6c,d). Furthermore, STMN2 expres-
that attenuates the NF-κB-mediated inflammatory response37. We sion and neuroepithelium integrity (ZO-1 expression) were affected by
infected 60-day-old organoids with low (0.2) and high (1.0) multiplicity both viruses, but here the effect was weaker in the mutant than in the
of infection (MOI) of HSV-1 GFP virus and treated them 8–12 h after wild-type strain (Extended Data Fig. 10d,e). These observations suggest
infection with different combinations of ACV, NEC-1 and CDDO-Me. that the activation of NF-κB and the related phenotypic consequences
As a readout, we analysed neuroepithelial integrity, neuronal dam- on tissue integrity were at least partially independent of viral replication.
age and viral load by measuring ZO-1, STMN2 and GFP, respectively, We further investigated the activation of upstream factors leading to
using immunostaining and protein quantification (Fig. 5a and Extended NF-κB p65 phosphorylation and observed that both wild-type and the
Data Fig. 8a–c). In organoids treated with ACV alone or co-treated with ICP27 deletion mutant lead to activation of p-AKT and ERK1/2 (Extended
anti-inflammatory agents, viral GFP expression was similarly reduced Data Fig. 10f), supporting the hypothesis that viral replication is
(Fig. 5b–d). Next, we assessed the localization of ZO-1 in infected orga- not prerequisite for activation of inflammatory responses.
noids. Notably, immunohistochemistry analysis of ZO-1 expression To assess whether p65 phosphorylation in infected organoids is
revealed that ACV alone only partially rescued the architecture of mediated by signalling molecules released during viral infection, we
organoid neuroepithelium, but co-treatment with both CDDO-Me or exposed naive brain organoids to the medium collected from infected
NEC-1 significantly improved neuroepithelium architecture and pre- organoids depleted of virus via two rounds of filtration with 0.1 µm
vented ZO-1 hyperactivation in both iPS cell lines (Fig. 5c and Extended filters (Extended Data Fig. 10g). No viral protein was detected in these
Data Fig. 8d–g). In addition, the downregulation of STMN2 expression, organoids by western blot analysis (Extended Data Fig. 10h,i) and only
which was one of the strongest observed responses to HSV-1 infection a few infective viral particles were detected in the medium by plaque
in our system (Fig. 2e and Supplementary Fig. 2a), was significantly assay (~30 viral particles per organoid). p65 phosphorylation in these
reduced in co-treated organoids, as shown by immunostaining and organoids was significantly elevated compared with mock medium
western blot analysis (Fig. 5d and Extended Data Fig. 9b,c). With a exposed organoids (Extended Data Fig. 10i,j), suggesting that infected
lower MOI, the phenotype was milder, and ACV alone showed stronger organoids release signalling molecules capable of NF-κB pathway
effect, although still superseded by co-treatments (Extended Data activation independently from viral infection.
Figs. 8d–g and 9a). Treatment with CDDO-Me or NEC-1 alone did
not prevent defects triggered by HSV-1, indicating that effectively Discussion
restoring tissue architecture requires blocking both viral replication A comprehensive understanding of the pathogenesis of HSV-1-driven
and inhibiting TNF signalling (Fig. 5c,d and Extended Data Fig. 9b,c). encephalitis and its association with brain tissue damage is still largely
To assess whether the combinatorial treatment suppresses HSV- elusive. We demonstrated that human brain organoids, combined with
1-driven neuroinflammation, we analysed p65 phosphorylation in a broad range of RNA, protein and electrophysiological analyses, can
treated and uninfected organoids. With both iPS cell lines, p65 phos- contribute to a better understanding of HSV-1 pathogenesis and serve
phorylation was strongly induced at 3 dpi and partially reduced by as a model to study potential therapeutic strategies.
Fig. 4 | ACV does not prevent HSV-1-driven inflammatory responses. virus expression (bottom). d, Semi-quantification of p-p65 protein expression
a, Enriched gene sets for the assessed cell types based on the comparison of by densitometry in uninfected, 3 dpi HSV-1-infected and 3 dpi HSV-1-infected
HSV-1-infected (1 and 3 dpi) and 3 dpi HSV-1-infected ACV-treated organoids ACV-treated organoids, normalized to GAPDH expression. *P < 0.05, **P < 0.01,
with uninfected organoids. Colour indicates normalized enrichment score Benjamini–Hochberg-corrected two-sided t-tests. n = 6 biologically independent
(NES) values and dot size represents −log10-adjusted P value for statistically samples; that is, six batches of pooled organoids from two iPS cell lines; colour
significant pathways according to gene set enrichment analysis (GSEA) analysis. indicates cell line of origin. Bar plots and error bars show mean ± standard
b, UMAP plots of cells coloured by the TNF pathway activity score in each cell error. e, Representative immunohistochemistry for SOX2 (top) and HSV-1–GFP
and separated in uninfected (CTRL), HSV-1-infected (1 and 3 dpi) and 3 dpi HSV- protein expression (middle) in uninfected, HSV-1-infected (1 and 3 dpi), and 3
1-infected ACV-treated organoids, with red and blue indicating increased and dpi HSV-1-infected ACV-treated organoids. Bottom: merged images (HSV-1–GFP,
decreased scores, respectively. Pooled data from two independent iPS cell lines, white; SOX2, red) co-stained with nuclear marker DAPI (blue). Scale bar, 100 µm.
n = 2 biological replicates for each line. c, Representative western blot analysis Representative image of three experiments (n = 3 independent organoids/
of p-p65 expression in uninfected, 3 dpi HSV-1-infected and 3 dpi HSV-1-infected condition/iPS cell line).
ACV-treated organoids; GAPDH serves as loading control, and GFP indicates

Our data indicate several HSV-1-induced defects at the molecular function. Secondly, activation of TNF–NF-κB signalling stood out in
level, which may be related to brain tissue damage and neuro an exploratory analysis of scRNA-seq data, indicating inflammation
degeneration observed in patients with encephalitis2. Firstly, down- as a factor contributing to brain tissue damage. The immune response
regulation of synaptic proteins and the key axonal maintenance factor induced by HSV-1 during the early phase of infection is able to limit
STMN2 may be associated with neuronal damage28,29. Calcium imaging viral replication, but can also lead to sustained and brain-damaging
data also supported that HSV-1 infection very rapidly alters neuronal neuroinflammation39. Antiviral ACV treatment, which is commonly
a −log10 adj. P value

Radial glia G1 phase Cortical neurons Mixed neurons Thalamic neurons 1
2
TNF signalling via NF-κB
4
UV response up
Apoptosis 6
Interferon-γ response 8
Myc targets v2
Unfolded protein response
E2f targets NES
p53 pathway
Myc targets v1 2
DNA repair 1
Wnt β-catenin signalling 0
Coagulation
−1
Haem metabolism
PI3K AKT mTOR signalling −2
Protein secretion
Apical surface
Xenobiotic metabolism
Bile acid metabolism
Oestrogen response late
Glycolysis
Peroxisome
mTORC1 signalling
Adipogenesis
Cholesterol homeostasis
Oestrogen response early
Fatty acid metabolism
Hedgehog signalling
Oxidative phosphorylation
L
L
L
L
L
L
L
L
L
L
TR
TR
TR
TR
TR
TR
TR
TR
TR
TR
TR
TR
C
C
C
C
C
C
C
C
C
C
us
us
us
us
us
us
us
us
us
us
us
us
rs
rs
rs
rs
rs
rs
rs
rs
rs
rs
rs
rs
ve
ve
ve
ve
ve
ve
ve
ve
ve
ve
ve
ve
i)
i)
i)
i)
i)
i)
V
V
i)
i)
V
dp
dp
dp
dp
dp
dp
dp
dp
AC
AC
AC
AC
(1
(1
(1
(3
(3
(3
(1
(3
+
+
+
+
-1
-1
-1
-1
-1
-1
-1
-1
i)
i)
i)
i)
SV
SV
SV
SV
dp
dp
dp
dp
SV
SV
SV
SV
H
H
H
H
H
H
H
H
(3
(3
(3
(3
-1
-1
-1
-1
SV
SV
SV
SV
H
H
H
H
TNF pathway score
0.10
0.05
0
UMAP2
UMAP2
UMAP2
UMAP2
UMAP1 CTRL UMAP1 HSV-1 (1 dpi) UMAP1 HSV-1 (3 dpi) UMAP1 HSV-1 (3 dpi) + ACV
c d P = 0.002 e
** CTRL HSV-1 (1 dpi) HSV-1 (3 dpi) HS1 (3 dpi) + ACV
HSV-1 (3 dpi) + ACV
* P = 0.023
SOX2
HSV-1 (3 dpi)
* P = 0.023
3
CTRL
p-p65/GAPDH
HSV-1 GFP
2
p-p65
70 kDa
1
GAPDH
Merge
35 kDa
25 kDa
HSV-1-GFP SOX2/HSV-1/DAPI
0
CTRL
HSV-1 (3 dpi)
HSV-1 (3 dpi) + ACV

a b HSV-1 (3 dpi)
HSV-1-GFP CTRL Mock ACV CDDO-Me NEC-1 ACV + CDDO-Me ACV + NEC-1
8–12 h 3 days Neuroepithelial integrity (ZO-1)
(a) DMSO
(b) ACV
(c) ACV+CDDO-Me Neuronal integrity (STMN2)
GFP
(d) ACV+ NEC-1
HSV-1 (3 dpi)
c CTRL Mock ACV CDDO-Me NEC-1 ACV + CDDO-Me ACV + NEC-1

Z0-1
GFP (HSV-1)
Merge
ZO-1/HSV1/DAPI
HSV-1 (3 dpi)
d CTRL Mock ACV CDDO-Me NEC-1 ACV + CDDO-Me ACV + NEC-1

STMN2
GFP (HSV-1)
Merge
STMN2/HSV-1/DAPI
Fig. 5 | Combinatorial treatment strategies for HSV-1-driven neuroinflam images (HSV-1–GFP, white; ZO-1, red) co-stained with nuclear marker DAPI (blue).
mation. a, Schematic overview of combinatorial treatment set-up. b, Exemplary Scale bars, 100 µm. Representative image of three experiments, two different iPS
images of intact organoids (top) and HSV-1–GFP protein expression (bottom) in cell lines (n = 3 independent organoids/condition/iPS cell line). d, Representative
uninfected (CTRL) and 3 dpi HSV-1-infected mock-treated (DMSO), ACV-treated, immunohistochemistry for STMN2 (top) and HSV-1–GFP (middle) protein
CDDO-Me-treated, NEC-1-treated, ACV and CDDO-Me-co-treated, and ACV and expression in uninfected and 3 dpi HSV-1-infected mock-treated, ACV-treated,
NEC-1-co-treated organoids. Scale bars, 800 µm. Representative image of three CDDO-Me-treated, NEC-1-treated, ACV and CDDO-Me-co-treated, and ACV
experiments from two different iPS cell lines (n = 3 independent organoids/ and NEC-1-co-treated organoids. Bottom: merged images (HSV-1–GFP, white;
condition/iPS cell line). c, Representative immunohistochemistry for ZO-1 (top) STMN2, red) co-stained with nuclear marker DAPI (blue). Scale bars, 100 µm.
and HSV-1–GFP (middle) protein expression in uninfected and 3 dpi HSV-1- Representative image of three experiments, two different iPS cell lines (n = 3
infected mock-treated, ACV-treated, CDDO-Me-treated, NEC-1-treated, ACV and independent organoids/condition/iPS cell line).
CDDO-Me-co-treated, and ACV and NEC-1-co-treated organoids. Bottom: merged
prescribed in clinical settings for patients suspected of encephalitis, It has been previously shown in a mouse model that HSV-1
was able to significantly reduce viral replication, but not TNF–NF-κB induces expression of TNF40, and that co-treatment with etanercept
signalling. Sustained inflammation can be toxic to neuronal cells, an (an antibody toTNF) or glucocorticoids increased the survival rate of
effect potentiated by activated immune cells. This could also explain HSV-1-infected mice compared with those only treated with antiviral
the fact that 70% of ACV-treated patients with HSE do not completely drugs41,42. Consequently, application of immunomodulatory strategies
recover neurological functions and suffer from cognitive deficits3. for HSE is currently being discussed3. Our data support this notion and

ICP27_MUT_rep2
ICP27_MUT_rep1
HSV-1_WT_rep2
HSV-1_WT_rep1
a +ACV c
MOCK_rep2
MOCK_rep1
+HSV-1
CDDO-Me
CDDO-Me
NEC-1
NEC-1
Mock
CTRL
ACV
p-p65 70 kDa
p-p65 70 kDa
ICP0
100 kDa
GAPDH 35 kDa
GFP (HSV-1) 25 kDa
GAPDH 35 kDa
b d ICP0 p-p65
5
GFP p-p65
4 4
P = 0.01
*
* P = 0.03
GFP or p-p65/GAPDH
Batch
Protein/GAPDH
3 3 P = 0.025 1
Line * 2
iPSC-1
2 3
iPSC-2 2
P = 0.013 4
*
1 1
0 0
CDDO-Me
CDDO-Me
CDDO-Me
CDDO-Me
Mock
Mock
NEC-1
NEC-1
NEC-1
NEC-1
CTRL
CTRL
ACV
ACV
ICP27_MUT
ICP27_MUT
WT_HSV-1
WT_HSV-1
Mock
Mock
+HSV-1 +HSV-1
+ACV +ACV
Fig. 6 | Combinatorial treatment reduces HSV-1-driven neuroinflammation all others; colour indicates cell line of origin). Error bars indicate the mean ±
in HSV-1-infected organoids. a, Representative western blots analysis of standard error. c, Representative western blots analysis of p-p65 expression in
p-p65 expression in uninfected (CTRL) and 3 dpi HSV-1-infected mock-treated uninfected (mock) organoids and 3-dpi organoids infected with wild-type (WT)
(DMSO), ACV-treated, CDDO-Me-treated, NEC-1-treated, ACV and CDDO-Me- HSV-1 KOS strain or HSV-1 KOS ICP27 mutant strain. GAPDH serves as a loading
co-treated, and ACV and NEC-1-co-treated organoids. GAPDH serves as a loading control and ICP0 indicates virus expression (bottom). d, Semi-quantification
control and GFP indicates virus expression (bottom). b, Semi-quantification of HSV-1-ICP0 (left) and p-p65 protein expression (right) by densitometry in
of HSV-1–GFP (left) and p-p65 protein expression (right) by densitometry in uninfected and 3-dpi organoids infected with wild-type HSV-1 KOS or HSV-1
uninfected and 3 dpi HSV-1-infected mock-treated, ACV-treate, CDDO-Me- KOS ICP27 mutant strain. P values were determined by Benjamini–Hochberg-
treated, NEC-1-treated, ACV and CDDO-Me-co-treated, and ACV and NEC-1-co- corrected paired two-sided t-tests (n = 5 biologically independent samples
treated organoids, normalized to GAPDH expression. P values were determined coming from n = 4 batches of medium preparation, indicated by colour).
by Benjamini–Hochberg-corrected two-sided t-tests (n = 5 biologically Bar plots and error bars show mean ± standard error.
independent samples for CTRL, HSV1 + CDDO-Me and HSV1 + NEC-1, n = 6 for
suggest that antiviral treatment alone is not sufficient to treat HSE, and complexity of the human brain, as they lack immune cells and blood
that a combinatorial therapy with anti-inflammatory drugs targeting vessels and better resemble foetal than adult brain tissues43. This, how-
the TNF–NF-κB pathway may be a valuable addition to current thera- ever, enables the study of a ‘pure’ neuronal immuno-defensive state,
peutic approaches. which is becoming a central topic in the field of neuroimmunology44.
So far, mostly intranasal or intracerebral infection mouse models Indeed, we did observe a strong activation of the innate inflammatory
have been used to study HSE and the infection pattern observed was cellular response, reflected by TNF–NF-κB pathway activation. This
more diffuse than that observed in patients with HSE, indicating was accompanied by a lower and cell-type-specific interferon pathway
that the mechanism of virus spreading can be different in humans40. activation. A complete lack of interferon activation was previously
For example, the virus did not spread to the temporal and frontal reported in early-stage HSV-1-infected organoids16, where treatment
lobes, which are the typical locations in HSE in humans40. HSV-1 is a with exogenous IFNα2 was able to inhibit viral replication and thus
human-specific pathogen, and therefore human iPS cell-derived brain prevent growth defects in this model. More complex brain organoids
organoids can be very valuable to more precisely model virus–host containing also immune cells may therefore be an important step
interaction, to develop preventive and therapeutic treatments. Of forward to better understand infections with neuroinvasive viruses.
note, organoids can also be an improvement to 2D neuron culture, as,
for example, prolonged infections or latency are difficult to model in Methods
2D culture but possible in organoids5. Study design
As a major limitation, we would like to emphasize that the In this study, we used human cerebral organoids as a model for acute
brain organoids used in this study capture only some aspects of the HSV-1. The aim of the study was as follows: (1) to assess cellular and

molecular signatures associated with viral infection; (2) to study the cells at the surface of the organoids, or MOI of 0.2), and 24 h after the
outcome of the current antiviral therapy used in clinical practice (ACV); medium was replaced with fresh organoid maturation medium and
(3) to identify and test better potential therapeutic approaches based organoids were cultured for an additional 1 or 3 days (1 dpi/3 dpi). The
on the results of 1 and 2. We used two genetically distinct human iPS control organoids were culture in parallel in the same conditions, but
cell lines and performed exploratory transcriptomic analyses (bulk without viral infection. For the anti-inflammatory treatment, around
RNA-seq and scRNA-seq) in two replicates and three conditions (con- five to eight organoids cultured in six-well plates were infected with
trols and two timepoints of infection, three replicates per condtion). an MOI of 1 or MOI of 0.2, and 8–12 h after the medium was replaced
For scRNA-seq, we additionally performed two replicates per line of with fresh organoid maturation medium supplemented with ACV
infected and ACV-treated organoids. Drug concentrations as well as (50 µM), NEC-1 (40 µM) or CDDO-Me (0.8 µM) or combination of ACV
timepoints of infection were determined on the basis of pilot experi- and organoids were cultured for an additional 1 or 3 days (1 dpi/3 dpi).
ments assessing virus spreading over time, with the aim of modelling For the infected medium exposure, we infected organoids with
acute infection. Validation experiments and additional characteriza- 75,000 viral particles per organoid (8 organoids per well) for 3 days.
tion experiments (western blots, immunhistochemistry and calcium We collected the medium from the infected organoids and filtrated
imaging), for example, for assessing the outcome of anti-inflammatory it twice thought a 0.1 µm filter and added the medium to naive (unin-
treatments, were performed in two to five biological replicates per fected) organoids by replacing half of the medium. Organoids were
cell line. Organoids were assigned blindly to the various experimental incubated with conditioned medium (medium HSV-1) for additional
groups. Data analysis could not be performed blind to the conditions of three days in the presence of ACV (to ensure no viral replication) before
the experiments due to the GFP signal detection in infected organoids. subequent analyses. The medium from mock-infected organoids,
No statistical methods were used to pre-determine sample sizes, but our which was processed in the same way from mock-infected organoids,
sample sizes are similar to those reported in previous publications16. served as control (medium mock).
For ultraviolet irradiation experiment, the HSV-1 GFP virus stock
iPS cell lines was diluted to about 5 × 106 PFU ml−1 in DMEM, and exposed to 254 nm
The human iPS cell lines iPSC-1 XM001 (ref. 45) and iPSC-2 (A18945, ultraviolet light in 35 mm plastic dishes without lid. We used 10 min
Thermo Fisher Scientific) were cultured in standard hypoxic conditions irradiation in a conventional Analytik Jena 254 nm crosslinker. After
(37 °C, 4% CO2, 4% O2 and 100% humidity), in E8 Flex medium (Thermo 10 min, the remaining activity of the virus was about 100 PFU ml−1.
Fisher Scientific).
Immunostaining of brain organoids
Generation of cerebral organoids For immunostainings, organoids were washed three times with
We generated iPS cell-derived cerebral organoids according to a phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde
protocol previously described with some modifications7. Shortly, for 20–60 min (depending on the organoid size) at 4 °C, then washed
after dissociation into single-cell suspension with accutase, we seeded with PBS three times for 10 min each. The tissue was incubated in 40%
6,000 cells per one well of 96-well plates in 100 µl of embryoid body sucrose (in PBS) until it sunk (overnight) and then embedded in 13%/10%
(containing Dulbecco’s modified Eagle medium (DMEM)/F12, 20% gelatin/sucrose. Frozen blocks were stored at −80 °C, before cryosec-
knockout replacement serum, 1× Glutamax, 1× Minimal Essential tions. Then 10–14 μm sections were prepared using a cryostat. Sections
Medium-Non Essential Amino Acid (MEM-NEAA), 2% embryonic stem were incubated with warm PBS for 10–15 min to remove the embedding
cell-qualified fetal bovine serum (FBS), 50 µM ROCK inhibitor, 10 µM medium and then fixed for additional 10 min with 4% paraformal
basic fibroblast growth factor or STEMdiff Cerebral Organoid Basal dehyde, washed three times with PBS and blocked and permeabilized
Medium supplemented with Supplement A (STEMCELL Technologies) in 0.25% Triton-X, 5% normal goat serum in PBS for 1 h. Sections were
medium. After 4 days, we replaced the medium with EB medium with- first incubated with primary antibodies (Supplementary Table 1) in
out basic fibroblast growth factor and ROCK inhibitor. On day 5, we 0.1% Triton-X, 5% normal goat serum overnight. The following primary
replaced the medium with a neural induction medium (DMEM/F12, antibodies were used: STMN2 (Novus Biologicals, NBP1-49461, 1:1,000),
1× N2 supplement, 1× Glutamax, 1× MEM-NEAA, 10 µg ml−1 heparin SOX2 (Merck, AB5603, 1:1,000), PAX6 (Biolegends 901301, 1:1,000),
solution). At days 7–9, we performed a liquid embedding of orga- ZO-1 (Invitrogen 339100, 1:1,000), NEUN (Sigma-Aldrich, MAB377,
noids into 2% Matrigel (Corning, 356234) and kept them in neural 1:500), MAP2 (Sigma-Aldrich, MAB3418A5, 1:1,000), SYN1 (Abcam
induction medium for 2 days, and in organoid differentiation medium ab8, 1:1,000), TUJ1 (Sigma-Aldrich T2200, 1:1,000), GFAP (Milipore/
containing 1:1 DMEM/F12: Neurobasal, 1× N2 supplement, 1× B27 + Merck, 1:1,000), TBR1 (Abcam, ab183032, 1:1,000), GFP (Abcam,
vitamin A supplement, insulin, 2-mercaptoethanol solution, Glutamax, ab13970, 1:1,000), phospho-Vimentin (Biozol Diagnostics, MBL-D076-3,
MEM-NEAA and CHIR99021 for another four days. Next, we transferred 1:2,000), SATB2 (Abcam ab34735, 1:1,000), C-Caspase3 (Cell Signaling
the organoids to ultralow-attachment six-well plates and cultured 9661, 1:500) and ICP0 (Santa Cruz SC53070, 1:500). They were then
them on an orbital shaker (80 r.p.m.) in organoid maturation medium washed three times for 10 min each with PBST (0.1% Triton X-100) and
containing 1:1 DMEM/F12:Neurobasal, N2 supplement, B27 + vita- incubated with secondary antibodies (Alexa 488 anti-chicken (Abcam
min A supplement, insulin, 2-mercaptoethanol solution, Glutamax 150169, 1:500), Alexa 568 anti-mouse (Abcam ab157473, 1:500), Alexa
supplement, MEM-NEAA, sodium bicarbonate, vitamin C solution, 647 anti-mouse (Abcam ab150115, 1:500), Alexa 568 anti-rabbit (Abcam
chemically defined lipid concentrate, brain-derived neurotrophic ab175471, 1:500), Alexa 647 anti-rabbit (Abcam ab150083, Alexa 647
factor (BDNF), glial cell line-derived neurotrophic factor (GDNF), anti-goat (Thermo Fisher A21-447 1:500)) at room temperature for 2 h,
cAMP and 1% Matrigel. Live organoid imaging was performed using followed by staining with 4′,6-diamidino-2-phenylindole (DAPI) (final
NICON Eclipse Ts2 microscope. 1 µg ml−1) for 10 min and washed three times with PBST. The images
were acquired using a Keyence BZ-X710 microscope.
Viral infections
The following HSV-1 strains were employed in this study: strain 17 Western blotting
containing GFP under an MCMV promoter46 KOS (VR-1493; ATCC), Organoids were lysed in RIPA buffer (150 mM NaCl, 5 mM EDTA, 50 mM
and a KOS-based ICP27 knockout virus38. Around five to eight orga- Tris, 1% NP-40, 0.5% sodium deoxycholate and 0.1% SDS) in the presence
noids cultured in six-well plates were infected with 75,000 or 15,000 of protease and phosphatase inhibitors. Subsequently, the samples
plaque-forming units (PFU) of HSV-1 particles per organoid (corres were incubated on ice for 30 min and centrifuged at 14,000g for 20 min
ponding to an MOI of 1 as calculated from the estimated number of at 4 °C. The total protein concentrations were determined using the

Pierce BCA assay (Thermo Fisher Scientific, 23225). Samples (10–20 µg) described31. Overlapping antisense transcripts from different RNA-seq
were boiled in standard SDS–PAGE sample buffer supplemented with libraries were collapsed together. All putative antisense transcripts
100 mM dithiothreitol, boiled at 95 °C and resolved in a 10–12% SDS– with an annotated upstream sense gene within 1 kb were discarded
PAGE. Proteins were transferred onto Transblot Turbo Midi PVDF as potential readthroughs, and those with a downstream sense gene
membranes (Bio-Rad, 1704157). The membrane was blocked with within 10 kb were discarded as potential transcripts from an upstream
Tris-buffered saline containing 0.1% Tween-20 (TBST, Sigma-Aldrich) transcription start site. We extended the GENCODE v27 genome anno-
supplemented with 5% skimmed milk for 1 h at room temperature. tation reference with detected antisense transcripts and counted the
Subsequently, the membranes were incubated with the primary anti- mapping reads using htseq-count 0.9.1, as described earlier.
bodies (Supplementary Table 1) diluted in TBST with skimmed milk
at 4 °C overnight. The following primary antibodies were used: AKT Tissue processing and Visium data generation
(Cell Signaling 9272, 1:1,000), p-AKT (Cell Signaling 9277S, 1:1,000), Brain organoid samples, frozen in isopentane and embedded in OCT
SYN1 (Abcam ab8, 1:1,000), TRAF6 (Santa Cruz 7221, 1:1,000), ICP0 (TissueTek Sakura) and cryosectioned at −12 °C (Thermo Cryostar).
(Santa Cruz SC53070, 1:800), GFP (Merck G1546, dilution 1:1,000), Sections were placed on chilled Visium Spatial Gene Expression Slides
pERK1/2 (Cell Signaling 9102, dilution western blot: 1:1,000), HOMER1 (2000233, 10x Genomics). Tissue sections were then fixed in chilled
(Synaptic Systems, 160011, 1:1,000), p-p65 (Cell Signaling 3033 T, dilu- methanol and stained according to the Visium Spatial Gene Expres-
tion western blot: 1:2,000) and GAPDH (Sigma G8795, 1:2,000). After sion User Guide (CG000239 Rev A, 10x Genomics). Libraries were
washing in TBST, the membranes were incubated with anti-rabbit or prepared according to the Visium Spatial Gene Expression User Guide
anti-mouse horseradish peroxidase-conjugated secondary antibody (CG000239; ref. 53). Libraries were sequenced on a NovaSeq 6000 Sys-
(secondary antibodies: goat anti-mouse IgG (H+L) HRP (Thermo Fisher tem (Illumina) using a NovaSeq S4 Reagent Kit (200 cycles, 20027466,
Scientific 31430, 1:8,000), polyclonal goat anti-rabbit HRP, Dako P0448, Illumina), at a sequencing depth of approximately 250–400 M read
1:5,000)) for 1 h at room temperature. Blots were then developed with pairs per sample. Sequencing of 10 µm organoid slices (four organoids
AmershamTM ECL select reagent (Cytiva, RPN3243) and visualized per condition) yielded, for uninfected organoids, on average 8,000
by Fusion FX (Vilber). Band intensities were quantified using ImageJ. UMIs, which quantified ~3,700 genes per spot. Data were analysed
using the Spaceranger 1.2.0 software (10x Genomics).
RNA sequencing
For the total RNA sequencing we used 100 ng of total RNA, where ribo- scRNA-seq
somal RNA was depleted using RNase H-based protocol. We mixed total Methanol-fixed cells were centrifuged at 3,000–5,000g for 5 min, rehy-
RNA with 1 μg of a DNA oligonucleotide pool comprising 50-nt-long drated in 1 ml PBS + 0.01% bovine serum album (BSA) supplemented
oligonucleotide mix covering the reverse complement of the entire with RNAse inhibitors (1 unit μl−1 RiboLock, ThermoFisher), pelleted
length of each human rRNA (28S rRNA, 18S rRNA, 16S rRNA, 5.8S rRNA, and resuspended again in 0.5 ml PBS + 0.01% BSA in the presence of
5S rRNA and 12S rRNA), incubated with 1 U of RNase H (Hybridase RNAse inhibitors. Cells were manually counted by means of a haemo-
Thermostable RNase H, Epicentre), purified using RNA Cleanup XP cytometer and diluted to a suspension of typically ~300 cells μl−1 in
beads (Agencourt), DNase treated using TURBO DNase rigorous treat- PBS + 0.01% BSA. Cells were encapsulated together with barcoded
ment protocol (Thermo Fisher Scientific) and purified again with RNA microparticles (Macosko-2011-10 (V+), ChemGenes Corp.) using the
Cleanup XP beads. We fragmented the rRNA-depleted RNA samples Dolomite Bio Nadia instrument, using the standard manufacturer’s
and processed them into strand-specific complementary DNA librar- dropseq-based54 scRNA-seq protocol.
ies using TruSeq Stranded Total LT Sample Prep Kit (Illumina) and Droplets were broken immediately after collection and cDNA
then sequenced them on NextSeq 500, High Output Kit, 2 × 76 cycles. libraries generated as previously described54. First-strand cDNA was
We performed differential gene expression analysis using the DESeq2 amplified by equally distributing beads from one run to 24 Smart
(version 1.20.00) R package. PCR reactions (50 μl volume; 4 + 9–11 cycles). Then 20 μl fractions of
each PCR reaction were pooled (total 480 μl), then double-purified
RNA-seq data analysis with 0.6× volumes of AMPure XP beads (Beckman Coulter). Amplified
Bulk RNA sequencing data were analysed using the PiGx-RNA seq pipe- cDNA libraries were assessed and quantified on a BioAnalyzer
line version 0.0.3 (ref. 47). Briefly, RNA-seq reads were mapped to High Sensitivity Chip (Agilent) and the Qubit dsDNA HS Assay system
the custom human-viral reference genome using STAR version 2.7.3a (ThermoFisher). Then 600 pg of each cDNA library was fragmented,
(ref. 48) and those uniquely mapping to exons have been used for amplified (13 cycles) and indexed for sequencing with the Nextera XT v2
gene expression quantification. Default settings were used, with DNA sample preparation kit (Illumina) using custom primers enabling
the exception of ‘–outFilterMismatchNoverLmax 0.05’. Reads were 3′-targeted amplification. The libraries were purified with AMPure XP
counted using htseq-count (version 0.9.1; ref. 49) with GENCODE v27 Beads, quantified and sequenced on Illumina NextSeq500 sequencers
genome annotation reference50. (library concentration 1.8 pM; NextSeq 500/550 High Output v2 kit (75
cycles) in paired-end mode; read 1 = 20 bp using the custom primer
Differential gene expression analysis Read1CustSeqB49 read 2 = 64 bp).
Differential gene expression analysis was done using DESeq2
version 1.34.0 (ref. 51), using default options, and multi-factor design scRNA-seq data analysis
that measures the effect of infection while accounting for differences Quality control and data pre-processing. Raw sequencing data were
between cell lines (design = ~ line + group). Significance threshold was converted to fastq format using bcl2fastq (v.2.19). Single-cell digital
set to adjusted P value of 0.05. The list of TNF ligands was obtained gene expression matrices (DGEs) were generated using the space-
from the HUGO Gene Nomenclature Committee52 database under the make pipeline (v.0.4.3)55 with Drop-seq tools (v.2.5.0) (https://github.
gene group ‘tumour necrosis factor (ligand) superfamily’. The list of com/broadinstitute/Drop-seq) and STAR (v.2.6.0)48. Briefly, read 1
synaptic genes was obtained from the Gene Ontology database under was used to determine the cell barcode (base pairs 1–12) and the UMI
the biological process ‘chemical synaptic transmission’. (13–20), while read 2 was mapped to a custom human-viral reference
genome built by simple concatenation of the human (hg38) and HSV1
Antisense transcription analysis genomes (barcode_flavors: default, run mode: scRNA_seq). For the
Antisense transcribed regions were identified and defined from HSV-1 genome, we started out from the known previously published
STAR-aligned reads using the running sum algorithm previously sequence24 and refined it on the basis of the total RNA-seq data. For

the mapping, we removed the terminal repeats to avoid multimappers. Differential gene expression and GSEA
The employed sequence and annotation is available on the National To identify differentially regulated genes upon infection in robust
Center for Biotechnology Information Gene Expression Omnibus entry and reproducible way, we performed a pseudobulk analysis using the
accompanying this manuscript (GSE163952). edgeR package62. For selected clusters, we generated a pseudobulk
Only reads mapping to exons have been used for gene expres- expression profile for each organoid by summing the expression of 50
sion quantification (count_intronic_reads: false) and only the 10,000 randomly selected cells (with replacement). We then used a multifac-
barcodes with the highest UMI counts have been used for downstream torial design to account for intercell line differences (function model.
analyses (n_beads: 10,000). DGEs were further analysed in R (v4.1) using matrix(~line + condition)). We then filtered genes with low expression
Seurat (v.4.0.4)56. We initially adopted a very conservative filtering values (filterByExpr function with min.count 10 and min.total.count
strategy removing cells with fewer than 250 detected genes and genes 50), normalized expression values by library size (calcNormFactors
detected in fewer than 5 cells (CreateSeuratObject function with min. function), estimated negative binomial dispersions (estimateDisp
cells of 5 and min.features of 250). We normalized gene expression function) and fitted a generalized linear model (glmQLFit function).
and scaled gene expression for each individual organoid (SCTrans- For each comparison (1 dpi versus CTRL, 3 dpi versus CTRL, 3 dpi ACV
form function with method ‘glmGamPoi’ and vst.flavor ‘v2’)57. We then versus CTRL), we identified differentially expressed genes (glmQLFTest
integrated cells from control organoids to analyse cell composition function) and adjusted P value was calculated using Bonferroni cor-
at baseline and integrated DGEs from all organoids to analyse the rection for multiple testing correction (p.adjust function). To identify
response to infection. To do so, we selected the 3,000 genes most enriched pathways associated with HSV infection, we performed GSEA
variable across the datasets (SelectIntegrationFeatures function with using clusterProfiler63. For each cluster and each comparison, genes
nfeatures 3,000), computed missing residuals (PrepSCTIntegration were ranked according to decreasing log fold change to perform gene
function), identified cells representing mutual nearest neighbours58 set enrichment against the Hallmark gen sets (GSEA function). Only
or pairwise anchors between datasets (FindIntegrationAnchors func- pathways with a Benjamini–Hochberg-adjusted P value lower than
tion)59 and finally performed iterative pairwise dataset integration 0.05 were considered as significantly enriched.
starting from the most similar ones (IntegrateData function). We then
performed dimensionality reduction (RunPCA function) and selected Calcium imaging
the first 30 principal components to identify shared nearest neighbours For calcium imaging experiments, organoids were plated in 35 mm
(FindNeighbors with dims 1:30), clustering (FindClusters function glass-bottom imaging dish coated with poly-d-lysine and Geltrex.
with resolution 0.4 for control organoids and 0.8 for all organoids) Two days after plating, organoids were loaded with 5 μM CalBryte590
and compute a two-dimensional embedding to visualize the AM (AAT Bioquest) and 0.02% pluronic acid in organoid maturation
data (RunUMAP with dims 1:30). We removed two control clusters medium for 30 min at 37 °C. Imaging was performed in a Dragonfly
(clusters 3 and 14) and four clusters across all organoids (clusters 6 spinning disk confocal microscope (Andor, Oxford Instruments) at
and 10) that were characterized by a substantially lower UMI count a frequency of 5 Hz for 5 min. The resulting images were then back-
and/or higher mitochondrial UMI percentage. Furthermore, to remove ground subtracted, and the relative changes in intensity over time were
low-quality cells in a cell-type-specific way, we removed cells with computed. Activity-based segmentation was performed as previously
UMI counts below the 25th percentile in each cluster. Afterwards, we described64. Spike detection was performed on the intensity traces
repeated gene expression normalization, integration, dimensionality from each region of interest (ROI) to calculate action potential firing
reduction and clustering as described above and identified 16 clusters frequency.
in control organoids (resolution 0.4) and 20 across all organoids
(resolution 0.8). Statistical analysis
No statistical methods were used to pre-determine sample sizes, but
Cluster annotation our sample sizes are similar to those reported in previous publications.
We identified marker genes in control organoids clusters (FindAllMark- Data distribution was assumed to be normal for datasets with a low
ers with only.pos = T) and annotated each cluster leveraging published number of replicates (n < 10), but this was not formally tested, while
references21,60,61. We relied both on overlapping marker genes between non-parametric tests were applied to larger datasets. Relevant details
our control clusters and published references and on label transfer of the statistical tests used for each comparison as well as multiple
results for ref. 21. Label transfer was performed using the standard testing corrections are provided for all relevant panels in the figure
Seurat pipeline. Briefly, a subset of ‘features’ (that is, genes) identified legends and reported below.
by the SelectIntegrationFeatures function with default parameter and
then pairs of ‘anchors’ (that is, cells) between the two datasets have Western blot quantification. Band intensities were quantified using
been identified using the FindTransferAnchors function (normaliza- ImageJ (version 1.52q, functions from Analyze/Gels toolbar). Statistical
tion.method ‘SCT’, reference.assay ‘SCT’, query.assay ‘integrated’, significance of the observed differences between each condition
reduction ‘pcaproject’, dims 1:20, features ‘features’, nn.method ‘rann’, was evaluated, after normalization to GAPDH, with a Benjamini–
eps 0.5) and finally the TransferData function (anchorset ‘anchors’, Hochberg-corrected t-test in RStudio Server v.1.4.1106 (R v.4.1.2).
prediction.assay ‘TRUE’, weight.reduction ‘pcaproject’, dims 1:20,
eps 0.5) was leveraged to score each of the cells in the query dataset Immunostaining quantification. Images were then analysed by quan-
for similarity with labelled cell types in the reference dataset. Finally, tifying the fluorescent signal in each channel with a jlm script within
LabelTransfer scores were inspected and analysed together with marker fiji imagej 2.0.0-rc-69/1.53 f/java1.8.8_172 (64 bit), after background
gene expression to annotate clusters in control organoids. To annotate subtraction with a rolling ball radius of 20. Statistical significance of
clusters in all organoids, we leveraged the annotation of control cells the observed differences between each condition was evaluated with
and used the same annotation for the whole cluster when control cells a Benjamini–Hochberg-corrected rank-sum test, with the pairwise.
accounted for at least 33% of the cluster or at least 10% but accounted wilcox.test, function in RStudio v.1.2.5033 (R v.4.1.2).
for at least 50% of the specific control cluster. The remaining clusters
have been annotated as ‘highly infected’ and ‘hindbrain’ based on the scRNA-seq. Marker genes used for the annotation of each single-cell
expression of viral and HOX genes, respectively. We further removed cluster were identified using a Wilcoxon rank-sum test comparing
control cells falling in ‘highly infected’ clusters as they exhibited the expression of all genes in the cluster of interest with all the
substantially lower UMI counts. remaining cells as implemented in the FindAllMarkers function in the

Seraut R package (v.4.0.4). Differentially expressed genes in the 15. Deguchi, S. et al. SARS-CoV-2 research using human pluripotent
different conditions were identified for selected clusters using a stem cells and organoids. Stem Cells Transl. Med 10, 1491–1499
quasi-likelihood negative binomial generalized linear model as imple- (2021).
mented in the glmQLFit function in the edgeR R package (v.3.36.0). 16. Krenn, V. et al. Organoid modeling of Zika and herpes simplex
Computed P values were adjusted using Bonferroni correction for virus 1 infections reveals virus-specific responses leading to
multiple testing. Enriched gene sets have been identified as imple- microcephaly. Cell Stem Cell https://doi.org/10.1016/j.stem.
mented in the GSEA function in the clusterProfiler R package (v.4.2.2), 2021.03.004 (2021).
and only pathways with a Benjamini–Hochberg-adjusted P value lower 17. Cairns, D. M. et al. A 3D human brain–like tissue model of herpes-
than 0.05 were considered as significantly enriched. induced Alzheimer’s disease. Sci. Adv. 6, eaay8828 (2020).
18. Qiao, H. et al. Herpes simplex virus type 1 infection leads to
Reporting summary neurodevelopmental disorder-associated neuropathological
Further information on research design is available in the Nature Port- changes. PLoS Pathog. https://doi.org/10.1371/journal.ppat.
folio Reporting Summary linked to this article. 1008899 (2020).
19. Nowakowski, T. J. et al. Spatiotemporal gene expression
Data availability trajectories reveal developmental hierarchies of the human
Single-cell and bulk RNA-seq data from human brain organoids infected cortex. Science 358, 1318–1323 (2017).
with HSV-1, at different timepoints post infection, are available at the 20. Su, X. et al. Human brain organoids as an in vitro model system of
National Center for Biotechnology Information under Gene Expres- viral infectious diseases. Front. Immunol. 12, 792316 (2021).
sion Omnibus identifier GSE163952. Source data are provided with 21. Kanton, S. et al. Organoid single-cell genomic atlas uncovers
this paper. human-specific features of brain development. Nature 574,
418–422 (2019).
Code availability 22. Camp, J. G. et al. Human cerebral organoids recapitulate gene
Scripts for the analysis of single-cell and bulk RNA-seq data and the expression programs of fetal neocortex development. Proc. Natl
generation of the related figure plots are publicly available on GitHub Acad. Sci. USA 112, 15672–15677 (2015).
(https://github.com/rajewsky-lab/HSV1_cerebral_organoids). 23. Qian, X. et al. Sliced human cortical organoids for modeling
distinct cortical layer formation. Cell Stem Cell 26, 766–781.e9
References (2020).
1. Thellman, N. M. & Triezenberg, S. J. Herpes simplex virus 24. Sandbaumhüter, M. et al. Cytosolic herpes simplex virus capsids
establishment, maintenance, and reactivation: in vitro modeling not only require binding inner tegument protein pUL36 but
of latency. Pathogens 6, 28 (2017). also pUL37 for active transport prior to secondary envelopment:
2. Marcocci, M. E. et al. Herpes simplex virus-1 in the brain: the dark the role of pUL36 and pUL37 for HSV1 capsid transport.
side of a sneaky infection. Trends Microbiol. 28, 808–820 (2020). Cell. Microbiol. 15, 248–269 (2013).
3. Piret, J. & Boivin, G. Immunomodulatory strategies in herpes simplex 25. Acuña-Hinrichsen, F. et al. Herpes simplex virus type 1 neuronal
virus encephalitis. Clin. Microbiol. Rev. 33, e00105–e00119 (2020). infection triggers disassembly of key structural components of
4. Rozenberg, F. Herpes simplex virus and central nervous system dendritic spines. Front. Cell Neurosci. https://doi.org/10.3389/
infections: encephalitis, meningitis, myelitis. Virologie 24, fncel.2021.580717 (2020).
283–294 (2020). Oct 1. 26. Acuña-Hinrichsen, F. et al. Herpes simplex virus type 1 enhances
5. D’Aiuto, L. et al. Modeling herpes simplex virus 1 infections in expression of the synaptic protein arc for its own benefit.
human central nervous system neuronal cells using two- and Front. Cell Neurosci. 12, 505 (2018).
three-dimensional cultures derived from induced pluripotent 27. Yin, J. & VanDongen, A. M. Enhanced neuronal activity and
stem cells. J. Virol. https://doi.org/10.1128/jvi.00111-19 (2019). asynchronous calcium transients revealed in a 3D organoid
6. Sehl, J. et al. An improved animal model for herpesvirus model of Alzheimer’s disease. ACS Biomater. Sci. Eng. 7, 254–264
encephalitis in humans. PLoS Pathog. https://doi.org/10.1371/ (2021).
2Fjournal.ppat.1008445 (2020). 28. Krus, K. L. et al. Loss of Stathmin-2, a hallmark of TDP-43-
7. Lancaster, M. A. et al. Cerebral organoids model human brain associated ALS, causes motor neuropathy. Cell Rep. 39,
development and microcephaly. Nature 501, 373–379 (2013). 111001 (2022).
8. Qian, X. et al. Generation of human brain region-specific 29. Melamed, Z. et al. Premature polyadenylation-mediated
organoids using a miniaturized spinning bioreactor. Nat. Protoc. loss of stathmin-2 is a hallmark of TDP-43-dependent
13, 565–580 (2018). neurodegeneration. Nat. Neurosci. 22, 180–190 (2019).
9. Quadrato, G. et al. Cell diversity and network dynamics in 30. Kadowaki, M. et al. N-cadherin mediates cortical organization in
photosensitive human brain organoids. Nature 545, 48–53 (2017). the mouse brain. Dev. Biol. 304, 22–33 (2007).
10. Dang, J. et al. Zika virus depletes neural progenitors in human 31. Wyler, E. et al. Widespread activation of antisense transcription
cerebral organoids through activation of the innate immune of the host genome during herpes simplex virus 1 infection.
receptor TLR3. Cell Stem Cell 19, 258–265 (2016). Genome Biol. 18, 209 (2017).
11. Garcez, P. P. et al. Zika virus impairs growth in human 32. Tunkel, A. R. et al. The management of encephalitis: clinical
neurospheres and brain organoids. Science 352, 816–818 (2016). practice guidelines by the Infectious Diseases Society of America.
12. Qian, X. et al. Brain-region-specific organoids using mini- Clin. Infect. Dis. 47, 303–327 (2008).
bioreactors for modeling ZIKV exposure. Cell 165, 1238–1254 33. Song, B., Liu, J. J., Yeh, K. C. & Knipe, D. M. Herpes simplex virus
(2016). infection blocks events in the G1 phase of the cell cycle. Virology
13. Watanabe, M. et al. Self-organized cerebral organoids with 267, 326–334 (2000).
human-specific features predict effective drugs to combat Zika 34. Squair, J. W. et al. Confronting false discoveries in single-cell
virus infection. Cell Rep. 21, 517–532 (2017). differential expression. Nat. Commun. 12, 5692 (2021).
14. Zhang, B. et al. Differential antiviral immunity to Japanese 35. Dostert, C., Grusdat, M., Letellier, E. & Brenner, D. The TNF family
encephalitis virus in developing cortical organoids. Cell Death Dis. of ligands and receptors: communication modules in the immune
9, 719 (2018). system and beyond. Physiol. Rev. 99, 115–160 (2019).

36. Pahl, H. L. Activators and target genes of Rel/NF-κB transcription 58. Haghverdi, L., Lun, A. T. L., Morgan, M. D. & Marioni, J. C. Batch
factors. Oncogene 18, 6853–6866 (1999). effects in single-cell RNA-sequencing data are corrected by
37. Wang, Y. Y., Yang, Y. X., Zhe, H., He, Z. X. & Zhou, S. F. Bardoxolone matching mutual nearest neighbors. Nat. Biotechnol. 36, 421–427
methyl (CDDO-Me) as a therapeutic agent: an update on its (2018).
pharmacokinetic and pharmacodynamic properties. Drug Des. 59. Stuart, T. et al. Comprehensive integration of single cell data. Cell
Dev. Ther. 8, 2075–2088 (2014). https://doi.org/10.1016/j.cell.2019.05.031 (2019).
38. Smith, I. L., Hardwicke, M. A. & Sandri-Goldin, R. M. Evidence 60. Velasco, S. et al. Individual brain organoids reproducibly form
that the herpes simplex virus immediate early protein ICP27 cell diversity of the human cerebral cortex. Nature 570, 523–527
acts post-transcriptionally during infection to regulate gene (2019).
expression. Virology 186, 74–86 (1992). 61. Uzquiano, A. et al. Proper acquisition of cell class identity in
39. Furman, D. et al. Chronic inflammation in the etiology of disease organoids allows definition of fate specification programs of the
across the life span. Nat. Med 25, 1822–1832 (2019). human cerebral cortex. Cell 185, 3770–3788 (2022).
40. Mancini, M. & Vidal, S. M. Insights into the pathogenesis of herpes 62. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a
simplex encephalitis from mouse models. Mamm. Genome 29, Bioconductor package for differential expression analysis
425–445 (2018). of digital gene expression data. Bioinformatics 26, 139–140
41. Boivin, N., Menasria, R., Piret, J., Rivest, S. & Boivin, G. The (2010).
combination of valacyclovir with an anti-TNF alpha antibody 63. Wu, T. et al. clusterProfiler 4.0: a universal enrichment tool for
increases survival rate compared to antiviral therapy alone in a interpreting omics data. Innovation 2, 100141 (2021).
murine model of herpes simplex virus encephalitis. Antivir. Res 64. Junek, S., Chen, T. W., Alevra, M. & Schild, D. Activity correlation
100, 649–653 (2013). imaging: visualizing function and structure of neuronal
42. Sergerie, Y., Boivin, G., Gosselin, D. & Rivest, S. Delayed but populations. Biophys. J. 96, 3801–3809 (2009).
not early glucocorticoid treatment protects the host during
experimental herpes simplex virus encephalitis in mice. J. Infect. Acknowledgements
Dis. 195, 817–825 (2007). We thank M. Korshevniuk and B. Beyersdorf for the initial data
43. Qian, X., Song, H. & Ming, G. L. Brain organoids: advances, analysis, R. Pareja, for the technical support with iPS cell culture
applications and challenges. Development 146(Apr 16), and western blotting, A. Boltengagen and S. Ayoub for the technical
dev166074 (2019). support with scRNA-seq, Ż. Zarębska for help with western blotting
44. Enquist, L. W. & Leib, D. A. Intrinsic and innate defenses of neurons: during revision, and M. Herzog and A. Tschernycheff for the lab
détente with the herpesviruses. J. Virol. 91, e01200–e01216 (2017). organizational support. We thank C. Quedenau, A. K. Secener,
45. Wang, X. et al. Genome-wide analysis of PDX1 target genes in T. Borodina and T. Conrad from the BIMSB Genomics platform for
human pancreatic progenitors. Mol. Metab. 9, 57–68 (2018). performing sequencing runs. We thank S. Diecke for providing and
46. Snijder, B. et al. Single‐cell analysis of population context biobanking iPS cell lines, and H. Lickert (for providing the XM001 iPS
advances RNAi screening at multiple levels. Mol. Syst. Biol. 8, 579 cell line. We thank A. Whisnant and L. Dölken for providing the ICP27
(2012). knockout virus and the corresponding KOS wild-type virus and the
47. Wurmus, R. et al. PiGx: reproducible genomics analysis pipelines Vero-ICP27 complementing cell line. We thank A. Cliffe, L. Dölken,
with GNU Guix. GigaScience 7, giy123 (2018). A. Whisnant, T. Hennig, C. Scheidereit, K. Ruprecht, P. Körtvélyessy,
48. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. D. Reinhold and M. Kolesnichenko for discussions and support. We
Bioinformatics 29, 15–21 (2013). also thank all the members of the N. Rajewsky laboratory for critical
49. Anders, S., Pyl, P. T. & Huber, W. HTSeq—a Python framework to and useful discussions. This work was supported by: BIH Grant:
work with high-throughput sequencing data. Bioinformatics 31, Multiscale Omics Projects (M.L). T.M.P. is financially supported by the
166–169 (2015). Berlin School of Integrative Oncology through the GSSP programme
50. Frankish, A. et al. GENCODE reference annotation for the human of the German Academy of Exchange Service (DAAD) and by the
and mouse genomes. Nucleic Acids Res. 47, D766–D773 (2019). Add-On Fellowship of the Joachim Herz Foundation. The funders
51. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold had no role in study design, data collection and analysis, decision to
change and dispersion for RNA-seq data with DESeq2. Genome publish or preparation of the manuscript.
Biol. 15, 550 (2014).
52. Seal, R. L. et al. Genenames.org: the HGNC resources in 2023. Author contributions
Nucleic Acids Res. 51, D1003–D1009 (2023). Conceptualization: A.R.-W., E.W. and N.R. Methodology: A.R.-W. and
53. Maynard, K. R. et al. Transcriptome-scale spatial gene expression E.W. Calcium-based imaging: A.O.M. and A.W. Western blot: A.R.-W. and
in the human dorsolateral prefrontal cortex. Nat. Neurosci. 24, A.L. Virus production: E.W. and B.B.K. Computational analysis: T.M.P.,
425–436 (2021). I.L. and S.J.K. (scRNA-seq), P.G. and T.M.P (bulk RNA-seq) and E.W.,
54. Macosko, E. Z. et al. Highly parallel genome-wide expression S.J.K. and T.M.P (spatial VISIUM). Writing—original draft: A.R.-W.
profiling of individual cells using nanoliter droplets. Cell 161, Writing—review and editing: I.L., E.W., P.G., M.L., T.M.P., A.L. and N.R.
1202–1214 (2015). All authors discussed and analysed the data. Supervision: N.R. and M.L.
55. Sztanka-Toth, T. R., Jens, M., Karaiskos, N. & Rajewsky, N.
Spacemake: processing and analysis of large-scale spatial Funding
transcriptomics data. GigaScience https://doi.org/10.1093/ Open access funding provided by Max-Delbrück-Centrum für
gigascience/giac064 (2022). Molekulare Medizin in der Helmholtz-Gemeinschaft (MDC).
56. Satija, R., Farrell, J. A., Gennert, D., Schier, A. F. & Regev, A.
Spatial reconstruction of single-cell gene expression data. Competing interests
Nat. Biotechnol. 33, 495–502 (2015). The authors declare no competing interests.
57. Hafemeister, C. & Satija, R. Normalization and variance stabiliza
tion of single-cell RNA-seq data using regularized negative Additional information
binomial regression. Genome Biol. https://doi.org/10.1186/ Extended data is available for this paper at
s13059-019-1874-1 (2019). https://doi.org/10.1038/s41564-023-01405-y.

Supplementary information The online version contains supplementary Open Access This article is licensed under a Creative Commons
material available at https://doi.org/10.1038/s41564-023-01405-y. Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format,
Correspondence and requests for materials should be addressed to as long as you give appropriate credit to the original author(s) and the
Agnieszka Rybak-Wolf or Nikolaus Rajewsky. source, provide a link to the Creative Commons license, and indicate
if changes were made. The images or other third party material in this
Peer review information Nature Microbiology thanks Deepak Shukla article are included in the article’s Creative Commons license, unless
and the other, anonymous, reviewer(s) for their contribution to the indicated otherwise in a credit line to the material. If material is not
peer review of this work. Peer reviewer reports are available. included in the article’s Creative Commons license and your intended
use is not permitted by statutory regulation or exceeds the permitted
Reprints and permissions information is available at use, you will need to obtain permission directly from the copyright
www.nature.com/reprints. holder. To view a copy of this license, visit http://creativecommons.
org/licenses/by/4.0/.
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations. © The Author(s) 2023

Extended Data Fig. 1 | Cerebral organoids development at single cell and (Rep) showing overlap between two biological replicates from control 60 days
spatial resolution. a, Representative images from the Resolve Biosciences’ old cerebral organoids derived from induced Pluripotent Stem Cell (iPSC) line
Molecular Cartography analysis of 60 days old organoid from iPSC line 2 for 1 (upper panel) and iPSC line 2 (lower panel). Scale bar - 100 µm d, UMAP plots
markers: EOMES – intermediate progenitors, SOX2/PAX6 – progenitors, TBR1 depicting normalized expression of cell-type specific markers: VIM, SOX2-
- deeper layer cortical neurons, BCL11B- upper layer cortical neurons, TUBB3- progenitors, FOXG1- telencephalic neurons and progenitors, MKI67-proliferating
neurons. Representative image of one experiment, n=2 biological replicates. radial glia, DCX- neurons, EOMES-intermediate progenitors, NEUROD2, NEUROD6
Scale bar - 100 µm. b, High magnification of exemplary immunostaining of – cortical neurons, BCL11B- upper layer cortical neurons, TCF7L2- Thalamic
60 days old cerebral brain organoids showing expression of TBR1 - deeper layer neurons, GFAP, SPARCL1-astroglia, RSPO2, RSPO3- cortical hem, TTR-choroid
cortical neurons, SATB2- upper layer cortical neurons, SOX2- progenitors, TUJ1/ plexus, DCT- retinal pigmented epithelium, SFRP2- retinal progenitors, COL1A2
MAP2 – neurons, GFAP- astroglia, ZO-1- neuroepithelial junctions. Representative – mural cells, RELN- Cajal Retzius neurons. All expression values are clipped to 2
of three independent experiments (n=3 independent organoids/ iPSC line / for a better and homogenous color representation. (Two independent iPSC lines,
experiment). Scale bar - 100 µm. c, UMAP plots with cells colored by replicate each n=2 biological replicates).
Nature Microbiology
Extended Data Fig. 2 | HSV-1 driven impairment of synaptic activity. image of three independent experiments, n=3 biological replicates. e, Mosaic
a, Spatial transcriptomics analysis of HSV-1 GFP expression in uninfected image of the footprint of an uninfected organoid (CTRL) and HSV-1-GFP
(CTR) as well as 1 and 3 days post infection (dpi). Colors indicate log2 GFP infected organoid (HSV-1) used in calcium imaging experiments showing TUJ1
mRNA expression. Representative image of one experiment, n=4 biological expression (red), HSV-1-GFP expression (green) and DAPI (blue). Scale bar – 1 mm.
replicates. b, Fold changes (Log2FoldChange) of gene expression between Representative image of one experiment, n=3 independent organoids/condition
control (CTR) and infected organoids 3 days post infection (+HSV-1 3dpi) versus were analyzed. f, Line chart showing the progression of the relative activity based
log10 normalized mean transcript counts (baseMean). HSV-1 transcripts (green) on the mean spike frequency of uninfected (CTRL) and infected (+HSV-1) after
and ‘CHEMICAL SYNAPTIC TRANSMISSION’ genes (red) are highlighted. Data 24 and 48 hours post-infection (hpi), n=3 independent replicates/condition,
from two independent iPSC lines, n=3 biological replicates/condition/iPSC three measurements per timepoint. Error bars show the mean ± s.d. g, Exemplary
line (see Methods). c, Exemplary immunohistochemistry for synaptic (SYN1 – images of HSV-1 GFP virus expression in sectioned infected and ACV treated
black, magenta) neuronal, nuclei (DAPI – black, blue) and HSV-1 (GFP - white) in organoids, immunostained with anti-GFP antibody. Infected (HSV-1 3dpi), ACV
uninfected and infected organoids. Scale bar - 100 µm. Representative image of treated organoids (HSV-1 3dpi + ACV). Right panel: merged images, GFP (white),
three independent experiments (n=3 independent replicates). d, Representative DAPI (blue). Scale bar - 500 µm. Representative image of three independent
western blots of synaptic proteins (HOMER1, SYN1) and a housekeeping protein experiments (n=3 biological replicates).
(GAPDH) of uninfected (CTRL) and infected (HSV-1) organoids. Representative
Nature Microbiology
Extended Data Fig. 3 | Acyclovir (ACV) treatment in control and HSV-1 UMAP) and ACV treated control organoids (right UMAP). Colors are mapped to
infected organoids. a, Exemplary images of HSV-1 GFP virus expression in the clusters (annotated on the right). c, UMAP plots of cells colored by replicate
sectioned infected and ACV treated organoids, immunostained with anti-GFP (Rep.) from uninfected organoids (CTRL), 1 day and 3 days post infection (HSV-1
antibody. Infected (HSV-1 3dpi), ACV treated organoids (HSV-1 3dpi + ACV). 1,3 dpi), and acyclovir treated, 3 days post infection (dpi) organoids derived from
Right panel: merged images, GFP (white), DAPI (blue). Scale bar - 500 µm. induced Pluripotent Stem Cell (iPSC) line 1 (left) and iPSC line 2 (right panel), n=2
Representative image of three independent experiments (n=3 independent independent replicates/condition/iPSC line.
replicates). b, UMAP plots of single cell RNA-seq data in control organoids (left
Nature Microbiology
Extended Data Fig. 4 | See next page for caption.
Nature Microbiology
Extended Data Fig. 4 | Cell-type specific responses in HSV-1 infected and and both by condition and cluster assignment. Cells are coloured by Condition,
Acyclovir (ACV) treated organoids. a, Proportion of cell types in uninfected, Clusters and Viral load. Data from two independent iPSC lines, each with n=2
infected(1, 3dpi) and ACV treated organoids derived from iPSC line 1 (upper) biological replicates. d, Barplot showing the proportion of cells from each
and iPSC line 2 (lower panel). b, Representative immunohistochemistry for infected condition assigned to each Highly infected cluster. e, Dotplot showing
mitotic marker (pVIM – white) and merged images for mitotic marker (pVIM the expression of selected marker genes in each Highly infected cluster. Color
- green), HSV-1 (GFP - white), nuclei (DAPI - blue) in uninfected (CTRL), 3 dpi indicates average, scaled expression values and dot size shows the percentage
infected (HSV-1) and ACV treated organoids. Scale bar - 100 µm. Representative of expressing cells in each Highly infected cluster. f, Violin plot showing the viral
image of three experiment (n=3 biological replicates). c, Viral gene expression load (% of viral transcripts, that is Unique Molecular Identifiers or UMI) in each
heatmap showing log10 expression values of detected viral genes. Genes (rows) Highly infected cluster. Each dot represents a cell. (All data derived from two
are ordered by decreasing overall expression and cells (columns) are ordered independent iPSC lines, each n=2, biological replicates).
Nature Microbiology
Extended Data Fig. 5 | Viral infection induces inflammatory responses. cluster. Only highly infected clusters and clusters included in the differential
a, Dotplot showing the expression of detected genes from the ‘TNF-α signaling gene expression and gene set enrichment analysis are shown. Data pooled from
via NF-κB’ HALLMARK gene set. Color indicates average, SCT-normalized two independent iPSC lines, each n=2, biological replicates.
expression values and dot size shows the percentage of expressing cells in each
Nature Microbiology
Extended Data Fig. 6 | Acyclovir (ACV) treatment does not prevent Data from two independent iPSC lines, n=2, biological replicates/condition/
inflammatory responses. a, TNF-α pathway activity score in each condition: iPSC line. b, UMAP plots of cells colored by the TNF pathway activity score in
uninfected (CTRL), uninfected ACV treated (CTRL +ACV), HSV-1-infected (HSV-1 each cell and separated in uninfected organoids, 1 day and 3 days post infection
1dpi, HSV-1 3dpi), and 3 dpi HSV-1-infected ACV-treated (HSV-1 3dpi +ACV) (dpi), and acyclovir treated, 3 days post infection (dpi) organoids, with red and
organoids. Boxplot centers represent median values, while the boxplot bounds blue indicating increased and decreased score, respectively. Data from two
represent the 25% and 75% quantiles. Boxplot whiskers represent the 25% quantile independent iPSC lines, n=2, biological replicates/condition/iPSC line.
−1.5× interquartile range (IQR) and the 75% quantile +1.5× IQR, respectively.
Nature Microbiology
Extended Data Fig. 7 | TNF ligands are upregulated upon infection. a, Fold by acyclovir treatment. Row-mean normalized expression of upregulated
changes (Log2FoldChange) of gene expression between control (CTR) and TNF ligands in scRNA-seq data (pseudo-bulk). Transcript counts are shown
infected organoids 3 days post infection (+HSV-1 3dpi) versus log10 normalized for 2 replicates per condition (group) and two cell lines (line). c, Exemplary
mean transcript counts (baseMean). HSV-1 transcripts (green) and TNF ligands images of Cleaved Caspase 3 and HSV-1 GFP virus expression in sectioned and
(red) are highlighted. Significantly differentially expressed TNF ligands with immunostained organoids, uninfected (Mock) infected with HSV-1 (HSV-1 1dpi,
a log2 fold change > 2.5 are labelled. Transcript counts were measured from HSV-1 3dpi) and ACV treated (HSV-1 3dpi+ACV). Right panel: merged images,
two independent iPSC lines, n=3 biological replicates/condition/iPSC line Cleaved Caspase 3(red), GFP (white), DAPI (blue). Scale bar - 100 µm, n=3
(see Methods). b, TNF ligands induced upon infection are not downregulated biological replicates/condition.
Nature Microbiology
Nature Microbiology
Extended Data Fig. 8 | Combinatorial treatment strategies for HSV-1- driven panel: merged images, ZO-1(red), GFP (white), DAPI (blue). Scale bar - 100 µm.
neuroinflammation. a, Exemplary images of HSV-1 GFP virus expression in Representative image of three independent experiments with two iPSC lines.
sectioned and anti-GFP antibody immunostained organoids. Uninfected (CTRL), d, e, ZO-1 fluorescent signal quantification in sectioned and immunostained
infected (HSV-1), ACV treated organoids (ACV), ACV and CDDO-Me co-treated organoids infected with high (d) or low (e) multiplicity of infection (MOI) of
(ACV+CDDO-Me) and ACV and NEC-1 cotreated (ACV+NEC-1). Lower panel: HSV-1, each dot represents one neuroepithelial loop, colors indicated cell
merged images, GFP (white), DAPI (blue). Scale bar - 500 µm. Representative lines: iPSC line 1 (dark blue), iPSC line 2 (bright blue). Lower panel, p-values
image of three independent experiments with two iPSC lines (n=3 biological for the fluorescent signal quantification f, g HSV-1-GFP signal quantification
replicates/condition/iPSC line). b, Scatterplots showing all comparisons in immunostained organoids infected with high (f) or low (g) multiplicity of
between viral GFP RNA, measured as Delta Ct with respect to the human DICER1 infection (MOI) of HSV-1 virus, each dot represents one neuroepithelial loop,
gene in cDNA from whole organoids lysate, viral GFP gDNA, measured as before colors indicated cell lines. Lower panel: p-values (Benjamini-Hochberg-corrected
in genomic DNA from the same organoids, and viral GP protein, measured from pairwise Wilcoxon Mann Whitney test for the fluorescent signal quantification,
Western blot densitometric analysis, normalized on GAPDH. All correlation n=14,34,32,17,23 individual neuroepithelial loops quantified from 3,5,6,4,4
coefficients (Pearson) are shown in the top left of each plot. Sample identity individual organoids per condition for panels d and f, n=43,31,33,22,21,38,29
is indicated next to each dot, representing the mean between three biological loops from 5,3,3,2,3,3,3 organoids for panels e and g, from two independent
replicates (n=3, biological replicates). c, Exemplary images of ZO-1 and HSV-1 hiPSC lines and two MOIs. Boxplot show lower quartile, median and upper
GFP virus expression in sectioned and immunostained organoids, infected with quartile, whiskers represent upper and lower quartile ± the smallest value within
high (MOI=1) and low (MOI=0,2) multiplicity of infection (MOI) of HSV-1. Lower 1.5-fold of the interquartile range respectively.
Nature Microbiology
Extended Data Fig. 9 | Combinatorial treatment reduces HSV-1- driven ACV and NEC-1 cotreated (HSV-1+ACV+NEC-1). GAPDH serves as loading control
neuroinflammation. a, Scatterplot showing the comparisons between viral and GFP indicates virus expression (lower panel). Number of replicates is shown
GFP protein measured from Western blot densitometric analysis (normalized in panel c. c, Semi-quantification of STMN2 (right panel) and GFP (left panel)
on GAPDH) or immunofluorescence (signal intensity minus background, see protein expression by densitometry in uninfected (CTRL), infected (HSV-1),
methods), in green for ‘low MOI’ infections and red for ‘high MOI’ infections. ACV treated organoids (+ACV), ACV and CDDO-Me cotreated (+ACV+CDDO-
Western blot was performed on ‘low MOI’ only. Sample identity is indicated Me) and ACV and NEC-1 cotreated (+ACV+NEC-1). P-values are indicated for the
next to each dot, representing the mean between all biological replicates from relevant comparisons, according to Benjamini-Hochberg-corrected two-sided
Supplementary Fig. 5C–F (IF) and Fig. 6B (WB). b, Representative western blots t-tests (n=4 biologically independent samples for CTRL, HSV1+CDDO-Me and
analysis of STMN2 expression in uninfected (CTRL), infected (HSV-1), ACV treated HSV1+NEC-1, n=6 for all others, that is 4-6 batches of pooled organoids; color
organoids (HSV-1+ACV), CDDO-Me treated (HSV-1+ CDDO-Me), NEC-1 treated indicates cell line of origin).Bar plots and error bars shown ± s.e.
(HSV-1+ NEC-1), ACV and CDDO-Me co-treated (HSV-1+ACV+CDDO-Me); and
Nature Microbiology
Nature Microbiology
Extended Data Fig. 10 | Infected organoids release signaling molecules in uninfected mock organoids, infected with HSV-1- WT virus and HSV-1-ICP27
capable of NF-κB pathway activation independently from viral infection. mutant virus. GAPDH -loading control, ICP0 -virus expression, n=3 biological
a, Exemplary western blots analysis of p-P65 expression in uninfected (MOCK), replicated. Samples were processed in parallel on several blots (see source data
mock treated (ACV, CDDO-Me, NEC-1), infected (HSV-1), infected and treated ED Fig. 10f). g, A scheme of experimental set-up. h, Exemplary images of HSV-1
(HSV-1+ACV, HSV-1+ CDDO-Me, HSV-1+ NEC-1) organoids. GAPDH – loading GFP expression in infected organoids or organoids exposed to the medium (2x
control and GFP -virus expression, n=3 independent experiments. b, Exemplary 0.1 µm filtered) collected from 3dpi infected organoids. i, Exemplary western
western blots analysis of p-P65 expression in uninfected (MOCK), infected blots analysis of p-P65 expression in uninfected mock (MOCK), HSV-1-GFP
(HSV-1), and organoids exposed to irradiated virus (HSV-1 IRR). GAPDH -loading infected and organoids exposed to mock or infected organoid medium (MOCK
control, GFP -virus expression (n=3 independent experiments). c, Exemplary medium, HSV-1 medium, respectively). GAPDH - loading control and GFP - virus
western blots analysis of p-P65 expression in uninfected (MOCK) and HSV-1- expression (lower panel). j, Semi-quantification of p-P65 (right panel) and GFP
WT virus (3dpi and 6dpi) or HSV-1-ICP27 mutant virus (3dpi and 6dpi) infected (left panel) protein expression by densitometry in uninfected mock (MOCK),
organoids. GAPDH – control, ICP0 - virus expression. n=3 independent HSV-1-GFP infected and organoids exposed to mock or infected organoid
experiments. d, e, Exemplary images of ZO-1 (d) and STMN2 (e) and HSV-1 medium (MOCK medium, HSV-1 medium, respectively). Data were normalized
ICP0 virus expression in: uninfected, mock treated organoids, and HSV-1- WT on the same sample in each batch. P-values are indicated according to Benjamini-
virus (3dpi) or HSV-1-ICP27 mutant virus (6dpi) infected organoids. Lower Hochberg-corrected paired two-sided t-tests (n=4 biologically independent
panel: merged images, ZO-1/STMN2 (green), ICP0 (white), DAPI (blue). Scale samples coming from n=3 batches of medium preparation, indicated by color).
bar - 100 µm. Representative of more than three independent experiments. f, Barplots and error bars show mean ± s.e.
Exemplary western blots analysis of p-P65, pAKT, AKT, TRAF6, ERK1/2 expression
Nature Microbiology

Modelling Viral Encephalitis Caused by Herpes Simplex Virus 1 Infection in Cerebral Organoids

Uploaded by

Copyright:

Available Formats

Modelling Viral Encephalitis Caused by Herpes Simplex Virus 1 Infection in Cerebral Organoids

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modelling Viral Encephalitis Caused by Herpes Simplex Virus 1 Infection in Cerebral Organoids

Uploaded by

Copyright:

Available Formats

nature microbiology

Modelling viral encephalitis caused

Received: 22 July 2022 Agnieszka Rybak-Wolf 1,17 , Emanuel Wyler 2,17,

Check for updates

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1252

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1253

iPS cell Day 0 Day 5 Day 11 Day 30

Immature cortical neurons 1

SOX2 ZO-1 ZO-1/SOX2/DAPI

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1254

a b CTRL HSV-1 (1 dpi) HSV-1 (3 dpi)

HSV-1 (GFP protein)

HSV-1 infection Validations

0.10 80 0.2 0.10 35 0.2

e STMN2 NEUN HSV-1 Merge f ZO-1 HSV-1–GFP Merge

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1255

Thalamic neurons 2 0.0 0.3 3.5 0.8

Highly infected Cortical neurons 0.0 1.6 1.1 0.7

Hindbrain neurons 1 0.0 1.5 2.3 1.2

Hindbrain neurons 2 0.0 0.4 1.8 0.7 60

Retinal Retinal pigmented cells 0.0 0.7 1.3 0.4

Astroglia 0.0 0.3 1.9 0.6

Highly infected 2 NA 46.2 5.2 3.2

Highly infected 4 0 NA 58.7 23.6 4.4

d CTRL HSV-1 (1 dpi) HSV-1 (3 dpi) HSV-1 (3 dpi) +ACV

UMAP1 UMAP1 UMAP1 UMAP1

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1256

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1257

a −log10 adj. P value

TNF pathway score

HSV-1 (3 dpi) + ACV

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1258

8–12 h 3 days Neuroepithelial integrity (ZO-1)

c CTRL Mock ACV CDDO-Me NEC-1 ACV + CDDO-Me ACV + NEC-1

d CTRL Mock ACV CDDO-Me NEC-1 ACV + CDDO-Me ACV + NEC-1

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1259

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1260

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1261

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1262

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1263

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1264

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1265

Nature Microbiology | Volume 8 | July 2023 | 1252–1266 1266

Extended Data Fig. 4 | See next page for caption.

Extended Data Fig. 8 | See next page for caption.

Extended Data Fig. 10 | See next page for caption.

You might also like