Abstract

   Astrocytes are central to the pathogenesis of multiple sclerosis (MS);
   however, their regulation by post-translational ubiquitination and
   deubiquitination is unresolved. This study shows that the
   deubiquitinating enzyme OTUD7B in astrocytes protects against murine
   experimental autoimmune encephalomyelitis (EAE), a model of MS, by
   limiting neuroinflammation. RNA-sequencing of isolated astrocytes and
   spatial transcriptomics show that in EAE, OTUD7B downregulates
   chemokine expression in astrocytes of inflammatory lesions, which is
   associated with reduced recruitment of encephalitogenic CD4^+ T cells.
   Furthermore, OTUD7B is necessary for glial fibrillary acidic protein
   (GFAP) expression of astrocytes bordering inflammatory lesions.
   Mechanistically, OTUD7B (i) restricts TNF-induced chemokine production
   of astrocytes by sequential K63- and K48-deubiquitination of RIPK1,
   which limits NF-κB and MAPK activation and (ii) enables GFAP protein
   expression by supporting GFAP mRNA expression and preventing its
   proteasomal degradation through K48-deubiquitination of GFAP. This dual
   action on TNF signaling and GFAP identifies OTUD7B as a central
   inhibitor of astrocyte-mediated inflammation.

   Subject terms: Autoimmunity, Ubiquitylated proteins
     __________________________________________________________________

   Astrocytes drive multiple sclerosis (MS) by enhancing inflammation.
   Here, the authors show that the deubiquitinase OTUD7B in astrocytes
   protects mice from experimental autoimmune encephalomyelitis (EAE), a
   model of MS, by reducing inflammation and stabilizing GFAP thereby
   physically restricting inflammation.

Introduction

   Astrocytes are a highly abundant cell population present in all regions
   of the central nervous system (CNS) and play an important role in the
   maintenance of CNS homeostasis and health. Under physiological
   conditions, astrocytes provide physical and metabolic support for
   neurons and blood-brain barrier (BBB) functions^[52]1–[53]3. Upon CNS
   damage, astrocytes are rapidly activated and undergo structural and
   functional changes, by a process known as reactive astrogliosis which
   critically regulates the pathogenesis, development and outcome of CNS
   disorders including autoimmune diseases, infections, neurodegenerative
   diseases and trauma^[54]4. Reactive astrogliosis is characterized by
   astrocyte hypertrophy and increased STAT3-mediated expression of the
   glia fibrillary acidic protein (GFAP)^[55]5.

   Depending on the underlying disease and the inflammatory milieu
   astrocytes can either support or suppress CNS inflammation, contribute
   to CNS damage but also promote regeneration. This functional plasticity
   of astrocytes is based on the activation of different signaling
   pathways leading to the respective production and secretion of
   disease-modifying proteins. In multiple sclerosis (MS), a human
   inflammatory demyelinating disease and its murine model experimental
   autoimmune encephalitis (EAE), reactive astrocytes upregulate GFAP
   protein expression and are the major producers of
   chemokines^[56]6,[57]7. These chemokines foster the recruitment of
   immune cells, primarily autoimmune CD4^+ T cells to the site of
   inflammation. The CNS infiltrating CD4^+ T cells orchestrate an attack
   on the myelin sheath resulting in demyelination and
   neurodegeneration^[58]8,[59]9. The production of chemokines by
   astrocytes in MS and EAE is driven by several signaling pathways,
   including NF-κB, MAPK and JAK-STAT pathways^[60]10–[61]12. Experimental
   studies have also shown that astrocyte-specific ablation of NF-κB
   signaling molecules attenuates EAE^[62]11,[63]12. In addition to
   disease promoting functions, astrocytes can exert protective roles, in
   particular, at later stages of EAE and MS. These protective mechanisms
   include the local restriction and resolution of neuroinflammation as
   well as the promotion of remyelination and axonal repair^[64]13–[65]15.
   In this regard, STAT3-mediated astrocyte survival, proliferation and
   up-regulation of GFAP expression ameliorates EAE by limiting the spread
   of encephalitogenic T cells and bordering of inflammatory
   lesions^[66]16. Depending on the subtype of MS and the EAE model,
   active and resolving inflammatory lesions exist in parallel throughout
   the CNS. In these different lesions astrocyte reactivity is diverse,
   shows a high plasticity, differs regionally and may also change over
   time^[67]17–[68]19.

   Astrocyte reactivity is significantly determined by the differential
   activation of signaling pathways which regulate inflammatory reactions
   and glia scar formation. These signaling pathways are critically
   modulated by ubiquitination, a posttranslational modification mediated
   by the covalent linkage of ubiquitin, a 76-aa large protein, to
   substrates. Ubiquitination is catalyzed sequentially by an
   ubiquitin-activating enzyme (E1), an ubiquitin-conjugating enzyme (E2),
   and an ubiquitin ligase (E3). Ubiquitin molecules can be added to
   proteins in the form of monomers and polyubiquitin chains, in which
   ubiquitins are linked through the N-terminal methionine residue (M1) or
   one of the seven lysine residues (K6, K11, K27, K29, K33, K48,
   K63)^[69]20. The fate of ubiquitinated substrates is determined by the
   type of ubiquitin linkage. K48- and K11 chains lead to proteasomal
   degradation of substrates, whereas K63-linked ubiquitin chains modify
   protein function and can trigger signal transduction^[70]21–[71]24. The
   process of ubiquitination is reversible and can be counteracted by
   deubiquitinating enzymes (DUBs). As fine-tuning modulators of cell
   signaling and activities, DUBs have emerged as important regulators of
   astrocytes and CNS autoimmunity mediating both protective and
   disease-promoting astrocyte functions^[72]25. Of note, the function of
   DUBs in CNS autoimmunity is also cell type-specific as illustrated by
   EAE promoting function of the DUB A20 (TNFAIP3) in T cells and its EAE
   inhibitory function in astrocytes^[73]26,[74]27

   OTUD7B is a DUB belonging to the OTU subfamily and is expressed in all
   human and murine tissues. OTUD7B is expressed in all regions of the
   CNS, in particular in astrocytes and oligodendrocytes but not in
   neurons and microglia
   ([75]https://www.proteinatlas.org/ENSG00000264522-OTUD7B). OTUD7B can
   hydrolyze K11-^[76]28, K48-^[77]29 and K63-^[78]30 linked ubiquitin
   chains from distinct substrates and regulates pro-inflammatory
   signaling by inhibiting TNF-mediated NF-κB activation through K63
   deubiquitination of RIPK1 and TRAF6^[79]30,[80]31. The regulation of
   TNF signaling by OTUD7B is cell type-specific since OTUD7B prevents
   TNF-mediated apoptosis in dendritic cells in vitro and in murine
   cerebral malaria by cleaving K48-ubiquitin chains from the E3 ubiquitin
   ligase TRAF2^[81]29. In T cells, OTUD7B inhibits TCR and CD28-mediated
   activation by regulating the non-degradative ubiquitination of ZAP70, a
   central molecule in proximal TCR signaling^[82]32. Consequently,
   OTUD7B-deficient mice are protected from myelin oligodendrocyte
   glycoprotein (MOG)-induced EAE due to impaired activation of
   encephalitogenic CD4^+ T cells^[83]32 further illustrating cell type-
   and disease-specific functions of OTUD7B in inflammatory diseases.

   In this study we investigated the unresolved astrocyte-specific
   function of OTUD7B in EAE and identified a dual role of OTUD7B. First,
   OTUD7B restricts TNF-induced pro-inflammatory astrocyte responses by
   sequential K63- and K48-deubiquitination of RIPK1. Second, OTUD7B
   fosters GFAP protein expression and astrogliosis by enhancing
   interleukin (IL) − 6 STAT3-dependent GFAP mRNA expression and
   preventing proteasomal degradation of GFAP through its
   K48-deubiquitination.

Results

Upregulation of astrocytic Otud7b during CNS autoimmunity

   To investigate the role of OTUD7B in CNS autoimmunity, we first
   compared the expression of Otud7b in the astrocytes of patients with
   progressive MS and healthy individuals. Analysis of public available
   microarray dataset (No. [84]GSE180759) revealed that Otud7b mRNA
   expression was upregulated in the lesion edges and periplaque white
   matter of MS patients as compared to control white matter of healthy
   individuals (Fig.[85]1A). Next, we determined the expression of Otud7b
   in MOG-induced EAE, which is characterized by a peak of disease around
   day 15 post immunization (p.i.) followed by a gradual decline of
   clinical symptoms^[86]16. In EAE, Otud7b mRNA expression was
   upregulated and highest in spinal cords at day 15 p.i. and declined up
   to day 22 p.i. (Fig. [87]1B). Immunofluorescence imaging of the spinal
   cord further demonstrated a marked increase of OTUD7b protein
   expression in the astrocytes at 15 p.i. as compared to non-immunized
   mice (Fig. [88]1C).

Fig. 1. Upregulation of Otud7b in MS and EAE lesions.

   [89]Fig. 1
   [90]Open in a new tab

   A mRNA levels of OTUD7b were analyzed from astrocytes from different
   lesional regions of patients with progressive MS (n = 5), and control
   white matter of astrocytes from healthy individuals (n = 3). Data from
   GEO database (accession number [91]GSE180759). B EAE was induced in
   Otud7b^fl/fl mice by MOG[35-55] peptide immunization combined with
   pertussis toxin. Spinal cord was isolated from non-immunized (d0) and
   immunized mice at day15 and day 22 days p.i., mRNA was isolated and the
   expression of Otud7b mRNA was analyzed by qRT-PCR. (n = 3). C
   Immunofluorescence staining showing presence of SOX2/SOX9^+GFAP^+
   OTUD7b expressing astrocytes (marked by yellow arrow) in spinal cord
   sections of unimmunized and day 15 p.i. Otud7b^fl/fl mice. All
   photographs are representative of three mice per group. Scale bars
   correspond to 10 μm. D Astrocytes were isolated from non-immunized (d0)
   and immunized mice at day15 p. i., using anti-ACSA-2 antibody for
   magnetic cell sorting. The expression of Otud7b mRNA in sorted
   astrocytes was analyzed by qRT-PCR (n = 3 mice per timepoint). (E-H)
   For spatial transcriptomics, EAE was induced in Otud7b^fl/fl mice by
   MOG[35-55] peptide immunization combined with pertussis toxin. 5 μm
   thick sections of the spinal cord from non-immunized and immunized mice
   at day15 (n = 3, both genotypes) were placed on Xenium slides and in
   situ RNA expression analysis was performed using a pre- designed mouse
   brain panel of 247 genes and a custom panel of 50 genes. (E, top panel)
   Representative hematoxylin and eosin (H&E)-stained sections from
   Otud7b^fl/fl mice (d0 and d15 p.i.) mounted on Xenium slides. E (bottom
   panel) Cell map from the same mice showing the distribution of the
   annotated cell populations in the spinal cord based on single-cell
   sequencing analysis using Xenium mouse brain slides in their native
   position. In mice with EAE, the inner area surrounded by black solid
   lines represents the lesion core compartment, he area between the black
   solid and dashed lines represents the lesion rim compartment and the
   outermost peri-lesion compartment is on the other site of the dashed
   line. F Dot plot showing Otud7b expression at day 15 p. i. in different
   lesion compartments compared to non-immunized mice. G The left bar
   graph shows the total number of cells in the spinal cord sections and
   the right stacked bar graph shows the relative cellular composition in
   different anatomical compartments (thin vertical lines indicate
   standard error across three replicates for the percentage of the
   respective cell type). (n = 3 mice per timepoint). H Chemokine gene
   expression by astrocytes at day15 p. i. compared to astrocytes of
   non-immunized mice in the different disease-associated anatomical
   compartments. All data represented as mean values ± SEM. *p  <  0.05,
   **p  <  0.01, ***p  <  0.001, ****p  <  0.0001. ns, not significant.
   Statistical analyses: two-tailed Student’s t test (B, D) Adjusted
   p-values were calculated by DEseq2 using Benjamini-Hochberg corrections
   of two-sided Wald test p values (F, G, H). Scale bar = 10 μm (C),(E)
   200 μm. Source data are provided as source data file.

   Consistently, analysis of Otud7b mRNA expression of astrocytes isolated
   from the normal and EAE-diseased spinal cord revealed a prominent
   upregulation of Otud7b mRNA expression in astrocytes of mice with EAE
   (Fig. [92]1D). For a detailed spatial analysis of the expression of
   astrocytic Otud7b in EAE, we performed spatial transcriptomics with
   single-molecule resolution on the Xenium platform. Upon immunization
   with MOG[35–55] peptide and a pertussis toxin boost, cell types were
   identified based on the transcriptome data using the pre-designed mouse
   brain panel with 247 genes in combination with a custom panel including
   additional 50 genes (Suppl. dataset. [93]1). We defined inflammatory
   lesions in the sections characterized by the accumulation of CD3^+
   CD4^+ and CD3^+ CD8^+ T cells (Fig. [94]1E). In addition, we defined a
   lesion rim in direct proximity to the lesions, with fewer T cells
   (200 µm), and as a third region defined peri-lesion with no T cells
   (Fig. [95]1E, black lines are defined lesion regions). We identified
   that Otud7b mRNA is upregulated about 1.5-fold in Sox2^+ Sox9^+ Aqp4^+
   astrocytes within the lesion core and lesion rim, but not in the
   peri-lesions as compared to non-immunized mice (Fig. [96]1F). As
   expected, the number of identified cells per section increased almost
   two-fold upon immunization (Fig. [97]1G), with a strong influx of
   CD68^+ microglia and macrophages into the lesions (Fig. [98]1G). We
   observed a strong induction of genes associated with inflammation
   including Isg15, the chemokines CXCL9, Ccl2, CXCL10 and the cytokines
   IL-6 and IL1β at d15 p.i. (Fig. [99]1H), with the induction being
   particularly strong within the lesion core and a gradual decline in the
   lesion rim and peri-lesion. Taken together, astrocytes upregulate
   Otud7b expression in CNS autoimmunity with the highest expression in
   the T cell-enriched lesions and a gradual decline with increasing
   distance from the lesions core.

Astrocytic OTUD7B-deficiency aggravates EAE

   To determine the astrocyte-specific function of OTUD7B in EAE, we
   crossed GFAP-cre mice with Otud7b^fl/fl to generate GFAP-cre
   Otud7b^fl/fl mice with deletion of Otud7b in astrocytes. The GFAP-cre
   strain used here expresses Cre late during embryonic development (day
   14.5^[100]33), has a high deletion efficacy in all spinal cord and
   brain astrocytes but only low deletion in
   neurons^[101]10,[102]11,[103]16,[104]33, and allows analysis of
   astrocyte functions in inflammatory CNS
   disorders^[105]10,[106]11,[107]16,[108]17,[109]34. Mice were genotyped
   using DNA extracted from ear biopsies. Knockout mice were confirmed by
   the presence of the floxed OTUD7B allele and Cre recombinase under the
   GFAP promoter, while wild-type controls carried only the floxed OTUD7B
   allele (Suppl. Fig. [110]1A).In vivo deletion of Otud7b in spinal cord
   astrocytes was validated by spatial transcriptomics showing a strong
   upregulation of Otud7b in astrocytes in EAE of Otud7b^fl/fl mice, which
   was absent in GFAP-cre Otud7b^fl/fl mice (Fig. [111]2A). In
   Otud7b^fl/fl mice, Otud7b was most strongly expressed in astrocytes,
   much weaker in neurons and marginally in pericytes, microglia,
   oligodendrocytes and endothelial cells (Fig.[112]2A). Western blot (WB)
   analysis of astrocytes isolated from spinal cord of Otud7b^fl/fl and
   GFAP-Cre Otud7b^fl/fl also demonstrated efficient deletion of OTUD7B in
   astrocytes of GFAP-cre Otud7b^fl/fl mice (Supplementary Fig. [113]1B).
   GFAP-cre Otud7b^fl/fl mice were born in a normal mendelian ratio and
   grew normally. Histological analysis showed that astrocytes in the
   spinal cords of GFAP-cre Otud7b^fl/fl mice lacked OTUD7B expression
   (Supplementary Fig. [114]1C). Despite this, the spinal cords appeared
   normal, with no signs of neurodegeneration or inflammation
   (Supplementary Fig. [115]1D). Major leukocyte populations in the spinal
   cord, spleen, and lymph node were comparable between Otud7b^fl/fl and
   GFAP-cre Otud7b^fl/fl mice (Supplementary Fig. [116]1E–H),
   consolidating that OTUD7B expression in astrocytes does not regulate
   inflammatory responses under homeostatic conditions.

Fig. 2. Astrocyte-specific OTUD7B ameliorates EAE.

   [117]Fig. 2
   [118]Open in a new tab

   EAE was induced in Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl by MOG[35-55]
   peptide immunization combined with pertussis toxin (A) Spatial
   transcriptomics of Otud7b mRNA expression in CNS-resident cell
   populations of non-immunized and immunized Otud7b^fl/fl and GFAP-cre
   Otud7b^fl/fl mice. B The mice were monitored daily for the clinical
   signs of the disease. Upper panel represents mean clinical score ± SEM
   (Otud7b^fl/fl n = 7, GFAP-cre Otud7b^fl/fl n = 11) and the lower panel
   shows the disease incidence, day of disease onset and maximum clinical
   score of all immunized mice per group. Data represented as mean
   values ± SEM. C The body weight was measured daily up to day 16 p.i.
   Graph represents percent body weight ± SEM (Otud7b^fl/fl n = 7,
   GFAP-cre Otud7b^fl/fl n = 5). All bars represent mean values ± SEM.
   Student’s t-test *p < 0.05, **p < 0.01. D Cresyl-violet luxol-fast blue
   staining showing inflammatory infiltrates (marked by dotted line) and
   myelin staining at day 15 p.i in spinal cord of Otud7b^fl/fl and
   GFAP-cre Otud7b^fl/fl mice. E H&E staining showing immune cell
   infiltration (marked by dotted line) in spinal cord sections from
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice at day 15 p.i. E–G At day
   15 p.i., reactive astrocytes of Otud7b^fl/fl mice exhibit strong GFAP
   immunoreactivity close to inflammatory lesions in the spinal cord,
   whereas GFAP immunoreactivity of astrocytes surrounding inflammatory
   lesions in spinal cord sections of GFAP-cre Otud7b^fl/fl mice was
   largely lost by immunohistochemistry (F) and immunofluorescence (G) All
   photographs are representative of three mice per group. Scale bars
   correspond to 100μm (D–G). H Immunofluorescence staining showing
   presence of SOX2/SOX9^+GFAP^- astrocytes surrounding inflammatory
   lesions (marked by yellow dotted line) in spinal cord sections of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice at day 15 p.i. All
   photographs are representative of three mice per group. Scale bars
   correspond to 20μm. I Dot plot showing GFAP mRNA expression in
   different lesion compartments at d 15 p.i. *p  <  0.05, **p  <  0.01,
   ***p  <  0.001, ****p <  0.0001. ns, not significant. Statistical
   analyses: Mann–Whitney U test (B) two-tailed Student’s t test (C),
   Adjusted p-values were calculated by DEseq2 using Benjamini–Hochberg
   corrections of two-sided Wald test p values (A, I). No adjustments were
   made for the multiple comparisons. Source data are provided as source
   data file.

   To explore the role of astrocytic OTUD7B in EAE, we immunized
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice with MOG peptide. Although
   the two genotypes had similar disease onset at around day 12 p.i.,
   GFAP-cre Otud7b^fl/fl mice showed significantly increased clinical
   scores with higher maximal clinical scores and disease incidence
   (Fig. [119]2B), and reduced body weight (Fig. [120]2C).

   In accordance with the aggravated clinical symptoms, and reduced myelin
   thickness was more pronounced in GFAP-cre Otud7b^fl/fl as compared to
   control mice at d15 p.i. (Fig. [121]2D). In addition, inflammatory
   infiltrates were larger, more confluent and infiltrated deeper in the
   spinal cord tissue in GFAP-cre Otud7b^fl/fl mice (Fig. [122]2D, E). In
   EAE, astrocyte morphology and reactivity was strongly altered. Adjacent
   to and bordering inflammatory lesions astrocytes of Otud7b^fl/fl mice
   strongly expressed GFAP and showed prolonged GFAP^+ astrocytic
   processes (Fig. [123]2F, G). On the contrary, astrocytes of the
   GFAP-cre Otud7b^fl/fl mice associated with inflammatory infiltrates
   only weakly expressed GFAP and, thus, did not form a GFAP^+ border
   surrounding the inflammatory infiltrate (Fig. [124]2F–H). In both
   control and GFAP-cre Otud7b^fl/fl mice, equal numbers of SOX2^+ SOX9^+
   astrocytes were present indicating that the reduction of GFAP
   expression was not caused by a loss of OTUD7B-deficient astrocytes
   (Fig. [125]2G–H). Interestingly, a comparison of GFAP mRNA levels by
   spatial transcriptomics showed upregulation particularly in the lesion
   core and the lesion rim, which was higher in Otud7b^fl/fl than in
   GFAP-cre Otud7b^fl/fl mice. (Fig. [126]2I) indicating that the
   diminished transcription of GFAP may contribute to the reduced GFAP
   protein in GFAP-cre Otud7b^fl/fl mice (Fig.[127]2F–H).

OTUD7B suppressed chemokine and cytokine production by astrocytes

   Reactive astrocytes contribute to the development of EAE by producing
   pro-inflammatory mediators, including chemokines, which induce the
   recruitment of encephalitogenic autoimmune CD4^+ T cells into the
   CNS^[128]35,[129]36. Since GFAP-cre Otud7b^fl/fl mice exhibited
   increased CNS inflammation, we next determined the impact of OTUD7B on
   the transcriptome of astrocytes isolated from the spinal cords of
   non-immunized and MOG-immunized (d15 p.i.) Otud7b^fl/fl and GFAP-cre
   Otud7b^fl/fl mice. We detected 1944 genes differentially regulated
   between OTUD7B-deficient and -sufficient astrocytes (Fig. [130]3A). Of
   these 1944 genes, 689 genes were differentially regulated under
   homeostatic conditions, 1090 genes were regulated at d15 p.i. and 165
   genes were regulated under both conditions (Fig. [131]3A). Clustering
   of the genes using the K-means clustering algorithm resulted in 10
   clusters. Ontology and KEGG pathway enrichment analysis showed that the
   pathways related to chemokine and cytokine signaling (Fig. [132]3B)
   were upregulated in cluster 6 and 2 of OTUD7B-deficient astrocytes upon
   EAE. In addition, OTUD7B-deficient astrocytes of cluster 6 and 2 had
   increased expression of genes regulating transendothelial migration of
   leukocytes, pro-inflammatory cytokine and chemokine signaling, cell
   adhesion and angiogenesis (Fig. [133]3C–D). A detailed analysis of
   chemokine genes showed that upon induction of EAE, expression of CCL
   and CXCL chemokines was upregulated in both genotypes but expression of
   CCL2, 3, 4, 5, 6, 7, 8, 11, 19 and CXCL 1, 2, 9, 10, 12, 13, 16 were
   higher in OTUD7B-deficient astrocytes. This suggests that the increased
   disease pathology in GFAP-cre Otud7b^fl/fl mice was due to increased
   chemokine and cytokine mediated infiltration of leukocytes into the
   spinal cord, (Fig. [134]3E). Only Ccl22, Cxcl16, and Cxcl12, which are
   stronger expressed by microglia and brain macrophages as compared to
   astrocytes ([135]https://www.proteinatlas.org) were increased
   upregulated in OTUD7B-comptetent astrocytes (Fig. [136]3E).

Fig. 3. OTUD7B deficient astrocytes show pro-inflammatory gene signature.

   [137]Fig. 3
   [138]Open in a new tab

   A–F EAE was induced in Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl by
   MOG[35-55] peptide immunization combined with pertussis toxin. Spinal
   cord and magnetic cell sorted astrocytes were isolated from
   non-immunized (d0) and immunized mice at day 15 p.i. and processed for
   RNA sequencing. A Venn diagram of 1944 upregulated genes (FC < 3) in
   Otud7b^fl/fl mice compared to GFAP-cre Otud7b^fl/fl at d0 and d15 p. i.
   Absolute and relative values are shown. Gene ontology (B, D) and KEGG
   pathway (C) analysis was performed using ClueGO software with
   right-sided hypergeometric test with Bonferoni-step-down p-value
   correction. Only terms with FDR < 0.05 and at least GO-level 6 are
   shown. Numbers in brackets represent number of enriched genes compared
   to total number of annotated genes in that term. E Heat map showing the
   expression profile for chemokine genes from astrocytes of non-immunized
   and immunized mice at day 15 p. i. F Spatial transcriptomics analysis
   of differential gene expression in various lesional compartments of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice at day 15 p.i. G RNA was
   isolated from the spinal cord of non-immunized and immunized mice (day
   15 p.i.). Expression of Cxcl1, Cxcl10, Cxcl11, Ccl2 and Ccl20 was
   analyzed by qRT-PCR. Data show fold-change increase in gene expression
   compared to non-immunized mice (n = 3 per group, all bars represent
   mean ± SEM. two-tailed Student’s t-test *p < 0.05, **p < 0.01).
   Statistical analyses: Adjusted p-values were calculated by DEseq2 using
   Benjamini-Hochberg corrections of two-sided Wald test p values (F) and
   two-tailed Student’s t test (G). Source data are provided as source
   data file.

   Spatial transcriptome analysis of spinal cord tissue provided further
   information on the anatomic distribution of chemokines and cytokines.
   Chemokine mRNA expression was higher in OTUD7B-deficient astrocytes in
   the lesion cores and in the rims surrounding the lesion in relation to
   peri-lesion areas where the chemokine levels were comparable
   (Fig. [139]3F). Among the cytokines studied, only IL-6 and IL-1β but
   not TNF, TGF-β1 were highly expressed in OTUD7B-deficient astrocytes in
   EAE. In contrast to astrocytes, chemokine and cytokine expression of
   microglia was equal between the two genotypes.

   Consistent with the astrocyte RNA-sequencing data, RT-PCR-based
   quantification of chemokine mRNA in spinal cords of mice with EAE
   revealed that GFAP-cre Otud7b^fl/fl mice had significantly higher
   levels of CXCL1, CXCL10, CXCL11, CCL2, and CCL20 mRNA at day 15 p.i.,
   which are major attractants for lymphoid and monocytic cells
   (Fig. [140]3G).

Increased recruitment of encephalitogenic CD4^+ T cells in GFAP-cre
Otud7b^fl/fl mice

   To determine whether the increased chemokine production of
   OTUD7B-deficient astrocytes resulted in an enhanced recruitment of
   leukocytes to the CNS in MOG-immunized mice, we performed a flow
   cytometry analysis of leukocytes isolated from the spinal cords. In
   contrast to non-immunized mice (Supplementary Fig. [141]1E), relative
   and absolute numbers of CD4^+ but not of CD8^+ T cells were
   significantly increased in Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice
   with EAE (Fig. [142]4A). Additionally, numbers of CD11c^+ dendritic
   cells, Ly6C^high CD11b^+ inflammatory monocytes and F4/80^+ CD11b^+
   macrophages/microglia but not of CD19^+ B cells were higher in GFAP-cre
   OTUD7b^fl/fl mice (Fig. [143]4B). Among CD4^+ T cell subsets,
   IFN-producing Th1 cells, IL-17-producing Th17 cells, and
   GM-CSF-producing CD4^+ T cells are the major effectors in EAE and each
   of them can induce EAE independently^[144]37–[145]39. We identified
   that the absolute but not relative numbers of infiltrating GM-CSF^+,
   IFN-γ^+, and IL17^+ CD4 cells were significantly increased in the
   spinal cord of GFAP-cre Otud7b^fl/fl mice with EAE (Fig. [146]4C, D).
   This indicates that astrocyte-specific OTUD7B limits the recruitment of
   encephalitogenic CD4^+ T cell subsets to the CNS but does not influence
   the differentiation and composition of the recruited CD4^+ T cell
   subsets during EAE. In accordance with the increased recruitment of
   encephalitogenic CD4^+ T cells, inflammatory monocytes and macrophages
   into the CNS of GFAP-cre Otud7b^fl/fl mice with EAE, mRNA production of
   IFN-γ,TNF, IL-17, GM-CSF and NOS2, which all contribute to
   demyelination in EAE^[147]40–[148]44, was also increased
   (Fig. [149]4E).

Fig. 4. Reduced leukocyte infiltration and inflammation in the spinal cord of
Otud7b^fl/fl mice during EAE.

   [150]Fig. 4
   [151]Open in a new tab

   EAE was induced in Otud7b^fl/fl (n = 8) and GFAP-cre Otud7b^fl/fl
   (n = 8) by MOG[35-55] peptide and spinal cord was isolated at day 0 and
   15 p.i. The infiltrating leukocytes were isolated by Percoll gradient
   centrifugation and analyzed by flow cytometry. A Representative dot
   plots show the percentage of CD4^+ and CD8^+ T cells and (B) absolute
   number of infiltrating leukocytes in the spinal cord of Otud7b^fl/fl
   and GFAP-cre Otud7b^fl/fl mice at day15 p.i. C, D GM-CSF, IFN-γ and
   IL-17 producing CD4^+ T cells were analyzed by flow cytometry.
   Representative dot plots (C) and absolute numbers of GM-CSF-, IFN-γ-
   and IL-17-producing CD4^+ T cells (D) are shown (Otud7b^fl/fl n = 6,
   GFAP-cre Otud7b^fl/fl n = 7). E Relative mRNA expression of IFN-γ,
   IL-17, GM-CSF, TNF and Nos2 was analyzed by qRT-PCR from spinal cord of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl (n = 3) mice at day 15 p.i. All
   bars represent mean values ± SEM (B, D, E).Statistical analyses:
   two-tailed Student’s t test (B,D,E). *p  <  0.05, **p  <  0.01,
   ***p  <  0.001, ****p  <  0.0001. ns, not significant No adjustments
   were made for multiple comparisons. Source data are provided as source
   data file.

   Collectively, OTUD7B expression in astrocytes suppressed astrocyte
   chemokine production, recruitment of encephalitogenic leukocytes,
   demyelination and disease symptoms in EAE illustrating the important
   neuroprotective function of astrocytic OTUD7B.

OTUD7B suppressed early TNF-induced pro-inflammatory signaling in astrocytes

   The increased astrocyte activation and chemokine production of GFAP-cre
   Otud7b^fl/fl mice under inflammatory but not under homeostatic
   conditions indicates that pro-inflammatory cytokines induced
   OTUD7B-dependent immunoregulatory astrocyte function. To identify
   whether astrocyte chemokine and cytokine production induced by the
   encephalitogenic cytokines TNF, IFN-γ and IL-17, respectively, is
   regulated by OTUD7B, we cultured OTUD7B-deficient and -competent
   astrocytes, stimulated them with the cytokines and determined chemokine
   and cytokine production by RT-PCR. Upon stimulation with TNF, mRNA
   expression of CXCL1, CXCL11, CCL20, IL-6, CCL2 and NOS2 were
   significantly increased in OTUD7B-deficient astrocytes (Fig. [152]5A).
   In contrast, OTUD7B did not regulate IFN-γ- and IL-17-induced chemokine
   and cytokine production (Fig. [153]5A) indicating that TNF signaling is
   the major pathway regulated by OTUD7B in astrocytes. Thus, both
   increased TNF mRNA expression in EAE (Fig. [154]4E) and TNF-induced
   astrocytic cytokine and chemokine production are under control of
   astrocytic OTUD7B.

Fig. 5. OTUD7B impairs TNF mediated pro-inflammatory responses in astrocytes.

   [155]Fig. 5
   [156]Open in a new tab

   A, B Primary astrocyte cultures were prepared from P0/1 pups of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice, respectively. A Cells were
   stimulated with TNF (20 ng/mL), IFN-γ (10 ng/mL) or IL-17 (50 ng/mL)
   for 16 h and lysed in buffer RLT for RNA isolation. Relative mRNA
   expression of Cxcl-1, Cxcl-10, Cxcl-11, Ccl-1, Ccl-20, Nos-2 and IL-6
   was determined by qRT-PCR (The experiment was independently repeated
   three times, n = 3). All graphs represent fold change over unstimulated
   control. All bars represent mean values ± SEM. Two-tailed student’s
   t-test *p < 0.05, **p < 0.01, ***p < 0.001. Proteins were harvested and
   analyzed by WB with the indicated antibodies. Representative WB are
   shown after (B) TNF (20 ng/mL), (C) IFN-γ (20 ng/mL) and (D) IL-17
   (50 ng/mL) stimulation, respectively. Blots represent one out of three
   independent experiments. Source data are provided as source data file.

   To directly assess the effect of OTUD7B on TNF, IFN-γ and IL-17
   signaling, we stimulated cultivated astrocytes with these cytokines and
   analyzed activation of the respective signaling pathways by WB. Upon
   stimulation with TNF, activation of both the NF-κB and MAPK pathway,
   indicated by increased phosphorylation of p65 and degradation of IκBα,
   and the phosphorylation of ERK, p38 and JNK in the MAPK pathways were
   stronger in OTUD7B-deficient astrocytes (Fig. [157]5B). Of note, these
   differences were detectable as early as 10 min post TNF stimulation. In
   contrast, stimulation with IFN-γ did not result in increased STAT1
   phosphorylation (Fig. [158]5C) and activation with IL-17 had no effect
   on phosphorylation of IκBα, ERK, p38 and JNK (Fig. [159]5D). Thus,
   OTUD7B is an inhibitor of TNF but not IFN-γ and IL-17 signaling in
   astrocytes. To further prove that OTUD7B inhibits TNF-induced
   inflammation in astrocytes, OTUD7B was overexpressed in both OTUD7B-
   sufficient and -deficient astrocytes (Fig. [160]6A). Overexpression of
   OTUD7b reduced levels of TNF-induced p-p65 (Fig. [161]6A) and Cxcl1,
   Cxcl11, Ccl20, Il6, Ccl2, and Nos2 (Fig. [162]6B) mRNA expression of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl astrocytes and completely
   abolished differences between the two genotypes. This confirms the
   critical inhibitory role of OTUD7B in TNF-induced astrocyte activation.

Fig. 6. Over-expression of OTUD7B impairs TNF mediated pro-inflammatory
responses in OTUD7B deficient astrocytes.

   [163]Fig. 6
   [164]Open in a new tab

   A, B Primary astrocyte cultures were prepared from P0/1 pups of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice, respectively. OTUD7B was
   over-expressed in both OTUD7B-sufficient and -deficient astrocytes
   using the pCMV6-AC-GFP-OTUD7b expression plasmid. A Cells were
   stimulated with TNF (20 ng/mL) for 24 h and proteins were harvested and
   analyzed by WB with the indicated antibodies. WB representative of one
   of the three independent experiments. B Astrocytes were lysed in RLT
   buffer for RNA isolation. Relative mRNA expression of Cxcl-1, Cxcl-10,
   Cxcl-11, Ccl-1, Ccl-20, Nos-2 and IL-6 was determined by qRT-PCR (The
   experiment was independently repeated three times, n = 3). All graphs
   represent fold-change over unstimulated control. All bars represent
   mean values ± SEM. Statistical analyses: two-tailed Student’s t test
   *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as source
   data file.

OTUD7B regulates TNF signaling by sequential K63- and K48- deubiquitination
of RIPK1

   Since chemokine production of OTUD7B-deficient astrocytes was highest
   in the lesion core (Fig. [165]3F) and OTUD7B-regulated TNF-dependent
   chemokine production, we characterized the cellular source of TNF. Our
   spatial transcriptome data identified CD68⁺ microglia/macrophages and T
   cells as the major cellular source of TNF. The percentage of
   TNF-producing cells was similar in both Otud7b^fl/fl and GFAP-cre
   Otud7b^fl/fl mice (Fig. [166]7A). We next analyzed the spatial
   distribution of TNF mRNA in Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice
   with EAE. In accordance with the chemokine data, expression of TNF was
   highest in the lesion core with a gradual decline towards the lesion
   rim and the peri-lesional tissue (Fig. [167]7B). The higher TNF levels
   in GFAP-cre OTUD7b^fl/fl mice at day 15 p.i. (Fig. [168]7B), correspond
   to the increased infiltration of macrophages and T cells in the spiunal
   cord of GFAP-cre Otud7b^fl/fl mice during EAE (Fig. [169]4B). The
   spatial transcriptome data also indicate that astrocytes will be
   exposed to TNF for several days based on the persistence of the
   lesions.

Fig. 7. Temporal regulation of TNF signaling by OTUD7B.

   [170]Fig. 7
   [171]Open in a new tab

   A Spatial transcriptome analysis of TNF producing cells in the spinal
   cord of non-immunized and immunized Otud7b^fl/fl and GFAP-cre
   Otud7b^fl/fl mice. B Spatial transcriptome analysis of TNF mRNA
   expression in EAE lesions of non-immunized and immunized Otud7b^fl/fl
   and GFAP-cre Otud7b^fl/fl mice. C–H OTUD7b-sufficient and -deficient
   primary astrocytes were stimulated with TNF (C–F: 20 ng/mL, G–H
   indicated concentrations). At indicated timepoints, astrocytes were
   lysed in RIPA lysis buffer for protein isolation. C Proteins were
   immunoblotted with the indicated antibodies. Protein expression was
   normalized to GAPDH and quantified based on WB data. Representative WB
   from one out of three experiments is shown. D Co-immunoprecipitation
   was performed with anti-RIPK1 antibody followed by immunodetection of
   the indicated proteins by WB. E OTUD7B-sufficient and -deficient
   astrocytes were transfected with non-specific control siRNA or A20
   siRNA (5 μM) and stimulated with TNF for 5 days. At day 5
   post-stimulation, astrocytes were lysed with RIPA buffer and proteins
   were analyzed by WB using indicated antibodies. RIPK1 and TRAF2
   expression was normalized to GAPDH and quantified based on WB data. F
   Astrocytes were treated with TNF for 3 days and MG132 was added on day
   3 for 6 hours to inhibit the proteasome. Astrocytes were lysed in RIPA
   buffer and the protein lysates were analyzed by WB for RIPK1 and GAPDH.
   G, H OTUD7B-sufficient and -deficient primary astrocytes were
   stimulated with 10 ng/mL, 20 ng/mL, and 50 ng/mL TNF for 5 days. On day
   5 post-stimulation, cells were lysed in RIPA lysis buffer and proteins
   were isolated for (G) analysis of protein expression by WB and (H)
   identification of interacting proteins by co-immunoprecipitation of
   protein complexes with anti-RIPK1 antibody and WB. All WB blots are
   representative of one of three independent experiments. Statistical
   analyses: Adjusted p-values were calculated by DEseq2 using
   Benjamini-Hochberg corrections of two-sided Wald test p values (B).
   Source data are provided as source data file.

   TNF signaling is dynamically regulated by K63 and K48
   poly-ubiquitination of RIPK1. Rapidly after TNF stimulation, RIPK1 is
   K63 poly-ubiquitinated by TRAF2 and cIAP1 leading to RIPK1-dependent
   activation of NF-κB and MAPKs^[172]29,[173]45. The activation of NF-κB
   induces expression of A20 (TNFAIP3), which cleaves K63 chains from
   RIPK1 and can also induce K48-dependent proteasomal degradation of
   RIPK1^[174]46, which also leads to degradation of TRAF2 and termination
   of NF-κB and MAPK signaling^[175]47,[176]48. Thus, we determined the
   impact of OTUD7B on RIPK1, TRAF2 and cIAP1 over a period of 5 days to
   reflect that under in vivo conditions astrocytes are exposed to TNF for
   several day. WB analysis showed that RIPK1, TRAF2 and cIAP1 protein
   levels remained unchanged in both genotypes but that phosphorylation of
   p65, p38 and JNK was increased in OTUD7B-deficient astrocytes at days 1
   and 2 of TNF stimulation (Fig. [177]7C). From day 3 to 5, RIPK1
   (Fig. [178]7C,) and from day 4 to 5 TRAF2 gradually declined in
   OTUD7B-deficient but not OTUD7B-competent astrocytes, whereas cIAP1
   remained unchanged (Fig. [179]7C). The decline of RIPK1 protein in
   OTUD7B-deficient astrocytes was preceded by a strong increase of A20,
   which was much weaker in OTUD7B-competent astrocytes (Fig. [180]7C). In
   OTUD7B-deficient astrocytes the decline of RIPK1 and TRAF2 protein was
   paralleled by strong reduction of p65, p38 and JNK phosphorylation,
   which was not detectable in OTUD7B-expressing astrocytes
   (Fig. [181]7C).

   Since RIPK1 activity and stability are regulated by its K63 and K48
   poly-ubiquitination, respectively, we analyzed next whether the DUB
   OTUD7B regulates the RIPK1 ubiquitination status. Immunoprecipitation
   of RIPK1 at the various time points up to day 5 of TNF stimulation,
   showed that the activating K63-ubiqutination of RIPK1 was higher in
   OTUD7B-deficient astrocytes at days 1 and 2 but reduced at days 3 to 5
   as compared to OTUD7B-expressing astrocytes (Fig. [182]7D). On the
   contrary, K48 poly-ubiquitination of RIPK1 strongly increased from days
   3 to 5 in OTUD7B-deficient but to a much lesser extent in
   OTUD7B-competent astrocytes (Fig. [183]7D). In the RIPK1 complexes,
   higher amounts of A20 and TRAF2 were present in OTUD7B-deficient as
   compared to OTUD7B-expressing astrocytes (Fig. [184]7D).

   To validate that OTUD7B counteracts A20-mediated proteasomal
   degradation of RIPK1, we silenced A20 by siRNA in TNF-stimulated
   astrocytes. Inhibition of A20 restored RIPK1 protein in
   OTUD7B-deficient astrocytes to the same level as in OTUD7B-competent
   astrocytes (Fig. [185]7E). In parallel to the increased RIPK1 protein,
   TRAF2 increased in A20 siRNA-treated OTUD7B-deficient astrocytes
   (Fig. [186]7E), which is explained by the stabilizing role of RIPK1 for
   TRAF2^[187]48. In addition, proteasome inhibition by MG132 prevented
   degradation of RIPK1 in OTUD7B-deficient astrocytes (Fig. [188]7F),
   which is in line with both the K48-dependent proteasomal degradation of
   RIPK1 in OTUD7B-deficient astrocytes and the prevention of proteasomal
   degradation by K48-deubiquitination of RIPK1 by OTUD7B in
   OTUD7B-competent astrocytes (Fig. [189]7F).

   Since spatial transcriptomics revealed an increase of TNF mRNA from the
   healthy tissue to the core of EAE lesions (Fig. [190]7B), we stimulated
   cultivated astrocytes with increasing TNF concentrations for 5 days
   (Fig. [191]7G). These studies revealed that TNF-stimulation reduced
   RIPK1 proteins in OTUD7B-deficient but not in OTUD7B-competent
   astrocytes in dose-dependent manner (Fig. [192]7G). In parallel TRAF2
   levels declined, whereas cIAP1 protein levels were not affected by
   OTUD7B-deficiency. In addition, the decrease of downstream p65, p38 and
   JNK phosphorylation in OTUD7B-deficient astrocytes was higher upon
   stimulation with increased amount of TNF (Fig. [193]7G). In both
   OTUD7B-competent and -deficient astrocytes, A20 was induced by
   TNF-stimulation and A20 protein levels were independent of the TNF
   concentration. Importantly and in agreement with the analysis of the
   kinetic of A20 expression (Fig. [194]7C), A20 levels were strongly
   increased in OTUD7B-deficient astrocytes, and, thus, were not affected
   by the reduced NF-κB activation. K48 poly-ubiquitination of RIPK1
   increased in OTUD7B-deficient astrocytes stimulated with higher
   concentrations of TNF and this was paralleled by a TNF-dependent
   increase of A20/RIPK1 complex formation (Fig. [195]7H). Collectively,
   these data identify that OTUD7B regulates both K63- and
   K48-ubiquitination of RIPK1 in a time- and TNF dose-dependent manner.

OTUD7B mediates GFAP-stability by its K48 deubiquitination

   Increased GFAP protein expression is a hallmark of reactive astrocytes
   but the mechanism regulating GFAP protein levels, i.e., the impact of
   GFAP mRNA transcription and, in particular, GFAP protein stability and
   turnover are largely unresolved. Histological analysis of GFAP showed
   that OTUD7B was required for GFAP protein expression in EAE
   (Fig. [196]2F–H). Moreover, the reduced GFAP mRNA of GFAP-cre
   -Otud7b^fl/fl mice with EAE (Fig. [197]2I) indicates that the reduced
   GFAP transcription may contribute to the reduction of GFAP protein in
   GFAP-cre OTUD7b^fl/fl mice.

   Since IL-6 activates STAT3 and both are important for the induction of
   GFAP mRNA, we analyzed first whether IL-6 mRNA production was reduced
   in GFAP-cre OTUD7b^fl/fl mice with EAE. Spatial transcriptomics showed
   that levels of IL-6 mRNA expression in the lesions were similar between
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice (Fig. [198]8A), indicating
   that differences in IL-6 production did not cause the variable GFAP
   mRNA expression in the two genotypes (Fig. [199]2I).

Fig. 8. OTUD7B prevents K48-ubiquitination of GFAP and its subsequent
proteasomal degradation.

   [200]Fig. 8
   [201]Open in a new tab

   A Spatial transcriptome analysisof IL-6 mRNA expression in EAE lesions
   of non-immunized and immunized Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl
   mice. B, C Primary astrocytes were stimulated with 50 ng/mL of TNF and
   lysed at the indicated time points in RIPA lysis buffer for protein
   isolation (B, F) or RLT buffer for RNA extraction (C). In (B), WB was
   performed to analyze the expression levels of OTUD7B, p-STAT3, STAT3,
   GFAP, and GAPDH. Protein expression was normalized to GAPDH and
   quantified based on WB data. Representative WB from one out of three
   experiments is shown. C Fold-change in mRNA expression levels of gfap
   over unstimulated control was analyzed by qRT-PCR and normalized to
   hprt. (The experiment was independently repeated three times, n = 3) D
   Immunofluorescence staining showing SOX2/SOX9^+GFAP^+ pSTAT3^+
   astrocytes in the spinal cord of Otud7b^fl/fl, while SOX2/SOX9^+
   astrocytes of GFAP-cre Otud7b^fl/fl mice are largely GFAP- and
   pSTAT3-negative at day 15 p.i. (Scale bars = 10μm). E OTUD7B-sufficient
   and -deficient astrocytes were stimulated with 20 ng/mL of TNF alone,
   or in combination with anti-IL6 neutralizing antibody for 3 days. On
   day 3 post-stimulation, 10 μg/mL of CHX and/or 10 μM of MG132 were
   added to the cells for 6 hours, and proteins were harvested for WB. The
   expression level of GFAP was analyzed by WB using GAPDH as a loading
   control. Relative protein expression of GFAP normalized to GAPDH is
   shown. F Protein complexes from total protein lysates were
   immunoprecipitated using anti-GFAP antibody, and precipitates were
   analyzed for GFAP, OTUD7B, and K48 ubiquitin chains by WB. All graphs
   represent fold-change over unstimulated control. All bars represent
   mean values ± SEM. Statistical analyses: Statistical analyses: Adjusted
   p-values were calculated by DEseq2 using Benjamini–Hochberg corrections
   of two-sided Wald test p values (A) two-tailed Student’s t test (C)
   *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as source
   data file.

   Next, we analyzed whether downstream of the IL-6 receptor, OTUD7B might
   regulate IL-6 signaling, in particular STAT3 activation. In
   TNF-stimulated OTUD7B-deficient astrocytes, the phosphorylation of
   STAT3 was increased in the first two days (Fig.[202]8B). and
   subsequently strongly reduced as compared to OTUD7B-competent
   astrocytes (Fig.[203]8B). GFAP protein and mRNA was regulated in the
   same kinetic with an increase in the first two days (Fig.[204]8B–C) and
   subsequent strong reduction (Fig.[205]8B–C). In good agreement, the
   reduction of pSTAT3 and GFAP was also observed in vivo in
   OTUD7b-deficient astrocytes (Fig.[206]8D), indicating that
   OTUD7B-dependent STAT3 phosphorylation is an important factor
   regulating GFAP mRNA expression. Of note, total STAT3 protein levels
   were OTUD7B-independent (Fig. [207]8B) demonstrating that OTUD7B did
   not regulate STAT3 stability.

   To determine whether GFAP protein stability might be additionally
   regulated by OTUD7B, we stimulated astrocytes with TNF, inhibited IL-6,
   blocked new protein synthesis by CHX and prevented proteasomal
   degradation of proteins by MG132 treatment (Fig. [208]8E). We limited
   these experiments until day three post TNF stimulation, since from that
   time point onwards the impaired p38 and JNK activity of
   OTUD7B-deficient astrocytes might impact on GFAP mRNA expression. These
   experiments revealed that (i) TNF-stimulation increased GFAP protein in
   OTUD7B-competent but not in OTUD7B-deficient astrocytes, (ii)
   additional inhibition of IL-6 reduced GFAP protein in both genotypes,
   (iii) additional inhibition of protein synthesis by CHX further reduced
   GFAP and that (iv) inhibition of proteasome by MG132 restored GFAP
   levels in OTUD7B-deficient astrocytes to the same level as in
   OTUD7B-competent astrocytes (Fig. [209]8E). This uncovers that OTUD7B
   stabilizes GFAP by preventing its proteasomal degradation. In good
   agreement, K48 poly-ubiquitination of GFAP was strongly increased in
   TNF-stimulated OTUD7B-deficient astrocytes (Fig. [210]8F).
   Collectively, OTUD7B supported GFAP protein levels/abundance by two
   independent but synergistic mechanisms: STAT3-mediated GFAP mRNA
   expression based on sustained RIPK1, p38 and JNK activation leading to
   continued STAT3 phosphorylation and by K48-deubiquitination of GFAP
   critical for stabilization and increased GFAP proteins in reactive
   astrocytes (Fig. [211]9A).

Fig. 9. Schematic representation showing the dual protective function of
astrocytic OTUD7B in EAE.

   [212]Fig. 9
   [213]Open in a new tab

   OTUD7B (i) prevents enhanced inflammation by limiting TNF-induced
   chemokine production in astrocytes through early K63- and late
   K48-deubiquitination of RIPK1, and (ii) fosters astrocyte bordering of
   lesions by preventing proteasomal degradation of GFAP through
   K48-deubiquitination. The figure is created in BioRender. Gopala
   Nishanth. (2025) [214]https://BioRender.com/rin8c1c.

Discussion

   Astrocytes are an important regulator of CNS inflammation, which can
   both limit and augment
   neuroinflammation^[215]10,[216]11,[217]27,[218]49–[219]51. However, the
   astrocyte intrinsic molecular mechanisms determining the pro- and
   anti-inflammatory function of astrocytes in CNS autoimmunity are
   incompletely understood. Analysis of public available transcriptome
   data of brain tissue from patients with progressive MS (^[220]52,
   Fig. [221]1A) and our spatial transcriptome analysis of mice with EAE
   (Fig. [222]2G) identified that OTUD7B is expressed in astrocytes under
   homeostatic conditions and upregulated in astrocytes associated with
   the autoimmune inflammatory lesions. The upregulation of OTUD7B is a
   protective astrocyte-intrinsic mechanism ameliorating CNS autoimmunity
   by dynamic modulation of pro-inflammatory TNF signaling through RIPK1
   deubiquitination and by upregulation of GFAP protein levels.

   A hallmark of reactive astrocytes in most CNS pathologies is an
   increased expression of GFAP protein. Mice with OTUD7B deletion in
   astrocytes had normal GFAP expression under homeostatic conditions but
   greatly reduced or even absent GFAP protein proximal to inflammatory
   lesions in EAE. Mechanistically, prolonged stimulation with TNF induced
   direct interaction of OTUD7B with GFAP and prevented K48-dependent
   proteasomal degradation of GFAP. Thus, in addition to cleavage of GFAP
   by caspase-3^[223]53, the deubiquitinating function of OTUD7B plays a
   non-redundant role for GFAP stability in reactive astrocytes.

   In addition, sustained GFAP mRNA expression was dependent on OTUD7B in
   vivo and in vitro. STAT3 is the major transcription factor inducing
   GFAP mRNA expression and can be activated by TNF and IL-6,
   respectively. Deletion of STAT3 or the cognate receptors for IL-6
   family cytokines leads to strongly diminished GFAP protein expression
   and impaired containment of inflammatory lesions by reactive
   astrogliosis^[224]5,[225]16,[226]54–[227]56.

   Since IL-6 mRNA levels did not differ in inflammatory lesions of
   OTUD7b^fl/fl and GFAP-cre OTUD7b^fl/fl mice but p38 and JNK were less
   activated due to impaired RIPK1 stability in OTUD7B-deficient
   astrocytes, OTUD7B might regulate GFAP mRNA expression indirectly. Key
   factors for the phosphorylation of STAT3 are the MAPK kinases p38 and
   JNK^[228]57–[229]62, which can be activated by RIPK1 upon TNF
   stimulation. OTUD7B regulated RIPK1 activity and downstream p38 and JNK
   activation (Fig. [230]6B). Thus, OTUD7B might regulate GFAP mRNA
   expression indirectly via the RIPK1-p38/JNK-STAT3 axis.

   In EAE and other inflammatory diseases of the CNS, reactive astrocytes
   can form borders around inflammatory lesions, which contributes to the
   local restriction of the neuroinflammation and CNS
   damage^[231]63–[232]65. Studies in GFAP-deficient mice revealed that
   deletion of GFAP leads to a more widespread inflammation and more
   severe disease in EAE and bacterial and parasitic CNS
   infections^[233]63,[234]66–[235]68. This contributed to the concept
   that bordering of inflammatory lesions by reactive astrocytes with
   increased GFAP expression contributes to the local containment of
   inflammation (reviewed by Sofroniew^[236]9). Of note, a low level of
   K48 ubiquitination and GFAP degradation were also detected in
   OTUD7B-competent astrocytes (Fig. [237]8E). This may be an important
   mechanism to prevent pathological GFAP accumulation in reactive
   astrocytes. Excessive accumulation of GFAP in astrocytes induced by
   gain-of-function mutations in the GFAP gene and by impaired proteasomal
   GFAP degradation is toxic and underlies human Alexander disease and its
   corresponding mouse models^[238]69. Perspectival, it would be
   interesting to explore OTUD7B in this astrocytopathy and to determine
   whether GFAP degradation induced by OTUD7B inhibition might have a
   therapeutic effect. In addition, it remains to be determined which E3
   ligases mediate K48 ubiquitination of GFAP. Collectively, these data
   imply that the regulation of GFAP abundance by OTUD7B-dependent
   deubiquitination is an important general mechanism potentially
   regulating the function of reactive astrocytes independent of the
   underlying disease.

   In addition to reduced GFAP protein, the dominant phenotype of
   OTUD7B-deficient astrocytes was increased chemokine production in the
   core and rim of inflammatory EAE lesions, which was associated with an
   increased recruitment of encephalitogenic T cells, more widespread
   inflammation and demyelination. Detailed analysis of the molecular
   functions of OTUD7B identified that OTUD7B rapidly interacted with
   RIPK1 upon TNF exposure. RIPK1 is a central signaling molecule
   regulating the activation of pro-inflammatory NF-κB and MAPK signaling
   as well as cell death pathways^[239]70,[240]71. The ubiquitination
   status of RIPK1 is critical to induce or inhibit these individual
   cellular pathways^[241]29,[242]45. OTUD7B limited NF-κB and MAPK
   activation by reducing RIPK1 K63-ubiquitination, which is important for
   NF-κB - and MAPK-dependent cytokine and chemokine production of
   TNF-stimulated astrocytes. The reduced chemo- and cytokine expression
   of OTUD7B-competent astrocytes was evident in vitro, in the bulk RNAseq
   analysis of ex vivo isolated astrocytes and spatial transcriptomic
   analysis of mice with EAE. In TNF activated cells, NF-κB activation
   leads to the expression of A20 which subsequently interacts with and
   inhibits RIPK1. In this negative feedback loop, A20 inhibits sustained
   RIPK1 activation by RIPK1 K63-deubiqutination and by inducing
   K48-dependent proteasomal degradation of RIPK1. Here, we identified
   that the A20-mediated dynamic change of RIPK1 ubiquitination, resulted
   in a shift of the targeted ubiquitin chains of OTUD7B from K63 to K48
   of RIPK1, and that OTUD7B interacted with A20 and diminished K48
   polyubiquitination of RIPK1. In the absence of OTUD7B, A20 induced K48
   ubiquitination of RIPK1 induced its proteasomal degradation.
   Functionally important and in agreement with the present study deletion
   of A20 in astrocytes results in an augmented and sustained NF-κB and
   also STAT1 activation leading to aggravation of EAE^[243]27.

   Of note, the dominant in vivo phenotype of OTUD7B-deficient astrocytes
   was increased activation and chemokine production at day 15 p.i., i.e.,
   when clinical symptoms of EAE already existed for several days. Thus,
   the OTUD7B-regulated in vitro shift from TNF-induced increased to
   reduced astrocyte activation was not detectable in vivo. In this
   regard, it should be stressed that (i) other NF-κB and MAPK activating
   pathways, in particular IL-17 signaling, were not regulated by OTUD7B
   (Fig. [244]5D), and (ii) IL-17 was increased expressed in the CNS of
   GFAP-cre Otud7b^fl/fl mice (Fig. [245]4E). In addition, disturbance of
   GFAP protein expression induces endoplasmic reticulum stress, increased
   activation of MAPK kinases and neuroinflammation as observed in murine
   models of Alexanders diseases^[246]69. Thus, the in vitro diminished
   RIPK1 signaling upon prolonged TNF exposure might be in vivo
   compensated by other pro-inflammatory signaling pathways contributing
   to the sustained increased chemokine production by astrocytes.

   TNF can induce RIPK1-dependent cell death, if K63 ubiquitination of
   RIPK1 is impaired. Histologically, the greatly diminished GFAP protein
   expression of OTUD7B-deficient astrocytes gave the impression that
   these cells might have been eliminated by apoptosis. However, we
   detected identical presence of SOX2^+ Sox9^+ astrocytes in both
   GFAP-cre Otud7b^fl/fl and Otud7b^fl/fl mice and spatial transcriptomics
   showed that astrocyte numbers did not differ in the core, rim and
   peri-lesion between the two genotypes. Thus, OTUD7B-deficient
   astrocytes with impaired K63-ubiqutination of RIPK1 were highly
   resilient against TNF-induced cell death. This is in contrast to
   OTUD7B-deficient dendritic cells, which rapidly undergo apoptosis upon
   exposure to TNF in murine cerebral malaria^[247]29. On the contrary,
   OTUD7B does not regulate apoptosis in T cells but facilitates proximal
   T cell receptor signaling by deubiquitination of the tyrosine kinase
   ZAP70 in EAE and murine listeriosis, two diseases characterized by the
   production of large amounts of TNF. At present, the cell type-specific
   differences underlying the differential impact of OTUD7B on TNF induced
   apoptosis are unresolved. Our current study on OTUD7B expands our
   previous understanding on the regulation of astrocyte reactivity by the
   DUBs A20^[248]27 and OTUB1^[249]10. All three DUBs are negative
   regulators of astrocyte-mediated inflammation in EAE, though via
   distinct mechanisms. A20 restricts NF-κB- and STAT1-mediated chemokine
   production by astrocytes in response to TNF, IL-17, and IFN-γ,
   respectively. Meanwhile, OTUB1 suppresses IFN-γ-driven astrocyte
   activation by deubiquitinating the suppressor of cytokine signaling
   SOCS1. In contrast, OTUD7B regulates both inflammatory and structural
   astrocyte functions by sequential K63- and K48-deubiquitination of
   RIPK1 and by directly stabilizing GFAP protein through
   K48-deubiquitination, respectively. Together, these findings highlight
   a critical and non-redundant role of DUBs in controlling astrocyte
   reactivity and CNS autoimmunity.

   This study illustrates that astrocyte activity is dynamically and
   interdependently regulated by the concentration of external factors
   such as TNF and by intrinsic signaling molecules, in particular OTUD7B.
   Although OTUD7B is a critical factor regulating astrocyte reactivity in
   EAE and unique in its capacity to regulate both GFAP and RIPK1, the
   interplay with other external factors including those derived from the
   microbiome, which also regulate astrocyte intrinsic
   responses^[250]72–[251]75, still has to be explored. Thus, it remains
   to be determined whether OTUD7B has the same function and central role
   under other experimental conditions and in human CNS disorders. At
   present the complex network of interdepend external and intrinsic
   factors regulating astrocyte reactivity cannot be resolved in its
   kinetic at the spatial level for single astrocytes. Thus, it remains to
   be determined whether OTUD7B regulates the plasticity of astrocytes or
   is a determinant for the development of specific astrocyte
   subpopulations under inflammatory conditions.

Methods

Mice

   OTUD7B^fl/fl mice with C57BL/6 background were generated with
   C57BL/6N-^Otud7btm1b(EUCOMM)Wtsi/Wtsi embryonic stem cells purchased
   from the European Mouse Mutant Archive. First, Otud7b mutant mice were
   crossed with B6.129S4-Gt(ROSA)26Sor^tm1(FLP1)Dym/RainJ (Stock No:
   009086, The Jackson Laboratory, Bar Habor, ME, USA) to delete the
   frt-flanked sequences. Thereafter, otud7b mutants were crossed with
   C57BL/6 GFAP-cre mice^[252]33 to obtain GFAP-cre Otud7b^fl/fl mice. The
   genotyping was performed by PCR of the tail DNA with primers specific
   for GFAP (GFAP Cre Sense 5’–GACACCAGACCAACTGGTAATGGTAGCGAC–3, GFAP
   Anti-Sense 5’–GCATCGAGCTGGGTAATAAGCGTTGGCAAT–3’) and Otud7b flox
   (Otud7b Sense 5’–CAGAATAAAGAGGTGGGTGAGC–3’, Otud7b Anti-Sense
   5’–CACATTGCAGTGATCATGC–3’) respectively. Wildtype C57BL/6 mice were
   obtained from Janvier (Le Genest-Saint Isle, France). Animals were kept
   under specific pathogen-free (SPF) conditions in animal facilities of
   the Otto-von-Guericke University Magdeburg (Magdeburg, Germany) and
   Hannover Medical School (Hannover, Germany). Animal care and
   experimental procedures were carried out according to the European
   animal protection law and approved by local authorities
   (Landesverwaltungsamt Halle, file number 42502-2-1260).

Induction of EAE and clinical assessment

   For active EAE induction, 8–12 weeks old mice were immunized with
   200 μg of myelin oligodendrocyte glycoprotein (MOG)[35-55] peptide
   mixed in complete Freund’s adjuvant containing 800 μg of killed
   Mycobacterium tuberculosis. In addition, 200 ng pertussis toxin
   dissolved in 200 μl PBS was intraperitoneally injected respectively at
   day 0 and 2 post immunization (p.i.). The symptoms and body weight were
   monitored daily in a double-blinded way according to a previously
   published score with increasing severity from 0 to 5 as
   follows^[253]10: 0, no signs; 0.5, partial tail weakness; 1, limp tail
   or slight slowing of righting from supine position; 1.5, limp tail and
   slowing of righting; 2, partial hind limb weakness; 2.5, dragging of
   hind limb(s) without complete paralysis; 3, complete paralysis of at
   least one hind limb; 3.5, hind limb paralysis and slight weakness of
   forelimbs; 4, severe forelimb weakness; 5, moribund or dead. Daily
   clinical scores were displayed as the mean of all individual disease
   scores within each group. Mice were euthanized using ketamine and
   xylazine. Both male and female mice were used in this study

Astrocyte isolation from adult mice

   Spinal cords were isolated from anaesthetized (ketamine and xylazine)
   and PBS-perfused naïve and EAE mice at day 15 d p.i. Single-cell
   suspension was generated using NeuroCult™ Enzymatic Dissociation Kit
   according to the manufacturer’s instruction. Astrocytes were purified
   from the single-cell suspension with the anti-ACSA-2 Microbead Kit and
   the purity was analyzed by flow cytometry with anti-ACSA-2-PE antibody.

Isolation of cells from the spinal cord and flow cytometry

   To obtain leukocytes from the spinal cord of non-immunized mice (d0)
   and mice with EAE (d15 p.i.) mice were first cardially perfused with
   0.1 M PBS (pH 7.4) in deep ketamine and xylazine anesthesia.
   Immediately thereafter, spinal cords were removed, minced through 70μm
   cell strainers followed by Percoll^® gradient centrifugation. Cells
   were counted with a hemocytometer and stained with fluorochrome-coupled
   antibodies against CD4, CD3, CD45, CD8, CD19, B220, Ly6C, Ly6G, CD11b
   and CD11c to differentiate T cells, B cells, inflammatory monocytes,
   granulocytes, macrophages, and dendritic cells respectively (see
   Supplementary dataset [254]2). For intracellular staining, cells were
   incubated with PMA (50 ng/ml), ionomycin (500 ng/ml), and Brefeldin A
   (1 μg/ml) in RPMI 1640 medium supplemented with 10% FCS, 1%
   Non-essential amino acids (NEAA), and 1% L-glutamine at 37 °C for 4 h.
   Thereafter, cells were stained with CD3, CD4, CD45 antibodies, fixed
   and permeabilized with Intracellular Fixation/Permeabilization Kit
   followed by staining with anti-IL-17, anti-GM-CSF, and anti-IFN-γ
   antibodies, respectively. Flow cytometry was performed on a Cytek
   Northern Light Flow Cytometer and data were analyzed with the FlowJo
   software.

Histology

   Mice anesthetized using ketamine and xylazine were perfused with 0.1 M
   PBS followed by 4 % paraformaldehyde in PBS. After embedding in
   paraffin, sections of brains and spinal cords were used for hematoxylin
   & eosin and cresyl violet-luxol fast blue (CV-LFB) staining. Expression
   of GFAP was demonstrated in an ABC protocol with 3,3’ diaminobenzidine
   and H[2]O[2] as substrate. For the immunofluorescent staining, the
   slides were deparaffinized for 1 h at 60 °C, followed by a removal of
   the residual paraffin with 2 × 15 min washing in xylene. Next, the
   tissue was rehydrated and incubated in 0.5% NaBH4 for 30 min at room
   temperature to reduce autofluorescence^[255]76. For antigen retrieval,
   the slides were incubated in 10 mM citric acid buffer (with 2 mM EDTA,
   0.05% Tween20) for 15 min at 95 °C and allowed to cool down for 20 min.
   After washing in PBS-T (1% TritonX-100) for 30 min and PBS for 10 min
   at room temperature, slices were incubated with the primary antibodies
   diluted in blocking solution (3% NDS, 0.5% TritonX-100) at
   4 °C^[256]77. As primary antibodies rabbit anti-GFAP (1:1000), mouse
   anti-Sox2 (1:500) goat anti-SOX9 (1:500) rat anti-Iba-I (1:800) rabbit
   anti-OTUD7b (1:1000) rabbit anti-PSTAT3 (1:1000) and rabbit anti-GFAP
   (1:1000) were used. After 2 × 15 min washing with PBS the slices were
   incubated with secondary antibodies diluted in blocking solution at
   4 °C overnight. As secondary antibodies Alexa488-conjugated donkey
   anti-goat (1:400), Alexa488-conjugated donkey anti-mouse (1:400),
   Cy3-conjugated donkey anti-rat (1:400) and Cy5-conjugated donkey
   anti-rabbit (1:400) were used. Subsequently, secondary antibodies were
   removed, and nuclei were stained using DAPI (1:10000). After 3 × 10 min
   washing in PBS, slides were mounted with Aquapolymount solution and
   stored at 4 °C. Images were taken using a Zeiss inverted Axio Observer
   seven with ApoTome.2 equipped with an Axiocam 503 and a Colibri 7 LED
   light source, and Zeiss LSM 780 with four lasers (405, 488, 559 and
   633 nm) and ×20, ×40 and ×63 objective lenses. For apotome acquisition
   on Zen2.6 pro software was used.

Primary astrocyte cultures and treatment

   Primary astrocytes were isolated from 1- to 2-day-old newborn mice and
   cultured in DMEM containing 1% glutamine, 10% FCS, and 1%
   penicillin/streptomycin as described before^[257]10. The purity of
   astrocyte cultures was more than 95%, as assessed by flow cytometry
   with antibodies against CD11b and ACSA-2. For the analysis of cytokine
   receptor-activated signaling pathways, astrocytes were stimulated at
   the indicated concentrations with TNF (10 ng/ml), IL-17 (50 ng/ml), and
   IFN-γ (10 ng/ml), respectively, for the indicated time points. For long
   term TNF treatment, primary astrocytes were stimulated with increasing
   concentrations of TNF (10 ng/mL, 20 ng/mL or 50 ng/mL) for 5 days.

CHX chase assay

   To detect the stability of GFAP protein, cells were stimulated with
   either TNF alone (50 ng/mL) or in combination with α-IL6 antibody
   (2 µg/mL) for 3 days. At d3 post-stimulation, cells were treated with
   10 µg/mL of cycloheximide (CHX) with/without 10 µM of proteasome
   inhibitor MG132 for 6 h.

siRNA transfection

   For siRNA-mediated knockdown of A20, primary astrocytes were
   transfected with 5 µM of A20-specific siRNA according to the
   manufacturer’s instructions. Thereafter, the cells were stimulated with
   50 ng/mL of TNF for 5 days, followed by protein isolation and WB
   analysis

Western blot

   Samples from primary astrocytes and mouse organs were lysed on ice in
   RIPA lysis buffer supplemented with PhosSTOP, phenylmethylsulfonyl
   fluoride (PSMF) and protease inhibitor cocktail. Cell lysates were
   pre-cleared by centrifugation at 10,000 g for 15 min at 4 °C.
   Supernatant was collected and quantified by BCA assay according to
   manufacturer’s protocol. Protein samples heated in lane marker reducing
   sample buffer at 99 °C for 5 min. Equal amounts of samples were
   separated by SDS-PAGE and subsequently transferred to polyvinylidene
   difluoride (PVDF) membranes, which were blocked with 5% BSA at room
   temperature for 1 h, followed by incubation with mentioned primary
   antibodies (Supplementary dataset 2) at 4 °C overnight.

   Blots were developed using the ECL Plus Kit and images were captured on
   Intas Chemo Cam Luminescent Image Analysis system (INTAS).
   Quantification and analysis of WB images was performed with the
   LabImage 1D software.

Co-immunoprecipitation (Co-IP)

   Whole cell lysates from astrocytes were precleared by incubation with
   GammaBind G Sepharose beads with gentle shaking at 4 °C for 2 h. After
   removal of beads by centrifugation, samples were incubated with
   specific antibodies under continuous shaking at 4 °C overnight.
   Following day, antibody-protein complexes were captured by incubating
   samples with GammaBind G Sepharose beads at 4 °C for 2 hours.
   Thereafter, the beads were washed with ice cold PBS thrice, resuspended
   in 2x lane marker reducing sample buffer and boiled at 99 °C for 5 min.
   Samples were centrifuged and supernatant was collected for WB analysis.

Quantitative RT-PCR

   Total mRNA was isolated from spinal cord tissue or astrocytes in buffer
   RLT using the RNeasy Mini Kit according to manufacturer’s protocol.
   mRNA was reverse-transcribed into cDNA with the SuperScript Reverse
   Transcriptase Kit. Quantitative RT-PCR was performed with a LightCycler
   480 system using TaqMan probes (Supplementary dataset [258]2). Gene
   expression levels were normalized to internal control Hprt and fold
   change increase in gene expression over naïve controls was calculated
   according to the ∆∆ cycle threshold (CT) method (Livak and Schmittgen,
   2001^[259]78).

Transcriptome analysis of isolated astrocytes

   Astrocytes were isolated by magnetic microbeads from spinal cords of
   Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice, respectively, at day 15
   p.i. Astrocytes isolated from naïve Otud7b^fl/fl and GFAP-cre
   Otud7b^fl/fl mice were used as control. For each experimental group,
   astrocytes from three mice (n = 3) were pooled to reduce biological
   variability, resulting in one sample per condition. RNA was isolated
   using RNeasy Mini Kit (Qiagen). RNA quality and integrity of total RNA
   was controlled on Agilent Technologies 2100 Bioanalyzer (Agilent
   Technologies). The RNA sequencing library was generated from 500 ng
   total RNA using Dynabeads mRNA DIRECT Micro Purification Kit (Thermo
   Fisher) for mRNA purification followed by ScriptSeq v2 RNA-Seq Library
   Preparation Kit (Epicenter) according to manufacture´s protocols. The
   libraries were sequenced on Illumina HiSeq2500 using TruSeq SBS Kit
   v3-HS (50 cycles, single ended run) with an average of 3 × 107 reads
   per RNA sample. FASTQ data quality was evaluated by FASTQC tool. Before
   alignment to reference genome each sequence in the raw FASTQ files were
   trimmed on base call quality and sequencing adapter contamination using
   Trim Galore. Reads shorter than 20 bp were removed from FASTQ file.
   Trimmed reads were aligned to the reference genome using open source
   short read aligner STAR. Feature counts were determined using R package
   Rsubread. Only genes showing counts greater than 5 at least two times
   across all samples were considered for further analysis. Gene
   annotation was done by R package bioMaRt. Data normalization was
   performed by calculating log2 counts per million (CPM) values and
   quantile normalization (edgeR). Based on the 20th percentile of
   distribution of the normalized log2 CPM values a CPM cutoff of 0.56 was
   derived to discriminate lowly abundant transcripts. Only transcripts
   with log2 CPM > 0.56 in at least one sample were used for further
   analysis. Differential expression between samples from Otud7b knockout
   versus Otud7b-competent mice isolated on day 0 and day 15 was
   calculated, respectively. Transcripts with fold change (FC) with
   |FC | > 3 were regarded relevant.

   The online platform DAVID Bioinformatics from National Institutes of
   Health (NIH) was used for the Gene Ontology and KEGG pathway enrichment
   analysis.

Spatial transcriptomics

   In situ RNA expression analysis at a single-cell level was performed on
   spinal cords of Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl mice,
   respectively, at day 15 p.i. using the Xenium system (10x Genomics).
   Astrocytes isolated from naïve Otud7b^fl/fl and GFAP-cre Otud7b^fl/fl
   mice were used as control (n = 3). 5μm thick sections were placed on a
   Xenium slide according to the manufacturer’s protocol, with drying at
   42 °C for 3 hours and overnight placement in a desiccator at room
   temperature, followed by deparaffinization and permeabilization to make
   the mRNA accessible. The Probe Hybridization Mix was prepared using a
   pre-designed panel with 247 genes (Xenium Mouse Brain Gene Expression
   Panel v1) and custom add-on panel with 50 genes (Xenium Custom Gene
   Expression panel, design ID: QZD68C) according to the user guide
   ([260]CG000582, Rev D, 10x Genomics). The staining for Xenium was
   performed using Xenium Nuclei Staining Buffer (10x Genomics product
   number: 2000762) as a part of the Xenium Slides & Sample Prep Reagents
   Kit (PN-1000460). Following the Xenium run, Hematoxylin and Eosin (H&E)
   staining was performed on the same section according to the Post-Xenium
   Analyzer H&E Staining user guide ([261]CG000613, Rev B, 10x Genomics).
   Spatial transcriptome data samples were processed using the VoltRon
   package ([262]GitHub - LandthalerLab/EAE_Otud7b). Differential
   expression was calculated using DESeq2 based on “pseudobulk” raw counts
   (i.e., per cell type and per sample, counts are added up for all
   cells). Count normalization per gene was done using the formula
   log((raw count for the gene)/(raw count sum for the cell) * (size
   factor) +1), with a size factor of 10000, and using natural logarithms.

Transfection of primary astrocytes

   Primary astrocytes were transfected with pCMV6-AC-GFP-OTUD7b expression
   plasmid using Lipofectamine 3000 (Thermo Fisher) according to the
   manufacturer’s instruction. After 48 hrs post transfection, cells were
   stimulated with 20 ng of TNF for 24 hr, followed by protein isolation
   and western blot analysis as described before.

Quantification and statistical analysis

   Quantification and analysis of WB images was performed with the
   LabImage 1D software. Statistical significance was determined by
   two-tailed t-tests using Prism v.10 (GraphPad) with p > 0.05=ns (not
   significant), *p ≤ 0.05, **p ≤ 0.01 and ***p ≤ 0.001. Error bars in the
   figure represent standard error of mean (SEM). N indicates the number
   of biological replicates. For spatial transcriptomics analysis adjusted
   p-values were calculated by DEseq2 using Benjamini-Hochberg corrections
   of two-sided Wald test p values.

Reporting summary

   Further information on research design is available in the [263]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [264]Supplementary Information^ (3.3MB, pdf)
   [265]41467_2025_65093_MOESM2_ESM.docx^ (11.5KB, docx)

   Description of additional supplementary file
   [266]Supplementary Dataset 1^ (22.7KB, xlsx)
   [267]Supplementary Dataset 2^ (31.9KB, docx)
   [268]Reporting Summary^ (1MB, pdf)
   [269]Transparent Peer review file^ (283.6KB, pdf)

Source data

   [270]Source data^ (31.1MB, xlsx)
   [271]Supplementary Source Data^ (9.1MB, docx)

Acknowledgements