Mice
Table of Contents
C57BL/6J, Aireflox/flox, Rag1−/−, Cd40−/− and Tcra−/− mice were obtained from Jackson Laboratories and bred in our facility. Aqp4−/− mice were provided by A. Verkman44, Aqp4flox/flox mice45 were provided by O. P. Ottersen, Foxn1-cre mice46 were provided by L. Klein, Mb1-cre mice47 were provided by M. Schmidt-Supprian and DEREG mice48 were provided by Tim Sparwasser. To generate mice with cell-type-specific excision of loxP-flanked cassettes, mice were bred with respective cre mice as specified. All mouse strains were on the C57BL/6J background. Mice were housed in a pathogen‐free facility at the Technical University of Munich. All experimental protocols were approved by the standing committee for experimentation with laboratory animals of the Bavarian state authorities and performed in accordance with the corresponding guidelines (ROB-55.2-2532.Vet_02-17-234, ROB-55.2-2532.Vet_02-20-01, ROB-55.2-2532.Vet_02-20-23, ROB-55.2-2532.Vet_02-21-154).
Generation of bone marrow chimeras
For the generation of bone marrow chimeras, recipient mice were lethally irradiated. A total dose of 9 Gray (Gy) was delivered as two 4.5 Gy doses 3 h apart. When Rag1−/− mice were used as bone marrow recipients, they were irradiated with 4.5 Gy only once. A total of 5 × 106 bone marrow cells was injected intravenously (i.v.) into recipients 1 day after irradiation. For generation of mixed bone marrow chimeras, bone marrow from two distinct donor mice was pooled and injected as specified. Reconstituted mice were maintained on antibiotic water (enrofloxacin, Bayer, 0.1 mg ml−1) for 2 weeks after cell transfer. Reconstitution of the haematopoietic system was tested in the peripheral blood.
Antigens
Mouse MOG(35–55) (MEVGWYRSPFSRVVHLYRNGK) and mouse AQP4(201–220) (HLFAINYTGASMNPARSFGP) were synthesized by Auspep or Biotrend, respectively. Human MOG protein was obtained from Biotrend, and mouse AQP4 protein was produced using a baculovirus-insect cell expression system and purified as previously described6.
EAE induction
Mice were immunized subcutaneously at the base of the tail with 200 μl of an emulsion containing 200 μg of MOG(35–55), 200 µg of AQP4(201–220), 100 µg of full-length AQP4 or 100 µg of full-length human MOG, all dissolved in PBS and emulsified with 250 µg Mycobacterium tuberculosis H37Ra (BD Difco) in mineral oil (CFA). Moreover, mice received 200 ng pertussis toxin (Sigma-Aldrich, P7208) i.v. on days 0 and 2 after immunization. Clinical signs of disease were monitored daily with scores as follows: 0, no disease; 1, loss of tail tone; 2, impaired righting; 3, paralysis of both hind limbs; 4, tetraplegia; 5, moribund state49.
Preparation of single-cell suspensions
Lymph nodes, spleens and thymi were passed through a 100 μm cell strainer (Greiner Bio-One), followed by gravity centrifugation (400g, 4 °C, 10 min). Spleen samples underwent erythrocyte lysis using BD Pharm Lyse (BD Biosciences).
Flow cytometry
Single-cell suspensions of lymphoid tissues were incubated with LIVE/DEAD fixable dyes (Aqua; 405 nm excitation) and mouse Fc Block in phosphate-buffered saline (PBS) for 15 min on ice. Cells were washed with FACS buffer (2% FCS in PBS) and incubated with antibodies against surface markers for 30 min on ice. For intracellular staining, cells were additionally fixed and permeabilized (Cytofix/Cytoperm and Perm/Wash Buffer; BD Biosciences) and stained with antibodies against intracellular markers overnight. A list of all of the antibodies used is provided in Supplementary Table 2.
Flow cytometry analysis was performed on the CytoFLEX flow cytometer (Beckman Coulter) with CytExpert (v.2.3.1.22) software or a FACSAria III (BD Biosciences) system with the BD FACSDiva (v.8.0.1) software, and flow cytometry data were analysed using FlowJo (v.10.5.1) software (BD Biosciences).
I-Ab–tetramer staining
For I-Ab–AQP4(205–215) and I-Ab–PLP(9–20) tetramer stainings, I-Ab tetramers were produced as described previously50,51. Cells were treated with 0.7 U ml–1 of neuraminidase (Sigma-Aldrich, N-2133) and 10 nM of dasatinib (Selleckchem) for 30 min at 37 °C for 30 min and 5% CO2, washed twice and treated with Fc block for 15 min on ice and subsequently stained with I-Ab tetramers for 2 h at room temperature with repeated resuspension, and the cells were finally surface and intracellularly stained for flow cytometry analysis. Thymocytes from naive mice were also enriched using anti-PE and anti-APC magnetic-activated cell sorting (MACS, Miltenyi Biotec) magnetic beads according to the manufacturer’s instructions before flow cytometry analysis.
Isolation of thymic APCs
TECs were isolated as described previously52. After euthanizing mice under deep anaesthesia by intracardial perfusion with PBS, thymi were dissected thoroughly to avoid blood contamination and adhering lymph nodes were removed immediately. Dissected thymi were placed into a six-well plate containing an enzyme solution with RPMI, 0.05% liberase TH (Sigma-Aldrich) and 100 U ml−1 DNase I (Sigma-Aldrich), and incubated at 37 °C with repeated mechanical dissociation by gently pipetting. After approximately 60 min, the supernatants were pooled, washed and passed through a 100 µm filter. Thymocytes were depleted with CD90.2 MACS beads according to the manufacturer’s instructions. Finally, TECs were FACS-sorted on live CD45–EPCAM+, thymic B cells on live CD45+EPCAM–CD19+ and thymic DCs on CD45+EPCAM–CD19–CD11c+MHC-IIhigh into PBS with 2% BSA.
Isolation of primary astrocytes
Primary astrocytes were isolated as previously described53. Brains from neonatal C57Bl/6J mice aged 1 to 3 days were dissected and cleaned from meninges, digested with DNase I (1 mg ml−1) and 0.25% trypsin-ethylenediaminetetraacetic acid (EDTA, Calbiochem) for 15 min, and passed through a cell strainer (70 μm). Single-cell suspensions were cultured at 37 °C on 175 cm2 cell culture flasks coated with 2 µg ml−1 poly-l-lysine (Sigma-Aldrich). After 7–10 days, the mixed glia cell culture reached confluence, and microglia were removed by sequentially shaking at 180 rpm for 30 min and 220 rpm for 2 h, changing the medium between and after the shaking steps.
qPCR
Total RNA was isolated from sorted cells using the RNAeasy Mini kit (Qiagen). The isolated RNA was transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Probes were purchased from Life Technologies and the assays were performed using the TaqMan Fast Advanced Master mix (Thermo Fisher Scientific) on 384-well reaction plates (Life Technologies). The quantitative PCR (qPCR) was performed on the Quant Studio 5 system (Life Technologies). In all of the experiments, GAPDH (Mm99999915_g1 and Hs00230829_m1, Thermo Fisher Scientific) was used as a reference gene to calibrate gene expression.
Histological analysis of mouse tissue
Mice were euthanized under deep anaesthesia by intracardial perfusion with PBS followed by perfusion with 4% (w/v) paraformaldehyde (PFA) dissolved in PBS. All of the organs were removed and fixed in 4% PFA overnight. Vertebral columns, including the spinal cords, were additionally decalcified with Osteosoft (Sigma-Aldrich) for 72 h before paraffin embedding; 2-μm-thick sections were prepared. Immunohistochemistry was performed using the Leica Bond Rxm System with the Polymer Refine detection kit (Leica). A list of all of the antibodies used is provided in Supplementary Table 2. DAB was used as chromogen, and counterstaining was performed with haematoxylin. The slides were then scanned on the Leica AT2 system, and the images were analysed using QuPath v.0.3.2 (https://qupath.github.io, University of Edinburgh, Scotland).
Quantification of histological samples was performed automatically with computer-assisted algorithms provided by QuPath. To detect the total cell counts, regions of interest were annotated and analysed automatically by positive nuclear detection. All annotations were performed in a blinded manner.
Histological analysis of human tissue
Human newborn thymic tissue was obtained from A. Büttner (approval by the local ethics committee, A2023-0038). Immunofluorescence was performed using EDTA buffer and steam cooking for antigen retrieval, followed by standard protocols. A list of all of the antibodies used is provided in Supplementary Table 2.
After deparaffinization, immunohistochemistry with anti-CD20 antibodies (DAKO) was conducted using 4 μm serial sections from formalin-fixed, paraffin-embedded (FFPE) tissues according to a standard protocol. Counterstaining was performed with 50% Gill’s haematoxylin 1 (American MasterTech), bluing with tap water and 0.02% ammonium hydroxide water. Slides were mounted with xylene and EcoMount (EcoMount, Biocare Medical).
To test colocalization of CD20 and AQP4, human thymus sections were sequentially costained for both markers after deparaffinization and blocking. Stained sections were analysed on an inverted TCS SP8 confocal microscope (Leica) using a HC PL APO CS2 ×40/1.30 NA objective. Confocal images were acquired as tiled image stacks with 1 µm z spacing and 3D image volumes were analysed using Imaris v.9.7 (Oxford Instruments).
B cell scRNA-seq analysis
In addition to surface staining for flow cytometry analysis, single-cell suspensions from different immune compartments (spleen, lymph node, bone marrow, thymus, blood) were labelled with TotalSeq-C anti-mouse Hashtags 1–3 and 5–6 (M1/42; 39-F11, BioLegend, 1:100 for all) according to the manufacturer’s instructions. A dump channel was defined by CD3, CD11b, F4/80 and NK1.1. From each compartment, 12.000 live dump−CD19+ cells were sorted separately into an FCS-coated 96-well plate.
scRNA-seq, scBCR-seq and cell hashing libraries were prepared using the 10x Chromium Single Cell 5′ Solution (Chromium Next GEM Single Cell VDJ v1.1 with feature barcoding technology for cell surface protein, 10x Genomics) as described previously54. In brief, sorted cells were transferred to the Chromium Next Gem chip (10x Genomics), and partitioning was performed automatically by the Chromium Controller (10x Genomics). Library preparation was performed according to the manufacturer’s instructions. For quality control and quantification, a Bioanalyzer 2100 (Agilent Technologies) was used. Finally, libraries were sequenced on the Illumina NovaSeq 6000 system as provided by Novogene.
Single-cell sequencing data processing
scRNA-seq reads, supplied by Novogene, were aligned to mouse reference genome mm10-2020-A using the Cell Ranger (v.7.1.0)55 count pipeline with the option ‘include introns’ disabled. Preprocessing, clustering, annotation and visualization were performed using SCANPY (v.1.9.2)56 according to established guidelines57. Specifically, for quality control, cells with more than 15% mitochondrial gene counts were excluded, as well as cells with more than 2% haemoglobin gene counts and cells with less than 5% ribosomal gene counts. Cells with less than 200 detected genes and cells with more than 1 × 104 counts per cell were also removed. Doublet exclusion was performed using scrublet (v.0.2.3)58. In total, quality control removed 1,926 cells. A total of 16,783 genes detected in less than three cells was excluded. To focus on subset variability beyond B cell specificity, 254 variable B cell receptor chain genes detected in the dataset were excluded. Gene counts were normalized to 1 × 104 total counts per cell, and variance was stabilized by log1p transformation. Data were not scaled to preserve the original weighing of gene expression and no regression was applied due to the low impact of confounding parameters in most compartments (that is, proliferation scoring and mitochondrial gene content). Highly variable genes for clustering were determined using the SCANPY default settings based on normalized dispersion, batched by organ, yielding 1,291 genes for further analysis. On the basis of these genes, principal component analysis was performed using the default SCANPY settings. A neighbourhood graph was computed based on 30 principal components and 20 neighbours. Clustering was performed using the Leiden algorithm (leidenalg package, v.0.9.1) with a resolution of r = 0.7 and UMAP dimensionality reduction was computed using the default SCANPY settings.
Gene signature scores were calculated in SCANPY using the default settings. Relevant gene signatures were as follows: genes upregulated by ex vivo B cells 2 h after exposure to CD40L40: Tnfaip3, Bcl2l1, Bcl2a1a, Gadd45b, Gadd45g, Fas, Slc16a1, Slc19a1, Prps1, St3gal6, Hmgcr, Ldlr, Fasn, Ccnd2, Stat5a, Nfkb1, Myc, Irf4, Cr2, Cd44, Il2ra, Ebi3, Lilrb4a, Fcer2a, Gpr65, Il1b, Traf1, Nfkbia, Nfkbib, Marcksl1, Icam1, Cd83, Jarid2, Bhlhe40, Gm4736* and Tfg. Gene signature of mouse splenic GC light zone B cells41,42: Cd52, Cd83, Hspd1, Ran, Mif, Atp5b, Myc, Dkc1, Lrrc58 and H2-Aa. Genes annotated with an asterisk were part of the published signature but not detected in our dataset.
For cell trajectory inference, spliced/unspliced reads were generated from Cell Ranger-aligned sequences using the velocyto (v.0.17.17) run10x pipeline59. Data were merged with gene expression analysis outlined above using scvelo (v.0.2.5)60 and trajectories were derived using UniTVelo (v.0.2.5.2)61 configured to run the model based on 1,500 top variable genes. For trajectory inference, the highly proliferative and transcriptionally active bone marrow clusters 6 and 7, respectively, were excluded from the dataset.
Isolation of B cell subsets
Single-cell suspensions from primary and secondary lymphoid tissues were enriched for CD19 with MACS beads according to the manufacturer’s instructions, followed by surface staining for flow cytometry cell sorting. For sorting CD19+ cells from thymi and bone marrow, we defined three distinct groups on the basis of the expression of IgM and IgD. Double-positive cells were defined as live CD19+B220+IgM+IgD+ cells. Furthermore, we isolated CD19+B220+IgM+IgD− and double-negative B cells, defined as CD19+B220+IgM−IgD− cells. For sorting CD19+ cells from secondary lymphoid tissues, we defined four groups on the basis of their maturation status. Naive B cells were specified as live CD19+B220+IgD+CD21+CD95−GL7− cells, GC B cells as live CD19+B220+CD95+GL7+ cells, marginal zone B cells as live CD19+B220+IgD−CD21+CD95−GL7− cells and memory B cells as live CD19+B220+IgD−CD21−CD95−GL7− cells. RNA from sorted cells was immediately isolated and processed for qPCR analysis as described above.
Primary mouse B cell cultures
Single-cell suspensions from secondary lymphoid tissues were FACS-sorted for live CD19+ cells and cultured at 37 °C and 5% CO2 for 2 days before RNA was isolated for qPCR. For stimulation, different combinations comprising 50 µg ml−1 anti-mouse CD40 (FGK4.5, BioXCell), 20 ng µl−1 recombinant IL-21 (Miltenyi Biotec), 10 µg ml−1 goat anti-mouse IgG + IgM (H+L), 10 µg ml−1 goat anti-human IgG (H+L, both Jackson Immuno Research) and 1 µg ml−1 LPS were used.
Primary human B cell cultures and stimulation
Human tonsillar tissue was obtained from routine tonsillectomies by the Department of Otorhinolaryngology of the University Hospital Klinikum rechts der Isar of the Technical University of Munich School of Medicine with patients’ informed consent. Single-cell suspensions were prepared from freshly collected human tonsil tissue and frozen. In brief, tonsil tissue was cut into small pieces and disaggregated through a cell strainer. After washing, cells were purified using Histopaque-1077 Hybri-Max (Sigma-Aldrich) gradient centrifugation according to the manufacturer’s instructions. Single-cell suspensions from human tonsils were FACS-sorted for live CD19+ cells and cultured at 37 °C and 5% CO2 for 2 days before RNA was isolated for qPCR. For sorting different B cell subsets, we defined three distinct groups based on the expression of CD38, CD27, IgD and CD10. Human naive B cells were defined as live CD3−CD19+CD27−CD38+ cells, memory B cells as live CD3−CD19+CD27+CD38− cells and GC B cells as live CD3−CD19+CD27+CD38+ cells. RNA from sorted cells was isolated and processed for qPCR analysis as described above.
For the in vitro stimulation of human CD19+ cells, we used a coculture system with immortalized follicular dendritic cells (YKL) that were equipped with membrane-bound CD40L (CD40Lg) kindly provided by D. Hodson62. CD15 MACS-beads-depleted human peripheral blood mononuclear cells were FACS-sorted for live CD3−CD19+CD27− cells and seeded on irradiated YKL cell layers for 5 days. For comparability, FACS-sorted cells were also seeded onto control YKL cells transduced with an empty vector instead of the CD40L vector. After 5 days, RNA was isolated and processed for qPCR analysis as described above.
Recall assay
To test AQP4-specific recall responses, we used a system established previously63, based on a T cell hybridoma cell line (A5 cells) that was equipped with a GFP reporter linked to NFAT—a downstream transcription factor of the IL-2 signalling pathway64. We transfected these cells with a high-affinity AQP4-reactive TCR (clone 4 or clone 6) to test APCs for their endogenous presentation of AQP4. The generation of AQP4-TCR-transduced A5 cells is described in “Transduction of A5 cells with AQP4-specific TCRs” below. As controls, we added either P41 exogenously (0.3 and 1.0 µg ml−1), anti-CD3 (1 µg ml−1) and/or anti-mouse MHC-II (I-A and I-E) blocking antibody (5 µg ml−1) to the coculture. We determined the fraction of NFAT–GFP+ cells after 20–24 h by flow cytometry analysis.
Bulk RNA-seq
Total RNA was isolated from FACS-sorted whole-thymic B cells and thymic B cell subsets using AmpureXP beads (Beckman Coulter). Library preparation for bulk-sequencing of poly(A)-RNA was performed as described previously65. In brief, barcoded cDNA of each sample was generated with Maxima RT polymerase (Thermo Fisher Scientific, EP0742) using oligo-dT primers containing barcodes, unique molecular identifiers (UMIs) and an adapter. The 5′-ends of the cDNAs were extended by a template switch oligo (TSO) and full-length cDNA was amplified with primers binding to the TSO-site and the adapter. The NEB UltraII FS kit was used to fragment cDNA. After end repair and A-tailing, a TruSeq adapter was ligated and 3′-end fragments were finally amplified using primers with Illumina P5 and P7 overhangs. In comparison to previous descriptions65, the P5 and P7 sites were exchanged to enable sequencing of the cDNA in read1 and barcodes and UMIs in read2 to achieve a better cluster recognition. The library was sequenced on the NextSeq 500 (Illumina) system with 65 cycles for the cDNA in read1 and 19 cycles for the barcodes and UMIs in read2. Data were processed using the published Drop-seq pipeline (v.1.12) to generate sample- and gene-wise UMI tables66. Reference genome (GRCm38) was used for alignment. Transcript and gene definitions were used according to GENCODE version M25.
Bulk RNA-seq data processing
Raw counts of two individual sequencing runs of the same library were merged, and non-overlapping genes were dropped. Differential expression analysis was performed using the EdgeR package (v.3.40.2)67. After excluding low-expressed genes (that is, genes for which an expression threshold of greater than 1 count per million was not attained in at least 4 samples), the negative binomial model was fitted. The resulting P values were adjusted for multiple testing using the FDR correction. To increase the power, we limited our analysis to genes in the serumantibodyome43. Genes were considered to be differentially expressed if they had a less than 5% probability of being false positive (Padj < 0.05). PCA was performed using the PCA function of the FactoMineR package on log-transformed counts per million (log[CPM]). Gene set enrichment analysis (GSEA) was performed on unfiltered DESeq2 normalized count data using the DESeq2 package (v.1.40.2)68 and GSEA v.4.3.2 software69,70 in conjunction with MSigDB (v.2023.1). The interrogated gene sets were derived from the M8 collection of cell type signature gene sets. Analysis was run with permutation-type phenotype and a FDR of 0.25.
Generation of AQP4 TCR retrogenic mice
Retroviruses containing a high-affinity AQP4-specific TCR were produced by calcium phosphate precipitation of Platinum-E virus packaging cells with retroviral vectors (pMP71, a gift from D. Busch). To generate retrogenic mice, bone marrow was collected from the tibia and femur of 8–20-week-old Cd45.1+/− × Rag1−/− mice. After red blood cell lysis, single-cell suspensions were stained with anti-mouse Ly6A/E (Sca-1, 1:300) and anti-mouse CD3/CD19 (1:300) antibodies. FACS-sorted SCA1+CD3−CD19− stem cells were expanded with mouse IL-3 (2 ng ml−1), IL-6 (50 ng ml−1) and SCF (50 ng ml−1) for 3–4 days. Retroviral transduction of expanded stem cells was achieved by spinoculation71. In brief, 400 µl of Platinum-E supernatants were centrifuged in a retronectin-coated 48-well plate at 3,000g for 2 h at 32 °C. Then, 200 µl of the medium was removed and filled up with expanded stem cells to a final concentration of 50,000 cells per 400 µl. After 2 days of culture, transduced stem cells were injected into irradiated recipient mice.
Generation of mixed bone marrow chimeras
To assess the negative selection of AQP4-specific thymocytes in vivo, mixed bone marrow chimeras were generated by grafting congenically marked Rag1−/− bone marrow engineered to retrogenically express an AQP4-specific TCR (clone 6) along with bone marrow from either WT or Aqp4ΔB mice (4:1) into lethally irradiated Aqp4−/− recipients so that maturing thymocytes from the polyclonal and the retrogenic (AQP4-specific) thymic compartments could be compared in the same host mouse. This setup also ensured that maturing thymocytes from both the polyclonal and retrogenic (AQP4-specific) thymic compartments encountered a thymic environment with either WT or AQP4-deficient B cells. AQP4 TCR retrogenic mice were generated as described earlier. Stem cells from the complementary donor were isolated, ex vivo stimulated in parallel and mixed with the retrogenic compartment before injection. Reconstitution was tested 5 to 6 weeks after transplantation.
GC characterization
For characterization of GC reactions, mice were immunized with 200 µl emulsion containing 50 µg of either full-length AQP4 or recombinant human MOG protein. Sera were collected before (day −1) as well as on day 10 and day 21 after immunization. On day 21, the mice were euthanized for further analysis. Half of the draining lymph nodes and the spleen were isolated for flow cytometry analysis before mice were perfused with 4% PFA. The remaining half of the spleen and draining lymph nodes were isolated for histology. TFH cells were quantified using flow cytometry analysis of live CD3+CD4+PD-1+BCL6+ cells. GC B cells were quantified histologically by immunoreactivity to BCL6 and by flow cytometry analysis of live CD3−CD19+B220+CD95+BCL6+ cells.
Anti-AQP4 and anti-MOG antibody detection assay
To detect anti-AQP4 or anti-MOG antibodies in the sera of protein-immunized mice, a cell-based flow cytometry assay was used72,73. Sera were diluted (1:50) in RPMI 1640 growth medium and added to a 96-well plate containing 30,000 LN18AQP4 or LN18MOG cells per well. As a control, every serum was tested on LN18CTRL cells (transduced with an empty vector). The plate was incubated on ice on an orbital shaker for 25 min. Cells were washed twice with FACS buffer. To stain mice for LN18-cell bound mouse IgG, 50 μl of diluted (1:100 in washing buffer) Alexa-Fluor-488-labelled goat anti-mouse IgG H+L (Life Technologies, Thermo Fisher Scientific) was added to each well. After incubation for 25 min on ice, cells were washed twice with FACS buffer.
Statistical analysis
Statistical evaluations of cell frequency measurements and cell numbers were performed using one-way-ANOVA and post hoc tests when more than two populations were compared. Two-way ANOVA followed by post hoc multiple comparison tests was used as indicated in the figure legends. Correlation and linear regression were calculated for fractions of thymic CD4 SP and thymic B cells. EAE incidence was calculated using Kaplan–Meier analysis and the P values were analysed with a log-rank test (Mantel–Cox). Age at immunization, day of disease onset, peak EAE scores and cumulative EAE scores were compared using Mann–Whitney U-tests. P < 0.05 was considered to be significant. Calculations and the generation of graphs were performed using Graph Pad Prism v.10.9.0 (GraphPad), RStudio (v.2023.06.01) and Python (v.3.9.16). Figures were prepared using Adobe Illustrator 2022 (v.26.0.1).
Materials and reagents
An extensive list of materials and reagents is provided in Supplementary Table 2.
Mouse histology heat maps
Heat maps representing CD19 density in mouse thymi were created using Visiopharm (v.2023.1). First, tissue regions with high expression of EPCAM were identified and outlined based on the marker intensity. CD19+ cells were identified based on DAPI and CD19 expression. For heat map creation, CD19+ cells overlapping within a radius of 50 µm were considered. Figures were prepared using Adobe Photoshop CS6 (v.13.0).
Analysis of published thymic scRNA-seq data for mouse and human
Annotated matrix files (h5ad) with precalculated UMAPs and cell type classifications of the original authors16 were downloaded from Zenodo (https://doi.org/10.5281/zenodo.5500511). All cells with any Aqp4 or AQP4 expression greater than zero were included in the analysis of positive cells. An MHC-II score was generated in SCANPY to summarize mouse H2-Aa, H2-Ab1, H2-Eb1, H2-Eb2 and human HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DQA2, HLA-DQB1, HLA-DQB2, HLA-DRA, HLA-DRB1 and HLA-DRB5 expression. Pre-annotated classifications as stored in the dataset’s .obs data frame ‘cell types’ and ‘Anno_level_fig1’ were used for mouse and human cell type labelling, respectively.
Transduction of A5 cells with AQP4-specific TCRs
Retroviruses containing AQP4-specific TCRs were produced by calcium phosphate precipitation of Platinum-E virus packaging cells with retroviral vectors (pMP71) as described previously71. Virus-containing supernatants were collected after 2 days, centrifuged to dispose of cell debris, and either used immediately for spinoculation or kept at 4 °C for a maximum of 4 weeks. For spinoculation, 400 µl of Platinum-E supernatants were centrifuged in a 48-well plate at 3,000g for 2 h at 32 °C. Then, 200 µl medium was removed and filled up with A5 cells to a final concentration of 50,000 cells per 400 µl. After 2 days of culture, the expression of TCR alpha and beta variable chains in transfected A5 cells was tested by flow cytometry analysis. A5 cells expressing both chains of the respective TCR were then enriched by FACS sorting, stimulated with P41-pulsed APCs and tested for their NFAT–GFP expression after stimulation using flow cytometry.
Treg cell depletion
DEREG mice provide an efficient in vivo model for inducible Treg cell depletion48. This model was investigated in EAE previously and required some modifications of the immunization procedure74. In contrast to the earlier described immunization protocol, mice received a reduced amount of pertussis toxin (50 ng intravenously on days 0 and 2 after immunization) and 0.5 µg of diphtheria toxin intraperitoneally (i.p.) two days before and on days 5 and 6 after immunization.
Transfer of mature T cells into Tcra
−
/
− mice
Mature CD4+ T cells were isolated from unmanipulated Aqp4ΔB mice by CD4 MACS bead enrichment (Miltenyi Biotec). In total, 5 × 106 T cells were injected i.v. into Tcra−/− mice. Then, 1 day after transfer, Tcra−/− recipient mice were immunized as indicated and analysed for AQP4-specific T cell frequencies in the transferred T cell repertoire by P41–I-Ab tetramer staining 2 days after EAE onset.
Adoptive transfer EAE
A total of 5 × 106 FACS-sorted CD4+ T cells was transferred intravenously into Tcra−/− mice (day 0), followed by subcutaneous immunization at both flanks with an emulsion containing PBS and CFA (day 1). Mice were scored and weighed daily before they were euthanized on day 32 after immunization. Serum was preserved after cardiac blood collection and tested for autoantibodies.
Autoantibody assay
To screen for autoantibodies, the preserved mouse sera were tested on histological cryosections from lymphocyte-deficient Rag1−/− mice using secondary anti-mouse IgG (H+L) AF488 antibodies (Thermo Fisher Scientific). Mice were euthanized by sequential intracardial perfusion with ice-cold PBS followed by 4% PFA. Dissected organs were then embedded in Tissue-Tek O.C.T. compound and immediately frozen with liquid nitrogen. After fixation with precooled acetone (−20 °C) for 1 min and subsequent blocking with normal goat serum (30% dilution in PBS) for 10 min, 10 µm cryosections were incubated with preserved mouse sera (1:50) for 1 h. The samples were next treated with secondary anti-mouse IgG (H+L) AF488 antibodies (1:500) for 30 min. Finally, the sections were washed and mounted with ProLong Gold Antifade Mounting reagent containing DAPI (Thermo Fisher Scientific, P36931). The staining was performed using the Leica Bond RXm device, all washing steps were performed with Bond Wash solution (Leica, AR9590) and all dilutions were prepared using Bond Primary Antibody Diluent (Leica, AR9352).
Quantification of AQP4 loss
AQP4 loss was determined in a semi-quantitative approach counting the extent of adjacent CNS lesions. The extent of CNS lesions was measured automatically using QuPath’s positive nuclear detection of CD45+ cells and their engaging area. The AQP4 signal was calculated using QuPath’s positive pixel count algorithm in the adjacent area with a defined radius of 100 µm to CD45+ infiltrates (region of interest). For reasons of comparison, the AQP4 signal was normalized to the extent of CNS lesions. All annotations were performed in a blinded manner.
Licenses
Parts of Figs. 2 and 4, as well as Extended Data Figs. 3, 6, 7 and 9 were drawn using pictures from Servier Medical Art. Servier Medical Art by Servier is licensed under a Creative Commons licence CC BY 3.0. Parts from Extended Data Fig. 9a were created using BioRender.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.