Myelin-reactive B cells exacerbate CD4+ T cell-driven CNS autoimmunity in an IL-23-dependent manner

Abstract

B cells and T cells collaborate in multiple sclerosis (MS) pathogenesis. IgH^[MOG] mice possess a B cell repertoire skewed to recognize myelin oligodendrocyte glycoprotein (MOG). Here, we show that upon immunization with the T cell-obligate autoantigen, MOG_[35-55], IgH^[MOG] mice develop rapid and exacerbated experimental autoimmune encephalomyelitis (EAE) relative to wildtype (WT) counterparts, characterized by aggregation of T and B cells in the IgH^[MOG] meninges and by CD4⁺ T helper 17 (Th17) cells in the CNS. Production of the Th17 maintenance factor IL-23 is observed from IgH^[MOG] CNS-infiltrating and meningeal B cells, and in vivo blockade of IL-23p19 attenuates disease severity in IgH^[MOG] mice. In the CNS parenchyma and dura mater of IgH^[MOG] mice, we observe an increased frequency of CD4⁺PD-1⁺CXCR5^- T cells that share numerous characteristics with the recently described T peripheral helper (Tph) cell subset. Further, CNS-infiltrating B and Tph cells from IgH^[MOG] mice show increased reactive oxygen species (ROS) production. Meningeal inflammation, Tph-like cell accumulation in the CNS and B/Tph cell production of ROS were all reduced upon p19 blockade. Altogether, MOG-specific B cells promote autoimmune inflammation of the CNS parenchyma and meninges in an IL-23-dependent manner.

Fig. 1. Active immunization with MOG_[35-55] induces severe EAE in IgH^[MOG] mice.

a Left, Disease curves and linear regression of males (WT, n = 8, IgH^[MOG], n = 8) and females (WT, n = 10, IgH^[MOG], n = 10) immunized with MOG_[35-55] and monitored for development of EAE. Linear regression was used for statistical analysis. Representative of > 20 experiments conducted in each sex. Right, area under curve (AUC) analysis of disease curves in left panels. Two-way ANOVA was used to assess the contribution of genotype. The contribution of sex was p < 0.9912 and there was no interaction effect between the variables. b Spinal cords from female WT and IgH^[MOG] mice at endpoints were sectioned and stained with hematoxylin & eosin (H&E; immune infiltration) and Luxol fast blue (LFB; myelin). 10X magnification. Representative of n = 4 mice of each genotype from a single immunization. Quantification in Supplementary Fig. 1b. c Draining LN cells were isolated at disease onset from female WT (n = 3) and IgH^[MOG] (n = 3) mice from a single immunization and IFNγ and IL-17 expression were determined by flow cytometry. Gated on live CD4⁺CD44⁺ events. Two-tailed t-test was used. d Female WT or IgH^[MOG] B cells were co-cultured with CellTrace Violet-labeled MOG_[35-55]-specific 1C6 T cells that were pulsed, or not, with the indicated concentrations of MOG_[35-55]. Gate frequencies indicate the percentage of cells undergoing at least one cell division. Representative of cultures derived from individual mice (n = 3 mice of each genotype). Data is from one experiment. e Sera were collected from male WT (n = 5) and IgH^[MOG] (n = 5) mice at endpoints and the concentration of MOG-specific IgG was assessed by ELISA. Representative of one of two immunizations. A two-tailed t-test was used. In all violin plots, horizontal lines represent medians. Each datapoint represents an individual mouse. Exact p-values can be found in the Source Data file. Linked to Supplementary Fig. 1.

Fig. 2. Increased lymphocytic aggregation in the meninges of immunized IgH^[MOG] mice.

a Brain and cerebellar meningeal sections from female MOG_[35-55]-immunized WT (n = 8) or IgH^[MOG] (n = 9) mice were assessed by immunofluorescence (IF) for expression of B220 (red), CD4 (blue), and CD11c (green) at endpoints, and lymphocyte aggregates were enumerated. Data are from two immunizations. Two-tailed t-test was used. In violin plot, horizontal line represents median. Each datapoint represents an individual mouse. b Assessment of the ECM near meninges in the cerebellum was determined by IF. Expansion on the meningeal ECM is shown by fibronectin (green) and PDGFRα/β (blue) which overlap with B220 (red) B cell clusters. Dotted lines outline the meningeal membrane and (*) denotes vascular endothelium. Representative of n = 4 female mice of each genotype. Data are from two immunizations. c Representative IF images of IgH^[MOG] meningeal lymphocytic clusters, stained for CD138 (yellow), AID (green), B220 (red), DAPI (blue). i., CD138⁺B220⁺, ii., CD138^-AID⁺, iii., CD138⁺B220^-, iv., CD138⁺AID⁺. The source meningeal cluster is presented in Supplementary Fig. 2c. Representative of 4 female mice. Data are from two immunizations. Exact p-values can be found in the Source Data file. Linked to Supplementary Fig. 2.

Fig. 3. Inflammatory B cells facilitate Th17 responses in IgH^[MOG] mice.

a The frequency of mature (IgM^hiIgD^hi), transitional (IgM^hiIgD^mid) and class-switched (CS; IgM^-IgD^-) B cells were assessed from female WT (n = 5) and IgH^[MOG] (n = 5) CNS at endpoints. Gated on live B220⁺ events. Representative of one of two immunizations. Two-tailed t-test was used. b Expression of the indicated cytokines was assessed from female WT (n = 5) and IgH^[MOG] (n = 6) CNS B cells at endpoints. Gated on live CD19⁺ events. Representative of one of two immunizations. Two-tailed t-test was used. c Female WT (n = 5) and IgH^[MOG] (n = 5) were immunized with MOG_[35-55] and CNS-infiltrating CD4⁺ T cells were isolated at endpoints. Expression of the indicated cytokines was assessed by intracellular flow cytometry. Gated on live CD4⁺ events. Representative of one of two immunizations. Two-tailed t-test was used. d Female NOD.Scid mice were passively infused with 2 × 10⁶ female WT or IgH^[MOG] B cells, and received 2×10⁶ 1C6 Th17 cells 7 days later. Mice were monitored for signs of EAE. Left, disease curve and linear regression from one of two transfer experiments; right, AUC comparison of mice from both transfers (n = 7, both groups of recipient mice). Two-tailed t-test was used. e CNS-infiltrating CD4⁺ T cells were isolated from adoptive transfer recipients in (d) at endpoints and the indicated cytokines were assessed by flow cytometry. Two-tailed t-test was used. Representative plots for data in Fig. 3e are presented in Supplementary Fig. 3g. In all violin plots, horizontal lines represent medians. Each datapoint represents an individual mouse. Exact p-values can be found in the Source Data file. Linked to Supplementary Fig. 3.

Fig. 4. CNS IgH^[MOG] B cells upregulate IL-23.

a Male WT (n = 6) and IgH^[MOG] (n = 6) mice were immunized with MOG_[35-55] and B cells were isolated from CNS tissues at endpoints. Expression of IL-23 heterodimer was assessed by intracellular flow cytometry. Data are presented from one of four immunizations. Two-tailed t-test was used. b IF detection of IL-23 (green), B220 (red), CD4 (aqua), DAPI (blue) in the meninges of female MOG_[35-55]-immunized WT or IgH^[MOG] mice. Representative of n = 3 mice of each genotype from a single immunization. c Female IgH^[MOG] mice were immunized with MOG_[35-55] and were treated on day (-1) and day 6 with 1 mg anti-p19 (n = 5) or rIgG1 isotype (n = 5). Disease curves representative of one of seven experiments. Two-tailed t-test was used. d Cerebellar meningeal sections from MOG_[35-55]-immunized female IgH^[MOG] mice, treated with isotype (n = 6) or anti-p19 (n = 6), were assessed for lymphocyte aggregation. Data are from two immunizations. Two-tailed t-test was used. e CNS-infiltrating mature, transitional and class-switched B cells were assessed by flow cytometry from anti-p19 (n = 4) or rIgG1 isotype (n = 4)-treated immunized female IgH^[MOG] mice. Gated on live B220⁺ events. Data are from a single immunization. Two-tailed t-test was used. f CD4⁺ T cells were isolated from p19 (n = 5) or rIgG1 isotype (n = 5)-treated immunized female IgH^[MOG] mice and the indicated cytokines were assessed by flow cytometry. Data are representative of one of two immunizations. Two-tailed t-test was used. g IL23A was assessed from CSF immune cells of MS-affected individuals (see individual information on pwMS in Supplementary Table 1) by qPCR and was correlated to the frequency of CD19⁺ B cells within the CD45⁺ CSF leukocyte population of each patient. Linear regression analysis was applied to calculate Spearman r and p-value. In all violin plots, horizontal lines represent medians. Each datapoint represents a biological replicate. Exact p-values can be found in the Source Data file. Linked to Supplementary Fig. 4.

Fig. 5. Tph cells infiltrate the IgH^[MOG] CNS.

a Male WT (n = 5) and IgH^[MOG] (n = 5) mice were immunized with MOG_[35-55] and T cells were isolated from CNS at endpoints. Expression of PD-1 and CXCR5 was assessed by flow cytometry. Data are representative of seven experiments. Two-tailed t-test was used. b Male WT or IgH^[MOG] mice were immunized and sacrificed at onset (n = 4 each), peak (n = 3 each), recovery (n = 3 each). CNS-infiltrating CD4⁺ T cells were isolated from these groups, as well as from mice that did not develop disease (n = 4 each) and the frequency of PD-1⁺CXCR5^- cells was assessed. Data are from one of two immunizations. Tukey’s post-hoc test after one-way ANOVA was used. c Expression of MHC class II (I–Ag⁷), IL-21, ICOS and CXCL13 were assessed from CNS-infiltrating CD4⁺ T cells taken from immunized male WT (n = 5) and IgH^[MOG] (n = 5) mice, and expression between PD-1^- and PD-1⁺ CD4⁺ subpopulations was compared. Data are representative of two immunizations. Representative plots are presented in Supplementary Fig. 5c. Sidak’s multiple comparisons test after repeated measures two-way ANOVA was used. d Expression of IFNγ and IL-17 was assessed from PD-1^- and PD-1⁺ CD4⁺ T cells from immunized male WT or IgH^[MOG] mice (n = 3 each). Data are representative of two immunizations. Sidak’s multiple comparisons test after repeated measures two-way ANOVA was used. e Male WT (n = 5) and IgH^[MOG] (n = 5) mice were immunized with MOG_[35-55] and T cells were isolated from the dura mater and subdural meninges (SDM) at endpoints. Expression of PD-1 and CXCR5 was assessed by flow cytometry. Data are representative of two experiments for dura mater and one immunization for SDM. Two-tailed t-test was used. f T cells were isolated from spleen and CNS at endpoints from anti-p19 (n = 5) or isotype-treated (n = 4) immunized male IgH^[MOG] mice. Expression of PD-1 and CXCR5 was assessed by flow cytometry. Data are representative of three experiments. Two-tailed t-test was used. In all violin plots, horizontal lines represent medians. Each datapoint represents an individual mouse. Datapoints connected by a line represent paired observations from the same mouse. Exact p-values can be found in the Source Data file. Linked to Supplementary Figs. 5, 6.

Fig. 6. CNS-infiltrating IgH^[MOG] B cells and Tph cells show increased production of reactive oxygen species.

a GO term analysis of the transcriptome of CNS-infiltrating B cells and Tph cells from female IgH^[MOG] (n = 3) vs WT (n = 3) from a single immunization. b Normalized enrichment scores (NES) for “pathways of neurodegeneration” (mmu05022) and “oxidative phosphorylation” (mmu00190) from B cell and Tph cell transcriptomes from the same samples as in (a). The weighted Kolmogorov–Smirnov statistical test with the Benjamini-Hochberg method was used to adjust for the false discovery rate. Uptake of ROS indicator CM-H2DCFDA from B cells (c, e) or Tph cells (d, f) from immunized female IgH^[MOG] vs WT mice (c, d; n = 5 each) or from immunized and anti-p19 (n = 5) vs isotype-treated (n = 6) immunized female IgH^[MOG] mice. Data in c, d are representative of two immunizations. Data in e, f are representative of two immunizations. Two-tailed t-test was used. In all violin plots, horizontal lines represent medians. Each datapoint represents an individual mouse. Exact p-values can be found in the Source Data file. Linked to Supplementary Fig. 7.