Turncoat antibodies unmasked in a model of autoimmune demyelination: from biology to therapy

Abstract

Autoantibodies contribute to many autoimmune diseases, yet there is no approved therapy to neutralize them selectively. A popular mouse model, experimental autoimmune encephalomyelitis (EAE), could serve to develop such a therapy, provided we can better understand the nature and importance of the autoantibodies involved. Here we report the discovery of autoantibody-secreting extrafollicular plasmablasts in EAE induced with specific myelin oligodendrocyte glycoprotein (MOG) antigens. Single-cell RNA sequencing reveals that these cells produce non-affinity-matured IgG antibodies. These include pathogenic antibodies competing for shared binding space on MOG's extracellular domain. Interestingly, the synthetic anti-MOG antibody 8-18C5 can prevent the binding of pathogenic antibodies from either EAE mice or people with MOG antibody disease (MOGAD). Moreover, an 8-18C5 variant carrying the NNAS mutation, which inactivates its effector functions, can reduce EAE severity and promote functional recovery. In brief, this study provides not only a comprehensive characterization of the humoral response in EAE models, but also a proof of concept for a novel therapy to antagonize pathogenic anti-MOG antibodies.

Keywords: Autoimmunity; CD138; antibody engineering; antibody-secreting cell; autoantibody antagonist; demyelinating autoimmune disease; live cell-based assay; mass cytometry; meningeal lymphoid follicle.

Figure 1.

Extrafollicular plasmablasts expand greatly and transiently in EAE induced with B cell-dependent antigens. (A) Representative flow cytometry plot of lymph node cells from a mouse immunized with bMOG 8 days earlier. Gating strategy used to identify plasmalasts (CD138^hiCD19^+/−) and B cells (CD138⁻CD19⁺) is shown. Dead cells, doublets, and other leukocytes (CD3⁺, Ly6G⁺, CD11b⁺, CD11c⁺) were excluded. (B) Quantification of plasmablasts in different tissues and at different time points after bMOG immunization. *Significantly different from the other time points (ANOVA, P < 0.0001; post hoc Tukey’s test, P < 0.0001). Sample size: 4 day 0, 11 day 8, 10 day 16, 10 day 24. (C, D) Quantification of plasmablasts in lymph nodes 8 days after immunization with B cell-dependent (bMOG, hMOG) or -independent (MOG_35–55) antigens. Mice untreated (naive) or injected with adjuvants alone (sham) were used as controls. *Significantly different from the controls (C: Kruskal-Wallis, P = 0.014; post hoc Dunn’s test, P ≤ 0.018; D: ANOVA, P < 0.0001; post hoc two-tailed Student’s t-test, P < 0.0001). Sample size: C, 3 naive, 10 hMOG_1–125, 10 bMOG; D, 2 naive, 3 sham, 5 MOG_35–55, 5 bMOG. (E) Flow cytometric gating strategy used to identify plasmablasts and B cells expressing the IL-10 reporter EGFP in lymph nodes of Vert-X^+/+ mice 8 days after bMOG immunization. (F, G) Percentage and relative number of plasmablasts and B cells expressing IL-10 in lymph nodes of Vert-X^+/+ mice after the indicated treatments. *Significantly different from the other groups (ANOVA, P ≤ 0.01; post hoc two-tailed Student’s t-test, P ≤ 0.05). ^†Tends to be significant (ANOVA, P = 0.0588; Student’s t-test: P ≤ 0.0477). Sample size: 2 naive, 3 sham, 5 MOG_35–55, 5 bMOG. (H) Confocal images at low (top) or high (bottom) magnification showing the distribution of CD138⁺ plasmablasts (red) in a lymph node of a CD19^cre × Ai14 mouse expressing the B cell reporter tdTomato (false color green) on day 8 after bMOG immunization. Note that these plasmablasts: 1) are predominantly located near lymphatic vessels (blue) outside the follicles (delineated by dashed lines); and 2) express lower amount of tdTomato than CD138^– B cells. Arrowheads: tdTomato^loCD138⁺ plasmablasts. Arrows: tdTomato^hiCD138^– B cells. Abbreviation: GC, germinal center. Scale bars: top, 100 μm; bottom, 10 μm. (I) Counts of plasmablasts and T cells in the spinal cord meninges from mice sacrificed on day 28 after immunization with bMOG or adjuvants alone (sham). *Significantly different from sham (two-tailed Student’s t-test, P < 0.0001). Sample size: 4 sham, 6 bMOG. (J) Confocal image of a spinal cord section on day 28 post-immunization showing a cluster of CD138⁺ plasmablasts (red) counterstained with DAPI (blue, nuclei) outside the parenchyma (P), within the meninges (M; delineated by a dashed line). Scale bar: 10 μm.

Figure 2.

Transcriptomic and protein profiles of bMOG-induced plasmablasts. (A) tSNE plot of lymph node cells enriched in CD138⁺ cells from four mice on day 8 after bMOG immunization. Cells were analyzed using 10× Genomics 5’ Single-Cell Gene Expression technology. The indicated cell subsets were identified using the tSNE distribution, K-means, and cell type-specific markers. Encircled cells were those selected for V(D)J profiling in Figure 3. Descriptive data on each subset are provided in table. Abbreviations: pDCs, plasmacytoid dendritic cells; cDCs, conventional dendritic cells. (B) Cells expressing the plasmablast markers CD138 and BCMA (blue). (C) Log₂ expression of Birc5 revealing proliferating plasmablasts (encircled). (D) Heat map showing the expression profile of genes differentially regulated in plasmablasts compared to B cells, as identified by significant feature comparison in Loupe Cell Browser (P ≤ 0.04). Median-normalized mean UMI counts were Ln(x+1)-transformed and centered without row scaling using Clustvis. (E) tSNE plot generated from mass cytometry data on lymph node cells on day 8 after bMOG immunization. (F) Markers used to identify plasmablasts (CD138), B cells (CD19), and type-2 cDCs (cDC2s; CCR7) in E. (G) Comparison of expression levels of proteins involved in antigen presentation between plasmablasts, B cells, and cDC2s. *Significantly different from plasmablast group (ANOVA, P < 0.0001; post hoc Tukey’s test (P ≤ 0.0004).

Figure 3.

Profile of antibodies produced by bMOG-induced plasmablasts. (A) tSNE plots of gene expression data in Figure 2 showing plasmablasts selected for V(D)J analysis (colors). (B) Abundance distribution of the top-100 clonotypes, as analyzed using 10× Genomics Single-Cell V(D)J technology. (C–E) Comparison of immunoglobulin gene segment usage between mice. (F) Sequencing data from gene expression libraries showing the presence of the transmembrane domain (blue) in Ighm, Igha, and Ighe, but not Ighg transcripts. (G) Percentage of somatic mutations in the Ighv gene segment of the main clonotypes with ≥ 5 cells (n = 89–123 clonotypes per mouse). Two outliers with mutation rates of 16 and 18 % were excluded in mouse 4. (H) Comparison of the CDR3 region of the heavy and light chains of nine highly similar clonotypes from mouse 1, collectively referred to as clonotype 2 in J. These clonotypes were identical except for the CDR3 regions. Alignment was performed with webPRANK. (I) Number of clonotypes shared between mice. (J) Honeycomb plots showing clonotypes (dot clusters) with ≥ 5 cells (dots) of any isotypes. Numbers indicate clonotypes selected for further analysis.

Figure 4.

C1 can react with wild-type MOG, but not when the latter is in its native, plasma membrane-bound form, making it a non-pathogenic antibody. In all panels, anti-MOG IgG1 clone 8–18C5 was used as a positive control. Antibodies were IgG1, except in D where three C1 isotypes were tested. (A) ELISA against bMOG (left) or wild-type mouse and human MOG_1–125 (mMOG and hMOG; right). Quantity of bound IgG is expressed as either raw absorbance minus background (left) or absorbance normalized to 8–18C5 (right). (B) Western blot for bMOG, mouse MOG_1–125 or full-length MOG from mouse spinal cord extracts. (C) Spinal cord sections stained with anti-MOG antibodies (red) and DAPI (blue). Scale bar: 25 μm. CC = central canal. (D) Flow cytometry of live GL261 cells transfected to produce full-length MOG (red), incubated with anti-MOG antibodies, and stained with anti-mouse IgG antibody. Non-transfected cells (blue) were used as a negative control. (E) Severity of EAE in mice immunized with MOG_35–55 and injected intravenously 8 days later with PBS or 200 μg of the indicated antibody. Data from two independent experiments are expressed as either daily scores from the day of disease onset (left) or area under the curve (right). Left panel: *significantly different from PBS (two-way ANOVA with repeated measures using rank-transformed data, P < 0.0001; post hoc Mann-Whitney tests, P ≤ 0.049). Right panel: *significantly different from the other groups (one-way ANOVA, P < 0.0001; post hoc Tukey’s test, P ≤ 0.0012). Sample size: 9 PBS; 20 clonotype 1; 20 8-18C5. (F) Disease incidence of experiments in E. Log-rank test, P = 0.0003. Sample size: 10 PBS; 20 clonotype 1; 20 8-18C5.

Figure 5.

Pathogenic anti-MOG IgG antibodies are produced in B cell-dependent EAE models. (A) ELISA detection of anti-MOG IgG1, IgG2b, and IgG2c in serum from mice immunized with either bMOG (top) or hMOG (bottom) through time. Sample size per group: bMOG, 16–20 mice; hMOG, 10 mice. (B) Detection of anti-MOG antibodies by live cell-based assay in serum from mice immunized with either bMOG (left) or hMOG (right). GL261 cells transfected to produce full-length mMOG were incubated with serum collected before or after immunization (days 0 and 14), and stained with isotype-specific secondary antibodies. Data are expressed as delta mean fluorescence intensity (ΔMFI). *Significantly different from day 0 group from same isotype (Mann-Whitney test, P < 0.0001). Sample size per group: bMOG, 12 mice; hMOG, 9 mice. (C) Severity of EAE in MOG_35–55-immunized mice injected with 150 μl of serum from either naive mice or mice sacrificed 8 days after immunization with bMOG. Data are expressed as either daily scores from the day of disease onset (left) or area under the curve (right). Left panel: *significantly different from naive group as determined by two-way ANOVA with repeated measures using rank-transformed data (P = 0.002), followed by post hoc Mann-Whitney tests (P ≤ 0.007). Right panel: Mann-Whitney test, P = 0.006. Data are from two independent experiments. Sample size: 12 naive serum; 14 bMOG serum. (D) Survival and (E) disease incidence rates for the experiment in C. P-values shown were calculated using the log-rank test. Sample size: as in C.

Figure 6.

8–18C5 competes with mouse and human pathogenic anti-MOG antibodies for binding to plasma membrane-bound MOG. (A) Live cell-based assay in which mMOG-expressing cells were sequentially incubated with 8–18C5 at the indicated concentrations, mouse serum collected on day 14 post-immunization with bMOG, and secondary antibodies to mouse IgG2b and IgG2c. Data are expressed as delta mean fluorescence intensity (Δ MFI). *Significantly different from the other concentrations (Kruskal-Wallis test, P ≤ 0.0003 for both IgG2b and IgG2c; post hoc Dunn’s test, P ≤ 0.0028). Sample size: 12 sera. (B) Same analysis as in A, except that the sera were from hMOG-immunized mice. *Significantly different from the other concentrations (Kruskal-Wallis test, P ≤ 0.0072; post hoc Dunn’s test, P ≤ 0.0042). Sample size: 6 sera. (C) Same analysis as in A, except that hMOG-expressing cells were incubated with serum from MOGAD patients and stained with an anti-human IgG secondary antibody. *Significantly different from the other concentrations (Kruskal-Wallis test, P < 0.0001; post hoc Dunn’s test, P < 0.0001). Sample size: 25.

Figure 7.

bMOG-induced EAE can be attenuated with 8–18C5Mut. (A) Severity of EAE in mice intravenously injected with PBS or 200 μg of the indicated antibody on day 9 post-immunization. Data are expressed as either daily scores from the day of disease onset (left) or area under the curve (right). Left panel: *significantly different from the isotype group, as determined by two-way ANOVA with repeated measures (P = 0.049), followed by Fisher’s LSD tests (P ≤ 0.0451), using rank-transformed scores. Right panel: *significantly different from the other groups (one-way ANOVA, P = 0.0036; post hoc Tukey tests, P ≤ 0.0220). Data are from two independent experiments. Sample Size: 18 8-18C5, 16 8-18C5Mut, 19 isotype, 17 PBS. (B) Kaplan-Meier plot showing the percentage of mice from panel A that had completely recovered by the end of experiment (log-rank test: overall, P = 0.0331; 8–18C5Mut vs isotype, P = 0.0229). Shaded areas: 95 % pointwise confidence interval. (C) Black gold-stained spinal cord sections on day 43 post-immunization at low (top) or high (bottom) magnification. Evidence of demyelination (lost or granular staining) is observable in a mouse treated with isotype antibody, but not in a mouse treated with 8–18C5Mut or not immunized (naive). The clinical score at the time of sacrifice is indicated for each mouse. Scale bars: top, 200 μm; bottom, 50 μm.