Spontaneous relapsing-remitting EAE in the SJL/J mouse: MOG-reactive transgenic T cells recruit endogenous MOG-specific B cells

Abstract

We describe new T cell receptor (TCR) transgenic mice (relapsing-remitting [RR] mice) carrying a TCR specific for myelin oligodendrocyte glycoprotein (MOG) peptide 92-106 in the context of I-A(s). Backcrossed to the SJL/J background, most RR mice spontaneously develop RR experimental autoimmune encephalomyelitis (EAE) with episodes often altering between different central nervous system tissues like the cerebellum, optic nerve, and spinal cord. Development of spontaneous EAE depends on the presence of an intact B cell compartment and on the expression of MOG autoantigen. There is no spontaneous EAE development in B cell-depleted mice or in transgenic mice lacking MOG. Transgenic T cells seem to expand MOG autoreactive B cells from the endogenous repertoire. The expanded autoreactive B cells produce autoantibodies binding to a conformational epitope on the native MOG protein while ignoring the T cell target peptide. The secreted autoantibodies are pathogenic, enhancing demyelinating EAE episodes. RR mice constitute the first spontaneous animal model for the most common form of multiple sclerosis (MS), RR MS.

Figure 1.

Characterization of a new MOG-specific TCR transgenic SJL/J mouse. (A) Thymocytes from 8–10-wk-old TCR transgenic mice or NTLs were stained with antibodies to CD4, CD8, TCR-Vβ4, and TCR-Vα8.3 and cells were analyzed by flow cytometry. Representative figures of three to seven analyzed mice are shown. Transgenic Vα8.3 and Vβ4 are shown on cells gated as indicated, either CD4⁺/CD8⁻ or CD4⁻/ CD8⁺. (B) Proliferative response to recombinant MOG protein (rMOG) or MOG peptide 92–106 measured as ³H-thymidine incorporation. Splenocytes from TCR transgenic mice (8–10-wk-old healthy) were cultured with indicated concentrations of rMOG (top) and MOG_92-106 (bottom). (C) Cytokine responses to rMOG. Indicated cytokines were measured in supernatants harvested 48 h after rMOG activation by ELISA. Pooled data from three independent experiments are shown. Error bars indicate SEM. B and C: TCR¹⁶⁴⁰, n = 3; TCR¹⁶³⁹, n = 3; TCR¹⁵⁸⁶, n = 3; NTL, n = 2.

Figure 2.

Spontaneous RR-EAE in TCR transgenic SJL/J mice. (A) Male and female single-transgenic TCR¹⁶⁴⁰ compared with double-transgenic TCR¹⁶⁴⁰ × IgH^MOG mice. Shown is the spontaneous incidence of first signs of ataxia or classical EAE-like symptoms in TCR¹⁶⁴⁰ (left) and double-transgenic TCR¹⁶⁴⁰ × IgH^MOG (right) mice. TCR¹⁶⁴⁰ females (f), n = 12; males (m), n = 27; TCR¹⁶⁴⁰ × IgH^MOG females, n = 20; males, n = 26; TCR¹⁶⁴⁰ × Mog^−/− mice, n = 6. Disease kinetic of genders in TCR¹⁶⁴⁰ mice was not statistically significant (P = 0.304) but differed significantly between sexes of TCR¹⁶⁴⁰ × IgH^MOG mice (P = 0.0004). (B) Histological analysis of cerebellum and spinal cord from a sick TCR¹⁶⁴⁰ mouse (RR mouse) with ataxia and classical paralysis. Cerebellum (I–IV) and spinal cord (V–VIII) showed severe infiltration, demyelination, and axonal damage as visualized by immunohistochemistry using anti-CD3 (I and V) and anti–Mac3 antibodies (II and VI), Luxol fast blue staining (III and VII), and Bielschowsky silver impregnation (IV and VIII). I, II, V, and VI were counterstained by hematoxylin and eosin (H&E). Magnification: ×17 (cerebellum) and ×30 (spinal cord). Bars, 1 mm. Data are representative of at least two independent experiments consisting of more than three mice per group.

Figure 3.

Inflammatory cell infiltrates in RR-EAE lesions. (A–C) Cellular infiltrate into the CNS of sick TCR¹⁶⁴⁰ mice (score 3) is composed of macrophages, T cells, and B cells. CNS mononuclear cells were isolated from a sick TCR¹⁶⁴⁰ mouse and stained against CD11b and CD45.1 (A), together with CD8 and CD4 (B) or together with CD19 and B220 (C). Cells in B and C were analyzed among gated CD45.1⁺CD11b⁻ cells (red region as indicated). (D–F) Activation and Th1/Th17 cytokine expression of infiltrating CD4⁺ T cells. (D) CNS infiltrate cells express CD25, but not CD62L, and partially down-modulate their TCR (Vα8.3 and Vβ4). Activation status and TCR expression was compared between splenocytes (left) and CNS-isolated cells (right). (E) CD4⁺CD3⁺ T cells infiltrating the CNS of sick TCR¹⁶⁴⁰ mice predominantly expressed the pathogenic TCR composed of Vα8.3 and Vβ4 chains in a low expression level (TCR^low) compared with the spleen. Numbers indicate the percentage of stained cells in the respective quadrant or region. (F) Intracellular cytokine staining after stimulation with PMA/ionomycin in brefeldin A. Th1 (IFN-γ⁺/IL-17⁻) and Th17 (IFN-γ⁻/IL-17⁺) cells are enriched in CNS infiltrates of sick TCR¹⁶⁴⁰ (score 3.5) mice compared with splenocytes. D–F were analyzed among gated CD45.1⁺CD4⁺ cells. Flow cytometry data are representative of three to five sick TCR¹⁶⁴⁰ mice analyzed in three to five independent experiments.

Figure 4.

Ig deposition and B cell infiltrates in spinal cord of sick TCR¹⁶⁴⁰ × IgH^MOG and TCR¹⁶⁴⁰ mice. (A–D) Histological analysis of spinal cord using anti-Ig antibodies shows deposition of Ig in sick TCR¹⁶⁴⁰ × IgH^MOG (A and B) and TCR¹⁶⁴⁰ (C and D) mice (B and D show magnifications of marked areas in A and C, respectively). (E) Anti-B220 staining revealed some B cells among cellular infiltrates in sick TCR¹⁶⁴⁰ mice. (F) Deposition of complement C9neo within CNS lesions of TCR¹⁶⁴⁰ mice. Mononuclear cells were stained with H&E. Magnifications: A and C, ×50; B, ×425; D and E, ×200; F, ×450. Bars: (A and C) 1 mm; (B and D–F) 100 µm. Data are representative of at least two independent experiments consisting of more than three mice per group.

Figure 5.

MOG-specific antibodies in double-transgenic TCR¹⁶⁴⁰ × IgH^MOG and single TCR transgenic mice. (A) MOG binding allotype a autoantibodies were detected by ELISA in sera of indicated groups (each five to six mice). (B) Spontaneous anti-MOG IgG1 allotype b antibodies in TCR¹⁶⁴⁰ and TCR¹⁶⁴⁰ × IgH^MOG but not in NTL (n = 5 for each group). Sera at the indicated dilutions were incubated with plates precoated with rMOG. Bound anti-MOG Ig was detected by allotype-specific antibodies as indicated. Mean absorbance at OD 405 nm is shown. Error bars indicate SEM. Data are representative of two to three independent experiments.

Figure 6.

MOG-specific antibodies in single TCR transgenic mice. (A) Endogenous MOG binding antibodies of allotype b were measured by ELISA in sera of NTLs, healthy and sick TCR¹⁶⁴⁰, and rMOG-immunized WT SJL/J (n = 6 for each group). (B) Anti-MOG Igs of IgG1^b isotype are formed in TCR¹⁶⁴⁰ and sick TCR¹⁵⁸⁶ but not in TCR¹⁶³⁹ or TCR¹⁵⁸⁶ mice. Control NTLs, as well as TCR¹⁶⁴⁰ mice deficient for MOG (Mog^−/−), are devoid of MOG binding autoantibodies (each three to five mice per group). (C) Autoantibodies in sera from TCR¹⁶⁴⁰ mice are specific for MOG. Anti-MOG Ig antibodies, but not anti-MBP or anti–DNA-specific antibodies, were detected by ELISA in sera from TCR¹⁶⁴⁰ mice (1/50 diluted) using IgG1^b-specific antibodies (n = 5). (D) Appearance of anti-MOG IgG1^b and IgG2a/b^b in TCR¹⁶⁴⁰ mice by 5 wk of age. MOG binding antibodies were measured in 1/100 diluted sera from mice of different ages, each with three to five mice per group. Sera at the indicated dilutions were incubated with plates precoated with rMOG. Bound anti-MOG Ig was detected by allotype-specific antibodies as indicated. Mean absorbance at OD 405 nm is shown. Error bars indicate SEM. Data are representative of two independent experiments.

Figure 7.

MOG-specific autoantibodies bind MOG-expressing cells and activate complement. (A) Binding of anti-MOG Ig from indicated mice to correctly conformed MOG on the cell surface of transduced EL4 cells (EL4-MOG) shown by flow cytometry. EL4 cells (top) and MOG-expressing EL4-MOG cells (bottom) were incubated with 1/200 diluted sera obtained from the indicated mice or 0.5 µg/ml 8.18-C5 mAb. Bound antibodies were detected by FACS using biotinylated anti–IgG1-specific antibody (allotype unspecific) and streptavidin-PE (SA-PE). A representative plot of three independent experiments is shown. (B) Complement activating capability of MOG binding antibodies in sera from transgenic mice. EL4 and EL4-MOG cells were incubated with sera (1/20 and 1/200 diluted) obtained from the indicated mice, each with three to five per group. Complement activating capability was measured by cell lysis using propidium iodide (PI) staining after incubation of sera-bound EL4 cells with rabbit complement. Background values were subtracted, and shown are mean values with the SEM of one experiment representative of three similar experiments.

Figure 8.

B cells and anti-MOG antibodies are essential for spontaneous EAE. (A) Spontaneously developed MOG-specific autoantibodies from TCR¹⁶⁴⁰ mice have pathogenic potential. EAE was induced in WT SJL/J mice by low-dose PLP 139–151 immunization. Serum from TCR¹⁶⁴⁰ mice, NTL mice, or 8.18c5 mAb were transferred after mice showed first clinical symptoms. Serum from TCR¹⁶⁴⁰ mice significantly increased disease severity in recipient mice compared with serum from NTL mice. Error bars indicate SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Data were pooled from two independent experiments, each six to seven per group. (B) B cell depletion protects TCR¹⁶⁴⁰ mice from spontaneous EAE. B cells were depleted from TCR¹⁶⁴⁰ mice by twice weekly injections of anti-CD20 antibodies from day 3 after birth, and control mice received mouse IgG2a antibodies. The development of spontaneous EAE was monitored regularly. Although 85% of isotype control antibody-treated mice developed spontaneous EAE, treatment with CD20 antibody protected TCR¹⁶⁴⁰ mice from disease development. Shown are the disease course (left) and the spontaneous EAE incidence (right) of TCR¹⁶⁴⁰ mice treated with these antibodies. Data were pooled from two independent experiments, each with six to seven per group.