|Home | About | Journals | Submit | Contact Us | Français|
Clinical studies indicate that anti-CD20 B cell depletion may be an effective multiple sclerosis therapy. We investigated mechanisms of its immune modulation using two paradigms of experimental autoimmune encephalomyelitis (EAE).
Murine EAE was induced by either recombinant myelin oligodendrocyte glycoprotein (rMOG), a model in which B cells are considered to contribute pathogenically, or MOG peptide (p)35–55, a model that does not require B cells.
In EAE induced by rMOG, B cells became activated and, when serving as antigen presenting cells (APC), promoted differentiation of proinflammatory MOG-specific Th1 and Th17 cells. B cell depletion prevented or reversed established rMOG-induced EAE, which was associated with less CNS inflammation, elimination of meningeal B cells, and reduction of MOG-specific Th1 and Th17 cells. In contrast, in EAE induced by MOG p35–55, B cells did not become activated or efficiently polarize proinflammatory MOG-specific T cells, similar to naïve B cells. In this EAE setting, anti-CD20 treatment exacerbated EAE, and did not impede development of Th1 or Th17 cells. Irrespective of the EAE model used, B cell depletion reduced the frequency of regulatory T cells, and increased the capacity of remaining APC to promote development of encephalitogenic T cells.
Our study highlights distinct roles for B cells in pathogenesis and regulation of CNS autoimmune disease. Clinical benefit from depletion of antigen-activated B cells may relate primarily to abrogation of proinflammatory B cell APC function. However, in certain clinical settings, elimination of unactivated B cells, which participate in regulation of T cells and other APC, may be undesirable.
The central nervous system (CNS) has traditionally been viewed as an immune-privileged compartment with limited and well-controlled access for immune cells. B cells and plasma cells, however, are commonly found in active MS lesions1 and the presence of oligoclonal antibodies within the cerebrospinal fluid remains a hallmark finding in the diagnosis of MS. Myelin-specific antibodies have been identified in areas of vesicular demyelination2, suggesting that they directly promote CNS damage. The observation that plasma exchange was beneficial in MS patients with histologic evidence of CNS antibody deposition3 provided further support for a pathogenic role of antibodies. Besides serving as the source for antibody-secreting plasma cells, B cells express major histocompatibility complex (MHC) class II molecules constitutively and may participate as antigen presenting cells (APC). B cells are capable of processing native antigen and are very efficient APC when they recognize the same antigen as the responding T cells4, 5. As processing of native myelin antigen by CNS resident or infiltrating APC is required for initiation of CNS autoimmune inflammation and clinical disease6, 7, myelin-specific B cells may have an important role for the activation of encephalitogenic T cells in the pathogenesis of CNS autoimmune disease.
With greater appreciation that B cells may have dual humoral and cellular roles in MS pathogenesis, interest in use of selective B cell-depleting agents for therapy has intensified8, 9. Promising results were obtained in clinical trials testing a monoclonal antibody targeting CD20 (RituxanR), a cell surface protein that is expressed on immature and mature B cells, but not on differentiated plasma cells. Treatment with Rituxan was beneficial in patients with relapsing-remitting MS8 and in a subgroup of primary progressive MS patients with evidence of active CNS inflammation9. Anti-CD20 mediated B cell depletion was also clinically beneficial in a small open-label study in neuromyelitis optia (NMO)10, a CNS demyelinating disease associated with aquaporin-4-specific antibodies. The purpose of our investigation was to elucidate the immunologic consequences of anti-CD20 therapy in two related models of experimental autoimmune encephalomyelitis (EAE)11. In one model, EAE was induced by immunization with recombinant myelin oligodendrocyte glycoprotein (rMOG), which generates a population of antigen-activated B cells and promotes development of antibodies against MOG protein. B cell depletion prevented rMOG-induced EAE and reversed paralysis when treatment was initiated after EAE onset. In established EAE, anti-CD20 depleted B cells within the CNS. B cell depletion decreased the frequency of peripheral and CNS encephalitogenic Th1 and Th17 cells and was associated with reduced serum titers of myelin-specific antibodies. These findings highlight the pathogenic role of activated B cells in CNS autoimmune disease, and provide mechanisms of action in support of B cell depletion for treatment of MS.
In the second model, EAE was induced by immunization with MOG peptide (p) 35–55, which binds MHC II directly on lymphoid APC without processing6, and leads to peripheral activation of encephalitogenic T cells6,12. Using this protocol, considered “B cell-independent”, MOG protein-specific B cells were not activated. In contrast to the benefit observed in EAE elicited by MOG protein, B cell depletion exacerbated clinical and histologic EAE in this model and development of Th1 and Th17 cells was not dampened. In both rMOG and peptide-induced EAE, CD20-mediated B cell depletion reduced the frequency of FoxP3+ regulatory T cells (Treg) and augmented pro-inflammatory function of remaining myeloid APC. These observations indicate that in the absence of proinflammatory B cell function, depletion of unactivated (naive) B cells may not be advantageous. The results of this study highlight key differences between MOG protein and MOG peptide EAE models, and underscore the importance of B-T cross-talk in pathogenesis and regulation of CNS autoimmunity.
C57BL/6 female mice, 5–8 weeks of age, as well as μMT mice were purchased from the Jackson Laboratories (Bar Harbor, MN). hCD20 transgenic (Tg) C57BL/6 mice13,14 were used for anti-CD20-mediated B cell depletion. In these mice, hCD20 recapitulates the expression of endogenous murine CD2015 and treatment with the murine anti-human CD20 antibody (clone 2h7) results in rapid depletion of B cells (13, supplementary Figure 1a). Untreated or isotype control-treated hCD20 Tg mice developed EAE indistinguishable from wild-type mice (supplementary Figure 1b). C57BL/6 MOG 35–55-specific T cell receptor (TCR) Tg mice16 were kindly provided by V.K. Kuchroo (Harvard). JHT mice17 were obtained from K. Rajewsky (Harvard).
Mouse myelin oligodendrocyte glycoprotein (MOG) peptide 35–55 (MEVGWYRSPFSRVVHLYRNGK) was synthesized by Auspep (Parkville, Australia). Recombinant (r) mouse MOG (1–117) was synthesized, purified and refolded as previously described18. Ovalbumin (OVA) p323–339 (ISQAVHAAHAEINEAGR) was synthesized by Abgent, Inc. (San Diego, USA). Intact OVA was purchased from Sigma-Aldrich (Saint-Louis, MI).
Eight to twelve week-old female C57BL/6 or hCD20 Tg C57BL/6 mice were injected subcutaneously with 25 μg MOG p35–55 or 100 μg rMOG 1–117 in complete Freund’s adjuvant (CFA) (DIFCO Laboratories, Detroit, Michigan, USA). After immunization and 48 hrs later mice received an intravenous injection of 200 ng pertussis toxin. Individual animals were observed daily and clinical scores were assessed as follows: 0 = no clinical disease, 1 = loss of tail tone only, 2 = mild monoparesis or paraparesis, 3 = severe paraparesis, 4 = paraplegia and/or quadraparesis, and 5 = moribund or death.
In order to ensure maximal B cell depletion when examining anti-CD20 in prevention of EAE, anti-CD20, mice received weekly intraperitoneal (i.p.) injections of 200 μg of a-hCD20 monoclonal antibody (m2h7) or anti-ragweed IgG2a-isotype control monoclonal antibody starting 21 days prior to immunization. For evaluation of anti-hCD20 treatment in established EAE, mice were randomized to weekly treatment once they developed EAE disease score ≥2.
Serum was obtained from mice treated with a-hCD20 or isotype control (IgG2a) prior to treatment onset and weekly thereafter. 96-Maxisorb plates (Costar) were coated with MOG p35–55 or rMOG 1–117 (10 μg/ml in PBS) and then blocked with BSA (Sigma-Aldrich). Plate-bound antibodies were detected with horseradish–peroxidase-conjugated anti-mouse IgG (cross-reactive with all Ig isotypes; 1:6,000; Sigma-Aldrich). The antibody titers were quantified at the serum dilution indicated using commercially available anti-MOG 8.18C-5 (Millipore) as the standard. Plates were read at 450-nm wavelength. SOFTmax ELISA plate reader and software (Molecular Devices Corporation, Sunnyvale, CA) were used for data analysis.
For B cell – T cell co-culture assays, B cells were MACS (Miltenyi Biotec)-separated from lymph nodes or spleens. Following separation, B cells were evaluated for purity (>99%) by FACS staining for B220. For T cell co-culture assays using remaining splenocytes as APC, hCD20 Tg mice received weekly injections of 0.2 mg of a-hCD20 or IgG2a-isotype starting 21 days prior to immunization with rMOG protein or MOG p35–55 peptide. 12 days after immunization, spleens were isolated and B220+ B cells and CD3+ T cells were removed by MACS separation. 5 × 105 B cells or remaining splenocytes were co-cultured with 1 × 104 naïve T cells isolated from MOG TCR Tg mice in the presence of rMOG 1–117 or MOG p35–55. After 72 hrs, T cell differentiation was evaluated by FACS or ELISA.
B cells were examined by FACS analysis after staining with antibodies specific for B220, CD95 (FAS), GL7 or CD21. B220, CD95 and FAS antibodies were purchased from Pharmingen, and anti-CD21 was purchased from eBioscience. Proinflammatory T cell differentiation was evaluated by surface staining for CD3 (Pharmingen) and ICS for IFN-γ and IL-17 (eBioscience). Activation of monocytic cells was evaluated by surface staining for CD11b (Pharmingen) and ICS for TNF or IL-10 (eBiosience). Induction of Treg was evaluated by FACS staining for CD4 (GK1.5), CD25 and FoxP3 (eBiosience).
Culture supernatants were collected for cytokine analysis at 48-hr (IFN-γ) or 72-hr (IL-17), respectively and analyzed by ELISA (Pharmingen, San Diego, CA). The results for ELISA assays are expressed as an average of triplicate wells +/− SEM. SOFTmax ELISA plate reader and software (Molecular Devices Corporation, Sunnyvale, CA) were used for data analysis.
Brains and spinal cords of mice were fixed in 10% neutral-buffered formalin, paraffin-embedded and sectioned. Representative sections were stained with Luxol fast blue (LFB), hematoxylin and eosin (H&E) or evaluated for B cell- or T cell infiltration by B220- or CD3-immunohistochemistry, respectively. H&E stained sections (inflammation) and LFB stained sections (demyelination) were scored on a scale of 0–4. B220 and CD3 stained sections were evaluated by morphometric image analysis. Final results were reported as B220 or CD3 stained cells per square mm of spinal cord area.
Data are presented as mean ± SEM. For clinical scores significance between groups was examined using the Mann-Whitney U test. A value of p < 0.01 was considered significant. All other statistical analysis was performed using a one-way multiple-range analysis of variance test (ANOVA) for multiple comparisons. A value of p < 0.01 was considered significant.
Two different EAE models were examined in this report. In EAE induced by immunization with MOG protein (recombinant (r) MOG 1–117)), internalization and processing by APC is required for presentation of its encephalitogenic determinant to pathogenic CD4+ T cells6. In this model, B cells become activated through recognition of MOG protein via B cell receptor (BCR) engagement. As shown in Fig 1a, when used as APC for presentation of MOG protein, B cells isolated from MOG protein-immunized mice efficiently stimulated MHC II-restricted CD4+ T cells that recognize the encephalitogenic MOG peptide (p) 35–55. Following activation, B cells developed into plasma cells that secreted antibodies directed against MOG (19, Fig 1b). Therefore, immunization by this protocol activates both cellular and humoral components of B cell immunity.
Unlike antigen presentation of rMOG, MHC II-restricted T cell recognition of the MOG p35–55, does not require internalization and processing by APC6. Instead, naive B cells, independent of their BCR specificity, are capable of presenting short peptides through direct binding to cell surface MHC II molecules. As shown in Fig 1a, B cells from mice immunized with MOG p35–55, like naive B cells, were capable of presenting MOG p35–55, but not MOG protein, to MOG-specific T cells. Further, immunization with MOG p35–55 did not efficiently lead to expansion of MOG-specific B cells, and was not associated with a significant antibody response (20, Fig 1c).
Anti-CD20 treatment was investigated in human (h) CD20 transgenic (Tg) C57BL/6 mice13,14. These mice develop EAE in a manner that is indistinguishable from wild-type C57BL/6 mice (Supplementary Fig 1). Data indicate that kinetics of B cell depletion in different tissue microenvironments may depend upon vascular access of anti-CD20 antibodies13. Depletion of mature (B220+CD21+) B cells was examined in blood, bone marrow, lymph nodes spleen and in the peritoneal cavity at various time points following a single anti-CD20 treatment of unimmunized hCD20 Tg mice. A hierarchy in tissue susceptibility to CD20-mediated B cell depletion was evident13; reduction of B cells was detected in blood and bone marrow at three hours, and in lymph nodes and spleen at two days (Fig 2). B cell depletion in the peritoneum was slower; at two days peritoneal B cells were reduced by approximately 30%, and at seven days by 95%. There was more than 99% of depletion B220+CD21+ B cells in all immune and non-immune tissues examined 14 days post injection. In order to ensure maximal B cell depletion when anti-CD20 treatment was evaluated for EAE prevention, this antibody was administered weekly starting three weeks in advance of immunization.
Given that B cells responded differently to MOG protein and MOG peptide, we postulated that anti-CD20 treatment might lead to divergent clinical and immunologic outcomes. Anti-CD20-mediated B cell depletion reduced clinical severity of MOG protein-induced EAE when treatment began prior to disease induction (Fig 3a and Table 1). Similarly, treatment of established EAE reversed paralysis. In these mice, anti-CD20 treatment depleted 60% of B cells within established CNS lesions, which was reflected by a 70% reduction of B cells within meningeal lesions (Fig 3b and 3c).
The potential influence of CD20 B cell depletion in MOG protein-induced EAE on proinflammatory T cell and humoral responses was examined. In general, in untreated mice with EAE, the frequency of IL-17-producing cells was much lower in the periphery than within the CNS, consistent with observations by other investigators21. In anti-CD20 treatment, the frequencies of Th1 and Th17 cells were reduced in the periphery and, to a greater extent, within the CNS (Fig 4a). The absolute numbers of CNS CD3+ T cells were not significantly altered in treatment of established EAE when paralysis was reversed (Supplementary Figure 2), suggesting that anti-CD20 treatment did not initially reduce CNS influx of T cells. Amelioration of established MOG protein-induced EAE by anti-CD20 was associated with a reduction of serum antibody titers directed against rMOG (Fig 4b). In addition to serving as a source for antibody-secreting plasma cells and as APC, B cells may participate in homeostasis of regulatory T cells22. In this regard, despite the clinical benefit of anti-CD20 treatment, B cell depletion was associated with reduced frequency of Treg in peripheral lymphoid organs as well as within the CNS (Fig 4a).
Anti-CD20 treatment was investigated peptide-induced EAE, a model that does not require B cells for development of EAE. In contrast to anti-CD20 treatment of EAE induced by MOG protein, B cell depletion initiated either prior to immunization with MOG p35–55, or after onset of paralysis, exacerbated EAE (Fig 5a and Table 1). Clinical worsening was associated with more severe CNS inflammation and demyelination (Fig 5b and 5c), despite the fact that anti-CD20 treatment sufficiently depleted B cells within the CNS (Fig 5d). B cell depletion did not dampen pathogenic Th1 and Th17 responses in this disease model (Fig 5e). In fact, clinical worsening in anti-CD20 treatment of MOG p35–55-induced EAE was generally associated with an increase in CNS Th1, Th17 cells and Th1/Th17 double positive T cells, which may represent a more pathogenic T cell phenotype 23, 24. Similar to our findings in rMOG-induced EAE, anti-CD20 treatment reduced the frequency of CD4+CD25+FoxP3+ Treg in secondary lymphoid organs as well as within the CNS.
The immunologic mechanisms contributing to the paradoxical clinical outcomes of anti-CD20 depletion in EAE induced by rMOG or MOG p35–55 were examined further. We hypothesized that activated B cells in EAE induced by MOG protein might promote development of proinflammatory T cells, which were eliminated by anti-CD20 treatment. B cells were examined for cell surface expression of FAS, a protein that is up-regulated on lymphocytes following antigen receptor engagement, and GL-7, a marker of antigen-primed germinal center B cells25. Immunization with MOG protein, but not with p35–55, generated a population of activated B cells that expressed FAS and GL-7 (Fig 6a). Similarly, B cells from mice immunized with the non-encephalitogenic control protein ovalbumin (OVA), but not its short peptide OVA p323–339, upgregulated FAS and GL7, indicating that B cell activation was a characteristic associated with immunization with protein. Most importantly, only B cells from MOG protein-immunized mice, but not from unimmunized mice or mice immunized with MOG p35–55, efficiently polarized Th1 and Th17 cells when presenting MOG protein (Fig 6b). Activated B cells from mice immunized with MOG protein were also more efficient in promoting Th1 and Th17 differentiation of naive MOG p35–55-specific T cells when stimulated with MOG p35–55. Collectively, these results indicate that activated myelin antigen-specific B cells, which are generated in MOG protein-induced EAE, can contribute to encephalitogenic T cell priming in vivo. Loading of encephalitogenic peptide onto MHC II molecules expressed on unactivated (naive) B cell APC alone does not efficiently promote differentiation of encephalitogenic T cells.
Data indicate that B cells may communicate with other APC. For example, it was observed that B cells can capture antigen from lymph node subcapsular macrophages via their BCR, and deliver it to follicular dendritic cells, establishing a role for B cells in antigen transport 26. Through secretion of anti-inflammatory cytokines, B cells may also locally regulate other APC27. Thus, we evaluated how B cell depletion influenced the function of remaining APC. In both EAE models used, CD11b+ cells isolated from mice receiving anti-CD20 treatment produced more pro-inflammatory TNF and less anti-inflammatory IL-10 (Fig 7a). We then investigated whether this cytokine shift could translate into altered APC function. For this purpose, we isolated spleen cells from CD20 B cell depleted or isotype (control)-treated mice, and cultured them with naïve MOG p35–55 specific T cells. When compared to APC from isotype-treated mice (after in vitro removal of B cells), APC remaining after in vivo depletion of B cells exhibited an increased capacity to promote development of encephalitogenic Th1 and Th17 cells. Again, this proinflammatory gain of function by remaining APC after B cell depletion occurred in both EAE models (Fig 7b). In summary, in addition to their role in T cell activation, these results suggest that B cells can regulate other APC, and that nonselective depletion of B cells could augment the proinflammatory function of remaining APC.
Recent studies suggest that CD20-mediated B cell depletion may be effective in reducing CNS inflammation in MS8, 9. In this report, we investigated the immunological consequences of anti-CD20 B cell depletion in EAE induced by MOG protein and MOG p35–55. In MOG-protein-induced EAE, but not in EAE induced by MOG p35–55, activated MOG-reactive B cells participated as APC and promoted differentiation of naïve MOG-specific T cells into proinflammatory Th1 and Th17 cells in vitro. Anti-CD20-mediated B cell depletion ameliorated EAE induced by MOG-protein and suppressed development of Th1 and Th17 cells in vivo. Anti-CD20 treatment initiated after MOG-specific antibodies were generated, led to subsequent reduction in titers. Investigations in rheumatoid arthritis28, 29 and systemic lupus erythematosus30 indicated that administration of anti-CD20 similarly dampened humoral responses although plasma cells, which do not express CD20, were not eliminated13. While reduction of myelin-specific antibodies may potentiate the therapeutic effect of B cell depletion in a subgroup of MS patients with CNS antibody deposition3,31, it should be recognized that the benefit of anti-CD20 B cell depletion observed in a six-month placebo-controlled trial in relapsing-remitting MS was not associated with a reduction in antibodies8. Further, in EAE antibodies elicited by immunization with mouse MOG protein, although self-reactive, are not considered pathogenic19. Thus, the clinical benefit of anti-CD20 treatment observed in this EAE model more likely reflects a reduction in pro-inflammatory cellular function of MOG-specific B cells.
In both EAE models, anti-CD20 treatment depleted B cells within the CNS of mice with established EAE. In MOG protein-induced EAE, B cells became activated, and a greater number of B cells infiltrated the CNS (Fig 1d). The capability to deplete B cells within the CNS is of particular therapeutic relevance in light of the discovery of ectopic B cell follicles32 within the meninges in some individuals that developed secondary progressive MS, and that formation of these lymphoid follicle-like structures may be associated with elevated risk for irreversible disability33. The observation that B cells were efficiently depleted within the meninges, suggests that anti-CD20 could be also an attractive candidate for treatment of a subset of patients with secondary progressive MS.
Exacerbation of MOG peptide-induced EAE by CD20 treatment highlights the complexity of B cell function in CNS autoimmunity. Immunization with MOG p35–55 did not promote B cell activation. In contrast to anti-CD20 depletion in MOG protein-induced EAE, which was associated with clinical benefit and reduction in proinflammatory Th1 and Th17 cells within the CNS, CD20-mediated depletion resulted in clinical worsening of MOG p35–55-induced EAE and increased numbers of CNS infiltrating Th1 and Th17 cells. Besides serving as the source for antibody-secreting plasma cells and as APC for T cell activation, some B cell subsets may have an important role in immune regulation of CNS autoimmune disease22, 27, 34. Evidence suggests that antigen-naive B cells exert anti-inflammatory properties27, 35, which may inhibit maturation and pro-inflammatory differentiation of other APC in vivo36. In this regard, it has been observed that dendritic cells isolated from B cell-deficient mice, produce higher levels of IL-12 and promote pro-inflammatory T cell differentiation37. In conjunction with our observation that after anti-CD20 B cell depletion remaining myeloid APC secreted more pro-inflammatory TNF and less anti-inflammatory IL-10, these findings collectively indicate B cells can regulate other APC and suggest that this B cell characteristic may be abrogated by nonselective anti-CD20-mediated B cell depletion.
Naïve B cells may play an important role in development and maintenance of Treg in vivo22, 38. Deficiencies in the Treg compartment have been identified in several autoimmune conditions, including MS39, 40 and one of the goals in MS therapy is to correct this imbalance41, 42. While some studies suggest that anti-CD20 depletion may be associated with a modest increase of Treg 14, 43, we observed a reduction in numbers of CD25+Foxp3+ Treg in anti-CD20 treatment of EAE induced by either rMOG or MOG peptide. This finding is further supported by our investigations using B cell-deficient μMT44 and JHT17 mice. Similar to unimmunized anti-CD20 B cell-depleted mice, we demonstrated that B cell-deficient μMT or JHT mice contained lower frequencies of CD25+FoxP3+ Treg (Fig 8), again indicating that B cells participate in Treg homeostasis. There were no obvious qualitative differences in Treg in wild-type and anti-CD20 B cell-depleted mice. In this regard, we did not detect intracellular IL-10 protein production in CD4+CD25+ForP3+ Treg in either isotype-treated, or B cell-depleted mice.
Anti-CD20 therapy has been examined in other EAE settings45, 46. B cell depletion prevented exacerbations in a murine model of spontaneous relapsing-remitting EAE in which Tg T cells and B cells both recognize MOG46. A recent publication by Matsushita and colleagues45 also demonstrated exacerbation of MOG peptide-induced EAE when B cell-depleting treatment began prior to disease induction. Elegantly, the authors attributed worsening of disease to the absence of an IL-10-producing (B10) regulatory B cell subset. When anti-CD20 treatment started 14 days after immunization, severity of MOG p35–55-induced EAE was ameliorated, leading the authors to conclude that although protective at the time of disease induction, at a later stage, B cells or B cell subsets may promote disease progression. The apparent divergence in outcome of B cell depletion in reversal of MOG peptide-induced disease in our study could reflect differences in experimental procedures, such as dose of MOG p35–55 used for EAE induction, or the nature of the anti-CD20 antibody used47. One striking difference, however, is that they detected a peptide-specific antibody response upon immunization with their MOG p35–55 preparation, which could have reflected the four-fold higher dose of p35–55 used for EAE induction in their study. While those antibodies did not likely contribute in a pathogenic manner, their appearance may be indicative of B cell activation and maturation following immunization with MOG peptide, which was not observed in this report. Also, in our investigation, B cell depletion in hCD20 Tg mice was achieved using a mouse anti-hCD20 monoclonal antibody. More recently, we tested a mouse anti-mouse (m) CD20 monoclonal antibody for prevention of EAE induced by MOG protein or MOG peptide in non-Tg mice. Consistent with our findings using mouse anti-hCD20, anti-mCD20 treatment suppressed development of proinflammatory T cells and clinical EAE induced by MOG protein, while it promoted development of proinflammatory T cells and exacerbated clinical EAE induced by MOG p35–55. Most importantly, our demonstration that B cells regulate secretion of pro-inflammatory cytokines by monocytes is in agreement with the observation by Matsushita, et al. that certain B cell subsets have regulatory function, while others support pathogenesis of CNS autoimmune disease. Unlike the results of Matsushita and colleagues, our data indicate that the immunological and clinical outcome of B cell depletion is determined by the activation status and antigen-specificity of B cells, rather than the time of treatment initiation.
While the paradoxical clinical outcomes of CD20-mediated B cell depletion in EAE induced by MOG p35–55 and MOG protein correlated with increased and decreased frequencies of proinflammatory T cells, respectively, it should be recognized that reduction in Treg and augmentation of proinflammatory cytokine expression by remaining APC were common features of CD20 B cell depletion in both models. B cells may undertake additional cellular immune functions, which could have been eliminated by anti-CD20 treatment. It was observed that B cells are capable of capturing protein via their antigen-specific BCR and delivering it to lymph node follicular dendritic cells, more professional APC26. Through this mechanism of antigen transport, B cells can contribute indirectly to proinflammatory T cell polarization. We have demonstrated that activated MOG-specific B cells, but not naive B cells, serve directly as APC and polarize proinflammatory T cells. Therefore, we favor the possibility that there is a balance, and that the benefit from eliminating MOG protein-activated B cells reflects inhibition of their proinflammatory cellular function, while exacerbation of p35–55-induced EAE relates to depletion of unactivated (naive) B cells that participate in regulation. As was previously observed for myeloid APC, which can be divided into proinflammatory “type I” or anti-inflammatory “type II” classes48, 49, B cells may exhibit proinflammatory “Be1” or anti-inflammatory “Be2” T cell-polarizing phenotypes50. In the absence of antigen-activated Be1 cells, CD20 B cell depletion may exacerbate autoimmune disease in some settings51. Recently, we created Tg mice that contain B cells that express membrane MOG-specific BCR, but cannot secrete antibodies (Molnarfi, N. et al., unpublished). These BCR Tg mice will permit us to distinguish between certain cellular functions of Ag-specific B cells and the role of antibodies in pathogenesis of MOG-induced EAE.
In this report, we studied two distinct EAE models. One cannot conclude that EAE induced by either MOG protein or MOG peptide more closely reflects MS. Each model has its virtues and may emphasize different aspects of pathogenesis52. APC must process MOG protein through the endocytic pathway for MHC class II-restricted presentation of its encephalitogenic determinant to CD4+ T cells, while MOG p35–55 can be loaded onto MHC II molecules directly6. We have demonstrated that activated MOG-primed B cells are capable of efficiently presenting MOG protein and promoting differentiation of pathogenic MOG-specific T cells. Immunization with MOG protein elicits a stronger antibody response than does priming to MOG peptide. Our results highlight key differences in cellular and humoral B cell responses to MOG protein and MOG peptide, which could be important when choosing an EAE model for preclinical testing of other novel B cell-targeting agents for MS.
In summary, this study supports the use of anti-CD20-mediated depletion of activated B cells in treatment of CNS autoimmune disease and establishes inhibition of B cell-dependent activation of pathogenic Ag-specific T cells as one immunological mechanism that may contribute to its clinical benefit in MS. In addition, the observations in this report may be relevant to B cell depletion therapy in NMO, which is associated with pathogenic AQP4-specific IgG1, a T cell-dependent antibody subclass53, 54. Our study cautions that non-selective elimination of B cells may prevent unactivated or regulatory B cells from exerting their beneficial anti-inflammatory influence on other immune cells. Selective depletion of antigen-activated B cells may be a valuable strategy to further improve efficacy of B cell targeted therapies in MS and other inflammatory CNS demyelinating diseases.
Supplementary Figure 1. C57BL/6 hCD20 transgenic mice develop EAE indistinguishable from wild-type mice and allow rapid depletion of B cells. a) hCD20 mice received 200 μg of anti-CD20 or isotype and spleens were analyzed for the presence of B220+ B cells after 72 hours. b) C57BL/6 hCD20 transgenic and C57BL/6 wild-type mice were immunized with MOG p35–55 and either received 200μg isotype control or no treatment.
Supplementary Figure 2. Anti-CD20 treatment does not alter the number of CNS infiltrating T cells. hCD20 transgenic mice were immunized with rMOG protein (left panel) or MOG p35–55 (right panel). Weekly treatment with anti-hCD20 or isotype (IgG2a, control) started when EAE was fully established (EAE score ≥2). Spinal cord sections were evaluated for the presence of T cells by CD3-immunohistochemistry. Shown is the average number of CD3+ cells per square mm of spinal cord tissue.
We thank Dr. Patricia A. Nelson for helpful discussion. We also thank the Canio Refino of Genentech and the Genentech Histology and Immunohistochemistry Laboratories for processing and staining brain and spinal cord specimens. M.S.W. is supported by the National Multiple Sclerosis Society (NMSS, RG 4450A1/T). T.P. received fellowship support from NMSS and Teva Neuroscience. N.M. is supported by an advanced researcher fellowship from the Swiss National Science Foundation (#PA00A—119532). Support for this work was provided to S.S.Z. by the National Institute of Health (NIH) (RO1 AI073737, RO1 AI059709, and RO1 NS063008), the NMSS (RG 3622 and 3913), Dana Foundation, Guthy Jackson Charitable Foundation and Maisin Foundation. C.C.A.B. was supported by grants from the Baker Foundation, the National Health, the Medical Research Council of Australia and the NMSS (CCAB: RG3844A2/1).