The existence of regulatory B cells (Bregs) was proposed more than 30 years ago1, 2 and their immunosuppressive potential has now been proven in experimental animal models.3, 4 In humans, the data on Bregs are limited and almost exclusively based on observational studies, showing exacerbation of disease following B cell-depleting therapy in patients suffering from various autoimmune disorders.5–8
B cell-mediated suppression of intestinal inflammation has been demonstrated in different murine models of colitis. B cells play a protective role in the TCRα−/− model of colitis,9 which requires B cell expression of IL-10 and CD1d.10 B cell suppression of Gαi2−/− and CD4+CD45RBhigh T cell transfer colitis is independent of CD1d but requires B cell intrinsic IL-10 and regulatory CD8+ T cells.11, 12 IL-10 was also found to be a prerequisite for B cell suppression of dextran sulfate sodium (DSS) colitis but not for suppression of spontaneous colitis in nuclear factor of activated T cells c2 (NFATc2) deficient RAG−/− mice.13 Bregs have also demonstrated IL-10-dependent suppression of experimental autoimmune encephalomyelitis (EAE),14–17 and collagen-induced arthritis (CIA)18 and transfer of lipopolysaccharide (LPS)-activated19 or BCR-stimulated20 B cells was shown to mediate suppression of diabetes in NOD mice.
The ability to secrete IL-10 seems to be a general trait for the various described subsets of Bregs.21 Other or subsequent mechanisms like secretion of TGF-β,22 FasL-mediated induction of apoptosis,19 suppression of dendritic cell function,23 and regulatory T cells (Tregs) activation16 have been indicated as well. Today, a true marker for Bregs, as Foxp3 for Tregs, is still needed but expression of CD5 and high expression of CD1d have been associated with IL-10-producing B cells and are phenotypic characteristics of Breg subsets.21 Recently, IL-10-competent splenic CD1dhighCD5+CD19high B cells, named B10,24 have shown suppressive effects in EAE,25 contact hypersensitivity (CHS),24 murine lupus,26 and colitis.27 In addition, marginal zone (MZ) B cells and transitional-2 marginal zone precursor (T2-MZP) B cells both express CD1d, secrete IL-10 upon activation,28, 29 and suppress development of CHS30 and CIA.31
Of importance, stimulation through Toll-like receptors (TLRs) triggers IL-10 production in B cells32 and may, together with BCR ligation, very well be necessary for optimal Breg activation and function. This feature, in particular, could have an important impact on colitis development, as intestinal barrier leakage might allow inflow of different TLR ligands and recognition of enteroantigenic components.33, 34 We have previously shown that B cells pulsed with an enterobacterial extract (ebx-B cells) are extremely efficient antigen-presenting cells for enteroantigen-specific CD4+ T cells in vitro.35 We also demonstrated activation of Tregs in vitro by ebx-B cells but not by ebx-pulsed DCs.35 Since ebx represents the enteroantigenic repertoire and may contain multiple TLR ligands, we were encouraged to investigate the potential regulatory effect of ebx-B cells in vivo during the development of T cell transfer colitis. Identification of a linkage between enterobacteria and activation of immunosuppressive Bregs could beneficially be targeted in IBD treatment.
- Top of page
- MATERIALS AND METHODS
In the present study we demonstrate that B cells exposed to ebx secrete IL-10 and are capable of suppressing T cell transfer colitis and that suppression is associated with skewed T helper responses and increased Treg levels.
Our finding that B cells secrete IL-10 upon exposure to ebx is in line with previous demonstrations of IL-10 production in naïve B cells upon activation through TLRs,15, 32 as ebx probably comprises several TLR ligands. Combined stimulation through multiple TLRs synergistically enhances IL-10 production in B cells,32 which is in accordance with an increased IL-10-inducing ability of ebx compared to TLR4 ligation alone by LPS.
We observed elevated expression of CD1d and CD5 in the IL-10-producing splenic B cells exposed to ebx. Expression of these surface molecules has previously been associated with immunosuppressive B cells.10, 18, 24, 31, 43–45 In particular, the simultaneous expression of CD1d and CD5 indicates that ebx activates the recently identified splenic Breg subset named B10 cells.24 B10 cells are characterized as IL-10 competent CD1dhighCD5+ B cells and they have demonstrated their suppressive capacity in various CD4+ T cell-mediated disease settings as EAE,25 CHS,24 murine lupus,26 parasitic infection,46 and colitis.27 In fact, B10 cell frequencies increased in mice upon colitis development,27 in line with the increase in CD1dhighCD5+ B cells after prolonged exposure to ebx.
The significance of Bregs in human IBD has yet to be determined. Only few studies indirectly suggest the involvement of Bregs in IBD. B cell depleting therapy has caused exacerbation8 and development7 of ulcerative colitis (UC). In addition, levels of B-1a cells (CD5+) seem to be decreased in both UC and Crohn's disease (CD) patients.47–49 B cell-mediated suppression of experimental colitis has been demonstrated in different mouse models.9, 12, 13 The data presented here offer new aspects to the suppressive role of B cells in colitis. Like in the Gαi2−/− and TCRα−/− transfer models, we demonstrated suppression of T cell-mediated colitis. However, suppression of Gαi2−/− colitis was dependent on regulatory CD8+ T cells,12 which are absent in our model. In fact, B cell-mediated suppression of CD4+CD45RBhigh T cell transfer colitis, which is comparable to our model, also required CD8+ T cells.12 Potentially regulatory cell subsets are present in TCRα−/− mice but involvement of such in suppression of TCRα−/− colitis has not been addressed. Importantly, TCRα−/− colitis is mediated by a TH2 response,50 in contrast to our model, and suppression in this model required B cell-expression of IL-12.51
Similar to the Gαi2−/− and TCRα−/− models, B cell-derived IL-10 is a prerequisite for suppression in our model. Although IL-10 seems to be important for Breg function in general,21 we observed a tendency of suppression in IL-10 KO transplanted colitic mice. In a spontaneous innate model of colitis (RAG−/−NFATc2−/− double knockout mice) splenic B cells have demonstrated suppressive capability independent of IL-10 but dependent on a native BCR repertoire.13 Similar to the high frequency of CD4+ T cells (≈5%) constitutively recognizing ebx antigens,37 a high frequency of ebx recognizing B cells might exist. Therefore, we cannot exclude that T cell suppression by ebx-B cells is in part mediated through transcytosis-dependent enteroantigen uptake and presentation by B cells carrying BCRs specific for ebx. In accordance with this idea, we showed previously a superior enteroantigen-dependent T cell stimulatory capacity mediated by ebx-B cells in vitro.35 In line with this notion, Breg-mediated suppression of EAE14 and CHS24 was shown to require specific antigen, suggesting the presence of B cells recognizing EAE relevant autoantigens.
Suppression of colitis by ebx-B cells was associated with a decreased TH1 response, also observed during B cell-mediated suppression of EAE,52 CIA,18 and diabetes.19 However, we also observed an increased frequency of TH17 cells in B cell transplanted mice, which is in contrast to findings in EAE.15 B cells can impede TH1 cell development by suppressing IL-12 secretion by DCs23 and, since IFN-γ strongly counteracts TH17 cell differentiation,53, 54 suppression mediated by Breg may polarize in favor of TH17 cell development. However, the relevance of this finding needs further analysis.
We observed elevated levels of Foxp3-expressing Tregs in mice transplanted with colitis-suppressing ebx-B cells. This is in accordance with similar findings in other studies, demonstrating increased activation of regulatory T cells in B cell-mediated suppression of colitis,12 murine lupus,26 and EAE.16 Furthermore, splenic B cells can induce regulatory CD4+ and CD8+ T cell subsets by MHC-restricted presentation of ocular autoantigen55 and human CD40 stimulated B cells have the capability of inducing56 and expanding57 Tregs in vitro. As ≈1% of CD4+CD25− T cells express Foxp358—a frequency we also observed in colitic mice transplanted with T cells only—it is uncertain whether the increase in Foxp3+ Tregs observed after ebx-B cell transfer is due to de novo differentiation of Tregs or expansion of transferred Foxp3+ T cells. However, since ebx-B cell, but not ebx-pulsed DCs, can activate Tregs in vitro,35 our present data suggest that ebx-B-mediated Treg activation also occurs in vivo. Notably, B cells failed to suppress colitis in SCID mice transplanted with CD4+CD45RBhigh T cells,12 which are naïve cells and comprise no Foxp3+ Tregs.58 Because transfer of B cells caused a decrease in IL-10-producing T cells, it appears that the B cell-mediated suppression of colitis demonstrated here did not rely on T cell-derived IL-10. In support, Foxp3+ Tregs can efficiently suppress T cell responses in vitro42 and even transfer colitis59 independent of IL-10. However, further experiments are needed to conclude on the involvement of Tregs in suppression of colitis by ebx-B cells.
In conclusion, we demonstrate the presence of IL-10 secretion by B cells exposed to enterobacterial components. Such B cells acquire an immunosuppressive potential capable of suppressing the development of experimental T cell transfer colitis. B cell-induced suppression of colitis is associated with a decreased TH1 response and elevated levels of Foxp3+ Treg.