Our data demonstrate that genetically modified L. lactis can be used for mucosal delivery of functional immunodominant antigens, and that this approach suppresses inflammatory antigen-specific T cell responses in gliadin-sensitized NOD AB° DQ8 transgenic mice. Furthermore, our results suggest that the induced suppression is antigen-specific and mediated by Foxp3+ CD4+CD25+ and CD25− regulatory T cells that possibly function through an IL-10 and TGF-β dependent mechanism.
Successful clinical application of antigen-specific mucosal tolerance for the treatment of human diseases has been difficult to achieve and critically depends on several factors, including the purity, source, dose and the mode of antigen presentation to the mucosal immune system (7
). Several protocols for induction of oral tolerance, including oral administration of the antigen with IL-10, have been shown to induce antigen-specific Tr1 cells that suppress undesired immune responses toward self-antigens, allergens, and food antigens. Although a previous clinical attempt to induce tolerance in celiac disease by the administration of rhIL-10 in refractory celiac disease patients was ineffective (26
), strategies to boost the number and/or function of Ag-specific Tr-1 cells may offer new therapeutic opportunities. This notion is supported by the finding that gliadin-specific mucosal regulatory T cells from celiac disease patients are able to suppress proliferation of pathogenic Th0 cells (27
Both dendritic cells (DC) and regulatory T cells are critically involved in tolerance induction (29
). We recently demonstrated that exposure to L. lactis
alters DC phenotype and function, which in the presence of simultaneous exposure to a DC-presented antigen might result in the generation of an antigen-specific regulatory T cell subset (Huibregtse et al submitted and (12
)). We hypothesize that induction of antigen-specific regulatory T cells in our experiments was mediated by altered presentation of the immunodominant peptide by dendritic cells, and our observation that LL-eDQ8d treatment interfered with IL-12 production of splenocytes, suggests that at least part of the tolerogenic effect is DC-mediated. However, it should be noted that the spleen is not the primary antigen-recognition site in our model, and alternatively, activation of regulatory T cells may have resulted in a reduced activation of antigen-presenting cells. In fact, our results with the lamina propria cells demonstrate that the GALT is a site of induction of regulatory cells in response to mucosal administration of L. Lactis
We further demonstrated that LL-eDQ8d treatment reduced peripheral DTH responses as well as eDQ8d-specific proliferation of lamina propria cells, bulk splenocytes and inguinal lymph node cells. LL-pT1NX treatment also somewhat reduced the DTH and splenic proliferative capacity but less pronounced than the LL-eDQ8d treated mice. The LL-eDQ8d treatment-mediated IL-10 secretion and the reduction of IL-12 and IFN-γ production that was found after ex vivo stimulation of splenocytes and inguinal lymph node cells respectively and was not observed in the LL-pT1NX treated mice. Moreover splenocytes of sensitized LL-eDQ8d treated mice were not sensitive to stimulation with an irrelevant antigen. These data confirm our previous findings that the tolerogenic effect, at least in part is L. lactis mediated but that the co-delivery of low-dose antigen, in this case DQ8d, greatly enhances the induction of antigen-specific oral tolerance.
In recent years it has become apparent that Tregs play a critical role in the induction and maintenance of oral tolerance (31
). Still many questions need to be answered concerning the phenotype and complexity of Tregs as well as the precise role and different overlaps in oral tolerance. Several phenotypically and functionally distinct Treg subsets have been implicated in suppression of intestinal inflammation and induction of oral tolerance, including adaptive Tregs (aTregs), comprising Th3 and Tr1 cells, and naturally occurring Tregs (nTregs), which maintain tolerance to self-antigen under normal physiological conditions. Although it is probable that nTregs play a central role in regulating gut immune homeostasis, their precise function remains to be characterized (5
). Furthermore, recently a separate category of Tregs has been described that acquires Foxp3
upon TGF-β stimulation. These so-called inducible Tregs (iTregs) have regulatory functions both in vitro
and in vivo
). This recently discovered subset mimics the Tregs induced in this model, as both subsets are induced in the periphery, express Foxp3
and are critically dependent on TGF-β/IL-10.
To map the Tregs that mediated oral tolerance in our experiments, we studied the functional importance of TGF-β, IL-10, and LAP (membrane-associated TGF-β) on the eDQ8d-specific splenic proliferative response using neutralizing antibodies. Interestingly only the combined neutralization of IL-10 and TGF-β interfered with the proliferative capacity of the splenocytes suggesting an interactive role for both anti-inflammatory cytokines. The exact mechanism by which these cytokines function is not completely understood but both cytokines frequently interact during regulatory Treg responses (37
Furthermore we found a significant Foxp3
upregulation in both the mucosal and the splenic CD4+
T-cell population. It is known that antigen-specific TGF-β producing Th3 cells drive the differentiation of antigen-specific Foxp3+
regulatory cells in the periphery (40
). Furthermore TGF-β-dependent conversion of peripheral CD4+
T cells into CD25+
suppressor cells has been reported (35
). Unfortunately specific markers for distinguishing between naturally occurring Tregs and inducible Tregs are lacking. Our data therefore suggest that either mucosal CD4+
cells induced by LL-eDQ8d treatment play a regulatory role or they can be eventually converted into CD4+
). It has been shown that oral tolerance induced by CTB-conjugated Ag is associated with increased TGF-β production and the generation of Foxp3+
and both Foxp3+
). The relationship between thymus derived natural CD4+
Tregs and other subpopulations induced in the periphery, such as peripherally generated CD4+
cells, as well as Tr1 and Th3 cells needs clarification, but these data suggest a significant overlap and interactive function in the induction of mucosal tolerance.
Direct in vivo
induction of tolerogenic DC or Treg is a major target for immunotherapy for allergic, autoimmune and several inflammatory diseases and can be achieved by exposing the mucosal immune system to low doses of antigen (7
We here report that oral supplementation of a genetically modified L. lactis secreting DQ8d peptides greatly reduces systemic immune responses induced by that antigen in DQ8d-immunized NOD AB° DQ8 transgenic mice. The suppression is mediated by the induction of Foxp3+ Tregs that are dependent on both TGF-β and IL-10. These observations support the development of gut delivered bacteria enhanced oral tolerance for the treatment of both mucosal and systemic autoimmune, inflammatory or allergic diseases by specific antigen-secreting L. lactis. With such studies, this type of therapy may be considered a viable option for treating gluten-sensitive disorders such as celiac disease.