The gut-associated lymphoid tissue (GALT) is constantly exposed to a variety of Ags and must therefore decipher a large number of distinct signals at all times. Responding correctly to each set of signals is crucial. When the GALT receives signals from the intestinal flora or food Ags, it must induce a state of nonresponsiveness (mucosal tolerance). In contrast, when pathogenic bacteria invade the intestinal mucosa, it is necessary to elicit strong T and B cell responses. The GALT is therefore in the position of constantly fighting intolerance to food and the commensal flora while effectively battling infectious microbes. Determining precisely which type of response to generate in each case is key to the prevention of immune dysregulation and tissue damage.
A number of factors regulate immune responses to intestinal Ags. First, it is hypothesized that the various Ags encountered by gut-associated lymphoid tissue (GALT) APC are processed in different ways. However, the mechanisms that lead to differential processing and presentation remain unclear. Second, certain populations of T and B cells that are unique to the GALT play a role in directing the responses to intestinal Ags. Although it is clear that the gut microenvironment affects the way that APCs, T cells, and B cells respond to Ags, the means by which it exerts its influence on the development of tolerance and immunity is still largely unknown. We will discuss the current understanding of the induction of nonresponsiveness to dietary and commensal Ags in this review, emphasizing the most recent findings.
Various routes of access
During the last few years, a number of new routes for Ag entry into the GALT have been identified. Peyer’s patches (PP; lymphoid follicles predominantly found in the ileum of the small intestine) are a primary immune inductive site of the GALT. One well-established route of Ag entry into the GALT involves Ag binding to specialized epithelial cells (M cells) in the follicle-associated epithelium and its subsequent transport to APC residing in the subepithelial dome of the PP (1). More recent work has shown that gut bacteria can also be taken up by Ulex europaeus agglutinin-1+ M cells in the non-follicle-associated epithelium of intestinal villi (2). Dendritic cells (DC) can extend their dendrites through epithelial cell tight junctions and sample both pathogenic and nonpathogenic bacteria directly (3). There are also Ab-mediated mechanisms that help Ags cross from the gut lumen into the GALT. In addition to its role in immune exclusion, secretory IgA (S-IgA) can bind to, and be transported by, M cells (4). Upon arrival in the subepithelial dome of the PP, S-IgA binds to DC, CD4+ T cells, and B cells (4). Whereas all three cell types bind S-IgA, only DC internalize it, suggesting a role for this IgA transport pathway in the promotion of luminal Ag sampling (4). Related work has shown that, in addition to transporting IgG across the intestinal epithelial barrier into the lumen, the human neonatal FcR can recycle bound luminal Ag back into the GALT for processing by DC and presentation to T cells in the mesenteric lymph node (MLN) (5).
Where are luminal Ags presented?
A number of studies have shown that orally administered Ags are presented in the GALT. However, there is good evidence that oral Ag can be presented in the periphery as well. The dose of Ag and whether it is administered in a tolerogenic (no adjuvant) or immunogenic (with adjuvant) form are important factors that affect Ag presentation. Low doses of orally administered peptide are predominantly presented by DC of the PP and MLN, whereas higher doses are presented by DC, macrophages, and B cells in local and systemic sites (6). Similarly, intragastric administration of whole protein Ags leads to peptide presentation by DC and B cells in both the gut and periphery (7). Even though DC and B cells can present both tolerogenic and immunogenic protein Ags, only B cell up-regulate costimulatory molecules in response to immunogenic Ags (7). A study using fluorescent microparticles suggested that particulate Ags can be presented in the PP as well (8). The data on Ag presentation correlates nicely with studies examining T cell activation and proliferation after the feeding of soluble protein Ags. Many groups, including ours, have shown that Ag-specific T cells in the MLN and PP up-regulate activation markers and proliferate after oral administration of soluble protein Ags (6, 9, 10, 11, 12, 13, 14). Large doses of orally administered Ag can induce the rapid activation of T cells in the spleen (SPL) and peripheral lymph nodes (LN), suggesting that food Ags can gain access to the bloodstream (9, 13, 15, 16). Interestingly, PP and MLN may not be necessary for the induction of oral tolerance or S-IgA; this was recently reviewed by Mowat (17) and will not be discussed further here.
The mucosal immune system responds to the intestinal flora
Although it is often stated that we are tolerant of the estimated 1014 bacteria that comprise our commensal gut flora, that label belies the dynamic interaction that characterizes this coexistence. Indeed, the intestinal flora has a profound impact on the development of both the mucosal and systemic lymphoid tissue (18). In germfree mice, the GALT is poorly developed; the PP have low numbers of IgA+ plasma cells and the lamina propria of intestinal villi (LP) lacks CD4+ T cells (18). Even peripheral lymphoid organs including the SPL and LN exhibit structural defects and serum IgG is reduced (18). Although systemic immune responses to the intestinal flora are not detectable in healthy individuals, it is clear that local mucosal responses are generated (18). A recent study demonstrating that intestinal DC contain commensal bacteria (19) is supported by previous work by Rescigno et al. (3) which showed that DC can extend their dendrites between intestinal epithelial cells to sample intestinal flora from the gut lumen. Although pathogens gain access to the systemic immune system, commensal bacteria are largely killed by mucosal macrophages. The few that remain are sequestered within DC that are restricted to the MLN and do not migrate to the SPL (19). However, in the absence of the MLN, the flora can gain access to the SPL (19). Interestingly, most flora are excluded from the PP as well, pointing to the MLN as the center for flora-DC interactions. Such territorial restrictions appear to be necessary to prevent undesired immune responses to the flora (19).
Among the many roles attributed to DC in immunity, one important to the gut is the role DC play in directing intestinal IgA responses to the flora. In fact, only DC loaded with commensal bacteria can induce IgA+ B cells, which help keep commensals in the lumen (19). Most of these B cells are B1 B cells, which explains why anticommensal IgA can be generated in μMT mice that lack conventional B cells (20, 21). In addition, this IgA response is present in TCRβ−/−δ−/− double-knockout mice and nude mice, albeit at 2- to 5-fold lower levels, suggesting that its production is partially T cell independent (19, 20).
A number of studies suggest a role for TGF-β and TGF-β receptors in isotype switching and the production of mucosal IgA by B cells (22, 23, 24). TGF-β has also been reported to enhance the ability of B cells to present Ag (25). TGF-β can be produced by GALT DC (26, 27), or by T cells that then promote the secretion of IgA by B cells (28). Orally tolerized T cells cannot provide help for B cell Ab (IgM, IgG1, and IgG2a) production during peripheral (29) or mucosal (30) challenge with Ag plus adjuvant. Surprisingly, inducing oral tolerance before mucosal challenge also decreases the production of serum and mucosal IgA (31). This idea is controversial, as it is reasonable to hypothesize that TGF-β-secreting T cells induced by orally administered Ag should promote IgA production by B cells upon Ag challenge. CD11b+ PP DC also play a special role in enhancing B cell production of IgA via secretion of IL-6 (32).
TLR involvement in responses to intestinal Ags
The role played by the TLRs of the innate immune system in sensing pathogens is well documented. The presence of microbial products on nonpathogenic, commensal bacteria that can also be detected by TLRs illustrates one of the fascinating conundrums implicit in the induction of nonresponsiveness to the luminal contents. Indeed, signaling through TLRs by the flora under steady-state conditions is necessary for the maintenance of gut homeostasis, strongly supporting the idea that the flora and the GALT constantly communicate with each other (33). In the absence of TLR signaling by the flora, injury to the gut is increased (33). This gut injury is rescued by feeding bacterial products like LPS (33). Recent data from our laboratory also support this idea by showing that allergic responses to food Ags are much greater in the absence of either TLR4 signaling or components of the intestinal flora (34). This suggests that TLR4 signaling by the intestinal microflora is necessary for proper immune homeostasis both in the gut and systemically (34). Our findings were foreshadowed by earlier studies, predating the identification of TLRs, which demonstrated that nonresponsiveness to SRBC could be induced in C3H/HeN but not in LPS-nonresponsive C3H/HeJ mice (35, 36). The authors insightfully proposed that the failure to induce nonresponsiveness was related to the absence of a flora-induced immunoregulatory cell population in C3H/HeJ mice. Interestingly, germfree mice were not tolerant to orally administered SRBC, but tolerance could be induced by feeding the mice LPS, suggesting a role for TLR4 signaling by the intestinal flora (37). If the absence of TLR4 signaling (either in mice lacking functional TLR4 or mice lacking the commensal flora) leads to stronger allergic responses to food Ags, one might expect aberrant responses to other Ags as well. The Bir substrain of C3H/HeJ develops a colitis characterized by the induction of Abs to bacterial flagellin, the ligand for TLR5 (38). This suggests the intriguing possibility that the inability to transmit flora-derived signals via TLR4 results in a “regulatory deficit” that can predispose to responsiveness to both dietary and flora-derived Ags. Other support for this idea comes from the observation that, in vitro, the suppressive activity of TLR4+CD4+CD45RBlowCD25+ T regulatory cells (Tregs) is enhanced by exposure to LPS (39).
Signaling through TLRs regulates DC activity
Current evidence suggests that, like Ags administered via other routes, orally administered Ags are presented by bone marrow-derived APC (10). This APC is likely to be a DC; the in vivo expansion of DC using Flt3 ligand enhances both the induction of oral tolerance and mucosal immunity (40, 41). Innate immune system signaling influences DC migration, cytokine production, and costimulatory molecule expression. When DC are activated through TLRs, they secrete inflammatory cytokines that promote productive T cell immunity. Recent work indicates that the DC migration and maturation induced by these cytokines are not sufficient for T cell activation. In particular, the secretion of IL-6 by DC protects effector T cells from suppression by CD4+CD25+ Tregs (42). When inflammatory cytokines like IL-6 are not produced by DC (such as in MyD88−/− mice), unchecked Treg-mediated suppression prevents T cell activation (43). However, other data suggest that the cytokines produced by TLR-activated DC reverse Treg anergy (inability to proliferate in response to TCR stimulation) without affecting their suppressive activity (44). Costimulatory molecule (CD40, CD80, and CD86) expression is also up-regulated on DC after TLR signaling. Unlike the expression of inflammatory cytokines, up-regulation of costimulatory molecules can occur in the absence of MyD88 (42, 43, 45) and even in the absence of TLR stimuli (44).
Role of CD40 expressed on DC
Studies performed using DC lacking CD40 strongly support a role for CD40 in the generation of productive immunity. First, DC lacking cell surface expression of CD40, due to inhibited RelB function, prevent immune priming, and can suppress ongoing immune responses (46). DC generated from CD40-deficient mice display similar capabilities (46). In addition, DC from CD40−/− mice do not make IL-12 or elicit CD4+ and CD8+ T cell responses, even though they are able to present peptide Ag (47). Instead, DC lacking cell surface CD40 have been shown to induce IL-10-secreting Tregs (46). Finally, CD40/CD40L interactions release immature DC from suppression by CD4+CD25+ T cells, further suggesting that CD40 ligation is necessary and sufficient to abrogate tolerance and inhibit the action of Tregs (48). Administration of an agonistic Ab to CD40 can prevent induction of nonresponsiveness to both peripherally and orally administered Ags, suggesting that signaling through CD40 promotes immunity to oral Ags as well (49). In agreement with this, irradiated mice reconstituted with bone marrow from CD40-deficient mice can be orally tolerized to OVA, suggesting that CD40/CD40L interactions are dispensable for oral tolerance induction (50). Prior work had suggested that oral tolerance could not be induced in CD40L-deficient mice, implying that CD40L expression (presumably on CD4+ T cells) is necessary for tolerance induction (51).
Role of CD80 and CD86 interactions with CD28 and CTLA-4
It is known that low levels of costimulatory molecule (CD80 and CD86) expression on APC leads to T cell anergy (52). This occurs because CTLA-4, which inhibits T cell responses, has a higher affinity for CD80 and CD86 than CD28, which promotes T cell responses (52). Recent work has demonstrated that the localization of CTLA-4 to the immune synapse is severely reduced when T cells are cultured with CD80−/− B cells in vitro (53). In contrast, the localization of CD28 to the immune synapse is not altered (53). In addition, when CD86−/− B cells are used as APC, the amount of CD28, but not CTLA-4, that localizes to the synapse is decreased (53). These data support a model where CD80 inhibits T cell function by binding to CTLA-4 and CD86 promotes T cell responses by binding to CD28 (54). However, studies of CD80/86 interactions with CD28/CTLA-4 in oral tolerance have produced different results. A study of low-dose oral tolerance showed that the administration of an anti-CD86, but not an anti-CD80, blocking Ab abrogated nonresponsiveness (55). These data suggest that CD86 must interact with CD28 during low-dose oral tolerance and that CD80/CTLA-4 interactions play no role. However, this study also showed that treatment with CTLA-4-Ig abrogated high-dose oral tolerance, suggesting that CD80/86 must engage CTLA-4 for tolerance induction to occur (55, 56). These results suggest that high- and low-dose oral tolerance occur through distinct mechanisms, which is in agreement with other studies (57).
CTLA-4 has a well-documented role in the negative regulation of T cells (58), and mice deficient in CTLA-4 have severe autoimmune disease (59, 60). Using a model of i.v. tolerance, Sharpe and colleagues (61) showed that CTLA-4−/− T cells are resistant to tolerance induction. Tolerized T cells are normally blocked at the late G1 to S restriction point, but the absence of CTLA-4 allows the T cells to enter S phase (61). A model of i.p. tolerance to a peptide Ag agreed with the idea that CTLA-4 is necessary for anergy induction, because blocking CD80/86 interactions with CTLA-4 Ig or neutralizing anti-CTLA-4 both prevent the induction of T cell tolerance (62). Although it is clear CTLA-4 is required for T cell anergy, the role of the G1S phase block has recently been questioned. Treatment of T cells with rapamycin (which blocks cells at G1) during T cell activation induced anergy, but treatment of T cells with sanglifehrin A (which also induces G1 arrest) did not induce anergy, suggesting that a block at G1 is not sufficient to induce T cell anergy (63).
Studies of the involvement of CTLA-4 in oral tolerance are largely in agreement with studies on its role in peripheral tolerance. Treatment with a neutralizing Ab to CTLA-4 abrogates oral tolerance, apparently by stimulating T cell division and cytokine production in the MLN after OVA feeding (56, 64, 65). A number of studies have clearly shown that CD4+CD25+ Treg cells constitutively express CTLA-4 (66, 67). Interestingly, T cells from DO11.10 × CTLA-4−/− mice transferred into irradiated recipients prevented the induction of oral tolerance, whereas the transfer of DO11.10 × IL-10−/− T cells did not (64). This study suggested that Ag-specific T cells need to express CTLA-4, but not IL-10, for tolerance induction to occur (64).
Regulatory DC and the modulation of T cell immune responses
Many different populations of GALT DC have been described. The maturation state of particular DC subpopulations appears to dictate the subsequent immune response. Therefore, recent work has focused on the identification of DC with tolerogenic properties. One study described a CD11clowCD45RBhigh population of DC in the SPL and LN that is increased in IL-10 transgenic mice (68). Interestingly, these DC display a plasmacytoid morphology and a stable immature phenotype (68). These immature plasmacytoid DC (pDC) secrete IL-10 and induce tolerance by promoting the differentiation of IL-10-secreting T regulatory 1-type (Tr1) Tregs (68).
A second study described airway pDC (CD11c+Gr-1+B220+CD45RB+) that suppressed the generation of effector T cells primed by myeloid DC (69). This pDC population also generated suppressive Tregs in vitro, suggesting that tolerogenic subpopulations of DC may specialize in the generation of Tregs (69). In this model, depletion of pDC in vivo with an anti-Gr-1 Ab led to airway inflammation and allergic responses to inhaled Ags (69). An earlier study by Umetsu’s group (27) described regulatory DC that played an important role in maintaining tolerance to airway Ags. This study also showed that DC from the MLN of OVA-fed mice secrete TGF-β and can direct CD4+ T cells to produce IL-4, IL-10, and TGF-β (27). This is in agreement with earlier studies by Iwasaki and Kelsall (26) showing that freshly isolated PP DC produce TGF-β and that PP DC activated via CD40 ligation in vitro produce IL-10. Recent studies have shown that the expression of TGF-β, either by a retrovirus (70) or a transgene (71), can direct T cells to make IL-10, strongly suggesting that these two tolerance-promoting cytokines are tightly linked. Several studies have now shown that T cells activated in the presence of various subsets of GALT DC (such as CD11b+ myeloid DC and CD11clowCD8α+ pDC) are directed to make IL-4 and IL-10 and acquire regulatory properties (26, 72, 73). These observations are supported by work demonstrating the requirement for IL-4 and IL-10 for the induction and maintenance of oral tolerance (74) and are consistent with the ability of transferred CD4+ T cells from OVA-fed mice to induce nonresponsiveness in naive recipients (75, 76). Recent work from Matzinger’s laboratory (77) has confirmed and extended these ideas using an elegant allophenic embryo aggregation technique to create chimeric recipient mice bearing APCs with two different MHC types. At adulthood, the chimeric mice were reconstituted with monoclonal populations of transgenic T cells of two different Ag specificities to examine the response to orally administered Ags. Splenic CD4+ T cells responded to orally administered Ag by secreting IL-4 and IL-10. DCs “educated” by the cytokines secreted by these T cells were then able to act as a “temporal bridge” to induce the same cytokine response in naive T cells (77).
The role of Tregs in the induction and maintenance of mucosal tolerance
During the past decade, the role of Treg cells in a variety of tolerance models has been explored. Many different markers, including latency-associated peptide (LAP) (78), glucocorticoid-induced TNFR (GITR) (79), CTLA-4 (66, 67), CD25 (80, 81, 82), CD45RBlow (83), and Foxp3 (84, 85, 86), have been used to describe Tregs. However, most investigators now recognize two major Treg subsets: a natural population that suppresses through cell-to-cell contact, and an induced population that secretes suppressive cytokines. The two markers that most clearly define natural Tregs are CD25 (80, 81, 82) and the forkhead transcription factor Foxp3 (84, 85, 86). Importantly, natural Tregs are anergic and can suppress T cell responses in vitro in an Ag-nonspecific manner (80, 82, 87, 88). The absence of natural Tregs (either CD25+ or Foxp3+) in both mice and humans results in severe autoimmune disease, emphasizing their vital role in the prevention of self-reactivity (84, 85, 86). Substantial evidence indicates that Foxp3+CD4+CD25+ Tregs represent a distinct thymic-derived lineage; retroviral expression of Foxp3 in naive T cells converts them into Foxp3+ Tregs that can suppress T cell responses (84, 85, 86). However, in addition to controlling autoimmunity, CD4+CD25+ Tregs have also been shown to play a role in regulating responses to pathogenic microbes (89, 90). Whether CD4+CD25+ Tregs act primarily to prevent pathology by limiting responses to infectious agents or to control autoreactivity is a matter of current debate (91). The development of Foxp3+ Tregs from CD4+CD25− T cells in the periphery is also controversial. von Boehmer’s group (92) used a subcutaneously implanted osmotic pump to deliver low doses of peptide to thymectomized TCR-transgenic mice on a RAG2−/− background. Strikingly, this peptide delivery system led to the development of peripheral CD4+CD25+ Tregs that were Foxp3+, CTLA-4+, and CD45RBlow and displayed suppressive function both in vitro and in vivo, showing that CD4+CD25+ Tregs can develop from CD4+CD25− T cells in the absence of a thymus (92).
The role of Tregs in regulating immune responses to pathogens has led many to ask whether they also maintain tolerance to the intestinal flora. One population of gut T cells, called mucosal-associated invariant T cells, is located in the LP of conventional but not germfree animals, suggesting that they arise in response to the flora (93). As mentioned above, LPS signaling through TLR4 on Treg cells can increase the suppressive ability of Tregs in the absence of APC (39). Therefore, it seems likely that the flora can communicate directly with T cells. However, some gut T cell populations that respond to the flora are apparently not induced by the flora. For example, CD4+CD25− T cells that respond to enterobacterial Ags in vitro are present in both conventional and germfree mice (94). These cells are normally prevented from reacting to the flora by a population of CD4+CD25+ Tregs, suggesting that T cell responses to the intestinal flora are tightly controlled (94). Earlier studies that described CD4+ Tr1-like T cells in the LP (95) and intestinal T cells that secrete IL-10 and TGF-β show that induced Tregs also reside in the gut (96).
Two major populations of induced Tregs have been described. Some are IL-10-secreting Tr1 cells (97), whereas others are TGF-β-secreting Th3 cells (98). Interestingly, Tr1 Treg clones can prevent colitis when transferred into SCID mice, showing that induced Tregs can play a role in maintaining intestinal homeostasis (97). A role for IL-10-secreting cells in the control of colitis was supported by a study showing that CD4+CD45RBlow cells from IL-10−/− mice could not transfer protection (99). In addition, Tr1 cells also arise at the airway mucosa in response to Bordetella pertussis infection in mice, consistent with the idea that Tregs respond to pathogenic bacteria (100). Studies of oral tolerance have suggested that IL-10-secreting CD4+ PP cells arise during and are responsible for active suppression in tolerance to low doses of orally administered Ags (101, 102). These IL-10-secreting T cells inhibit T cell proliferation in vitro (which can be reversed by Abs to IL-10), and T cell-mediated inflammation in vivo (101). Orally administered Ag also induces TGF-β-secreting Th3 cells with Ag-nonspecific suppressive activity that can affect many different cell types (98). More recent studies of TGF-β in Treg function have emphasized a role for surface-bound TGF-β, particularly on subsets of CD4+CD25+ T cells (78, 103). This raises the possibility that the regulatory activity originally ascribed to Th3 cells may be contained within a subset of CD4+CD25+ Tregs; additional work is needed to clarify this point.
It is currently unclear whether IL-10-secreting Tregs can become CD4+CD25+ Tregs (and vice versa) or whether these two Treg populations arise from the same lineage (104). However, most of the current data suggest that the two populations are distinct and that one cannot arise from the other. For example, even though CD4+CD25+ Tregs can produce IL-10 (89, 99), it is not required for their effector function (105). IL-10-secreting Tregs have been reported that do not express Foxp3 (104) and can arise in the absence of CD4+CD25+ Tregs (106). However, recent work has shown that TGF-β can induce regulatory function in T cells by inducing Foxp3 expression (71, 107, 108), suggesting that induced Tregs can promote the generation of Tregs with a natural phenotype.
Studies examining the response to intragastrically administered Ag have shown that CD4+CD25+ cells arise during the induction of oral tolerance in mice adoptively transferred with Ag-specific T cells (12). Other work showed that feeding OVA to DO11.10 TCR transgenic mice increased the percentages of CD4+CD25+ T cells in the SPL and LN and that the transfer of CD4+CD25+, and not CD4+CD25−, T cells from mice fed OVA could transfer suppression to naive mice that were challenged with OVA/CFA (109). Surprisingly, CD4+CD25+, but not CD4+CD25−, T cells from unfed mice suppressed these responses (109). This finding was probably due to the transfer of natural Tregs from the unfed mice. A different study suggested that both CD4+CD25+ and CD4+CD25− T cells from the PP and MLN, but not the SPL, of adoptively transferred mice fed OVA transferred suppression to naive mice (14). Depletion of CD25+ cells from normal mice prevents the induction of oral tolerance (110). The mechanisms by which CD4+CD25+ Tregs function in oral tolerance need further investigation.
Conclusions and future questions
Despite much recent progress, many unanswered questions remain. It is still unclear whether tolerance to the gut flora is induced and maintained the same way as tolerance to soluble food proteins. However, the available data suggest that they are not the same. First, immune exclusion by S-IgA helps maintain tolerance to the enteric flora, whereas S-IgA Ab responses are not generated to most dietary Ags in the absence of adjuvants (52). This suggests that different Ag uptake pathways and/or different Ag presentation strategies generate responses to each type of Ag. Second, although soluble protein Ags can gain access to the peripheral LN and SPL (probably via the bloodstream), the intestinal flora cannot travel beyond the GALT if the MLN is intact (19). This, along with the production of S-IgA, suggests that tolerance to the flora is more “active” than tolerance to soluble proteins. However, it is currently unclear whether the presentation of soluble protein Ags in the SPL and peripheral LN plays a role in inducing Ag-specific nonresponsiveness.
Another example of how tolerance to the flora is active highlights the role of the innate immune system in mucosal tolerance. Unlike soluble proteins, the commensal bacteria can signal through TLRs. The current data create a picture where TLR signaling is necessary for nonresponsiveness to the gut flora but not dietary protein Ags (Fig. 1). However, the in vivo situation is obviously more complicated because signaling by the flora can affect responses to dietary proteins. As shown in our laboratory, when peanut proteins are orally administered in the absence of TLR signaling, there is a breakdown of tolerance that leads to severe allergic reactions. As mentioned above, it seems that this breakdown of tolerance in TLR4−/− mice is due, at least partially, to the absence of certain Treg cell populations.
What is the role of Tregs in oral tolerance? Many studies have implicated a role for Tregs in the induction and maintenance of mucosal tolerance to both soluble proteins and the flora. However, a variety of Treg populations (such as TGF-β-secreting Th3 cells and CD4+CD25+ cells) with different mechanisms of action have been shown to participate in tolerance. Therefore, future studies are necessary to clearly define the mechanisms by which tolerance is induced and how different conditions give rise to distinct Treg populations.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Abbreviations used in this paper: GALT, gut-associated lymphoid tissue; PP, Peyer’s patch; DC, dendritic cell; S-IgA, secretory IgA; LN, lymph node; MLN, mesenteric LN; SPL, spleen; LP, lamina propria of intestinal villi; Treg, T regulatory cell; pDC, plasmacytoid DC.