T cells and intestinal commensal bacteria-ignorance, rejection, and acceptance

Abstract

Trillions of commensal bacteria cohabit our bodies to mutual benefit. In the past several years, it has become clear that the adaptive immune system is not ignorant of intestinal commensal bacteria, but is constantly interacting with them. For T cells, the response to commensal bacteria does not appear uniform, as certain commensal bacterial species appear to trigger effector T cells to reject and control them, whereas other species elicit Foxp3⁺ regulatory T (Treg) cells to accept and be tolerant of them. Here, we review our current knowledge of T cell differentiation in response to commensal bacteria, and how this process leads to immune homeostasis in the intestine.

1. Introduction

The intestines are home to hundreds of commensal bacterial species that provide many benefits to the host, including important contributions to the metabolism of food [1]. However, as commensal bacteria are foreign to the host, they can trigger unwanted immune responses such as inflammatory bowel disease (IBD) [2]. Yet, the gastrointestinal tract is also an important portal for bacterial infections. Thus, the immune system must protect against pathogenic bacteria while avoiding inappropriate immune responses to commensal bacteria.

In broad strokes, this balancing act is analogous to that employed by the immune system to protect against infections while avoiding autoimmunity. However, the mechanisms that prevent immune responses to self-antigens may not be applicable to commensal bacterial antigens. An important mechanism of self-tolerance is the purging of auto-reactive cells while they are still immature and unable to cause immunopathology For T cells, this occurs in the thymus, an organ that is far removed from the intestine. Gut antigens are therefore unlikely to be present in the thymus to induce tolerance, implying that different mechanisms are utilized to prevent unwanted T cell immune responses to commensal bacteria. Here, we will review our current understanding of T cell interactions with commensal bacteria that preserve immune homeostasis and avoid immunopathology.

2. Regulatory T cells play a crucial role in maintaining tolerance to commensal bacteria

Although the mucosal lining of the intestine provides an important protective barrier against commensal bacteria, it has become clear that this barrier is imperfect, resulting in a need for immune tolerance to gut antigens. One of the first clues came from early studies showing that a subset of CD4⁺ T cells [3], now known to be Foxp3⁺ regulatory T (Treg) cells, is required to prevent other CD4⁺ T cells from inducing colitis upon transfer into lymphopenic hosts (reviewed in [2, 4, 5]). Subsequent studies showed that commensal bacteria are important for driving intestinal pathology in the context of Treg cell deficiency For example, ablation of IL-2 leads to a loss of Treg cell number and function, which results in a variety of immunopathology [6]. However, only colitis is markedly reduced in germ-free (GF) IL-2-deficient mice [7–9]. Similarly, GF conditions limit colitis, but not other inflammatory manifestations, that arise with depletion of Foxp3⁺ cells in Foxp3^DTR mice [10], or with T cell deletion of Uhrf1 (Ubiquitin-like, with pleckstrin-homology and RING-finger domains 1), an epigenetic regulator that facilitates colonic Treg proliferation and maturation [11]. As memory T cells from conventionally housed mice are more efficient at inducing colitis [12], these data suggest that Treg cells act to restrain normal effector responses to commensal bacteria. Thus, Treg cells are important in establishing a tolerogenic environment to maintain immune homeostasis to commensal bacteria in the intestine.

Substantial progress has been made in the past several years regarding the origin and specificity of Treg cells involved in intestinal tolerance. Although it is still debated [13], several lines of evidence suggest that colonic Treg cells arise primarily from peripheral Treg cell differentiation from naïve T cells (pTreg, [14, 15]), as opposed to arising from the thymus during T cell development (tTreg, [16, 17]). First, adoptive transfer of naïve T cells into normal hosts suggested that the intestines are sites that facilitate peripheral Treg cell selection [18]. Second, markers that may distinguish thymic Treg cells such as high expression of Helios or Neuropilin-1 (Nrp-1) imply that ~80% of colonic Treg cells arise from peripheral selection [19–21]. Third, one T cell receptor (TCR) repertoire study suggested that most colonic Treg TCRs do not facilitate thymic Treg cell selection [22]. Fourth, a study of Foxp3-deficient mice devoid of Treg cells revealed that adoptive transfer of natural Treg cells was not sufficient to restore immune homeostasis, and required transfer of Foxp3⁻ cells capable of undergoing peripheral Treg cell selection [23]. Finally, studies of the Foxp3 locus revealed that mice deficient in conserved noncoding sequence-1 (CNS-1) had fewer Treg cells in the intestine, which was correlated with a defect in peripheral Treg cell selection [24, 25]. Taken together, these data support an important role for peripheral Treg cell selection in the gut.

However, other reports have argued against this interpretation. For example, the sensitivity and specificity of markers of thymic versus peripheral Treg cell selection such as Helios has been called into question [23, 26, 27]. Moreover, a different TCR repertoire study suggested that most colonic Treg cells are of thymic origin based on cross-referencing TCR sequences [28]. While future experiments are required to determine the exact proportions of pTreg versus tTreg in the colon, it is clear that the intestines are enriched in pTreg cells relative to other tissues in the body.

In regards to the antigen specificity of colonic Treg cells, recent studies suggest that they are often reactive to commensal bacteria. First, several studies reported that some, but not all, bacterial species can increase the frequency of Treg cells in the gut. For example, Clostridium clusters XIVa and IV [21, 29] or altered Schaedler flora (ASF) [30] introduced into GF mice markedly increased the frequency of colonic Treg cells. Lactobacillus reuteri [31] has also been associated with increased Treg cell percentage in the intestines. Second, analyses of TCR repertoires suggested that a substantial fraction and perhaps the majority of colonic Treg TCRs recognize commensal bacterial antigens. Since the TCR repertoire is extremely complex, mice with limited TCR repertoires were used to assess the colonic Treg TCR repertoire [22, 28]. In these studies, it was observed that colonic Treg cells utilized TCRs that were different than those used by Treg cells in other tissues or secondary lymphoid organs [22]. Further analysis of colonic Treg TCRs revealed direct recognition of antigens present in colonic contents, or in several cases, individual bacterial isolates [22, 28]. Consistent with TCR recognition of commensal bacterial antigens, antibiotic treatment could markedly change the colonic Treg repertoire [28], Taken together, this body of evidence strongly suggests that commensal bacteria routinely trigger antigen specific Treg cell responses.

3. Mechanisms that facilitate Treg cell selection in the gut

The enhancement in peripheral Treg cell selection to commensal bacteria may result from a number of mechanisms (Fig. 1). First, antigen presenting cell (APC) subsets in the intestine such as CD103⁺ dendritic cells (DCs) have been reported to favor Treg cell selection, as they express higher levels of retinal dehydrogenase (RALDH) to produce the vitamin A metabolite retinoic acid (RA) [18, 32, 33]. RA may inhibit effector cell cytokine production [34], as well as act directly on T cells [35] to promote Treg cell selection. Moreover, CD103⁺ DCs can generate transforming growth factor β (TGFβ) that acts in concert with RA to induce Treg cells [36]. Finally, it has been reported that these DC functions may be related to WNT/β-catenin signals that are delivered to intestinal, but not splenic DCs [37].

Figure 1. Mechanisms that facilitate Treg cell selection in the gut.

Commensal bacteria can promote the induction of Treg cells via direct sensing of microbial products through TLRs or metabolites such as SCFAs. Commensals or SCFAs can also induce tolerogenic DCs that favor Treg cell differentiation through the production of RA and TGF-β. Mucus and intestinal WNT can also trigger a β-catenin dependent tolerogenic program in DCs. The sites of interactions are not specified in this Figure, but may be in the mesenteric lymph nodes.

What might be the signals localized to the gut that promote these DCs to facilitate Treg cell selection? One recent report suggested that mucus from the intestinal lumen itself can activate tolerogenic pathways in DCs by triggering WNT signaling via β-catenin [38]. Mucus triggered WNT-signaling might be predicted to affect only the subset of DCs close to the mucosal surface. However, a TCF reporter of WNT signaling suggests that intestinal DCs receive a fairly uniform degree of WNT signaling [37]. Future studies are required to address the relative contributions of mucus versus other sources of WNT signaling in intestinal DCs.

Another potential intestinal signal for Treg cell selection may come from the microbiota itself. Short-chain fatty acids (SCFAs) arising from bacterial fermentation can act on DCs to promote tolerance [39–41]. SCFAs may also act directly on T cells themselves to promote colonic Treg cell expansion [42]. Another bacterial product that can affect Treg cells is polysaccharide A (PSA) from Bacteroides fragilis (B. fragilis), which triggers toll-like receptor (TLR) 2 on Treg cells to induce IL-10 production [43]. Thus, commensal bacteria themselves provide potent signals that are interpreted by the intestinal immune system to facilitate tolerance.

4. Effector cell generation to commensal bacteria

Despite the plethora of intestinal factors that favor Treg cell differentiation or expansion, Treg cells represent only 20–40% of CD4⁺ T cells in the colon. Effector cells therefore represent the dominant T cell population within the intestine during homeostasis [44, 45]. The generation of effecter T cell subsets in the intestine has been associated with signals derived from commensal bacteria, as GF mice harbor fewer Th1 and Th17 cells compared to conventionally housed mice [12, 46]. Also, DNA from gut bacteria plays an important role in the induction of gut resident Th1 and Th17 cells at steady state through the engagement of TLR9 [47]. Thus, the generation of intestinal effector T cell population is clearly influenced by the commensal bacteria.

It is less clear, however, whether intestinal effector T cells recognize commensal bacterial antigens. Indirect support comes from the observation that effector T cells induced in the presence of commensal bacteria are more efficient at inducing colitis than cells from GF mice [12]. However, this does not necessarily mean that effector cells are specific to commensal bacteria, as bacterial products may also cause antigen-non-specific changes in the environment that affect T cell trafficking, differentiation, and function. Direct support for commensal-specific effector cell generation comes from murine studies of Th17 cells, which were shown to be critically dependent on a single bacterial species-segmented filamentous bacteria (SFB) [48]. SFB is a spore-forming Gram-positive anaerobe residing primarily in the terminal ileum that makes intimate interaction with the mucosal barrier through tight attachments to epithelial cells. This interaction is a unique feature of SFB compared with other commensal bacteria that may underlie its ability to elicit Th17 cells. Recently, it was shown that the majority of Th17 cells are specific to SFB antigens [49, 50], suggesting that SFB provides both the dominant T cell epitopes as well as signals that facilitate Th17 differentiation. Thus, both Treg and effector cells can be elicited to commensal bacteria.

5. Effector versus regulatory T cell selection

The induction of both Treg and effector cells by commensal bacteria raises a fundamental question as to how the immune system determines Treg versus effector cell selection to bacterial antigens. Available data suggests that this is not a stochastic process, but may be instructed by specific bacterial species. For example, Th17 selection to SFB is not accompanied by Treg selection, as SFB-specific TCR transgenic cells become mostly RORγt⁺ and not Foxp3⁺ cells [49]. Similarly, it has been reported that CD44^hi effector T cells utilize different TCRs than Treg cells [22], indicating cell-fate determination based on TCR specificity.

The usage of different TCRs between Treg and effector cells could suggest that a biophysical property of TCR interaction with peptide:MHC, such as affinity or the amount of antigen, may determine peripheral T cell selection [51, 52]. In addition, recent studies indicate that environmental factors contribute to peripheral T cell differentiation to commensal bacteria. For example, SFB-specific TCR transgenic cells undergo Th1, and not Th17, development when their cognate antigen is expressed by Listeria and not SFB [49]. Moreover, TCR transgenic cells specific for the commensal CBir1 flagellin antigen adopt a Th1 phenotype upon infection with Toxoplasma gondii (T. gondii), but a Th17 phenotype after dextran sodium sulfate (DSS) injury [53]. Although studies of Treg versus effector cell differentiation to commensal bacteria are currently unavailable, OT-II TCR transgenic cells undergo peripheral Treg cell development in response to oral feeding of cognate antigen [18, 54], whereas they undergo effector cell generation in the context of infection [55]. Therefore, TCR affinity for antigen, antigen dose, and the cytokine milieu may all contribute to the signals that direct T cell lineage commitment.

Innate stimulators from commensal bacteria may direct T cell differentiation via selective activation of cytokine production from APCs by TLRs, which are important sensors that recognize conserved molecular motifs on bacteria. TLRs have been suggested to promote pro-inflammatory effector responses. For instance, mice deficient in MyD88 (Myeloid differentiation primary response gene 88), an important signaling component for many TLRs, are relatively resistant to intestinal inflammation resulting from Treg cell [56] or IL-10 [57] deficiency. Consistent with these observations, TLR9 sensing of commensal DNA has been suggested to favor intestinal effector (Th1/Th17) rather than Treg development [47]. However, a role for TLRs in Treg selection has also been suggested. TLR2 signaling has been reported to induce IL-10 and RA to promote Treg cell selection [58]. DNA of the probiotic Lactobacillus species is enriched in suppressive motifs that signal through TLR9 to prevent DC activation and maintain Treg-cell conversion during inflammation [59]. Moreover, B. fragilis PSA can trigger TLR2 signaling on Treg cells to induce IL-10 expression and promote tolerance [43]. One explanation for some of these contrasting reports is that TLRs are found on many cell types and contribute to facets of intestinal homeostasis not directly related to T cell differentiation signals, including epithelial integrity [60]. Thus, while it is likely that TLRs play an important role in T cell responses to commensal bacteria, their precise function in determining peripheral T cell differentiation to specific bacterial species remains to be determined.

APC subsets have also been implicated in Treg versus effecter cell determination [61, 62]. Lamina propria DC subsets may provide different cytokine environments that favor Treg versus Th17 selection. As discussed above, CD103⁺ DCs have been suggested to favor Treg cell selection via production of RA and TGFp [18, 32, 33]. Interestingly, the CD11b⁺ subset of CD103⁺ DCs have been found to be potent producers of IL-6 and IL-23 that favor Th17 differentiation [47, 63, 64]. Consistent with these observations, genetic depletion of the CD103⁺CD11b⁺ DC subset showed a decrease in Th17 cells [64–66], whereas depletion of both CD11b⁺ and CD11b subsets of CD103⁺ DCs were required to see a decrease in the number of Treg cells [65]. However, it is unclear whether the impact on the number of Th17 versus Treg cells is related to specific response to bacteria species or global effects on the overall cytokine environment, as the decreased number of Th17 cells was independent of MHC II expression on the CD103⁺CD11b⁺ cells [65].

Other types of APC have also been described to be involved in gut homeostasis [67]. Macrophages, perhaps via cytokine production, can affect the regulatory to effector cell balance in the gut [68, 69]. Recently, another cell type in addition to the traditional APCs (DCs and macrophages) has been suggested to perform an antigen presenting role in the gut. Innate lymphoid cells (ILCs) were recently shown to express MHC II and negatively regulate CD4⁺ T cell responses [70]. However, instead of inducing Treg cells, it was suggested that ILC antigen presentation inhibits the differentiation of effector T cells via an unknown mechanism.

In summary, there are many parameters that can control T cell differentiation to commensal bacteria. There are global factors such as SCFAs, WNT, TGFp, and RA that favor Treg cell selection. There are local factors such as TLR ligands, access to mucus, and different APC subsets that differentially favor Treg versus effector cell selection. Moreover, it is possible that some global factors are also modulated at the local level, and vice versa. Anatomic factors, such as small intestine versus colon may also play a role by affecting the resident bacterial constituents (e.g. SFB is mostly ileal), mucus amount and composition, mucosal barrier function, and APC types. Thus, much remains to be discovered regarding how individual commensal bacterial species trigger the specific mechanisms that determine Treg versus effector T cell differentiation.

6. Effector versus regulatory T cell function during homeostasis

During homeostasis, both regulatory and effector T cells with differing antigen specificities are generated to commensal bacteria. While thymically-derived Treg cells are likely to be involved at some level in intestinal tolerance to commensal bacteria [28], their functional role has not been clearly established. However, the importance of peripheral Treg cell selection in maintaining tolerance to commensal bacteria is supported by the aforementioned studies of T cell transfers into Foxp3-deficient mice [23], as well as studies of CNS-1-deficiency in the Foxp3 locus [24].

By contrast, the role of effector T cells in gut homeostasis to commensal bacteria has been more difficult to demonstrate. One hypothesis is that effector T cells represent an active immune response against certain commensal bacterial species. For example, it has been shown that CD4 T cells are required in the regulation of IgA production, and thus indirectly prevent bacterial translocation in mice lacking junctional adhesion molecule A (JAM-A, encoded by F11r) [71]. Similarly, the aberrant expansion of SFB in IgA deficient mice suggests that adaptive immunity can restrain certain commensal bacteria [72]. Other examples come from studies of innate cells, where depletion of ILCs revealed a marked dissemination and systemic infection of commensal bacteria Alcaligenes [73]. Furthermore, absence of T-bet, a transcription factor important for Th1 type responses, in innate cells in T and B cell-deficient mice results in loss of commensal bacterial control and inflammation [74]. However, T-bet deficiency in mice with T and B cells does not result in spontaneous colitis [74]. Similarly, spontaneous pathology has not been reported for mice deficient in the signature Th17 cytokine, IL-17 [75] or transcription factor RORγt [76]. Thus, effector T cells, perhaps via induction of IgA, may act to constrain certain species, but do not appear required for preventing large scale invasion by commensal bacteria during homeostasis.

Another possibility is that effector T cells have no major role in homeostasis, and are byproducts of an immune system that cannot perfectly determine whether each bacterial epitope is associated with a benign versus pathogenic bacteria [77, 78]. This is reminiscent of the observation that some self-reactive cells do escape thymic tolerance mechanisms that induce deletion or tTreg cell selection. While these autoreactive escapees of thymic selection appear quiescent during homeostasis, their pathogenic potential can be revealed after adoptive transfer into lymphopenic hosts [79] or Treg cell depletion [80].

A third possibility is that effector T cells to commensal bacteria exist to prime the immune system against pathogens. For example, mice harboring SFB are more resistant to Citrobacter Rodentium infection [48]. However, it remains possible that commensal bacteria can influence ILCs or other innate immune cells in addition to altering the effector T cell population. Thus, a clear role for commensal bacteria-specific effector T cells during homeostasis remains to be established.

7. T/B collaboration in response to commensal bacteria

The data discussed above suggests that both effector and regulatory T cells routinely interact with commensal bacteria under homeostatic conditions [45]. Similarly, the B cell arm of the adaptive immune system has also been found to interact with commensal bacteria during homeostasis [81, 82], with the majority of B cells in gut associated lymphoid tissue (GALT) directed towards bacterial antigens [83]. As suggested by a study of a commensal species Bacteroides thetaiotaomicron [84], IgA may limit the ability of commensal bacteria to trigger an inflammatory response.

Although T cell help is not essential for IgA class switching in the intestines [85], it appears that the paradigm of T cell help for B cells is applicable to intestinal IgA production during homeostatic conditions, as the presence of T cells is important for generating somatic hypermutation [86]. Moreover, TCR transgenic cells specific to a commensal bacterial flagellin antigen CBir1 induce IgA to that antigen [87]. Thus, these data suggests that commensal bacterial antigens are presented to both the B and T cell arms of the adaptive immune system during homeostasis.

One interesting observation from the TCR transgenic experiments was that depletion of Treg cell using anti-CD25 lead to a decrease in anti-CBir1 IgA, suggesting that Treg cells provide help for B cells [87]. Foxp3⁺ Treg cells have been reported to convert to follicular B helper T cells (Tfh) in lymphopenic settings [88]. However, an alternative possibility is that CD25 depletion may also affect effector cells. Additionally, Treg cells have also been reported to convert into follicular regulatory T cells (Tfr) [89, 90], which lack CD40L expression and instead secrete IL-10 and may act to inhibit, rather than provide, T:B collaboration. Thus, it remains unclear whether Treg cells, perhaps via transdifferentiation into other T cell subsets, provide T cell help for B cells.

Th17 cells appear to be an important source of B cell help, as large numbers of both antigen specific and nonspecific Th17 cells were reported to migrate to B cell germinal centers in the diffuse lamina propria during SFB colonization[91]. Other groups utilized cell fate reporters have shown that Th17 cells can also convert to Tfh cells in the gut [92]. However, as Th17 cells have not been reported to be readily induced to bacterial species other than SFB, the mechanism of T cell help for IgA class switching and affinity maturation that react to the rest of the microbiota remains unclear.

8. Are adaptive immune responses to commensal bacteria dynamically regulated?

An intriguing observation regarding IgA produced in response to commensal bacteria is that it appears to be dynamically regulated [93]. While colonization of GF mice with one species resulted in a specific and long-lived IgA response, this response was lost with the addition of other commensals. Moreover, repeated exposure to the commensal did not result in the booster effect that might be expected based on the classic immunization paradigm. These data suggest that IgA responses to commensal bacteria may not fit the pattern of classic long lived plasma cells that can last the life of an individual, but rather are dynamically regulated by the current commensal bacterial population.

Indeed, there are other peculiarities of the GALT B cell biology that distinguishes the IgA response from a typical systemic B cell response. For example, mucosal associated IgA plasma cells do not seem to home to the bone marrow (BM), and are instead retained in the gut, where APRIL- (a proliferation inducing ligand) secreting neutrophils have been suggested to aid in B cell maintenance and survival [94]. Additionally, gut IgA plasma cells, unlike BM plasma cells, retain functional B cell receptors on their surface [95], which may provide survival signals in lieu of BM stroma-derived cytokines. Whether these differences in IgA B cell biology account for the dynamic nature of IgA responses to commensals is unclear.

An intriguing question is whether all adaptive immune responses to commensal bacteria are dynamic and relatively short lived. Treg cells have been suggested to be regulated by the amount of self-antigens, as the Treg TCR repertoire changes with the anatomic location of secondary lymphoid organs [96, 97]. Similarly, the number of small intestinal Th17 cells has been shown to decrease when the mice are treated with vancomycin, an antibiotic that kills SFB [98]. When SFB is reintroduced, Th17 numbers go up, consistent with dynamic regulation of the T cell population. These observations in Treg and Th17 cells support the notion that continual competition for antigen may drive the intestinal immune system. This does not preclude the possibility that T cell memory can co-exist with dynamic regulation, as strong responses may elicit memory in the effector [53, 99] and regulatory T cell population [100].

In addition to the provision of antigens that maintain T cell “fitness”, commensal bacteria may also tune the T cell population through other factors. For example, it has been shown that commensal bacteria-stmulated IL-23 [101] and IL-15 [102] promote inflammatory T cell responses. Similarly, RA was required to elicit proinflammatory T cell responses to infection and mucosal vaccination [103]. In addition, IL-12 can enhance the antigen responsiveness of T cells to sustain an ongoing autoimmune responses [104]. Finally, mice with intestinal epithelial cell-specific deficiency of caspase-8 or FADD (Fas-Associated protein with Death Domain) have been shown to spontaneously develop terminal ileitis or colitis associated with loss of Paneth cell and epithelial cell necrosis [105, 106]. In these mice, commensal bacteria-induced TNF-α trigger the programmed necrosis of intestinal epithelial cells which drives the inflammatory response and epithelial inflammation. In aggregate, these data suggest that the T cell population is also dynamically tuned by antigen-independent factors elicited by commensal bacteria.

Why might dynamic regulation of adaptive immunity to commensal bacteria occur? One possibility is that pathogenic bacteria are encountered episodically, requiring memory, whereas commensal bacteria are always in contact with the host. A dynamically regulated Treg cell population would be tuned to the current commensal microbiota and be more responsive to a large influx of commensal antigens from mucosal injury, thereby limiting effector T cell activation and excessive inflammation. By contrast, bacteria that are new to the intestinal tract would not trigger pre-existing Treg or effector T cells, but require de novo selection of effector versus regulatory T cell responses. This would allow more rapid effector responses to pathogens and limit the buildup of effector T cells responsive to commensal bacteria that might trigger IBD. Future experiments are required to address the conjecture that T cell responses to commensal bacteria are amnestic and dynamically regulated by clonal competition for antigen.

9. How is tolerance to commensal bacteria broken?

For most individuals, tolerance to commensal bacteria remains intact throughout their lifetime. However, approximately 0.44 % of individuals break tolerance and develop IBD [107]. Genome wide association studies (GWAS) have clearly identified a number of genetic risk factors for IBD that affect intestinal epithelial function, general immune regulation, and innate immune function [108, 109]. Defects in non-T cell related genes may increase bacterial translocation or outgrowth, triggering an inflammatory response from otherwise normal T cells. Defects in T cell related genes, such as HLA, suggest that T cell interactions with commensal bacteria may also be abnormal in IBD [108]. Thus, breakage in T cell tolerance to commensal bacteria could include a primary T cell defect or arise secondary to other perturbations of intestinal immune homeostasis.

In addition to genetic factors, an intriguing question is whether environmental insults can break tolerance. This is suggested by the observation that IBD is increasing in geographic areas such as North America [107]. An increased risk of IBD has also been associated with recent gastrointestinal infections [110]. While commensal bacterial exposure during mucosal breaches may be reduced by a containment structure referred to as intraluminal cast comprised of monocytes and neutrophils [111], intestinal infections can expose the immune system to a variety of commensal bacteria [112]. Moreover, gut infections are associated with dysbiosis with significant shifts in microbiota composition. For example, γ-proteobacteria can become dominant due to their ability to thrive under inflammatory condition [113, 114]. Thus, infection may result in the exposure to the immune system of previously encountered commensals as well as those that typically do not have access during homeostasis.

A recent study convincingly showed that exposure of commensals to the immune system can occur during mucosal breaches [53]. They used CBir TCR transgenic T cells that recognize flagellin expressed by certain commensals including the Clostridium cluster XIV class of bacteria. Interestingly, these T cells typically possess a naïve phenotype during homeostasis, suggesting that they are not routinely exposed to antigen. However, barrier breach results in their development into Th1 cells during a highly Th1 polarizing T. gondii infection, or Th17 cells after DSS administration. These T cells persist long term in both the intestinal and secondary lymphoid tissues, and respond like memory T cells upon rechallenge. Although physical segregation between the microbiota and immune system is rapidly restored after resolution of infection, the long-term presence of commensal-specific memory T cells may fundamentally alter the balance of tolerance. One could speculate that over time, multiple infections may lead to the expansion of commensal-bacteria reactive effector T cells, potentially leading to a feed-forward loop in which these T cells can themselves overwhelm Treg cells and cause enough inflammation to lead to barrier breach and further antigen presentation [115]. This could create a chronic and self-sustaining inflammatory state, perhaps culminating in IBD.

And yet, intestinal infections associated with diarrhea are common and often do not trigger IBD [110, 116]. Nor do other intestinal disorders such as diverticulitis or appendicitis that might be predicted to result in increased exposure to commensal bacteria. Perhaps the experience with CBir is not representative of most T cell responses during infection as its antigen is not constitutively presented to the immune system. Antigens that are constitutively presented may have already triggered the appropriate effector or regulatory T cell response. Additionally, the CBir antigen is flagellin, a TLR5 ligand, and may not be representative of most commensal bacterial ligands. Thus, it remains unclear the extent that infection or other injuries that result in mucosal barrier breakdown affects the balance between effector and regulatory T cell responses to commensal bacteria.

10. Summary

Recent work has demonstrated that the adaptive immune system has continuous interactions with commensal bacteria to induce both regulatory and effector T cells that promote tolerance and immunity, respectively The differences in the Treg versus effector cell TCR repertoires suggest that these T cell populations perform unique functions to maintain host health. However, many questions remain. What are the mechanisms by which the immune system determines regulatory versus effector cell differentiation? How is this affected by infections or other diseases of the intestinal tract? What is the role of effector cells in maintaining a healthy immune environment? Addressing these questions is of considerable interest to understanding the fundamental immunologic question of how we establish immune tolerance to foreign antigens, but also raises the possibility that such knowledge may be useful for the treatment or prevention of IBD or other allergic and autoimmune diseases of the intestines.