The battle against immunopathology: infectious tolerance mediated by regulatory T cells

Abstract

Infectious tolerance is a process whereby one regulatory lymphoid population confers suppressive capacity on another. Diverse immune responses are induced following infection or inflammatory insult that can protect the host, or potentially cause damage if not properly controlled. Thus, the process of infectious tolerance may be critical in vivo for exerting effective immune control and maintaining immune homeostasis by generating specialized regulatory sub-populations with distinct mechanistic capabilities. Foxp3⁺ regulatory T cells (T_regs) are a central mediator of infectious tolerance through their ability to convert conventional T cells into induced regulatory T cells (iT_regs) directly by secretion of the suppressive cytokines TGF-β, IL-10, or IL-35, or indirectly via dendritic cells. In this review, we will discuss the mechanisms and cell populations that mediate and contribute to infectious tolerance, with a focus on the intestinal environment, where tolerance induction to foreign material is critical.

Tolerance takes multiple forms

One of the most fundamental questions in immunology is how the immune system can target the diverse array of environmental pathogens we encounter on a daily basis while simultaneously maintaining non-reactivity, or tolerance, to self. A detailed understanding of tolerance mechanisms is beginning to emerge, presenting new therapeutic applications for the restoration of tolerance to self-antigens in autoimmune disease, promoting tolerance or ignorance to innocuous non-self-agents such as food antigens and commensal microbiota, or breaking tolerance to chronic pathogens or aberrant self (e.g., cancer) [1–3]. Research over the last few decades has led to the realization that the immune system employs many strategies to exert tolerance, from the deletion of auto-reactive T cells in the thymus, termed ‘central tolerance’, to dominant mechanisms of ‘peripheral tolerance’ where suppressive populations, such as Foxp3⁺CD4⁺ regulatory T cells (T_regs), halt the activation of auto-reactive T cells. Given that T_regs are generally present at low numbers, a key process that appears to be required for the peripheral control of the immune system is termed ‘infectious tolerance’ whereby one suppressive lymphoid population can confer suppressive capabilities on another [4, 5]. A complete understanding of the mechanisms of infectious tolerance will lead to advanced therapeutics that will impact the fields of transplantation biology, infectious disease, autoimmunity, and cancer research. The focus of this review is to detail our current understanding of infectious tolerance with a particular focus on the ability of T_regs to generate a regulatory microenvironment that nurtures the induction of infectious tolerance. This is especially relevant in mucosal sites, where a barrage of foreign, mostly innocuous antigens are present and must be tolerated in order to avoid inflammatory disease [6].

The origins of infectious tolerance

Much of our understanding of peripheral and infectious tolerance has come from studies in transplantation immunology. Seminal studies by Billingham and Medawar [7] on skin grafting in laboratory animals revealed that newborn mice could tolerate allogeneic skin grafts that adults would reject. Later, it was realized that suppressive cells mediate this tolerance and that their suppressive capacity could be transferred from one population of cells to another. Gershon and Kondo [4] were the first to describe this phenomenon in the 1970s, coining the phrase ‘infectious immunological tolerance’. Subsequent studies by Gershon and colleagues, and later by North and colleagues, suggested that the suppressive population was a Ly1⁺ (CD5) Ly2⁻ (CD8) CD4⁺ T cell subset [8–11]. However, it was the work of Waldmann et al. [12–14] which firmly established the concept of infectious tolerance with the demonstration that tolerance to allogeneic skin grafts could be therapeutically induced by pre-treating recipients with non-depleting, blocking antibodies to the T cell co-stimulatory molecules CD4, CD8, or CD154. T cell adoptive transfer at various points during the tolerance induction protocol revealed that tolerance took several weeks to initiate, but once established was long lasting and could be transferred [15]. Subsequent studies with congenically marked T cells demonstrated that tolerance could be transferred from one T cell population to another [16]. Amazingly, this infectious tolerance appeared to occur sequentially and seemingly indefinitely in this model.

Insight into the mechanism of infectious tolerance was provided by experiments which demonstrated the phenomenon of linked suppression (Fig. 1). The observation was that a recipient mouse of strain A tolerized to a graft from donor strain B was also capable of accepting a second graft of an F1 cross of a strain B mouse with a strain C mouse, but a strain C graft alone was rejected [17]. After the tolerance-inducing period, a new graft from a pure strain C mouse could then be tolerated in the absence of strain B antigens. This implied that infectious tolerance was expanded from strain B antigens to strain C antigens due to them being present on the same graft. Even placing grafts from strain B and strain C side by side in the same graft bed was not sufficient to induce infectious tolerance to graft C. This led to the notion that a highly localized regulatory microenvironment is critical for the spread of tolerance from one T cell population to another.

Fig. 1.

Infectious tolerance through linked suppression. Allogeneic transplantation models have revealed that tolerance to a particular graft induced through co-stimulatory molecule blockade can be transferred to unrelated antigens. This is possible if a second graft from an F1 cross of the tolerated strain and an unrelated strain is given, highlighting the importance of a localized regulatory environment [17]. Linked suppression may be the result of tolerated antigens and foreign antigens being presented by the same DCs or in a highly localized cytokine environment conducive to T_reg induction. After tolerance induction to the F1 graft is established, a graft from the unrelated strain can be tolerated in the absence of the initially tolerated antigens

Foxp3⁺ regulatory T cells and tolerance

Suppressor T cell studies received early skepticism due to the tainted legacy of the suppressor cell era. The seminal discovery that the suppressive CD4⁺ T cell population could be uniquely identified by expression of the alpha chain of IL-2 receptor (CD25), led to the demonstration that the transplantable, suppressive T cell population was restricted to CD4⁺CD25⁺ T cells [18]. Adoptive transfer of CD25-depleted CD4⁺ T cells into T cell-deficient, nude mice resulted in systemic autoimmunity, which could be prevented by co-transfer with CD4⁺CD25⁺ T cells within a limited period of time [19]. The discovery of the forkhead transcription factor Foxp3 as the defective gene in human immunodysregulation polyendocrinopathy enteropathy X-linked (IPEX) syndrome patients and scurfy mice, which both develop severe systemic autoimmunity, provided the first tangible marker for CD4⁺CD25⁺ T_regs [20–23]. Subsequent experiments proved that Foxp3 expression is restricted to the suppressive CD4⁺CD25⁺ T cell subset, is necessary for the development of suppressive cells, and when introduced via retroviral transduction can confer suppressive capacity on naïve CD4⁺ T cells [24, 25]. Collectively, these seminal studies established Foxp3 as a master regulator of T_reg development and function. Several studies have subsequently demonstrated that Foxp3⁺ T_regs are the primary suppressive population that mediates immunological tolerance to skin grafts. Female A1 Rag1 ^−/− TCR transgenic mice, specific for the male antigen Dby, fail to develop thymus-derived ‘natural’ Foxp3⁺ T_regs and readily reject male skin grafts. However, under tolerance induced through co-receptor blockade, Foxp3⁺ T_reg induction was observed that coincided with the establishment of tolerance [26]. These ‘induced’ T_regs (iT_regs) within the tolerated tissue were required to maintain tolerance to the allograft. This was demonstrated through transfer of the tolerated graft into Rag1 ^−/− recipient mice, which lack the capacity to reject the graft on their own [26, 27]. Treatment with the CD25-depleting antibody, PC61, led to T_reg depletion within the transferred graft which was rapidly rejected due to the removal of suppression and restoration of alloreactivity by the graft-resident T cells. These observations were recently supported by studies in transgenic mice expressing human CD2 under the control of the Foxp3 promoter, which allowed targeted deletion of T_regs using ablative anti-human CD2 antibodies [28]. Thus, Foxp3⁺ T_regs are required for the induction and maintenance of tolerance.

A variety of CD4⁺ iT_reg populations have been shown to mediate dominant, infectious tolerance in a variety of model systems, and thus they will be the primary focus for the remainder of this review. However, other suppressive lymphoid populations have been described, such as regulatory B cells and CD8⁺ regulatory T cells. However, their role in mediating infectious tolerance has not been extensively examined and they will not be discussed further. The diversity of iT_reg populations that can mediate infectious tolerance should not be surprising since the types of immune responses and environments in which regulatory populations must act are also diverse. However, evidence is emerging that there are several common characteristics of what constitute an environment capable of suppressing immune responses and inducing infectious tolerance.

Infectious tolerance mediated by T_reg-derived cytokines

Reductionist in vitro studies have provided considerable insight into the suppressive mechanisms used by T_regs and their ability to transfer their suppressive capacity to another T cell population (Fig. 2). Considerable attention has focused on TGF-β, a pleiotropic cytokine that appears to play a central role in immune tolerance [29]. Genetic deletion of TGF-β1, which is predominantly expressed in the immune system, results in multi-organ immunopathology [30]. TGF-β induces Foxp3 expression by TCR-stimulated mouse and human T cells and confers suppressive capacity [31, 32]. TGF-β has also been suggested to regulate Foxp3⁺ T_reg development in the thymus [33]. Upon activation, the latent form of TGF-β is detected on a high percentage of mouse and human T_regs [34]. However, the role of TGF-β as a T_reg effector cytokine remains controversial [35]. Given the in vitro effects of TGF-β on T_reg conversion, and the observation that latent TGF-β is expressed on activated T_regs, it was hypothesized that T_reg-derived TGF-β could be a key mediator of infectious tolerance, even in the absence of antigen presenting cells. Indeed, it was shown that Foxp3 is induced in effector T cells stimulated in the presence of pre-activated, latent TGF-β-expressing T_regs in a TGF-β- and cell contact-dependent manner, and that these iT_regs were suppressive in vitro and in vivo [36]. In these experiments, prior activation of T_regs and latent TGF-β expression were key, as freshly activated T_regs were not strong mediators of infectious tolerance. Lastly, human T_regs have been shown to mediate infectious tolerance in vitro by induction of a regulatory population in a TGF-β-dependent manner [37].

Fig. 2.

Cytokine-mediated infectious tolerance by T_regs in the presence or absence of DCs. Direct infectious tolerance via the cytokines IL-10, TGF-β, and IL-35 have been described in vitro in the absence of a requirement for antigen presenting cells. Foxp3⁺ T_regs have also been described to interact with DCs to promote a tolerogenic phenotype. This occurs through modulation of co-stimulatory molecule expression and promotion of tolerance-inducing factors like IDO and immunosuppressive cytokines, leading to an environment that promotes the induction of T_regs

The immunosuppressive cytokine IL-10 has been shown to induce a population of functionally suppressive CD4⁺ T cells, sometimes referred to as Tr1 cells [38, 39]. They share in common dependence on IL-10 and TGF-β to mediate their suppressive effects but differ from conventional T_regs by lacking expression of Foxp3 [39]. Human T_regs have been shown to induce Tr1 cells in vitro that are capable of suppressing third party T cells in an IL-10-dependent manner [40, 41]. In the experimental autoimmune encephalomyelitis (EAE) mouse model of multiple sclerosis, CD4⁺CD25⁺ T_regs have also been observed to convert antigen-specific pathogenic effector T cells into IL-10-producing Tr1 cells [42, 43]. These converted cells are capable of suppressing EAE in an IL-10-dependent manner. The development of IL-10 reporter mice has provided increased spatiotemporal resolution of IL-10 expression [44]. In vivo, neither Foxp3⁺ nor Foxp3⁻ IL-10-producing T_regs are dependent on IL-10 expression to develop, as reporter expression can still be detected in IL-10 genetic knockouts. However, IL-10⁺ T_reg development is dependent on TGF-β signaling as TGF-β antibody neutralization can abrogate their induction. Recently, it has been shown that Foxp3⁺ T_regs respond to IL-10 by increasing production of IL-10 through a feed-forward mechanism that is critical for controlling T_H17 responses, particularly during intestinal inflammation [45–47]. It has also recently been observed that pathogenic T_H17 cells can be converted in the intestinal environment into suppressive T_H17 cells that express high levels of IL-10 and are partially dependent on IL-10 for their suppressive capacity [48]. However, it is unclear if T_reg-derived IL-10 acts directly to confer suppressive capacity on this pathogenic T_H17 population through infectious tolerance or if other cellular and molecular mediators are required.

IL-35 is a heterodimeric cytokine composed of IL12a/p35 and Ebi3 that suppresses both mouse and human T cells. IL-35 secretion appears to be restricted to T_regs and is required for their maximal suppressive capacity in vivo [49, 50]. Furthermore, T_reg-derived IL-35 can induce the conversion of CD4⁺ effector T cells into a regulatory population that mediates suppression via IL-35 production, termed iTr35 [49]. iTr35s do not express Foxp3, TGF-β, and IL-10. Adoptive transfer studies with iTr35s have revealed their remarkably protective capacity and stability in five distinct autoimmune and inflammatory models in vivo. Importantly, T_reg-mediated iTr35 induction appears to contribute about half the suppressive microenvironment in B16 melanoma, suggesting that iTr35 may contribute to infectious tolerance in vivo [49]. Human T_regs can also express IL-35 and mediate iTr35 induction in an IL-35-dependent fashion [51]. Further understanding of the role of T_reg-derived IL-35 in vivo in mediating infectious tolerance is clearly required.

Infectious tolerance mediated via dendritic cells

Experiments suggesting that infectious tolerance may be the result of linked suppression led to the hypothesis that antigen presenting cells in the tolerated tissue may be playing a critical role in this process. Dendritic cells (DCs) contribute to the generation of iT_regs in the periphery. DCs are the most potent cellular inducer of T_regs and DC immaturity is correlated with increased efficiency of T_reg generation [52, 53]. CD8⁺ DCs in the spleen are more capable than CD8⁻ DCs at inducing T_regs in the presence of TGF-β [54]. This conversion process also requires signaling through the programmed death ligand 1 (PD-L1). It has also been observed that plasmacytoid DCs (pDCs) are particularly good inducers of T cell tolerance [55]. CCR9⁺ pDCs have been shown to induce Foxp3⁺ T_regs and suppress antigen-specific immune responses both in vitro and in vivo [56]. Bone marrow-derived DCs preconditioned with TGF-β, IL-10, or vitamin D3 have also have been reported to induce tolerance in a skin graft model [57, 58]. In humans and mice, a subset of IL-10-producing DCs have been characterized in vivo, and were found to be potent inducers of Tr1 cell in vitro [59, 60]. Antibody-mediated blockade of IL-10 signaling could inhibit the induction of Tr1 cells, indicating that DC production of IL-10 was critical. In addition to being found naturally, these IL-10-producing DCs can also be induced in vitro from monocytes in the presence of IL-10 [59]. It is tempting to speculate that these IL-10-producing DCs may be a required intermediate in vivo for IL-10-mediated infectious tolerance via Tr1 cells.

Data also show that DCs are able to induce human iTr35 generation. Rhinovirus-infected DCs cultured with T cells caused profound anergy and resulted in up-regulation of IL-35 [61, 62]. These Foxp3⁻ iTr35s were able to suppress naïve T cell proliferation, comparable to results in the mouse. Furthermore, induction of human iTr35s by infected DCs was dependent on PD-L1 and CD169 expression on DCs. It will be interesting to determine if iTr35s generated via virus-exposed DCs or in the presence of Foxp3⁺ T_regs possess a similar phenotypic and functional profile.

A common observation in several studies of DC-induced tolerance is that an environment which strongly promotes T cell anergy or unresponsiveness often potentiates T_reg induction (reviewed in [52, 63]). This has led some to hypothesize that sub-optimal T cell activation by DCs under certain conditions is required for effective induction of infectious tolerance. However, the molecular details of this remain elusive. Tolerogenic DCs that preferentially induce iT_reg induction over effector T cell activation tend to exhibit low constitutive MHC expression, a low ratio of co-stimulatory to co-inhibitory molecules, and have an impaired ability to synthesize T_H1-inducing cytokines (such as IL-12) [64]. It is not yet clear if anergy induction precedes iT_reg development in a linear and dependent fashion, or if these two events are independent outcomes of an anergy-inducing environment. Although questions and challenges remain, these findings have led some to attempt to utilize in vitro tolerized DCs in the clinic to drive immunosuppression [65].

The factors that induce T cell anergy have been expanded from the classic two-signal paradigm. Initially, anergy was attributed to an antigen-specific signal through the TCR (signal 1) in the absence of a co-stimulatory signal mediated by CD28 ligation (signal 2), which resulted in lack of IL-2 transcription (reviewed in [66]). Upon full T cell activation, signaling through the IL-2R activates the PI3K/AKT/mTOR pathway leading to cell cycle entry [67]. It is now recognized that several pathways can provide this second signal. Thus, any signaling events that negatively affect the mTOR pathway and metabolism in stimulated T cells can induce anergy. For instance, inhibition of mTOR signaling with rapamycin is sufficient to induce anergy in the presence of antigen plus co-stimulation [68, 69]. Since mTOR is downstream of several nutrient sensing pathways, and the energy-sensing AMP-activated protein kinase complex, which inhibits mTOR upon detection of a low ATP:ADP ratio (reviewed in [70, 71]), the convergence of metabolic information is a determining factor in the induction of anergy versus activation. Conditional knockout of mTOR in CD4⁺ T cells has been shown to result in preferential induction of Foxp3, highlighting the importance of this signaling pathway in tolerance induction [72]. Likewise, anergy can result from mTOR independent signaling pathways such as the GCN2 amino acid-sensing pathway [73]. GCN2 is a kinase that becomes activated upon binding to uncharged transfer RNAs, acting as a detector of amino acid deprivation. GCN2 activation results in S phase arrest and inhibition cell cycle entry [74]. Extracellular adenosine, which is generated by ATP hydrolysis, can act as a sensor of hypoxia. Adenosine signals through the A2A receptor and stimulates adenyl cyclase to elevate intracellular levels of cAMP (reviewed in [75]). Adenosine-mediated reduction in Ras/MAPK signaling results in T cell anergy in vitro and in vivo and the generation of induced T_regs [76]. Thus, an anergy-inducing environment conducive to T_reg generation is one that lacks co-stimulatory molecules, nutrients, and/or energy.

T_regs alter DC functionality

In order to confirm that T_regs drive the generation of new iT_regs through an APC intermediate, experiments needed to be performed which demonstrate that DC function is altered by T_regs. Indeed, there is mounting evidence showing that Foxp3⁺ T_regs are capable of altering DC phenotype such that they promote tolerance induction of newly primed T cells (Fig. 2). Two-photo laser scanning microscopy experiments have confirmed that strong interactions occur between T_regs and DCs in lymph nodes [77, 78]. In the absence of T_regs, T cells were shown to have reduced locomotion and longer contact duration with DCs, which correlated with elevated cytokine production in the lymph nodes. Interactions between T_regs and naïve T cells were not observed in these studies, suggesting that the primary target of T_regs in these systems was the DC. However, these studies do not rule out the possibility that T_regs directly alter T cell function and iT_reg conversion, but their interaction is very brief and/or cytokine-driven and thus does not require direct contact.

T_regs may also mediate linked suppression and infectious tolerance by altering the ability of DCs to provide co-stimulation. Indeed, early in vitro studies using mouse and human DCs found that co-stimulatory molecules were down-regulated in the presence of T_regs [79, 80]. T_regs have also been shown to physically outcompete naïve T cells for DC-presented antigen and co-stimulation in vitro. Antigen-specific T_regs form stable clusters around DCs inhibiting their maturation and choking off productive interactions with naïve T cells [81]. These interactions resulted in decreased expression on CD80 and CD86 in the DC, which is partially dependent on T_reg expression of CTLA-4 [82]. Recently, CTLA-4 on T_regs was shown to interact with CD80 and CD86 on DCs and mediate their trans-endocytosis into the T_regs where they were subsequently degraded [83]. This mechanism has been shown to occur both in vivo and in vitro, with a high proportion of T_regs utilizing this mechanism in vivo. The cell surface inhibitory molecule LAG-3 has been shown to be required for maximal T_reg function and is expressed by activated CD4⁺ and CD8⁺ T cells [84–86]. Interestingly, LAG-3 has been shown to inhibit DC maturation and their ability to present antigen by transducing negative signals in an MHC class II-, FcγRγ-, and SHP-1-dependent manner [87]. Via these mechanisms, the resulting paucity of TCR signal strength and costimulation can result in incomplete activation, anergy, and potentially iT_reg conversion. One caveat is that many, but not all, of these proposed mechanisms have not been demonstrated in vivo, thus their physiological impact has yet to be assessed.

T_regs may also broaden their influence by transducing signals into DCs that alter nutrient availability. The ligation of CTLA-4 on T_regs with CD80/CD86 on DCs induces their expression of Indoleamine 2,3-dioxygenase (IDO), a tryptophan-metabolizing enzyme [88, 89]. In vivo IDO induction in response to CpG DNA is dependent upon PD1-expressing T cells, as Rag1 ^−/− splenic DCs do not up-regulate this enzyme [90]. Other catabolic enzymes may also be involved in T_reg-mediated infectious tolerance as it has been shown that T_regs specifically induce expression of several essential amino acid-depleting enzymes both within skin grafts and in cultured DCs [91]. In response to the resulting amino acid depletion, T cells fail to proliferate due to reduced mTOR signaling and instead up-regulate Foxp3 expression in a TGF-β-dependent manner. IDO may also have a non-enzymatic role in promoting long-term pDC-mediated tolerance [92]. TGF-β can signal independently of Smad to induce both IDO and TGF-β expression in pDCs. These tolerogenic pDCs can induce Foxp3⁺ T_regs in vitro in a TGF-β-dependent manner and retain their tolerogenic phenotype in vivo for at least 3 months, implying that this may be a mechanism through which infectious tolerance occurs. Since TGF-β can be produced by many cell types, and is present in a latent form bound to the extracellular matrix, it would be interesting to determine if TGF-β-derived from activated T_regs is sufficient to drive IDO and TGF-β production from pDCs and mediate infectious tolerance.

The intestinal mucosa as a specialized environment for inducing tolerance

The gut is recognized as being an anatomical location rich in immune suppressive factors such as TGF-β and IL-10, and possessing multiple regulatory lymphocytic and tolerogenic DC populations [93, 94]. This bias towards a tolerogenic environment is critical for avoiding potentially disastrous inflammatory responses to beneficial bacterial symbionts and food antigens. Tolerogenic DCs in the Peyer’s patches and intestinal lamina propria express high levels of IL10 and thus may shape the intestinal microenvironment [95, 96]. Furthermore, in the presence of exogenous TGF-β, DCs isolated from the lamina propria of the small intestine and from the mesenteric lymph node are superior to splenic DCs at inducing Foxp3 expression by activated T cells [97, 98]. A high proportion of DCs isolated from mucosal tissues express the αE integrin CD103 and are potent inducers of iT_regs compared with other DC populations. Importantly, this conversion is mediated without the addition of exogenous factors due to their ability to release large amounts of bioactive TGF-β [97, 98]. This feature is clearly critical, as α_vβ₈ integrin-deficient DCs, which cannot release bioactive TGF-β, fail to generate T_regs in culture, and mice with a leukocyte-restricted deletion of α_vβ₈, which is predominantly expressed by DCs, lack colonic Foxp3⁺ T_regs and thus develop colitis [99]. The gut is also an environment rich in retinoic acid (RA), which enhances the capacity of TGF-β to induce T_regs [97, 98, 100, 101]. The target population for RA during iT_reg induction has been contentious. RA can act directly on naïve CD4⁺CD25⁻CD44^loCD62L^hi cells to promote their conversion in vitro, even in the absence of endogenous and exogenous IL-2 [102]. However, RA can also partially act through the inhibition of cytokine production by CD4⁺CD44^hi effector/memory cells, which functionally block iT_reg conversion [103]. Importantly, RA receptor antagonists can abrogate the T_reg induction seen in the presence of small intestine lamina propria DCs, and naïve OT-II T cells transferred into RA receptor α-deficient mice exhibit limited iT_reg conversion, indicating that RA signaling plays a critical role in the peripheral T_reg induction process [97, 98, 103]. In addition, the mesenteric lymph node-resident DCs have constitutively elevated levels of the tolerance-inducing enzyme IDO relative to splenic DCs [88].

Several studies have highlighted the central role of IL-10 in maintaining a suppressive microenvironment in the gut. Using IL-10 reporter mice, high frequencies of Foxp3⁺IL10⁺ T_regs have been noted in the lamina propria, while a particularly large proportion of Foxp3⁻IL10⁺ Tr1-like cells were observed in the Peyer’s patches and within the intraepithelial lymphocyte population [44]. Anti-TGF-β treatment results in reduced conversion of CD4⁺Foxp3⁻ to CD4⁺Foxp3⁺IL10⁺ T_regs in the gut, implicating TGF-β as an important factor in this process. Recent evidence highlights an interesting new mechanism to eliminate inflammatory cells in the gut. In vivo CD3-specific antibody treatment, which leads to T_H17 induction, results in an exodus of T_H17 cells from the body via the lumen of the small intestine [48]. However, T_H17 cells remaining in the small intestine acquired a suppressive phenotype that could be abrogated by blocking the IL10, CTLA-4, and TGF-β pathways simultaneously. Taken together, these observations suggest that the intestinal microenvironment possesses many unique cellular and molecular features which predispose it toward the generation of tolerogenic responses.

The contribution of commensal bacteria in promoting tolerance

Many of the tolerogenic signals in the gut may be derived from interactions between the immune system and intestinal microbes. The hygiene hypothesis postulates that early encounter with microbial agents in life inhibits the later development of allergic and autoimmune responses [104]. There are substantial epidemiological data to support this notion. Several recent reports have highlighted the importance of commensal bacteria on the development of pro- and anti-inflammatory responses (reviewed in [105]). Germ-free mice have profoundly underdeveloped gut-associated lymphoid tissues including Peyer’s patches, mesenteric lymph nodes, and intestinal lamina propria [106]. Several recent studies have shown that specific microbes, such as Clostridium spp., can mediate the induction of Foxp3⁺ T_regs [107–109]. This mechanism of Foxp3 induction is likely crucial for expanding the specificity of the T_regs repertoire to include commensal bacteria and dietary antigens to promote intestinal homeostasis.

Tolerance-inducing signaling pathways modulated by microbes can act by altering DC function or by directly acting on T cells. The bacterium Bacteroides fragilis has recently emerged as a model for characterizing microbiota-derived tolerogenic signals. The B. fragilis molecule polysaccharide A (PSA) signals through TLR2 on the surface of T cells, which results in immunoregulation rather than inflammation [107, 110]. In the absence of PSA, B. fragilis was unable to colonize the mucosa and instead elicited a T_H17 response. PSA-mediated induction of IL-10 expression by CD4⁺ cells was observed, which appears to be critical as Il10 ^−/− mice cannot be protected from IL-17-mediated colitis. Likewise, resistance to T_H17-mediated pathology was dependent on T_regs, as their depletion abrogated PSA-mediated protection [110]. Recent evidence corroborates the important role of IL-10 production by T_regs in suppressing T_H17-mediated intestinal pathology [45]. Autocrine IL-10:IL10R signaling in T_regs has been shown to amplify IL-10 production via a Stat3-dependent pathway [45]. Thus, it is possible that microbial-derived signals propagate IL-10 signaling and thus promote a suppressive microenvironment, which is particularly effective in suppressing T_H17 cells. This finding has led to the hypothesis that signals from commensal microbes result in defined immunoregulatory populations and pathways that have co-evolved to perfect a mutualistic partnership between the host and symbiont [105].

Helminth parasites as immunoregulators

Helminth parasites have also been shown to induce T_regs and suppress allergic disease in several models [111]. The capacity of helminths to act as immune suppressors would argue against the notion that allergic disease reduction is simply due to a systemic deviation away from T_H2 responses, since many helminths that dampen allergy also induce strong T_H2 responses. Helminths have been reported to induce a variety of regulatory populations including Foxp3⁺ T_regs [112, 113], CD8⁺T_regs [114], and regulatory B cells [115]. In addition to the induction of regulatory populations, helminths also appear to augment the regulatory capacity of the resident Foxp3⁺ T_regs population [112, 113].

Unlike symbiotic microbes that are beneficial to the host, helminth parasites have likely evolved tolerance-inducing mechanisms as a mode of immune evasion. Only live parasites are capable of expanding T_regs, suggesting that they are actively involved in inducing immunosuppression [116] and that helminth excretory/secretory (E/S) products mediate T_reg induction [117]. These findings argue against T_reg induction as merely a response to intestinal inflammatory signals in an effort to limit pathology. During helminth infection, there is evidence that T_reg recruitment and induction occurs. T_reg depletion prior to infection with Litomosoides sigmodontis has shown that resident Foxp3⁺ T_regs contribute to filarial worm survival [118]. However, iT_regs are generated at a high rate during Heligmosomoides polygyrus infection [117], a model in which T_regs are known to protect against immune pathology [119]. Recent studies have shown that H. polygyrus modulates the immune system by secreting molecules that mimic TGF-β signaling [117]. Indeed, H. polygyrus E/S products plus TCR stimulation in vitro are sufficient to induce Foxp3⁺ T_reg development in the absence of antigen presenting cells. Interestingly, H polygyrus infection during ovalbumin-mediated oral tolerance induction resulted in greatly augmented frequencies of Foxp3⁺ iT_regs specific for ovalbumin [117]. This finding is further evidence that tolerance initiated by H. polygyrus is ‘infectious’ to non-parasite-related antigens. Schistosoma mansoni soluble egg antigen (SEA) can also act directly on CD4⁺ T cells via TLR2 to induce the secretion of bioactive TGF-β which is required for their induction of iT_regs [120].

In addition to direct induction of T_regs by helminths, there are examples where DCs provide a critical link in mediating suppression. For instance, S. mansoni SEA requires DCs to mediate TGF-β-dependent Foxp3⁺ T_reg conversion and also elevates DC IL-10 production, potentially augmenting T_reg development and function [45, 121]. Thus, the direct effects of SEA on both the DC and CD4⁺ compartments may act synergistically to maximize T_reg conversion. Likewise, H. polygyrus E/S product can mediate T_reg induction through a DC-dependent mechanism. Antigen-pulsed bone marrow-derived DCs treated with H. polygyrus E/S product were deficient in their ability to up-regulate co-stimulatory molecules in vitro and initiate antigen-specific immune responses in vivo [122]. These treated DCs induced IL-10-producing Tr1-like cells in vitro. H. polygyrus has recently been shown to expand a tolerogenic DC population characterized by low CD11c expression [123]. These DCs could induce T_regs in vitro in the presence of TGF-β in an RA-dependent manner. Furthermore, despite depletion of CD11c^hi DCs in vivo, H. polygyrus infection could still effectively expand T_regs implicating a CD11c^lo population in tolerance induction.

Not only do tolerance-promoting microbes ensure their own survival by dampening immune responses, but they also promote tolerance to non-related antigens through the creation of a tolerogenic environment. Thus, the intestinal microbiota appear to contribute substantially to systemic peripheral tolerance. For instance, PSA from B. fragilis limits the development of EAE [124]. Likewise, helminth parasites can protect against immunopathology in autoimmune diabetes and asthma mouse models [125, 126]. Despite the clear expansion of tolerance by certain microbes, determining the contribution of infectious tolerance in these systems will be a challenge moving forward.

Conclusions

Dominant tolerance by regulatory T cell populations occurs through multiple mechanisms. This mechanistic diversity has likely evolved to match the plethora of distinct immune responses elicited to a multitude of different pathogens. Infectious tolerance represents an effective way for T_regs to establish a regulatory milieu to maintain immune balance. Immune tolerance may be maintained by strength in numbers, or through commissioning regulatory populations that can provide unique functional capabilities to ensure an effective immune response and prevent adverse immunopathology.

Future studies will be needed to fully understand the interrelationship between distinct T_reg populations. It will be particularly important to clarify these issues in vivo and to define their mechanism of induction. For example, the relative contribution of direct cytokine-mediated conversion of suppressive populations, as seen in vitro, versus conversion that requires a cellular intermediate, such as a DC, will be challenging but important to ascertain. With further advances in our understanding of T_reg development and mechanisms of suppression, it may be possible to design targeted immunotherapies to modulate the immune response during cancer, allergy, and autoimmunity.

Abbreviations

DC

Dendritic cells

EAE

Experimental autoimmune encephalomyelitis

E/S

Excretory/secretory products

IDO

Indoleamine 2,3-dioxygenase

IPEX

Immunodysregulation polyendocrinopathy enteropathy X-linked

iT_regs

Induced T_regs

PSA

Polysaccharide A

RA

Retinoic acid

SEA

Soluble egg antigen

T_regs

Regulatory T cells