|Home | About | Journals | Submit | Contact Us | Français|
Foxp3+ regulatory T cells (Tregs) contribute significantly to the maintenance of peripheral tolerance, but they ultimately fail in autoimmune diseases. The events that lead to Treg failure in controlling autoreactive effector T cells (Teffs) during autoimmunity are not completely understood. In this review, we discuss possible mechanisms for this subversion as they relate to type 1 diabetes (T1D) and multiple sclerosis (MS). Recent studies emphasize (i) the role of inflammatory cytokines, such as IL-6, in inhibiting or subverting Treg function, (ii) the issue of Treg plasticity, (iii) the possible resistance of autoimmune T cells to Treg-mediated control, and (iv) Treg-associated inhibitory cytokines TGFβ, IL-10 and IL-35 in facilitating Treg suppressive activity and promoting Treg generation. These recent advances place a large emphasis on the local tissue specific inflammatory environment as it relates to Treg function and disease development.
Autoimmunity ensues when central and/or peripheral tolerance barriers are overcome, thereby allowing the activation of self-reactive T cells, which induce tissue destruction. Most self-reactive T cells are deleted in the thymus, a process referred to as central tolerance. The few remaining self-reactive T cells that enter the periphery are controlled by peripheral tolerance mechanisms, in which Foxp3+ Tregs have emerged as the primary mediators . Autoimmune disease develops when this last tolerance barrier is compromised. It has been shown that many individuals with no autoimmune manifestations harbor self-reactive T cells . Thus, the continual battle between self-reactive T cells and suppressive Tregs is critical in determining whether autoimmunity commences. Many autoimmune diseases manifest themselves in an oscillating fashion, which is likely due to the aforementioned struggle between the regulatory and inflammatory arms of the immune system, suggesting that Tregs are still actively involved in regulating this self-reactive response. There are three mechanisms that might contribute to a breakdown in Treg control: (a) Treg numbers are reduced and/or Tregs are dysfunctional due to inherent deficiencies in autoimmune susceptible individuals, (b) Treg suppressive function is inhibited, diverted or converted by the chronic inflammation that occurs in autoimmunity, and/or (c) self-reactive effector T cells (Teff) become unusually aggressive and are refractory to regulation by otherwise functional Tregs because they either overwhelm regulatory control or express molecules that render them resistant. Here we will discuss how these possibilities might contribute to this loss of Treg control, illustrated by two key autoimmune diseases Type 1 Diabetes (T1D) and Multiple Sclerosis (MS), with occasional reference to other autoimmune or inflammatory diseases.
Even though multiple mechanisms of Treg-mediated suppression have been identified, the relative importance of each mechanism in vivo remains unclear . Current dogma, based mostly on in vitro experiments, suggests that Treg function is contact dependent. However, a recent study illustrates that this may be inaccurate as Tregs that are optimally stimulated by contact with Teff can mediate potent third party suppression via soluble factors in vitro . These data, together with significant support from in vivo observations, suggest that inhibitory cytokines such as TGFβ, IL-10, and IL-35 are major contributors to Treg function [3,5]. Consequently, this review will also focus on the contribution of these three inhibitory cytokines in mediating (or failing to mediate) Treg function in autoimmune disease.
Under normal conditions, Tregs effectively inhibit excessive inflammation and autoimmune manifestations . However, in autoimmune diseases Tregs fail to control the initial inflammatory insult and subsequent disease progression due to cell-intrinsic and/or cell-extrinsic factors. Studies performed in a variety of autoimmune models suggest that Tregs are still actively and continuously involved in inflammatory regulation after disease onset [7,8], although this needs to be more definitively assessed. When naturally occurring Tregs are absent in Foxp3-deficient islet antigen-specific BDC2.5 TCR transgenic NOD mice, disease onset is significantly earlier . Interestingly, the rate of islet infiltration in BDC2.5 mice is similar in the presence or absence of Tregs, however, the infiltrating cells cause immediate damage in their absence . In experimental autoimmune encephalomyelitis (EAE), an induced murine model of MS, Foxp3+ Tregs accumulate in the central nervous system (CNS) as the disease progresses  and in some instances aid in natural recovery from EAE . Depletion of CD25+ T cells prior to disease induction reduces recovery rate while their depletion after recovery obliterates their protection against secondary autoimmune disease induction . Although Tregs have some control over autoimmune inflammation, there may be several reasons why they ultimately fail to prevent disease progression.
A possible explanation for Treg failure is due to a reduction in their number and/or function in autoimmune predisposed individuals. This may not necessarily manifest systemically, and may be restricted to the site of autoimmune insult. Some studies have reported an age-related defect in NOD Treg function , while others were unable to detect any major deficiency . In both studies Tregs were identified and isolated based on CD25 expression, which may have resulted in contamination with activated Teff thereby altering data interpretation. A more recent study took advantage of NOD Foxp3-GFP reporter mice to cleanly separate the two populations and detected no intrinsic defects in NOD Treg numbers or suppressive capabilities . There is some evidence that Treg function in human patients with an early onset of diabetes might be compromised . However, human studies have to rely on CD25 as a marker for Treg separation, and thus are not entirely conclusive. More recent studies were able to exclude contaminating Teff from Treg analysis based on CD127 (IL7Rα) expression that has restricted expression on activated T cells and is absent from Tregs [15,16]. Although these studies are preliminary, at least in the early diagnosed T1D patients Treg numbers appear to be decreased .
Another possibility for Treg failure is that the chronic inflammation and consequent molecular milieu at the site of autoimmune destruction could have a negative effect on Treg function. This could be manifested as a reduction in Treg inhibitory function, apoptosis, or conversion into a pro-inflammatory T cell lineage. Recently, the fate of Foxp3+ Tregs over the course of diabetes development in NOD mice was followed. Treg function at the site of inflammation was found to be diminished due to decreased levels of IL-2, a critical Treg growth factor . These tissue-infiltrating Tregs were more prone to apoptosis and thus failed to control autoimmune responses. This study underlines the dependence of Treg function on local IL-2 production and the inflammatory milieu.
Treg stability and plasticity during autoimmune response has recently become a hot issue, especially since administration of in vitro differentiated Tregs is being considered for human therapy. Recent studies addressing the stability of Foxp3+ T cells have shown that both lymphopenic and inflammatory environments can result in the loss of Foxp3 expression in Tregs [18–20]. The apparent plasticity of Treg lineage might result in the instability of this population during chronic autoinflammatory responses and lead to loss of Treg function in an autoimmune environment.
Clearly, Tregs possess other modes of action in addition to the three inhibitory cytokines discussed here [3,21]. There is still some controversy over the relative importance of soluble factors versus cell contact-dependent inhibition in mediating Treg function. Substantial in vivo data, together with a recent study indicate that cytokines contribute significantly to Treg function when cell contact is absent . However, to unequivocally show the importance of cytokines in mediating Treg function one would have to assess inhibition in their absence. The extent of redundancy within these three inhibitory cytokines remains unclear. At first glance, it seems that there is significant overlap between the functions of the three Treg inhibitory cytokines (Figure 1). Both IL-10 and TGFβ can inhibit Th1 type responses, both are able to convert T cells into a regulatory population, albeit with quite distinct phenotypes, and all three cytokines seem to be play a role in the gut homeostasis. However, recent studies suggest that different transcriptional ‘programs’ are utilized by Tregs depending on the Th subset they are attempting to suppress which may likewise lead to the utilization of distinct suppressive functions [22,23]. Interestingly, Tregs appear to utilize similar transcription factors as the T helper cells they are suppressing, raising the possibility of a co-evolutionary progression. IRF4 deficiency in Tregs resulted in selective ablation of IL-10 transcription, partial decrease in EBI3 expression, and no affect on TGFβ. These findings suggest that Tregs may respond with a distinct inhibitory cytokine profile depending on the inflammatory milieu . Thus, it is possible that further examples of Treg functional plasticity may emerge. Therefore, it is possible that although all three cytokines are utilized by Tregs in the same tissue they may control different cellular populations or that different cytokines may be more important in different organs, in a manner that cannot be fully gleaned from the studies conducted thus far. Given the brevity of cellular targets and unique microenvironments within which Tregs have to operate, it would not be surprising if there is limited functional redundancy for these cytokines, as has been suggested for IL-10 . It will also be important to assess the effect of losing one inhibitory mechanism on the remaining functional landscape as it is conceivable that Tregs may make adjustments to compensate for such a loss given the functional plasticity already described.
A third possibility is that highly activated self-reactive Teffs become resistant to Treg-mediated control. Studies performed in both spontaneous T1D and induced EAE models report normal Treg function during disease, but show that the Teffs are hyper activated and do not capitulate to suppression by Tregs [7,9,13,25,26]. NOD Treg function, as measured by in vitro assays, was reported to be fairly comparable to B6 Tregs. In contrast, NOD Teffs exhibited an increased ability to proliferate and were unresponsive to Treg-mediated suppression . Similarly, Teffs from human T1D patients are more refractory to inhibition by Tregs than Teffs from healthy controls .
Mounting evidence shows that inhibitory actions of Treg-associated cytokines can be subverted in highly inflammatory environments. In particular, IL-6 has been shown to have a negative effect on Treg function , and in the CNS of EAE mice a typically inhibitory cytokine TGFβ synergizes with IL-6 to induce pro-inflammatory Th17 cells [29,30]. Moreover, when directly compared to Th1 and Th2 cells, Th17 cells were more likely to overcome the inhibition by Tregs . These data have raised the possibility that Th17 cells are more resistant to regulatory control than other T helper subsets. The molecular basis for this resistance remains obscure.
Thus it is possible that Treg function is normal during autoimmune disease, but the chronic inflammatory environment at the tissue site of autoimmune damage counteracts and overwhelms the suppressive Treg activity. While the issue of Treg fitness during autoimmunity remains a hot topic of discussion, it is possible that multiple pathways lead to the breakdown of peripheral tolerance. It is clear, however, that Tregs function has to be considered and assessed in a context of the inflammatory environment where they operate rather than in the periphery.
As discussed in the previous section, the local inflammatory milieu at the site of autoimmune destruction can have a significant effect on Treg function. Cytokines secreted by Tregs, Teffs and other cellular infiltrates contribute to the generation of this environment. Inhibitory cytokines, such as TGFβ and IL-10, can directly inhibit cells that participate in this autoimmune destruction and/or can mediate the generation of induced T regulatory cells (iTregs) . More recently IL-35 has also been shown to directly inhibit T cell proliferation [4,32], although its ability to affect other cell populations or its capacity to mediate/influence the induction of iTregs has yet to be established.
The in vivo role of Treg-derived cytokines in modulating autoimmune diseases has been clearly established. Mice deficient in TGFβ succumb to spontaneous autoimmune disease at 4–5 weeks of age [33,34], while the IL-10 knock out animals are more susceptible to induced autoimmune disease . In TGFβ deficient mice, the dysregulated inflammatory response is alleviated after the depletion of CD4+ and CD8+ T cells [36,37]. Although, Tregs are not the only source of inhibitory cytokines, it has become more evident that Treg secreted cytokines contribute significantly to Treg function in vivo.
TGFβ has been implicated in the function of Tregs in T1D and inflammatory bowel disease (IBD) mouse models [38,39]. Depletion of TGFβ with neutralizing antibodies results in a decrease in both mouse and human Treg function [39–41]. T cell transfer and bone marrow chimeric studies revealed that TGFβ has both cell-intrinsic and cell-extrinsic regulatory effects [42,43]. These data show that TGFβ helps to preserve peripheral tolerance by induction of iTregs and maintenance of nTregs in the periphery.
IL-10 is a homodimeric cytokine with a wide range of inhibitory actions. IL-10 is produced by a variety of cells, including monocytes and T cells, and can exert its effects on both myeloid and lymphoid cells . Its regulatory activity is mediated largely through its effects on APCs . Interestingly, IL-10 and IL-10 receptor-deficient mice, unlike their TGFβ counterparts, do not develop spontaneous autoimmune symptoms which only becomes apparent following inflammatory insult. In the absence of IL-10, mice develop colitis in response to some intestinal microorganisms [45,46]. In other murine autoimmune models, such as EAE, mice lacking IL-10 exhibit an exacerbated form of the disease . An increase in IL-10 production by Foxp3+ Tregs is associated with the recovery phase of EAE , and transfer of in vitro generated iTregs or exogenously purified natural Tregs was shown to prevent the induction of EAE through the production of IL-10 . IL-10 plays a significant role in intestinal homeostasis. Although, IL-10-deficient Tregs are not devoid of regulatory activity, their function is particularly affected in controlling already established gut inflammation , and mice with Treg-specific deletion of IL-10 are susceptible to colitis induced by commensal flora . Similarly to TGFβ, IL-10 can also mediate the generation of an induced regulatory T population, Tr1, that is characterized by secretion of IL-10 .
IL-35 is a recently discovered heterodimeric cytokine that is produced by Foxp3+ Tregs and contributes to their suppressive function . IL-35 is composed of two chains, IL-12α (p35) and EBI3 that it shares with two other heterodimeric cytokines IL-12 and IL-27, respectively. IL-35 is required for maximal regulatory function in vivo as Tregs deficient in either chain are unable to control homeostatic T cell expansion and IBD . Given that IL-35 emerged more recently, we still have a limited understanding of its biological activity which has also been complicated by the lack of appropriate tools for its functional dissection. Recombinant IL-35 has proven particularly challenging to generate and purify, compared with IL-12 and IL-27. It is tempting to speculate that the apparent poor stability of IL-35 might underlie important physiological features of this cytokine, such as limited potency over short-range. Alternatively, DCs that secrete IL-12 or IL-27 may be precluded from generating IL-35, due to preferential pairing of IL-12 and IL-27.
It is clearly important to determine the bioactivity of IL-35. Whether IL-35 can inhibit all T helper subsets and other cellular populations, such as B cells, macrophages or DCs, or has more selective targets remains to be determined. Significant insight is likely to be gained from the determination of the IL-35 receptor and its expression pattern. It is tempting to speculate that since IL-35 represents a novel pairing of the IL-12 cytokine chains, its receptor will also be a new combination of the known family receptor chains . Alternatively, the receptor might be composed of novel subunits. If it is the former, a key question will be how cells can translate similar signaling pathways into completely different downstream functional consequences, i.e. stimulatory signaling induced by IL-12 and IL-27 vs inhibitory signal delivered by IL-35. Interestingly, precedence does exist for a similar signaling cascade mediating quite different functional outcomes, as exemplified by the IL-6 and IL-10 receptor which both use STAT3 . Lastly, an open question is whether IL-35 can mediate the generation of an induced regulatory T cell population. This is clearly a feature of the other two inhibitory cytokines IL-10 and TGFβ and thus it remains plausible that IL-35 might have a similar capacity. TGFβ-inuced iTregs (Th3) and IL-10-induced Tr1 clearly have very distinct transcriptional and functional profiles [3,52] and so it is possible that any IL-35-induced regulatory populations may also be quite distinct.
It is clear that Foxp3+ Tregs play a major role in most autoimmune diseases. What remains unknown is the relative contribution of natural, thymus-derived Tregs (nTregs) versus iTregs in mediating control of autoimmunity. Furthermore, recent studies suggest that the importance of iTregs in controlling autoimmunity may vary depending on the site of tissue inflammation [4,32].
Recent studies have shown that Tregs and Treg-derived cytokines have long lasting tolerance effects in vivo [4,32]. Even though diabetes onset can be prevented by Treg transfer into NOD mice, the endogenous Foxp3+ Treg population dominates at the site of the original autoimmune response in the pancreas while the transferred population is significantly reduced . In a mouse model of IBD, while TGFβ is essential for downregulating the autoimmune response, Treg-derived TGFβ was not critical for protection [39,42]. This suggests that the main function of TGFβ may be to mediate infectious tolerance (ie. generation of iTregs) which may not necessarily be Treg-derived . It is possible that in order to achieve infectious tolerance several simultaneous signals are required in addition to TGFβ. These can be other inhibitory cytokines, such as IL-10, or cell surface molecules that require cell contact for ligation .
Recent findings suggest that different organs or tissues are more suitable or permissive to Treg differentiation due to their cellular and/or molecular composition. Mucosal CD103+ dendritic cells are exceptionally adept at inducing Tregs through production of TGFβ and the Vitamin A metabolite retinoic acid (RA) [55,56]. RA counteracts the effects of inflammatory cytokines and aides in skewing T cells towards a regulatory phenotype [57–59]. CD8+CD205+ dendritic cells in the spleen can also mediate iTreg development in part through production of TGFβ . The CNS, however, does not seem to favor the generation of Foxp3+ iTregs . This is in part due to Teff-derived IL-6, which in combination with TGFβ induces Th17 cells rather than Foxp3+ iTregs . DC phenotype is initially induced by microbial interactions leading to the induction of either inflammatory or inhibitory/toleragenic DCs . It is conceivable that the complete Freund’s adjuvant used to induce EAE, which contains mycobacterial cell wall components, results in the generation of DCs that are suboptimal for iTreg generation. In support to this, mice immunized with myelin peptide in incomplete Freund’s adjuvant that lacks Mycobacterium tuberculosis exhibited a decrease in IL-6 production and an increase in antigen specific Foxp3+ T cells in the spleen. However, there was only a modest increase in iTregs and whether these were present in the CNS is unknown . Taken together, these studies suggest that the stimulus received by DCs may be more critical than the particular site or organ that is affected in determining iTreg development. Thus, under alternate stimulatory conditions the CNS might be an acceptable location for iTreg differentiation.
Recent advances have placed significant emphasis on dissecting Treg function in autoimmunity and assessing their therapeutic utility. The local inflammatory environment, consisting of specialized APCs, metabolic components, growth factors and inhibitory and/or inflammatory cytokines, can significantly alter the efficacy and generation of Tregs. Given the complexity of Treg function in vivo, several key questions remain before we fully understand molecular and cellular defects that occur in autoimmunity, as they pertain to Tregs, and how this knowledge might be utilized in a clinical setting.
DAAV is supported by the Juvenile Diabetes Research Foundation International (1-2004-141 [The Robert and Janice Compton Research Grant, In Honor of Elizabeth S. Compton] and 1-2006-847), the NIH (AI072239), the St Jude Cancer Center Support CORE grant (CA-21765) and the American Lebanese Syrian Associated Charities (ALSAC). MB is supported by a Juvenile Diabetes Research Foundation International post-doctoral fellowship (3-2009-594).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Papers of particular interest, published within the past two years, have been highlighted as:
• of special interest
•• of outstanding interest