The immune system evolved to protect the host against the attack of foreign, potentially pathogenic, microorganisms. It does so by recognizing antigens expressed by those microorganisms and mounting an immune response against all cells expressing them, with the ultimate aim of their elimination. Various mechanisms have been reported to control and regulate the immune system to prevent or minimize reactivity to self-antigens or an overexuberant response to a pathogen, both of which can result in damage to the host. Deletion of autoreactive cells during T- and B-cell development allows the immune system to be tolerant of most self-antigens. Peripheral tolerance to self was suggested several years ago to result from the induction of anergy in peripheral self-reactive lymphocytes1. More recently, however, it has become clear that avoidance of damage to the host is also achieved by active suppression mediated by regulatory T (Treg) cell populations2,3,4,5. We discuss here the varied mechanisms used by Treg cells to suppress the immune system.
Treg cells may be defined as CD4+ T cells that inhibit immunopathology or autoimmune disease in vivo. Specifically, Treg cells include those able to suppress naive T-cell proliferation in vitro and to control CD4+ or CD8+ T-cell numbers in vivo, in lymphopenic hosts. Two major Treg populations have been described so far, which we will designate as naturally occurring Treg and interleukin (IL)-10 secreting Treg.
Originally CD4+CD45RBlow cells (which comprise the compartment of activated CD4+ cells found in normal mice) were shown to be able to control the colitis induced by the transfer of CD4+CD45RBhigh (naive) cells6. It was later found that control of colitis was achieved by a subset of the CD4+CD45RBlow cells that express CD25, the α-chain of the IL-2 receptor4. We refer to these CD4+CD25+ T cells as naturally occurring Treg. They were identified due to their ability to control various manifestations of autoimmunity secondary to neonatal thymectomy2,3. Within the CD4+CD45RBlow population, some CD25− T cells also have regulatory activity7, but little work has been published on the further characterization of this population.
More recently, Treg cells producing IL-10 (IL-10 Treg) and secreting transforming growth factor (TGF)-β were described8,9,10,11,12. These cells are derived in culture and also express CD25, due to their activation in vitro. IL-10 Treg cells were shown to have regulatory activity in systems classically used to study the naturally occurring Treg populations. In particular, this novel type of regulatory T cell also inhibits naive T-cell proliferation in vitro, suppresses experimentally induced autoimmune disease and controls CD4+ and CD8+ T-cell numbers in vivo in lymphopenic hosts8,9,10,11,12. Because these IL-10 Treg were derived via different protocols, it is unclear at this stage whether, or how they are related to each other, or to the naturally occurring CD4+CD25+ Treg cells with respect to their development and function in the inhibition of the immune response at various levels. It should also be kept in mind that, although Treg cells clearly have an important role in the regulation of immune responses and the inhibition of immunopathology, T-helper type 1 (TH1) and type 2 (TH2) cells, by their production of specific cytokines, can also regulate immune responses (Table 1).
Suppressor T cells were first proposed to exist in the mouse in the early 1970s (ref. 13) and were thought to cause suppression by secreting antigen-specific suppressor factors. Subsequently, human regulatory T cells were also described that caused suppression by means of antigen nonspecific mechanisms14. In those early days the experimental systems used were poorly characterized and, due to the paucity of cell surface markers, lymphocyte subpopulations were difficult to isolate consistently. The field eventually floundered because these cells could not be grown reliably and attempts to clone the putative suppressor factors failed. In addition the major histocompatibility (MHC) antigen thought to be their restriction element, I-J, was not found as expected in the MHC locus, and many suppressor T-cell clones were also found to not express a T-cell receptor (TCR)15. The concept of immune regulation by specialized T cells remained, however, such that when better cellular and molecular tools became available the field was reborn. Under a new name, supported by more solid experimental systems and admittedly better data, suppressor T cells are back in business16. Today's regulatory T cells appear to require antigen-specific activation to act, but function through antigen-independent mechanisms.
Naturally occurring CD4+CD25+ Treg cells
The CD4+CD25+ T cell subset is currently the focus of intensive research. These cells represent 5–10% of the CD4+ T lymphocytes in healthy adult mice and humans and are thought to perform a specialized role in controlling both the innate and the adaptive immune system2,3,17. Although easily identified and isolated from unmanipulated mice and humans on the basis of CD25 expression, this chain of the IL-2R is also expressed on activated T cells2,3,17. Thus far, no characteristic stable surface marker has been ascribed to Treg cells. Additional markers expressed by these cells include cytotoxic T-lymphocyte–associated protein 4 (CTLA-4)18,19 and glucocorticoid-induced tumor necrosis factor receptor20,21, which were initially implicated in the mechanism of Treg action. However, both of these molecules are also expressed by nonregulatory T cells, after activation. Secretion of TGF-β and IL-10 is also a common feature of Treg, and the activity of these two cytokines is involved in their ability to control some, but not all, immune responses in vivo22. Again, neither TGF-β nor IL-10 secretion is a unique feature of Treg cells4.
The forkhead/winged helix transcription factor Foxp3 was shown to be specifically expressed by CD25+ Treg cells, as well as by CD25− T cells with regulatory activity23,24,25. This transcription factor is thought to program the development and function of this subset and so far is the most unambiguous marker available to identify naturally occurring Treg cells. Loss-of-function mutations in this gene, in both humans and mice, caused the absence of Treg cells (but, importantly, not of other CD25+ activated T cells) and a prominent phenotype that includes autoimmune endocrinopathy, early onset type 1 diabetes, thyroiditis and, in some cases, severe atopy and food allergy26,27. In humans the disease can be corrected, or at least ameliorated, by bone marrow transplantation28. In mice, the hematopoietic component that reverts the disease, or at least prevents it, is the CD4+CD25+ Treg population25. Ectopic expression of Foxp3 as a transgene significantly delayed the onset of the proliferative disease observed in mice lacking the CTLA-4 gene24. Furthermore, forced expression of Foxp3 in nonregulatory CD4+ T cells caused their acquisition of properties known to be characteristic of Treg cells, including inhibition of colitis and gastritis, inhibition of the lymphoproliferative and wasting disease observed in the natural Foxp3 mutant mice, scurfy, and control of the size of the CD8+ T-cell compartment in vivo23,25. In addition Foxp3-transduced CD4+ T cells can inhibit naive T-cell proliferation in vitro23,25. Expression of Foxp3 is thus a reliable marker for naturally occurring Treg cells, although, as discussed below, not all populations displaying regulatory activity express this gene.
It is as yet unclear whether Foxp3 is induced exclusively at a specific period of development in the thymus, or whether it can be upregulated in the periphery. Induction of Foxp3 in antigen-specific peripheral T lymphocytes would open new therapeutic strategies in the treatment of immunopathologies and may provide a solution to the problem of transplantation tolerance. Some suggestions that Foxp3 can be upregulated on nonregulatory T cells have been reported29,30, but it is difficult to interpret whether this upregulation was on naturally occurring Treg cells that had lost the expression of CD25. Recently it was shown that CD4+CD25− cells from TCR-transgenic mice on a recombination of activating gene (RAG-2)-deficient background, which do not contain thymic-derived regulatory T cells, develop into CD4+CD25+ regulatory T cells in the periphery and upregulate Foxp3 expression upon encounter with peptide antigen by means of osmotic pumps31. However, whether Foxp3 is the marker that uniquely distinguishes all Treg populations still remains unclear.
CD4+CD25+ Treg cells protect lymphopenic mice from inflammatory bowel disease4 and diabetes32,33, and can also prevent transplant rejection34,35. Conversely, a role damaging to the host is their ability to inhibit the immune response to tumors36,37. IL-10 production by Treg cells contained within the CD4+CD45RBlow compartment is essential to protect immunocompromised mice from colitis38 and allograft rejection34,35 initiated by CD4+ T cells contained within the CD45RBhigh compartment. IL-10 and TGF-β have been shown to be important mediators of Treg-mediated suppression of not only colitis22,38, but also autoimmune or allergic pathologies8,11,39,40,41,42. CD4+CD25+ Treg cells inhibit the proliferation and expansion of naive T cells both in vitro and in vivo3,7. However, the in vivo suppression of proliferation7,10, as already mentioned above, and wasting disease7 is IL-10 dependent, in contrast to the inhibition of gastritis22 and of naive T-cell proliferation in vitro, which are IL-10 independent and are thought to require cell-cell contact3,10,43. As yet the mechanisms for this cell-contact–mediated suppression seem to be undefined3, but it should be mentioned that membrane-bound TGF-β, found on the surface of activated CD4+CD25+ Treg cells, has been implicated as a mechanism by which these cells may suppress T-cell proliferation44.
IL-10–producing Treg cells
Regimens of antigen administration known to generate anergy/tolerance in vivo also induce the appearance of IL-10–producing T cells10,40,45,46,47. Sundstedt et al.10 have demonstrated that IL-10–producing Treg cells can be induced by repeated intranasal administration of an antigenic peptide of myelin basic protein (MBP). These regulatory cells inhibit the proliferation of naive MBP-specific T cells both in vitro and in vivo10. Interestingly, and in keeping with previous findings7,34,38 using CD4+CD25+ Treg cells, IL-10 mediated the suppression of naive T cells by the IL-10 Treg in vivo10, but not in vitro3,10. Moreover, naive CD4+ T cells can be made to generate in vitro a homogeneous population of cells secreting IL-10 (ref. 11), representative of those in vivo-derived populations described previously. These IL-10 Treg populations block experimental autoimmune encephalomyelitis, and their development and function in vivo are IL-10-dependent. However, the inhibition of in vitro T-cell proliferation by these IL-10–Treg cells is not43. Thus, it seems that both types of Treg cells use different mechanisms to regulate proliferation in vitro or in vivo and to control in vivo immune responses and/or pathologies, which may be accompanied by different levels of inflammation22,38.
In vitro assays used to detect inhibition of naive T-cell proliferation by Treg cells may not reveal a role for IL-10 as a result of the duration of the assay and the type of antigen-presenting cell (APC) used. In addition, the extent of APC maturation/activation that occurs in vivo may not be reflected in in vitro systems. This was clearly observed by Pasare and Medzhitov, who showed that activation, through Toll-like receptor (TLR)-4 or TLR-9, of dendritic cells derived from bone marrow overcomes the inhibition of naive T-cell proliferation mediated by CD4+CD25+ Treg cells. This antisuppressive action of activated dendritic cells was not fully dissected, but is in part due to the production of IL-6 (ref. 48). As the major effects of IL-10 are to inhibit APC function, including the production of proinflammatory cytokines49, it is likely that the function of Treg cells is dependent on IL-10 only when cells of the innate immune system are involved; indeed, Treg cells inhibit innate immunity17. In addition, although CD4+CD25+ Treg cells can be induced to produce IL-10 and can inhibit colitis, it is clear that IL-10 Treg, TH2 or TH1 cells and APC also produce IL-10 and thus can control the magnitude of an immune response and limit immunopathology49.
IL-10 Treg and CD4+CD25+ Treg cells are independent
Yet to be defined is the relationship between the two types of regulatory T cells: naturally occurring CD4+CD25+ Treg cells2,3,4, which, depending on the system, may or may not require the suppressive action of IL-10, and IL-10–producing Treg cells, which can be developed under different regimens of antigenic stimulation, both in vitro8,11,50 and in vivo10,40,45,46. After transfer into lymphopenic hosts, Foxp3-transduced CD4+CD25− T cells expressed enhanced amounts of IL-10 mRNA, comparable to those of naturally occurring CD4+CD25+ Treg cells and higher than the amounts found in untransduced cells25. Based on these findings it was suggested that Foxp3 directly upregulates IL-10 production, but it is likely that the cells require additional signals in vivo (besides the expression of Foxp3) for induction of the IL-10 gene. These studies did not support a direct relationship between Foxp3 and IL-10 expression. It has been demonstrated that IL-10 Treg cells, generated through in vitro or in vivo strategies, do not express the transcription factor Foxp3 (ref. 43), which distinguishes them from the classic CD4+CD25+ Treg cells. As they inhibit naive T cell proliferation, this shows that expression of Foxp3 is not always necessary for suppressor function. Furthermore IL-10 Treg cells can be obtained in vitro using antigen-specific TCR transgenic T cells that do not contain naturally occurring (that is, Foxp3-expressing) Treg cells11. Despite their lack of Foxp3 expression, however, homogeneous populations of IL-10–Treg cells inhibited the proliferation of naive CD4+CD25− T cells in vitro with an efficiency similar to that of CD4+CD25+ Treg cells43. It is notable that enhanced IL-10 expression in IL-10 Treg cells was always mirrored by a significant decrease in IL-2 expression and that IL-2 expression is also absent from Foxp3-expressing CD4+CD25+ Treg cells24,43. Foxp3 has been suggested to act as a negative modulator of IL-2 transcription51, but clearly this can be achieved in IL-10 Treg cells by other mechanisms, probably dependent on antigenic stimulation in the presence of immunosuppressive drugs or secondary to peptide antigen regimens leading to anergy induction10,11,47,52,53.
What is the mechanism of action of Treg cells?
What seems to be a prerequisite of the ability of Treg cells to inhibit naive T-cell proliferation in vitro is a lack of IL-2 (and probably IL-4) production, which would fit with the hypothesis that anergic T cells may regulate naive T-cell proliferation, and thus are active “suppressors”52, as opposed to working passively. The mechanism whereby Treg cells bring about the cell-contact–dependent inhibition of naive T cells is as yet unclear, although in some cases membrane-bound TGF-β may be involved in suppression mediated by naturally occurring CD4+CD25+ Treg cells44. It is of interest that, although naturally occurring Treg cells are anergic, they can expand in vivo7,54,55.
Similar to CD4+CD25+ Treg cells, the regulatory capacity of IL-10 Treg cells to inhibit naive CD4+CD25− T-cell proliferation in vitro is independent of their intrinsic production of IL-10 and was overcome by exogenous IL-2 (refs. 3,10,43). We propose that the inhibition of naive T-cell proliferation in this in vitro system represents the first layer of regulation, before the innate immune response is activated, when IL-10 then becomes involved in the mechanism of suppression3. This layer of regulation, which is independent of IL-10 expression, may have as its in vivo equivalent the competition-dependent regulatory activity that can be observed in lymphopenic mice56, perhaps at the level of competition for APC or for survival factors. Inhibition of gastritis mediated by naturally occurring (Foxp3-expressing) Treg cells22 is also IL-10 independent, perhaps reflecting low levels of inflammation. In our view, higher layers of regulation of immune responses exist, which require production of IL-10 and/or TGF-β (Fig. 1). Control of T-cell expansion in vivo by CD4+CD25+ Treg is dependent on IL-10 (ref. 7), and the higher inflammation associated with activation of the innate immune response may demand that the regulatory system be notched up. That is why in many in vivo systems, as discussed above, where strong inflammation—often induced directly by infectious agents—is the norm, both CD4+CD25+ Treg and IL-10 Treg cells inhibit inflammation through IL-10– or TGF-β–dependent mechanisms. This requirement for stronger suppression may also be related to differences in systemic versus local inflammation.
Treg and other T cells in infection
It has been shown that CD4+CD25+ Treg cells can regulate the immune response to the pathogen Leishmania major, in part through the suppressive action of IL-10 (ref. 57). It had been previously demonstrated that a TH2 response, observed for example in BALB/c mice infected with L. major, was completely incapable of protecting the host. In this system, IL-4 prevented L. major clearance by TH1 cells producing interferon-γ, which have been shown to be required for the eradication of this pathogen58 (Table 1). Furthermore, regulatory cells within the CD45RBlo compartment were shown to inhibit the clearance of L. major by CD4+CD45RBhi T cells through the action of IL-4, but to inhibit wasting disease in the same model via the action of TGF-β6,59 and IL-10 (ref. 38), independently of IL-4. Thus, while TH2 cells produce IL-10 (they are distinguished from Treg because TH2 cells produce in addition IL-4, IL-5 and IL-13) and may contribute to suppression of responses to L. major, IL-4 production by TH2 cells can also mediate suppression in certain models of this disease6,58. Anergic IL-10–producing T cells obtained from chronic Mycobacterium tuberculosis infected patients who are nonreactive to purified protein derivative (PPD) can be made reactive to PPD by the addition of antibodies to IL-10 to in vitro cultures. These cells have also been shown to act as Treg cells and inhibit the proliferation of allo-specific CD4+ T cells through the action of IL-1060. More recently, Helicobacter hepaticus-induced colitis was shown to be inhibited by CD4+ T cells (both CD25+ and CD25−) isolated from mice infected with this bacterium through an IL-10–dependent mechanism61. This work suggested in addition that regulatory T cells arise after H. hepaticus infection and require activation by a specific antigen to manifest their function.
IL-10 was originally described as a cytokine produced by TH2 cells49. It has since become clear that this suppressive cytokine is produced by numerous cell types, including other CD4+ and CD8+ T cells, macrophages, B cells and dendritic cells49,62. IL-10 suppresses the production by dendritic cells and macrophages of proinflammatory cytokines such as IL-12 required for TH1 development49, and yet IL-10 is produced, albeit in varying levels, by TH1 cells62. The relative levels of IL-10 and interferon-γ produced by TH1 cells may be an important determinant of the balance between clearance and persistent infection with certain pathogens49,58,62. In this regard, IL-10 may also have the important role of feedback regulator to control the pathology associated with an overexuberant, albeit efficacious, TH1 response63. Thus, the source of IL-10 during immune responses to pathogens or allergens may be TH1 or TH2 cells, respectively49,58,62, or regulatory T cells41,57,60,64. Although regulatory T cells were originally implicated in the regulation of autoimmune and/or inflammatory pathologies2,4,15, it is clear that they may also have a role in regulating the response to infectious pathogens and/or allergens to minimize immune pathology4,6,26,27,41,49,57,60,64.
It is of note, however, that although TH1 and TH2 cells can clearly function as “regulatory” cells by their production of IL-10 (and of IL-4 in the case of TH2 cells) and by their ability to reciprocally regulate each other58, they also mediate effector functions. Such effector functions have not as yet been attributed to the two types of regulatory T cells discussed here, as they seem to have no function other than regulation (that is, suppression) of immune responses. However, active roles have been defined for TGF-β and IL-10 in effector immune responses49.
Clinical applications of Treg cells
The use of Treg cells in the treatment of immunopathologies such as colitis, rheumatoid arthritis or multiple sclerosis is appealing and, if possible, would represent a major advance in the treatment of such diseases. The development of antigen-specific Treg cells would avoid generalized immunosuppression with obvious benefits, for example in the fields of organ transplantation or allergy/asthma, where the antigen is often known. The goal should be, rather than deriving Treg cells in vitro and transplanting those cells into patients, to directly induce their development in vivo. Although IL-10–Treg cells may be obtained in vivo by repeated antigen administration10, there is the danger that proinflammatory cytokines may also be induced and the risk of pathologic side effects, as can be occasionally observed in the treatment of multiple sclerosis with peptide therapy. The combination of antigen with immunosuppressive drugs, based on the studies mentioned above11, may allow the development in vivo of IL-10–Treg cells in the absence of (proinflammatory) cytokine-dependent side effects, hopefully leading to long-term regulation. Conversely, in cancer or chronic infectious diseases, neutralization of IL-10 or inhibition or deletion of Treg cells might enhance the eradication of the tumor or pathogen36,37,60.
A major hurdle in the application of Treg cells in the treatment of inflammatory and autoimmune diseases, however, will be to selectively target antigen-specific Treg cells and thereby disarm the pathogenic T cells, while leaving undisturbed T cells directed against infectious pathogens and/or tumors. Similarly, treatment of chronic infectious diseases with antagonists of Treg cells or their products must be achieved without the activation of autoreactive T cells.
Several questions remain to be answered with respect to the development and function of Treg cells. It still remains to be clarified when each type of Treg cell develops and when each has a role in the regulation of immune pathology. Additionally, are the naturally occurring CD4+CD25+ Treg cells that develop in the thymus and are Foxp3-dependent mainly important for responses to self-antigens and/or antigens encountered in the thymus, or can they also regulate inflammatory pathologies associated with gut flora4,17, alloantigens34,35, tumors36,37 and infectious diseases57,60? If so, is the latter achieved by cross-reactivity of CD4+CD25+ Treg cells between self- and microbial antigens? The work of Kullberg et al.61 suggests that this is not the case, because Treg cells from uninfected mice do not protect from H. hepaticus-induced colitis. Also, how do naturally occurring CD4+CD25+ Treg cells expressing Foxp3 compare with the antigen-driven IL-10–producing Treg cells obtained by in vitro and in vivo regimens of antigenic stimulation? Are these IL-10–Treg cells long-lived, as has been shown for CD4+CD25+ Treg cells53,65? CD4+CD25+ Treg cells and possibly other Treg cells producing IL-10 can regulate immune responses to pathogens and/or effector T cells10,11,57,60. Regulation in vivo has been shown to be mediated by relatively small numbers of antigen-specific Treg cells32. Whether these IL-10–Treg cells are of the same “lineage” as the naturally occurring Foxp3-dependent CD4+CD25+ Treg cells, which prevent lymphoproliferative disorders and endocrinopathies such as diabetes in mice and humans, is as yet unclear.
It is likely that the functional mechanisms involved in regulating immune responses to self-antigens are similar to those regulating immune responses to pathogens and other antigens. Different Treg cells seem to develop by means of different pathways, however, and the molecular mechanisms involved in regulation may depend on the level of inflammation accompanying an immune response. The ability to modulate Treg cell function in disease is promising but as yet unachieved. This will clearly be the focus of therapeutic strategies over the course of the next few years.