CD8+CD28– and CD3+CD4–CD8– regulatory T cells
The concept of CD8+ suppressor T cells was originally proposed in the 1970s, and fell into disrepute in the 1990s. The re-appreciation of the existence of regulatory cells in vivo in the past years has led to the resurgence of CD8+ suppressor T cells in the field of transplantation tolerance. CD8+CD28– T cells that were capable of suppressing CD4+ T cells with allo- or xenospecifity could be generated by repetitive allo- or xenostimulation in vitro. The resulting CD8+CD28– T cells were able to induce the up-regulation of immunoglobulin-like transcript 3 (ILT3) and ILT4 on monocytes and DCs, rendering these APCs tolerogenic and unable to stimulate CD4+ T cells. Data from human cardiac transplant patients were also correlated with this finding (17).
CD3+CD4–CD8– (DN) T cells with regulatory properties could be isolated and cloned after pretransplant donor lymphocyte transfusion (18). The DN regulatory cells were able to prevent skin allograft rejection mediated by CD8+ T cells in an antigen-specific manner. The underlying mechanism of suppression was attributed to the ability of the DN regulatory cells to cause Fas-Fas l-mediated apoptosis of the effector CD8+ T cells. The intricate immune regulatory capacity of CD3+ DN T cells has been further highlighted by a recent study showing that CD3+ DN T cells from lymphoproliferative (lpr) mice were able to down-regulate allogeneic immune responses mediated by syngeneic CD4+ and CD8+ T cells from wild-type mice both in vitro and in vivo (47). It will be interesting to see whether CD3+ DN T cells from peripheral blood in normal rodents and in humans possess regulatory properties.
Anergic CD4+ T cells as regulatory cells
Probably the earliest example implying the potential for anergic T cells to act as suppressor cells came from a SAg in vivo model in which Mls-1b mice were injected with Mls-1a cells. The Lyt-1+ cells (probably CD4+) acted as efficient suppressor cells that were antigen-nonspecific in their effector function (48). More recently we explored the regulatory properties of anergic T cells derived from human CD4+ T-cell clones in vitro (49). These cells acted as potent suppressor cells in an antigen-specific fashion. Furthermore, they were able to effect linked suppression if the APCs expressed the antigens for which the anergic and the target T cells were specific (50). The suppression was cell-contact dependent, and could not be reversed by the addition of neutralizing antibodies against regulatory cytokines, IL-4, IL-10 and TGF-β. Subsequently, we were able to demonstrate that the anergic T cells could prolong skin allograft survival in vivo, providing the first evidence that anergic T cells were capable of effecting immunoregulation in vivo (51). In our further studies we found that the anergic T cells were able to render DCs almost totally incapable of stimulating T-cell proliferation (52). This loss of immunogenicity by DCs was accompanied by reduced expression of MHC class II and B7 molecules. These data suggest that inhibition of DC maturation and function may be one of the ways in which anergic T cells exert their regulatory functions, although the precise mechanism is yet to be elucidated. Given the mode of action of the anergic T cells, these cells resemble a spontaneously occurring population of T cells that regulate autoimmune inflammation in the mouse, CD4+CD25+ cells.
Naturally occurring CD4+CD25+ regulatory T cells
The first series of experiments to demonstrate that CD4+CD25+ cells act as regulatory cells was reported in the mid-1990s in a mouse model of organ-specific autoimmune disease. When BALB/c athymic nude (nu/nu) mice were injected with CD4+ T cells depleted of CD25+ subset from BALB/c nu/+mice, all recipients spontaneously developed autoimmune diseases such as thyroiditis, gastritis, insulitis, sialoadenitis, adrenalitis, oophoritis, glomerulonephritis, and polyarthritis (5–7). Most importantly, reconstitution of CD4+CD25+ cells 10 days after transfer of CD4+CD25– cells could prevent these diseases. However, the first hint that CD4+CD25+ cells possess regulatory capacity was provided by Hall et al. several years earlier in a rat model involving the adoptive transfer of transplantation tolerance (8). Given that CD25 is also expressed on activated effector T cells, and the latter are unable to act as regulatory cells in most studies (53), it is obvious that CD25 is not an ideal cell-surface marker to identify CD4+ cells with regulatory properties. CD4+CD45RBlow cells in mice (54) and CD4+CD45RC– cells in rats (55) have also been identified as regulatory cells.
The data from Papiernik et al. have established that CD4+CD25+ regulatory cells naturally arise in the thymus and then migrate to the periphery (56). This is in line with the findings of Modigliani et al. that athymic mice grafted at birth with allogeneic thymic epithelium could induce life-long tolerance to skin grafts of donor haplotype (57). It appears that the generation of CD4+CD25+ cells is dependent on MHC class II-positive thymic cortical epithelium. Indeed, Bensinger et al. have shown that CD4+CD25+ regulatory cells were present in mice when thymic cortical epithelium expressed MHC class II molecules, but not in MHC class II-deficient mice (58). One attractive hypothesis was that the intrathymic positive selection of CD4+CD25+ cells requires TCRs with higher affinity for self-peptide-MHC class II complexes than those expressed by CD4+CD25– cells. Compelling evidence to support this assumption comes from Jordan et al. in an influenza hemagglutinin (HA) double-transgenic mouse model (59). They showed that almost 50% of HA-specific CD4+ cells were CD25+ regulatory cells in the double-transgenic mice. In contrast, when a TCR transgenic mouse with low affinity for HA antigen was used, CD4+CD25+ thymocytes were not detected in the double-transgenic mice, suggesting that intrathymic selection of CD4+ cells with high affinity for self-antigen leads to the generation of CD4+CD25+ regulatory cells.
Although CD4+CD25+ cells are developed in the thymus they exert their suppressive function in the periphery. The peripheral homeostasis of CD4+CD25+ cells appears to be dependent on IL-2 and constant stimulation by self-MHC-peptide complexes. An earlier study showed that in IL-2–/– knockout mice CD4+CD25+ regulatory cells were detected neither in the thymus nor in the periphery (56). Furthermore, transgene-encoded thymic expression of IL-2Rβ in IL-2Rβ–/– mice could prevent the autoimmune disease normally seen in these animals (60). IL-2 was also required for CD4+ regulatory T-cell function (61). CD28/B7 costimulation and TRANCE/RANK (TNF-related activation induced cytokine and receptor activator of NF-κB, respectively) may also play important roles in the peripheral maintenance of CD4+CD25+ cells. Indeed, CD28–/– or B7–/– knockout NOD mice manifested a profound defect in CD4+CD25+ regulatory cells and developed spontaneous diabetes (62). Blockade of TRANCE-RANK signaling also led to a decreased frequency of CD4+CD25+ regulatory cells in the periphery and rapid progression of diabetes (63). Another key aspect of these cells is their trafficking properties. Peripheral CD4+CD25+ regulatory cells are capable of trafficking to inflammatory sites to control pathological immune responses through chemoattractant CCL-4, as anti-CCL-4-treated mice accumulated autoantibodies (64). Chemokine receptor CCR7 may also be involved in the trafficking machinery of murine CD4+CD25+ cells (65). Human CD4+CD25+ cells specifically express chemokine receptors CCR4 and CCR8, and engagement with their ligands may be responsible for the recruitment of regulatory cells to sites of inflammation in man (66).
First described in mice (5–7), we and others have reported the existence of CD4+CD25+ regulatory cells in the peripheral blood, cord blood and in the thymus of humans (67–72). In vitro studies showed that CD4+CD25+ cells were anergic, as they failed to proliferate after a variety of TCR stimuli, except when exogenous IL-2 was present. They suppressed the proliferation of CD4+CD25– T cells by inhibiting their IL-2 transcription after activation. CD4+CD25+ cells also inhibited the response of CD8+ T cells (12,63,68,71,73,74). Addition of exogenous IL-2 was able to abrogate the suppression. The suppression seemed to require physical contact between CD4+CD25+ cells and the cells to be regulated. Interestingly, once activated CD4+CD25+ cells were capable of suppressing the response of CD4+ T cells activated through a distinct TCR via the same or different APCs (75). Although some clues exist concerning the mechanism underlying the suppressive functions of CD4+CD25+ cells, the precise mechanism of action remains to be defined.
Role of CTLA-4. Freshly isolated murine or human CD4+CD25+ cells, but not CD4+CD25– cells constitutively express CTLA-4, a negative regulator of T-cell activation. Some studies have implied that anti-CTLA-4 antibody or its Fab were able to abrogate the suppression of murine CD4+CD25+ cells in vitro (76) and in vivo in inflammatory bowel disease (77) and in transplantation tolerance (12, 78). CTLA-4 blockade, however, was unable to reverse the suppressive function of CD4+CD25+ cells from human peripheral blood. Nevertheless, anti-CTLA-4 antibody could partially inhibit the suppression of human CD4+CD25+ thymocytes (72). It seems likely that the high-affinity interaction between CTLA-4 and B7 molecules expressed on APCs prevents or diminishes the delivery of costimulatory signals by APCs to CD4+CD25– T cells. The precise role of CTLA-4 in the function of CD4+CD25+ regulatory cells requires further investigation.
Role of IL-10 and TGF-β. IL-10 and TGF-β are well-known immunosuppressive cytokines, although the molecular basis underlying their mode of action has yet to be determined. CD4+CD25+ cells seem to have more potential than CD4+CD25– cells to produce both cytokines after TCR activation (6). Studies in vitro have concluded that neutralizing anti-IL-10 antibody has no effect on the suppression by both murine and human CD4+CD25+ cells. Two lines of evidence have indicated that CD4+CD25+ cells express cell-surface TGF-β, and high concentrations of neutralizing anti-TGF-β antibody were capable of abrogating the suppression (72,79). Most other studies in vitro, however, have failed to confirm this finding. On the other hand, several in vivo studies have suggested that both IL-10 and TGF-β are required for the effector function of CD4+CD25+ regulatory cells. Anti-IL-10 and anti-TGF-β antibodies could abrogate the protective role of CD4+CD25+ cells in autoimmune thyroiditis (55), and inflammatory bowel disease (80), and blockade of IL-10 could abolish the suppression of skin allograft rejection mediated by CD4+CD25+ cells in an adoptive tolerance transfer model (78). The conflicting data concerning the role of IL-10 and TGF-βin vitro and in vivo seem hard to reconcile. Interestingly, data from Jonuleit et al. (81) and from Dieckmann and colleagues (82) have suggested that CD4+CD25+ regulatory cells are capable of inducing ‘infectious tolerance’ in cocultured CD4+CD25– cells. The latter were able to become TGF-β-secreting or IL-10-producing regulatory cells. These findings appear to provide a rationale for the involvement of IL-10 and TGF-β in the suppression of CD4+CD25+ cells in experimental animal models. At sites of inflammation such as in an allograft, CD4+CD25+ regulatory cells could suppress effector CD4+CD25– cells in a cell-contact-dependent (cell-surface TGF-β may be involved at this stage) and soluble cytokine-independent manner. The latter cells could differentiate into IL-10/TGF-β-producing cells after being suppressed, and thereby inhibit other CD4+CD25– cells. This may also explain the fact that adoptive transfer of a small number of CD4+CD25+ cells is sufficient to achieve transplantation tolerance or the prevention of autoimmune disease. Based on this assumption and a large body of evidence including our own we further propose that the mode of suppression by naturally occurring CD4+CD25+ cells per se would be IL-10-independent, however, suppression by ‘CD4+CD25+’ cells induced from CD4+CD25– cells in the presence of CD4+CD25+ naturally occurring suppressor cells in vivo or in vitro would be mostly IL-10-dependent (83–86).
Role of GITR. Two studies have provided evidence that glucocorticoid-induced TNF receptor (GITR) is preferentially expressed on murine CD4+CD25+ cells when compared to resting CD4+CD25– cells (87, 88). Importantly, anti-GITR antibody was able to abrogate CD4+CD25+ T-cell-mediated suppression and to induce spontaneous organ-specific autoimmune disease (87). Furthermore, GITR over-expression in CD4+CD25+ cells was observed in mice bearing an accepted skin allograft, but not in rejecting mice (89). Activated CD4+CD25– cells, however, can also up-regulate surface expression of GITR, and did not acquire suppressive properties spontaneously. Moreover, addition of anti-GITR antibody to previously activated CD4+CD25+ cells had no inhibiting effect in their effector function. This suggests that GITR may be involved in regulating the development of suppressive functions by murine CD4+CD25+ cells rather than in mediating their suppressive effects. In support of this notion T cells from GITR knock out mice showed increased proliferation and IL-2 production compared with wild-type cells upon anti-CD3 activation (90), and it will be interesting to see whether CD4+CD25+ cells from GITR–/– mice have impaired suppressive function. On the other hand, two lines of evidence and our own unpublished data have indicated that anti-GITR antibody was unable to abolish the suppression effected by human CD4+CD25+ cells or clones or lines despite the fact that these cells expressed GITR (91,92). The disparate results concerning the role of GITR in human and murine CD4+CD25+ T-cell effector function need to be clarified.
Antigen specificity. Limited data exist concerning the specificity of CD4+CD25+ T cells in the regulation of pathological immune responses. Several lines of evidence suggest that these cells have specificity for tissue-specific self-antigens. This was illustrated by the observation that CD4+CD25+ T cells from the pancreatic lymph nodes of prediabetic mice could transfer protection against diabetes if given in sufficient numbers to other prediabetic animals. No protection was conferred by the transfer of the equivalent population from anatomically distant lymph nodes (63). Such observations were compatible with one study in thyroid autoimmunity that CD4+ (probably CD4+CD45RC–) regulatory cells from athyroid rats were unable to control thyroid targeted autoimmunity, but these cells could still suppress diabetes (93). Despite this probable specificity for antigens derived from self-proteins it appears that this regulatory population has the capacity for allorecognition. Some of the most compelling evidence for this comes from adoptive transfer studies in models of transplantation tolerance. In many such models the tolerance can be transferred by CD4+ T cells from animals carrying a tolerated allograft to a naïve host, and the transferred tolerance appears to be specific for the alloantigens carried by the initial graft. The relevance of CD4+CD25+ T cells to this phenomenon has recently been demonstrated by Hara et al. and van Maurik et al. who showed that tolerance could be transferred selectively by CD4+CD25+ T cells and that this was alloantigen-specific (9,94).
Interestingly, it appears that the regulatory CD4+ T cells that maintain transplantation tolerance have indirect rather than direct allospecificity, although CD4+CD25+ cells with direct allospecificity have been generated ex vivo and could prevent GVHD after allogeneic bone marrow transplantation (95). These experimental findings are mirrored by recent observations in clinical transplant recipients. While we and others have failed to detect regulation of direct pathway antidonor alloresponses in stable transplant patients, depletion of CD4+CD25+ T cells could reveal significant indirect pathway antidonor alloresponses in a fraction of stable transplant patients (Game, Salama, Sayegh and Lechler, to be published). Probably the best evidence comes from our recent ex-vivo study in which we were able to raise and expand CD4+CD25+ T-cell lines with indirect allospeficity against a defined HLA-A2 allopeptide from human peripheral blood CD4+CD25+ T cells. The ex vivo-cultured CD4+CD25+ T cells retained suppressive properties. They suppressed CD4+CD25– cells with the same allopeptide specificity, and were also able to inhibit the alloresponse of naïve T cells to semiallogeneic DCs in the presence of the peptide (‘linked suppression’) (Jiang and Lechler, to be published). These findings suggest that peripheral CD4+CD25+ regulatory cells are a precommitted cell lineage from which cells with specificity for nonself-peptides can be selected. This may pave the way for inducing and expanding peptide antigen-specific regulatory T cells ex vivo for cell therapy in transplantation, allergy and autoimmune disease.