Naturally occurring CD4+CD25+ regulatory T (TR) cells, a component of the innate immune response, which play a key role in the maintenance of self-tolerance, have become the focus of numerous studies over the last decade. These cells inhibit the immune response in an Ag-nonspecific manner, interacting with other T cells. Much less is known about adaptive TR cells, which develop in response to chronic antigenic stimulation, and act directly on professional and nonprofessional APC, rendering them tolerogenic and able to elicit the differentiation of CD8+ and CD4+ T cells with suppressive activity. In this review, we will discuss data pertaining to the bidirectional interaction between Ag-specific TR with APC and their clinical relevance.
The primary function of the immune system is to protect the organism from pathogens. A highly effective and dynamic cellular network has evolved to signal the presence of pathogens and initiate a response that is specific for the invading pathogen yet maintains tolerance to self. APC are key regulators of the immune response, promoting or suppressing T cell responses depending on their functional state. Dendritic cells (DC) 2 are highly specialized APC that integrate a wide array of incoming signals and convey them to lymphocytes, directing the appropriate immune responses. Bidirectional interactions between DC and Ag-experienced T cells initiate either a tolerogenic or an immunogenic pathway. The outcome of this interaction is of crucial importance in malignancy, transplantation, and autoimmune diseases. There is increasing evidence that the immune response can be inhibited by various CD4+ and CD8+ regulatory T (TR) cells, which participate in innate or adaptive immunity (1, 2, 3, 4, 5, 6, 7).
Naturally occurring CD4+ and CD8+ TR cells develop during the normal process of T cell maturation in the thymus, play an essential role in preventing autoimmune responses, and are characterized by the constitutive expression of CD25 and forkhead-winged helix transcription factor FOXP3. After TCR triggering, natural TR cells inhibit immune responses in vivo and in vitro in an Ag-nonspecific and MHC-nonrestricted manner, via an APC-independent process (1, 2, 3, 8).
Adaptive TR cells are Ag induced, develop in the periphery, and exert their function either by secreting inhibitory cytokines (such as IL-10 and TGF-β) or tolerizing APC by direct cell-to-cell interactions. CD4+ and CD8+ IL-10-secreting TR cells, best known as TR1 cells, are FOXP3− and recognize Ags presented by myeloid as well as plasmacytoid DC (pDC) (9, 10, 11). Because IL-10 induces the up-regulation of inhibitory receptors and down-regulation of costimulatory molecules on APC, it disables the latter from efficient T cell priming (6). However, TR1 cell-mediated inhibition is not Ag specific because IL-10 induces bystander suppression of all immune components (7).
A distinct category of CD8+ and CD4+ Ag-induced TR act in a MHC class I- and class II-restricted manner, respectively (12, 13, 14). These Ag-specific TR do not secrete cytokines and interact directly with APC rendering them tolerogenic. In turn, tolerogenic APC trigger the generation of new waves of TR, which inhibit the immune response (14). We review data pertaining to mechanisms involved in the generation of tolerogenic APC and adaptive suppressor/TR cells.
Both in humans and mice, at least two distinct subsets of DC, myeloid DC (mDC) and pDC, have been described previously. mDC and pDC differ in phenotypic markers and functional properties (15, 16, 17).
mDC are generated in the bone marrow from CD34+ hemopoietic progenitor cells under the control of stromal growth factors or by direct contact with bone marrow stromal cells. Human CD11c+ mDC precursors include CD1a+ epidermal Langerhans cell precursors, CD1a− interstitial DC precursors, and CD14+ monocytes. These mDC precursors migrate into the skin epidermis and other tissues and become immature DC (iDC). iDC capture Ags in the peripheral tissues, process them into peptides, and then migrate to lymphoid organs where they present MHC-peptide complexes to T lymphocytes. During migration to lymphoid organs, DC mature into potent APC with increased immunostimulatory properties and decreased Ag-capturing capacity (15). The initial uptake and phagocytosis of microbes by DC is facilitated by TLR, which are key regulators of both innate and adaptive immune responses (18). TLR, the best characterized class of pattern recognition receptors, signal to the host the presence of infection. Other important pattern recognition receptors include mannose receptors (MR) and C-type lectins (18). TLR detect pathogen-associated molecular patterns, which are unique to the microbial world and invariant among entire classes of pathogens. Human mDC precursors express TLR2 and TLR4, whereas pDC precursors express TLR7 and TLR9. Signals through TLR induce mDC and pDC activation and maturation as measured by up-regulation of MHC class II, costimulatory and adhesion molecules (such as CD80, CD86, CD40, LFA3/CD58, and ICAM-1/CD54), CD83, a characteristic maturation marker, and production of large amounts of immune enhancing cytokines such as IL-12 and IFN-α, respectively. pDC differentiate in lymphoid organs from CD11c− pDC precursors, which derive from common lymphoid progenitors. They lack the myeloid markers CD13, CD14, and CD33 yet express high levels of CD123 (the IL-3R α-chain), CD4, and CD62L. Immature pDC undergo maturation upon recognition of viral components via TLR. The maturation process is mediated by type 1 IFN, acting in an autocrine manner (15, 17).
The concept has been proposed that in the steady state, iDC play a major role in silencing self-reactive T lymphocytes and avoiding “horror autotoxicus” (19). They may present self-Ags to developing T cells in the thymus, inducing negative selection of autoreactive cells, which accounts for central tolerance to self. iDC may also mediate Ag-specific peripheral tolerance, capturing soluble Ags as well as apoptotic cells and inducing either the deletion of responding T cells or the generation of TR cells upon migration to regional lymph nodes (19). This view is supported by evidence that CD83− human iDC, used for in vitro allostimulation of cord blood CD4+ T cells, induce CTLA4high T cells, which act as non-Ag-specific T suppressor (TS) cells (20). Infusion of donor iDC prolongs cardiac allograft survival in nonimmunosuppressed recipient mice (21). Similarly, human iDC pulsed with influenza matrix peptide induce the in vivo generation of IL-10-producing CD8+ TR cells (22). The tolerogenic function of iDC both in innate and adaptive immunity may be related to expression of the endocytic MR, a member of the type I C-lectin receptor superfamily, which can deliver negative signals to the cells. MR levels are up-regulated by anti-inflammatory molecules (IL-10, corticosteroids, and vitamin D3) or Th2 cytokines (IL-4 and IL-13) and PG, yet they are down-regulated by proinflammatory stimuli (LPS and IFN-α) (23, 24). There is evidence that Ag targeting through C-type lectin receptors under steady-state conditions (i.e., in the absence of inflammatory stimuli) renders APC capable of triggering anergic T cells with suppressive/regulatory activity. Engagement of MR on iDC induces the release of anti-inflammatory molecules, such as IL-1R antagonist and nonsignaling (decoy) receptor IL-1RII (both of which neutralize IL-1 activity), the increase of IL-10, and the decrease of IL-12 production in maturing mDC (23). In the presence of mucins (MUC1) released by tumor cells, monocyte-derived DC also differentiate into high IL-10- and low IL-12-producing DC with tolerogenic activity (25).
Similarly, immature pDC induce anergy in human CD4+ T cell clones with various antigenic specificities and elicit the generation of IL-10-producing CD4+CD25+ TR1 cells. The anergizing effect of pDC on Ag-specific Th cells is not mediated by IL-10, IFN-α, or TGF-β because it cannot be abrogated by the corresponding neutralizing mAbs. Instead, it seems to be attributable to the expression of inhibitory receptors on these immature pDC (17).
However, there is evidence that not only immature but also mature mDC are able to induce CD4+ T cell tolerance. Thus, murine pulmonary DC, with a mature phenotype, produce IL-10 and elicit the generation of CD4+ TR1 cells (26). Evidence has also been provided that pathogen-specific TR1 cells can be induced in the respiratory tract by a bacterial molecule that stimulates IL-10 production by DC (27). In other studies, repetitive injections of DC matured with TNF-α were found to induce IL-10-producing TR1 cells, which mediated Ag-specific protection of mice from autoimmune encephalomyelitis (28).
Furthermore, it has been demonstrated recently that splenic stroma drives mature mDC to differentiate into a distinct population of regulatory DC, which induce T cell activation without proliferation and inhibit proliferation while sustaining activation (29).
Similar to mature mDC, pDC can also display tolerogenic activity after maturation. Thus, CD40L-activated pDC induce the generation of human CD8+ and CD4+ IL-10-producing TR cells (11, 17). The finding that mature DC are able to induce TR cells indicates that tolerogenicity is not an exclusive attribute of iDC.
Functional licensing of DC
License to kill.
Effective priming of CD4 T lymphocytes requires cell-to-cell interaction between APC and T cells. This interaction allows bidirectional costimulatory signals, which activate the APC presenting the processed Ag and those T cells with cognate-specific TCR. TCR triggering results in up-regulation of CD40L (CD154) expression on activated T cells. CD40L interacts with CD40 on APC inducing them to up-regulate B7 (CD80 and CD86) costimulatory molecules. These B7 molecules interact with their counter-receptor (CD28) augmenting CD40L expression and inducing CD4 Th cell proliferation (15, 16, 17, 18). Signaling along the B7 pathway is bidirectional and enables the conditioning of DC as evidenced by induction of IL-12, IL-6, and IFN-γ production, expression of IDO, and activation of the MAPK pathway (30). T cell activation in the absence of B7 costimulation results in T cell paralysis or anergy (31).
In 1998, three groups demonstrated independently that the differentiation of CD8+ CTL is contingent upon recognition of MHC class I/peptide complexes on APC that have been “licensed” or conditioned by CD4+ Th cells, via the CD40-signaling pathway. For effective CTL priming, different peptides of the same protein must be recognized on the same APC by CD4+ and CD8+ T cells (32, 33, 34, 35). Emerging evidence indicates that DC have a specialized capacity to process exogenous Ags into the MHC class I pathway, being able to cross-prime CD8+ T cells (36). Stimulatory pathogens and inflammatory cytokines can also license DC to stimulate T cell responses (32, 33, 34, 35).
License to heal.
It was also in 1998, when our group first demonstrated that human CD8+ TS cells, generated by multiple rounds of in vitro allostimulation, interact with APC down-regulating the expression of costimulatory molecules and the capacity to trigger CD4+ Th cell proliferation (12). Multiple stimulation of human T cells with xenogeneic APC or with peptide-pulsed APC resulted also in the generation of CD8+ TS cells capable of inhibiting Th proliferation (37, 38). In later studies, we showed that CD8+ TS cells derive from an oligoclonal population, are MHC class I restricted, lack CD28, and express FOXP3, GITR, CTLA-4, CD25, OX40, CD103, CD62L, 4-1BB, and TNFRII at about the same level as in CD4+CD25+ natural TR cells for which they represent specific molecular markers (1, 39). These CD8+ TS cells interact directly with APC yet produce no IFN-γ, IL-10, TGF-β, or other known cytokines. CD8+CD28− TS cells inhibit CD40-mediated up-regulation of costimulatory molecules such as CD80 and CD86 on APC that present the peptide-MHC class I complexes to which the CD8+CD28− TS cells have been primed previously. The suppressed APC are rendered unable to induce and sustain the full program of CD4+ Th cell activation due, at least in part, to inhibition of NF-κB activation and transcription of costimulatory molecules in APC (12, 13, 14, 37, 38, 39, 40, 41).
After exposure to CD8+CD28− TS cells, monocytes and DC show increased expression of the genes encoding Ig-like transcript (ILT)3 and ILT4. The inhibitory receptors ILT3 and ILT4 (13), which are expressed by monocytes and DC, belong to a family of Ig-like inhibitory receptors that are structurally and functionally related to killer cell inhibitory receptors. The subset of ILTR, which includes ILT3 and ILT4, displays a long cytoplasmic tail containing ITIM. These receptors mediate inhibition of cell activation by recruiting tyrosine phosphatase Src homology region 2 domain-containing phosphatase 1 (42, 43, 44, 45). Coligation of ILT in monocytes inhibits Ca2+ mobilization and tyrosine phosphorylation triggered by Ab ligation of FcγRII (CD32), HLA-DR, and FcγRI (CD64). Although the ligand for ILT3 is unknown, ILT4 binds to the α3 domain of HLA class I (HLA-A, HLA-B, HLA-C, and HLA-G), competing with CD8 for MHC class I binding (42, 43, 44, 45).
We demonstrated that overexpression of ILT3 is associated with inhibition of NF-κB activation and with a reduced capacity of tolerogenic APC to transcribe NF-κB-dependent costimulatory molecules (13, 41). When human CD4+CD25+ T cells are allostimulated with ILT3highILT4high DC, they differentiate into MHC class II-specific CD4+CD45R0+CD25+ TR cells (14). Similar to CD8+CD28− TS cells, CD4+ TR cells are Ag specific and act on APC in a cytokine-independent, contact-dependent manner, inducing the up-regulation of the inhibitory receptors ILT3 and ILT4 (14). Hence, the interaction of iDC with HLA class I-restricted CD8+CD28− TS cells or HLA class II-restricted CD4+CD25+ TR cells induces them to become tolerogenic and licensed to inhibit the activation of Th and CTL. Based on these findings, we conclude that ILT3highILT4high DC are licensed to heal, providing a useful tool for treatment of autoimmune diseases and prevention of allograft rejection.
We found that ILT3high ILT4high tolerogenic DC can also be induced by treating monocyte-derived iDC with IL-10, IFN-α, or a mixture of IL-10 and IFN-α (16). Such tolerogenic DC display low allostimulatory capacity and costimulatory molecules. T cells primed with IL-10- and IFN-α-treated DC become anergic. Their reactivity can be restored by adding IL-2 or Abs to ILT3 and ILT4 to the culture (13, 14). DC tolerized by cytokine treatment induce the generation of CD4+CD45R0+CD25+ TR cells, which inhibit naive T cell proliferation in an allorestricted manner (14). In affymetrix studies, other investigators showed that monocyte-derived DC, after treatment with IL-10, overexpressed eight ITIM-containing inhibitory molecules (ILT2, ILT3, ILT4, ILT5, DCIR, PILRA, FcγRIIB, and SLAM). Phenotypic studies of the ILThigh subset of IL-10-treated DC revealed expression of CD14, TLR2, DC-SIGN, and CD123 and the lack of lymphocyte, monocyte/macrophage, and plasmacytoid markers such as CD3, CD8, and CD68 (46). The allostimulatory capacity of ILThigh DC was reduced by 89% compared with control iDC (46). However, no evidence was provided that inhibitory receptors other than ILT3 and ILT4 are important for suppression of T cell reactivity by tolerogenic DC.
The finding that IL-10 renders human DC tolerogenic and able to elicit the generation of regulatory CD4+ and CD8+ T cells is supported by numerous studies (47, 48, 49), which, however, attributed this phenomenon to the lack of immunogenic properties and immature phenotype of IL-10-treated DC (CD83−, CD80low, CD86low, and MHC class IIlow). Because ILT3highILT4high tolerogenic DC differ from both immature and mature DC with respect to chemokines, cytokines, transcription factors, apoptosis-related proteins, cell growth regulators, and molecules involved in signal transduction, it is more likely that rather than being immature, they are the product of an alternative pathway of differentiation (50, 51).
It is important to notice that CD4+ TR1 regulatory cells differentiate in the presence of IL-10 and regulate T cell responses via their ability to produce IL-10 and TGF-β (9, 10). Therefore, it is quite likely that their suppressive activity derives from their capacity to secrete IL-10, which triggers ILT3 and ILT4 up-regulation in APC.
Recently, it has been shown that the combination of IL-16 plus thrombopoietin (TPO) induces the in vitro generation of tolerogenic DC from CD34+ stem cells. These tolerogenic DC induce anergy in T cells, display high levels of ILT2, ILT3, and ILT4, low levels of the C-type lectin DC-SIGN (CD209, an adhesion receptor that regulates DC trafficking and T cell interactions), and produce high amounts of the immunosuppressive cytokines TGF-β and IL-10. The tolerogenic properties of IL-16/TPO-treated DC could not be attributed to an immature state because the level of expression of HLA and costimulatory molecules was similar to that displayed by immunogenic DC, which had been matured with other cytokine combinations (52). TNF-α (28) and TGF-β2 (53) also render mDC capable to trigger IL-10-producing regulatory CD4+ or CD8+ T cells although it is unknown whether they induce the up-regulation of inhibitory receptors.
DC can also be licensed to become tolerogenic by use of several immunosuppressive agents (54). Of particular interest is 1α,25-dihydroxyvitamin D3 and its analogues, which render DC tolerogenic inducing the up-regulation of ILT3 and secretion of IL-10, while down-regulating the expression of costimulatory molecules and the production of IL-12 both in vitro and in vivo. Tolerogenic DC generated by treatment with this agent induce the generation of CD4+CD25+ TR cells, which mediate tolerance in experimental models of transplantation and autoimmunity (55, 56).
Recently, the concept has been advanced that cells expressing IDO, an enzyme that degrades the amino acid tryptophan, can suppress T cell responses and promote tolerance (57). In mice, a subset of DC expresses detectable levels of IDO protein and induces potent T cell suppression (57). Certain CD4+ TR cells can induce IDO expression, which can be further potentiated by IL-10 (57). It has been suggested that tolerogenic signals such as CD80/CD86 ligation by CTLA4high TR cells could induce IDO expression eliciting a tolerogenic DC phenotype, whereas immunogenic signals such as CD40 ligation would promote an immunogenic phenotype and down-regulate IDO expression. This hypothesis predicts that both immunogenic and tolerogenic DC are fully competent mature APC, which serve opposing functions (30, 57).
Tolerogenic endothelial cells (EC)
Evidence has been provided that human vascular EC present microbial Ags to memory T cells as a mechanism of immune surveillance. In turn, activated T cells provide both soluble and contact-dependent signals to regulate the functions of EC, including recruitment of inflammatory leukocytes and Ag presentation, which leads to activation of T cells (58). Because EC are positioned ideally to come into contact with circulating T cells, they may play an important function in regulating the immune response. In earlier studies, we have demonstrated that xenospecific human TS cells inhibit the transcriptional activation of CD86 in porcine APC and EC (37, 59). More recently, we showed that alloantigen-specific CD8+CD28− TS cells interact in a HLA class I-specific manner with activated or nonactivated EC, inducing up-regulation of the inhibitory receptors (ILT3 and ILT4), down-regulation of costimulatory (CD40), adhesion (CD54, CD58, and CD62E), MHC (HLA-DR), and maturation (CD83) cell surface molecules. These ILT3highILT4high EC acquire tolerogenic properties eliciting the in vitro differentiation of CD8+CD28−FOXP3+ TS cells, which inhibit Th1 proliferation and CTL function. Modulation of EC in favor of promoting tolerance may be important for maintenance of tolerance to autologous and allogeneic tissues (60).
The functional and phenotypic overlap between tolerogenic DC and EC is not surprising because both in mice and human monocytes and bone marrow-derived DC undergo endothelial-like differentiation in the presence of angiogenic growth factors (61). While the generation of tolerogenic DC and EC is obviously undesirable in cancer, it could be highly beneficial in autoimmunity and transplantation.
Clinical relevance of tolerogenic APC
Progress in the treatment of several clinical conditions, including transplantation, autoimmunity, infectious diseases, and cancer, can be greatly influenced by the development of agents that empower APC to down-regulate or up-regulate the immune response. Research in this area is in progress in many laboratories.
The development of tolerance to alloantigens is a dynamic process involving many mechanisms that might contribute at different stages. Persistence of Ag is known to be essential, because in the absence of alloantigen, tolerance is lost either immediately or gradually (4, 51). Although T cell deletion, clonal exhaustion, and anergy induced under suboptimal conditions of allostimulation (including costimulation blockade) have long been known to play an important role in tolerance, the involvement of TR cells and tolerogenic APC has been documented only recently in several experimental models (13, 60, 62).
1α,25-Dihydroxyvitamin D3 induces the differentiation of DC with a tolerogenic phenotype, characterized by increased expression of ILT3 (in human), low IL-12, and enhanced IL-10 secretion (24, 55, 56, 63). Treatment with this agent promotes tolerance to fully mismatched mouse pancreatic islet and vascularized heart allografts. Graft acceptance is associated with an increased percentage of CD4+ TR cells in the spleen and draining lymph nodes.
An active role for CD4 TR and tolerogenic DC has been demonstrated also in a mouse model of heart transplantation in which tolerance was induced by use of anti-CD45RB mAb and an analogue of the antirejection drug 15-deoxyspergualine, which suppresses NF-κB activation. Tolerogenic DC isolated from tolerant mice induced the in vitro generation of CD4 TR cells, whereas TR cells from such recipients induced tolerogenic DC when coincubated with DC progenitors (64).
Studies from our laboratory have documented the capacity of CD8+ TS cells to induce inhibitory receptors in EC and DC in a heart allotransplantation model in rats (65). In this model, ACI rats were rendered tolerant to Lewis heart transplants by multiple transfusions of UVB-irradiated blood. Tolerance was transferred from primary to secondary, syngeneic recipients by infusion of CD8+ T cells, which induced the up-regulation of the paired Ig-like receptor B (PIR-B, the murine orthologue of ILT4) in donor strain DC and EC (65). We demonstrated that EC from long-term surviving heart allografts expressed PIR-B and were tolerated indefinitely when retransplanted in secondary hosts, syngeneic to primary recipients (65). The role of PIR-B in tolerance induction has been also suggested in other studies showing that transfer of allogeneic splenocytes into PIR-B-deficient mice induced augmented activation of recipient DC, exacerbating graft-vs-host disease (66).
The clinical significance of tolerogenic DC and EC with high ILT3 and ILT4 expression was documented in studies of human heart transplant recipients. CD8+CD28− T cells from quiescent patients inhibited CD40 signaling of donor monocytes and EC carrying the donors’ HLA class I Ags and induced up-regulation of ILT3 or ILT4 (14, 60, 67). EC from endomyocardial biopsies obtained from rejection-free patients displayed ILT3 and ILT4. These inhibitory receptors were not expressed on EC from patients with rejection and could not be induced in donor APC using CD8+CD28− T cells from these patients. The bidirectional interaction between TS cells and EC perpetuated long-term quiescence as demonstrated by the persistence of TS cells 3 yr following transplantation in patients without chronic rejection (60). Such studies permit the identification of patients who could benefit from partial or complete withdrawal of immunosuppression. This is an important aim in view of the morbidity and mortality associated with the long-term use of immunosuppressive drugs.
It is still unknown how many inhibitory receptors contribute to the tolerogenicity of DC or EC exposed to TS cells. However, PIR-B in rodents and ILT3 and ILT4 in humans serve as markers for the identification of tolerogenic APC and as a tool for recognizing TS/TR activity.
The maintenance of tolerance to self is largely attributable to CD4+CD25+ natural TR cells (1). However, breakdown of self-tolerance can occur, resulting in the development of organ-specific autoimmune diseases. The search for therapeutic strategies, which can inhibit the immune response in an Ag-specific manner, is of paramount importance in the field. Agents that induce inhibitory receptors on APC such as 1α,25-dihydroxyvitamin D3 were shown to prevent systemic lupus erythematosus in mice, experimental allergic encephalomyelitis, collagen-induced arthritis, Lyme arthritis, inflammatory bowel disease, and autoimmune diabetes in NOD mice. Furthermore, this agent inhibits the recurrence of autoimmune diseases after allotransplantation in NOD mice and ameliorates significantly chronic relapse of experimental allergic encephalomyelitis (63).
In other murine autoimmune disease models, CD8+CD28− TR cells were shown to inhibit the up-regulation of costimulatory molecules on APC through a cell-cell contact mechanism, thus inhibiting activation and clonal expansion of encephalitogenic CD4+ Th1 cells (68). These studies support the notion that the bidirectional interaction between TR cells and APC plays a critical role in prevention of autoimmunity.
It has become increasingly evident that defects in DC function play an important role in nonresponsiveness to tumors. The dysfunction of DC in malignancy has been attributed to abnormalities in the differentiation of mature mDC caused by tumor-produced cytokines such as endothelial growth factor, IL-10, GM-CSF, IL-6, and M-CSF. There is evidence that IL-10 is produced early at tumor sites and that it induces subsequent generation of CD4+ TR cells causing a systemic collapse of antitumor immunity (69). The possibility exists that IL-10 and other cytokines produced by tumors induce the differentiation of tolerogenic DC.
The consequences of viral infection on DC maturation depend on the type of virus. While infection with influenza and dengue virus leads to DC maturation, infection with herpes simplex, vaccinia, hepatitis C viruses, and HIV inhibit DC maturation. CMV and EBV inhibit monocyte differentiation to iDC. Measles virus and Ebola also impair DC function (70). Currently, it is unknown whether any of these viruses with the exception of HIV induce the expression of inhibitory receptors on APC, which may trigger the generation of TS or TR cells. However, we have shown that HIV+ subjects display increased ILT4 expression on CD14+ monocytes, mostly due to the elevated levels of IL-10 in their serum. As a consequence, their monocytes have deficient Ag-presenting and allostimulatory capacities (71).
There is ample evidence that naturally occurring CD4+CD25+ TR cells play a major role in establishing and maintaining self-tolerance. Natural TR cells are produced by the normal thymus as a distinct subpopulation and persist in the periphery with stable function. Congenital deficiency or experimental depletion of natural TR cells lead to autoimmunity (1, 2). However, because natural TR cells are not Ag specific, it is difficult to envisage advantages emerging from their use for treatment of transplant rejection, autoimmunity, or allergy compared with iatrogenic immunosuppression.
In contrast, Ag-specific CD8+CD28− TS and CD4+CD25+ TR cells may be ideally suited for inhibition of the immune response. They seem to act through a common mechanism inducing the expression of inhibitory receptors ILT3 and ILT4 in professional or nonprofessional APC (13, 14, 60). The interaction between TS/TR cells and APC is bidirectional, creating an immunoregulatory loop in which tolerized APC further elicit the differentiation of TS and TR cells (Fig. 1). Molecular characterization of Ag-specific TS and TR cells has shown that they share specific molecular markers with natural TR cells (39, 62). This phenotypic overlap indicates that none of these markers controls the selective interaction of natural TR cells with other T cells or of adaptive TS/TR cells with professional and nonprofessional APC. The transcriptional repressor FOXP3 is highly expressed by both natural and adaptive TS/TR cells, yet it is not up-regulated in IL-10-producing TR1 cells. Because the inhibitory effect of TR1 cells is mediated by IL-10 rather than by cell-to-cell contact, it is possible that expression of FOXP3 endows natural and adaptive TS/TR cells with the capacity to tolerize the cells with which they interact. In contrast, because IL-10 induces the up-regulation of ILT3 and ILT4 on DC and EC, the ultimate effect of TR1 cells on APC may be the same.
The finding that tolerogenic DC display a distinct phenotype characterized by the up-regulation of ILT3 and ILT4 fits with the recently proposed concept that “mature but quiescent DC induce tolerance and activated mature DC induce immunity” (72). Our results are consistent with the hypothesis that immunologic tolerance is the end result of a cascade of events that promote the induction of T cell anergy (5). In our “tolerogenic cascade” model, Ag-specific, MHC class I-restricted CD8+CD28− TS cells set the stage by generating the first wave of tolerogenic DC. These tolerized DC anergize CD4+ T cells that spread infectious tolerance to other naive T cells capable of recognizing MHC peptide complexes on their membrane (4, 5, 14, 51) (Fig. 1).
Because in our experimental models, multiple (chronic) antigenic stimulation was required to generate TS cells and set in motion the immunoregulatory cascade, it is conceivable that an efficient immune response mediated by the interaction of Th and cytotoxic cells with APC is ultimately down-regulated by the emergence of TS and TR cells (12, 14, 60, 65). The delicate balance between immunity and tolerance may depend on the size of the effector populations or prevalence of inflammatory or inhibitory cytokines that may affect the immunogenic or tolerogenic phenotype of APC.
However, mature DC can further differentiate under the influence of stromal cells into regulatory DC, which inhibit T cell proliferation, through the production of NO (29). Furthermore, pDC matured in vitro by CD40 ligation (11, 17) and mature mDC treated with TNF-α (28), IFN-α (73), TPO plus IL-16 (52), or TGF-β2 (53) produce IDO and/or IL-10, inhibiting T cell proliferation and eliciting the differentiation of CD4+ or CD8+ TR1 cells. In turn, the emerging TR1 cells can tolerize APC and initiate a suppressor cell cascade (Fig. 2). However, it appears that tolerogenic mDC or EC maintain a stable phenotype and function, suggesting that they may be terminally differentiated.
The tolerogenic cascade model does not detract from the importance of thymic deletion of self-reactive T cells and/or from the importance of apoptosis as an effective mechanism that limits the immune response. It only brings together various aspects of peripheral tolerance by emphasizing the central role played by tolerogenic APC.
The concept emerges that regulation of the immune system occurs in a self-sustaining loop, whereby TS/TR cells induce the differentiation of tolerogenic APC and vice versa. Therefore, breaking or initiating this process will be of great value for treatment of various diseases. Breaking the chain of tolerogenic events is of obvious importance in malignancies and infectious diseases. It is believed that progression of malignancies is caused by the inefficiency of the immune response against tumor-specific peptides/MHC class I complexes. Notoriously, numerous tumors display a decrease in the level of expression of MHC Ags escaping recognition by Ag-specific T cells. To create conditions that optimize licensing of APC to kill rather than heal, it may be necessary to block their capacity to transcribe inhibitory receptors. Treatment with immunostimulatory agents and/or depletion of natural and adaptive TS/TR cells may accomplish this goal.
Conversely, in autoimmunity and transplantation, tolerogenic DC may serve as a vehicle for vaccines that prevent progression of the immune response. The list of inhibitory cytokines and pharmaceutical agents, which induce tolerogenic APC, is certain to increase as more compounds are found to induce ILT3 and ILT4 up-regulation in DC and EC. Large amounts of autologous DC could be sorted and/or differentiated from leukophoresis-derived monocytes and CD34+ hemopoietic stem cells (74).
In transplantation, recipient DC can be tolerized ex vivo in the presence of apoptotic donor cells. Tolerized DC may inhibit both the direct and indirect pathway of allorecognition eliciting the differentiation of CD8+ TS cells. This outcome is suggested by experiments in which CD8+ TS cells, which mediated tolerance to allogeneic heart transplants, were induced in rats by chronic exposure to apoptotic (UVB-irradiated) donor blood. CD8+ TS cells inhibited rejection via direct or indirect alloreactivity and transferred tolerance to secondary hosts (65).
It is also possible that tolerance to allogeneic transplants can be induced by treatment with soluble ILT3 (sILT3). This hypothesis is supported by our recent finding that recombinant sILT3 inhibits T cell alloreactivity and converts activated CD8+ T cells in FOXP3+ TS cells. sILT3 is unlikely to induce nonspecific immunosuppression because it acts only on activated T cells (N. Suciu-Foca et al., manuscript in preparation).
Another field that may benefit from the use of tolerogenic DC is autoimmunity. It is believed that DC, which present tissue- or organ-specific peptides, derived from cells undergoing necrosis under inflammatory conditions, can prime autoreactive T cells. The progression of the autoimmune response may be prevented by treatment with autologous tolerized DC that have processed ex vivo apoptotic cells containing the putative autoantigen. Alternatively, administration of sILT3 may convert proliferating autoreactive T cells into TS cells, which initiate and maintain the cascade of suppressive events.
It is our belief that licensing APC to heal, by inducing them to express inhibitory receptors, opens new avenues to induction of immunological tolerance in transplantation and autoimmune diseases. The finding that APC and, indeed, EC are the true source of immune regulation explains what otherwise would be a paradoxical and counterproductive aspect of immune regulation. If immune regulation depended on non-Ag-specific T cells alone, it would be broadly suppressive and thus possibly more detrimental to well-being than the autoimmune reactions it presumably controls. In contrast, the focus on APC and EC allows immune regulation to be tissue specific and site specific, and in this way, it provides much more efficient control mechanism.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Abbreviations used in this paper: DC, dendritic cell; TR, regulatory T; pDC, plasmacytoid DC; mDC, myeloid DC; iDC, immature DC; MR, mannose receptor; TS, T suppressor; ILT, Ig-like transcript; TPO, thrombopoietin; EC, endothelial cell; PIR-B, paired Ig-like receptor B; sILT3, soluble ILT3.