Regulatory T cells—the renaissance of the suppressor T cells

Abstract

Immune reactions are stringently regulated and balanced by complex interactions of stimulating and suppressing mechanisms. Dysfunctions of this sophisticated immune regulatory network can lead to a variety of diseases such as autoimmunity, allergy, cancer, and pregnancy disorders. The rediscovery of suppressor T cells a decade ago—now designated as T regulatory cells—set off a huge avalanche of research activities leading to a multitude of preclinical and clinical studies. Herein, we give a comprehensive review about this research on T regulatory cells and the relevance of this suppressive T cell population for the development of innovative immune therapeutic strategies.

• Autoimmunity
• cancer
• immunotherapy
• T regulatory cells
• transplantation
• Tregs

Abbreviations
IL	=	interleukin
APC	=	antigen presenting cells
TGF	=	transforming growth factor
cAMP	=	cyclic AMP
MHC	=	major histocompatibility complex
GITR	=	glucocorticoid‐induced TNF receptor
CTLA	=	cytotoxic T lymphocyte antigen
TLR	=	toll‐like receptor
CCR5	=	CC chemokine receptor 5
IFN	=	interferon
NK	=	natural killer cells
TNF	=	tumor necrosis factor

Introduction

History

The revelation of the thymus function and of thymus‐derived lymphocytes (T cells) during the 60s and 70s of the last century clearly indicated the vital importance of T cells for the regulation of an immune response 1,2. Besides the immune stimulating function of T cells it was quite obvious that a certain fraction of this population has suppressive properties and therefore was referred to as suppressor T cells 3,4. A multitude of studies convincingly demonstrated the crucial role of suppressor T cells for the maintenance of immunologic tolerance 5. Unfortunately, the characterization of I‐J determinants on suppressor T cells that subsequently could not be approved by genomic analyses discredited this field of immunology for a long time 6–9. Nevertheless, many studies using autoimmune and transplantation models consistently demonstrated that suppressor T cells exist although it was not possible to unequivocally characterize such cells. Finally, a ground‐breaking study by Sakaguchi and colleagues definitively identified suppressor T cells as CD4⁺CD25⁺ T cells that emerged from the thymus and could be isolated from the periphery of nonimmunized mice 10. In order to avoid the discredited expression suppressor T cells these cells were designated as T regulatory cells (Tregs) or more precisely as naturally occurring Tregs (nTregs). nTregs comprise 5%–10% of whole CD4⁺ T cells, and depletion of them before transfer of CD4⁺ T cells into T cell‐deficient mice resulted in a large panel of different autoimmune diseases 10. Thus, these findings indicate that a fraction of the positively selected peripheral CD4⁺ T cells is potentially autoreactive and requires being tolerized by nTregs. Initial analyses revealed that the reduction of interleukin (IL)‐2 expression in responder T cells is the main target of nTreg‐mediated suppression 11. Regarding their characterization, it should be noted that CD25 is not a unique membrane‐bound marker of nTregs but also expressed by activated conventional T cells. Thus, CD25 is not exclusively expressed by nTregs. Moreover, despite very intensive endeavors an exclusive surface marker for nTregs could not be identified yet, but the transcription repressor FoxP3 was found to be expressed selectively by murine nTregs 12,13.

In the beginning of 2000 it was shown that mutations in the gene encoding for FoxP3 were responsible not only for the phenotype seen in the so‐called scurfy mouse but also were the cause for IPEX (immune dysregulation, polyendocrinopathy, enteropathy, X‐linked syndrome) in humans 14–17. Recent studies clearly revealed that FoxP3 is not only a transcriptional repressor but also essential for the development and function of murine and human Tregs 12,18,19. In contrast to mice, human conventional CD4⁺ and CD8⁺ T cells have been shown to upregulate FoxP3 mRNA and protein upon in vitro activation 20–24, suggesting a cell‐intrinsic role, e.g. in regulating cytokine production and proliferation in these cells rather than the conversion into adaptive regulatory T cells. Moreover, the existence of IPEX‐like individuals which lack mutations within the FoxP3 gene but are phenotypically not distinct from IPEX‐patients casts doubt on the role of FoxP3 in the development and function of human Tregs. On the other hand, the result that transduction with FoxP3 of human conventional CD4⁺ T cells leads to T cells with suppressive properties 25, resumes the discussion whether the mechanisms of nTreg development and function differ in humans.

The characterization of human Tregs as well as the description of various T cell populations with suppressive activities, provisionally termed induced Tregs (iTregs), in vitro and in vivo, eventually led to a multitude of studies, which resulted in a huge amount of data. This review is primarily focused on the function of nTregs/iTregs in several preclinical animal models and their potential exploitation for novel immune therapeutic strategies.

Treg subpopulations

It was clearly shown by thymectomy and by using T cell receptor transgenic mice that nTregs develop in the thymus to a discrete T cell population 10,26. However, it is still a prominent question, at least with respect to humans, if and to what extent such cells with suppressive potencies can arise in the periphery 27. Indeed, there are several evidences for the peripheral induction of CD4⁺ T cells with suppressive properties (iTregs, Table I). However, in the human system, as CD25 and even FoxP3 are not reliable markers for human nTregs, it is a desirable approach to determine whether large numbers of iTregs are generated in humans. As it is therefore not possible to discriminate between nTregs and iTregs in the periphery of an individual, these populations are in the following referred to as Tregs.

Table I. Characteristics of naturally occurring (nTregs) and induced T regulatory cells (iTregs).

	nTregs	iTregs
	CD4^+CD25^+ Tregs	Tr1/Tr1‐like	Th3/Th3‐like
Origin	Thymus	Periphery	Periphery
Development	• CD28‐CD80/86	• IL‐10	• TGF‐β, IL‐4, IL‐10
	• Intermediate TCR‐signal	• Tolerogenic DC, nTregs	• Tolerogenic DC, nTregs
	• Unknown molecules	• Vitamin D3 + dexamethasone	• Oral/mucosal tolerance
FoxP3 Expression	+++	Variable	Variable
Suppression	Contact‐dependent, IL‐10 in vivo	Mainly IL‐10	Mainly TGF‐β

In vitro, conventional CD4⁺ T cells can be converted to so‐called Tr1 (T regulatory type 1) or Th3 (T helper type 3) cells by using different approaches. In general, all these approaches favor the activation of T cells under suboptimal or repressing conditions, e.g. repetitive stimulation with nonprofessional antigen presenting cells (APCs) or in the presence of immune‐suppressive drugs or cytokines like transforming growth factor (TGF)‐β or IL‐10. Tr1 cells were initially generated by chronic stimulation of conventional CD4⁺ T cells in the presence of IL‐10 28. In addition, the in vitro treatment with a combination of vitamin D3 and dexamethasone resulted also in the generation of T cells with suppressive properties 29. Complementary to Tr1 cells, Th3 cells which produce preferentially TGF‐β have been first generated in vivo following oral tolerance induction 30. Besides these studies, it was recently shown that FoxP3⁺ iTregs develop from conventional CD4⁺ T cells after activation in the presence of TGF‐β and IL‐2 31,32. Furthermore, there is growing evidence that cyclic AMP‐elevating agents, like prostaglandin E₂ (PGE₂) or the vasoactive intestinal peptide (VIP), not only induce the expression of FoxP3 in conventional T cells but also promote the generation of T cells with suppressive properties in vivo and in vitro33–35. Finally, also dendritic cells (DC) can play a decisive role in the extrathymic induction of T cells with regulatory properties. The activation of conventional CD4⁺ T cells in vitro and in vivo with immature DC or with DC treated with immune‐suppressive cytokines (e.g. IL‐10, TGF‐β) results in the induction of anergic T cells with regulatory properties 36.

The number of iTreg family members is still growing and iTregs cannot be discriminated from nTregs by specific surface markers. However, some iTregs are distinct from nTregs as the suppressive potency of the former is generally based on immune‐suppressive cytokines (e.g. IL‐10 or TGF‐β) and not on cell contact‐dependent mechanisms as shown for nTregs. Furthermore, iTregs are heterogeneous in their ability to express CD25, glucocorticoid‐induced TNF receptor (GITR) and FoxP3. However, despite the heterogeneity of iTregs and their mechanism of action, the therapeutic potential of these peripherally generated iTregs is of great interest especially in the human system. As the isolation of large numbers of nTregs from the peripheral blood of individuals is challenging, the generation of great amounts of antigen‐specific iTregs ex vivo might represent a sanguine approach for the control of autoimmune responses.

Suppressive mechanisms

Central to the understanding of nTregs is the molecular basis of their mechanism of action. Although great effort has been made to unravel these mechanisms, even the basics of this process are far from definitive 37. Surveying the hitherto performed in vitro and in vivo studies the picture gets even more disconcerting. In vitro it was clearly shown that nTregs exert their suppressive properties via a cell contact‐dependent mechanism which requires antigen‐specific T cell receptor stimulation but is independent from APCs 38. In contrast to the in vitro obtained results, the in vivo suppressive mechanism of nTregs remains controversial. In particular all hitherto performed in vitro studies clearly vote for the colocalization and stable T cell interactions between nTregs and effector T cells. However, recent studies using two‐photon laser‐scanning microscopy to directly visualize nTregs and effector T cells within the same lymph node suggested only short‐term interactions between nTregs and effector T cells 39. Moreover, Tang et al. showed that nTregs formed persistent cell contacts with dendritic cells (DC) preceding a shortened contact between the effector T cells and DC, assuming that a cell contact‐dependent mechanism of suppression by nTregs in vivo might preferentially target DC. Moreover, in vitro studies clearly failed to identify a soluble factor mediating the suppressive function of nTregs. However, in vivo there is growing evidence that TGF‐β and IL‐10 play an important role in the immune regulatory network orchestrated by nTregs. Thus, it is likely that in vivo nTregs provide the basis of peripheral tolerance by initially suppressing other T cells via cell contact‐dependent mechanisms, and simultaneously conveying suppressive properties to these T cells in a process termed infectious tolerance 40,41. The resulting iTregs were found to produce the suppressive cytokines TGF‐β and/or IL‐10. Hence, the secretion of these inhibitory cytokines by iTregs considerably expands the Treg‐prearranged suppressive network, thereby manifesting also far‐ranging peripheral tolerance. By virtue of all recent data, the molecular mechanisms underlying nTreg‐mediated suppression in vitro and in vivo require further investigation.

Key messages

• The enhancement of Treg function is an innovative immunotherapeutic strategy for the treatment of autoimmune, allergic, and infectious diseases and for the prevention of transplant rejection and pregnancy disorders as well. Conversely, for treatment of cancer the transient abrogation of Treg function is a desirable approach.

Preclinical mouse models

Autoimmunity

Nearly 40 years ago neonatal thymectomy was found to lead to oophoritis suggesting that thymus‐derived cells can suppress autoimmunity 42. This study provided the first experimental evidence that thymocytes and their progeny are involved in the prevention of organ‐specific autoimmune diseases. Subsequently, the transfer of nTreg‐depleted CD4⁺ T cells in T cell‐deficient mice was found to result in the development of various autoimmune diseases, among them gastritis, oophoritis, and thyroiditis, indicating the crucial role of nTregs for the maintenance of peripheral tolerance 10. These findings led to a reevaluation of results from established autoimmune models and caused a huge wave of studies to clarify the role of nTregs in autoimmune diseases.

Inflammatory bowel disease

In a murine inflammatory bowel disease (IBD) model that corresponds to human diseases morbus Crohn and colitis ulcerosa it was shown that nTregs can prevent the induction of the disease 43. The analysis of the suppressive mechanism revealed that TGF‐β as well as IL‐10 were involved. However, the influence of IL‐10 was found to be rather complex depending on the IBD model used 44. IL‐10 could be produced by nTregs as well as by other hematopoietic cells among them Tr1 and Th2 cells. Very recently, it was shown that transferred nTregs can cure intestinal inflammation 45. The progeny of these nTregs were found to be selectively enriched in the colonic lamina propria and to express IL‐10 and FoxP3 indicating the important function of nTreg‐derived IL‐10 in vivo. TGF‐β also plays a crucial role but has not to be produced by nTregs themselves since TGF‐β‐deficient nTregs are fully functional after transfer in wild‐type mice 46. However, their suppressive capacity was inhibited when TGF‐β was neutralized indicating that in vivo this cytokine is provided by non‐Treg cells under the influence of nTregs 47. In addition, colitogenic T effector cells which cannot respond to TGF‐β cannot be controlled by nTregs, underscoring the importance of TGF‐β for Treg‐mediated suppression 48. Recently, it was published that CD4⁺FoxP3+⁺ T cells which were generated in vitro in the presence of TGF‐β can prevent the induction of Th1‐mediated experimental colitis 49. In analogy to the preclinical studies, increased numbers of human Tregs could be identified in the lamina propria of patients suffering from morbus Crohn suggesting that in the human system Tregs can also play an important role for the regulation of IBD 50. Human Tregs from inflamed mucosa could be expanded in vitro indicating that such Tregs represent an attractive option for an adaptive immunotherapy of IBD 51.

Experimental autoimmune encephalomyelitis

Experimental autoimmune encephalomyelitis (EAE) is a model for multiple sclerosis and can be induced by immunizing mice with myelin oligodendrocyte glycoprotein (MOG), myelin basic protein (MPB), or proteolipid protein (PLP) leading to the development of encephalitogenic Th1 cells 52. Feeding MBP before induction of EAE could prevent outbreak of the disease via the induction of oral tolerance. Further analyses revealed that orally administered MBP induced the development of TGF‐β‐producing Th3 cells in mesenteric lymph nodes 30. Furthermore, depletion of CD25⁺ T cells could aggravate PLP‐induced EAE while the transfer of nTregs from naïve SJL (Swiss James Lambert) mice alleviated symptoms of the disease 53. Transferred Tregs from IL‐10‐deficient mice were ineffective suggesting that nTregs control EAE via an IL‐10‐dependent mechanism. These data imply that a complex interaction between nTregs, Tr1, and Th3 cells is required in combination with IL‐10‐ and TGF‐β‐dependent mechanisms to prevent autoimmune reactions directed to the central nervous system. Recently, it was found that administration of vasoactive intestinal peptide (VIP) to EAE mice results in the expansion of nTregs which inhibit encephalitogenic T cell activation. Interestingly, the suppressive mechanism of such nTregs was based partially on direct cell contact and on TGF‐β and IL‐10 as well 34. Hence, these data strongly suggest that EAE can be profoundly alleviated through preferential activation and recruitment of distinct Treg populations. Therefore, transient enhancement and expansion of T cells with suppressive potencies will certainly be a promising strategy for treatment of EAE.

Type 1 diabetes

An important animal model of human type 1 diabetes (T1D) is the nonobese diabetic mouse (NOD) which spontaneously develops autoimmunity that is caused by autoaggressive CD4⁺ and CD8⁺ T cells 54,55. These T effector cells ultimately destroy β cells of the pancreas. T1D could be induced by transfer of Th1 cells while Th2 cells were found to inhibit a diabetogenic response in NOD mice 56. Therefore, an initial strategy for treatment of T1D was the promotion of Th2 cells. Unfortunately, Th2 cells subsequently were also described to induce T1D indicating that other T cell populations have to be involved in the control of T1D 57. Meanwhile, it has been found that nTregs play a central role in NOD mice because it was shown that their decrease allows the development of T1D in such mice 58,59. Disruption of the CD28‐B7 interaction prevented the development of nTregs and resulted in exacerbated diabetes in NOD mice. Transfer of nTregs in CD28‐deficient NOD mice could prevent or at least delay diabetes. Thus, development of diabetes depends on the balance of diabetogenic T cells and nTregs and is regulated by costimulatory mechanisms. nTregs that had been antigen‐specifically expanded in vitro could very efficiently inhibit diabetes in an early phase of disease 60,61. Intranasal application of proinsulin in combination with intravenously administered anti‐CD3e F(ab')2 fragments induced iTregs and prevented autoimmune diabetes 62,63. Similar to other autoimmune models IL‐10 and TGF‐β are decisively involved in the suppressive mechanism of such iTregs. These data suggest a profound therapeutic potential for this combination of polyclonal T cell activation and antigen‐specific stimulation of iTregs and nTregs as well.

Transplantation/pregnancy

Transplantation

About half a century ago G. Snell discovered the major histocompatibility complex (MHC) and elucidated its importance for tissue transplantation originally using a tumor transplant model 64,65. Simultaneously, P. Medawar and colleagues—inspired by R. Owen's work on vascular anastomoses between bovine twins—demonstrated that skin graft tolerance can be induced by neonatal transfer of spleen cells prepared from the strain of the graft 66,67. This finding indicated that transplantation tolerance can be actively acquired. Subsequently, treatment with an antilymphocyte serum enables the survival of skin heterografts implying the involvement of T cells in transplantation tolerance 68. The description of suppressor T cells at that time also suggested a crucial role of such T cells for acceptance of transplants via manipulation of the host immune system 4,69. Alternatively, immunosuppressive agents like the calcineurin inhibitor cyclosporine A (CSA) were found to significantly improve the early survival of transplants. The analysis of mechanism of action revealed that treatment with CSA induced suppressive CD4⁺ T cells 70. In analogy, the treatment of adult mice with a combination of nonlytic anti‐CD4 and anti‐CD8 antibodies led also to tolerant T cells which, in addition, conveyed their suppressive properties to naïve T cells. Hence, these findings which were paraphrased by the expressions bystander suppression, linked suppression and infectious tolerance explained the observation that no further immune suppressive treatment was required for permanent transplantation tolerance 71,72. Subsequently, these data were doubtlessly corroborated in different models of transplantation tolerance indicating a crucial role for nTregs as well as iTregs 73–77. In a study using skin allotransplants nTregs and most possibly iTregs could also be identified in the transplant 78. After retransplantation of the skin graft to a second host these Tregs prevented rejection antigen‐specifically without hindrance of rejection of third party skin. These results demonstrate that nTregs and/or iTregs are directly active in a populated graft far beyond the lymphoid tissue. For therapeutic applications it is important that large numbers of murine nTregs can be obtained by activation and expansion in vitro. Such expanded nTregs were found to significantly inhibit a disease in a murine graft versus host disease (GVHD) model 79. Very recently human Treg cell lines could be generated from cord blood cells which strongly inhibited mixed lymphocyte reactions in vitro underscoring the therapeutic potency of human Tregs for the induction of transplantation tolerance 80.

Pregnancy

The maternal immune system has to accept a semiallogeneic fetus during pregnancy and tolerate the paternal alloantigens. Hence, pregnancy represents a state of immunological tolerance 81,82. Abortion‐prone mice were found to harbor significantly higher numbers of Th1 cells and simultaneously lower frequencies of nTregs as compared to females with normal pregnancy 83. In addition, abortion could be inhibited by transfer of nTregs from normal pregnant females, indicating the essential role of nTregs for the maintenance of transplantation tolerance during pregnancy. In addition, the number of human Tregs was elevated during early pregnancy in decidua and circulation, peaked in the second trimester, and declined postpartum84–87. Thus, these data suggest that human Tregs in analogy to murine nTregs can prevent pregnancy disorders via suppression of antipaternal immune responses.

Cancer

Besides their beneficial role in protecting the host from autoimmune responses, nTregs may unfortunately suppress adequate antitumor immune responses. Hence, in murine tumor models, transient depletion of nTregs by anti‐CD25 monoclonal antibodies (mAb) in vivo can elicit an antitumor response through tumorantigen‐specific cytotoxic T lymphocytes (CTL) and tumorantigen‐nonspecific CD4⁻CD8⁻ cytotoxic cells with natural killer (NK)‐like activity 88,89. Administration of the immunomodulator cytotoxic T lymphocyte antigen (CTLA)‐4‐Ig which acts as a competitive inhibitor of CD28‐B7 costimulation can profoundly enhance this antitumor response. These findings emphasize the hypothesis that likewise as self‐antigens, tumor cells are protected by nTregs. Moreover, tumors seem to be able to actively exploit the suppressive capacity of nTregs to escape an effective antitumor immune response. In this context, transfer of nonfractionated draining lymph node cells from tumor‐bearing mice 9 days post tumor inoculation resulted in a robust antitumor immune response against established tumors. In contrast, the same transfer on day 12 post tumor inoculation seldom prevented lethal tumor progression. As the lymph node cells from the tumor‐draining lymph node 12 days post tumor inoculation showed a four‐fold increase in cell numbers, this experiment suggested that preferentially T cells with regulatory properties were induced or that the number of Tregs was strongly increased at the tumor site 90.

Furthermore, tumor‐derived IL‐10/TGF‐β as well as tumor‐induced endogenous IL‐10/TGF‐β locally favor the conversion of conventional CD4⁺ T cells to a T cell with regulatory properties analogous to iTregs 91,92. In addition, solid tumors may produce chemokines (CCL22) that preferentially attract nTregs, which constitutively express several chemokine receptors 93. Both mechanisms—induction of iTregs and/or attraction of nTregs—will prevent efficient antitumor responses. Thus, transient depletion of nTregs and iTregs is a promising therapeutic approach to enable a curative and long‐lasting antitumor immune response. However, abrogation of tumor‐mediated suppression by depletion of Tregs is a potentially risky strategy since only little is known about the repopulation of nTregs and iTregs upon depletion. Moreover, as also tumor‐specific T effector cells may express CD25, the use of anti‐CD25 mAbs requires sophisticated kinetics and dosage studies during the course of an antitumor immune response. Recently, Treg‐affecting antitumor therapeutic strategy was ameliorated by application of nondepleting anti‐GITR mAb 94. Injection of agonistic anti‐GITR mAb into tumor‐bearing mice led to a strong antitumor response that was further improved by the simultaneous application of nondepleting anti‐CTLA‐4 mAb. Although the authors suggested that administration of anti‐GITR mAb primarily could abrogate the suppressive potency of nTregs, it is more likely that stimulation via GITR leads to a strong coactivation of antitumor‐specific T effector cells, which thereby escaped nTreg‐mediated suppression. However, this concept is not easily transferable into the human system, since the function of human GITR is not well characterized, and cross‐linking of GITR poses the risk of a polyclonal T effector cell stimulation potentially causing fatal autoimmunity.

Besides the application of anti‐CD25 mAb, administration of cyclophosphamide was assumed to primarily reduce the number of nTregs 95,96. Alternatively to direct manipulation of nTregs and T effector cell activity, immunologic suppression can be broken by conditioning of mature DC with specific toll‐like receptor (TLR)‐ligands. Several approaches clearly demonstrate that such a treatment, based on the usage of in vitro conditioned DC or through direct application of TLR ligands in vivo (that is to say reanimating an approach that was elaborated and applied by Busch, Fehleisen, and Coley a century ago 97–100) can neutralize or overcome the suppressive effects of nTregs leading to strong cytolytic activities and antitumor responses 101–103. Using a murine model of established tumor tolerance, it was found that a persistent TLR signal is needed in combination with a DC‐based vaccine for an effective antitumor response 104. As outlined below in the ‘Clinical translation’ section, Tregs were identified in many patients suffering from different tumors suggesting that the transient impairment of immune suppression by Tregs should profoundly improve specific tumor therapies.

Infection

Although Tregs seem to have far‐reaching beneficial effects on our health by influencing autoimmunity, transplantation, and even allergy, these cells always act as suppressors of immune responses, leading not only to salutary but also to detrimental effects. For example, inhibition of antimicrobial T effector cells by nTregs may lead on the one hand to severe and chronic infections, but on the other hand nTregs can prevent collateral damage of host tissue caused by vigorous antimicrobial immune responses 105,106. A very well investigated indicative of this ambivalent role of nTregs is their involvement in gastrointestinal homeostasis 107. During gastrointestinal inflammation commensal gut bacteria can enhance the pathology of such infections. However, in different mouse models of colitis nTregs play a crucial role in preventing excessive immune reactions 18,108. In addition, in an infectious mouse model using Leishmania major it was found that these parasites can be totally eradicated in the absence of nTregs 109. However, these mice did not develop a protective memory response. Obviously, in this infectious model long‐lasting immunity of the host is based on a compromise that depends on a persisting low‐level parasitic infection, which is enabled by a limited nTreg‐mediated suppression. Moreover, nTregs, migrating in response to CCR5 ligands, rapidly accumulate within chronic dermal sites of Leishmania major infection 110. Apparently, nTregs favor low‐level parasite persistence by a partial suppression of T effector cells. Thus, in the course of an antimicrobial immune response, nTregs play a very delicate and sophisticated role. During the initial phase, nTregs should ideally not inhibit an antimicrobial immune response in order to immediately allow for a full‐blown response. Recent data strongly suggest that this is accomplished by two mechanisms, induced by the invading microbes themselves. DC especially conditioned by microbial TLR‐ligands penetrate the intrinsic suppressive shield of nTregs by strongly costimulating the proliferation of T effector cells and nTregs as well 111–113. Moreover, secretion of growth factors (e.g. IL‐2) by T effector cells potentiates the suppressive capacity of nTregs and their survival/proliferation 114. Simultaneously, TLR2 ligands directly induce the proliferation of nTregs thereby transiently concealing their suppressive potency 115. In the late phase of the response—after clearance of the microbial stimulators—DC remain in a nonactivated stage and the expanded nTregs display again their suppressive properties for effector T cells and DC as well. In addition, according to results from in vitro studies, demonstrating that preactivated nTregs have a much higher suppressive capacity than freshly isolated nTregs, expanded nTregs in the late phase have even stronger suppressive properties as compared to nTregs during initiation of the microbial infection 38,116. Very recently, it was suggested that nTregs can not only suppress but actively kill their target cells 117. Thus, the successful combat of a microbial infection is a result of a dynamic regulation and modulation of Tregs, T effector cells and DC under the influence of microbial components, in particular TLR ligands.

Collectively, all preclinical models presented herein convincingly indicate that nTregs as well as iTregs are very promising targets for innovative therapeutic strategies to improve treatment of autoimmune and infectious diseases, cancer, transplant rejection, and pregnancy disorders. Figure 1 summarizes these preclinical data and illustrates future prospects for clinical applications. However, translation to the human system—as discussed below—certainly bears high risks and therefore requires great additional efforts to further characterize the distinct Treg populations, clarify their development and interrelationship, and to study their relations to other hematopoietic cells.

Figure 1. Tregs—immune modulators in health and disease. Naturally occurring CD4⁺CD25⁺ Tregs (nTregs) develop directly from CD4⁺ T cell precursors during positive selection in the thymus. Induced Tregs (iTregs) develop from naïve conventional CD4⁺ T cells either as a result of cell contact‐dependent interaction with nTregs or under the influence of suppressive agents like IL‐10, TGF‐β, dexamethasone, vitamin D3, vasoactive intestinal peptide, tolerogenic DC, and potentially other inhibitory mechanisms. Modulation of Treg (nTregs/iTregs) function is a promising intervention strategy to either improve responses in tumor and microbial diseases or suppress those unwanted in autoimmunity, allergy, transplantation, and pregnancy disorders. The transient depletion of Tregs as well as their manipulation, especially via toll‐like receptor (TLR) ligands, allow a transient reduction of Treg activity and enforce antitumor responses and immunity against viral infections. On the other hand their selective activation could diminish chronic pathological immune and autoimmune responses.

Clinical translation

As mentioned above, Tregs contribute to many immune‐mediated diseases either by inhibiting effective immune responses (e.g. cancer) or by failing to control unwanted immune responses in case of autoimmunity, allergy, certain infections, or transplantation. Therefore, manipulation of Treg activity becomes an attractive approach for immune intervention strategies.

Several evidences from cancer patients suggest that increased Treg activity is associated with an ineffective antitumor immune response. Enhanced numbers of Tregs have been observed in patients with lung, breast, pancreatic, and skin cancer either in peripheral blood or in the tumor itself 118–121. These Tregs inhibit proliferation and interferon (IFN)‐γ production of CD4⁺ and CD8⁺ effector cells and block cell‐mediated cytotoxicity of NK cells. Furthermore, the tumor tissue itself can favor the induction of Tregs by the production of suppressive cytokines such as IL‐10 and TGF‐β that convert immunostimulatory DC into tolerogenic APC. These DC further enhance the immunosuppressive state in cancer patients 122–124. The greatest barrier for Treg‐specific immunointervention of cancer is the lack of specific markers for targeting of human Tregs and to distinguish Tregs from activated effector T cells. Depletion of CD25⁺ Tregs by anti‐CD25 mAb, for example, includes the risk to deplete also tumor‐specific CD25⁺ T effector cells. Furthermore, IL‐10‐producing iTregs induced during tumor development by mechanisms of infectious tolerance or by tolerogenic DC induced by the tumor microenvironment might not be affected by this strategy. Therefore, anticancer therapies should target both nTreg and iTreg populations. An attractive alternative to direct manipulation of Treg activity in cancer patients comes from experiments showing that strong inflammatory conditions can break immunological tolerance. It has been demonstrated that conditioning of mature DC with specific TLR ligands can neutralize or overcome Treg‐mediated immunosuppression leading to strong antitumor responses 36,103,113,125. Thus, for novel anticancer immunotherapeutic strategies, transient depletion or functional inactivation of Tregs in combination with an effective, antigen‐specific immune stimulation by conditioned tumorantigen‐presenting DC is a promising therapeutic objective.

Several studies have investigated whether human autoimmune diseases are associated with defective function of Tregs. In patients with multiple sclerosis, type I diabetes, psoriasis, and rheumatoid arthritis, isolated CD4⁺CD25⁺ T cells showed a reduced suppressive activity in vitro suggesting a defect in Treg function rather than reduced ratios of Tregs that contributes to the disease in these patients 126–130. However, these results might also reflect the inability to distinguish between activated CD4⁺CD25⁺ T effector cells and CD25⁺ Tregs in the periphery and/or increased migration of functional active Tregs to the inflamed tissues. Additionally, increasing evidence exists that autoimmune diseases can develop as a consequence of an enhanced resistance of autoaggressive T cells to regulation. For example, a recent study in patients with rheumatoid arthritis showed that isolated Tregs from these patients displayed an anergic phenotype and suppressed proliferation of T effector cells in vitro. In contrast, they were unable to inhibit cytokine release of activated T cells and production of proinflammatory cytokines by monocytes in a TNF‐α‐dependent manner 130. Furthermore, Tregs derived from patients were unable to convey a suppressive phenotype to CD4⁺ T effector cells. Interestingly, in patients that received an anti‐TNF‐α therapy, these defects appeared to be restored, suggesting that anti‐inflammatory therapies can also restore the functional activities of Tregs. Additionally, as the ratios of Tregs are unchanged or increased at the side of autoimmune lesions, therapeutic intervention strategies on the basis of Treg manipulation should be concentrated to restoration of Treg activity rather than simply increasing Treg numbers. Anti‐inflammatory drugs that enhance Treg activity and might also restore the sensibility of T effector cells to regulation have been documented. Well established examples are glucocorticoids, which are able to induce anti‐inflammatory cytokines such as IL‐10 and to modify human Treg activity 131,132. Furthermore, immunosuppressive drugs can induce tolerogenic DC that are able to convey nonregulatory CD4⁺ T helper cells into IL‐10‐producing iTregs. Although most effects of the drugs are nonspecific, anti‐inflammatory treatment might improve the efficacy and functional activity of Tregs as shown for anti‐TNF‐α therapy in patients with rheumatoid arthritis. More recent reports showed the potential of another nonspecific therapy of autoimmunity using non‐Fc‐binding anti‐CD3 antibodies in patients with new‐onset type I diabetes 133,134. Patients treated with anti‐CD3 mAb early after disease onset maintained an enhanced β cell function and required lower doses of insulin compared to the control group for up to 18 months. Although the mechanisms of action are not fully known, anti‐CD3‐treated patients showed a decrease of autoaggressive T cells together with a TGF‐β‐depending increase of Tregs 135. The oldest example of an efficient immunotherapy is the antigen‐specific desensitization of allergic patients. Injection of increasing concentrations of specific allergen in selected patient groups is able to induce antigen‐specific immunological tolerance. Two recent studies 136,137 provide evidence that this tolerance is associated with the induction allergen‐specific IL‐10‐producing iTregs.

Modulation of imbalanced immunity in immune‐mediated pathologies by directed manipulation of Treg activity has the potential to work therapeutically. However, many questions remain, and strategies that specifically and efficiently target Tregs in vivo must be defined or optimized.

Acknowledgements

The authors are grateful to Drs Karen Lingnau and Matthias Klein for critical reading of this manuscript and helpful discussions. This work was supported by the Deutsche Forschungsgemeinschaft grant SFB548‐A6 (to T. Bopp, E. Schmitt) and grant SFB548‐A8 (to H. Jonuleit).

Notes

1. It is difficult, especially in the human system, to decide whether ex vivo isolated Tregs belong to the nTreg or iTreg pool. Therefore, in case of doubt, we used the general word Tregs instead of iTregs or nTregs.