Regulatory T cells and toll‐like receptors: regulating the regulators

Abstract

Regulatory T cells (Treg) play a crucial role in maintaining control of leucocytes. Several studies have shown that in vivo Treg depletion results in autoimmune syndromes like thyroiditis, gastritis, diabetes mellitus and colitis, but at the same time, may also result in improved anti‐tumour vaccination. Although Treg are recognised to maintain peripheral tolerance in healthy individuals, recent research has shown that Treg also suppress immune responses during infections to prevent tissue damage. How the Treg themselves are regulated is still under investigation. Their suppressive activity must be regulated in order to allow for the effective elimination of pathogens. Until recently, this control of Treg function was found to be through modulation via cytokines or by stimulation via co‐stimulatory molecules on antigen‐presenting cells. It is now demonstrated, however, that the presence of pathogens can be communicated to Treg directly through toll‐like receptors (TLRs). Up until now, Treg have been reported to respond to ligands for TLR2, 4, 5 and 8, and different TLRs can have alternative effects on Treg resulting in more suppression or, in contrast, abrogation of suppression. As TLRs can also recognise endogenous proteins, such as heat shock proteins, it is tempting to speculate on the role of these proteins in modulating Treg function during chronic inflammation. In this review, we will discuss the implications of TLR engagement on Treg and any consequences this may have for chronic autoinflammatory diseases like rheumatoid arthritis (RA).

The most difficult task for the immune system is to establish an optimal defence against pathogens as well as cancer, while at the same time avoiding damage to healthy tissues. In order to accomplish this task, the immune system has several checkpoints and regulatory systems available. One of the most important brakes of our immune defence system is constituted by the intrinsic regulatory T cells (Treg).

In the 1970s it was hypothesised that T cells could act as suppressor cells and inhibit immune effector cells.¹ However, when none of the so‐called suppressor genes were found, the “suppressor cell” was considered a mistake and was banned from immunology.² For a number of years suppressor T cells were not investigated until in 1995 Sakaguchi and colleagues reported that a subset of CD4 T (Treg) cells, which constitutively expresses the IL2 receptor α‐chain (CD25), exhibits suppressive capacities.³ Since then, CD4 CD25⁺ Treg cells have received enormous attention from immunologists,⁴ which resulted in the elucidation of the many mysteries that surrounded the Treg. As illustrated by the lethal IPEX syndrome in humans in which development of Treg cells is absent,⁵ lack of the suppressors is deadly. In contrast, overactive Treg may prevent effective anti‐tumour immunity. It appears that our immune system can be an uncontrolled force that needs to be guided by the suppressors. To understand the factors that control Treg functions will provide us with novel therapeutic intervention strategies for allergies, autoimmune diseases and cancer.

Treg and cancer

Excitingly, murine studies on Treg and cancer showed that limiting Treg depletion to short time periods induced “spontaneous” rejection of immunogenic tumours, without giving rise to lethal autoimmune disease.⁶,⁷ Possibly the lack of suppressive regulation by the Treg cells was sufficient for the immune system to mount an effective anti‐tumour response before newly formed Treg could again suppress the anti‐tumour T cells. In addition, we have shown that the combination of Treg depletion and CTLA‐4 blockade operated synergistically to enhance immunity to a melanoma vaccine.⁸ The potential of temporal depletion of Treg prior to vaccination is further emphasised by data showing that naive CD4 T cells can develop into IL10‐producing suppressor T cells themselves when activated in the presence of Treg (“infectious tolerance”).⁹ Strikingly, in a murine glioma model,¹⁰ Treg made up the majority of tumour‐infiltrating lymphocytes and their depletion resulted in complete tumour rejection. Curiel and colleagues¹¹ have shown that ovarian carcinomas are highly infiltrated by Treg and that increased infiltrating Treg numbers correlate with poor survival. Moreover, prevalence of Treg is increased in the peripheral blood of cancer patients,¹² as well as in the tumours of patients with invasive breast or pancreas cancers.¹³ Wang and colleagues¹⁴,¹⁵ succeeded in isolating tumour‐infiltrating Treg and identified LAGE‐1 and ARTC1 (both tumour proteins) as target antigens. Collectively, these findings could possibly explain why even in tumours found to be infiltrated with CD4 and CD8 T cells, tumour progression is seemingly unhindered.¹⁶,¹⁷,¹⁸,¹⁹,²⁰,²¹ Regulating the powerful immune‐modulating activity of Treg is anticipated to become an important tool in clinical practice tumour immunology.

Mechanisms of Treg‐mediated suppression

There has been much discussion in the literature on the exact mechanism of Treg‐mediated suppression, but it becomes clear that several mechanisms are available that equip the Treg with their suppressor capacity.²² Activated intrinsic Treg are found to strongly suppress effector T cells via a cell–cell contact‐dependent mechanism and inhibition was shown not to depend on the secretion of cytokines like TGF‐β or IL10.²³,²⁴,²⁵,²⁶ In contrast, however, to in vitro experiments, it has become clear that both IL10²⁷,²⁸ and TGF‐β can be responsible for Treg‐mediated suppression of effector T cells in vivo. Furthermore, TGF‐β has been reported to contribute to the suppression of autoimmune diabetes,²⁹,³⁰ autoimmune thyroiditis³¹ and in the generation of oral tolerance by Treg.³² Moreover, TGF‐β blocking antibodies were able to prevent tumour‐specific tolerance, resulting in complete tumour elimination.³³ Several groups have reported that membrane‐bound TGF‐β (mTGF‐β) is involved in the suppressive effect of Treg.³⁰,³⁴,³⁵,³⁶,³⁷ Of note, TGF‐β knockout mice still have some functional Treg, implicating that mTGF‐β is not the only machinery by which Treg employ their suppressor function, as IL10²⁷,²⁸ and CTLA‐4³⁸ are also reported to be involved.

Regulating the regulators

It is of vital importance to have a large pool (up to 20% of the total CD4 T cell population) of functional Treg, as is illustrated by the lethal autoimmune disease that spontaneously arises in Treg‐deficient individuals. It was found that the transcription factor Foxp3 is essential for the development of suppressor T cells.³⁹,⁴⁰ Indeed, mice and humans deficient for this transcription factor suffer from severe lymphoproliferative disease (the IPEX syndrome in man and the scurfy phenotype in mice⁴¹). The pivotal role that Treg play in immunomodulation (see above) stresses the need for a Treg to know exactly when to suppress and when to be non‐interfering with the initiation or elongation of immune responses. Treg have to distinguish a situation in which, for example, commensal bacteria are present in the gut from a pathogenic event in which bacteria are entering the body via the skin. It makes sense that one of the major discriminators of an immune response will be the local environment of the first encounter with any antigen; a response in the gut will be different from that in the skin. The local environment has a number of means to communicate with the Treg and we can now discriminate three major classes of molecules that directly regulate the Treg to alter their function: co‐stimulatory molecules, cytokines and danger signals mediated by toll‐like receptors (TLRs). Below these three Treg modulators will be discussed.

Regulation by cytokines and co‐stimulatory molecules

The regulation of Treg by cytokines and co‐stimulatory molecules has recently been reviewed and is summarised below.⁴² Various cytokines are able to influence the function of Treg directly. One of the first described to directly affect Treg function was IL2.⁴ The addition of IL2 to in vitro suppression assays led to the abrogation of suppression.⁴³ IL2 was also found to affect other aspects of Treg function, such as their development and homeostasis in vivo.⁴⁴ The importance of IL2 for Treg is illustrated by the finding that IL2‐deficient, as well as IL2 receptor‐α‐chain (CD25)‐deficient mice, fail to generate functional regulatory T cells in the periphery, lose CD4 T cell homeostasis and suffer from lethal autoimmune disease.⁴⁴,⁴⁵ IL15 is able to replace IL2 as a growth factor in vitro.⁴⁶ However, it appears not to be as important as IL2 for regulating Treg homeostasis in vivo as it was found that IL15‐deficient mice do not display lymphoid hyperplasia.⁴⁷ IL4 and IL7 are also reported to promote growth and survival, respectively.⁴⁸,⁴⁹ IL4, however, also appears to increase the Treg inhibitory function. Though IL2, 4, 7, and 15 have the ability to modulate Treg directly, it should be noted that these cytokines may affect suppression indirectly by affecting effector T cell function as well.

IL1, IL6 and IL12, produced by matured antigen‐presenting cells (APC), can signal the Treg directly to potentiate the responsiveness of Treg to IL2 and, therefore, increase proliferation.⁵⁰ IL6 and IL12 also act indirectly by releasing effector T cells from suppression by Treg.⁵¹,⁵²,⁵³ Notably, these cytokines are induced in APC upon TLR stimulation, indicating that TLR‐mediated APC maturation also indirectly releases effector T cells from suppression.

A second mechanism to control Treg activity takes place during the direct interaction between APC and Treg. Mature APC express high amounts of co‐stimulatory molecules like CD40 and CD80/CD86, and recent studies have revealed that they can induce the expansion of Treg.⁵²,⁵⁴ In addition, CD80/CD86‐deficient mice and CD40‐deficient mice contain significantly less Treg compared to wild‐type mice, suggesting an additional role of co‐stimulation in the development of the peripheral Treg pool.⁴ Besides stimulating Treg proliferation, certain co‐stimulatory molecules found on APC reduce Treg‐mediated suppression. Glucocorticoid‐induced TNFR family‐related gene (GITR), OX‐40, 4‐1BB, and RANK have all been demonstrated to prevent Treg‐mediated suppression,⁵⁵,⁵⁶,⁵⁷,⁵⁸,⁵⁹ but it is pivotal to acknowledge that co‐stimulatory molecules can also block Treg suppression by directly interfering with the effector T cells themselves. Taken together, co‐stimulatory molecules and cytokines are able to modulate Treg function, by promoting their development, increasing their expansion or by affecting their suppressive capacities.

Treg and TLRs

TLRs constitute an important family of pattern recognition receptors that are able to recognise pathogen‐associated molecular patterns (PAMPs) associated with microbes and viruses.⁶⁰ TLR signalling plays a crucial role in innate immune activation, but its role in T cell biology has long been unappreciated until four years ago, when the first comparative study on TLR expression in murine Treg versus effector T cells was reported. Both Treg and effector T cells expressed TLR1, 2 and 6. However, the Treg subset expressed significantly more TLR4, 5, 7 and 8 than effector T cells.⁶¹ Interestingly, triggering of Treg was reported to induce higher Foxp3 expression, and in co‐culture suppression experiments TLR4 and TLR5 ligands enhanced the suppression of effector T cells by Treg.⁶¹,⁶² With respect to TLR2 and 5, effector T cells also expressed significant amounts of these TLRs and the addition of stimulating TLR2 or 5 ligands resulted in more proliferation and cytokine production by the effectors.⁶²,⁶³

We recently reported a crucial role for TLR2 in regulating Treg expansion and suppression by acting directly on Treg themselves.⁶⁴ Using ultra‐pure Treg, we demonstrated that TLR2 triggering on Treg in combination with IL2 and T cell receptor (TCR) stimulation results in expansion of the otherwise non‐proliferating Treg, both in vitro and in vivo.⁶⁴ Moreover, we and others⁶⁵ showed that in the presence of TLR2 ligands, the suppressive phenotype of Treg was temporarily abrogated, thereby not interfering with the immune response in vitro and in an acute infection model in vivo. Interestingly, we further showed that TLR2 signalling cooperated with TCR triggering, potentiating the Treg response.⁶⁴ We hypothesise that antigen recognition renders the Treg more sensitive to the presence of PAMPs and a possible infection, thus allowing for an adequate immune response. Interestingly, Cohen et al.⁶⁶ showed that human Treg are modulated by TLR2 as well. They found that low concentrations of the endogenous heat shock protein 60 (Hsp60) triggered TLR2 on the Treg, resulting in more suppressive Treg in the absence of Treg proliferation.⁶⁶ It appears that low and high concentrations of Treg can have opposing effects, but the exact nature of Treg responses to different TLR2 ligand concentrations in vivo needs to be further addressed.

Peng and colleagues⁶⁷ have shown that human Treg express high levels of TLR8 and that TLR8 triggering of Treg also prevents their inhibitory phenotype. This study used specific siRNA technology to neutralise the TLR8 mRNA directly in the Treg. Their results showed that TLR8 triggering directly by the Treg, and not the effector T cells, abrogates suppression. It appeared that TLR8 stimulation of human Treg did not induce proliferation.

It is still not known how TLR triggering modulates the Treg suppressive phenotype. One explanation could be the upregulation or downregulation of Foxp3 expression by the different TLR stimulations,⁶²,⁶⁵ but how TLR signalling could alter Foxp3 expression is still unclear. One other option to explain the reduced suppressor function but enhanced proliferative capacity of Treg after TLR2 stimulation, is in line with reports demonstrating that Treg lose their ability to suppress after receiving strong co‐stimulatory signals.⁶⁸ We assume that TLR2 signalling on murine Treg may act as a strong co‐stimulatory signal. This would drive Treg into the proliferative pathway, which might coincide with a temporal reversal of suppressive capabilities. In addition, TLR2 stimulation increased CD25 expression on Treg and IL2 production by effector T cells and resulted in IL2‐mediated abrogation of suppression. Taken together, we hypothesise that TLR stimulation would enable an unrestricted immune response, resulting in the successful control of an acute infection. Once returned to the resting state, the Treg would recover their suppressive capacities.⁴² Finally, although the results from different groups may vary and different methods may be used, it appears that T helper cells express increased levels of TLRs as compared to cytotoxic T lymphocytes. In the same line, it seems that Treg express increased numbers and levels of TLRs as compared to T helper cells, but the highest expression is found on the professional APC.⁴² With respect to the regulatory role these cells play in the immune system, one can hypothesise that the level of TLR expression positively correlates with immunoregulatory potential. In other words, cells that are more involved in regulation of immunity are equipped with increased levels of TLRs (fig 1). We expect that additional comparative studies on the exact TLR expression by different T cell subsets (protein level) (including the recently described Th17 cells⁶⁹) will elucidate the direct role of TLRs on T cell subset regulation.

Figure 1 Hypothetical relationship between TLR expression and the modulatory potential of immune cells. With respect to the different numbers and levels of TLR expression, it appears that immune cells that are generally assumed to play a more prominent role in immune regulation (dendritic cells and Treg) are found to display increased TLR expression. We hypothesise that the “decision‐makers” should be equipped with the proper tools to make the right choice on the exact nature of the immune response when the body encounters a pathogen or tumour.

Treg, TLRs and chronic inflammation

In line with the above hypothesis, is the report by Belkaid and colleagues,⁷⁰ which reports that Leishmania‐specific Treg contribute to chronic infection and that increased numbers of pathogen‐specific Treg were observed near the site of the infection. We theorise that Leishmania‐derived TLR2 ligands could induce Treg expansion and abrogate suppression in the acute phase of the infection. Then the increased numbers of Treg regain their suppressive phenotype and enable the chronic persistence of the pathogen when the bulk of the pathogens (and TLR ligands) have been cleared.

Since TLRs appear to be so important for the function of Treg, one can imagine that TLRs also bind self‐proteins (eg, heat shock proteins,⁷¹) or ligands expressed by normal bacteria present in the gut, thereby contributing to the induction of tolerance and/or maintenance of Treg. Although the endogenous TLR ligand hypothesis is still controversial, several reports suggest that TLR2 and TLR4 recognise endogenous molecules.⁷¹ If the recognition of endogenous ligands is indeed occurring, TLR triggering may modulate the Treg population in a pathogenic or non‐pathogenic, stress‐induced chronic inflammatory environment that can be found, for instance, in RA (fig 2). In this hypothesis, endogenous TLR ligands may drive the inflammation through the modulation of Treg and APC, resulting in abrogation of suppression and inflammatory cytokine secretion. These cytokines result in increased autoimmune tissue damage and further release of stress factors and endogenous TLR ligands. We hypothesise that breaking this circle of inflammation through inhibition of TLR signalling may be beneficial in the treatment of autoimmunity as is already indicated by the treatment of patients with RA by a TLR2/4 antagonist.⁷² The in‐depth investigation of TLR‐mediated control of T cell responses will likely provide us with new therapeutic potential to treat autoimmune disorders and cancer.

Figure 2 TLR‐mediated progression of chronic inflammation. When during an immune response endogenous or low amounts of exogenous TLR ligands are released, these ligands may drive the continuation of inflammation by affecting Treg and APC. Treg may be affected directly by the TLRs to reduce their suppressive potential, resulting in increased immune responses. In addition, the TLR ligands induce APC maturation, cytokine secretion and expression of co‐stimulatory molecules, resulting in increased inflammatory responses. The environment of cytokines, co‐stimulation and TLR ligands may further neutralise the Treg's suppressive potential and this environment may result in additional tissue damage. Tissue damage, in its turn, will cause the release of more stress factors and endogenous TLR ligands that further contribute to the continuation of inflammation.

Abbreviations

APC - antigen‐presenting cells

PAMPs - pathogen‐associated molecular patterns

RA - rheumatoid arthritis

TCR - T cell receptor

TLRs - toll‐like receptors

Treg - regulatory T cells