Toll-like receptor signalling on Tregs: to suppress or not to suppress?

Abstract

To balance self-tolerance and immunity against pathogens or tumours, the immune system depends on both activation mechanisms and down-regulatory mechanisms. Immunologists have long been focusing on activation mechanisms, and a major breakthrough was the identification of the Toll-like receptor (TLR) family of proteins. TLRs recognize conserved molecular patterns present on pathogens, including bacteria, viruses, fungi and protozoa. Pathogen recognition via TLRs activates the innate as well as the adaptive immune response. The discovery of a suppressive T-cell subset that constitutively expresses the interleukin (IL)-2 receptor α-chain (CD25) has boosted efforts to investigate the negative regulation of immune responses. It is now well appreciated that these regulatory T cells (Tregs) play a pivotal role in controlling immune function. Interestingly, recent studies revealed that TLR2 signalling affects Treg expansion and function. This review will focus on the presence and influence of different TLRs on T lymphocytes, including Tregs, and their role in cancer.

Introduction

Induction of protective T-cell responses requires naïve T cells to receive signals via their T-cell receptor (TCR), costimulatory molecules and cytokine receptors. These signals can be provided by professional antigen-presenting cells, such as activated dendritic cells (DCs). DCs are stimulated when they encounter pathogen-associated molecular patterns (PAMPs) that are recognized by pathogen recognition receptors, such as Toll-like receptors (TLRs). TLR engagement alerts the immune system to danger and leads to the activation of innate immune cells,^1,2 which results in, for example, production of pro-inflammatory cytokines, and the induction of phagocytosis and other innate effector mechanisms. Furthermore, TLR triggering induces DC maturation, which is essential for the induction of adaptive immune responses.¹ DC maturation leads to the up-regulation of costimulatory molecules and major histocompatibility complex (MHC) molecules, secretion of immune modulatory cytokines [i.e. interleukin (IL)-12 and IL-23] and chemokines, and enhanced migration from the periphery to draining lymph nodes.^2–4 In mice and humans, 13 TLRs have now been identified that recognize distinct conserved PAMPs.^3,5 For example, TLR4 and TLR5 recognize the gram-negative bacterial component lipopolysaccharide (LPS) and the bacterial flagellin protein, respectively.^4–6 TLR2 can interact with either TLR1 or TLR6.^4,5 The heterodimer TLR1/2 recognizes bacterial triacyl lipopeptides, while TLR2/6 recognizes bacterial diacyl lipopeptides. Recently, profilin on uropathogenic Escherichia coli and Toxoplasma gondiiwas identified as the ligand for TLR11.^7,8 These bacteria-sensing TLRs are largely located on the cell surface of immune cells. In contrast, TLR3, TLR7, TLR8 and TLR9 are present inside immune cells and can recognize nucleic acids such as RNA and DNA or derivates thereof.^4–6 They sense the presence of intracellular pathogens or virally infected cells following phagocytosis. The ligands for TLR10, TLR12 and TLR13 are as yet unknown. In addition to pathogen-derived exogenous ligands, some TLRs can also become activated by recognition of so-called endogenous ligands. TLR2 and TLR4 have both been reported to interact with heat-shock proteins^9–11 and necrotic cells, and TLR3 with mRNA.¹² TLR4 has also been proposed to interact with fibronectin, fibrinogen and murine β-defensin 2.^6,13,14 Finally, TLR9 has been shown to become activated by chromatin–immunoglobulin G (IgG) complexes found in the autoimmune disease systemic lupus erythematosus (SLE).¹⁵ The finding that TLRs can recognize endogenous ligands, especially ligands that are released following tissue destruction/pathology, as well as PAMPs indicates that they are not only key molecules in immunity against micro-organisms but may also play a role in autoimmune diseases and cancer.⁶

Immune suppression: regulatory T-cell subsets

To prevent extensive immune-mediated tissue damage or autoimmune diseases, the initiation, expansion and retraction of effector T-cell (Teff) responses need to be closely controlled. For this purpose, multiple feedback control mechanisms are in place within the Teff itself, such as induction of suppressor of cytokine signalling (SOCS) genes and activation-induced cell death following T-cell activation.^16,17 In addition, Teff immune responses are highly regulated by immune-suppressive regulatory T-cell (Treg) subsets. The significance of Tregs in maintaining immune homeostasis is illustrated by the development of autoimmune symptoms in individuals lacking functional Tregs.¹⁸ Furthermore, the occurrence of autoimmunity in spontaneous autoimmune models in mice could be prevented upon transfer of Tregs,^19,20 while temporal depletion of Tregs improved cancer vaccine efficiency by enhancing Teff responses.²¹

Several distinct immunosuppressive Treg subsets have been described, which can be broadly subdivided into two groups: (i) cells that originate from the thymus,²² referred to as ‘naturally occurring Tregs’, and (ii) Tregs that have been induced in the periphery, also called ‘adaptive Tregs’. The best characterized Tregs are the naturally occurring CD4⁺CD25⁺ Tregs that constitute 5–15% of the total CD4⁺ T-cell population.²² Once activated, the CD4⁺CD25⁺ Tregs are able to suppress T-cell proliferation and cytokine production as well as antigen-presenting cell function.²³ The suppressive activity of these cells requires TCR triggering by MHC class II molecules presenting either self or non-self peptide epitopes. CD4⁺CD25⁺ Treg-mediated suppression is known to be antigen non-specific and involves cell contact-dependent mechanisms.²⁴ Naturally occurring Tregs express the transcription factor forkhead box protein (FoxP3),^25,26 which has been shown to be induced by the cytokine transforming growth factor (TGF)-β.^27,28 Recently, the AKT [also known as protein kinase B (PKB)] signalling pathway has been identified as a strong repressor of Treg differentiation in the thymus through reduction of TGF-β-induced FoxP3 expression. AKT-mediated signals thus represent a major determinant with broad impact on the onset of Treg specification.²⁹ Recently, Ono et al. found that FoxP3 physically interacts with the transcription factor acute myeloid leukaemia 1/Runt-related transcription factor 1 (AML1/Runx1), thereby preventing IL-2 and interferon (IFN)-γ production by Tregs while inducing Treg-cell-associated molecules and suppressive activity.³⁰ The importance of FoxP3 in Treg development and function was further demonstrated in FoxP3 knockout mice as well as in scurfy mice that carry a natural mutation in the FoxP3 gene. Interestingly, both mice show autoimmune symptoms resembling the human immunodysregulation, polyendocrinopathy, enteropathy, X-linked (IPEX) syndrome that is caused by a mutation in the human FoxP3 gene.¹⁸ Furthermore, transduction of FoxP3 into naïve CD4⁺ T cells resulted in a suppressive phenotype in mice.^26,31 Although several other markers have been identified in human and/or murine CD4⁺CD25⁺ Tregs, such as cytotoxic T-lymphocyte activation antigen 4 (CTLA-4),³² the glucocorticoid-induced tumour-necrosis factor receptor (TNFR) family-related gene (GITR),³³ integrin α_Eβ₇ (CD103),³⁴ CCR8 and the absence of CD127,³⁵ FoxP3 represents the most specific marker for naturally occurring Tregs available to date.

CD4⁺ Tregs that develop in the periphery are referred to as adaptive Tregs. At least two types of adaptive Tregs have been characterized. Type 1 regulatory T cells (Tr1) arise after repeated TCR stimulation in the presence of IL-10.^36,37 More recently, IL-27 produced by DCs upon interaction with FoxP3-expressing Tregs was shown to be a key factor in the generation of Tr1 cells which could be further enhanced by TGF-β.³⁸ Tr1 cells have been identified in humans and in mice, and are able to inhibit T-cell responses in vitro and in vivo.³⁶ Tr1 cells do not express FoxP3 and mediate suppression by secreting high amounts of IL-10.³⁶ Altered Tr1 function has been reported for patients with multiple sclerosis,³⁹ and adoptive transfer of Tr1 cells inhibited the development of murine experimental allergic encephalomyelitis in vivo.⁴⁰ Th3 cells represent the second subset of adaptive Tregs. They differentiate from naïve CD4 T-cell precursors through repeated TCR stimulation in combination with exposure to high amounts of TGF-β.^41,42 Th3 cells also secrete high amounts of TGF-β themselves, thereby suppressing immune responses.^41,42 Antigen-specific TGF-β-producing Th3 cells have recently been shown to be important in inducing and maintaining peripheral tolerance by driving the differentiation of adaptive antigen-specific FoxP3⁺ regulatory cells in the periphery in mice.^43,44 We note that human CD4⁺CD25⁺ Tregs expressing α₄β₇ integrin or α₄β₁⁺ have also been shown to induce Tr1-like or Th3-like T suppressor cells from naïve CD4⁺ T cells, respectively.⁴⁵

In addition to the CD4⁺ Treg subsets expressing the αβ TCR, recent studies have shown that CD8⁺CD122⁺ Tregs⁴⁶ and T cells carrying the γδ TCR can also have immune-suppressive functions. T cells expressing the γ and δ TCR chains are divided into two different subsets, namely Vδ1 (γδ1 T cells or intraepithelial lymphocytes) and Vδ2 (residing in peripheral blood).⁴⁷γδ1 T cells isolated from breast tumour tissue were shown to be highly suppressive in a trans-well system, but the soluble factor causing the suppression remains to be identified.⁴⁷ These tumour-infiltrating γδ1 T cells do not express CD25, GITR or FoxP3.

In conclusion, the field of Tregs is still expanding and new regulatory subsets are likely to be discovered in the coming years. The existence of multiple Treg subsets underscores the importance of immune-suppressive cells within the immune system, and addressing their role in tolerance and immunity will be a very active area of future research.

Suppressive mechanisms of naturally occurring Tregs

Tr1, Th3 and γδ T cells are known to exert suppression via the secretion of soluble factors, such as anti-inflammatory cytokines. Naturally occurring CD4⁺CD25⁺ Tregs exert suppression via contact-dependent mechanisms. In the past few years, multiple cell contact-dependent mechanisms exploited by these Tregs have been reported (Fig. 1). Cell contact-dependent suppression was shown to involve CTLA-4 and membrane-bound TGF-β (mTGF-β) expressed on the cell surface of CD4⁺CD25⁺ Tregs,^35,48in vitro secretion of granzyme B⁴⁹ or perforin,⁵⁰ and modulation of IL-2 responsiveness.^35,51 Consistent with the finding that neither of the aforementioned suppressive mechanisms could entirely explain Treg suppression, several new suppression mechanisms have recently been defined. In mice, CD4⁺CD25⁺FoxP3⁺ Tregs preferentially express the ectonucleotidases CD39 and CD73 on their cell surface and their expression is amplified and stabilized by FoxP3.^52,53 CD39 degrades nucleoside tri- and diphosphates such as ATP into adenosine monophosphates (AMPs) and CD73 catabolizes the conversion of AMP into adenosine. The presence of extracellular ATP is regarded as an indicator of tissue damage and can function as a natural immune adjuvant and danger signal by binding to the purinergic receptors. In contrast, adenosine is known to exert an immunosuppressive effect on immune cells, for example inhibiting the proliferation of T helper type 1 (Th1) cells and their synthesis of tumour necrosis factor (TNF) or IFN-γ,^52–54 upon binding to adenosine receptors. Thus, CD39-mediated removal of the pro-inflammatory ATP and its conversion into immunosuppressive AMP by CD73 represent another mechanism by which CD4⁺CD25⁺FoxP3⁺CD39⁺CD73⁺ Tregs can suppress immune responses.^52,53

Figure 1.

Schematic representation of the different suppressive mechanisms employed by naturally occurring regulatory T cells (Tregs). ATP, adenosine triphosphate; AMP, adenosine monophosphate; CTLA-4, cytotoxic T-lymphocyte activation antigen 4; GZB, granzyme B; IL-35, interleukin-35; PDE3B, phosphodiesterase 3B; TGF, transforming growth factor; TRAIL, tumour necrosis factor (TNF)-related apoptosis-inducing ligand.

In vitro and in vivo activation of the TNF-related apoptosis-inducing ligand (TRAIL)/death receptor 5 (DR5) pathway is a known mechanism for the induction of apoptosis. Recently, TRAIL was found to be up-regulated upon Treg activation, while activated Teffs were found to express increasing levels of DR5.⁵⁵ Addition of DR5-blocking antibodies significantly reduced the suppressive capacity of the Tregs in vitro and in vivo.⁵⁵

Proteomic analysis of human CD4⁺CD25⁺ Tregs revealed galectin-10 to be a specific intracellular marker of CD4⁺CD25⁺ Tregs.⁵⁶ Further analysis revealed that short interfering RNA (siRNA)-mediated down-regulation of galectin-10 abrogated Treg suppressive capacity, but further characterization is necessary to elucidate its exact physiological role in suppression.

The existence of another mechanism became apparent following the discovery that the cAMP-cleaving enzyme phosphodiesterase 3B (PDE3B) is strongly reduced in Tregs as compared with conventional CD4⁺ T cells. As a consequence, Tregs contain elevated levels of cAMP, a second messenger known to regulate a wide variety of cellular functions in a large group of cell types.⁵⁷ In T lymphocytes, increased levels of endogenous cAMP inhibit cell activation, cytokine production, and cell proliferation by interfering with the activation of Ras and Rap1 (both small GTPases).⁵⁷ Bopp et al. then demonstrated that contact-dependent suppression by naturally occurring Tregs can occur via a well-known mechanism, namely the intercellular transport of cAMP via gap junctions.⁵⁸ Upon Treg interaction with a target cell, cAMP levels within the target cell increased, resulting in immune suppression, which could be blocked by the addition of a gap junction inhibitor.⁵⁸ These two examples link the previously described role of nucleotide catabolites in immune regulation to the cell contact-dependent suppressive activity of Tregs.

Functional genomics analysis comparing CD4⁺CD25⁺FoxP3⁺ Tregs with CD4⁺CD25⁻ Teff also led to the discovery of the inhibitory cytokine IL-35, a novel member of the IL-12 heterodimeric cytokine family.⁵⁹ Epstein–Barr-virus-induced gene 3 (Ebi3) was shown to be preferentially expressed in the Treg subset along with the IL-12α chain (or p35) to form the heterodimer IL-35. Subsequent in vitro experiments showed that Tregs from Ebi3 knockout or IL-12α knockout mice have a significantly reduced suppressive capacity.⁵⁹ Teffs retrovirally transduced with IL-35 gained suppressive activity and also recombinant IL-35 inhibited Teff proliferation. Subpopulations of γδ T cells and CD8⁺ T cells may also express small amounts of IL-35, suggesting that IL-35 may also be involved in their regulatory function. Whether or not IL-35 can be induced in FoxP3⁺ Th3 cells is not known. The discovery of IL-35 secretion as an additional factor required for maximal Treg-mediated immune suppression is intriguing, but also raises questions regarding the contact dependence of Treg-mediated suppression and the target cells expressing the putative IL-35 receptor. The presence of multiple different suppressive mechanisms exploited by Tregs further raises the question of which suppressive mechanism(s) is most important for inhibition of a particular function in a given target cell or pathological condition.

Modulation of the Treg function

Tregs play a central role in the suppression of immune reactions and prevention of autoimmune responses harmful to the host. However, during acute infection, Tregs might hinder Teff activity directed towards the elimination of the pathogenic challenge. Therefore, Treg-mediated suppression needs to be tightly controlled. Control of Treg function is known to occur through cytokines such as IL-1, IL-6 and IL-12, and multiple costimulatory molecules expressed by antigen-presenting cells.^60,61 These cytokines and costimulatory molecules are efficiently induced upon TLR stimulation of antigen-presenting cells and act either by direct stimulation of Treg proliferation and/or inhibition of Treg suppression or indirectly by rescuing Teffs from Treg-mediated suppression. Another key cytokine that supports Treg development and maintenance is IL-2. Furthermore, IL-2 plays a dominant role in regulating Treg-mediated suppression.^61,62 IL-15, which signals through the common IL-2 receptor β and γ chains, is able to substitute IL-2 as a growth factor in vitro, while IL-4 and IL-7 can act as growth and survival factors, respectively.⁶⁰

More recent TLR expression profiling studies revealed that multiple TLRs are expressed in CD4⁺ T cells as well as CD4⁺CD25⁺ Tregs.^6,60,63–65 Interestingly, murine and human CD4⁺CD25⁺ Tregs express higher levels of TLR4, TLR5, TLR7 and TLR8 in comparison with CD4⁺ Teffs.^6,60,63,64 Several independent studies have now highlighted the importance of TLRs on CD4⁺CD25⁺ Tregs (Fig. 2). TLR5 is expressed on both CD4⁺ Teffs and CD4⁺CD25⁺ Tregs.⁶⁶ Interaction of flagellin with its receptor TLR5 on Teffs increased their proliferation and production of IL-2, while on Tregs flagellin/TLR5 increased their suppressive capacity.^63,66 The influence of TLR4 on Tregs is not yet clear. Initially, it was reported that LPS could enhance murine Treg-mediated suppression by binding to TLR4 on Tregs,⁶⁴ but this direct effect of LPS on purified Tregs could not be confirmed by others.^65–67 Nevertheless, the confined expression of TLR4 on Tregs warrants further examination of the effects of TLR4 ligands, including endogenous TLR4 ligands, on Tregs. We reported an important role for TLR2 in regulating murine Treg-mediated suppression.⁶⁷ TLR2 activation on Tregs using the synthetic ligand PAM3Cys, in combination with IL-2 and TCR triggering, can induce Treg proliferation and results in temporary loss of suppression.⁶⁷ Upon removal of the TLR2 ligand, the Tregs regained their suppressive function.⁶⁷ Experiments with TLR2 knockout and MyD88 knockout mice showed that these effects on Treg function were indeed TLR2 and MyD88 dependent.⁶⁷ Similar findings were reported by Liu et al.⁶⁸ They further suggested down-regulation of FoxP3 as a putative mechanism for the abrogation of Treg suppression. Surprisingly, using the endogenous TLR2 ligand heat-shock protein 60 (hsp60), opposite effects of TLR2 triggering on Tregs were observed by Zanini-Zhorov et al.⁹ Hsp60-activated Tregs enhanced their suppressive capacity via both cell contact-dependent mechanisms and TGF-β and IL-10 production.⁹ This discrepancy could possibly be explained by the nature of the TLR ligands used, PAM3Cys being a TLR1/2 ligand and hsp60 a TLR2/? ligand, the concentrations of the ligands used or differences in the ways endogenous and exogenous ligands interact with TLR2. TLR8 is strongly and preferentially expressed on human Tregs as compared with human Teffs.⁶⁹ Triggering of TLR8 on Tregs resulted in the specific abrogation of suppression without affecting Treg proliferation, while no effects on human Teffs were observed. Applying siRNA technology to knock-down TLR8 in Tregs completely blocked the effect, demonstrating a crucial role for TLR8.⁶⁹

Figure 2.

Effects of Toll-like receptor (TLR) ligands proposed to directly modulate the function of naturally occurring regulatory T cells (Tregs) upon direct interaction of the TLR ligand with the Treg. Pretreatment of Tregs with endogenous hsp60 (TLR2), lipopolysaccharide (LPS) (TLR4) or flagellin (TLR5) has been reported to enhance the Treg suppressive capacity. In contrast, PAM3Cys (TLR1/2) and single-stranded RNA (ssRNA) (TLR8) abrogate the suppressive capacity. The exposure of Tregs to PAM3Cys induces Treg proliferation and is mediated via TLR1/2. CpG, guanosine-containing DNA oligonucleotides; hsp60, endogenous 60-kDa heat-shock protein.

The positive and negative effects of TLR ligands on Tregs themselves are intriguing, but further research is required to completely decipher the role of TLR triggering on Tregs. Crucial questions regarding the dynamics of TLR expression on immune-suppressive Treg subsets upon inflammation or in relation to the type of pathogen encountered are largely unexplored. In addition to pathogen-derived TLR ligands, it will also be important to elucidate the impact of endogenous TLR ligands on Tregs to shed light on the multifactorial regulation of Treg homeostasis in health and disease.

Tregs, TLRs and cancer

Multiple studies have shown that immune-suppressive T cells can infiltrate tumours and dampen the anti-tumour immune response in mice.^70,71 Increased levels of Tregs have been documented in the peripheral blood of cancer patients and especially in the local tumour microenvironment.⁷² Naturally occurring CD4⁺CD25⁺ Tregs as well as adaptive CD25⁺FoxP3⁺ Tregs, Tr1 cells and Th3 cells have all been detected in tumours.⁵¹ In addition to these CD4⁺ suppressor T cells, other suppressive cell types reported to be involved in tumour immune escape are IL-10-secreting CD8⁺ cells,⁷³ invariant natural killer T (NKT) cells^74,75 and γδ T cells.⁴⁷ Wang et al. succeeded in isolating human tumour-infiltrating Tregs and identified LAGE-1 (also known as NY-ESO-1) and Ag recognized by Treg cells 1 (ARTC1) as the first natural tumour ligands for these Tregs.^76,77 Collectively, these findings could possibly explain why, even in tumours found to be infiltrated with leucocytes,^78–80 tumour progression is apparently unhindered.

The detrimental effect of Tregs in anti-tumour immunity is emphasized by murine cancer models showing that Treg depletion with monoclonal antibodies against CD25 led to significantly increased anti-cancer immunity.^81–85 Moreover, Treg depletion improves the efficacy of anti-cancer vaccines. As TLRs provide an important link between innate and adaptive immunity, TLR ligands are increasingly being used in the development of cancer vaccines.⁸⁶ However, in addition to innate immune cells and now T lymphocytes, non-immune cells such as epithelial cells and keratinocytes have been shown to express TLRs. Moreover, recent reports indicate that TLRs can also be expressed on tumour cells.^87–90 It might be especially rewarding to investigate the effects of TLR ligands on the function of different inhibitory T-cell subsets. To date, the application of TLR8 ligands has been shown to enhance the immune function of DCs and at the same time to reduce the suppressive capacity of human Tregs.⁶⁹ In contrast, the immune stimulatory potential of LPS and flagellin might be counteracted by the direct enhancement of the suppressive function of murine Tregs via TLR4 and TLR5, respectively.^63,64,66 TLR2 plays a crucial role in both Treg expansion⁶⁷ and the suppressive capacity of Tregs.⁶⁸ Hence, stimulation via this receptor may lead to temporal immune stimulation but also a substantial increase in the number of Tregs. Selecting the optimal combination of TLR ligands for a given vaccine may turn out to be a crucial component in maximizing the anti-tumour immune response.

In summary, understanding the functional control of immune-suppressive T cells, including the role of TLR signalling, may offer new opportunities to shift the balance between immunity and tolerance. This, and the identification of specific targets on immune-suppressive T cells that allow their elimination from the tumour microenvironment, represent major challenges in the development of effective cancer immunotherapy.