Translational Mini-Review Series on Th17 Cells: CD4⁺ T helper cells: functional plasticity and differential sensitivity to regulatory T cell-mediated regulation

Abstract

CD4⁺ T cells display considerable flexibility in their effector functions, allowing them to tackle most effectively the range of pathogenic infections with which we are challenged. The classical T helper (Th) 1 and Th2 subsets have been joined recently by the Th17 lineage. If not controlled, the potent effector functions (chiefly cytokine production) of which these different cells are capable can lead to (sometimes fatal) autoimmune and allergic inflammation. The primary cell population tasked with providing this control appears to be CD4⁺ regulatory T (T_reg) cells expressing the forkhead box P3 (FoxP3) transcription factor. Here we consider the comparative capacity of FoxP3⁺ T_regs to influence the polarization, expansion and effector function of Th1, Th2 and Th17 cells in vitro and in vivo as well as in relation to human disease. This remains a particularly challenging series of interactions to understand, especially given our evolving understanding of T_reg and T effector interrelationships, as well as recent insights into functional plasticity that cast doubt upon the wisdom of a strict categorization of T effector cells based on cytokine production.

Introduction

The study of CD⁺ T cells has been greatly facilitated by their division into functional subsets. The basis for this division was the identification of distinct cytokine production profiles among T cell clones, giving rise to T helper (Th) 1 and Th2 subsets [1]. The developmental and functional relationship between these prototypic Th subsets was subject to intense study and provided the framework for classifying T cell responses for almost two decades. These ‘classical’ subsets exemplify the characteristics required to claim subset status. They can be differentiated from naive T cells under the influence of exogenous cytokines and they each express unique transcription factors which confer subset-specific expression profiles of cytokine production and effector function. In recent years T cell biology has been enriched and enlivened by the description of two further subsets. Interleukin (IL)-17-producing T cells were identified as important drivers of autoimmune pathology, forcing the re-evaluation of the role of Th1 cells in models of autoimmunity [2–4]. Elucidation of the factors promoting development of these Th17 cells [transforming growth factor (TGF)-β, IL-6 and IL-21][5–8] and the regulators of their transcriptional profile (RORγt and RORα[9,10]) established Th17 cells as a third effector T cell subset (reviewed in [11]). The three effector subsets appear to have evolved to cope with the threat posed by distinct classes of pathogen. Th1 cells are associated classically with intracellular bacteria and viral infections, Th2 responses are elicited by parasitic helminths, while Th17 responses are protective against certain extracellular bacterial and fungal infections [11]. Dysregulated Th2 responses promote the development of allergy and asthma, while uncontrolled Th1 and Th17 responses can result in autoimmune inflammation; therefore, the actions of these effector CD4⁺ cells need to be controlled strictly.

The identification of a minor subpopulation of CD4⁺ cells capable of preventing the development of autoimmunity [12,13] revolutionized our concept of T cell regulation. Identification of forkhead box P3 (FoxP3) as the lineage-specific transcriptional regulator determining this suppressive phenotype [14,15] confirmed the status of FoxP3⁺ regulatory T cells (T_regs), as distinct from previously described effector subsets [16]. In the scurfy mouse, a frameshift mutation in FoxP3 results in production of non-functional product and a lethal lymphoproliferative disorder [17,18] caused by over-activation of CD4⁺ T cells [19]. Similarly, the human condition immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) is caused by mutations of FoxP3 [20]. ‘Natural’ T_reg (nT_reg) provide the thymically derived FoxP3⁺ cells that prevent spontaneous inflammatory disease and provide the T_reg population that are assessed in vitro when using naive mice [21]. In addition, T cell receptor (TCR) stimulation of naive T cells in the presence of TGF-β can drive de novo expression of FoxP3 in uncommitted naive T cells, providing a population of ‘induced’ T_regs (iT_regs). Antigenic stimulation, therefore, can drive the polarization of naive T cells to become Th1, Th2, Th17 and/or iT_reg cells, in addition to the activation of antigen-responsive nT_regs. The balance of (and timing in the appearance of) these different populations is dependent upon the nature of the antigen presentation and the cytokine milieu.

While it is certain that T_regs form a principle component of peripheral tolerance [22], whether they are equally capable of preventing the over-expansion of Th1, Th2 or Th17 effectors is unclear. Indeed, as the subtle nuances of the intimate developmental relationships between T cell subsets continue to emerge [23,24] it becomes apparent that T_regs are not equally suppressive of all subsets or the functions thereof. In fact, in certain circumstances T_regs can promote and potentially stabilize the Th17 developmental programme [6], thus fully warranting their description as ‘regulatory’ rather than simply ‘suppressor’ cells.

Levels of suppression

Influence of FoxP3 on lineage decisions

It appears that FoxP3 can protect against pathology at various levels. Technological advances, in particular the generation of FoxP3 and RORγt reporter mice [15,25], have provided greater finesse, allowing the unequivocal identification of iT_regs[26–28] and dissection of the lineage relationships between iT_regs and Th17 cells [5]. These experiments therefore identified the possibility that ‘suppression’ could not only be mediated via the action of established T_regs on responder cells, but could also operate at the level of lineage commitment. Mice with conditional cell-specific deficiencies in targeted elements of the suppressive machinery used by T_regs are now allowing the relative importance of these elements to be addressed with increased precision [29–31]. For example, FoxP3 can interact directly with elements involved in both Th17 (RORγt) and Th2 interferon regulatory factor-4 (Irf-4) lineage commitment [25,32]. Thus FoxP3 can act to suppress inflammation directly, by physically preventing the activation of proinflammatory programmes in the cell in which it is expressed.

The TCR repertoire of T_regs is thought to be enriched for self-reactive TCRs [33]. Therefore, T_regs may represent a significant pool of autoreactive cells if they were able to gain proinflammatory effector function. Bearing this in mind, it is unclear whether the pathologies seen in the scurfy mutant or FoxP3 knock-out mouse reflect a gain of effector function by ‘T_regs’ expressing non-functional FoxP3 or from the activation of self-reactive naive T cells from the FoxP3^– peripheral repertoire. Selective depletion of FoxP3-expressing cells can be achieved by administering diphtheria toxin to mice engineered to express the human diphtheria toxin receptor in FoxP3⁺ cells [34]. T_reg depletion via this system induced the rapid onset of fatal autoimmune disease, indicating that autoaggressive T cells arising from the FoxP3^– pool are sufficient to recapitulate the scurfy phenotype. However, other studies have indicated that there is also pathogenic potential within the T_reg compartment. FoxP3 function is not binary in nature, and T_regs expressing an attenuated level of FoxP3 were found to display a reduced expression of T_reg‘signature’ genes and an increased propensity to differentiate into Th2 effectors [35]. While pathology in mice completely lacking FoxP3 is non-polarized in nature, mice expressing lower than normal levels of FoxP3 show Th2-driven pathology [35]. This may suggest that while high levels of FoxP3 expression are required to prevent Th2 differentiation, a reduced level of FoxP3 expression is still sufficient to prevent the emergence of Th1 and potentially Th17 responses. Indeed, mature T_regs in which FoxP3 expression has been ablated (due to an induced cre-mediated deletion of a floxed FoxP3 allele) develop a capacity to produce considerable amounts of IL-2, tumour necrosis factor (TNF)-α, IFN-γ and IL-17 [36]. Furthermore, upon transfer to lymphopenic hosts, T_regs in which FoxP3 had been deleted failed to show suppressive function, but rather contributed to inflammation and predominated among tissue infiltrating lymphocytes.

How is ‘suppression’ measured?

Any scientific readout is only as robust as the assay used to achieve it, and the assays used to measure suppressive potential in vitro and in vivo have different strengths and weaknesses. This must be borne in mind because, like many biological phenomena, T_reg activity in vivo cannot always be predicted accurately from their behaviour in vitro and vice versa [37–39]. The techniques used to interrogate T_reg activity have changed over time, reflecting our changing understanding of how T_regs function. The initial identification of the role of T_regs in preventing autoimmunity came from observations of autoimmune pathology in mice lacking CD25⁺ T cells [13]. Subsequently, assaying the capacity of CD25⁺ T_regs to suppress the proliferation of their CD25^– counterparts in vitro became the gold standard measurement of suppressive potential (see below [40]) and antibody-mediated depletion of CD25⁺ T cells in vivo was adopted as an imperfect but practical strategy to assess the role of T_regs in models of infection, allergy and autoimmunity [41–44]. These in vitro and in vivo experiments identified many of the suppressive pathways utilized by T_regs– IL-2 deprivation [40], expression of CTLA-4 and glucocorticoid-induced TNF receptor-related protein (GITR) [45,46], cell contact-dependent suppression [40], production of anti-inflammatory cytokines such as IL-10, TGF-β and IL-35 [31,47–51] and the expression of enzymes promoting tryptophan catabolism and adenosine production [52–54]. Throughout this time the role of T_regs was seen primarily as preventing the activation and differentiation of autoreactive T cells and the main arena for suppressive activity was considered to be the draining lymph node during naive T cell priming [39,55,56]. Their potential to modulate ongoing responses, or to display suppressive activity at sites of inflammation, was harder to address using such assays, although promising findings have been reported [57–59]. On this point, it is important to remember that T_regs can have controlling effects on inflammation through actions on a range of immune cell populations, not simply T cells. Notably, T_regs are also capable of suppressing the recruitment and function of innate immune cells [natural killer (NK) cells, dendritic cells (DCs), monocytes]in vitro and in vivo[57,60–63]. However, here we concentrate on evidence for differential sensitivity as measured by T cell effector functions.

In vitro assays of suppression

Thornton and Shevach described a co-culture system to measure T_reg-mediated suppression that not only provided important mechanistic data on the requirements for suppression, but also laid down a template for demonstrating the functional activity of T_regs. The classical suppression assay involves the co-culture of CD25⁺ T_regs and CD25^– responder T cells over a range of suppressor : responder ratios and measurement of the extent to which T_regs restrain the proliferation of CD25^– T cells [40]. There is almost no area of T_reg biology which has not been assessed by some modification of this basic technique. This assay has been used to compare the regulatory function of different subsets of T_regs[64], of in vitro-activated versus freshly explanted T_regs[65–68], of T_regs from sites of inflammation [69], of nT_regs and iT_regs[26] and of T_regs in infected versus healthy mice [70] and humans [71]. The findings of many of these studies informed further in-vivo experiments and they have greatly enhanced our knowledge of T_reg function. However, the specificity and activation status of regulatory and effector T cell populations as well as the cytokines present in the microenvironment and the activation status of antigen-presenting cells (APCs) will influence the capacity of T_regs to suppress in vivo. These conditions are often not well modelled in vitro and this caveat represents the greatest limitation of this type of assay. Particularly in mice, most often the responder population used for in vitro suppression assays are CD4⁺CD25^– T cells from naive mice, and such cells are highly susceptible to T_reg-mediated suppression. Indeed, it has been suggested that the window of susceptibility to T_reg-induced suppression in vitro is regulated tightly and restricted to the first 12 h of stimulation [72]. Limiting the proliferation or cytokine production of highly activated polarized T cells is a much more demanding task, and this may be why a clear comparison of the capacity of T_regs to limit the activity of polarized Th1, Th2 and Th17 cells is missing from the literature. It has been shown, however, that while T_regs can suppress the priming of Th2 responses, they are unable to suppress the proliferation or cytokine production of established Th2 effectors unless they themselves are pre-activated in vitro[73]. The importance of the comparative activation status of effectors and T_regs has been well illustrated. T_regs at sites of inflammation, for example, are typically more highly activated than peripheral T_regs[74,75], and this draws into question the extrapolation of functional assays carried out using mismatched responder : suppressor co-cultures and argues in favour of sampling both T_regs and effector T cells from the tissue of interest wherever possible [44,69,76]. Following this principle, data from human studies using responder and regulatory T cells from the inflamed joints versus the peripheral blood of rheumatoid arthritis or juvenile idiopathic arthritis patients demonstrated that T_regs from the site of inflammation are more suppressive than their peripheral blood counterparts but, simultaneously, the activated responder T cells from the inflamed joints are more resistant to suppression [77,78].

Modifying the classical suppression assay to measure cytokine production by the responder (CD4⁺CD25^–) T cells activated has revealed that there may be a hierarchy of suppression, with down-regulation of IFN-γ mRNA occurring earlier than suppression of Th2 cytokine production [79]. A similar study examining the transcriptional profile of T cells activated in the presence or absence of T_regs revealed down-regulation of factors promoting both Th1 and Th2 development [IL-12Rα, IL-12Rβ2 and Irf-4 as well as T-bet and GATA binding protein 3 (GATA-3)] in ‘suppressed’ T cells [80]. Notably, expression of IL-21, a Th17-associated cytokine, was also suppressed upon co-culture, suggesting that T_regs can down-regulate at least one element of Th17 effector function.

In vivo suppression

Suppression in multi-organ autoimmunity

Sakaguchi et al. reported that mice lacking CD25⁺ T cells develop exacerbated responses to non-self antigens and eventually develop various autoimmune pathologies [13]. This seminal observation implicated T_regs in governing the magnitude of immune responses and setting the threshold for the development of clinical autoimmune disease.

If T_regs are particularly important in restraining one type of effector T cell response, this might be revealed by looking at what type of pathology is most prevalent in the absence of T_regs or when their regulatory function is impaired. Several mouse models have been utilized to investigate defects in FoxP3 function with varied degrees of severity, from global impairment [18,81,82] and inducible ablation [34], to attenuated expression of FoxP3 [35], to T_reg specific-disruption of selected suppressive mechanisms [30,31] or homing mechanisms [29]. T cells from scurfy mice are hyperproliferative to TCR ligation [17] and produce higher levels of cytokines than wild-type littermates [83]. Heightened production of IL-2, IL-4, IL-5, IL-6, IL-7, IL-10, IFN-γ and TNF-α in scurfy mice indicated that components of both Th1 and Th2 responses are exacerbated in the absence of functional T_regs, while pathology results from an excessive ‘non-polarized’ response. Because IL-17 was not recognized as an important proinflammatory product of T cells at the time the scurfy mouse was characterized, levels of IL-17 were not determined. Mice with a targeted disruption of FoxP3 recapitulated the phenotype of scurfy mice displaying allergic airway inflammation and hyperproduction of immunoglobulin (Ig)E, indicative of overactive Th2 responses. However, both Th1 and Th2 cytokines were overproduced in FoxP3 knock-out mice, suggesting a non-selective dysregulation of both Th1 and Th2 responses [82].

In a systemic graft-versus-host disease-like model, co-transfer of CD4⁺CD25⁺ T_regs at the disease induction stage resulted in complete disease prevention due to suppression of pathogenic T cell priming and expansion, but this was not successful when T_regs were transferred at a later stage. Strikingly, while IFN-γ production was suppressed potently, an increase in IL-17⁺ T cells was observed [84]. These data suggest that Th17 and Th1 cells may differ in their susceptibility to T_reg-mediated suppressive signals.

Organ-specific inflammatory diseases

The pivotal influence of T_regs in determining whether a pathological autoimmune response develops following immune challenge was confirmed using T_reg depletion and reconstitution strategies in various induced models of organ-specific autoimmune disease, including collagen-induced arthritis (CIA) [85] and experimental autoimmune encephalomyelitis (EAE) [44,86–88]. In these models depletion of T_regs was associated with more vigorous immune responses and particularly increased levels of IFN-γ production [87], illustrating that T_regs suppress Th1 responses effectively which, at the time, were considered the driving force in these models. An elegant series of experiments dissecting the comparative roles of IL-12 and IL-23 in promoting autoimmunity prompted a dramatic change in emphasis, highlighting the pathogenic roles of IL-23 in promoting the expansion of IL-17-producing effector T cells and their critical importance in autoimmune inflammation [89,90]. Most studies using anti-CD25-mediated T_reg depletion strategies were carried out before the implications of these studies were realized fully. However, there is evidence that T_regs suppress production of both Th1 and Th2 responses in models of arthritis [91], and that T_reg depletion heightens production of IL-17 and IL-6 (both associated with Th17 responses) as well as IFN-γ during EAE [92]. Thus, it appears that T_regs have at least some capacity to hold the development of Th17 responses, as well as Th1 and Th2 responses, in check.

A hierarchy of receptiveness to suppression of Th 1, 2 and 17

Most models of organ specific autoimmunity are associated with definitively polarized immune responses. Unusual in this respect is autoimmune gastritis (AIG), which can be induced by Th1-, Th2- or Th17-polarized CD4⁺ T cells. Pathology in AIG is orchestrated by CD4⁺ T cells recognizing the alpha chain of the H⁺K⁺adenosine triphosphatase (ATPase) expressed in gastric parietal cells [93]. Disease can be induced in immunodeficient nude mice by transfer of antigen-specific transgenic T cells and this can be suppressed by the co-transfer of T_regs[94]. It has now been shown that while Th1, Th2 and Th17 polarized populations can all induce AIG, they differ in their pathogenicity and in their susceptibility to suppression. Th1 cells appear to be those suppressed most easily by freshly explanted polyclonal T_regs, while Th2 cells were slightly less well controlled [95]. Th17 cells, on the other hand, were the most pathogenic population tested and the pathology they induced could not be prevented by the transfer of polyclonal nT_regs[95]. The hierarchy of resistance to suppression described in this AIG model has implications for the design of T_reg-based therapies in terms of which responses can be targeted effectively by T_regs, and which type of T_regs are most appropriate for the job. This was highlighted by a further study in this experimental system, which illustrated once again the additive effects of activation status and antigen specificity in determining the capacity of T_regs to modulate autoaggressive responses. Only antigen-specific (not polyclonal) iT_reg can suppress the development of Th17-induced pathology in the gastritis model [96].

A similar pattern of responsiveness to T_reg-induced suppression has been observed in several other model systems. The ameliorative effect of all trans-retinoic acid treatment on the development of type 1 diabetes is dependent upon an expansion of FoxP3⁺ T_regs which suppress the generation of IFN-γ but not IL-17 responses [97]. We have found that T_regs isolated from the central nervous system (CNS) of mice with EAE suppress IFN-γ production efficiently by CNS-derived effector T cells in co-culture, but are unable to suppress their production of IL-17 [76]. Our own unpublished studies also suggest that polarized myelin-responsive Th17 populations are relatively resistant to T_reg-mediated suppression of their proliferation in vitro, compared to their Th1 counterparts. Consistent data from human studies show that Th17 cells are resistant to T_reg-mediated suppression at the level of proliferation [98], as well as cytokine production [99]. Extrapolation of these in vitro studies would suggest that Th17 cells might preferentially resist T_reg-mediated control of their clonal expansion in vivo. As yet, this has not been tested formally.

Factors overcoming suppression

It therefore appears that Th1 responses are perhaps the most acutely sensitive to T_reg-mediated suppression, while Th17 responses appear most resistant. The basis for differential sensitivity to regulation remains unclear. However, factors associated with Th17 responses (IL-6, IL-21, TNF-α and potentially IL-17 itself) impair the suppressive capacity of T_regs and may thus prevent suppression of Th17 responses selectively.

Several studies have presented persuasive arguments that the suppressive function of T_regs must, at times, be subverted to allow inflammatory immune responses to effectively eliminate pathogens. Central to this hypothesis is the ability of the innate immune system to sense the presence of a pathogen via Toll-like receptor (TLR) signalling and respond by producing proinflammatory cytokines such as IL-6, which overcome T_reg-mediated suppression [100]. IL-6 blockade has been shown to restrain the development of both Th1 and Th17 responses following immunization [101]. IL-6 influences the development and expansion of effector and T_reg cell responses as well as T_reg function, and this has been demonstrated most elegantly in the EAE model. IL-6 knock-out mice are resistant to EAE, and immunization with myelin oligodendrocyte glycoprotein (MOG) results in the expansion of T_regs in the absence of a Th17 response. Depletion of T_regs facilitated the emergence of an IL-17 response [102], proving that IL-6 is critical in determining the outcome of immunization. Generation of a Th17 response in IL-6 knock-out mice also established the existence of an IL-6-independent route to Th17 priming which is dependent upon the autocrine production of IL-21 by T cells [102,103]. The role of TLR-stimuli in inducing the IL-6 production that determines whether T_reg or Th17 responses develop was illustrated further by the fact that immunization with MOG in incomplete Freund's adjuvant (IFA) leads to a MOG-reactive T_reg response, while immunization with MOG in complete Freund's adjuvant (CFA) (in which heat-killed Mycobaterium tuberculosis is the source of the TLR ligands) results in Th17 polarization [104]. However, TLR-stimulation may not be required to promote IL-6 production once Th17 effector cells have been generated; therefore, effector cytokine production in the absence of infection may exacerbate inflammation, both directly and by retarding T_reg function. In this respect, production of IL-6 has been implicated in preventing efficient regulation of effector responses in the CNS during EAE [69]. Other proinflammatory cytokines that have been shown to overcome T_reg-mediated suppression are TNF-α, IL-7 and IL-15 [69,105–108], all of which have also been suggested to promote Th17 responses [6,109,110], emphasizing further the tight regulation between T_regs and Th17 cells.

Changes in the balance of effector versus regulatory T cells on a local basis precede the development of diabetes in non-obese diabetic (NOD) mice. Onset of disease correlates with a progressive decrease in the T_reg : T effector cell ratio in the inflamed islets which is not reflected in the draining pancreatic lymph node [111]. Whether this change resulted from the selective death of T_regs[111], ineffective regulatory function or resistance to regulation within the effector population was unclear. It has since been reported that the poor efficiency of T_reg-mediated suppression in NOD mice or patients with type 1 diabetes is not due to intrinsic T_reg defects, but rather to effector T cells becoming resistant to regulation [112–114]. This resistance was associated with IL-21 production by effector T cells, which could block T_reg function both in vivo and in vitro[115]. IL-21 has also been shown to prevent the TGF-β-induced expression of FoxP3 in naive T cells and favour Th17 development [116]. It seems pertinent that cytokines produced by and promoting the development of Th17 cells – IL-6 and IL-21 [117]– inhibit T_reg function.

Distinct degrees of susceptibility to a particular means of suppression may also provide the basis of differential responsiveness among effector subsets. An early model for the function of T_regs was that they competitively deprived effector cells of IL-2. In the first paper to describe the use of an in vitro system for assaying the suppressive function of T_regs it was demonstrated that T_regs suppress production of IL-2 by effector T cells and that the provision of exogenous IL-2 could overcome T_reg-mediated suppression [40]. A recent study revisited this theme, demonstrating cytokine deprivation-induced apoptosis in effector T cells co-cultured with T_regs[118]. Although IL-2 is important in supporting the expansion of Th1 cells and the differentiation and survival of iT_regs[27], it is now recognized that, at least in mice, IL-2 acting via signal transducer and activator of transcription 5 (STAT5) constrains the development of Th17 responses [119]. In this sense, a mechanism acting to suppress the development of a Th1 response could facilitate simultaneously the expansion of a Th17 response, which is supported further by the findings that IFN-γ blockade promotes Th17 responses [120,121]. Furthermore, exposure to IL-2 during T cell activation is known to predispose cells for activation-induced cell death (AICD) [122] via the up-regulation of Fas and FasL expression [122–124]. Sensitivity to AICD is enhanced by IFN-γ[125], which may underlie the increased sensitivity of Th1 cells to AICD compared to their Th2 counterparts [126]. The fate of ‘suppressed’ effectors and the comparative sensitivity of Th17 effectors to AICD deserve further study.

T cell polarization and the implications of plasticity

It is clear that T_regs can modulate both Th1 and Th2 effector responses during infection [41,127,128] as well as in models of autoimmunity and allergy [43,85,86]. However, the impact of T_regs on Th17 responses in autoimmunity and infection requires more detailed study. This may be because many of our infectious and autoimmune models were constructed and characterized during the tenure of the Th1/Th2 dichotomy and have been described consequently in its limited parlance. Even in those diseases in which Th17 cells are now considered key players (for example, CIA and EAE [129]), many experiments looking at the effects of T_regs on immune responses in vivo and in vitro were carried out before the full significance of the emerging Th17 subset was realized, and have not been revisited in its new light. Finally, and perhaps most significantly, the apparent lack of data on the regulation of Th17 cells by FoxP3⁺ T_regs may be due to our increasing recognition that these two subsets share overlapping pathways of differentiation, and it is at this level that we have focused upon T_reg/Th17 interplay. A full examination of the Th17/T_reg developmental relationship is reviewed elsewhere in this series [130,131]; however, the central observations are pertinent to the topic considered here. The primary observations that factors known to promote FoxP3 expression can also promote Th17 differentiation [5,6,132], followed by evidence that exogenous factors such as retinoic acid [133–135] and aryl-hydrocarbon receptor agonists [136–138] regulate the balance between the T_reg and Th17 generation, established an especially close relationship between these subsets. Further observations confirmed this and indicated that Th17 and T_reg responses arise in parallel, that a subset of FoxP3⁺ cells also express ROR-γt [139,140], and that ROR-γt and FoxP3 can interact directly [25,140,141] and indirectly [142] to suppress Th17 differentiation. It is now also apparent that IL-6 and IL-1β, acting via STAT3, promote a loss of FoxP3 expression and the induction of ROR-γt expression and IL-17 production in nT_regs[25,143]. Whether iT_regs are similarly prone to transdifferentiate into Th17-like cells is controversial [25,144]. The intimate relationship described between murine T_reg and Th17 cells is also present in humans, as FoxP3⁺ T_regs capable of IL-17 production have now been identified in humans [145,146]. Proinflammatory cytokines, in particular IL-1β, also promote IL-17 production by human T_regs[145–149]. It is currently unclear whether FoxP3⁺RORγt⁺ T cells retain their suppressive activity [146] or undergo a reversible loss of suppressive activity during the switch to IL-17 production [149]. What is clear, however, is that T_regs display a higher than suspected degree of phenotypic-plasticity and may at times perform proinflammatory effector functions. This is leading some authors to question their accepted status as a lineage of committed T_regs[150]. It is notable that Th17 cells also display a degree of phenotypic instability and can convert to a Th1 phenotype in a STAT-4- and T-bet-dependent fashion [151–156]. It is tempting to speculate upon the functional significance of this plasticity in relation to the anti-inflammatory properties of T_regs. If the net effect of Th17/iT_reg-inducing factors favours Th17 development during the initiation of a response, an initial wave of IL-17-producing cells generated during an acute response might be resistant to nT_reg-mediated suppression. Indeed, via production of IL-6 and IL-21 they may subvert T_reg-mediated suppression actively and facilitate expansion of Th1/Th2 polarized responses. However, if inflammation is not resolved, and the Th17 cells repeatedly re-encounter their antigen, their subsequent transition towards a Th1-like phenotype may increase their susceptibility to T_reg-mediated suppression facilitating the resolution of inflammation.

Coda

It seems almost incredible now that the Th1/Th2 paradigm sufficed to describe the majority of T cell responses for so long, and with the continuing discovery of new subsets [157,158] it appears that the mirage of the four-subset paradigm will be quick to pass. The high degree of plasticity inherent in certain phenotypes is becoming more apparent and the dynamic relationship between subsets more complex. In a situation reflective of the modern career pathway, it no longer seems reasonable to expect that a single decision early in development (no matter how well informed) will suffice to ensure that a productive, relevant and valued contribution will be made by an individual throughout their working life. Long-term success can be secured only by adaptability. It is increasingly clear that to cope with our expanding knowledge of T cell biology, immunologists must be as flexible as the cells they love to study.

Disclosure

The authors declare no conflict of interest.