Origins of γδ T Cell Effector Subsets: A Riddle Wrapped in an Enigma

Two types of T cells—αβ T cells and γδ T cells—are present in vertebrates and are defined by surface expression of either the αβ or the γδ TCR complex. Both of these subsets develop in the thymus and have AgRs generated by V(D)J recombination; however, there are several key differences between αβ and γδ T cells. First, although αβ T cells are primarily localized in secondary lymphoid organs, γδ T cells are relatively rare in lymphoid organs and, instead, predominate at epithelial surfaces (1, 2). Second, αβ T cells recognize peptide ligands in the context of MHC class I and class II molecules, whereas γδ T cells do not recognize ligand in an MHC-restricted manner. Instead, γδ T cells recognize and respond to a broad range of Ags, including nonclassical MHC molecules, heat shock proteins, and lipids (3). Third, αβ T cells are naive upon exit from the thymus and acquire effector function in the periphery, but many γδ T cells acquire their effector fate during development in the thymus and exhibit limited plasticity after arrival in the periphery (4–6). Fourth, the role of various αβ T cell subsets in immune responses is relatively well understood, whereas much remains to be learned about the functional complexity of γδ T cells and its impact on immune responses. Nevertheless, recent studies demonstrated that γδ T cells can function in both adaptive and innate modes and play indispensible roles in certain immune responses. For example, γδ T cells are crucial for the eradication of certain bacterial infections, such as Nocardia asteroides and Klebsiella pneumoniae, because mice deficient for γδ T cells die upon infection, whereas wild-type mice survive (7, 8). γδ T cells are also crucial for the maintenance of epithelial barriers and the eradication of cutaneous tumors (9–11), and γδ T cell agonists have been investigated for possible therapeutic use in breast and prostate cancer (12, 13). Despite the growing appreciation for the specialized roles of γδ T cells during immune responses, the processes that influence their development and acquisition of effector function remain largely unknown.

αβ and γδ T cells arise from a common CD4⁻ CD8⁻ (double negative; DN) precursor in the thymus (14, 15). The DN compartment can be subdivided by CD44 and CD25 expression into DN1 (CD44⁺ CD25⁻), DN2 (CD44⁺ CD25⁺), DN3 (CD44⁻ CD25⁺), and DN4 (CD44⁻ CD25⁻) subsets (16). Rearrangements at the Tcrd, Tcrg, and Tcrb loci are initiated at the DN2 stage (17, 18), and αβ and γδ lineage divergence is thought to be complete upon arrival at the DN3 stage (14, 19, 20). Although the earliest stages of αβ T cell development have been well defined by surface expression of various cell markers, one of the main hindrances to analysis of the early stages of γδ T cell development is the lack of adequate cell surface molecules to distinguish subsets at various stages of development. Historically, the best marker of irreversible commitment to the γδ T cell lineage has been HSA/CD24 (21–23). Indeed, most immature γδ T cells in the thymus are CD24^hi and retain the capacity to undergo fate switching to the αβ T cell lineage (21, 22, 24); however, the ability to fate switch is lost by γδ progenitors upon downregulation of CD24. Recently, we identified an additional cell surface molecule, CD73, which is induced upon TCR stimulation and is present on a subset of immature CD24^hi γδ T cells (25). CD73 marks γδ T cell commitment, because CD24^hi CD73⁺ γδ T cell progenitors cannot undergo fate switching to the αβ T cell fate, whereas CD24^hi CD73⁻ γδ T cells still retain that capacity. Furthermore, >90% of peripheral γδ T cells are CD73⁺, suggesting that TCR stimulation is a common occurrence during γδ T cell development. Therefore, the identification of CD73 as a marker of γδ T cell commitment provides a useful tool for exploration of the molecular mechanisms that underlie γδ T cell development.

Two models have been advanced to explain the role of TCR complexes in lineage commitment: the instructional model and the stochastic model (26, 27). According to the instructional model, the common αβ/γδ T cell progenitor has the potential to adopt either lineage, and signaling through either the pre-TCR or the γδ TCR complex specifies lineage fate. Alternatively, the stochastic model posits that early T cell progenitors are preassigned to either the αβ or γδ T cell lineage, and only those progenitors that successfully rearrange the TCR complex that matches this predetermined fate are selected to survive. Several key observations have called these early models into question, and the signal strength model was proposed to reconcile these inconsistencies (28). According to the signal strength model, weak TCR signals promote adoption of the αβ fate, whereas strong signals induce γδ T cell commitment, regardless of the TCR complex from which they originate. Crucial studies using a single γδTCR transgene demonstrated that both the αβ and γδ T cell lineages could be generated by modulating the ability of the γδTCR to transduce either a weak or strong signal, respectively, and the signal strength model has become the prevailing model to describe the role of the TCR in lineage commitment (22, 29). Importantly, the concept that fate can either be predetermined prior to TCR expression or influenced by TCR signaling also was proposed to describe the acquisition of effector fate during γδ T cell development, and the evidence supporting these models will be discussed below.

One of the key differences between αβ and γδ T cells is the acquisition of effector fate during development. Although conventional αβ T cells exit the thymus in a naive state and acquire effector function in the periphery, many γδ T cells acquire their effector fate during development in the thymus and exhibit limited plasticity upon arrival in the periphery. γδ T cells can be divided minimally into three subsets based on their effector function: IL-17 producers, IFN-γ producers, and innate-like γδ T cells, which are defined either by expression of the transcription factor PLZF and simultaneous production of IFN-γ and IL-4 (30) or the ability to rapidly produce either IL-17 or IFN-γ in response to cytokine stimulation alone without TCR engagement (31, 32). γδ T cell effector subsets can be delineated by surface expression of the TNFR superfamily member CD27 (5). γδ T cells that produce IFN-γ, either in response to TCR stimulation or cytokines, are characterized by the expression of CD27, whereas IL-17–producing γδ T cells are restricted to the CD27⁻ subset. In contrast, the IFN-γ/IL-4–coproducing innate-like γδ T cell subset is characterized by coexpression of CD27 and the NK cell marker, NK1.1.

Interestingly, the IFN-γ– and IL-17–producing γδ T cell subsets are localized in different anatomical areas, with a large proportion of γδ T cells in the spleen producing IFN-γ, whereas the majority of lymph node γδ T cells produce IL-17 (4). The sublocalization of IFN-γ–producing and IL-17–producing γδ T cells is also evident within tissues themselves: dendritic epidermal T cells (DETCs), an IFN-γ–producing γδ T cell subset, localize to a distinct layer of the skin (epidermis), whereas IL-17–producing γδ T cells localize to the dermis (33). Notably, there is a high correlation between effector fate and Vγ usage. For example, Vγ2⁺ γδ T cells are enriched for IL-17 producers, whereas Vγ1⁺ γδ T cells are primarily IFN-γ producers (34). However, the IFN-γ/IL-4–producing innate-like γδ T cell subset has an even more restricted TCR repertoire, expressing Vγ1.1 Vδ6.3/6.4. How the expression of a particular Vγ-chain is linked to either anatomic location or an effector fate remains an active area of research, and several models have been proposed to explain this phenomenon.

The predetermination model for γδ T cell effector fate posits that the ultimate effector fate for γδ T cells is not influenced by TCR signaling; it is specified either prior to or concurrent with γδ TCR expression (Fig. 1A). Some observations consistent with this model have emerged recently, particularly relating to the development of IL-17–producing γδ cells. Extensive gene expression profiling of immature γδ T cells based on Vγ expression revealed that subsets defined by Vγ usage were enriched for fate-specifying transcription factors linked to the effector fate that they ultimately adopt (35). For example, the Vγ2 subset, which frequently adopts the IL-17–producing effector fate, expressed elevated levels of the transcription factor RORγt, whereas IFN-γ “biased” Vγ3⁺ progenitors expressed elevated levels of the transcription factor Egr2 (35). Further studies identified the genetic network of transcription factors required for the generation of Vγ2 IL-17⁺ γδ T cells. These cells require a network composed of Sox13, Sox4, Tcf1, and Lef1 for their generation (36). Interestingly, this is highly restricted to the Vγ2 subset, because the Vγ4 subset, which also produces IL-17, was only marginally impacted by the loss of Sox13, suggesting that these two subsets arise from distinct developmental programs. Collectively, these studies establish a correlation between expression signatures suggestive of effector fate and Vγ usage. Because the expression profiling was performed on immature CD24^hi γδ T cells, the investigators interpreted these results as being supportive of the predetermination model, because immature CD24^hi γδ T cells presumably have not yet been influenced by TCR signaling. However, it remains possible that, despite their immature phenotype, these cells may have experienced and been influenced by γδ TCR signaling.

Further studies reported that γδ T cell effector fate is tied to the environment in which γδ T cells develop. In particular, the generation of IL-17–producing γδ T cells is restricted to the fetal thymus, because the adult thymus is incapable of supporting their generation (37). Adoptive transfer approaches suggested that both the fetal thymus and fetal hematopoietic progenitors were required to support the generation of IL-17–producing γδ T cells. Moreover, the presence of IL-17 produced by αβ T cells appears to be an important determinant in this mode of developmental control, because IL-17–producing γδ T cells were able to develop in adult animals if production of IL-17 by αβ T cells was prevented. Collectively, these experiments suggested that the fetal thymus provides a unique environment for fetal thymocytes to adopt the IL-17–producing γδ T cell fate and that the IL-17⁺ γδ T cells present in adults represent a long-lived population.

The above studies suggest that the environment from which γδ T cell progenitors originate plays a crucial role in the permissiveness to adopt the IL-17–producing effector fate. This raises key questions about what factors may be influencing the generation of IL-17–producing γδ T cells. The Notch-signaling pathway is one of the key signaling pathways that has been implicated in the development of IL-17⁺ γδ T cells (38). In the absence of Hes1, a transcriptional repressor regulated by Notch, there is a severe decrease in IL-17–producing γδ T cells in the fetal thymus, suggesting that Notch signaling is crucial for the generation of IL-17⁺ γδ T cells. Numerous cytokines also have been linked to IL-17⁺ γδ T cell development. CD27⁻ thymocytes, which are highly enriched for IL-17–producing γδ T cells, have higher IL-7Rα surface expression than do their CD27⁺ counterparts (39). In line with these observations, IL-7 was demonstrated to promote the expansion of IL-17–producing γδ T cells, as well as priming them for IL-17 production following TCR stimulation. TGF-β is another cytokine that may be required for the generation of IL-17⁺ γδ T cells (40). γδ T cells within the thymus express both type I and type II TGF-β receptors, and Tgfb1^−/− mice have a reduction in thymic IL-17–producing γδ T cells but do not exhibit a defect in the generation of γδ T cells. An additional study demonstrated that IL-17⁺ γδ T cells are able to produce IL-17 in the absence of TGF-β, but the production of IL-17 was enhanced by the addition of TGF-β (5). Finally, the cytokine lymphotoxin also has been implicated in the development of IL-17⁺ γδ T cells (41, 42). The lymphotoxin β receptor is expressed on thymic γδ T cells, and double-positive thymocytes act like lymphoid tissue–inducer cells and produce lymphotoxin that acts on γδ T cell progenitors to enforce a γδ T cell genetic signature and production of IL-17–producing γδ T cells. Collectively, these reports suggest that external factors, such as cytokines and surface receptors (e.g., Notch), may influence the generation of IL-17⁺ γδ T cells. At least in some cases, these factors appear to be active prior to TCR expression, consistent with tenets of the predetermination model.

In contrast to the predetermination model, TCR-dependent models of effector fate specification have been advanced, which posit that a gradient of TCR signaling from weakest to strongest is involved in the generation of IL-17, IFN-γ, and innate γδ T cells, respectively (Fig. 1B, 1C). Efforts to test this model entailed manipulation of either the γδ TCR–selecting ligand or signaling molecules downstream from the TCR.

Although few γδ T cell ligands are known, one of the best-studied γδ T cell ligands is the nonclassical MHC class I molecule T10/T22, which requires β2-microglobulin (β2m) for its surface expression (43). To assess the role of T10/T22 ligand engagement in effector fate specification, B2m^−/− mice were used, in which surface expression of T10/22 is indirectly attenuated. One caveat is that β2m deficiency does not completely eliminate the expression of T10/22 on the cell surface (22, 29), and the low levels of T10/T22 that remain might influence the development of T10/T22-reactive γδ T cells. Interestingly, staining of thymus and spleen with a T22 tetramer revealed that the number of T10/T22-reactive γδ T cells was not markedly altered in B2m^−/− mice; however, the effector fate of the T10/T22-reactive γδ T cells was affected (4). Indeed, the T10/T22-reactive γδ T cells that developed in the presence of ligand produced IFN-γ, but those cells that developed in β2m-deficient mice in the absence of ligand primarily produced IL-17. These results suggest that the relatively weak TCR signals generated in the absence of cognate ligand were important for adoption of the IL-17–producing effector fate, whereas the more intense TCR signals induced by ligand engagement diverted the tetramer-binding progenitors to the IFN-γ–producing fate. A similar analysis was performed on the Vγ3Vδ1⁺ DETC subset, which is dependent upon the Skint1 protein for development (44–46). In mice harboring a mutation in Skint1, the number of Vγ3Vδ1⁺ DETCs in the skin is markedly reduced (44, 45). Importantly, the absence of Skint1 diverts the Vγ3Vδ1⁺ progenitors to the IL-17–producing fate and redirects their homing to the peritoneum and uterus (45, 46). These analyses also linked the nature of the TCR signal to the transcription factors required to specify effector fate. Indeed, Skint1 expression is thought to lead to strong signals that activate the Egr3 transcription factor. Egr3, in turn, induces the differentiation of these cells and their adoption of the IFN-γ effector fate, while repressing RORγt activity, which is required for the generation of IL-17–producing cells (46). Conversely, in response to weak TCR signals transduced in the absence of Skint1, RORγt expression is retained and, along with Sox13, induces adoption of the IL-17–producing effector fate. Accordingly, these data demonstrate that the effector fate of a progenitor can be altered by changes in TCR signaling and suggest that more intense signaling promotes adoption of the IFN-γ–producing fate, whereas weaker signals facilitate adoption of the IL-17–producing fate. However, whether the weak signals that facilitate development of IL-17–producing γδ T cells are truly ligand independent or involve interactions with ligands of lower affinity remains in question.

The final effector subtype for which differences in TCR signaling intensity have been implicated is the so-called “innate” γδ T cell subset. Innate γδ T cells comprise two distinct subsets that are defined based either on their expression of the PLZF transcription factor or their rapid responsiveness to cytokines without TCR stimulation. One subset of innate-like γδ T cells is NK γδ T cells, which are characterized by restricted use of the Vγ1.1⁺ Vδ6.3/6.4⁺ γδ TCR, expression of NK markers, and functional dependence (i.e., coproduction of IL-4 and IFN-γ) on the transcription factor PLZF (6, 47). Their generation is dependent upon SLAM receptor signaling, and accumulating evidence suggests that their generation also is dependent upon strong TCR signaling. Indeed, in the KN6 γδ TCR-transgenic model system, progenitors adopt the PLZF-expressing innate effector fate in response to high-affinity T10/T22^b ligand stimulation but not the lower-affinity T10^d ligand (23, 48). Moreover, high-affinity Ab stimulation is reported to induce PLZF expression in multiple Vγ subsets, which is inconsistent with the notion that the capacity to adopt the innate fate is inherently restricted to progenitors expressing a particular Vγ. PLZF also was recently identified as a direct target of the transcription factors Egr1 and Egr2, which are induced in proportion to TCR signal strength (49). Along with PLZF, the zinc-finger transcription factor ThPOK, which is also induced downstream of strong TCR signaling in γδ T cells, influences the generation of Vγ1.1⁺ innate-like γδ T cells, in that ThPOK deficiency impairs their generation, whereas ectopic expression of ThPOK increases their number (47, 48). Therefore, the IFN-γ/IL-4–producing innate-like γδ T cell subset is highly dependent on the expression of two key transcription factors that are induced upon strong TCR stimulation, suggesting a direct link between TCR signal strength and effector fate.

Recent analysis revealed a role for TCR signaling in the selection of other innate γδ T cell subsets defined by their rapid responsiveness to cytokine stimulation alone. Indeed, strong TCR signaling was implicated in the development of innate IL-17–producing and IFN-γ–producing γδ T cells that are responsive to IL-1β + IL-23 and IL-12 + IL-18, respectively (31). Two lines of evidence support the hypothesis that these innate γδ subsets depend on strong TCR signaling. First, both the innate IL-17– and IFN-γ–producing cells exhibit an activated phenotype that is typified by elevated CD45RB expression. Second, the generation of these cells is impaired by attenuating the capacity of the γδTCR to transduce signals, either by mutating a key signaling molecule (i.e., Zap70 for IL-17 producers) or mutating a presumptive ligand/TCR cofactor (i.e., Skint1 for DETCs) (31). Interestingly, although the prior TCR stimulation required for development of these innate cells appears to license them for rapid responsiveness to particular cytokines, it appears to impair their responsiveness to TCR-based stimuli. This pattern of responsiveness was observed for both the IL-17– and IFN-γ–producing cells, suggesting that this might be a common mechanism to achieve the innate-like characteristics displayed by γδ T cells. Nevertheless, it should be noted that, in response to certain stimuli, such as skin injury, DETCs are induced to produce IL-17 in response to TCR and TLR signaling, as well as IL-1β and IL-23, suggesting that the hyporesponsiveness of these innate-like γδ T cells may be overcome in some scenarios (50, 51). Therefore, although the importance of the attenuation of TCR signaling to their function remains unclear, it is clear that the nature of TCR signaling during development in the thymus is able to influence the ultimate effector fate adopted by γδ T cells.

Because there is substantial support for the signal strength model of αβ/γδ lineage commitment, which proposes that commitment to the γδ fate is dependent upon stronger/more prolonged TCR signals than is commitment to the αβ fate, an additional question relating to the role of TCR signaling in influencing the effector fate adopted by a γδ T cell is whether effector fate specification occurs coincident with γδ lineage commitment or at a subsequent, separate step (Fig. 1B, 1C). Recently, several reports suggested that lineage commitment and effector fate specification are indeed separable. Mapping of the histone methylation patterns in peripheral CD27⁺ and CD27⁻ γδ T cell subsets revealed identical histone marking on genes required for γδ T cell development, whereas loci that encode for the IFN-γ and IL-17 genes (as well as their respective key transcription factors RORγt and T-bet) had significantly different H3K4me2 patterns (52). These results are consistent with a process whereby distinct effector subsets share a common early developmental program but diverge later during development into functional effector cells. The influence of TCR signaling on the gene expression profiles of immature γδ T cells is illustrated by the elimination of Itk, a key signaling molecule downstream of the γδ TCR, which caused the immature Vγ2 subset to cluster with the Vγ1 subset, which was distinct in Itk-expressing progenitors (36). Interestingly, however, the specific transcription factors associated with the IL-17–producing effector fate (RORγt and Sox13) were not substantially affected by the loss of Itk, suggesting that their elevated expression in immature cells is either not dependent upon TCR signals or is orchestrated through a TCR-mediated, but Itk-independent, signaling cascade. This study suggests that gene signatures pre-exist in immature cells but are responsive in some cases to TCR signals, which is consistent with a developmental program where there is separation between commitment to the γδ T cell lineage and acquisition of effector fate function. Further exploration of this possibility recently was made possible through the identification of CD73, a TCR ligand–inducible molecule that appears to mark the γδ T cell lineage–committed progenitors among the broader CD24^hi immature γδTCR⁺ DN population (25). Although immature CD24^hi CD73⁻ Vγ subsets exhibited the expression signatures described above that are characteristic of the ultimate effector fate linked to a particular Vγ subset (i.e., Vγ1 with Egr2 and the potential to produce IFN-γ, or Vγ2 with RORγt and the potential to produce IL-17), this linkage was severed upon induction of CD73 and commitment of CD24^hi CD73⁺ cells to the γδ fate. Both the Vγ1 and Vγ2 subsets exhibit increased expression of Egr2 and PLZF and decreased expression of RORγt upon induction of CD73, thereby severing the link between Vγ usage and effector fate potential. The link between transcription factor expression and effector fate was then re-established upon maturation into CD24^lo cytokine-producing γδ T cell effectors. Collectively, these results are consistent with a two-step process to produce functional γδ T cells: an initial signal induces commitment to the γδ T cell lineage, whereas a secondary signal(s) through the TCR and/or other environmental cues dictates effector fate specification.

Despite the discovery of γδ T cells nearly 30 years ago, their important and distinct functions during the immune response have only relatively recently been appreciated. Because of their functional importance, there has been a growing interest in understanding the mechanisms behind the generation of γδ T cell effector subsets. There is accumulating evidence that differences in TCR signal strength play a crucial role in both the commitment to γδ T cell fate and the acquisition of effector fate. However, key questions remain regarding how differences in TCR signaling are generated and whether ligands of differing affinity play a critical and general role in this process or whether the γδ TCR complex itself is able to generate diverse signaling outputs in an autonomous manner. Furthermore, how TCR signaling cooperates with other environmental cues during development, such as Notch and cytokines, to influence effector fate specification also remains unknown. Finally, although substantial evidence exists that TCR signaling can influence effector fate, there is also evidence that certain γδ progenitors exhibit a predisposition to particular effector fates. Consequently, to unravel the riddle wrapped in enigma that surrounds the specification of γδ T cell effector fate, it will be critical for future efforts to focus on elucidating the molecular basis for establishing potential predispositions so that their dependence on and influence by differences in TCR signaling can be assessed in a rigorous manner. Because γδ T cells are increasingly understood to play critical roles in both normal and pathologic immune responses, a more complete understanding of their development is critically important, because it may enable manipulation of γδ effector generation for therapeutic benefit. Nevertheless, our understanding of human γδ T cell development is rudimentary, and it remains unclear to what extent insights gained from murine models can be extrapolated to humans. Several studies suggested that many human γδ T cells are naive upon emigration into the periphery (53, 54). Moreover, there are some suggestions that cytokines have different effects on the function of human γδ T cells than they do on murine γδ T cells. For example, stimulation of human adult Vγ9Vδ2 T cells with TCR agonists in the presence of IL-23 augments IFN-γ production, not IL-17 production as it does in mouse (55). Accordingly, it will be critical to carefully determine whether the principles governing the acquisition of effector fate by γδ T cells in model organisms also apply to human γδ T cells.

The authors have no financial conflicts of interest.