Mammalian Target of Rapamycin Integrates Diverse Inputs To Guide the Outcome of Antigen Recognition in T Cells

The two-signal model of TCR stimulation as Signal 1 and costimulation via CD28 and other receptors as Signal 2 has provided a useful paradigm for dissecting the differences in stimuli leading to T cell activation versus tolerance. Over the past two decades, it has become apparent that the outcome of Ag recognition is not merely determined by activation or tolerance; rather, there is plasticity of Th cells such that TCR engagement can lead to a variety of different CD4⁺ effector phenotypes, depending on the environmental milieu (1–5). In this regard, some have referred to cytokine exposure as Signal 3 (6). More recently, it has become apparent that other environmental cues such as nutrient availability, oxygen, growth factors, and chemokines can all make significant contributions to molding the outcome of TCR engagement. Although this broad range of signals can activate a complex array of signaling pathways, one common feature they share is an ability to modulate the activity of the evolutionarily conserved serine/threonine kinase mammalian target of rapamycin (mTOR).

In this Brief Review, we highlight the diverse inputs that can modulate mTOR activity in T cells and how this can subsequently guide the outcome of TCR engagement. In the first part of this review, we provide a general overview of mTOR signaling and the emerging role of mTOR in regulating T cell activation, differentiation, and trafficking. As there have been a number of in-depth reviews on this topic, our goal is not to exhaustively catalog these pathways (7, 8). Rather, we hope to provide a framework for the second part of this review that seeks to explore the diverse inputs that can modulate mTOR in T cells. In doing so we hope to demonstrate how: 1) known immunologic signals mediate their effects in part by regulating the mTOR pathway; and 2) environmental cues not previously associated with regulating T cell function may change the outcome of Ag recognition in part through their ability to regulate mTOR.

mTOR is a large (289 kDa), highly conserved serine/threonine kinase initially defined as the mammalian target of the natural macrolide rapamycin (9). Although initially developed as an antifungal antibiotic, rapamycin is a potent immunosuppressive agent, has been employed clinically in a wide range of transplantation procedures, and has shown great promise in several experimental models of autoimmunity (10–12). The exact mechanism by which rapamycin facilitates systemic immunosuppression is still an area of active investigation, but the compound has been shown to influence cellular proliferation, differentiation, and cytokine secretion of cells belonging to both the innate and adaptive immune systems (7).

In mammalian cells, mTOR exists as one gene but forms two distinct protein complexes: mTOR complex (mTORC) 1 and mTORC2, which differ in their inputs and substrates (Fig. 1) (13). mTORC1 consists of the regulatory-associated protein of mTOR (Raptor), mLST8, PRAS40, and DEPTOR. mLST8 and DEPTOR are also found in the mTORC2 complex, with the addition of RICTOR, mSIN1 proteins, and PROTOR (13). Upstream of the mTORC1 complex is the small activating GTPase Ras homolog enriched in brain (Rheb), the function of which is regulated by the GAP activity of tuberous sclerosis complex 1 (TSC-1) and TSC-2 (14, 15). The GAP activity of TSC-1/2 can be inhibited via phosphorylation by the kinase Akt, thereby permitting the GTP-bound form of Rheb to activate mTOR (16). The activation of Akt is facilitated by receptor-mediated activation of PI3K, which, through the production of phosphatidylinositol 3,4,5-triphosphate (PIP₃), activates phosphoinositide-dependent kinase-1 (PDK1), which in turn activates Akt. Although the activation of AKT by PDK1 has long been thought to be critical to the activation of mTORC-1, recent evidence has suggested that mTORC1 can be activated in T cells independently of AKT (17) (J.D. Powell, unpublished observations). Additionally, AKT-mediated inhibition of PRAS40 has been shown the promote mTORC1 activity independently of TSC-1/2 (18). The activity of mTORC1 is commonly assessed by measuring the phosphorylation of its substrates p70 S6-kinase and 4E-BP1 (19). mTORC1 plays a critical role in regulating mRNA translation, glucose and lipid metabolism, mitochondrial biosynthesis, and autophagy (20–23).

Although the upstream signals that regulate mTORC1 activity have been very well defined, identification of the precise signals regulating mTORC2 is still an active area of investigation. Recent studies have shown that mTORC2 is strongly and specifically activated following association with ribosomes, whereas its kinase activity is inhibited by endoplasmic reticulum stress and the glycogen synthetase kinase-3β (GSK-3β) (24, 25). Downstream targets of mTORC2 include Akt, serum and glucocorticoid-inducible kinase 1, and protein kinase C (26, 27). It should be noted that Akt acts as both an upstream regulator of mTORC1 activity (as indicated by the PI3K/PDK1-dependent phosphorylation at the T308 residue) as well as a downstream target of mTORC2 (as indicated by phosphorylation at S473 residue). Akt-dependent inhibition of TSC2 (upstream of mTORC1) does not require mTORC2 (27–29).

To specifically address the potential role of mTOR in CD4⁺ T cell differentiation, our group selectively knocked out mTOR in T cells (30). Interestingly, CD4⁺ T cells lacking mTOR fail to differentiate into Th1, Th2, or Th17 effector cells when cultured in appropriate conditions in vitro. Rather, the mTOR null T cells become Foxp3⁺ regulatory T cells (Tregs). The inability of mTOR-deficient CD4⁺ T cells to differentiate toward an effector phenotype is accompanied by decreased STAT4, STAT3, and STAT6 phosphorylation in response to IL-12, IL-6, and IL-4, respectively (30). Pharmacological inhibition of mTOR signaling in naive CD4 T cells by rapamycin treatment also facilitates the development of Foxp3⁺ Tregs, and Foxp3⁺ CD4 T cells exhibit lower levels of mTOR activity than their effector counterparts (31–34). Interestingly, although genetic deletion and pharmacological inhibition of mTOR signaling can result in the induction of a large population of Foxp3⁺ regulatory CD4 T cells in the absence of high concentrations of exogenous cytokines, this process is still dependent on the low levels of TGF-β found in serum-containing media (35).

Rapamycin has classically been held to be a selective inhibitor of mTORC1 signaling due to its avidity in a complex with FKBP12 for the Raptor component of mTORC1. However, recent data indicate that prolonged exposure to higher doses leads to inhibition of mTORC2 signaling as well (28, 36). Therefore, it has taken recent genetic approaches to clarify precise roles of mTORC1 and mTORC2 signaling in T cell effector function. Selectively deletion of Rheb in T cells specifically inhibits mTORC1 activity but maintains mTORC2 activity (28). As was the case with the mTOR null T cells, Rheb null T cells fail to become Th1 and Th17 cells when activated under appropriate culture conditions. However, somewhat surprisingly, the Rheb null T cells still maintain the ability to differentiate into Th2 cells. Conversely, examination of T cells lacking mTORC2 activity via selective deletion of Rictor reveals that Rictor null T cells fail to become Th2 cells in response to IL-4 but, unlike the Rheb null T cells, Rictor null T cells still maintain the ability to become Th1 and Th17 cells. Another group has also conditionally deleted Rictor in T cells using a different Cre transgene and likewise observed these cells fail to become Th2 cells, but interestingly, this was accompanied by a decrease in Th1 differentiation as well in this system (29). Importantly, elimination of either mTORC1 or mTORC2 signaling alone in T cells did not lead to the spontaneous generation of Tregs following activation under non-Treg culture conditions (as was seen from mTOR null T cells lacking both mTORC1 and mTORC2). These observations support the view that inhibition of both mTORC1 and mTORC2 is necessary to promote generation of Foxp3⁺ T regulatory cells. Such data suggest that the new class of mTOR kinase inhibitors (that simultaneously inhibit mTORC1 and mTORC2 activation) might prove to be potent immunosuppressive agents (37).

These data lead us to propose a model in which mTOR integrates diverse inputs to coordinate the downstream signaling programs that are responsible for regulating the ultimate outcome of Ag recognition. For example, in addition to directly regulating IL-12–induced STAT4 activation, mTORC1 also regulates the activity of the glycolytic machinery (38). Normal T cell activation has been shown to rely heavily on oxidative glycolysis (39, 40). By standing at the crossroads of these many critical signals for the activated T cell, mTOR may serve as a biochemical traffic cop to coordinate the development of effector T cells.

CD8⁺ T cell Ag recognition leads to a marked increase in proliferation along with a switch from catabolism to anabolism and an increase in glycolysis, similar to that seen in CD4⁺ T cells (41). CD8⁺ effector generation requires increased protein synthesis; thus, it is not surprising that Ag recognition in CD8⁺ T cells leads to both mTOR- and MAPK signaling-induced S6 phosphorylation (42). If this is blocked by inhibition of mTOR, the consequence is actually promotion of memory CD8⁺ T cell generation. In a lymphocytic choriomeningitis virus model, it was shown that low-dose rapamycin treatment during infection promotes the generation of memory T cells (43). Similarly, long-lived memory cells could be generated by culturing lymphocytic choriomeningitis virus-specific T cells with rapamycin and then adoptively transferring them into mice (44). Rao and colleagues (45) were able to demonstrate that treating CD8⁺ T cells with rapamycin promoted memory generation in part by inhibiting T-bet expression and facilitating the expression of eomesodermin. Likewise, in a model of homeostatic proliferation-induced memory, this group was able to show that blocking mTOR with rapamycin abrogated the need for IL-15 signaling in upregulating eomesodermin and thus promoting memory (46).

Metabolically, rapamycin-treated CD8⁺ T cells demonstrate an increase in oxidative phosphorylation (44). Along these lines, Pearce et al. (47) observed that when TNFR-associated factor 6 was specifically deleted in T cells, CD8⁺ memory cell generation was markedly impaired. The failure of the effector cells to transition into memory cells was associated with an inability to switch to catabolism relating to fatty acid oxidation. Based on these observations, they went on to show that activating AMP-activated protein kinase (AMPK) with metformin or inhibiting mTOR with rapamycin led to an increase in fatty acid oxidation and a consequent increase in memory generation.

The ability of naive T cells to circulate through secondary lymphoid tissue is facilitated by the expression of a number of cell-surface receptors, including CD62L and the chemokine receptor CCR7 (48). Mechanistically, the expression of CD62L, CCR7, and the memory marker IL-7Rβ (CD127) has been linked to the FOXO family of transcription factors and Kruppel-like factor 2 (KLF2) (49, 50). mTORC2 activation of Akt, inhibits activation of the FOXOs, leading to decreased KLF2 expression (51, 52). Because KLF2 positively regulates the transcription of these trafficking molecules, the expression of CD62L and CCR7 declines upon mTOR and Akt activation. In this regard, a critical role for Akt in the regulation of CD8⁺ T cell trafficking has been described (17). Likewise, the G protein-coupled receptor sphingosine 1-phosphate receptor 1 (S1P1) is also regulated by KLF2 (53). S1P1 plays a critical role in promoting T cell egress from lymph nodes (54). Mechanistically, the regulation of these homing molecules by mTOR serves to coordinate activation status with trafficking out of lymphoid tissues.

As shown above, a critical role is emerging for mTOR in integrating signals and regulating the outcome of Ag recognition in T cells. In the second part of this review, we highlight the diverse array of environmental cues that can regulate mTOR in T cells. By examining these inputs, summarized in Table I, two interesting themes emerge. First, it is clear that a number of well-established immunologic mediators, CD28 and programmed death-1 (PD-1), for example, exert their effects in part by regulating mTOR activity. Second, there is mounting evidence that nutrient availability and metabolic regulators play a critical role in directing T cell differentiation and function in part by their ability to regulate mTOR (51, 55).

The specificity of the TCR cannot distinguish between self- or pathogen-associated peptides. However, the concentration of the peptide presented by an APC, as well as the affinity of the peptide–TCR interaction, can convey biochemical information that can influence the outcome of Ag recognition. For example, lower affinity or altered peptide ligands can lead to the induction of T cell anergy (56). Likewise, it has been shown that low doses of peptide can promote Th2 responses in the absence of skewing cytokines and very low doses of peptide can promote the generation of Foxp3⁺ Tregs (57, 58).

PI3K activation is downstream of TCR engagement, and thus, Ag recognition can in fact lead to mTOR activation (59). However, when compared with mTOR activation induced by CD28 engagement, TCR-induced mTOR activity is relatively weak and short-lived. Nonetheless, the modulation of PI3K, and hence mTOR via the strength of TCR stimulation can result in functional consequences. For example, the ability of low-dose Ag to induce Foxp3⁺ T cells has been attributed in part to weak TCR-induced mTOR activity (35, 58). This is particularly prominent when immature dendritic cells are used as APCs (60). Similarly, it has been shown that premature termination of TCR engagement promotes Foxp3⁺ expression due to antagonized PI3K–mTOR signaling (34). Katzman et al. (61) have been able to correlate the duration of TCR signaling with the induction of T cell activation or tolerance. In their model, short-lived T cell–APC interactions leading to tolerance are correlated with decreased mTOR activation.

Although it is now clear that the Signal 2 is in reality comprised of multiple ligand receptor interactions, perhaps the best-described costimulatory signal on T cells is the interaction between CD28 and its two known ligands B7.1 and B7.2. CD28 facilitates the nuclear translocation of NF-κB and enhances transcription and translation of IL-2 (62). Thus, one means CD28 ligation can promote mTOR activity is in an autocrine fashion through IL-2 signaling. The ligation of CD28 on an activated T cell can also directly activate PI3K. PI3K binds the phosphorylated cytoplasmic tail of CD28 at a conserved YMNM motif and mediates Akt activation (63). Ab-mediated ligation of CD28 can induce Akt activity independently of TCR stimulation (64), and constitutively active Akt can overcome the inability of CD28-deficient cells to secrete IL-2 but cannot restore their proliferative capacity (64).

The sustained activation of PI3K and mTOR resulting from CD28 activation has been shown to promote proliferation in T cells independently of IL-2 production (65). This is the consequence of optimal expression of cyclin D3 and downregulation of the cell-cycle inhibitor p27 (66). In addition to T cell activation, CD28-mediated costimulation plays an important role in enhancing glycolysis and glucose uptake (67). This process has been shown to be dependent on PI3K/Akt signaling and involves the rapid upregulation in expression of the Glut1 glucose transporter (67).

The costimulatory signal provided by CD28 ligation on naive T cells is important for the initiation of a T cell response, but additional receptor–ligand interactions can also provide a costimulatory signal and fine-tune the T cell activation profile at the time of initial activation. The ICOS/ICOS ligand interaction is a potent inducer of PI3K activation. In fact, studies suggest that the direct binding of PI3K to the conserved YMFM motif on the ICOS cytoplasmic tail leads to more robust activation than that induced by CD28 engagement (68). Detailed studies examining the role of ICOS on regulating mTOR activity in T cells have yet to be performed. However, given the prominent role that ICOS plays in PI3K activation in T cells, one would predict that ICOS will also play an important role in regulating mTOR.

The surface receptor OX40 (CD134) has recently gained recognition as a potent costimulatory molecule that complements the activity of CD28 and ICOS. A member of the TNF-α receptor superfamily, OX40 expression is strongly, though transiently, induced following TCR stimulation in both CD4 and CD8 T cells, peaking in expression 48 h after stimulation and returning to baseline by 120 h (69). Ligation of this receptor, either through interaction with its APC-restricted ligand OX40L (CD252) or by Ab-mediated cross-linking, facilitates increased clonal proliferation, cell survival, cytokine secretion, and memory generation (70). In part, these effects are mediated by the ability of OX40 to stimulate the activity of PI3K activity, thereby promoting AKT activity upstream of mTOR (71–73). Interestingly, ligation of OX40 on the surface of naive T cells facilitates the generation and proliferation of Foxp3⁺ Tregs. However, regulatory cells generated under OX40 stimulation are poorly suppressive and display an exhausted phenotype, which can be reversed with IL-2 treatment (71).

CTLA-4 is an inhibitory member of the CD28 receptor family. CTLA-4 ligation can lead to decreased mTOR activity by inhibiting IL-2 production and hence autocrine IL-2–induced mTOR activity. From a signaling perspective, the mechanism by which CTLA-4 ligation inhibits T cell activation is complex and incompletely understood. Ligation of CTLA-4 on the surface of T cells following TCR/CD28 stimulation does not result in a reduction in PI3K activity, but does reduce Akt phosphorylation in a process that appears to be dependent on the phosphatase PP2a (74, 75). However, other studies have shown that CTLA-4 ligation induces PI3K and Akt activation that in turn inhibits apoptosis and thus sustains T cell anergy while simultaneously preventing cell death (76).

The surface receptor PD-1, and its associated ligands programmed death ligand-1 (PD-L1; B7-H1) and PD-L2 (B7-DC), provide another inhibitory counterbalance to the costimulatory signals induced by the interaction of CD28 and its ligands B7.1/B7.2 or ICOS with ICOS ligand (77, 78). As is the case for CTLA-4, the mechanism by which PD-1 modulates T cell activation, effector differentiation, and the development of Tregs is multifaceted and incompletely understood. Association of SHP-1 and/or -2 to the immunoreceptor tyrosine-based switch motif of the cytoplasmic tails of PD-1 can directly antagonize TCR-induced phosphorylation of ZAP70 (79, 80). The exposure of CD4⁺ T cells to PD-L1–coated microbeads has been shown to result in an increase in expression of the phosphatase PTEN, which antagonizes PI3K/mTOR function by facilitating the degradation of PIP₃ (81, 82). PD-1 can inhibit the PI3K–Akt axis by preventing CD28-mediated activation of PI3K (74). Additionally, it has been shown that the ability of PD-1/PD-L1 interaction to promote the development, maintenance, and function of inducible Tregs is dependent upon the inhibition of mTOR (81). That is, the ability of PD-L1 to promote inducible Tregs is mediated through the downregulation of the Akt–mTOR axis signaling.

mTOR signaling plays a role in regulating the downstream consequences of a number of immunologically relevant cytokines. Early studies identified mTOR activity as being increased upon IL-2–induced stimulation (83). IL-2–induced mTOR activation was shown to be important for facilitating cell-cycle progression and proliferation (83). These observations led to a series of studies examining the ability of mTOR to regulate T cell anergy (65, 84–86). It has been shown that the ability of IL-2 to both prevent and reverse T cell anergy is dependent upon mTOR activation (85, 87). Other common γ-chain cytokine receptors also activate mTOR. Like IL-2, IL-4R signaling is another potent inducer of T cell proliferation but has the added ability to skew naive CD4 T cell to a Th2 phenotype. The cytoplasmic tail of the IL-4R possesses five evolutionarily conserved tyrosine residues that have been shown to differentially regulate STAT5 and PI3K activity (88). The loss of the Y1 residue inhibits the ability of IL-4 treatment to induce PI3K activity and downstream mTOR activation, but leaves intact the ability of IL-4 to induce STAT5 and STAT6 phosphorylation (88). The ability of the IL-4R to induce STAT5/6 activity appears to be dependent on the Y2-4 residues on the cytoplasmic tail and acts independently of PI3K/mTOR signaling (88).

The IL-7R also activates the PI3K/Akt/mTOR axis (89). IL-7 plays an important role in maintaining T cell metabolism and survival. Interestingly, it has been shown that the ability of IL-7 to promote Bcl2 expression is mTOR independent (90). In contrast, IL-7R–induced increases in size and glucose metabolism are dependent on mTOR signaling.

IL-1R–dependent mTOR activation has recently been shown to be indispensible for the generation and proliferation of Th17 CD4 T cells (91). Gulen et al. (91) have demonstrated that Th17 differentiation induces the expression of SIGIRR, a negative regulator of IL-1 signaling that acts as a damper to continued IL-17 secretion. The deletion of SIGIRR results in an increase in IL-17 production under Th17 culture conditions and a corresponding increase in mTOR activity. Importantly, the T cell-specific deletion of mTOR negates the ability of IL-1 treatment to enhance Th17 proliferation. With regard to CD8⁺ effector generation, IL-12 has been shown to prolong mTOR activation upon stimulation (45). This in turn leads to an increase in T-bet expression. Likewise, both type I and type II IFNs have been shown to induce mTOR activity via PI3K activation (92–94). Stimulation of type I IFN receptors results in the rapid phosphorylation of insulin receptor substrates 1/2, resulting in the recruitment of the p85 regulatory subunit of PI3K and the induction of downstream Akt and mTOR activity.

Chemokine receptors regulate cellular migration primarily through the β/γ subunits of the G-protein coupled receptor’s activation of PLCγ2/γ3 and PI3K (95, 96). The link to mTOR was made by the observation that the addition of rapamycin can inhibit the migration of neutrophils in response to GM-CSF, as well as smooth muscle cells in response to fibronectin (97, 98). Subsequently, it has been shown that many G-protein coupled chemokine receptors rely on mTOR signaling for at least some aspects of their migratory effects. Naive T cells use mTOR signaling to respond to CXCL12 stimulation (99). For activated Th1/Th2 CD4 T cells, mTOR activity is required for CCR5/CCL5 (RANTES)-mediated migration that is dependent upon 4EBP-mediated translation (100). However, not all chemokines depend on mTOR activation. For example, mTOR signaling is dispensable for CCL19 (Mip3β)-mediated migration (99).

Although best known for regulating appetite and energy expenditure, the adipokine leptin also plays a significant role in regulating the functionality and proliferative capacity of T cells through its ability to stimulate mTOR activity (101–103). In the absence of leptin receptor stimulation, autoreactive CD4⁺ T cells exhibit decreased expression of the antiapoptotic factor Bcl2, an impaired ability to skew to a Th1/Th17 phenotype, and a failure to upregulate mTOR activity (103). Further, leptin acting via the mTOR signaling pathway has been shown to provide a link between energy status and Treg function (104).

The lysophospholipid S1P is another potent inducer of mTOR activity in T cells via its G-protein coupled receptor S1P1 (105, 106). Although S1P1 signaling is able to induce mTOR activity in T cells, the receptor facilitates its own downregulation due to the ability of mTORC2 activity to suppress the activity of the transcription factor KLF2 (48, 107). S1P1 signaling has canonically been thought to regulate T cell migration from the thymus and secondary lymphoid organs (54). However, it has recently been recognized that S1P1-dependent modulation of mTOR activity plays a critical role in regulating CD4⁺ T cell differentiation and the functionality of Tregs (105, 106). Overexpression of S1P1 in CD4⁺ T cells facilitates the development of Th1-polarized cells while inhibiting Foxp3⁺ Treg development in an mTOR-dependent process. Conversely, the deletion of the S1P1 receptor facilitates Treg development and enhances their suppressive capacity (105).

Lack of nutrients or oxygen deprivation all lead to the inhibition of mTOR activity (23). A cell normally maintains a very high intracellular ATP to AMP ratio. Increased AMP activates AMPK, directly phosphorylating the TSC1/2 complex, thereby increasing its GAP activity and decreasing Rheb-dependent mTORC-1 activity. In addition, activation of AMPK can result in the direct phosphorylation of Raptor, inhibiting mTORC1 activity in a TSC1/2-independent fashion (108). Pharmacologic activation of AMPK by AICAR inhibits T cell function and has been shown to block the induction of experimental autoimmune encephalomyelitis and promote anergy by inhibiting mTOR (109–111). Likewise, activation of AMPK, by the glucose analog 2-deoxyglucose (2-DG), leads to the inhibition of mTOR (109, 112, 113). 2-DG is readily taken up by T cells via the GLUT-1 transporter; however, 2-DG–6-phosphate cannot be processed further by the cellular glycolytic machinery and therefore competitively inhibits the process of glycolysis. Given the well-defined role for mTOR inhibition in facilitating the development of memory T cells, one might hypothesize that many of the clinically approved AMPK agonists, such as metformin, may turn out to facilitate the generation of memory T cells.

The phosphorylation and activation of the GAP activity of TSC can also be promoted by GSK-3β (114). This is a mechanism of action analogous to that observed for AMPK-mediated mTOR inhibition, and it appears that the phosphorylation of TSC2 by GSK-3β is dependent on prior phosphorylation of the substrate by AMPK at the S1345 residue (114). The interaction of Wnt with its receptor on the plasma membrane of most mammalian cells inhibits the activity of GSK-3β. As such, Wnt signaling can promote mTORC1 activity.

Low oxygen tension (as might be experienced in a tumor microenvironment) can also regulate mTOR activity. It has been shown that in the setting of low oxygen, the hypoxia-induced factor protein regulated in the development of DNA damage response 1 (REDD1) can inhibit mTOR by promoting the assembly and activation of TSC (115). Cells lacking REDD1 show continued mTOR activity even under conditions of nutrient withdrawal (116), whereas hypoxia facilitates the REDD1-mediated activation of the TSC1/2 by facilitating the stabilization of TSC1/2 by 14-3-3 protein (117). Although hypoxia is a potent regulator of mTOR activity, mTORC-1 regulates the expression of the canonical hypoxia response element hypoxia inducible factor-1 (118, 119). Hypoxia inducible factor-1 expression has recently been shown to facilitate the development of Th17 CD4 T cells via the formation transcriptionally active complex with RORγT and the induction of a highly glycolytic metabolic phenotype while simultaneously inhibiting the development of Tregs by facilitating the degradation of Foxp3 (120, 121).

Availability of amino acids also regulates mTOR activity. Specifically, branch chain amino acids such as leucine promote mTOR activity (122). This is accomplished by promoting the interaction between Rheb and mTORC1. The ability of branch chain amino acids to activate mTOR has immunologic consequences. For example, Tregs can facilitate the generation of infectious tolerance in part by depleting branch chain amino acids, leading to mTOR inhibition and further Treg generation (123). Likewise, it has been shown that the leucine analog NALA can inhibit T cell function, and TCR engagement in the presence of NALA promotes T cell anergy by inhibiting mTOR (109, 124)

Although the two-signal model provides a framework for understanding the generation of the adaptive immune response, it is clear that the inputs that influence the outcome of Ag recognition are varied and complex. Likewise, there is a greater appreciation for the diversity of outcomes upon TCR engagement. In this regard, mTOR has emerged as a critical integrator of environmental cues in T cells. Concomitant with our increasing appreciation for mTOR to influence T cell activation, differentiation, and tolerance is a greater appreciation for the diversity of environmental inputs that can influence these processes by regulating mTOR. In a number of cases (for example CTLA-4), a connection between receptor ligand interaction and PI3K signaling has been made but the precise connection to downstream mTOR signaling has yet to be defined. Nonetheless, the role of a diversity of inputs in regulating mTOR and the increasing role of mTOR in regulating T cell function suggest that these pathways may prove to be potent pharmacologic targets for suppressing, redirecting, and enhancing T cell responses.

We thank Emily Heikamp, Sam Collins, and Kristen Pollizzi for suggestions and Christopher Gamper for invaluable editorial assistance.

The authors have no financial conflicts of interest.