When the T cell receptor (TCR) is engaged, tyrosine phosphorylation of the TCR-associated CD3 chains recruits kinases and scaffold proteins leading to the formation of a supramolecular complex that promotes activation of signaling cascades, generation of second messengers, and initiation of transcriptional events, which lead to T cell differentiation programs (1). These signaling pathways synergistically promote glycolysis and anabolic metabolism to support T cell clonal expansion and effector cell generation (2–4). It is now well understood that metabolic mediators function as intermediates between these signaling events and the functional outcome of T cell activation. Costimulatory receptors, engaged simultaneously with the TCR, have a major impact on signaling events and a decisive role in the differentiation program of T cells. Thus, their role in altering the immunometabolic programming mediated by TCR-mediated activation is unequivocal (5–7).

Our understanding about the functional role of costimulation has evolved from the two-signal model proposed by Lafferty and Cunningham to explain the activation of naïve T cells (8, 9). Although T cell costimulatory pathways were envisioned as stimulators of T cell responses by that model, it is now clear that both stimulatory (costimulatory) and inhibitory (coinhibitory) second signals exist and mediate their impact not only in naïve (T_N) but also in effector (T_EFF), memory (T_M), and regulatory (T_REG) T cells (10–12). These receptors are key regulators of T cell activation, tolerance, and exhaustion, and therapeutic modulation of costimulatory and coinhibitory pathways provides effective new treatment strategies in cancer, autoimmunity, infectious diseases, and allogeneic transplantation (13–15).

The first costimulatory receptor CD28 and the first coinhibitory receptor CTLA-4 and their shared ligands CD80 (B7-1) and CD86 (B7-2) constitute the best-characterized pathway, which serves as a paradigm for other costimulatory and coinhibitory pathways. The massive lymphoproliferation and organ infiltration by activated T cells in the CTLA-4-deficient mice (16, 17) was the first indication that coinhibitory receptors have a mandatory role in the regulation of T cell tolerance. Today, many additional costimulatory and coinhibitory receptors have been identified, and their role in the induction and maintenance of T cell tolerance has been established (Figure 1). These pathways fall into two major families: the Ig superfamily, which includes the B7–CD28, TIM, CD226–TIGIT–CD96 families as well as LAG-3, and the TNF–TNF receptor superfamily (18–20). Coinhibitory receptors provide a balance on the activation and expansion of antigen-specific T cells upon encounter with antigen but also regulate T cell tolerance by restraining the initial activation of naïve self-reactive T cells and/or responses of harmful self-reactive T cells. Coinhibitory pathways also regulate the generation and function of thymic-derived T_REG and T_REG generated at peripheral sites (21, 22). Ligands for various coinhibitory receptors are expressed on antigen-presenting cells (APCs) but also in non-hematopoietic cells (18–20). The expression of coinhibitory ligands on non-hematopoietic cells has a key role for the maintenance of tissue tolerance by suppressing the expansion and function of self-reactive T cells. Importantly, tumors and infectious agents that cause chronic infections exploit these natural tolerance mechanisms to evade surveillance and attack by the immune system (13, 23).

Coinhibitory receptors have a major impact on the differentiation fate of T cells as well as on their ability to proliferate. These functional endpoints are regulated by T cell metabolic reprogramming (24, 25). Therefore, altering metabolic reprogramming might be a key mechanism by which coinhibitory receptors modify T cell differentiation and function. Because various coinhibitory receptors mediate distinct effects on the activation of signaling pathways downstream of the TCR, the role of coinhibitory receptors on altering the metabolic programs of T cells is also anticipated to be different. Identification and targeting the specific immunometabolic pathways relayed by coinhibitory receptors might have significant clinical implications. For example, selective receptor targeting might be employed in order to achieve specific desired modifications in the metabolic programs and functional fates of T cells. In this review, first, we will discuss the current understanding of the mechanisms underlying metabolic programs induced during T cell differentiation and how these are affected by cytotoxic lymphocyte antigen-4 (CTLA-4) and programmed death-1 (PD-1), the two most clinically relevant coinhibitory receptors thus far, for which therapeutic targeting has achieved major success in cancer treatment. Next, we will review briefly other inhibitory pathways, and we will discuss lymphocyte activation gene-3 (LAG-3), T cell-immunoglobulin–mucin domain 3 (TIM-3), and T cell immunoglobulin and ITIM domain (TIGIT), which are additional promising therapeutic targets. Finally, we will briefly discuss the CD160/B and T lymphocyte attenuator (BTLA)/LIGHT/herpesvirus entry mediator (HVEM) pathway, which has the rare property to engage members of the Ig superfamily and the TNF receptor superfamily and to deliver both coinhibitory and costimulatory signals (Please see glossary in Supplementary Material).

Similar to the metabolism of other non-proliferating cells, naïve T cells depend on catabolic metabolism to meet their energetic needs, relying predominantly on oxidative phosphorylation (OXPHOS) to generate ATP (26). Extrinsic IL-7 survival signal maintains naïve T cell quiescence, a mechanism that relies on tuberous sclerosis (TSC) function to keep mTOR activation in check (27–29). Upon antigen encounter, T cells switch to anabolic metabolism to meet the increased metabolic demands of biomass accumulation, proliferation, and biosynthesis of mediators of effector functions (Figures 2 and 3A). Signaling via the TCR is sufficient to upregulate expression of the glucose transporter Glut1 via Myc (30). However, the metabolic reprogramming accompanying the differentiation into T_EFF requires costimulation. Signaling via CD28 costimulation activates the PI3K/Akt/mTOR pathway, which is crucial to activate aerobic glycolysis required for T cell differentiation and function (6, 31, 32). mTOR has a key role in regulating metabolism coupling nutrient availability to cell growth and division (33, 34).

After pathogen clearance, antigen-specific T cells undergo contraction, and most T_EFF cells undergo cell death. A small subset survives and goes on to become T_M cells. T_M cells exhibit a distinctive increase in mitochondrial mass and higher spare respiratory capacity (SRC) (35). This is a consequence of fusion in the mitochondria of T_M cells, which configures associations of the electron transport chain and favors OXPHOS (36). In contrast to T_EFF, which rely predominantly on glycolytic metabolism, T_M rely on mitochondrial fatty acid oxidation (FAO) of de novo synthesized FA, rather than uptake of lipids (37). During TCR engagement with APCs, T cells can divide asymmetrically. It has been determined that the APC-proximal daughter generated during cell division is more likely to become a T_EFF while the APC-distal daughter cell is more likely to differentiate into a T_M (38). In addition to its role in promoting glycolysis, asymmetric Myc sorting between the daughter cells plays a role in the distribution of amino acid content and transporters as well as mTORC1. This results in altered metabolism and thus altered proliferation and differentiation that may give rise to effector versus memory fates of the daughter cells (39). This is likely not the only mechanism to generate T_M cells, as other studies have shown that T_EFF cells are the ones that can give rise to T_M (40, 41). Enforcing the metabolic switch to FAO by enhancing AMPK or by inhibiting mTOR results in increased numbers of T_M cells (42, 43). In mammalian cells, mTOR forms functionally distinct complexes—mTORC1 and mTORC2. Raptor and Rictor are the core adaptor subunits of mTORC1 and mTORC2, respectively. Interestingly, T cells deficient in Rictor show enhanced CD8 memory formation with enhanced recall responses without dampening effector functions (44).

Since the activation of T cells requires accompanying metabolic reprogramming for differentiation, it is no surprise that metabolism drives the differentiation of CD4 T helper (Th) subsets as well. Th1, Th2, Th17, and T_REG cells are the best metabolically defined thus far. Initial evidence showing their metabolic distinction came when suppression of mTOR by rapamycin prevented T_EFF proliferation and instead promoted T_REG generation, even in Th17 polarizing conditions (45). Genetic deletion of mTOR in T cells confirmed the results of pharmacological inhibition and caused production of only T_REG, illustrating the obligatory role of mTOR for T_EFF production (46). Furthermore, T_EFF show a strong preference for glycolytic rather than mitochondrial metabolism, whereas T_REG rely more on the oxidation of lipids (47).

Th17 cells are particularly dependent on glycolysis. Although Myc is dominant over the hypoxia-inducible factor 1α (HIF1α) in regulating differentiation of naïve T cells into T_EFF cells (30), HIF1α has a key role in the metabolic switch to aerobic glycolysis that influences the balance of Th17/T_REG in favor of Th17 cells. Genetic deletion of HIF1α in mice or blocking glycolysis with 2-deoxyglucose showed reduced Th17 and enhanced T_REG development (48). HIF1α in coordination with RORγt and p300 promotes the development of Th17 cells while concurrently weakening the development of T_REG cells by promoting proteasomal degradation of FoxP3 (49). Furthermore, differentiation of Th1 and Th17 cells is selectively regulated by mTORC1 while Th2 differentiation is regulated by mTORC2 (50).

It should be noted that glycolytic enzymes themselves might also have an active role in regulating T_EFF cell function. For example, it has been proposed that when glycolytic rates are low in Th1 CD4⁺ T cells, glyceraldehyde phosphate dehydrogenase (GAPDH) binds to and suppresses IFN-γ mRNA translation (51). In effector memory CD8⁺ T cells, glycolysis was observed to impact chromatin remodeling in the promoter region of the IFN-γ gene, thus facilitating gene transcription (52). In addition, naïve T cells express both isoforms of the glycolytic enzyme pyruvate kinase, PKM1 and PKM2, which catalyzes the conversion of phosphoenolpyruvate to pyruvate. After activation, the M2 isoform rapidly accumulates and is the dominant form expressed in T_EFF cells (53, 54). The M2 isoform is less efficient, but this skews glycolysis toward biosynthetic pathways and may give cells a growth advantage during rapid proliferation (55). Because coinhibitory receptors are induced after T cell activation, a process that is dependent on glycolysis, there might be a correlation between the expression and function of glycolytic enzymes and coinhibitory receptors. These findings provide evidence for an intimate causative link among TCR-mediated signaling, T cell metabolism, and T cell differentiation.

Signaling pathways activated by engagement of the TCR and costimulatory receptors and CD28, the prototype costimulatory receptor, activates the PI3K/Akt/mTOR and MEK/Erk MAPK pathways, and their simultaneous activation is required to promote T cell growth, division, and cytokine production (56) (Figure 3A). Autocrine and paracrine cytokine signaling loops induce further PI3K and MAPK activation, together with JAK/STAT signaling (57, 58). PI3K catalyzes the production of the second messenger phosphatidylinositol 3,4,5-trisphosphate [PI(3,4,5)P3], leading to the activation of several downstream kinases, including Akt, its most prominent effector. In naïve T cells, active Akt leads to increased Glut1 surface expression, improved coupling of HKII to the mitochondria, and increased rates of glycolysis, which correlates with proliferative capacity (6, 7, 53). Akt also promotes aerobic glycolysis by direct phosphorylation and activation of glycolytic enzymes, such as PFK2 (59), by regulating Glut1 trafficking, and by inducing several transcription factors (60). Akt can activate mTORC1, a key driver of anabolic metabolism (61).

MAP kinases have a central role in metabolic regulation during T cell activation by affecting both glucose and glutamine uptake and metabolism (2, 62). One target of MAPK with a direct role on metabolic regulation is HIF1α. MAPK promotes the ability of HIF1α to activate transcription and expression of many glycolytic genes, while mediating suppression of genes involved in OXPHOS (63). HIF1α levels are also upregulated by PI(3,4,5)P3 signaling through mTOR (61) providing a mechanistic explanation for the requirement of simultaneous activation of PI3K and MAPK pathways for optimal effector T cell expansion (56). The critical role of HIF1α in T cell differentiation and function is underlined by the observation that T cells with constitutively elevated HIF1α display sustained increase of aerobic glycolysis and constitutively maintain effector function (64).

After TCR and costimulatory receptor engagement, T cell activation is maintained by sustained signaling from IL-2 and other cytokines acting on common gamma chain (γc) cytokine receptor complexes (65). This effect of IL-2 leads to glycolysis and anabolic metabolism by activating transcriptional upregulation and trafficking of Glut1. This is mediated by Jak-STAT5 signaling (66) and also by activation of PI(3,4,5)P3/Akt (57).

T cell receptor signaling pathways also lead to the generation of second messengers, diacylglycerol (DAG), and calcium. Calcium flux from intracellular stores stimulates the calcium-activated mitochondrial dehydrogenases that drive the TCA cycle (67). In addition, calcium flux downstream of the TCR causes a short-term phosphorylation of AMPK mediated by CAMKK (68). AMPK acts as an important metabolic checkpoint to control cellular anabolic pathways and to induce energy-conserving mechanisms under conditions of nutrient shortage or cellular stress. AMPK activation early during the initiation of T cell stimulation may facilitate transient increase of OXPHOS before the cells switch their metabolism toward glycolysis thereby supporting the energy demands at this early T cell activation stage. Mechanistically, AMPK activation is induced by the upstream kinase LKB1 when the cellular levels of ATP are diminished and the levels of AMP and ADP are increased in response to physiological stress (69). During T cell activation, in addition to regulation of energy, AMPK also provides metabolic fitness to the differentiating T_EFF and T_M cells (70). After activation, AMPK phosphorylates several substrates involved in the metabolism of glucose, lipid, and proteins, in autophagy, cell cycling, and cell growth (71). Activation of the TSC1/TSC2 complex by AMPK can negatively regulate the mTOR pathway (72, 73) not only by inhibiting mTOR activation via Rheb (74), a member of the Ras family of GTPases, but also by mediating phosphorylation of the mTORC1 partner Raptor (75).

The key signaling pathways that are activated by antigen encounter are targets of costimulatory and coinhibitory receptors (Figure 3B). Costimulatory receptors amplify TCR-mediated signaling and impact T cell metabolic reprogramming. For example, CD28-mediated costimulation mediates transcriptional upregulation of the glucose receptor Glut 1 (6) and the glutamine receptors Snat1 and Snat2 (62) after TCR ligation. Costimulation also amplifies MAPK and PI3K/Akt activation, and the simultaneous engagement of these signaling pathways is mandatory for cell cycle progression and cytokine production (56). Conversely, coinhibitory receptors oppose the signaling and functional effects induced by TCR and costimulatory pathways, thereby limiting T cell expansion and cytokine production and altering the ability of T cells to differentiate into T_EFF and T_M cells. Importantly, coinhibitory receptors do not induce global signal inhibition but rather selective targeting of signaling, resulting in imbalanced activation of pathways that are normally engaged by TCR-plus-costimulation. As a consequence, the implications of coinhibitory pathways on the metabolic and functional programs of T cells might significantly vary, dependent on the signaling targets of these inhibitory receptors.

The extensive studies regarding the expression and function of coinhibitory receptors provide currently emerging evidence that these receptors have synergistic roles in suppressing T cell function and provide evidence for their non-redundant roles in inhibiting the responses of activated T cells. Although the evolutional purpose for the coexpression of multiple coinhibitory receptors is not clearly understood, several observations provide challenging hypotheses. It is possible that coinhibitory receptors are expressed in distinct combinations on T cell subsets depending on the type of activation signals that such T cells receive, which might depend on the engagement of distinct costimulatory receptors or on the activation of autocrine and paracrine cytokine loops. For example, expression and recruitment of CTLA-4 to the immunological synapse depend on the strength of the TCR signal (107). It is also possible that ligands for the different coinhibitory receptors are differentially expressed in various tissues or tissue resident APC and, as a consequence, distinct coinhibitory receptors control T cell responses in distinct microenvironments. For example, parenchymal tissues and organs express ligands for PD-1 but not for CTLA-4 (20). As nutrients and metabolites provide important environmental cues with decisive role in T cell activation (187), it is also possible that distinct metabolic programs of various microenvironments selectively promote the expression and function of different coinhibitory receptors. For example, protection from oxidative damage mediated by reactive oxygen species (ROS) by an SOD-mimetic compound, resulted in elevated expression of LAG-3 in T cells and inhibition of TCR-mediated Th1 cell activation and effector function (139). Such metabolism-driven events might also differentially regulate the expression of coinhibitory receptors in T cell subsets (53).

Currently, there is little known about the metabolic repercussions of coinhibitory receptors on T cells and the metabolic consequences by targeting these immune checkpoint pathways in pathogen-specific and tumor-specific T cells. The immunometabolic effects mediated by the most clinically relevant coinhibitory receptors, CTLA-4 and PD-1, are starting to be elucidated. Laboratory studies have provided evidence that inhibitory checkpoint pathways can alter the metabolic program of T cells (126, 188). PD-1 engagement makes T cells unable to engage in glycolysis, glutaminolysis, or amino acid metabolism but increases FAO (126). This effect is due to inhibition of transport and utilization of glucose and glutamine. In contrast, PD-1 promotes FAO of endogenous lipids by inhibiting the lipid oxidation PI3K pathway, resulting in increased expression of carnitine palmitoyltransferase 1A (CPT1A), the rate-limiting enzyme of FAO, which plays an important role in the utilization of fatty acids as an energy source. In parallel, PD-1 abrogates the induction of fatty acid synthase and, as a consequence, decreases anabolic lipid metabolism and biosynthesis of lipids, which normally occurs during T cell activation.

Compared to T cells activated without PD-1 ligation, activated T cells receiving PD-1 signals have lower extracellular acidification rate (ECAR), an indicator of glycolysis, and lower oxygen consumption rate, an indicator of OXPHOS. These findings indicate that T cells receiving PD-1 signals are rather metabolically quiescent and preferentially use OXPHOS over glycolysis as indicated by the higher OCR/ECAR ratio. As fat-based metabolism is associated with longevity (189), the augmentation of FAO may explain the longevity of T cells, which engage the PD-1 pathway, in patients with chronic infections and cancer and their ability to be reinvigorated by blockade of PD-1. The role of PD-1:PD-L1 pathway in restricting glucose metabolism in vivo was also reported in a study of graft-versus-host disease in mice, where recipients of allogenic bone marrow deficient in PD-L1 had higher levels of Glut1 and lactate production (190).

Because of the increased FAO and the elevated OCR/ECAR ratio, it is possible that PD-1-mediated inhibitory effects in T cells might be related to a more oxidative environment. Although depletion of the antioxidant glutathione (GSH) was observed in T cells stimulated with or without PD-1 ligation, consistent with the expected increase of ROS during TCR-mediated activation (191, 192), PD-1 ligation resulted in more pronounced decrease in the levels of reduced GSH. However, under these conditions, T cells had higher levels of cysteine-GSH disulfide and ophthtalmate, a GSH-like product that is synthesized by the same enzymes (126). These changes might be suggestive of a more oxidative environment in T cells receiving TCR signals in conjunction with PD-1 ligation. However, several other tentative mechanisms might be implicated in these observations. In alloreactive T cells after allogeneic bone marrow transplantation, PD-1 expression has been correlated with an oxidative environment, ROS production, and higher susceptibility of these cells to metabolic inhibition by F1F0-ATP synthase complex inhibitors (193). Thus, expression and function of PD-1 and possibly other coinhibitory receptors might be linked with the T cell oxidative state, which can serve as a fine-tuning system of cellular differentiation.

A key mediator of oxidative detoxification is the PPARγ coactivator-1a (PGC-1α) (194, 195). Induction of PGC-1α during caloric restriction mediates selective cell adaptations that provide modifications permissive for cell survival and function under increased OXPHOS. Because PGC-1α is regulated by mTOR (196), which is downstream of PI3K/Akt pathway that is targeted by PD-1 (120, 121), it was hypothesized that by dysregulating PGC-1α expression and function, PD-1 might impair oxidative detoxification. Moreover, different subsets of T_EX cells might have distinct metabolic and bioenergetics properties such as mitochondrial biogenesis, which might correlate with glycolysis, SRC, CPT1A, and PGC-1α (126). Consistent with this hypothesis, studies in chronic LCMV infection, which leads to the generation of various subsets of T_EX cells, showed that the development of intermediate PD-1 (PD-1^int) T_EX cells correlated with reduced glucose uptake and glycolysis, dysregulated mitochondrial bioenergetics, and suppressed PGC-1α expression. PGC-1α expression was enhanced in the absence of PD-1, and forced expression of PGC-1α partially reverted some of the metabolic derangements of PD-1^Int—but not in PD-1^Hi—T_EX cells and improved effector function (197). Similarly to the observations of in vitro stimulated human T cells (126), transcriptional analysis during T_EX development in response to LCMV infection revealed enrichment of genes involved in OXPHOS, particularly CPT1A (197). While these results point to a strong connection between PD-1-mediated metabolic reprogramming and induction of T cell exhaustion, it should be pointed out that similar bioenergetics mechanisms of T cell exhaustion in vivo mediated by chronic infection and cancer can also occur in a PD-1-independent manner (198, 199). Thus, other coinhibitory receptors might mediate similar immunometabolic effects, thereby resulting in differentiation of functionally impaired T cells.

As mentioned above, PD-1 together with several other coinhibitory receptors coregulate T cell exhaustion during cancer and chronic infections. Pathogen-specific CD8⁺ T cells in mouse models and humans with chronic viral infections as well as TILs from mouse tumor models and patients with cancer can coexpress multiple repressors, including PD-1, LAG-3, TIM-3, TIGIT, and others. The pattern and number of coinhibitory receptors expressed on the same cell can dramatically influence the severity of exhaustion (114, 115). T cell exhaustion is an active process and can lead to measurable consequences on T cell function. Studies have shown that distinct subsets of T_EX cells exist, which have different potentials for recovery after PD-1 blockade. Although T_EX cells with PD-1^int expression can be reinvigorated by PD-1 blockade, T_EX cells with high PD-1 (PD-1^high) expression cannot (115). These distinct subsets of T_EX cells may also have different metabolic and bioenergetics properties depending on the nature of coinhibitory receptors expressed and engaged on their surface. Identification of the unique bioenergetics profiles of T_EX subsets might provide new tools to identify the level of exhaustion mediated by the combined and synergistic effects of coinhibitory receptors. It might also provide novel targets to reverse exhaustion and might serve as meaningful biomarkers to assess response to checkpoint inhibitor therapies.

In contrast to the effects of PD-1 in T cell metabolic reprogramming, CTLA-4 inhibits glycolysis without augmenting CPT1A and FAO, suggesting that CTLA-4 maintains immune quiescence by preserving the metabolic profile of non-stimulated cells (126). RNA-seq of TILs from progressor sarcoma tumors in mice that were either treated with control, anti-CTLA-4, anti-PD-1 blocking antibodies, or a combination of anti-CTLA-4 and anti-PD-1 antibodies revealed differential gene enrichment. PD-1 blocking showed enrichment predominantly in metabolic genes, while CTLA-4 blocking showed enrichment in cell cycle and effector memory genes. The combination of PD-1 and CTLA-4 antibody treatment revealed synergism of metabolic and T cell effector genes that correlated with the strongest enhancement of effector molecules (108). The distinct properties of CTLA-4 and PD-1 in affecting T cell metabolic reprogramming might be related to their differential ability to oppose signals mediated by the TCR. CTLA-4 mediates a potent inhibition of T cell activation, and CTLA-4 recruitment at the immune synapse is regulated by the strength of TCR signal (107). In contrast, PD-1 inhibits weak but not strong TCR signals (200). As a consequence, TCR-mediated biochemical events might be impaired but not fully abrogated (124, 201). Notably, the strength of PD-1 signaling also differentially affects T_EFF cell function (201). This rheostat effect of PD-1 on TCR signaling that likely depends on the level of PD-1 expression might mediate fine tuning of metabolic reprogramming leading to distinct T cell functional fates. The different immunometabolic implications of CTLA-4 and PD-1 in T cells, serve as a paradigm that provides evidence for the purpose of multiple coinhibitory receptors, which likely have unique roles in regulating immunometabolic programs and functional commitment of T cells. Mechanistic understanding of the immunometabolic effects of coinhibitory receptors will shed light not only to the state of T cell functional impairment and exhaustion in chronic infections and cancer but also to the metabolic regulation of T cell tolerance.

It has become increasingly understood that coordinated metabolic switches in T cells and innate immune cells modulate cellular activities and contribute to the progression of cancer and chronic infections. Metabolic communication among T cells, innate immune cells, and cancer might contribute to immunometabolic regulations and fine tuning of the activation and antitumor responses of T cells and myeloid cells (202) (Figure 4). In this regard, immunometabolic regulations mediated by coinhibitory receptors can impact T cell immune responses due to direct T cell-intrinsic signaling and by altering the metabolic properties and the differentiation program of innate immune cells (203). Not only the coinhibitory receptor ligands but also the coinhibitory receptors themselves are also present in various types of innate immune cells in addition to T cells. Intriguingly, it was recently reported that the PD-1: PD-L1 axis is implicated in immune metabolic dysfunctions of monocytes in chronic lymphocytic leukemia (203). In that system, triggering PD-1 checkpoint on monocytes hampers glycolysis and phagocytosis. Conversely, disrupting PD-1: PD-L1 signaling reverses these immune metabolic dysfunctions. Taken together, these findings imply a novel metabolic interplay between cancer cells and monocytes and that blocking PD-1: PD-L1 might restore metabolic together with antitumor activity of monocytes/macrophages.

In a sarcoma tumor model, it was reported that expression of PD-L1 on cancer cells was associated with cell-intrinsic signaling via the PI3K/Akt pathway and mTOR, leading to expression of glycolysis genes and enhanced glycolytic metabolism (204). Although it is unclear whether PD-L1 can trigger reverse signals to tumor cells by its short cytoplasmic tail that does not have obvious signaling motifs, PD-L1 functions as an inhibitory receptor that protects cancer cells from immune-mediated cancer cell destruction and Fas-mediated killing (205). Because cancer cells have elevated activation of the PI3K/Akt pathway and are highly glycolytic, expression of PD-L1 might result in concomitant increase of PI3K/Akt activation and elevated rate of tumor-intrinsic glycolysis as a consequence of improved survival. It has also been observed that subpopulations of certain human and murine melanoma cell lines and subpopulations of malignant cells from patients with melanomas express PD-1 (206, 207). However, in contrast to T cells, in which PD-1 ligation inhibits the activation of PI3K/Akt, MAPK, and mTOR pathways, ligation of PD-1 in melanoma cells was reported to activate these pathways and to promote expression of glycolytic enzymes, which correlate with tumor growth (206). To date, such finding of cancer cell-intrinsic growth-promoting effects of PD-1 have not been validated in tumor models and primary cancer cells. It is also unclear whether PD-L1 naturally expressed on tumor cells can trigger activation of glycolytic metabolism in cancer cells and whether such effect might be associated with tumor growth. However, together these findings (203, 204, 206) suggest that cell-intrinsic metabolic signaling mediated by the PD-1:PD-L1 pathway in cancer or myeloid cells of the tumor microenvironment might subsequently alter T cell immunometabolism and immune function due to nutrient competition and by generating metabolic products with detrimental effects on T cell immune function (208).

Coinhibitory receptors have a major impact on the metabolic programs that drive the differentiation and functional fate of naïve T cells. These coinhibitory receptors have a major role in reprogramming the metabolic preferences of T_EFF cells leading to metabolic and functional switches. Because coinhibitory receptors mediate distinct effects on the activation of signaling pathways, the role of coinhibitory receptors on altering T cell immunometabolism is also anticipated to be distinct. Identifying and targeting the specific immunometabolic determinants relayed by different coinhibitory receptors might have significant clinical implications as selective receptor targeting might be employed in order to achieve specific desired modifications in the metabolic programs and functional fates of T cells.

Metabolic determinants of different microenvironments provide important cues to drive differentiation and function of T cells by metabolism-based mechanisms. Expression and engagement of coinhibitory receptors after T cell activation can provide the necessary signals to terminate immune responses by altering the metabolic programs of antigen-specific T_EFF cells. Such metabolic reprogramming might lead to conversion of T_EFF to T_REG and T_M for termination of immune activation and maintenance of the T cell repertoire, respectively. Understanding and recapitulating the metabolic programs mediated by coinhibitory receptors will have implications for the induction of self-tolerance and the treatment of autoimmune diseases. Conversely, blocking and reversing metabolic programs mediated by coinhibitory receptors will promote T_EFF differentiation and function and will improve immune-based therapies mediated by checkpoint inhibitors and by adoptive immunotherapy of ex vivo engineered T cells for the treatment of cancer and chronic infections. Exploiting the role of the coinhibitory receptors in T cells, innate immune cells and cancer will be a major challenge.

NP and JW wrote the main body of the manuscript and generated figures; LS wrote several sections of the manuscript; CH participated in the preparation of the manuscript and generated the Supplementary Material; PS participated in the writing of various sections of the manuscript; and VB had the overall supervision in the preparation of the manuscript and figures.

Vassiliki A. Boussiotis has patents on the PD-1 pathway licensed by Bristol-Myers Squibb, Roche, Merck, EMD-Serono, Boehringer Ingelheim, AstraZeneca, Novartis and Dako. The authors declare no additional potential conflicts of interest.