MenTORing Immunity: mTOR Signaling in the Development and Function of Tissue-Resident Immune Cells

Abstract

Tissue-resident immune cells must balance survival in peripheral tissues with the capacity to respond rapidly upon infection or tissue damage, and in turn couple these responses with intrinsic metabolic control and conditions in the tissue microenvironment. The serine/threonine kinase mammalian/mechanistic target of rapamycin (mTOR) is a central integrator of extracellular and intracellular growth signals and cellular metabolism and plays important roles in both innate and adaptive immune responses. This review discusses the function of mTOR signaling in the differentiation and function of tissue-resident immune cells, with focus on the role of mTOR as a metabolic sensor and its impact on metabolic regulation in innate and adaptive immune cells. We also discuss the impact of metabolic constraints in tissues on immune homeostasis and disease, and how manipulating mTOR activity with drugs such as rapamycin can modulate immunity in these contexts.

Introduction

Although there are immune system-specific organs, such as the lymph nodes and thymus, many immune cells (of which there are numerous cellular types) are disseminated to reside in tissues throughout the body. Some immune cells exist in these tissues during homeostasis, while others become tissue infiltrating in settings of tissue damage or infection. Thus, unlike most cell types that differentiate to exist in a highly orchestrated environment within a given tissue, immune cells may exist in numerous distinct tissue environments. This implies a certain metabolic flexibility on the part of immune cells. Moreover, most immune cells are inherently capable of shifting from resting to activated states as they respond to danger signals or antigen and become engaged in the immune response. These transitions involve large-scale changes in gene expression and therefore of cellular function and may irreversibly change fate and lifespan expectancy. Recent work has established that enactment of these events requires substantial metabolic reprogramming (Buck et al., 2015; O’Neill and Pearce, 2016; Pearce et al., 2013), and this, as well as the realization that living in diverse tissue-specific niches may have metabolic consequences for immune cells, is serving to refocus attention on immune cell metabolism (for detailed discussion of bioenergetics in immune cell, please see Olenchock et al., 2017, in this issue). Integral to this area of research is the question of how immune cells assess their metabolic status. In this context, we focus here on the role of the metabolic sensor mTOR in tissue-resident immune cells.

mTOR: A Central Integrator of Cellular Metabolism

Central to metabolic control in eukaryotic cells is the mechanistic/mammalian target of rapamycin (mTOR), a serine/threonine kinase with high evolutionary conservation from yeast to humans (for in-depth reviews on mTOR function, see Albert and Hall, 2015; Laplante and Sabatini, 2012). mTOR responds to both extracellular signals, such as hormones and growth factors (insulin, IGF-1), ligation of pattern recognition and antigen-specific receptors (TLR, TCR, BCR activation), and cytokines (IL-2, IL-4, IL-12), and intracellular cues including nutrient (i.e., amino acid) abundance and cellular energy charge (AMP:ATP ratio) to regulate cell growth and proliferation (Howell et al., 2013; Morita et al., 2015; Pollizzi and Powell, 2015). mTOR exists in two structurally distinct complexes in cells—denoted mTOR complex 1 (mTORC1) and mTORC2—that mediate separate but overlapping cellular functions (Figure 1). Among the defining features of these complexes are unique structural components—Raptor for mTORC1 and Rictor for mTORC2—that mediate substrate specificity for each complex, which have facilitated the generation of genetically engineered mouse models to examine the function of each complex in immune cell subsets. Acute treatment with rapamycin inhibits mTORC1 activity while enhancing mTORC2 activity (Sarbassov et al., 2005), whereas active site mTOR inhibitor (asTORi) compounds such as Torin1 target both complexes (Thoreen et al., 2009). The central function of mTORC1 is to direct cellular growth and proliferation by regulating pathways of anabolic metabolism, most notably mRNA translation, while mTORC2 regulates downstream signal transduction by AGC family kinases (including Akt and SGK1) and the actin cytoskeleton (Figure 1).

Figure 1. mTORC1 and mTORC2 Mediate Separate but Overlapping Cellular Functions.

In resting cells, or when extracellular amino acid concentrations are low, mTORC1 is dissociated from lysosomes and is inactive (left side of lysosome). When immune cells become activated, they express amino acid transporters to allow them to more efficiently acquire extracellular amino acids. This is coordinated with Akt-dependent signals that alleviate TSC2-dependent inhibition of Rheb, to allow recruitment of mTORC1 to lysosomes, where it becomes activated by Rheb (right side of lysosome). Active mTORC1 phosphorylates S6K and 4EBP, with the net result of increasing ribosomal biogenesis and the translation of mRNA subsets coding for a variety of proteins, but especially proteins involved in anabolic metabolic pathways and immune mediators. This enables activated cells to generate more metabolic intermediates for biosynthesis to support cell growth, proliferation, and effector functions. mTORC2, which can be directly activated by PI3K, also promotes metabolic reprogramming through Akt-mediated activation of hexokinase 2 (HK2) or inhibition of Foxo1. Downstream targets of mTORC2 such as SGK also play direct roles in Th cell differentiation and in cytoskeletal dynamics important for cell movement.

Both mTORC1 and mTORC2 exert effects on cellular metabolism. mTORC1 activation downstream of receptor-coupled PI3K signaling leads to the phosphorylation of ribosomal protein S6 kinase (S6K) and eukaryotic translation initiation factor 4E (eIF4E) binding proteins (4E-BPs) to stimulate cap-dependent translation initiation (Morita et al., 2015). This promotes the translation of metabolic enzymes required for proliferation, such as the nucleotide biosynthesis enzyme phosphoribosyl-pyrophosphate synthetase 2 (PRPS2) (Cunningham et al., 2014), nuclear encoded mitochondrial proteins involved in mitochondrial homeostasis and electron transport (TFAM, Complex I and V components) (Morita et al., 2013), and the synthesis of transcription factors such as Myc and HIF-1α critical for metabolic reprogramming (Figure 1) (Barnhart et al., 2008; Düvel et al., 2010). mTORC1 plays additional roles in metabolic regulation through S6K-mediated phosphorylation of the CAD enzyme complex to stimulate pyrimidine biosynthesis (Ben-Sahra et al., 2013; Robitaille et al., 2013), ATF4-dependent stimulation of the mitochondrial tetrahydrofolate (mTHF) cycle to enhance purine biosynthesis (Ben-Sahra et al., 2016), and stimulation of lipid and sterol biosynthesis through activation of the sterol regulatory binding element proteins (SREBPs) (Düvel et al., 2010). mTORC2, by activating Akt through Ser473 phosphorylation, can enhance glycolytic metabolism (Hagiwara et al., 2012). This may be due in part to the association of TORC2 with the mitochondria-associated ER membrane (MAM), where mTORC2-Akt signaling stimulates hexokinase 2 (HK2) activity to drive the entry point of glycolysis (Betz et al., 2013). Foxo1 stimulates the expression of negative regulators of the Myc transcription factor, leading to reduced Myc-dependent metabolic reprogramming and suppression of both glycolytic and oxidative metabolism (Wilhelm et al., 2016). mTORC2-dependent acetylation and phosphorylation of Foxo1, which relieves Foxo1-mediated suppression of Myc, acts as a switch toward pro-growth metabolism (Figure 1; Sato et al., 2013).

Given the central role of mTOR in metabolic control, many inputs regulate its function. The canonical activation of mTORC1 by growth factors occurs through Akt-mediated phosphorylation of the tuberous sclerosis complex (TSC1-TSC2-TBC1D7) (Inoki et al., 2002). TSC2 is a GTPase-activating protein (GAP) for the Ras family GTPase Rheb (Inoki et al., 2003a; Tee et al., 2003), with TSC2 functioning as an inhibitor of Rheb and subsequently mTORC1 activation. Amino acids, notably leucine, promote mTORC1 recruitment to lysosomes (Sancak et al., 2010). Lysosomal recruitment of mTORC1, combined with dissociation of the TSC complex with the lysosome after TSC2 phosphorylation (Menon et al., 2014), leads to its activation by Rheb (Figure 1). Factors that enhance TSC complex stability, such as LKB1-AMPK signaling (Corradetti et al., 2004; Inoki et al., 2003b), antagonize mTORC1 activation. Inhibition of mTORC1, either by rapamycin or Raptor deletion, stimulates mTORC2-dependent Akt phosphorylation (Sarbassov et al., 2005), although prolonged rapamycin treatment can impact mTORC2 complex assembly, ultimately leading to reduced mTORC2 activity (Sarbassov et al., 2006). The existence of feedback loops for mTOR kinase activity indicates that caution should be used when evaluating distinct metabolic effects of each kinase complex, as we will discuss further.

mTOR in the Adaptive Immune System

Adaptive immunity is mediated by CD4⁺ and CD8⁺ T cells, which function primarily but not solely by secreting cytokines, and by B cells whose principal but not only role is to make antibodies. These cells clonally express antigen-specific receptors. In resting cells, previously unstimulated and positioned in lymphoid organs, ligation of these receptors plus additional critical signals from other immune cells leads to clonal expansion during which the cells gain effector functions. During this process, some cells assume short-lived roles as effector cells, which may enter peripheral tissues to mediate their effector function, while others become persistent quiescent memory cells in the case of T cells or B cells or antibody-secreting plasma B cells. These long-lived cells may circulate in the Bloodstream and/or reside in secondary lymphoid tissues or bone marrow. In the case of CD8⁺ T cells, it is also clear that memory cells can permanently reside in most tissues (Mueller and Mackay, 2016). Memory cell persistence and ability to become reactivated upon re-exposure to antigen is the basis of long-lived immunological memory.

Not surprisingly given its central role as a regulator of cell function, and fueled initially by the availability of rapamycin and other mTOR inhibitors, the role of mTOR in immune cell development and function after activation has received considerable attention. Initial studies have been focused on the ability of rapamycin to inhibit T cell proliferation and IL-2 production (Abraham, 1998; Peng et al., 2002) and to promote the development of anergy even in T cells that have been appropriately costimulated (Powell et al., 1999). From these studies emerged the finding that rapamycin can promote the development of regulatory CD4⁺ T cells (Treg cells) from naive T cells (Valmori et al., 2006), which is consistent with the fact that constitutively active Akt suppresses regulatory T (Treg) cell development (Haxhinasto et al., 2008). Treg cell-specific deletion of the phosphatase PTEN decreases Treg cell stability and function, due to excessive Akt and mTORC2 activity (Shrestha et al., 2015). The negative effects of mTOR signaling on Treg cell development has been recapitulated by the deletion of mTOR in CD4⁺ T cells, which promotes CD4⁺ T cell differentiation to Treg cells as a default pathway (Delgoffe et al., 2011). However, mTORC1 activity is critical for Treg cell function. CD4⁺ Treg cells display elevated steady-state mTORC1 activity, and loss of mTORC1 (via Treg cell-specific deletion of Raptor) leads to loss of Treg cell suppressive function and autoimmunity (Zeng et al., 2013). One implication of these results is that TCR- and/or IL-2-dependent signals to mTORC1 are necessary to maintain Treg cell survival and functional competency in peripheral tissues.

At about the same time that the importance of mTOR in Treg cell function finding was reported, it became clear that rapamycin treatment during immune response initiation can—paradoxically—promote the development of central CD8⁺ T memory (Tmem) cells (Araki et al., 2009; Pearce et al., 2009). Interestingly, we now understand that this same treatment prevents the accumulation in tissues of tissue-resident CD8⁺ Tmem cells (Sowell et al., 2014). The effects of rapamycin on central CD8⁺ Tmem cell development has been shown to be due to a pivotal role for the catabolic process of fatty acid oxidation (FAO) and autophagy in this process and the ability of rapamycin to promote these metabolic pathways (Pearce et al., 2009; Xu et al., 2014). It is now clear that Treg cells like CD8⁺ Tmem cells also depend on FAO (Michalek et al., 2011), indicating that the effects of rapamycin to promote the outgrowth of these cells is also the result of an induced metabolic bias. Deletion of TSC2 promotes constitutive mTORC1 activity by removing the molecular brake on Rheb and promotes the development CD8⁺ T effector cells associated with exaggerated glycolytic metabolism, with a related failure of Tmem cell development (Pollizzi et al., 2015). Interestingly, deletion of Tsc1 in T cells, while promoting enhanced mTORC1 activity and reduced mTORC2 activity, is required for naive T cell quiescence, leading to a loss of conventional T cell and invariant natural killer T (iNKT) cells and reduced T cell-mediated immunity (Yang et al., 2011). In contrast, Rheb deletion, which ablates mTORC1 activity, prevents CD8⁺ effector T cell (Teff) development and allows Tmem cell expansion, although these cells are incapable of developing into secondary Teff cells upon antigen challenge (Pollizzi et al., 2015). In this system, mTORC2 inhibition also promotes Tmem cell development. The differential effect of mTORC1 on Teff versus Tmem cell development raises the concept that metabolic reprogramming may underlie these fates.

Embedded in the early CD8⁺ T cell work were Raptor loss-of-function experiments that have revealed beneficial effects on CD8⁺ Tmem cell development by targeting mTORC1 (Araki et al., 2009). More recently, deletion of Rheb in CD8⁺ T cells has been shown to reduce glycolysis and increase FAO, leading to diminished Teff cell function but enhanced Tmem cell development. In contrast, Tsc2 deletion promotes the opposite phenotype: increased glycolysis, diminished FAO, exaggerated Teff cell function, and diminished Tmem cell development (Pollizzi et al., 2015). These data emphasize the pivotal role of mTORC1 in regulating metabolic reprogramming coupled to cell fate, with glycolysis associated with Teff cell function and FAO linked to Tmem cell development. In contrast, loss of mTORC2 activity by deletion of Rictor has no effect on CD8⁺ Teff cell function but results in long-lived cells that exhibit more robust Tmem cell characteristics than conventional memory T cells. This is associated with enhanced expression of Cpt1a, which regulates the rate-limiting step of mitochondrial FAO and enhanced spare respiratory capacity (Pollizzi et al., 2015), a mark of Tmem cells (van der Windt et al., 2012). These effects may be secondary to the ability of mTORC2 to inhibit FOXO-driven expression of the receptor for IL-15, which is known to regulate Tmem cell development by promoting FAO. These findings imply that an inhibitor capable of targeting mTORC2 might allow enhanced Tmem cell development in the absence of any deleterious effects on Teff cell function, a therapeutically attractive possibility for promoting long-lived T cell memory.

Dependent on environmental conditions, particularly the cytokine milieu generated by antigen-presenting cells (APCs) during immune priming, CD4⁺ T cells become activated and differentiate into functionally distinct T helper 1 (Th1), Th2, Th17, Th9, Treg, or T follicular helper (Tfh) cells. Both rapamycin and genetic loss of mTOR have been shown to inhibit Th1, Th2, and Th17 Teff cell differentiation while promoting Treg cell differentiation (Delgoffe et al., 2009). In general this reflects the role of mTOR in promoting glycolysis and repressing FAO, which are critical metabolic pathways for the development of Teff and Treg cells, respectively (Michalek et al., 2011; Buck et al., 2015). In terms of the relative roles of mTORC1 and mTORC2 in Th cell differentiation, there is consensus in some areas but not in others. For instance, there is agreement from experiments in which Raptor or Rictor were deleted, that mTORC1 but not mTORC2 is essential for Th17 cell differentiation (Delgoffe et al., 2011; Kurebayashi et al., 2012; Lee et al., 2010). Rheb deletion also supports a requirement for mTORC1 in Th1 cell differentiation (Delgoffe et al., 2011). However, these data have not recapitulated in studies using Raptor deletion (Kurebayashi et al., 2012). Moreover, through Rictor deletion, mTORC2 has also been implicated in Th1 cell development (Lee et al., 2010).

A simple interpretation of the differential requirements for Rheb, Raptor, and Rictor in Th1 cell differentiation is that they reflect distinct effects in the different systems of Cre-mediated excision driven by Cd4 or Lck promoters, which are temporally separated during T cell development. A more interesting possibility, yet to be fully explored in T cell biology, is that Raptor and Rheb exert different effects on mTORC1 function. One possibility is that Rheb or Raptor mutations differentially affect the translation of mRNAs that define Th cell differentiation fate, as observed in Treg cells (Bjur et al., 2013). More controversial is the direct impact of mTORC1 on transcription. Studies have shown mTORC1 nuclear localization and association with chromatin (Cunningham et al., 2007; Tsang et al., 2010), and recent work has demonstrated mTORC1 localization to TCA cycle and lipid biosynthesis genes (Chaveroux et al., 2013), both of which may be affected by Raptor deletion.

The relative roles of mTORC1 and mTORC2 in Th2 cell differentiation is somewhat unclear. While one report has implicated Raptor (but not Rheb) in Th2 cell differentiation (Yang et al., 2013), the majority of reports claim a critical role for mTORC2 in these cells (Delgoffe et al., 2011; Heikamp et al., 2014; MacIver et al., 2013). Interestingly, signaling downstream of mTORC2 in Th2 cells has been reported to be mediated by PKCq rather than Akt (Lee et al., 2010; Yang et al., 2013), a finding that is consistent with earlier reports linking PKCq with Th2 cell effector responses (Marsland et al., 2004; Salek-Ardakani et al., 2004). In contrast, in Th1 cells mTORC2 has been reported to signal through Akt (Lee et al., 2010). In contrast to the situation for Th1 and Th17 cells, which exert their functions primarily within non-lymphoid tissues and which rely on one or other of the mTOR complexes for their differentiation, Tfh cells, which are restricted to germinal centers in secondary lymphoid organs, are dependent on both mTORC1 and mTORC2, since Tnfrsf4-cre-mediated deletion of Raptor or Rictor is sufficient to prevent their development (Zeng et al., 2016).

Treg cells are among the most common tissue-resident lymphocytes, providing protection against excessive inflammatory or autoimmune responses in tissues. The long-lived nature of tissue-resident Treg cells infers that they must manage metabolic resources effectively, and so perhaps it is not surprising that Treg cells maintain a more oxidative metabolic profile than Teff cells (Michalek et al., 2011). Results indicating that mTORC1 signaling is dispensable for Treg cell differentiation are consistent with findings that Treg cells are less glycolytic than effector CD4⁺ T cells and the recognized role of mTORC1 in promoting glycolysis (Düvel et al., 2010). However, detailed analysis of mTORC1 activity has indicated that there is a dynamic role for mTORC1 in Treg cell differentiation and proliferation, such that oscillating waves of mTORC1 activity allow Treg cell proliferation at the expense of suppressive activity (Procaccini et al., 2010). In a similar vein, recent work has shown that signaling through toll-like receptors (TLRs) on Treg cells promotes glycolytic metabolism and cellular proliferation at the expense of suppressive ability through an mTORC1-dependent pathway (Gerriets et al., 2016). Thus, mTORC1 and Foxp3 drive opposite T cell fates through enforcement of distinct metabolic programs: Foxp3 suppresses anabolic metabolism and boosts catabolic pathways in Treg cells to reinforce Treg cell suppressive capacity, whereas triggering mTORC1 activity in Treg cells promotes their expansion at the expense of suppressive capacity. Loss of both mTORC1 and mTORC2 promotes Treg cell differentiation as a default fate, indicating a commonality between these two pathways in this regard (Delgoffe et al., 2011), possibly due to the ability of mTORC2 to stimulate Akt-dependent glycolytic metabolism (e.g., Albert et al., 2016; Huang et al., 2016).

A critical step forward in understanding fate determination in T cells comes from the finding that initial cell division after activation is asymmetric. Seminal work by Chang et al. (2007) has demonstrated that, during the first cell division, the daughter CD8⁺ T cell proximal to the point of stimulation is destined to become a Teff cell, whereas the distal daughter cell becomes a Tmem cell. Recent findings have revealed that this process is marked by the segregation of key molecular components into each daughter cell, with mTORC1, c-Myc, and amino acid transporters accumulating in the cell destined to become an effector cell (Figure 2). This is accompanied by an increase in glycolytic rate, and contrasts with the other daughter cell destined to become a Tmem cell, where mTORC1 and c-Myc are by comparison sparse and oxidative phosphorylation is emphasized (Adams et al., 2016; Pollizzi et al., 2016; Verbist et al., 2016). Asymmetric inheritance of mTORC1 has an impact on lymphocyte metabolic fitness, with mTORC1^lo daughter cells displaying increased SRC and better long-term in vivo survival than Teff cell counterparts with high mTORC1 activity (Pollizzi et al., 2016). These data argue that anabolic programs enhanced by mTORC1, while good for exponential growth, can be ultimately self-limiting. Thus, it may be advantageous for tissue-resident Tmem cells to maintain lower mTOR activity in favor of metabolic regulators such as AMPK, which can help lymphocytes manage stress and metabolic resources to promote long-term survival until recalled (Blagih et al., 2015). Consistent with this, activated B cells with asymmetric clustering of AMPK activity display increased mitochondrial turnover and escape immune senescence (Adams et al., 2016). AMPK may achieve this in part by suppressing mTORC1 activity in T cells (Blagih et al., 2015; Rolf et al., 2013). Inhibiting mTORC1 activity (or enhancing AMPK activity) at the time of T cell priming may shift the balance of Teff versus Tmem cell fate by limiting the impact of asymmetric segregation on metabolic reprogramming of differentiating daughter cells (Figure 2).

Figure 2. Asymmetric Lymphocyte Division Marked by Differential mTORC1 Segregation Can Dictate Cell Fate.

During the first cellular division of a naive CD8⁺ T cell (green), emerging daughter cells assume distinct properties, marked initially by the segregation of mTORC1 into the daughter cell proximal to the antigen-presenting cell. Amino acid transporters (Slc7a5, Slc1a5) are also enriched in proximal cells (red), promoting increased mTORC1 activation and mTORC1-dependent c-Myc expression. Distal daughter cells (yellow) display less mTORC1 activity and increased AMPK activity. This imbalance of key metabolic regulators in the emerging cells leads to metabolic differences in proximal and distal daughter cells such that the former is more glycolytic and anabolic, whereas the latter is more catabolic. Divergent metabolic profiles, enforced by asymmetric mTORC1 distribution, have far-reaching consequences for daughter cell fate, with proximal cells destined to assume effector functions (short-lived Teff cells) and distal cells more likely to persist as long-lived memory cells. Surface markers and intracellular signaling molecules associated with each state are highlighted.

In comparison to T cells, less is known about the roles of mTOR in B cells. To some extent this reflects challenges associated with reconstructing the essential structural and physiological aspects of germinal center (GC) conditions in tissue culture, and the corresponding difficulty of exploring B cell differentiation in vitro. Nevertheless, Mbl1-cre-driven deletion of Raptor has revealed that mTORC1 is critical for metabolic programming required to support B cell development and pre-B cell survival in vivo (Iwata et al., 2016). Moreover, data from experiments in which B cell Raptor has been deleted during the course of an ongoing immune response support an essential role for mTORC1 in B cell survival and class switch recombination as well as persistence of GCs and survival of plasma cells (Jones et al., 2016). Other studies reveal a role for mTOR in somatic hypermutation and antibody affinity (Zhang et al., 2013), which is consistent with the fact that rapamycin treatment allows the emergence of a less focused antibody response, which in some infections can have beneficial effects due to the increased potential for the production of protective cross reactive antibodies (Keating et al., 2013). There is evidence that mTORC2 also plays a role in antibody class switching (Limon et al., 2014) and mature B cell homeostasis (Lee et al., 2013), which may reflect contributions of Akt-Foxo1 signaling to this process. Readers are directed to a recent review on PI3K signaling in B cells for broader discussion of the role of mTOR in these cells (Jellusova and Rickert, 2016), as well as the impact of metabolic regulation on antibody responses by Boothby and Rickert (2017) in this issue.

mTOR in the Innate Immune System

Broadly speaking, the innate immune system is comprised of a set of different immune cell types (monocytes, macrophages, dendritic cells, neutrophils, eosinophils, mast cells, NK cells, innate lymphoid cells), some (the innate lymphoid cells [ILCs]) of which were discovered only within the last decade. Innate immune cells do not clonally express antigen-specific receptors, but rather engage in immune responses through a series of receptors that include pattern recognition receptors (PRRs) that allow the recognition of pathogens, or damage, and cytokine receptors that allow the cells to respond to messages being generated by other cells. In many, though not all, situations the activation of innate immune cells leads to inflammation, which in controlled settings is host protective and can set the scene for adaptive immune response development. The administration of rapamycin in various inflammatory settings in vivo has shown that mTORC1 can either promote or inhibit inflammation, depending on the setting (Cejka et al., 2010; Kirsch et al., 2012), illustrating that the role of mTOR in innate immunity is complex and context dependent. In this section we will discuss the roles of mTORC1 and mTORC2 in the biology of innate immune cells. Readers are directed to a comprehensive recent review on this subject for additional information (Weichhart et al., 2015).

Macrophages are seeded into tissues during embryogenesis but can also differentiate from monocytes that have entered tissues, a process which can occur in homeostatic conditions or due to damage and/or infection. Cells in the monocyte-macrophage lineage shift from resting to variously activated states in response to ligation of pattern recognition receptors or cytokine receptors, and consequently can play roles in a spectrum of critical functions such as wound healing, adipose tissue homeostasis, and control of infections and cancers. Stimulation of monocytes or macrophages with PRR agonists from bacteria typically induces the production of proinflammatory mediators, such as IL-6, IL-12, IL-23, and TNF-α in combination with the anti-inflammatory cytokine IL-10, which in an autocrine and paracrine fashion can temper the activation state. The monocyte to macrophage transition is inhibited in the absence of Rheb in vivo (Wang et al., 2016) and in MCSF-1-stimulated cultures of bone marrow from mice in which mTOR is deleted in the macrophage lineage (Hallowell et al., 2017). However, perhaps surprisingly, mTOR deletion in this lineage has not been found to have overt effects on the numbers of macrophages in peripheral tissues (Hallowell et al., 2017).

mTORC1 plays a role in shaping the balance of pro- and anti-inflammatory effects in monocytes and macrophages. Inhibition of mTORC1 with rapamycin promotes proinflammatory cytokine production while inhibiting IL-10 production, whereas mTORC1 activation due to TSC2 deletion in monocytes results in enhanced production of IL-10 and diminished production of proinflammatory cytokines in response to LPS (Chen et al., 2012; Mercalli et al., 2013; Weichhart et al., 2008). Recent work has shown that chronic mTORC1 signaling in macrophages due to TSC2 deletion promotes macrophage hypertrophy and proliferation, leading to excessive granuloma formation in vivo (Linke et al., 2017), indicating that excessive mTORC1 activity in macrophages can promote immunopathology.

Early work has shown that mTORC1 limits proinflammatory cytokine production by inhibiting the activation of NF-κB (Weichhart et al., 2008), the transcription factor that promotes inflammatory cytokine production. The understanding of this facet of mTORC1 biology has come into focus recently with the finding that, in macrophages, ligation of TLR4 by bacterial LPS activates a PI3Kγ-Akt-mTORC1-dependent pathway to promote C/EBPβ transcription factor activation and thereby the expression of genes that encode immunosuppressive proteins, such as Arginase 1 and IL-10 (Kaneda et al., 2016). Inhibition of PI3Kγ reduces mTORC1 activity and consequently increases NF-κB activity, marked by increased IKKβ phosphorylation with increased degradation of IkBα, sustained p65-RelA phosphorylation with increased RelA binding to DNA, and subsequent inflammatory cytokine production. In this way, inactivation of PI3Kγ in macrophages allows them to take on an immunostimulatory role and promote anti-tumor CD8⁺ T cells responses (Kaneda et al., 2016).

Consistent with these observations, loss of TSC2 function in monocyte-derived macrophages results in diminished IL-12 production and enhanced IL-10 production, correlating in vivo with an enhanced effect on tumor growth that is a reflection of the increased angiogenic properties of these cells (Chen et al., 2012). In contrast, rapamycin or TSC2 overexpression-mediated inhibition of mTORC1 results in enhanced IL-12 production, diminished IL-10 production, and the ability of cells to inhibit angiogenesis in tumors (Chen et al., 2012). The effects of mTORC1 on IL-10 production are linked indirectly to increased phosphorylation of STAT3, an mTORC1 substrate, but also to increased degradation of programmed cell death protein 4 (Pdcd4), which then releases the Twist family transcription factor 2 (Twist2) to promote Maf-dependent expression of IL-10 (van den Bosch et al., 2014).

While data from many studies indicate that mTORC1 (via TSC2 deletion) can limit inflammatory activation, other studies using bone marrow-derived macrophages (BMDMs) have revealed that TSC1 deletion, which promotes constitutive mTORC1 activity, can promote enhanced inflammatory activation in response to LPS (Byles et al., 2013; Saric et al., 2016). These data may reflect specific functions for TSC1 versus TSC2 in macrophages. Although most evidence points to TSC1 and TSC2 operating in a common pathway to regulate mTORC1 activity, some phenotypic differences between Tsc1^+/− and Tsc2^+/− animals have been observed (Kobayashi et al., 2001). Relevant to this discussion, and consistent with a role for mTORC1 in inflammatory activation, Rheb deletion in macrophages in an OVA-induced asthma model results in diminished inflammatory activation and enhances “alternative activation,” which in this setting caused more severe disease (Li et al., 2017).

Macrophage alternative (or M2) activation is induced by signaling through the IL-4R via STAT6. These cells have a regulatory role associated with the suppression of adaptive immune responses, tissue healing, and homeostasis, but are also linked to allergic and helminthic diseases (Van Dyken and Locksley, 2013). M2 macrophages express a set of genes that defines the alternatively activated state, including Retnla, Arg1, Chil3, and Mrc1. TSC1-deficient macrophages are found to fail to alternatively activate in response to IL-4, reflecting the fact that pronounced mTORC1 signaling is a strong negative regulator of alternative activation (see also Li et al., 2017). This was not due to effects on STAT6 phosphorylation, but rather to feedback inhibition of Akt phosphorylation through effects on insulin receptor substrate 2 (IRS2), which is also engaged by IL4R signaling (Byles et al., 2013; Wills-Karp and Finkelman, 2008). This pathway leads to Akt activation at both Thr-308 and Ser-473. Interestingly, in TSC1-deficient macrophages, low-dose rapamycin (which preferentially inhibits mTORC1 and not mTORC2) allows alternative activation in response to IL-4. However, recent work using higher concentrations of rapamycin-, which also inhibit mTORC2, Torin- (which inhibits both mTORC1 and mTORC2), and Rictor-deficient macrophages has revealed that mTORC2 is critical for alternative activation (Hallowell et al., 2017; Huang et al., 2016). Activation of mTORC2 and Akt Ser-473 phosphorylation is driven not only by IL-4R but also by events downstream of CSF-1R. In this manner, mTORC2 acts in parallel with pSTAT6 to promote alternative activation, in part by cooperating in events that lead to expression of IRF4 (Figure 1; Huang et al., 2016). mTORC2 may additionally serve to inhibit the synthesis of suppressor of cytokine signaling-1 (SOCS1) and SOCS5, both of which can inhibit STAT6 phosphorylation (Hallowell et al., 2017). Interestingly, deletion of Rictor in macrophages has no marked effect on inflammatory activation driven by TLR agonists, either by themselves or in combination with IFN-γ or ATP (Hallowell et al., 2017; Huang et al., 2016).

Despite a clear role for mTORC2 in alternative activation, there is also an impact of mTORC1 in this pathway, primarily through its ability to modulate IL-4-driven effects in the context of its canonical role as a sensor of amino acid availability. Mechanistically, this involves the ability of mTORC1 to promote the effects of Akt on ATP-citrate lyase (Acly), thereby affecting gene-specific histone acetylation and the regulation of expression of a subset of genes that mark the alternatively activated state (Covarrubias et al., 2015). Consistent with this, in separate studies, Lamtor1 (a key component of the amino acid-sensing machinery), is found to play an essential role in the ability of macrophages to become alternatively activated in response to IL-4 (Kimura et al., 2016).

As in T cells, mTOR plays a dominant role in the regulation of metabolism. In macrophages, the inflammatory phenotype associated with stimulation by TLR agonists is accompanied by enhanced glycolysis, and under the most polarizing conditions, Warburg metabolism, whereas alternative activation induced by IL-4 is associated with enhanced FAO and accompanying increases in glycolysis (O’Neill and Hardie, 2013). In both cases, inhibition of metabolic reprogramming inhibits activation. Of particular interest in this regard is the finding that stimulation of human monocyte-derived macrophages with β-glucan, which is a fungal ligand for Dectin1, followed by a period of rest, allows cells to subsequently exhibit an exaggerated proinflammatory cytokine response to the same ligand or to unrelated TLR agonists (Cheng et al., 2014). This process of trained immunity is accompanied by a switch to Warburg metabolism and increased epigenetic marks at promoters of genes in the mTOR pathway (see the review of metabolism and immune cell genomics by Phan et al., 2017, in this issue), increased expression of mTOR itself, and downstream targets of mTOR including targets of HIF1α. This process is inhibited by metformin, an activator of AMPK, or by low-dose rapamycin, strongly suggesting a dominant role for mTORC1 (Cheng et al., 2014).

One intriguing facet is that modulating mTOR activity in tissue-resident macrophages may have different health outcomes depending on their location. For example, macrophages associated with adipose tissue influence obesity-induced insulin resistance. Silencing mTORC1 (by Raptor deletion) in macrophages improves insulin sensitivity in mice fed a high-fat diet, and this is accompanied by reduced inflammatory cytokine production in the liver and adipose tissue (Jiang et al., 2014). Macrophages filled with lipid droplets (foamy cells) are proinflammatory and a mark of atherosclerosis. In this setting autophagy plays a positive role by promoting the clearance of intracellular lipid stores. In this condition, promoting autophagy by inhibiting mTORC1 with rapalogues or Raptor deletion reduces disease incidence that is associated with diminished chemokine gene expression and increased plasma LDL (Wang et al., 2014). There is reason to believe therefore that targeted inhibition of mTORC1 in macrophages could aid in preventing artherosclerosis.

Like macrophages, DCs populate tissues and exist in resting and activated states. The primary role of all conventional DC (cDC) subsets (distinct from plasmacytoid DCs [pDCs] which will be discussed below) is to act as a conduit through which T cells sense antigen. DCs proteolytically process proteins into peptides that associate with MHCI or MHCII glycoproteins and are displayed at the DC surface for recognition by T cells expressing the appropriate T cell receptors and co-receptors. In the steady state, this is not sufficient to activate T cells. However, in conditions where DCs have been activated through PRRs to upregulate expression of costimulatory molecules and secrete cytokines, interacting antigen-specific T cells are pushed toward full activation and differentiation into distinct effector subsets (as discussed above). In order to meet T cells, activated peripheral tissue DCs must migrate to secondary lymphoid organs where they join DCs resident in T cell-rich zones and together initiate T cell activation.

Signaling through Flt3 is critical for cDC and pDC development (Waskow et al., 2008). In culture, Flt3L is able to promote the differentiation of DC precursors in bone marrow into pDCs and equivalents of CD8⁺ and CD8⁻ lymphoid tissue-resident cDCs, and this process is inhibited by rapamycin at doses that preferentially target mTORC1 (Sathaliyawala et al., 2010), a finding that correlates with the ability of Flt3L to induce S6K phosphorylation in DCs. DCs can also differentiate from bone marrow under the influence of GM-CSF, but these cells are distinct from Flt3L-stimulated cells and their differentiation is not affected by rapamycin. Outgrowth of cDCs is accentuated in Flt3L-cutures by deletion of Pten, the negative regulator of PI3K and therefore of Akt signaling, and this process is inhibited by rapamycin. In vivo, deletion of Pten in the DC lineage results in excessive numbers of tissue-resident CD103⁺ DCs, which are closely related to lymphoid tissue-resident CD8⁺ DCs (Sathaliyawala et al., 2010). Deletion of late endosomal/lysosomal adaptor, MAPK and MTOR activator 2 (Lamtor2), which is a component of the Ragulator complex that is involved in amino acid sensing and mTOR activation (Sancak et al., 2010), has a similar effect, causing expansion of the cDC and pDC pools largely as a result of overexpression of surface Flt3, and therefore presumably excessive mTORC1 signaling (Scheffler et al., 2014). In this model inhibition of mTORC1 by rapamycin is able to reverse the defect conferred by Lamtor2 deletion. Recent evidence has implicated PI3Kγ upstream of mTORC 1 in the development of CD103⁺ DCs in the lungs (Nobs et al., 2015).

As in macrophages, there is evidence that mTORC1 negatively regulates DC activation in response to TLR agonists. Early studies have shown that loss of PI3K function in DCs results in diminished IL-10 production and enhanced IL-12 production in response to TLR agonists (Fukao et al., 2002). Later, rapamycin was shown to have similar effects and in addition to increase costimulatory molecule expression and longevity in LPS-stimulated GM-CSF-derived DCs (Amiel et al., 2012). This is due in part to the ability of rapamycin to inhibit LPS-induced Nos2 expression in these cells, which thereby circumvents inhibitory effects of NO on respiration (Amiel et al., 2014). Intestinal CD11b⁺ DCs from Raptor-deficient mice are also less able to make IL-10 while simultaneously expressing more costimulatory molecules, a situation that predisposes toward intestinal inflammation (Ohtani et al., 2012). In Langerhans cells, skin-resident cells with characteristics of both macrophages and DCs, Raptor deficiency causes spontaneous migration from the skin to draining lymph nodes (Kellersch and Brocker, 2013). On the other hand, TSC1 deletion impairs proliferation and survival during DC development and therefore negatively affects pDC and cDC numbers in Flt3l bone marrow cultures and causes diminished DC numbers in vivo. This is associated with dysregulated Myc expression and metabolism that result in a spontaneous increase in expression of costimulatory molecules and in the expression of markers typical of other lineages (F4/80 and Ly6G) (Wang et al., 2013). These deficits are largely correctable by rapamycin treatment.

Functionally, pDCs differ from cDCs in that their primary role is to produce large amounts of type 1 interferons (IFNs). In these cells, rapamycin, or siRNAs that target S6K, or PI3K inhibitors all strongly inhibit CpG-induced type 1 IFN production (Cao et al., 2008). This occurs through a mechanism that prevents TLR9 interacting with MyD88, the key downstream protein in the signaling pathway that leads to IRF7 transcription factor phosphorylation and type 1 IFN production.

There have been few studies on the effects of mTORC2 on DC biology. However, LPS-stimulated mTORC2-deficient DCs display enhanced ability to promote Th1 and Th17 cell responses (Raïch-Regué et al., 2015), associated with increased production of IL-12 and IL-23 (Raïch-Regué et al., 2015; Wei et al., 2015). In addition, mTORC2-deficient DCs have an enhanced ability to promote anti-tumor CD8⁺ T cell responses when injected directly into tumors in mice (Raïch-Regué et al., 2016). The mechanisms underlying the increased immunogenicity of mTORC2-deficient DCs is not clear. However, given the fact that mTORC2 inhibition favors both DC immunogenicity and long-lived Teff cell function, selective inhibition of mTORC2 may have therapeutic value for boosting cell-mediated immunity in cases such as vaccination.

Neutrophils leave the Blood and enter tissues at sites of damage and inflammation, and there can contribute to inflammatory and anti-microbial responses. In these cells, loss of mTORC2 function caused by silencing of Rictor results in a failure of the cells to polarize in response to chemoattractants (Liu et al., 2010). Consistent with the known role of mTORC2 in the actin cytoskeleton (Jacinto et al., 2004), this effect in neutrophils is at least in part due to effects of Rictor (or mTOR, but not Raptor) deletion on the ability of Rac and Cdc42 to regulate actin assembly and organization at the leading edge that is required for directed movement (He et al., 2013).

Neutrophil effector functions against microbial pathogens include the production of extracellular traps (NETs). NET production is inhibited by HIF-1α silencing, and pharmacologic inhibition of mTOR (at high doses of Rapamycin that do not allow conclusions to be drawn about relative contributions of mTORC1 or mTORC2) emulates this due to resultant deficiencies in HIF1α expression (McInturff et al., 2012). Likewise, inhibition of mTOR (again at high doses of Rapamycin) prevents the ability of IL-23 to stimulate neutrophils to make IL-17 and IL-22 (Chen et al., 2016).

Mast cells in tissues align with Blood vessels. When antigen binds high-affinity FcRε1-bound IgE on the surface of mast cells, they rapidly degranulate to release cytokines, histamine, and a series of additional mediators that alter endothelial permeability, driving local inflammation and immunity. mTORC1 is activated downstream of FcRε1 signaling. However, rapamycin and other mTOR inhibitors have been found to block cytokine production, but not degranulation per se (Kim et al., 2008). mTORC1 also plays a role in mast cell survival, whereas mTORC2 has been found to play a role in immature mast cell proliferation (Smrz et al., 2014).

Through their ability to produce type 2 cytokines, eosinophils play important roles in adipose tissue beiging and liver and muscle regenerative responses, and in disease settings that are prominent components of various inflammatory diseases. Eosinophils are a tell-tale indicator of helminth infection, where they have been shown in certain settings to play host-protective roles. Despite their importance, little is known about the role of mTOR in their biology, other than the fact that rapamycin can inhibit the IL-5-dependent proliferation and differentiation of bone marrow precursors into eosinophils (Hua et al., 2015).

NK cells and iNKT cells are innate lymphoid cells with cytotoxic effector functions and produce large amounts of cytokines in response to cytokine or lipid antigen, respectively. mTORC1-mediated metabolic reprogramming of NK cells is a prerequisite for the acquisition of normal effector functions by these cells (Donnelly et al., 2014; Finlay, 2015). NK cell activation by innate signals such as poly(I:C) or cytokines such as IL-2 and IL-12 promote blastogenesis and a glycolytic shift to Warburg metabolism, both of which are rapamycin sensitive (Donnelly et al., 2014). IL-15 promotes NK cell development and activation during inflammation and infection through a rapamycin-sensitive mTOR-dependent pathway of development (Marçais et al., 2014; Yang et al., 2016) that requires regulation by TSC1 in order to prevent excessive NK cell activation and cell death (Wu et al., 2014; Yang et al., 2016). iNKT cell development is also highly dependent on mTORC1, as Raptor deficiency results in iNKT cells failing to reach the periphery from the thymus (Shin et al., 2014; Wei et al., 2014; Zhang et al., 2014). In contrast, deletion of TSC1 affects cellular function, resulting in the development of iNKT cells that make IL-17 rather than IFN-γ, a process that is mediated by the TSC1-dependent maintenance of Tbet expression. In contrast, mTORC2 activity is critical for iNKT cell development. Rictor is required for the survival and proliferation of iNKT cells at the earliest stages of development (Prevot et al., 2015) and regulates the development of IL-17-producing iNKT cells (NKT-17 cells) (Sklarz et al., 2017). Similarly, deletion of Tsc1, which enhances mTORC1 but decreases mTORC2 activity in T cells, leads to reduced iNKT cell survival (Yang et al., 2011).

An exciting and expanding area of immunology is understanding how tissue-resident ILCs regulate tissue homeostasis and first-line defense (Klose and Artis, 2016). Relatively little is known of mTOR function in these cells, and this is likely to be an active area of future research.

Metabolic Sensing by mTOR and Competition for Nutrients within Tissue-Specific Niches

Receptors for cytokines acting as growth factors, or that recognize non-self or altered self, converge on PI3K-Akt signaling to promote the activation of Rheb at the lysosomal surface. If amino acids are present at sufficient concentrations, particularly leucine, mTORC1 is also recruited to the lysosome where it can be activated by Rheb (Figure 1). The requirement for mTORC1 localization at the lysosome ensures that mTORC1 activation can proceed only in the presence of sufficient nutrient abundance that can support anabolic growth. In this regard, activation signals (i.e., TCR triggering plus costimulation) must be “licensed” by nutrient availability. Once triggered, mTORC1 initiates a coordinated anabolic metabolic program that supports cell growth and proliferation, and acquisition of effector functions. This is broadly consistent with the observed importance of mTORC1 in CD8⁺ Teff cell activation and in the differentiation and activation of some Th cell subsets. Proteomic analysis of Raptor-deficient T cells shows that expression of key transcription factors in metabolic reprogramming (MYC, YY1, GABPA, SREBF1) and components of the translational machinery are dependent upon mTORC1 (Tan et al., 2017), highlighting the central role of mTORC1 in anabolic T cell growth. Moreover, in cytotoxic CD8⁺ T cells, proteomic analyses have shown that rapamycin treatment to inhibit mTORC1 causes a reduction in cell size and protein content associated with changes in abundance (up as well as down) of approximately 10% of the proteome, including reductions in glucose transporters, and enzymes in the glycolysis and cholesterol synthesis metabolic pathways, as well as of the PtdIns(3,4,5)P3 phosphatase PTEN, and key effector proteins such as IFN-γ, perforin granzyme, and TNFα (Hukelmann et al., 2016). These changes are generally related to rapamycin-induced reductions in transcription of the encoding genes, although outside of these pathways clear mTORC1-dependent differences in protein levels in the absence of effects on transcription are measurable. Additional intracellular sensors can act to short circuit mTORC1 activation when environmental conditions are poor. For example, AMPKα1 actively suppresses the translation of IFN-γ mRNA in T cells under low glucose conditions (Blagih et al., 2015).

This model predicts that sufficient quantities of amino acids are essential for cellular activation of mTOR (Figure 3). The veracity of this argument is illustrated by early findings, which predated the discovery of mTOR, demonstrating that the absence of just one of the branch chain amino acids from tissue culture medium was sufficient to inhibit protein synthesis in lymphocytes (Calder, 2006). Amino acid limitation can also be anticipated in certain infections, where prokaryoctic and eukaryotic pathogens can compete for available amino acids (Gogoi et al., 2016), and in other disease such as liver cirrhosis, where plasma branched chain amino acid abundance is diminished (Kakazu et al., 2007). Recently, Sinclair et al. (2013) has demonstrated that Slc7a5, a transporter for large neutral amino acids that is upregulated in activated T cells, is required for full T cell activation, and that Slc7a5-deficient T cells fail to respond to antigen due in part to loss of mTORC1-dependent metabolic reprogramming. These observations raise key questions about whether there is critical competition for amino acids within tissues. At one level, the ability to acquire nutrients will be regulated by the expression of amino acid transporters, as is likely the case during asymmetric T cell division, where mTORC1 and amino acid transporter expression is biased toward the proximal daughter cell destined for a Teff cell fate (Figure 2). On the other hand, neighboring cells may deplete amino acids that are essential for mTORC1 activation in tissue-resident immune cells (Figure 3). One scenario in which this occurs is when activated innate immune cells expressing Arginase 1 consume arginine, and in this way assume a regulatory role by depriving T cells of arginine and thereby blunting their activation (Rodriguez et al., 2017). Mechanistically, arginine sensing in T cells has been reported to depend on the binding of arginine, transported via Slc38a9, to Castor1, thereby disrupting its inhibitory interaction with Gator2, a protein that positively regulates the RAG GTPase-mediated recruitment of mTORC1 to lysosomes (Chantranupong et al., 2016). Depriving T cells of arginine is one of the core mechanisms through which myeloid-derived suppressor cells (MDSCs) and alternatively activated macrophages limit Teff cell function (Figure 3; Murray, 2016; Rodriguez et al., 2017).

Figure 3. Amino Acid Sensing Is a Critical Control Point for Determining mTORC1 Activation, Metabolic Reprogramming, and Cellular Fate.

Amino acid sufficiency, particularly branched chain amino acids (BCAAs) and arginine, is sensed by immune cells through a process coupled to mTORC1 activation. When mTORC1 is active, cellular metabolism is anabolic, and catabolic pathways are inhibited. This allows cellular growth, proliferation, and the development of effector functions to proceed. Amino acid limitation, due to low supply in the microenvironment or consumption by other immune cell types (M2 alternatively activated macrophages, MDSCs, or Treg cells), reduces mTORC1 activation and prevents metabolic reprogramming to support cellular activation. Expression by regulatory immune cells of enzymes that break down amino acids is a primary immune system intrinsic mechanism through which effector immune cell activation is regulated. Other events, such as infection or cancer, that lead to changes in amino acid availability, can effectively limit immune cell activation through the same pathway.

The loss of effector functions of tumor-infiltrating T cells (TILs) is often assigned to this type of negative regulation, but it is more broadly true that nutrient availability in general can be poor within tumors (Hirayama et al., 2009). In these settings, tumor-infiltrating lymphocytes (TILs) exhibit low S6K phosphorylation (Chang et al., 2015), and competition for amino acids as well as for glucose and other nutrients can negatively affect T cell mTORC1 activation and impact mTORC1-dependent metabolic reprogramming essential for TIL function and anti-tumor immune responses (Chang et al., 2015; Ho et al., 2015). Activating mutations in the mTORC1 pathway, such as loss of the PTEN and LKB1 tumor suppressors, are themselves oncogenic (Shackelford and Shaw, 2009; Song et al., 2012) and can account for the ability of cancer cells to effectively maintain anabolic metabolism and mTORC1 activity in settings where T cells struggle to activate this pathway and maintain cellular function. Conditioning T cells to adopt metabolic plasticity to maintain mTORC1 activity under poor growth conditions may be one way to enhance the response of TILs in the tumor microenvironment.

In comparison with the wealth of data on mTORC1, relatively little information is available for mTORC2. However, recent reports have shown that T cells deprived of arginine cell cycle arrest in G1 is mediated by mTORC2, which under these conditions acts to sense amino acid availability and control cellular proliferation (Van de Velde and Murray, 2016). In these experiments, arginine-deficient cells have shown to proliferate normally if Rictor was deleted, indicating that mTORC2 actively inhibits cell cycling, although the underlying mechanism is unclear. Our perception of how mTOR functions has come into sharper focus as we have realized the extent to which metabolic reprogramming underpins immune cell activation. In this light, a unifying model of mTORC1-mediated coordination of activating signals, amino acid availability, and metabolic reprogramming to support anabolic pathways is attractive, and for the most part fits the available data on immune cell activation. However, there are exceptions, and in these cases mTORC2 assumes a critical role, although how it does is not always clear. Understanding the impact of mTORC2 on immune cell function and how it helps shape immune responses is an exciting challenge for the near future. In this context, the availability of inhibitors with the specificity to discriminate between mTORC1 or mTORC2 would be of immense research benefit. Drugs of this type could also have great clinical potential in situations where the inhibition of specific types of immune cells, such as DCs or Tmem cells, is desirable.