Nutrient Inputs and Social Metabolic Control of T Cell Fate

Summary

Cells in multicellular organisms experience diverse neighbors, signals, and evolving physical environments that drive functional and metabolic demands. To maintain proper development and homeostasis while avoiding inappropriate cell proliferation or death, individual cells interact with their neighbors via “social” cues to share and partition available nutrients. Metabolic signals also contribute to cell fate by providing biochemical links between cell extrinsic signals and available resources. In addition to metabolic checkpoints that sense nutrients and directly supply molecular intermediates for biosynthetic pathways, many metabolites directly signal or provide the basis for post-translational modifications of target proteins and chromatin. In this review, we survey the landscape of T cell nutrient sensing and metabolic signaling that supports proper immunity while avoiding immunodeficiency or autoimmunity. The integration of cell extrinsic microenvironmental cues with cell intrinsic metabolic signaling provides a social metabolic control model to integrate cell signaling, metabolism, and fate.

eTOC Blurb

T cells require signals from their microenvironment for viability and drive activation and differentiation in the “social control” model. Available nutrients also contribute as metabolic checkpoints and signals through “social metabolic control” that integrates metabolic state with cell fate.

Introduction

Signals that guide cell survival, activation, and function depend on access to appropriate nutrients to meet evolving energetic and biosynthetic demands. Cell metabolism must, therefore, be dynamic and responsive to these extracellular cues. Sufficient nutrients, however, are not always available and can be systemically or locally limiting. High levels of waste products may also limit metabolism or cause cells to activate stress response pathways that ultimately result in cell death. Conversely, excess nutrients should not induce cell proliferation in multicellular organisms. Single cells in complex tissues must therefore share and distribute nutrients for the good of the tissue and organism. These limitations based on microenvironmental nutrient availability are further complicated as distinct cell types in the same tissues often use different nutrients [1]. The “social control” model of development proposed by Martin Raff posits that all nucleated cells in multicellular organisms rely on their cell neighbors for growth factor signals to proliferate or even survive [2]. Each cell, thus, depends on a social network of other cells for signals that balance tissue development and homeostasis. A revised model of “social metabolic control”, however, suggests that metabolites and nutrients in a cell’s surroundings play key roles in concert with growth factor cues to ultimately program cell metabolism and fate [3].

While Raff’s social control model focused on signaling, cells sense and share local nutrients to direct signaling and gene expression[3]. Indeed, metabolites and metabolic pathways are key to these processes and provide the biochemistry of cell biology and epigenetics. There are three general themes that illustrate social control through the lens of cell metabolism as social metabolic control (Figure 1). First, growth factor signals promote specific metabolic programs alongside processes that guide cell survival, proliferation, or differentiation. Second, established metabolic checkpoints sense specific nutrients or waste products in their microenvironment and modulate signals in their presence or absence. Third, metabolites themselves signal as ligands, co-factors, and substrates for post-translational and epigenetic modifications. In these themes of metabolic signaling, nutrient availability and metabolites directly influence cell signaling and gene expression. The importance of these pathways is clear, as failure to sense nutrients can prevent growth or lead to degenerative diseases while excessive nutrient signaling can drive cell proliferation and hyperplasia. T cells, for example, migrate to different tissue sites and must tailor their metabolism to local nutrient microenvironments with various degrees of damage, hypoxia, or inflammation.

Figure 1. Social metabolic control of cell signaling and fate.

Cells in complex tissues are dependent on cues from neighboring cells and microenvironment to regulate gene expression and survival. These social signals incorporate direct growth factor signals to promote specific gene expression programs, sensing of nutrients and waste products that can provide metabolic checkpoints if conditions are not appropriate, and a requirement for metabolites in signaling and as post-translational epigenetic modifications. These processes are connected, as growth factor stimuli also directly promote metabolic pathways necessary for the proscribed cell fate. Social metabolic control ensures that cells, such as T cells, integrate metabolic and growth factor signaling mechanisms to support and shape cell responses when nutrients are adequate and prevent those responses when inadequate.

The local availability of nutrients and metabolites or metabolic processes within cells provides a key element by which social metabolic control regulates cell fate. Beyond simple nutrient concentrations, uptake and metabolic processing of these signaling metabolites also varies across cell types. Indeed, tumors have been proposed for nearly a century to consume more glucose than normal tissues [4]. This has been the basis for Positron Emission Tomography (PET) imaging in cancer diagnosis and monitoring. Fractionation of tumors into different cellular compartments, however, showed that cancer cells are poor at taking up glucose and instead consume amino acids and lipids. Immune cells, macrophages in particular, were the dominant consumers of intratumoral glucose [1]. These differences were programmed in each cell type rather than defined by limited nutrient availability. Here we review key aspects of nutrient and metabolic signaling in the context of social metabolic control. While these concepts are widely applicable, we highlight T cells and how metabolites shape how cells interpret their surroundings and how the social control model is fundamentally a metabolic model for homeostasis and disease.

Nutrient Sensing and Metabolic Checkpoints

The changing landscape of nutrients and cell demands necessitates that cells have mechanisms to assess the availability of those nutrients. For the purposes of this review, we define nutrients as extracellular molecules that can be used for cell growth and signaling, and metabolites as largely intracellular molecules used in metabolic pathways and processes but may also contribute to cell growth and signaling. Cells constantly evaluate their needs and sense nutrients available in their microenvironment [5–7]. Like cell cycle checkpoints when DNA replication is disrupted, metabolic checkpoints engage when essential nutrients are not sufficiently available to match stimulatory cues to proliferate or survive. Lack of sufficient nutrients may cause cells to revert to nutrient recycling and survival-focused strategies such as autophagy. In extreme cases, lack of nutrients may cause cells to undergo apoptosis or ultimately necrosis [8, 9]. Conversely, excess nutrients can drive pathways that may promote cell growth or nutrient storage. Nutrients are sensed in many ways that vary based on the nutrient and context, and several fundamental nutrient sensitive signaling pathways have been identified (Figure 2). In addition to direct receptor-mediated nutrient or metabolite sensing through G protein-coupled receptors or nuclear receptors, we highlight several key pathways that play critical roles in cell survival and fate.

Figure 2. Nutrient sensing pathways and metabolic checkpoints.

Nutrients and intracellular metabolites are directly sensed through a variety of pathways to ultimately regulate gene expression and cell fate. These mechanisms include direct recognition through cell surface or nuclear receptors, regulation of kinase signaling pathways, oxygen and ROS sensing, and generation of intermediates to control glycosylation and protein trafficking. Metabolic sensing proteins that stimulate growth are shown in green and those that inhibit growth are shown in red.

mTORC1 and AMPK in Energy and Amino Acid Balance

The PI3K-Akt-mTOR signaling axis provides a nutrient sensing super pathway for anabolic growth that is balanced by the catabolic AMP-activated protein kinase (AMPK) [10]. The mechanistic Target of Rapamycin kinase (mTOR) complex 1 (mTORC1) pathway is stimulated by nutrient abundance and inhibited by the Tuberous Sclerosis Complex proteins (TSC1 and TSC2). This pathway integrates signals from growth factor receptors that activate Phosphatidyl-inositide 3-Kinase (PI3K) and Akt signaling [11–13] to inhibit the TSC complex [14–16] with nutrient sensing of amino acids, glucose, and nucleosides [11, 14]. This “and gate” provides a metabolic checkpoint to ensure that signals are only transduced when nutrients match growth stimuli. When both growth signals and nutrients are present, the mTORC1 pathway potently stimulates anabolic metabolism. mTORC1 promotes induction or translation of master metabolic regulatory transcription factors such as MYC, sterol regulatory binding protein 1 (SREBP1), and other targets that promote gene expression for anabolic cell growth and proliferation. In contrast to mTORC1, AMPK is activated upon binding AMP and phosphorylation by Liver Protein B Kinase 1 (LKB1) when ATP levels become depleted [17]. Once activated, AMPK inhibits mTORC1-driven anabolic metabolism by phosphorylating TSC2 and the mTORC1 component Raptor. This shifts cells into a catabolic state and with enhanced autophagy that limits biosynthesis while maximizing energy generation.

T cell metabolism and fate are tightly regulated by the mTORC1 and AMPK signaling axis. Indeed, the mTORC1 inhibitor rapamycin and related compounds are common immunosuppressants. Amino acids or glucose may become limiting as T cells enter different microenvironments, such as tumors or inflamed tissues. These settings may restrict nutrients and thus prevent mTORC1 activation or deplete ATP to activate AMPK. The role of mTORC1 in different CD4 T cell subsets, however, is selective. A seminal study by Delgoffe and Powell showed that T cell-specific deletion of mTOR kinase impaired the differentiation of effector T cells while Treg differentiation remained intact [18]. Treg are not wholly independent of mTOR, however, as Treg-specific deletion of amino acid sensing components in the mTORC1 pathway reduced effector Treg abundance and led to a fatal autoimmune disease [19]. Conversely, glucose limitation to activate AMPK can reduce T cell anabolic metabolism but lead to increased immune memory [20]. This affect may be mediated in part through disruption of mitochondrial function, which can activate the kinase upstream of AMPK, LKB1 to promote catabolic metabolism and inhibit mTORC1 [21, 22]. Together, mTORC1 and AMPK are central and balanced regulators of T cell nutrient and energy sensing [23].

Oxygen Sensing and Hypoxia-Inducible Factors

Oxygen (O₂) is a molecular nutrient central to mitochondrial metabolism and energy generation. A key function of oxygen is to serve as an electron acceptor during ATP synthesis via oxidative phosphorylation as electron transport complex IV produces water. This water is then used to produce malate from fumarate in the tricarboxylic acid (TCA) cycle and acts as an electron acceptor to generate reactive oxygen species (ROS), such as hydrogen peroxide, superoxide, and hydroxyl radicals. The role of oxygen in the formation of ROS further entangles O₂ with metabolism, as free radicals modify lipids and proteins and directly interact with the glutathione detoxification pathway. Beyond these metabolic roles, O₂ is required by dioxygenase enzymes that metabolize the TCA cycle intermediate α-ketoglutarate (α-KG) to hydroxylate target proteins and produce succinate and CO₂ in reactions that mediate a wide range of functions related to oxygen sensing and epigenetic modifications. The synthesis of prostaglandins and leukotrienes, which are crucial signaling molecules in inflammatory responses, also require oxygen through Cyclooxygenase and Lipoxygenase.

Because of these roles for oxygen, cells have a robust capacity to monitor and adapt to O₂ availability through the Hypoxia Inducible Factor (HIF) family of transcription factors. Under normal conditions, prolyl hydroxylase proteins (PHDs) use oxygen to hydroxylate proline residues on HIF-α proteins. These hydroxylated prolines are then recognized by the E3 ligase pVHL, which targets HIF1/2-α for ubiquitin-dependent proteasomal degradation. When O₂ levels are insufficient, PHDs are unable to hydroxylate HIF1/2-α proteins, which enables their stabilization and dimerization with HIF-β for transport to the nucleus to dock at hypoxia response elements and drive transcription. Genes that are managed by HIFs are involved in a wide range of processes such as angiogenesis (ex. VEGF and EPO), glycolysis (ex. GLUT1, HKII, PDK1, and LDHA), apoptosis (BNIP3), proliferation (IGF2), and differentiation (Oct4). Activation of hypoxia signaling allows a cell to increase vascularization to promote blood flow and oxygen saturation. This program also promotes the uptake and metabolism of glucose to continue ATP production without utilizing oxidative phosphorylation. Oxygen levels and HIF signaling are tightly linked to glucose metabolism and offer the cell another method to sense and regulate glucose availability. PHDs and other dioxygenases are closely linked to the TCA cycle because their enzymatic activities are dependent on α-KG and generates succinate. Thus, the ratio of α-KG to succinate is crucial, and accumulation of succinate can inhibit HIF hydroxylation, preventing HIF degradation and leading to HIF-induced gene expression. PHD proteins are thus also affected by amino acid levels, where depleted conditions yield reduced α-KG, thereby inhibiting PHD activity and promoting HIF signaling [24]. PHD proteins are also activated by iron and ascorbate and inhibited by ROS [25, 26].

Because T cells migrate throughout the body and enter inflamed tissues and tumors, hypoxia may become a frequent microenvironmental challenge as tissue growth and metabolic demands exceed the capacity of vascular exchange to replenish O₂. Deletion of HIF1α selectively impaired Th17 cells to suggest this subset is especially sensitive to hypoxia [27, 28]. Conversely, increasing HIF1α activity may provide a broad benefit, as deletion of pVHL to stabilize HIF1α increased T cell function and maintained T cell polyfunctionality even in tumors [29]. Increased HIF1α activity may come with a cost, however, if T cells also experience chronic antigen receptor stimulation. Exhausted phenotypes were increased with combined hypoxia and antigen exposure, suggesting that tumor microenvironments may lead to T cell exhaustion in part through hypoxic signaling and chronic neoantigen stimulation [30].

NRF2 and Redox sensing

ROS are generated as byproducts of high rates or inefficient mitochondrial respiration and accumulate in cells with poor ability to neutralize these byproducts. While mediating beneficial mitochondrial outward signaling, ROS can also cause cellular damage, including DNA and lipid oxidation. Cells thus deploy protective mechanisms to detoxify and prevent the accumulation of free radicals. The transcription factor NRF2 plays a central role to sense this metabolic byproduct and maintain redox balance [31]. In homeostatic conditions, NRF2 is bound by KEAP1 and constitutively targeted for proteasomal degradation [32]. When ROS accumulate, cysteines on KEAP1 are oxidized and NRF2 is released and transported to the nucleus, where it recognizes antioxidant response elements[31, 33]. This allows NRF2 to induce a host of antioxidant genes, including superoxide dismutase, thioredoxin, peroxiredoxin, and genes involved in cysteine uptake and glutathione synthesis [34–36]. Inducing these transcriptional targets culminates in a robust ability to sequester and detoxify free radicals, minimizes cellular damage, and contributes to the role of this pathway in oncogenesis. Hypermethylation-dependent silencing of KEAP1 is commonly associated with lung adenocarcinomas [37, 38]. By silencing KEAP1, these tumors maintain active NRF2, which promotes the expression of pro-survival genes that help the cancer cells evade apoptosis when stressed or upon treatment [37, 39].

The NRF2/KEAP1 pathway and redox regulation play key roles in T cells. Glutathione acts as a central buffer to ROS for which the NRF2 transcriptional target Glutamate Cysteine Ligase Catalytic subunit (Gclc) provides a rate limiting synthetic step. While not essential early in T cell activation, Gclc is required to maintain glutathione production needed to buffer ROS and prevent inhibition of mTORC1 activity [40]. Ultimately, Gclc-deficient T cells do not proliferate or gain effector functions and mice are protected from Experimental Autoimmune Encephalomyelitis (EAE) and fail to generate anti-viral immune responses. Selective deletion of Gclc in Treg led to autoimmunity, as Treg increased serine uptake and glucose metabolism, which increased mTORC1 activity and decreased FoxP3 [41]. Conversely, deletion of KEAP1 in Treg stabilized Nrf2 and antioxidant mechanisms, but also led to autoimmune phenotypes with reduced FoxP3 [42]. Deletion of Nrf2 also reduced effector T cell-mediated inflammation in models of liver transplantation or graft-vs-host disease [43, 44]. Thus, cells must maintain a certain level of ROS, but too much is detrimental. This proper redox setpoint differs for each cell subset, with Treg and effector T cells each requiring distinct levels of ROS and NRF2 activity.

Hexosamine Biosynthetic Pathway

The hexosamine biosynthetic pathway (HBP) integrates glucose and glutamine availability and generates a precursor to glycosylate proteins and regulate protein trafficking and signaling. Upon entry into the HBP, glucose and glutamine are enzymatically converted into glucosamine via Glutamine fructose-6-phosphate amidotransferase (GFAT) [45, 46]. Following the input of acetyl-CoA and UTP, a series of reactions convert glucosamine into the end-product of the pathway, UDP-GlcNAc. Receiving precursory material from glycolysis, amino acids, fatty acid metabolism, and nucleotide metabolism, the HBP is an integrative nutrient sensing and signaling pathway that requires a range of inputs [47]. O- and N-linked GlcNAcylation of targets can alter protein localization, stability, and function. This nutrient-derived signal has been documented to directly modify many glycolytic and pentose phosphate pathway enzymes [48]. Changes to O-GlcNAcylation levels also regulate HIF-1α-dependent glucose uptake in breast cancer cells by modifying α-KG availability [49]. O-GlcNAcylation influences acetyl-CoA and lipid synthesis in several cancer models [50, 51]. In the absence of adequate glucose or glutamine for glycosylation, protein trafficking is altered to reshape the landscape of receptor-mediated signaling and expression patterns. T cells fail to traffic the IL-2 receptor to the cell surface if glycosylation is deficient, thus preventing a potent T cell growth and differentiation signal to shift T cell fate and favor inflammatory Th17 cells over suppressive regulatory T cells[52].

Metabolites as Signaling Molecules

In addition to providing fuel and building blocks for biosynthesis, metabolites are receptor ligands, secondary messengers, and biochemical sources of post-translational modifications (PTM). Our understanding of metabolites as signaling molecules has evolved as new techniques define how metabolites directly bind to, interact with, and alter proteins and nucleic acids structure and function [53]. Some nutrients or metabolites serve as ligands for cell surface G protein-coupled receptors or nuclear receptors, such as succinate or lactate and lipids, respectively (Figure 3). Metabolite binding can alter molecular interactions, such as polyamine binding to DNA to promote Z-form DNA that is poorly recognized by cGAS and therefore less inflammatory[54]. Secondary messengers such as cyclic AMP (cAMP), IP₃, and nitric oxide (NO) are prominent examples of metabolites that function as signaling molecules. Energy molecules such as ATP, GTP, NAD⁺, and NADP are directly engaged and required substrates or co-factors used to power enzymatic reactions, regulate downstream signaling cascades, and biosynthesis[55]. In other instances, such as in the case of UDP-GlcNAc and acetyl-CoA, metabolites are the fundamental units used to construct more complex molecules in glycosylation and lipid synthesis. Together, the cycling of metabolites and their use as messengers is central to cell signaling and has been comprehensively summarized[56]. The covalent modification of proteins with metabolites also provides cells with tools to post-translationally regulate targets and gene networks. Metabolites serve as the basis for most common PTMs, which can alter the conformation and localization of a protein to regulate ultimate interactome and function [50, 57–59].

Figure 3. Metabolites as signaling molecules.

Metabolites are critical signaling molecules and the sources of many post-translational modifications. These include phosphorylation, acetylation, methylation, and glycosylation that affect wide ranging molecular targets and gene expression patterns. Metabolite-dependent enzymes that modify proteins or DNA to stimulate growth are shown in green and those that inhibit growth are shown in red.

Since Conrad Waddington first described the epigenetic landscape, technological developments have enabled a growing appreciation for this dynamic layer of chromatin regulation. In the 1960s, perhaps the most understood and thoroughly investigated epigenetic events of methylation and acetylation of lysine residues on histone tails were reported [60, 61]. Depending on how these methyl and acetyl groups decorate the genomic landscape, the structure of chromatin either condenses or relaxes, impacting accessibility for transcription factors and gene expression [61–65]. More recently, previously unknown modifications, such as succinylation and lactylation, have been identified for which the functional consequences are still being explored [66]. These metabolic sources of epigenetic marks provide direct links between social metabolic control or nutrient availability and gene expression to control cell function and fate.

Technical advances in sequencing have confirmed the incorporation of hydroxyl, methyl, and formyl groups, among other modifications, directly into DNA bases [67]. Modified DNA bases influence the stability of the DNA double helix structure and determine which DNA-binding proteins associate with that genomic region and have been associated with cellular stress responses, DNA damage, and disease. This branch of epigenetics expands further into RNA molecules, also known as epitranscriptomics. To date, approximately 100 DNA and 150 RNA base modifications have been identified, many of which remain functionally uncharacterized [68]. Modifications to RNA lead to changes in charge, stability, docking by readers, writers, and erasers, and secondary structure that can facilitate three-dimensional rearrangement of associated complexes. This metabolite-based circuitry enables a cell to deliver widespread, global regulation of gene expression or highly specific responses that focus on a single gene or pathway.

Methylation

Epigenetic methylation is mediated by the addition of a single carbon methyl group to a nucleoside ring or histone lysine tail yet generating the substrate for this reaction requires a diverse range of nutrients. This is mediated through one-carbon metabolism, which consists of two integrated cycles across mitochondria and cytosol to combine glucose, folate, essential amino acids, and other micronutrients from dietary sources with de novo synthesis of other intermediates [69]. As glucose is enzymatically processed through glycolytic reactions, 3-phosphoglycerate can divert to become 3-phosphohydroxypyruvate and ultimately serine essential for one-carbon metabolism. Dietary or microbiome derived folate is essential for the folate cycle of one-carbon metabolism and is first enzymatically reduced to tetrahydrofolate (THF). The glucose-derived serine and THF combine at serine hydroxymethyltransferase (SHMT) and use vitamin B6 as a co-factor to produce glycine and 5,10-methylenetetrahydrofolate [70]. To complete the folate cycle, 5,10-methylenetetrahydrofolate is converted into 5-methyltetrahydrofolate, which initiates the methionine cycle of one carbon metabolism to produce the methyl donor in the synthesis of methionine [71]. ATP and methionine are then converted into S-adenosylmethionine (SAM), the primary donor for methylation events (Figure 3) [72]. Following methylation reactions, the resulting S-adenosylhomocysteine (SAH) is converted into homocysteine, which can enter the transsulfuration pathway for glutathione production or be recycled into methionine to continue one-carbon metabolism flux by consuming betaine and vitamin B12 [73, 74].

In contrast to SAM that is essential for methylation, de-methlyation reactions, such as mediated by Ten Eleven Translocation (TET) family of DNA-dioxygenases, require α-KG. Different cell types and subsets are differentially regulated by α-KG mediated epigenetic events. One explanation rests in the three-dimensional remodeling of chromatin in response to fluctuating α-KG levels. CCCTC-binding factor (CTCF) is a zinc finger DNA binding protein that spatially compartmentalizes the genome to fine-tune transcriptional regulation [75]. Recently, it was shown that the selectivity of CTCF is responsive to α-KG availability [76]. By analyzing IL-2 sensitivity in CD4⁺ Th1 and CD8⁺ cytotoxic lymphocytes, Chisolm et al. observed an accumulation of α-KG associated with IL-2 treatment. Exogenous α-KG selectively activates effector programs and glycolysis, while inhibiting the programs required for memory and T follicular helper (Tfh) cells. Accordingly, ChIP-seq for H3K27me3 following treatment with high IL-2, low IL-2, or low IL-2 + α-KG showed decreased overall methylation with α-KG, as well as subsets of α-KG sensitive genes. Notably, a significant number of α-KG inducible genes were in proximity to α-KG inducible CTCF binding sites. Lastly, α-KG inhibits methylation within the CTCF sites near α-KG responsive genes and to open chromatin and functionally induce gene expression. CTCF is instrumental to restructure the 3D genome and regulate metabolism to promote proper developmental progression [77]. Taking these studies together, α-KG and other metabolites influence the spatial organization of the genome and play major roles in determining precise responses to upstream signals.

Consistent with key roles for α-KG in chromatin remodeling and gene expression, activated T cells display elevated glutamine uptake [78–80]. Importantly, this pathway differentially regulates Th1 and Th17 cell subsets. In response to glutaminase (GLS1) inhibition with CB839, production of glutamate and α-KG decrease, and these metabolites can be come limiting[80, 81]. While Th1 cells adapt to decreased glutaminolysis with increased glycolysis, Th17 cells instead initiate metabolic checkpoints that broadly decrease metabolic activity. These opposite responses lead to corresponding differences in differentiation, as Th1 cells exhibit a decrease in H3K27me3 and increased effector differentiation, while CB839-treated Th17 cells have elevated H3K27me3 and decreased differentiation. The decreased differentiation and function of Th17 cells with GLS1 inhibition can protect against a wide range of inflammatory disease models[80, 82, 83]. Conversely, the increased differentiation observed in Th1 cells also occurs in cytotoxic CD8 T cells, in part by inhibiting recruitment of myeloid derived suppressor cells, which has been exploited in new approaches to enhance cancer immunotherapy [84, 85]. Increased effector differentiation may, however, come at a cost and ultimately increase exhaustion [80]. Together, these results suggest that individual cell subsets can be differentially regulated by a single metabolite through distinct metabolic pathways and epigenetic outputs.

Acetylation and Coenzyme A-linked modifications

Cytosolic acetyl-CoA is primarily derived from the TCA cycle metabolite citrate and extracellular acetate and CoA. In addition to a central role in mitochondrial metabolism, acetyl-CoA can donate acetate as a two-carbon unit for lipid and cholesterol biosynthesis and provide the substrate for acetylation modifications. CoA synthesis relies on a β-mercaptoethylamine group, pantothenate (vitamin B5), and ADP. Pantothenate is an essential nutrient that is taken up by cells where it undergoes an energy-laden, 5-step enzymatic conversion into CoA[86]. Once CoA synthesis is complete, it can combine with acyl groups through a high energy thioester linkage. In the case of acetyl-CoA, the sulfhydryl group-containing β-mercaptoethylamine portion of CoA bonds with an acetyl group. In nutrient-rich conditions, cytosolic acetyl-CoA is primarily generated from either glucose-sourced carbon via pyruvate or citrate derived from the TCA cycle by the activity of ATP citrate lyase (ACLY). Glutamine-derived carbon can also contribute to acetyl-CoA production via TCA cycle reversal, known as reductive carboxylation, to generate citrate that is typically associated with hypoxia or in cells with defective mitochondria [87]. Cells can also directly incorporate acetate from the diet or microbiome and transported into the cell via monocarboxylate transporters (MCTs) or be generated intrinsically through deacetylation reactions [88–90]. In a de novo pathway of acetate production, glucose-derived pyruvate is converted into acetate through a ROS-dependent reaction catalyzed by keto acid dehydrogenases [91]. Ultimately, intracellular acetate provides a substrate for the acetyl-CoA synthetase (ACSS) family of enzymes to perform the ATP-dependent union of acetate-derived acetyl groups and CoA to form cytosolic acetyl-CoA [92].

While acetyl-CoA is often the product of anabolic metabolism, catabolic reactions, such as the degradation of branched-chain amino acids and fatty acids, can also provide significant sources of cytosolic acetyl- and other acyl-CoAs [93, 94]. Leucine, for example, contributes to protein synthesis and cytosolic acetyl-CoA production [95]. Altogether, acetyl-CoA synthesis from both anabolic and catabolic reactions ensures that cells have multiple pathways to generate this crucial epigenetic substrate that can feedback and self-regulate metabolic behaviors. In addition to acetyl-CoA, several other acyl-CoA species, such as succinyl-CoA and lactyl-CoA, are involved in epigenetics (Figure 3) [66, 96]. Many other intermediates of fatty acid metabolism form acyl-CoA derivatives (i.e. butyrate, isobutyrate, propionate, crotonate, and malonate) and have also been described to modify histones[97]. These pathways may be particularly important in T cells, where mitochondrial citrate export and the malate aspartate shuttle play key roles to support histone acetylation in Th1 cell differentiation and effector function[98].

Glycosylation

Sugars are catabolized for energy and contribute to the synthesis of nucleic acids, amino acids, and various other cell components. Sugars also become linked to lipids via glycosidic bonds to provide critical interfaces for cell-cell contacts and other signaling mechanisms. The synthesis of signaling sugars such as N-Acetylglucosamine (GlcNAc), N-Acetylgalactosamine (GalNAc), and poly(ADP-ribose) (PAR) requires inputs from other major metabolic pathways. The GlcNAc sugar moiety is synthesized by the hexosamine biosynthetic pathway (HBP), a shunt of glycolysis where up to 5% of all glucose enters [46]. Glucose, glutamine, acetyl-CoA, and UTP feed into the HBP, and through a series of enzymatic conversions, the charged sugar UDP-GlcNAc is produced [46, 99–101]. UDP-GlcNAc is the substrate used for N- and O-GlcNAcylation modifications and the precursor for UDP-GalNAc via epimerase conversion. These charged sugars are intimately involved in the assembly of complex polysaccharide structures through N-linked glycosylation, which serves a notable signaling capacity. O-GlcNAc Transferase (OGT) can modify serine and threonine residues with GlcNAc as a PTM, which can be reversed by O-GlcNAcase (OGA) to regulate a variety of pathways, including proteolysis and mTOR activation [102] and epigenetic modification of histones and epigenetic modifiers [103–105]. An additional way that OGT and O-GlcNAc modifications can affect T cells is through the mTORC1 pathway. CD8 T cells lacking OGT had hyperactive mTORC1 and mitochondrial dysfunction[102]. In this way, O-GlcNAc provides a modification as feedback to moderate mTORC1 and maintain metabolic homeostasis.

Integrating nutrients, signaling, and T cells through social metabolic control

T cell activation and differentiation provide a model to illustrate nutrient sensing, metabolite-dependent signaling, and how extracellular stimuli integrate with available metabolites to guide gene expression and social metabolic control of cell fate [106]. Resting T cells rely on IL-7 from stromal cells to maintain a basal catabolic metabolism [107]. Similarly, Chimeric Antigen Receptor (CAR)-T cells expressing IL-7 maintain greater metabolic fitness [108]. T cells stimulated with co-stimulation by antigen presenting cells undergo metabolic reprogramming to increase nutrient uptake and induce a wide range of anabolic processes [109–112]. Importantly, CD4 T cells develop distinct metabolic profiles as they differentiate into specialized effector or regulatory subsets, with a series of metabolic checkpoints and signaling events guiding these various differentiation paths [113].

Access to nutrients provides the first set of metabolic checkpoints. Individual cells rely on programmed uptake of nutrients from their environment [1] and T cell activation leads to upregulation of a range of nutrient transporters [114]. Of these, glucose, methionine, and glutamine transporters have been shown to play essential roles to provide substrates for T cell metabolism and to regulate T cell differentiation. In the absence of the glucose transporter Glut1 (Slc2a1), glucose uptake is decreased, the mTORC1 pathway is not activated, and T cells fail to efficiently differentiate into inflammatory effectors [115]. Amino acid uptake is broadly increased in T cell activation, most notably by induction of the glutamine transporter ASCT2 (Slc1a5) [116]. If ASCT2-mediated glutamine uptake is insufficient, mTORC1 activity is diminished and effector T cells fail to differentiate. Treg, however, are less dependent on ASCT2 and mTORC1 and are maintained. Glutamine and glucose are also essential for the HBP pathway and GlcNAc synthesis [45]. Similarly, T cell receptor activation upregulates LAT1 (Slc7a5), which is essential for methionine uptake and supports the methionine cycle of one-carbon metabolism [117].

Uptake of these nutrients then contributes to a variety of signaling events. Glucose is a major substrate for de novo phosphoinositide synthesis that is essential for TCR and costimulatory signaling [118]. By generating pyruvate or fueling the TCA cycle and citrate, glucose supports cytosolic acetyl-CoA through PDH or ACLY that is used for protein modifications such as histone acetylation [119, 120], including the IFN-γ locus [121]. Glutamine metabolism through GLS supports anaplerotic α-KG for the TCA cycle and epigenetic regulation. GLS-deficient T cells alter ROS regulation and histone methylation patterns that promote mTORC1 signaling and differentiation of Th1 cells while suppressing Th17 cells [80]. Broader blockades of glutamine metabolism also affect T cell epigenetic marks to inhibit Th17 [122] and promote anti-tumor CD8 T cells [84, 85]. The one-carbon metabolism pathway plays a key role to regulate T cell epigenetics and differentiation, as MTHFD2 in the mitochondrial folate cycle is rate limiting for nucleotide synthesis and both DNA and histone methylation for Th17 cells [123]. Ultimately, these and many other metabolic signaling events shape T cell activation, proliferation, and differentiation.

Conclusions and Perspectives

Metabolic signaling through nutrient sensing and direct signaling by metabolites link cells to their microenvironment through social metabolic control [2, 3]. Fundamentally, epigenetic modifications integrate metabolic processes into regulatory codes on histones, nucleotide bases, or RNA molecules to integrate cell signaling and metabolism with gene expression and protein translation. The epigenetic landscape is dynamic and responds to upstream signals and environmental conditions, including nutrient availability, and provides feedback loops where these signals create layers of regulation over metabolic processes. In the context of therapeutic development and intervention, this reciprocity introduces another variable to an already complicated formula. However, it becomes more likely that we will be able to therapeutically exploit this regulatory interface as we continue to improve our understanding of how these processes cooperate.

The concept of social metabolic control integrates metabolic signaling with social control of cells[2] within distinct tissue microenvironments (Figure 1). This model builds on the dependence of cells on growth factor signals from their neighbors with the addition of the need for cells to sense and partition available nutrients. Epigenetic marks are key mediators of social metabolic control and better understanding of how metabolites and metabolic pathways influence gene expression will help explain the complex biology of T cell-dependent immunity, inflammation, and cancer. Many questions remain to understand how metabolism, microenvironment, and metabolic signaling control cell fate. How do different cells in the same tissue respond to the same microenvironment? If metabolic cues lead to epigenetic and chromatin modifications, how reversible may they be? And how might social metabolic control signals be targeted to modify cell fate and selectively impact specific cell types in specific microenvironmental conditions? Together, the social metabolic control model provides a framework to address these questions and understand the biology of how distinct cell types in each tissue may respond differently based on their nutrient access, growth signals, and fundamental biology.