CD8⁺ T cell exhaustion driven by metabolic reprogramming

Abstract

CD8⁺ T cells are central to adaptive immunity, with their function tightly linked to cellular metabolism. Naïve T cells rely on oxidative phosphorylation, effector T cells shift to aerobic glycolysis for rapid proliferation, and memory T cells depend on fatty acid oxidation and efficient mitochondrial respiration. In chronic infections and tumors, CD8⁺ T cells often enter an exhausted state, marked by sustained inhibitory checkpoint expression, impaired proliferation and cytotoxicity, and a distinct transcriptional profile driven by factors such as TOX. The tumor microenvironment (TME) imposes metabolic constraints-nutrient competition, lipid accumulation, hypoxia, and metabolite stress-promoting exhaustion. Key metabolic dysregulations include reduced glycolysis, mitochondrial dysfunction, lipid-induced ferroptosis, amino acid scarcity, and hypoxia-mediated HIF-1α signaling. These factors create a self-reinforcing metabolic imbalance that limits T cell plasticity and anti-tumor activity. Emerging strategies aim to restore T cell function by targeting metabolism, including enhancing glycolysis and mitochondrial fitness, modulating lipid and amino acid pathways, and combining metabolic interventions with immune checkpoint blockade or adoptive cell therapy. Understanding the interplay between metabolism and T cell exhaustion provides opportunities to improve durable anti-tumor immunity. This review highlights recent advances in CD8⁺ T cell metabolic regulation, the impact of the TME, and therapeutic approaches to reinvigorate exhausted T cells.

Introduction

T cells play a central role in adaptive immune responses. Based on their developmental and functional states, T cells can be classified into multiple subsets, including naïve, effector, memory, and exhausted T cells [1]. Naïve T cells reside in a quiescent state, patrolling peripheral lymphoid organs, and primarily rely on mitochondrial oxidative phosphorylation (OXPHOS) to maintain low metabolic activity and long-term survival [1]. Upon antigenic stimulation, T cells rapidly proliferate and differentiate into effector phenotypes, a process accompanied by a metabolic shift from OXPHOS-dominant to glycolysis-dominant pathways [2]. This metabolic reprogramming meets the high energetic and biosynthetic demands required for rapid proliferation and immediate effector function [3]. Following pathogen clearance, a subset of effector T cells transitions into memory T cells, which restore their metabolic profile and rely on fatty acid oxidation (FAO) and efficient mitochondrial function to achieve long-term survival and rapid responsiveness [3]. Collectively, these observations indicate that distinct T cell fates and functional states are closely linked to specific metabolic profiles, whereby metabolism not only provides energy but also actively regulates cell fate and function through metabolic intermediates and signaling pathways, such as mTOR, AMPK, and HIF-1α [4]. Under conditions of chronic antigen stimulation, such as persistent infections or tumors, CD8⁺ T cells may progressively enter a state of functional decline known as T cell exhaustion [4]. Exhausted T cells are characterized by sustained high expression of multiple inhibitory immune checkpoint molecules, including PD-1, TIM-3, and LAG-3, as well as a distinctive transcriptional program mediated by factors such as TOX that drive epigenetic and transcriptional remodeling [5, 6]. Functionally, exhausted T cells exhibit limited proliferation, markedly reduced cytotoxicity and cytokine production, impaired memory formation, and decreased responsiveness to immune checkpoint blockade [6]. Importantly, recent studies have revealed that exhaustion is not a singular static endpoint but a heterogeneous and stageable process. Early-stage, partially reversible “pre-exhausted/progenitor” phenotypes and late-stage deeply exhausted phenotypes differ substantially in metabolic characteristics, transcriptional landscapes, and plasticity [6]. The development of exhaustion is influenced not only by persistent T cell receptor (TCR) stimulation, inflammatory microenvironments, and inhibitory signaling, but also profoundly by intracellular and extracellular metabolic states.

Metabolic reprogramming within the tumor microenvironment (TME) represents a key non-genetic factor driving T cell exhaustion and immunosuppression [7]. Malignant cells, to support rapid growth and invasiveness, often reconfigure their energy metabolism and biosynthetic pathways, displaying enhanced glycolysis (the Warburg effect), altered lipid metabolism, amino acid utilization, and mitochondrial dysfunction, while concurrently producing and accumulating metabolic byproducts such as lactate and adenosine, resulting in local acidification and hypoxia [7]. These changes not only satisfy the physiological demands of tumor cells but also impair tumor-infiltrating T cells through dual mechanisms of “resource competition” and “inhibitory metabolic byproducts.” Specifically, glucose deprivation caused by tumor consumption limits T cell glycolysis; lactate accumulation and acidic pH suppress T cell activity and adhesion/migration; lipid overload and lipid peroxidation induce T cell dysfunction [7]; amino acid scarcity (e.g., arginine, tryptophan) hinders proliferation and cytokine synthesis; and chronic hypoxia triggers HIF-1α-mediated metabolic and immune regulatory programs, thereby promoting immune checkpoint expression and accelerating the acquisition of exhaustion phenotypes [8]. Concurrently, exhausted T cells themselves exhibit impaired mitochondrial function, reactive oxygen species (ROS) accumulation, and loss of metabolic flexibility [3]. This combination of intrinsic and extrinsic factors creates a vicious cycle of “metabolic imbalance,” which is difficult to reverse through singular immune activation strategies.

Based on these insights, recent studies have explored metabolic modulation as a novel strategy to reverse or delay T cell exhaustion [3]. Approaches such as directly restoring T cell mitochondrial function, enhancing metabolic flexibility, or targeting tumor metabolic pathways-such as inhibiting tumor glycolysis, reducing lactate accumulation, and supplementing or restricting specific amino acid pathways-can partially restore T cell energy metabolism and antitumor function. Moreover, combining metabolic interventions with immune checkpoint inhibitors (ICIs) or cellular therapies (e.g., CAR-T cells) has demonstrated potential synergistic effects, offering new avenues to overcome clinical immunotherapy resistance and improve therapeutic efficacy [3].

In this review, we aim to systematically summarize recent advances in understanding how tumor metabolic reprogramming regulates CD8⁺ T cell exhaustion. We focus on key metabolic dimensions, including glucose metabolism, lipid metabolism, mitochondrial metabolism, amino acid metabolism, hypoxia, and metabolic waste accumulation, and discuss their mechanistic impact on T cell fate and function. Furthermore, we critically evaluate recent progress and challenges in integrating metabolic-targeted strategies with existing immunotherapies. Finally, we highlight future research directions, including the application of single-cell metabolomics to reveal TME metabolic heterogeneity, the interplay between metabolism and epigenetics, and the rational design of metabolism-immunotherapy combination strategies to achieve more durable antitumor immune responses.

T cell subsets and functional states

Naïve T cells

Naïve T cells originate from thymically matured T cell precursors and reside in peripheral blood and secondary lymphoid organs in a quiescent state, having not yet encountered their specific antigens [9]. Their primary functions are survival and immune surveillance, with relatively low demands for energy and biosynthetic precursors. Metabolically, naïve T cells rely predominantly on mitochondrial OXPHOS and a minor contribution from FAO, generating ATP at low levels to maintain homeostasis rather than support rapid proliferation [9]. At this stage, T cells exhibit high metabolic flexibility, enabling swift metabolic reprogramming upon activation (Fig. 1).

Fig. 1.

Metabolic reprogramming during T cell differentiation and exhaustion Naïve T cells primarily rely on oxidative phosphorylation (OXPHOS) for energy production. Upon antigen stimulation, T cells differentiate into effector T cells, which shift their metabolism toward aerobic glycolysis to support rapid proliferation and effector functions. Effector CD8⁺ T cells exhibit notable functional plasticity and can be reprogrammed into CD8⁺ regulatory T cells (Tregs) under specific immunological conditions, such as chronic antigen stimulation or exposure to tolerogenic cytokine environments. These induced CD8⁺ Tregs exert immunosuppressive functions not only through cell–cell contact–dependent mechanisms but also via the secretion of inhibitory cytokines, including IL-10, TGF-β, and in certain contexts IFN-γ, thereby contributing to the maintenance of immune homeostasis and peripheral tolerance. After pathogen clearance, a subset of effector T cells transitions into memory T cells, which depend on fatty acid oxidation and OXPHOS for long-term survival and rapid recall responses. Under conditions of chronic antigen exposure, such as in tumor microenvironments or chronic viral infections (e.g., LCMV), T cells undergo metabolic dysfunction and become exhausted, characterized by impaired mitochondrial metabolism and loss of effector function

Upon antigen recognition and receipt of co-stimulatory signals, naïve T cells undergo activation, proliferation, and effector differentiation [10]. During this process, a significant metabolic shift occurs: cells transition from an OXPHOS-dominant “quiescent metabolic state” to a glycolysis-dominant “high-activity metabolic state [10].” Although glycolysis is less energy-efficient than OXPHOS, it provides rapid ATP production and generates metabolic intermediates for nucleotide, lipid, and amino acid biosynthesis, thereby supporting rapid cell division and effector differentiation [10]. The mTOR signaling pathway plays a pivotal role in this metabolic reprogramming by upregulating glycolytic enzymes and transporters, such as Glut1, to facilitate T cell activation and biosynthesis, establishing the energetic and material foundation for subsequent immune responses [10].

Effector T cells

Effector T cells arise under persistent antigen stimulation and inflammatory conditions, representing a metabolically highly active population responsible for target cell killing and cytokine secretion [11]. Given their role in the immediate clearance of pathogens or tumor cells, effector T cells have markedly increased energy and biosynthetic demands. Their metabolism is dominated by enhanced aerobic glycolysis, whereby glucose is rapidly converted to lactate while providing intermediates for lipid and amino acid synthesis, thereby supporting cell proliferation and effector molecule production [11].The mTOR (mammalian target of rapamycin) signaling pathway is central to the metabolic reprogramming of effector T cells [11]. Activation of mTORC1 promotes expression of HIF-1α (hypoxia-inducible factor-1α), upregulates glycolysis-related genes (e.g., HK2, LDHA, PFKFB3), and enhances glucose uptake. HIF-1α can be activated not only by hypoxia but also by immune activation signals, forming a positive feedback loop that further reinforces glycolytic metabolism [11]. This metabolic state supports rapid production of cytotoxic molecules, such as Granzyme B, IFN-γ, and TNF-α, maintaining high killing capacity. However, this “high metabolism, high output” state is typically unsustainable over the long term; prolonged antigen stimulation can induce energy exhaustion and metabolic stress, eventually leading effector T cells toward an exhausted state [11].

Memory T cells

During the late phase of immune responses, a subset of effector T cells undergoes metabolic and transcriptional remodeling to differentiate into memory T cells [12]. Memory T cells are characterized by long-term survival and rapid responsiveness upon re-exposure to antigens [12]. Compared with effector T cells, memory T cells adopt a “metabolically economical and stable” state, relying primarily on mitochondrial OXPHOS and FAO to sustain long-term homeostasis [12]. The AMPK (AMP-activated protein kinase) signaling pathway is critical for memory T cell formation and maintenance. AMPK senses cellular energy status and activates PGC1α (peroxisome proliferator-activated receptor gamma coactivator 1-alpha), promoting mitochondrial biogenesis and function, thereby enhancing metabolic efficiency and antioxidant capacity [12]. Studies have demonstrated that memory T cells possess greater mitochondrial reserve, higher oxidative capacity, and robust metabolic flexibility, allowing them to survive and function under nutrient-limited or inflammatory conditions [12]. Active FAO enables memory T cells to efficiently utilize endogenous fatty acids as an energy source during prolonged quiescence, providing the metabolic foundation for durable immune memory [12].

Exhausted T cells (Tex)

Exhausted T cells (T cell exhaustion, Tex) represent a functionally impaired subset of T cells that arises under conditions of chronic antigen stimulation, such as persistent viral infections or tumors [13]. Characteristic features of Tex include sustained high expression of inhibitory receptors (e.g., PD-1, TIM-3, LAG-3, TIGIT), reduced cytokine secretion (e.g., decreased IFN-γ and IL-2 production), limited proliferative capacity, and distinct transcriptional and epigenetic programs [13]. The transcription factor TOX (thymocyte selection-associated high mobility group box) plays a central role in Tex formation by remodeling chromatin to maintain the stable expression of exhaustion-associated genes [14].

Metabolic dysregulation is a key driver of Tex formation and maintenance. Compared with effector T cells, Tex exhibit impaired mitochondrial function, decreased mitochondrial membrane potential, accumulation of reactive oxygen species (ROS), and reduced glycolytic activity, resulting in insufficient energy production [14]. Concurrently, PD-1 signaling further suppresses glycolysis and lipid synthesis by inhibiting the PI3K-Akt-mTOR pathway, exacerbating the metabolic inhibitory state. Tex display limited metabolic flexibility, making them poorly adaptable to glucose, oxygen, or amino acid deprivation in the tumor microenvironment, thereby restricting their functional recovery [15]. In contrast, memory T cells can switch to FAO or OXPHOS to sustain energy supply, demonstrating superior metabolic adaptability [15]. Overall, T cell subsets exist in a dynamic functional equilibrium, which depends on precise regulation of intracellular metabolic pathways. In Fig. 1, we depicts the differentiation trajectory of CD8⁺ T cells in relation to metabolic pathways. Naive T cells rely on OXPHOS for homeostasis, while effector T cells engage glycolysis for rapid energy and biosynthesis upon antigen stimulation. After pathogen clearance, memory T cells preferentially utilize fatty acid oxidation and mitochondrial metabolism to sustain long-term survival. However, under chronic antigen exposure and immunosuppressive conditions, T cells fail to maintain proper metabolic activity, leading to the exhausted phenotype. Additionally, a subset of CD8⁺ regulatory T cells (CD8⁺ Tregs) can emerge from effector or memory T cells under metabolic and immunosuppressive stress, contributing to immune regulation in the tumor or chronic infection context.

CD8⁺ regulatory T cells

CD8⁺ regulatory T cells (CD8⁺ Tregs) are a subgroup of regulatory T cells that have gained increasing attention in recent years for their important role in peripheral immune tolerance [16]. Peripheral tolerance is crucial for preventing self-reactive T cells, which escape thymic selection, from causing autoimmune reactions. As a core component of peripheral tolerance, regulatory T cells protect the body’s own tissues by suppressing excessive inflammatory responses and limiting abnormal activation of immune cells [17]. Compared with the well-studied CD4⁺ Tregs, the phenotype, origin, and mechanisms of action of CD8⁺ Tregs are still under investigation [17]. Current evidence suggests that CD8⁺ Tregs are not a uniform population but are highly heterogeneous. They display diverse phenotypes and arise through multiple pathways, including CD122⁺CD8⁺ T cells, CD28^low CD8⁺ T cells, CD103⁺CD8⁺ T cells, Ly49⁺CD8⁺ T cells, CD8⁺ TAb-supp cells, CD45RC^low CD8⁺ T cells, CD8αα⁺CD4⁻ T cells, as well as CD8⁺ Tregs induced in vitro by B cells or plasmacytoid dendritic cells (pDCs) [17]. Based on the expression of the classical regulatory transcription factor Foxp3, CD8⁺ Tregs can be divided into Foxp3⁺ CD8⁺ Tregs and CD8⁺ non-Foxp3 Tregs. Although different subsets may overlap in molecular features and functions, CD8⁺ Tregs generally contribute to immune regulation through mechanisms such as secreting immunosuppressive cytokines, inhibiting effector T cell activation, and modulating antigen-presenting cell functions [18]. Human CD8⁺ regulatory T cells (Tregs) are predominantly CD8⁺CD28⁻, although both CD8⁺CD28⁺ and CD8⁺CD28⁻ subsets can be induced in vitro [19]; functionally, CD8⁺CD28⁻ Tregs can be classified into three types, with type II and type III cells exerting immunosuppressive effects primarily through cytokine secretion, including IFN-γ, IL-6, and IL-10, whereas type I cells mediate suppression via direct cell–cell contact with dendritic cells by modulating CD80/CD86 expression [20]. They may play important roles in autoimmune diseases, chronic inflammation, and tumor immune tolerance.

T cell exhaustion

T cell exhaustion is a state of functional decline commonly seen in chronic infections or tumor environments [21]. It is characterized by reduced effector functions, impaired proliferation, and sustained high expression of inhibitory receptors such as PD-1 and CTLA-4 [21]. Studies show that continuous antigen stimulation is a major driver of exhaustion. When T cells are exposed to antigens for a long time without clearing the pathogen or tumor, intracellular stress accumulates and their function gradually declines [22]. Metabolic restrictions also play a key role in exhaustion. Exhausted T cells show altered glucose, fatty acid, and amino acid metabolism. Energy supply becomes insufficient, mitochondrial function is impaired, and reactive oxygen species (ROS) levels increase, which further weakens their effector functions [22]. In other words, persistent stimulation combined with limited nutrients and energy in the microenvironment together drive T cells into an exhausted state, reducing their ability to clear pathogens or tumors [22].

CD8⁺ T cell exhaustion is closely influenced by complex intercellular metabolic interactions within the tumor microenvironment, which play a critical role in its formation and maintenance [23]. First, CD8⁺ T cells engage in both metabolic cooperation and competition with CD4⁺ helper T cells [24]. CD4⁺ T cells can support CD8⁺ T cell proliferation and effector function by providing cytokines or metabolic substrates, but under nutrient-limited or high-glycolysis conditions, they may also compete for glucose, amino acids, or lipids, affecting the metabolic adaptability of CD8⁺ T cells [24]. Second, CD8⁺ T cells interact with regulatory T cells (Tregs), tumor-associated macrophages (TAMs), and myeloid-derived suppressor cells (MDSCs) in a complex network of competition and cooperation. Tregs, through high glycolytic or fatty acid metabolism, consume local metabolic resources and secrete inhibitory factors such as IL-10 and TGF-β, promoting CD8⁺ T cell exhaustion [25]; TAMs and MDSCs restrict CD8⁺ T cell activation via lactate, adenosine, or amino acid depletion [25]. Third, cancer-associated fibroblasts (CAFs), the vascular system, and extracellular matrix components also influence CD8⁺ T cell metabolism. Beyond hypoxia or acidic stress, they can secrete glutamine, lactate, lipids, and growth factors that shape local energy availability, altering CD8⁺ T cell metabolic fitness and exhaustion [26]. Taken together, CD8⁺ T cell exhaustion is not solely a cell-intrinsic metabolic phenomenon but is also significantly shaped by intercellular metabolic interactions within the TME. These findings underscore that future immunometabolic interventions should consider the broader network of cell–cell metabolic crosstalk to optimize antitumor immunity.

Metabolic reprogramming and CD8⁺ T cell exhaustion

These figures illustrate the interplay between metabolic regulation and CD8⁺ T cell fate during chronic infection or tumor exposure. In Fig. 2, we summarizes the main mechanisms by which metabolic restrictions drive CD8⁺ T cell exhaustion. In the tumor microenvironment (TME) or during chronic viral infection, limited glucose availability and high lactate levels impair glycolysis, reducing effector cytokine production (IFN-γ, TNF-α) and enhancing PD-1 expression, which suppresses mTOR signaling. Concurrently, mitochondrial dysfunction and oxidative phosphorylation (OXPHOS) defects lead to increased reactive oxygen species (ROS), DNA damage, and apoptosis. Disruption of fatty acid oxidation (FAO) further compromises energy supply and T cell survival. Metabolic stress also triggers AMPK/HIF1α signaling, upregulating inhibitory receptors such as PD-1 and TIM-3, which create a negative feedback loop that further suppresses T cell receptor (TCR) signaling and metabolic activity, culminating in functional exhaustion. To improve clarity and provide a structured overview of the complex metabolic mechanisms discussed above, we summarize the major metabolic pathways that regulate CD8⁺ T cell exhaustion in Table 1. This table integrates glucose metabolism, lipid metabolism, amino acid metabolism, mitochondrial stress, hypoxia, and metabolic byproducts into a unified framework, highlighting their key molecular regulators, functional consequences, and overall impact on T cell exhaustion. By organizing these pathways side-by-side, we aim to reduce conceptual redundancy, clarify context-dependent effects, and distinguish well-established mechanisms from emerging evidence. This overview facilitates comparison across metabolic modules and provides a concise reference for understanding how intrinsic metabolic reprogramming and extrinsic tumor microenvironment constraints converge to drive CD8⁺ T cell dysfunction.

Fig. 2.

Metabolic restrictions drive CD8⁺ T cell exhaustion. Chronic antigen exposure, immunosuppressive cells, and soluble inhibitory molecules promote functional decline. Glucose limitation, mitochondrial stress, fatty acid metabolism disruption, and upregulation of inhibitory receptors collectively impair CD8⁺ T cell effector functions, leading to exhaustion

Table 1.

Metabolic Pathways Regulating CD8⁺ T Cell Exhaustion in the Tumor Microenvironment

Metabolic pathway	Key mechanisms	Effect on CD8⁺ T cells	Exhaustion	References
Glucose Metabolism	Glucose deprivation, high lactate, PD-1/CTLA-4 signaling suppress HK2, PFK1, LDHA; Treg-mediated MondoA-TXNIP axis	↓ Glycolysis, ↓ IFN-γ/TNF-α, impaired proliferation	Yes	[27– [31, 33]
Lipid Metabolism	Lipid droplet accumulation, cholesterol & fatty acids; FAO modulation via PPAR/CPT1A; FASN–STAT3 axis; LPA-LPAR5 signaling; ROS & ferroptosis	Lipid peroxidation, ↑ ROS, reduced survival and effector function	Yes	[34–45]
Mitochondrial Metabolism & Oxidative Stress	↓ Membrane potential, impaired OXPHOS, ↑ ROS; DRP1/OPA1/PGC1α dysregulation; ATF4 activation; PGE2, impaired FA metabolism	Reduced cytotoxicity, proliferation; impaired metabolic flexibility	Yes	[46–52]
Amino Acid Metabolism	Glutamine, arginine, tryptophan depletion; IDO/TDO catabolism; Kyn–AhR–Siglec15 axis; serine & methionine metabolism; impaired transporter expression (SLC7A5/SLC38A1)	Impaired mTOR/GCN2 signaling, reduced proliferation, effector function	Yes	[53–58]
Glutamine Metabolism	Substrate for glycolysis & OXPHOS; regulates mTOR; 5-HT mediated GAPDH serotonylation; competition with tumor/DCs	Supports glycolysis, cytokine production; inhibition can prevent early exhaustion	Partial / context-dependent	[60–64] [65]
Hypoxia / TME	HIF-1α stabilization, lactate accumulation, ROS; CD39/CD73-mediated adenosine; stromal cell interactions	Dysregulated glucose, lipid, amino acid metabolism; ↑ PD-1/PD-L1; impaired T cell survival	Yes	[66–72]
Accumulation of Metabolic Byproducts	Lactate, adenosine, high extracellular K⁺, NAD⁺/NADH imbalance; metabolic enzyme dysregulation (CD39/CD73, LDH)	Inhibits TCR signaling, reduces cytokine secretion, lowers mitochondrial ATP	Yes	[73–78]

Glucose metabolism

Glucose metabolism plays a central role in CD8⁺ T cell activation, proliferation, and effector function [27]. Under normal immune responses, activated T cells increase glycolytic flux to meet high energetic and biosynthetic demands. However, within the tumor microenvironment (TME), multiple extrinsic and intrinsic factors converge to suppress glycolysis. Tumor cells with elevated glycolysis sequester available glucose, while lactate accumulation creates an acidic environment that directly inhibits T cell function by modulating pH-dependent signaling and suppressing mTOR activity [27–29]. Immune checkpoint signaling via PD-1 and CTLA-4 further restricts glycolytic capacity by downregulating key enzymes such as HK2, PFK1, and LDHA [29]. While many studies demonstrate correlations between glucose deprivation and T cell exhaustion, the causal strength varies. For instance, lactate-mediated suppression of T cell effector function is well-established in vitro, yet its quantitative impact in human tumors remains less clear [29]. Similarly, tumor-derived exosomes (e.g., IL-8–PPARα–UCP1 axis) and Treg metabolic remodeling via the MondoA-TXNIP pathway suggest intercellular metabolic interactions contribute to exhaustion [30, 31]. However, the relative importance of these extrinsic factors compared with cell-intrinsic glycolytic defects is not fully resolved. Genetic or pharmacologic modulation of T cell metabolism provides more direct mechanistic insights. PKM2 deletion in CD8⁺ T cells enhances pentose phosphate pathway activity, promotes TCF1⁺ progenitor-like T cells, and improves responsiveness to PD-1 blockade [32]. PSGL-1 deficiency maintains glycolytic activity and expands stem-like T cell pools, enhancing effector function [27–29]. Similarly, CD73 deletion increases glucose uptake and mitochondrial respiration, improving antitumor activity in vivo [33]. MondoA deficiency promotes heightened glycolysis in Tregs, resulting in Th17-like high-glycolytic Tregs that induce CD8⁺ T cell exhaustion via IL-17 A, thereby facilitating tumor progression. In MondoA-deficient mice, IL-17 A blockade mitigates tumor growth and reduces CD8⁺ T cell exhaustion, highlighting the pivotal role of the MondoA-TXNIP axis in Treg metabolic regulation and antitumor immunity [30]. Nevertheless, most evidence derives from murine models or ex vivo experiments, and translation to human TILs requires caution. Overall, these findings indicate that glucose metabolism is a critical driver of CD8⁺ T cell exhaustion, operating at the intersection of intrinsic metabolic regulation and extrinsic TME-imposed constraints. Future work integrating single-cell metabolomics and careful distinction between murine and human systems, as well as in vitro versus in vivo contexts, is needed to identify metabolic interventions that are likely to sustain T cell antitumor function clinically.

Lipid metabolism

Aberrant lipid metabolism within the TME has emerged as a central driver of CD8⁺ T cell functional suppression [34]. Tumor cells often upregulate lipid synthesis and uptake, leading to lipid droplet accumulation and elevated cholesterol and long-chain fatty acid levels in the stroma [34]. While these lipids support tumor growth by supplying energy and membrane components, they concurrently exert inhibitory effects on T cells through both direct cytotoxic mechanisms and indirect metabolic reprogramming [35]. Excess intracellular lipids induce peroxidation and ferroptosis, ultimately compromising CD8⁺ T cell viability and effector function [35]. Key lipid metabolism pathways modulate T cell activity. The PPAR axis regulates fatty acid oxidation (FAO), CPT1A mediates mitochondrial fatty acid transport, and SREBPs control lipid synthesis [36]. Enhanced FAO generally promotes T cell survival and memory formation, whereas FAO inhibition accelerates exhaustion, highlighting the therapeutic potential of modulating lipid flux to restore T cell vitality in the TME [36]. Recent mechanistic studies provide further insight. In hepatocellular carcinoma (HCC), Riplet deficiency—often caused by promoter hypermethylation—blocks K48-linked ubiquitination of fatty acid synthase (FASN), increasing fatty acid synthesis. Excess palmitic acid enhances STAT3 palmitoylation in T cells, driving terminal exhaustion and weakening antitumor immunity [37]. Notably, FASN inhibition reverses this effect, underscoring the Riplet–FASN–STAT3 axis as a potential immunotherapeutic target [37]. Exogenous lipids can also modulate CD8⁺ T cell function. Linoleic acid (LA) enhances mitochondrial function, promotes mitochondria–ER contact formation, and induces memory-like CD8⁺ T cells with superior effector capabilities [38, 39]. Conversely, chronic type I interferon (IFN-I) signaling disrupts lipid homeostasis, increasing ROS and lipid peroxidation in Tex cells and impairing responses to immune checkpoint blockade (ICB) [40]. Bioactive lipids, such as lysophosphatidic acid (LPA), further drive exhaustion via LPAR5-mediated metabolic reprogramming [41]. Nutrient competition and metabolic checkpoints also exacerbate exhaustion. Tumor overexpression of Slc7a11 limits cystine availability, triggering ferroptosis and increasing CD36-mediated lipid uptake in T cells [42]. Restoring glutathione synthesis via Gclc improves T cell metabolic fitness and antitumor activity [42]. In HCC, S100A10 drives lipid reprogramming through the cPLA2/5-LOX axis, elevating LTB4 levels and promoting CD8⁺ T cell exhaustion, while its inhibition restores effector function and suppresses tumor progression [43]. Iron-mediated lipid peroxidation and cholesterol metabolism are additional modulators. CD36 enhances iron uptake through TfR1, and NRF2 activation can mitigate lipid peroxidation, preserving T cell function [44]. Aurora kinase B (AURKB) increases intratumoral cholesterol via epigenetic regulation of NCEH1, inducing CD8⁺ T cell exhaustion; both AURKB inhibition and simvastatin restore T cell effector capacity in cholangiocarcinoma models [45]. Collectively, these findings illustrate that dysregulated lipid metabolism, through peroxidation, signaling perturbation, and energy reprogramming—exacerbates CD8⁺ T cell exhaustion and compromises antitumor immunity.

Mitochondrial metabolism and oxidative stress

Mitochondrial dysfunction is a hallmark of CD8⁺ T cell exhaustion. Chronic antigen stimulation and TME stress reduce mitochondrial membrane potential, impair respiratory chain efficiency, and increase ROS accumulation, collectively impairing proliferation and cytotoxicity [46]. DRP1-mediated fission, OPA1-mediated fusion, and PGC1α-dependent biogenesis are disrupted in exhausted T cells, limiting metabolic flexibility and ROS clearance [47]. Restoration of mitochondrial fitness—via PGC1α agonists or targeted interventions—enhances both glycolysis and FAO, improving effector function and survival [48]. PTPMT1 deficiency illustrates the metabolic vulnerability of exhausted T cells. Loss of this mitochondrial phosphatase shifts substrate utilization toward FAO, reduces flexibility, and induces oxidative stress, DNA damage, and apoptosis, ultimately accelerating exhaustion [49]. The transcription factor IRF-5 integrates lipid and mitochondrial metabolism, limiting exhaustion; IRF-5-deficient T cells show increased inhibitory receptor expression, elevated ROS, and reduced OXPHOS [50]. Extrinsic factors in the TME exacerbate mitochondrial stress. PGE2 accumulation promotes ROS and anti-apoptotic signaling, while impaired hepatic FA metabolism reduces mitochondrial potential and drives exhaustion in HCC[40. Persistent ATF4 activation induces metabolic polarization and terminal exhaustion in TILs, whereas ATF4 inhibition preserves mitochondrial homeostasis and improves response to PD-1 therapy [51]. Ferroptosis, mediated by ROS and lipid peroxidation, limits T_EX survival, but GPX4 restoration rescues T cell numbers and function, offering a therapeutic avenue [52]. Overall, mitochondrial integrity is central to maintaining CD8⁺ T cell effector function in metabolically hostile TMEs.

Amino acid metabolism

Amino acid scarcity in the TME significantly impairs CD8⁺ T cell function. Glutamine, arginine, and tryptophan are critical for proliferation, cytotoxicity, and signaling via mTOR and GCN2 pathways [53, 54]. Tumors deplete these substrates directly or through catabolic enzymes (e.g., IDO, TDO), generating immunosuppressive metabolites [54]. Mechanistic insights highlight the Kyn–AhR–Siglec-15 axis in HNSCC, where kynurenine accumulation suppresses CD8⁺ T cell infiltration and effector function, and its blockade restores antitumor activity [55]. Methionine metabolism in HCC similarly drives exhaustion: elevated SAM and MTA impair function, while MAT2A knockout rescues T cells and inhibits tumor growth [56]. Epigenetic modifications link amino acid metabolism to exhaustion. For example, H3K9 lactylation activates IL-11 and immune checkpoint genes, promoting dysfunction, whereas IL-11 inhibition restores cytotoxicity [57]. Tumor-mediated suppression of amino acid transporters (SLC7A5, SLC38A1) and exosome-induced YAP1 ubiquitination further impairs T cell metabolic signaling and glycolysis [58]. Serine metabolism competition in cervical cancer similarly contributes to exhaustion [59]. Collectively, amino acid deprivation impairs energy production, signaling, and transcriptional regulation, reinforcing CD8⁺ T cell dysfunction.

Glutamine metabolism

Glutamine metabolism plays a key role when CD8⁺ T cells shift from oxidative phosphorylation to aerobic glycolysis [60]. Glutamine not only serves as an important energy source for mitochondria, supporting CD8⁺ T cell proliferation and effector functions under high energy demand, but also regulates cell activation and immune responses through the mTOR signaling pathway [61]. Recent studies show that CD8⁺ T cells can accumulate glutamine or related metabolites through both internal synthesis and external uptake [62]. This supports glycolytic reprogramming and cytokine production. For example, serotonylation of GAPDH by 5-hydroxytryptamine (5-HT) promotes glycolysis in CD8⁺ T cells and enhances their antitumor activity, a process that depends on glutamine as an energy and metabolic substrate [62]. In the tumor microenvironment, CD8⁺ T cells compete with dendritic cells and tumor cells for glutamine uptake, which affects their antitumor function [63]. Clinically, glutamine abundance correlates with IFN-γ production in CD8⁺ T cells, and high expression of the glutamine transporter SLC25A22 is associated with an immunosuppressive microenvironment and resistance to immunotherapy [64]. Overall, glutamine metabolism not only provides energy for CD8⁺ T cells but also enhances their antitumor effects through mTOR signaling and metabolic regulation.

In the tumor microenvironment, activated CD8⁺ T cells often exhibit functional limitations due to metabolic unfitness. This metabolic stress is closely associated with T cell exhaustion and impaired memory formation, representing a major barrier to effective antitumor adoptive immunotherapy [65]. Recently, one study showed that in a mouse tumor model, tumor-specific CD8⁺ T cells cultured under glutamine-restricted (dGln) conditions, or CD8⁺ T cells treated with specific inhibitors of glutamine metabolism, more efficiently cleared tumors and significantly improved the survival of tumor-bearing mice [65]. Compared with control cells, dGln-cultured CD8⁺ T cells exhibited lower PD-1 expression and higher Ki67 positivity in tumor-infiltrating T cells, indicating that inhibition of glutamine metabolism can prevent premature T cell exhaustion [65]. Moreover, these T cells expanded more efficiently upon secondary antigen stimulation, and the expression of pro-survival and memory-associated transcription factors was significantly higher than in control cells. Consistently, previous studies have reported that restricting glutamine supply during the initial activation of mouse CD8⁺ T cells enhances glycolysis and spare respiratory capacity (SRC), prevents early exhaustion, and improves T cell persistence without compromising antitumor effector function [65]. Together, these findings suggest that modulating glutamine metabolism can not only maintain the antitumor effector function of CD8⁺ T cells but also delay or prevent exhaustion, providing a potential strategy to optimize adoptive immunotherapy.

Hypoxic tumor microenvironment

Hypoxia is prevalent in solid tumors, stabilizing HIF-1α and shifting metabolism toward glycolysis while suppressing oxidative phosphorylation [66]. Hypoxia promotes immune checkpoint expression (PD-1/PD-L1) and exacerbates CD8⁺ T cell exhaustion by dysregulating glucose, lipid, and amino acid metabolism [51]. Lactate accumulation and ROS further compromise function and survival. Therapeutic strategies include alleviating hypoxia or targeting HIF-1α to restore metabolic and effector function [67]. Mechanistically, Blimp-1 suppresses PGC1α-dependent mitochondrial reprogramming under hypoxia, reducing mitochondrial function and elevating ROS, while P4HA1 upregulation perturbs TCA cycle metabolites, limiting progenitor T cell expansion [68]. Terminally exhausted T cells acquire regulatory-like profiles, producing immunosuppressive adenosine via CD39; hypoxia drives CD39 expression, and its inhibition enhances T cell infiltration and immunotherapy response [69]. Tumor-associated stromal populations, such as PLXDC1⁺ PSCs in PDAC, exacerbate hypoxia-induced exhaustion, representing potential targets [70]. Hypoxia-inducible factors like RCOR2 and HIF-1α in macrophages promote immunosuppressive axes and metabolic reprogramming, further reinforcing CD8⁺ T cell exhaustion and resistance to PD-1 therapy [71, 72]. Loss of VHL in T cells promotes tissue-resident memory-like accumulation, highlighting hypoxia-mediated metabolic rewiring as a key determinant of exhaustion [72].

Accumulation of metabolic byproducts

Metabolic byproducts in the TME—including lactate, adenosine, and high extracellular K⁺—potently suppress CD8⁺ T cell function [73]. Lactate and acidity reduce cytokine secretion, adenosine inhibits TCR signaling via A2A receptors, and elevated K⁺ diminishes membrane potential and metabolism [74]. NAD⁺/NADH imbalance further restricts mitochondrial ATP production and antioxidant capacity [75]. Therapeutic strategies target these metabolites: inhibiting CD39/CD73, LDH, or other metabolic enzymes alleviates suppression. MFSD2A overexpression in gastric cancer reduces TGFβ1 release, reprograms the TME, enhances CD8⁺ T cell activation, and mitigates exhaustion [76]. In NPC, FLI1 drives IDO1–Kyn production, promoting exhaustion and Treg differentiation; its inhibition restores antitumor immunity [77]. TAM metabolic reprogramming via OXCT1 enhances succinate and Arg1 expression, impairing T cells, whereas its inhibition promotes cytotoxicity [62]. Elevated circulating methylmalonic acid (MMA) induces CD8⁺ T cell exhaustion by impairing mitochondrial NADH regeneration and upregulating TOX [78]. Together, these findings underscore that metabolic waste accumulation reinforces immunosuppression and limits antitumor immunity.

Immunotherapeutic strategies targeting metabolic reprogramming

With advances in tumor immunometabolism, targeting T cell or tumor metabolic pathways has emerged as a promising approach for antitumor immunotherapy [79]. These strategies primarily include small-molecule metabolic modulators to enhance T cell metabolic adaptability, combination with immune checkpoint inhibitors (ICIs) to boost antitumor efficacy, direct regulation of T cell metabolic flexibility, and targeted interventions addressing tumor-intrinsic metabolic abnormalities [79]. Small-molecule metabolic activators or inhibitors have been widely employed to restore exhausted T cell function. For example, metformin activates the AMPK signaling pathway, improves mitochondrial function, and enhances OXPHOS, thereby increasing T cell metabolic flexibility and survival [80]. Etomoxir inhibits CPT1 to regulate fatty acid oxidation (FAO), optimizing T cell energy metabolism [80]. PGC1α agonists promote mitochondrial biogenesis, reduce ROS accumulation, and improve metabolic status, enhancing effector function in exhausted T cells [81]. Such metabolic modulators provide a supportive framework to partially reverse T cell exhaustion and augment immunotherapy. Combination therapy with metabolic interventions and ICIs has demonstrated synergistic potential. Co-administration of PD-1 or CTLA-4 inhibitors with metabolic activators can simultaneously restore T cell metabolic fitness and relieve immune inhibitory signaling, significantly enhancing T cell cytotoxicity and antitumor capacity [82]. This dual strategy improves T cell function and survival within the TME, thereby increasing clinical immunotherapy efficacy. Direct modulation of T cell metabolic flexibility represents another key approach. Enhancing FAO or mitochondrial function enables T cells to adapt to nutrient scarcity and metabolic competition in the TME [83]. Activation of the AMPK–PGC1α axis or expansion of mitochondrial reserve allows T cells to switch energy pathways dynamically, maintaining long-term antitumor activity. Targeting tumor-intrinsic metabolic abnormalities is also feasible. Approaches such as inhibiting tumor glycolysis, reducing lactate production, or limiting lipid accumulation alleviate metabolic suppression within the TME. These strategies mitigate the inhibitory effects of lactate and acidic metabolites on T cells, improve energy acquisition, and indirectly reverse exhaustion. Metabolic optimization of CAR-T cells offers a clinically actionable engineering strategy. Through genetic or metabolic engineering, enhancing mitochondrial function, boosting FAO capacity, or limiting ROS accumulation can improve CAR-T cell resilience and persistence in nutrient-depleted, high-stress TMEs [84]. “Metabolically optimized” CAR-T cells hold promise for overcoming energy limitations in solid tumors, enhancing efficacy and durability. Overall, targeting metabolic reprogramming provides both a theoretical and practical basis for restoring exhausted T cell function, optimizing combination immunotherapy, and guiding CAR-T engineering [85]. Integration of single-cell metabolomics and TME metabolic monitoring is anticipated to facilitate individualized metabolic–immunotherapy strategies, improving clinical antitumor immunity.

Although many reports showed that multiple strategies to reverse T cell exhaustion or enhance antitumor immunity through metabolic modulation, including small-molecule metabolic activators (such as metformin, PGC1α agonists, and Etomoxir), combination therapy with immune checkpoint inhibitors, direct modulation of T cell metabolic flexibility, and targeting tumor-intrinsic metabolic abnormalities—these approaches have primarily shown efficacy in preclinical models and face significant challenges in clinical translation [86, 87]. First, several clinical trials of metabolic interventions, such as IDO inhibitors and certain metabolic enzyme inhibitors, have demonstrated limited efficacy; single-agent or combination treatments with immunotherapy did not significantly improve overall survival or response rates [88]. This may be closely related to the complexity of the tumor microenvironment, multi-level metabolic competition, and heterogeneity between T cells and tumor cells. Second, metabolic interventions may carry potential toxicity, as long-term or broad-spectrum inhibition of metabolic pathways could impair energy supply in normal tissues, leading to systemic side effects, which have not been fully assessed in early studies. Third, targeting a single pathway or broadly modulating metabolism may be insufficient to counteract the multiple mechanisms driving T cell exhaustion [88]. Therefore, multi-layered combination strategies, integrating checkpoint blockade, T cell metabolic optimization, and dynamic monitoring of microenvironmental metabolism, may be required [89]. Overall, while metabolic modulation provides both theoretical and experimental support for restoring T cell function and optimizing immunotherapy, future studies should pay greater attention to clinical feasibility, reasons for intervention failure, and safety evaluation, and leverage single-cell metabolomics and tumor microenvironment metabolic monitoring to develop individualized and precise metabolism–immunity intervention strategies, thereby improving clinical antitumor efficacy.

In this review, the distinction between progenitor and terminally exhausted CD8⁺ T cells provides an important framework for understanding metabolic regulation. We propose that metabolic interventions may have markedly different effects depending on the stage of disease progression. Progenitor exhausted T cells possess higher proliferative capacity and metabolic plasticity, making them more amenable to metabolic optimization—such as modulation of glycolysis, oxidative phosphorylation, or glutamine utilization—to restore antitumor function [90]. In contrast, terminally exhausted T cells exhibit metabolic rigidity and limited proliferative potential, and whether their function can be fully rescued through metabolic remodeling remains to be determined [91]. Importantly, CD8⁺ T cell metabolic states are dynamically regulated during immunotherapy: long-term antigen stimulation, metabolic competition within the tumor microenvironment, and therapeutic interventions collectively alter glycolysis, fatty acid oxidation, mitochondrial reserve, and key signaling pathways such as mTOR and AMPK. These observations highlight that effective metabolic interventions should integrate exhaustion stage, treatment timing, and the dynamic evolution of T cell metabolism to enhance both efficacy and durability of antitumor responses.

Future perspectives

Despite progress in immunometabolism, multiple avenues remain to be explored in T cell exhaustion and tumor metabolic regulation. First, single-cell multi-omics provides powerful tools to dissect metabolic heterogeneity within T cells and the TME. Integrating single-cell transcriptomics, metabolomics, and epigenomics can identify metabolic features and adaptive responses of distinct T cell subsets during exhaustion, revealing intracellular and extracellular metabolic networks and guiding precise intervention strategies. Second, metabolic competition in the TME is a key driver of T cell exhaustion, but its dynamic regulation remains incompletely understood. Future studies could develop spatiotemporal models of nutrient competition, combined with mathematical modeling and experimental validation, to quantify T cell–tumor cell competition for glucose, amino acids, and lipids. Such approaches will help predict the effects of metabolic interventions on T cell functional recovery, providing theoretical support for metabolic–immune combination therapies. Third, spatiotemporal metabolic imaging and immune monitoring will become critical. Techniques including metabolic probes, optical imaging, and magnetic resonance imaging enable in vivo tracking of T cell and tumor metabolic states, providing visualized readouts to evaluate therapeutic efficacy and optimize dosing strategies. These advances will allow metabolic interventions to be precisely integrated with immunotherapy, enabling real-time monitoring and individualized adjustment. Fourth, personalized metabolic–immunotherapy holds great promise. Given interpatient heterogeneity in TME metabolism, combining single-cell metabolomics with patient-specific immune profiling may enable tailored metabolic interventions. For instance, selecting modulators based on tumor glycolytic activity, lactate levels, or lipid metabolism could be combined with ICIs or cellular therapies to achieve precision antitumor immunity. Finally, the interplay between metabolism and epigenetics in T cell fate decisions represents a key research frontier. Metabolites serve as substrates for epigenetic modifications, such as acetyl-CoA for histone acetylation and S-adenosylmethionine (SAM) for methylation, regulating expression of genes associated with T cell exhaustion. Elucidating these metabolic–epigenetic interactions may uncover precise molecular targets for reversing exhaustion and inform novel therapeutic strategies. In summary, future research integrating single-cell multi-omics, metabolic imaging, mathematical modeling, and epigenetics will comprehensively elucidate the complex network linking TME metabolism and T cell exhaustion, paving the way for individualized and precision metabolic–immunotherapy. These approaches will deepen understanding of exhaustion mechanisms and inform optimization of clinical immunotherapy.

In the future, we emphasize the need to further explore the direct role of metabolic intermediates in regulating T cell function. Metabolites such as acetyl-CoA, S-adenosylmethionine (SAM), α-ketoglutarate (α-KG), and lactate not only serve as energy and biosynthetic substrates but also influence chromatin accessibility through histone acetylation, methylation, and other epigenetic modifications, thereby directly modulating gene expression and T cell activity. This highlights a tight coupling between metabolic state and epigenetic plasticity. Additionally, the epigenetic fixation mediated by the key transcription factor TOX during T cell exhaustion limits the potential for metabolic and functional remodeling. Emerging studies are investigating whether metabolic reprogramming or specific metabolic interventions can overcome TOX-driven epigenetic constraints to restore the proliferative capacity and effector function of exhausted T cells. We believe this line of investigation represents a critical frontier in immunometabolism and warrants comprehensive integration and in-depth discussion in the review, rather than being limited to a brief forward-looking paragraph.

Conclusion

T cell functional states are tightly regulated by metabolic processes, with metabolic reprogramming spanning the entire T cell lifecycle from naïve, activated, and effector to exhausted states. Distinct T cell subsets exhibit unique metabolic profiles: naïve T cells rely on oxidative phosphorylation to maintain quiescence; effector T cells utilize glycolysis to meet high energy demands; memory T cells employ FAO and mitochondrial homeostasis for long-term survival; exhausted T cells display energy deficiency and metabolic imbalance. The TME exerts multifaceted metabolic pressure on T cells, including nutrient competition, lactate and acidic metabolite accumulation, lipid and amino acid dysregulation, and hypoxia-induced signaling, collectively driving immune suppression and T cell exhaustion. Understanding the mechanisms of metabolic–immune crosstalk provides a foundation for targeted metabolic interventions to restore T cell function. Through metabolic activators, pathway-specific interventions, enhancement of T cell metabolic flexibility, or CAR-T metabolic optimization, T cell adaptability and effector function within the TME can be improved, enhancing antitumor immunity. Integration with single-cell omics, metabolic imaging, and personalized therapeutic strategies offers the potential for precision metabolic–immunotherapy, opening new avenues for clinical immune interventions.

Author contributions

Wenzhi Deng, Xiulin Jiang, Yongzhi Xie wrote the main manuscript text and prepared the figures and analyzed the literature in depth and optimized the topic and structure. Chunhong Li and Qiang Wang developed the conception of the topic, Xiulin Jiang and Yongzhi Xie supervised, reviewed, and revised the written manuscript. All authors reviewed the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (82501674).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This declaration is not applicable as the study does not involve human or animal subjects.

Competing interests

The authors declare no competing interests.