Metabolism, autophagy, and cell death: The triangular axis in tumor survival and therapeutic resistance Authors: Wen-Dong Wan, Muhammad Tufail, Can-Hua Jiang, Ning Li Published: 2026 ABSTRACT AbstractTumor cells face constant environmental and therapeutic stresses, relying on interconnected survival mechanisms involving autophagy, metabolism, and cell death regulation. This review explores the triangular relationship between these processes and how their dynamic interplay supports tumor growth and drives resistance to various treatments. We discuss key molecular pathways regulating autophagy and metabolic reprogramming, and their impact on different cell death modalities. Finally, we address current challenges in translating this knowledge into clinical therapies, emphasizing the need for precise biomarkers and combination strategies to overcome resistance. Understanding this complex network offers promising avenues for developing more effective cancer treatments.Keywords: Metabolic reprogramming, Autophagy, Cell death, Tumor survival and resistance, Therapeutic approaches Background Cancer cells constantly face harsh conditions such as nutrient shortage, low oxygen, and challenges from treatments. To survive and grow, they rely on a complex network of processes that link metabolism [1], autophagy [2], and cell death [3]. While these processes are discussed separately for clarity, they function together seamlessly to drive tumor progression and shape responses to therapy [4]. Cancer cells change the way they process nutrients to meet their high energy needs. These metabolic changes not only fuel cell growth but also influence autophagy and how easily cells die [5]. Autophagy is a highly conserved cellular process that helps cancer cells maintain balance when nutrients are scarce or when they are under stress. While autophagy can prevent cancer by removing damaged parts early on, in established tumors it often supports growth and helps the cells resist treatment [6,7]. Cell death in cancer is not limited to just one pathway. Besides the well-known apoptosis, there are other forms of programmed cell death like necroptosis, ferroptosis, pyroptosis, and autophagy-dependent death. Each of these is shaped by the cell's metabolic state and autophagic activity. This allows cancer cells to avoid death and adapt to the stresses they face [8]. In this review, we explore how autophagy, metabolism, and cell death interact in cancer cells. Understanding this triangular relationship sheds light on why tumors survive and become resistant to treatments. We also discuss new ways to target these processes together, aiming to improve therapies and patient outcomes (Fig. 1). The challenges and future perspectives of targeting the triangular axis of autophagy, metabolism, and cell death in cancer. At the center, the triangle highlights their interconnected nature through crosstalk between metabolism and autophagy [[9], [10], [11]], the dual roles of autophagy in cell survival and cell death [12,13], and feedback loops between cell death pathways and metabolic reprogramming [14,15]. Surrounding this axis are major challenges, including context dependency, metabolic plasticity, tumor heterogeneity, and the complexity of dynamic microenvironments, which hinder therapeutic progress. Extending outward are future perspectives, such as precision approaches, combinatorial therapies, temporal targeting, biomarker discovery, and the use of new technologies, which together provide promising directions to overcome current barriers and guide innovative therapeutic strategies. Metabolic reprogramming in tumors Hallmarks of cancer metabolism: glycolysis, OXPHOS, and beyond Metabolic reprogramming is one of the defining features of cancer, enabling tumor cells to thrive in conditions that would compromise the survival of normal cells. Instead of following the conventional patterns of energy production, cancer cells develop the ability to adjust and diversify their metabolic strategies. This flexibility allows them not only to sustain rapid growth but also to adapt to fluctuating nutrient supplies, oxygen availability, and therapeutic stress. Three central hallmarks characterize this altered metabolic landscape: reliance on glycolysis, dynamic use of oxidative phosphorylation (OXPHOS), and the exploitation of additional metabolic pathways that extend beyond these two core processes (Fig. 2). Hallmarks of cancer metabolism. This figure highlights the integration of glycolysis, OXPHOS, and auxiliary pathways. Glucose, glutamine, fatty acids, and lactate enter the cell through specific transporters such as GLUT1, ASCT2, and MCT1/4. Glycolysis, driven by key enzymes including HK2 and PKM2, provides pyruvate for mitochondrial oxidation and branches into the PPP and serine–glycine/one-carbon metabolism for nucleotide and redox support. Within mitochondria, the TCA cycle and electron transport chain (ETC I–V) generate ATP and metabolic intermediates. Lactate export contributes to an acidic TME, while outputs such as ATP, NADPH, biomass, and metabolic by-products support tumor growth, immune evasion, and therapy resistance. Upstream regulators including HIF-1α, MYC, mTOR, AMPK, and p53 coordinate pathway activity in response to microenvironmental cues. To understand how this metabolic flexibility is implemented at the molecular level, glycolysis represents the most intuitive starting point. A hallmark feature of cancer metabolism is the preference for glycolysis even in the presence of oxygen, a phenomenon known as the Warburg effect. Glycolysis serves as both an energetic and anabolic platform rather than a metabolically inferior alternative to mitochondrial respiration [16]. Glucose-derived metabolites enter pathways such as the pentose phosphate pathway (PPP). The coordinated upregulation of PPP flux and its key enzymes establishes a robust PPP–NADPH axis that buffers oxidative stress, enabling malignant cells to tolerate high glycolytic flux while maintaining redox homeostasis [17]. Despite the central role of glycolysis, cancer metabolism cannot be fully explained without considering the continued importance of mitochondrial function. OXPHOS provides a continuous supply of ATP in nutrient-limited conditions and generates intermediates for biosynthetic processes, such as citrate for lipid synthesis and aspartate for nucleotide production [18]. Beyond its bioenergetic role, mitochondrial metabolism also functions as a signaling hub. Mitochondrial respiration produces reactive oxygen species (ROS), which, at moderate levels, promote oncogenic signaling and adaptive stress responses. For example, an increase in mitochondrial superoxide anion release promotes colorectal cancer (CRC) progression by activating YAP via a post-transcriptional positive feedback loop between PGAM5 and MST3 [19]. However, excessive ROS production is hazardous, even to cancer cells. Importantly, while cancer stem-like cells exhibit elevated OXPHOS, they maintain paradoxically low ROS levels through mechanisms such as FoxM1-driven Prx3 expression and fatty acid oxidation (FAO)-mediated NADPH regeneration [20]. These observations illustrate how mitochondrial metabolism supports both tumor growth and redox homeostasis in aggressive cellular subpopulations. This nuanced redox tuning where highly oxidative cells simultaneously suppress ROS to avoid lethal damage represents a mechanistic novelty that underpins therapeutic resistance and highlights the need to view OXPHOS not merely as an energy source but as a dynamic redox-sensitive switch regulating cancer cell fate. Nevertheless, reliance on glucose metabolism alone—whether glycolytic or oxidative—is insufficient to meet the full biosynthetic demands of proliferating cancer cells. Consequently, cancer cells extend their metabolic repertoire by exploiting alternative nutrient sources. Among these, glutamine metabolism is especially prominent. Through glutaminolysis, glutamine replenishes tricarboxylic acid (TCA) cycle intermediates, fuels nucleotide synthesis, and supports antioxidant defense by generating NADPH [21]. Many tumors are described as “glutamine addicted,” underscoring its centrality in supporting malignancy. Experimental data confirm this dependency, as inhibition of glutaminase (GLS)—a key enzyme in glutamine catabolism—reduces tumor growth in multiple models [22]. Lipid metabolism emerges as another critical axis of metabolic reprogramming. Specifically, de novo fatty acid synthesis supports membrane biogenesis. Overexpression of fatty acid synthase (FASN) has been associated with tumor progression and metastasis [23]. Moreover, FAO becomes particularly important under metabolic stress, supplying energy and maintaining redox balance when glucose metabolism is compromised [24]. In summary, cancer metabolism encompasses more than just a shift to glycolysis. Tumor cells fine-tune multiple interconnected pathways to support growth, resist stress, and avoid cell death. This metabolic reprogramming lies at the heart of tumor progression and represents a critical node in the triangular axis connecting autophagy and regulated cell death. Metabolic plasticity: adaptations to nutrient and oxygen deprivation One of the most remarkable traits of tumor cells is their metabolic plasticity, the ability to adjust fuel usage and energy-generating pathways according to the availability of nutrients and oxygen. Unlike normal cells, which follow predictable metabolic routes, cancer cells operate in an environment that is frequently deprived of oxygen and nutrients due to poor vascularization and rapid proliferation. To survive under such conditions, they develop versatile metabolic programs that enable them to meet energy demands, maintain biosynthesis, and counteract cellular stress [25]. Among the environmental pressures shaping metabolic plasticity, hypoxia represents one of the most dominant and pervasive drivers of metabolic reprogramming in tumors [26,27]. This adaptation is largely governed by hypoxia-inducible factor-1α (HIF-1α), which upregulates glucose transporters and glycolytic enzymes, thereby ensuring continuous ATP production. In addition to promoting glycolysis, HIF-1α indirectly reduces the entry of pyruvate into the TCA cycle. This shift decreases mitochondrial oxygen consumption and helps limit excessive ROS production under hypoxic conditions [28,29]. However, mitochondria are not entirely abandoned; they adapt by relying more on glutamine and other substrates for biosynthetic reactions, particularly nucleotide and amino acid synthesis, ensuring that essential metabolites are available for cell growth and survival. Experimental data from tumor spheroids and hypoxic tissues confirm these adaptations, showing enhanced glutamine metabolism and altered mitochondrial function in hypoxic cancer cells [30]. This HIF-1α–mediated rewiring exemplifies an innovative redox-sensitive strategy: instead of simply shutting down mitochondria, cancer cells selectively reconfigure mitochondrial metabolism to minimize oxidative damage while preserving biosynthetic capacity, thereby creating a survival advantage that normal cells cannot recapitulate [31]. In addition to hypoxia, nutrient scarcity represents a parallel and equally formidable challenge within the tumor microenvironment (TME). Cancer cells often compete with each other and with stromal and immune cells for glucose, amino acids, and lipids. To overcome this limitation, tumors develop several adaptive strategies. One prominent mechanism is the increased utilization of glutamine. Studies show that tumor cells can outcompete T cells for glutamine, impairing immune function, while inhibiting glutamine metabolism suppresses tumor growth in multiple models. In pancreatic tumors, glutamine synthetase expression is induced during chronic nutrient deprivation, enabling de novo glutamine synthesis in a mechanistic target of rapamycin complex 1 (mTORC1)-dependent manner and supporting continued cell survival [32,33]. Some cancers also rely on alternative amino acids to supports nucleotide production and redox control. Glioblastoma cells upregulate serine and glycine pathways under glutamine-deprived conditions [34]. These nutrient-scavenging and amino acid–diverting strategies highlight a key innovation in cancer metabolism: the selective co-option of serine–glycine–one-carbon and glutamine pathways to sustain biosynthetic flux while dynamically reinforcing redox buffers under glucose and oxygen limitation [35]. Although arginine metabolism is frequently upregulated during starvation, current evidence does not substantiate a direct role in redox buffering; rather, arginine likely functions as a contextual metabolic node that intersects with serine and glutamine networks to maintain proliferative capacity [36]. Consequently, rather than a unified “glutamine–arginine” redox axis, tumors deploy a hierarchical and microenvironment-dependent amino acid program that prioritizes one-carbon and glutamine metabolism for oxidative stress resistance, with arginine serving a compensatory or synergistic role in specific tumor contexts. In parallel, many tumors enhance lipid metabolism under nutrient-limiting conditions. FAO provides both ATP and reducing power, while de novo fatty acid synthesis ensures the availability of membrane components for rapidly dividing cells. In breast cancer, repeated nutrient deprivation leads to a lipogenic phenotype characterized by lipid droplet accumulation, upregulation of adipogenic markers, and enhanced macropinocytosis and autophagy, allowing cells to scavenge extracellular material and fuel lipogenesis [37]. Similarly, cancer cells undergoing peritoneal metastasis directly harvest fatty acids from adjacent adipocytes, which are then used for energy production and membrane biosynthesis [38]. Together, these adaptations illustrate the remarkable plasticity of cancer metabolism in the face of nutrient deprivation, ensuring continued growth, biosynthesis, and survival in a highly competitive microenvironment (Fig. 3). A central control hub of key regulators in tumor cells adapting their metabolism to hypoxia and nutrient deprivation. In hypoxic conditions, HIF signaling enhances glycolysis, inhibits PDH through PDK1, promotes lactate production and export, activates pH regulators such as CAIX, and induces mitophagy and reductive carboxylation to sustain biosynthesis. Under nutrient scarcity, AMPK activation and mTORC1 suppression stimulate autophagy, macropinocytosis, and lipid mobilization, while tumor cells switch to alternative fuels such as fatty acids, lactate, acetate, and branched-chain amino acids. Together, these adaptive programs maintain energy production, redox balance, and biosynthetic capacity, enabling tumor survival in hostile microenvironments. Importantly, metabolic plasticity is not confined to individual cancer cells but often involves cooperation within the TME. For example, glycolytic cancer cells produce and release lactate, which neighboring oxidative cells can take up and use through OXPHOS. This symbiotic interaction optimizes the use of limited resources and helps sustain intratumoral heterogeneity. Cancer-associated fibroblasts (CAFs) and immune cells also contribute to this metabolic cooperation by supplying key metabolites such as lactate and amino acids [39]. Beyond passive metabolite exchange, tumors actively reprogram stromal metabolism. Real-time imaging has shown oscillatory metabolic exchanges between cancer cells and CAFs through autophagy and exosome-mediated communication, reflecting dynamic and bidirectional metabolite flow. This metabolic symbiosis not only allows tumors to thrive under nutrient and oxygen limitations but also supports functional specialization, with distinct populations within the tumor engaging in glycolysis or OXPHOS [40]. The division of metabolic labor and metabolic crosstalk contribute to tumor heterogeneity, aggressive behavior, and immune suppression, further complicating therapeutic targeting in the heterogeneous tumor niche [41]. Fundamentals of autophagy in cancer Autophagy plays a complex role in cancer. The main types of autophagy are macroautophagy, microautophagy and chaperone-mediated autophagy (CMA). The most widely studied form of autophagy is macroautophagy, often simply referred to as autophagy. In this process, cytoplasmic macromolecules, damaged organelles or pathogens are transported to lysosomes, and degraded by lysosomal hydrolases to produce nucleotides, glucose, amino acids, fatty acids, and ATP, which are subsequently recycled to support biosynthesis and energy production, thereby enhancing cell survival under stress conditions [42]. Role of autophagy in tumorigenesis: friend or foe? Autophagy occupies a paradoxical position in cancer biology. Depending on the cellular context, tumor stage, and microenvironmental conditions, it can either act as a guardian that suppresses tumor initiation or as an ally that promotes tumor progression and survival. This dual role has made autophagy both a promising therapeutic target and a complex challenge in cancer treatment. Importantly, this duality also reflects an evolving redox function of autophagy, where the same pathway can either restrain oxidative damage in premalignant cells or help established tumors survive oxidative and metabolic stress [43]. Understanding the context-specific functions of autophagy is therefore critical to deciphering its role in tumorigenesis. In the early stages of tumorigenesis, autophagy primarily functions as a protective mechanism that preserves genomic stability and prevents malignant transformation. For example, monoallelic deletion of the BECN1 gene is common in breast and ovarian cancers. Mouse models with heterozygous Beclin 1 deletion show increased spontaneous tumors [44]. Mechanistically, by degrading damaged organelles, particularly mitochondria through mitophagy, autophagy reduces the accumulation of ROS that can drive DNA damage and mutagenesis. The loss of BNIP3-dependent mitophagy in mammary tumor models led to elevated cellular superoxide, which drives HIF stabilization and tumor progression in vivo, underscoring the tumor-suppressive role of mitochondrial quality control [45]. In addition, autophagy prevents the buildup of misfolded proteins and toxic aggregates, which could otherwise create a cellular environment favorable to oncogenic transformation. Research using autophagy-deficient cells and mouse models has shown that disruption of core autophagy genes such as autophagy related 5 (ATG5) results in genomic instability, accumulation of DNA damage, and higher rates of malignant transformation, confirming the protective effect of autophagy in vivo [46]. Once a tumor has been established, the role of autophagy often shifts from suppressive to supportive. Rapidly proliferating cancer cells face a harsh microenvironment characterized by hypoxia, nutrient deprivation, and oxidative stress. In these conditions, autophagy provides a survival advantage by recycling intracellular components into essential metabolites. The amino acids, fatty acids, and nucleotides generated through autophagic degradation are redirected to fuel biosynthesis and sustain energy production, ensuring continued growth even when external nutrient supply is limited. This metabolic support has been confirmed in experimental models, where inhibition of autophagy in Ras-driven tumors impaired growth and increased cell death, while host-specific deletion of autophagy genes reduced circulating nutrients and limited tumor progression [47]. Moreover, enhanced Beclin 1 activity can provide metabolic support by sustaining autophagic flux. This supports tumor progression by supplying nutrients and energy under stress conditions [48]. A more critical interpretation here is that autophagy is not just a passive recycling pathway, but a stress-adaptation program that rewires tumor metabolism to preserve mitochondrial fitness and redox balance under conditions that would otherwise trigger cell death. In addition to metabolic adaptation, autophagy also plays a critical role in maintaining the viability of cancer stem-like cells, which are often resistant to conventional therapies. By preserving mitochondrial quality and providing substrates for oxidative metabolism, autophagy enables these cells to endure therapeutic stress and repopulate tumors after treatment. Elevated expression of autophagy markers such as Beclin 1 and ATG5 has been detected in cancer stem-like cells, and experimental inhibition of autophagy diminished their self-renewal capacity and sensitized them to chemotherapy in cancer models [49,50]. Furthermore, autophagy can modulate the immune microenvironment. For instance, autophagy-derived metabolites may support the secretion of pro-tumorigenic cytokines. In ovarian cancer, autophagy-dependent secretion of IL-6 has been shown to enhance tumor invasion. Autophagy in tumor-associated stromal cells also contributes, as studies revealed that mesenchymal stem cells within nutrient-deprived tumor cores upregulate autophagy, survive metabolic stress, and secrete growth-promoting factors. This stromal autophagy supports a “battery” model, where recycled nutrients such as amino acids and fatty acids are transferred to neighboring cancer cells, sustaining their proliferation [51,52]. Overall, autophagy in cancer acts as both a friend and foe, depending on the timing and context. In early stages, it helps prevent malignant transformation by maintaining cellular quality control. Yet, in advanced tumors, it supports cancer survival under stress, often aiding resistance to therapy. Inhibiting autophagy may make tumors more sensitive to treatment, while in some cases, stimulating it can reduce oncogenic stress. Because it can both protect against and promote tumor growth, targeting autophagy requires careful consideration of the tumor's stage, genetics, and environment. A stronger concluding message would be that autophagy should be framed as a context-dependent redox and metabolic checkpoint, not merely as a binary tumor suppressor or promoter. The triangular crosstalk: how autophagy, metabolism, and cell death interact in tumor survival In tumors, autophagy, metabolism, and cell death are interconnected processes that shape cancer survival. Autophagy helps cells cope with stress by recycling nutrients, but it can also promote cell death under certain conditions. Cancer cells often alter their metabolism to fuel rapid growth, and these metabolic shifts influence autophagy and cell death pathways (Fig. 4). Critically, this triangular crosstalk reframes autophagy not only as a recycling pathway but as a central node that integrates metabolic status, redox balance, and cell-death fate decisions, thereby determining whether a stressed cancer cell survives or self-destroys. The complex interplay between metabolic stress, autophagy, and cell death pathways in cancer cells. (a) depicts autophagy as a metabolic adaptation, where nutrient deprivation activates AMPK, which phosphorylates ULK1 to initiate autophagosome formation, delivering macromolecules to lysosomes for recycling into amino acids, fatty acids, nucleotides, and substrates for ATP, NADPH, and TCA cycle activity. (b) shows metabolic control of autophagy, highlighting how growth signals and nutrients activate mTORC1 to inhibit ULK1, whereas low ATP/high AMP triggers AMPK-mediated ULK1 activation, and acetyl-CoA/NAD+ modulates ATG protein function. (c) illustrates how autophagy regulates cell death thresholds, including mitophagy removing damaged mitochondria to prevent cytochrome-c release, selective degradation of pro-apoptotic factors to enhance survival, and hyperactive autophagy leading to autophagic cell death; BCL-2 family proteins interact with Beclin 1 to coordinate autophagy and apoptosis. (d) emphasizes feedback loops, showing how metabolic reprogramming such as enhanced glycolysis and hexokinase II localization to mitochondria suppresses apoptosis, while cell death releases DAMPs that remodel the TME and select for metabolically adapted cells. Autophagy as a metabolic adaptation to stress Autophagy is essential for physiological adaptation to metabolic stress in normal tissues, including the liver and heart, where it preserves organelle quality control and systemic energy balance [53]. In the hostile and fluctuating environment of tumors, autophagy functions as a core metabolic adaptation that enables cancer cells to survive and proliferate under conditions of nutrient scarcity, hypoxia, and therapeutic pressure. By degrading and recycling intracellular components—including proteins, lipids, and organelles—autophagy sustains both energy production and biosynthetic capacity when external resources are limited or damaged. This recycling process supports multiple metabolic demands. Autophagy-derived amino acids and fatty acids feed into central pathways such as the TCA cycle and β-oxidation, maintaining ATP production, redox balance, and nucleotide pools during metabolic stress. For example, in Ras-driven lung cancer, deletion of Atg7 causes accumulation of dysfunctional mitochondria, reduced oxygen consumption, and depletion of TCA intermediates—changes that impair tumor growth but can be partially rescued by metabolic supplementation [54]. Similarly, in pancreatic ductal adenocarcinoma (PDAC) with activating Kirsten Rat Sarcoma Viral Oncogene Homolog (KRAS) mutations, high autophagy activity preserves mitochondrial function and supports proliferation; pharmacological autophagy inhibition reduces oxygen consumption and suppresses tumor progression in preclinical models [55,56]. These findings underscore that autophagy is not merely compensatory, but actively sustains mitochondrial metabolism and metabolic fitness during oncogenic stress. Functionally, this adaptation provides tumors with a survival advantage, buffering against metabolic and oxidative stress. In the context of nutrient deprivation and hypoxia, intracellular ROS (e.g., hydrogen peroxide) are direct signals activating the ATM/CHK2 axis, which phosphorylates the Ser90/Ser93 residues of Beclin 1 and stimulates autophagic flux. Importantly, CHK2-mediated autophagy clears damaged mitochondria through the induced NIX and BNIP3, which in turn limits ROS levels. This feedback loop constitutes a pivotal adaptive strategy that protects cells from oxidative stress under metabolic fluctuations and hypoxia [57]. Moreover, autophagy/mitophagy protects cancer cells from oxidative and mitochondrial damage induced by therapies that target nutrient supply or mitochondrial function [12]. For instance, hypoxia-induced autophagy supports glioblastoma survival during antiangiogenic treatment, while mitophagy helps pancreatic cancer cells resist chemotherapy [58,59]. Although autophagy is predominantly cytoprotective, overwhelming or sustained activation under persistent nutrient deprivation can precipitate autophagic cell death, highlighting the necessity for tight regulatory constraints on this pathway [60]. In sum, autophagy in cancer acts as more than a passive survival mechanism—it is an active remodeler of cellular metabolism, linking nutrient recycling, mitochondrial quality control, and energy homeostasis. This integration reinforces the triangular axis between autophagy, metabolism, and cell death, shaping tumor progression and influencing therapeutic response. Metabolic control of autophagy induction and execution Autophagy is a highly regulated process that responds dynamically to the metabolic state of the cell. The induction and execution of autophagy are closely governed by metabolic cues that reflect energy availability, nutrient status, and cellular stress (Fig. 5). Tumor cells exploit this interplay to balance biosynthetic demands, energy production, and stress responses within the nutrient-and oxygen-limited TME. Regulation of autophagy in cancer cells. Metabolic enzymes, nutrient pathways, and stress sensors influence autophagy initiation. Core autophagy machinery (ULK complex, PI3K complex, ATG conjugation systems, LC3 lipidation, autophagosome closure, and fusion via LAMP1/2 and RAB7) executes the process. Key signaling pathways integrate growth, nutrient, and stress signals to activate or inhibit autophagy. Functional outcomes include stress survival, quality control, antigen presentation, and therapy resistance. Glycolysis is not only a pathway for ATP production but also a regulatory hub for autophagy. In cancer models such as glioblastoma, elevated glycolytic enzyme expression correlates with enhanced autophagy, particularly in stem-like tumor cells. Hexokinase 2 (HK2), which is often upregulated in cancer, modulates glycolytic flux and links autophagic turnover to energy regulation [61]. Similarly, pyruvate kinase M2 (PKM2), a cancer-associated isoform, contributes to metabolic rewiring and influences autophagic flux. FGFR1-mediated phosphorylation at Y105 alter PKM2's oligomeric state, promoting aerobic glycolysis while modulating energy-sensing cascades and autophagy-related signaling. LDH-A also contributes by converting pyruvate to lactate, with downstream effects on both metabolism and the TME. Phosphorylation of LDH-A by ULK1 enhances its activity and elevates lactate production, which in turn promotes Vps34 lactylation—an event that facilitates autophagic flux and lysosomal trafficking [62,63]. Additionally, enzymes such as GAPDH respond to energy stress by relocating to the nucleus under AMPK control, where they activate Sirt1 to initiate autophagy [64]. Other glycolytic regulators—including PGK1, which phosphorylates Beclin 1 under hypoxia or glutamine deprivation—directly activate autophagic programs essential for tumorigenesis [65]. These glycolytic checkpoints ensure that autophagy is engaged when glucose availability fluctuates or when mitochondrial function is impaired, thus maintaining energy homeostasis. Furthermore, mitochondria also produce metabolites that regulate autophagy. For example, citrate and acetyl-CoA levels influence acetylation of autophagy-related proteins, altering their activity [66]. Cytosolic acetyl-CoA, generated from mitochondrial citrate via ATP citrate lyase, acts as a key donor for protein acetylation, including histones and autophagy effectors. A decline in mitochondrial function reduces acetyl-CoA availability, which leads to global changes in acetylation that affect autophagy induction [67]. Accumulation of mitochondrial ROS (mtROS) under stress triggers mitophagy. This feedback mechanism maintains mitochondrial quality and promotes cell survival under metabolic stress [68]. Beyond ROS, functional mitochondrial complex I is required for optimal autophagy, as its inhibition disrupts mitochondrial-associated membrane dynamics and phospholipid synthesis necessary for autophagosome formation. During energy deprivation, a mitochondrial module involving Snf1, Mec1, and Atg1 activates autophagy by supporting respiration [69]. Furthermore, mitochondrial calcium transport plays a regulatory role: while NCLX knockdown impairs autophagosome formation and reduces autophagic flux, MCU suppression enhances autophagy [70]. The regulation of autophagy in cancer is tightly integrated with nutrient-sensing metabolism-adaptation and stress-response signaling pathways. AMPK acts as a cellular energy sensor that promotes autophagy through phosphorylation of the ULK1 complex when the cellular AMP/ATP ratio rises, enabling tumor cells to survive metabolic stress by removing damaged organelles and providing nutrients. AMPK also indirectly promotes autophagy by inhibiting mTORC1. In the tumor context, AMPK-driven autophagy under glucose starvation promotes tumor cell survival by maintaining metabolic homeostasis through mitochondrial biogenesis and metabolic pathway modulation. In contrast, mTOR functions as a negative regulator of autophagy under nutrient-rich conditions, and aberrant mTOR activation in many cancers suppresses autophagy, influencing tumor growth and therapeutic response. However, when nutrient availability declines or cells face therapeutic stress, reduced mTOR activity allows autophagy to proceed, enabling tumor cells to recycle intracellular components for survival. This duality highlights why mTOR is a central hub in the balance between growth and catabolism in cancer cells [71,72]. In summary, the metabolic control of autophagy induction and execution exemplifies the intricate crosstalk between metabolism and cellular recycling processes. Metabolic sensors such as AMPK and mTORC1, along with metabolites like acetyl-CoA and ROS, not only determine when autophagy begins but also shape how it is carried out. This regulation equips tumor cells with the flexibility to cope with environmental challenges and underscores the importance of the triangular relationship between autophagy, metabolism, and cell death in cancer survival and therapy resistance. Feedback loops between metabolic reprogramming and cell death pathways Cell death in cancer involves various pathways, each impacting tumor growth and treatment. Metabolic changes can modulate the activation or suppression of cell death (Fig. 6), while cell death signaling can, in turn, reshape metabolic pathways. This bidirectional communication creates a regulatory network that allows cancer cells to finely adjust their fate under environmental and therapeutic stresses. Metabolic and stress-regulated control of cell death pathways. The top panel shows the metabolic context, where key metabolites (ATP/AMP, NADPH/ROS, acetyl-CoA, iron, glucose, glutamine, and fatty acids) and master regulators (AMPK, mTOR, p53, and HIF-1α) set the energy and redox tone of the cell. At the center, the stress hub integrates DNA damage, hypoxia, cytokines, and nutrient, ER, or oxidative stress as upstream triggers. Apoptosis branches into extrinsic signaling through death receptors, FADD, and caspase-8, and intrinsic signaling through BAX/BAK-mediated MOMP, cytochrome c, APAF1, and caspase-9. Autophagic cell death is controlled by ULK1–ATG13–FIP200, BECN1–VPS34, ATG5–ATG12–ATG16L1, and LC3-II, with mTOR as an inhibitor, AMPK as an activator, and TFEB/TFE3 promoting transcriptional programs. Pyroptosis is mediated by NLRP3 inflammasome assembly with ASC and caspase-1, which cleaves GSDMD to form pores and releases IL-1β and IL-18. Necroptosis proceeds through RIPK1/RIPK3 and MLKL pore formation but is restrained by caspase-8. Ferroptosis arises when System Xc-/SLC7A11 is suppressed, leading to GSH depletion, GPX4 inactivation, iron accumulation, and lipid peroxidation, with ACSL4 and LPCAT3 priming membranes for damage. In contrast to other forms of cell death, cuproptosis is characterized by DLAT aggregation-triggered proteotoxic stress and impaired TCA cycle enzyme function. Apoptosis is a programmed form of cell death essential for maintaining tissue homeostasis and eliminating damaged or abnormal cells. While the molecular machinery of apoptosis has been extensively studied, several studies have shown that cellular metabolism profoundly influences both the intrinsic and extrinsic apoptotic pathways in cancer. The intrinsic apoptotic pathway, also known as the mitochondrial pathway, is initiated by intracellular stress signals. Central to this process is the regulation of mitochondrial outer membrane permeabilization (MOMP), governed by the Bcl-2 family of proteins. Importantly, metabolic status plays a decisive role in modulating this pathway. High glycolytic activity produces ATP and biosynthetic intermediates that support tumor cell survival. This activity also increases expression of anti-apoptotic proteins like Bcl-2 and Bcl-xL, suppressing pro-apoptotic signaling. For example, in nasopharyngeal carcinoma cells, glycolysis driven by the IRF2/CENP-N/AKT axis inhibited apoptosis [73]. In addition, lactate accumulation, another result of high glycolytic flux, stabilizes anti-apoptotic proteins and helps preserve mitochondrial integrity. In A549 lung cancer cells deprived of glucose, lactate increased Bcl-2 expression, and Bcl-2 knockdown sensitized cells to apoptosis under lactate deprivation. Lactate protects against glucose starvation-induced apoptosis through the Akt/mTOR/Bcl-2 pathway [74]. In contrast, superoxide anions emanating from the mitochondrial electron transport chain (ETC) can tip the balance toward apoptosis by activating pro-apoptotic members such as Bax and Bak to induce MOMP, following cytochrome c release, apoptosome formation, and caspase activation. Cancer cells with inherent mitochondrial defects are particularly vulnerable to ROS-driven apoptosis [75]. Thus, redox stress emerges as a double-edged sword: it can be a potent pro-apoptotic trigger in genomically unstable cells, yet tumor cells frequently rewire metabolism to limit ROS below the threshold required for MOMP, effectively escaping this death signal. Glutamine addiction creates a redox-dependent survival niche, preventing oxidative stress-induced apoptosis. Glutamine metabolism sustains GSH production to directly scavenge intracellular ROS. Moreover, glutamine and its metabolic derivatives (e.g., α-KG) promote c-MYC activity through post-translational modification (PTM), thereby transcriptionally upregulating anti-apoptotic Bcl-2/Bcl-xL and downregulating pro-apoptotic PUMA/Bax [76,77]. Thus, the balance between nutrient supply and metabolic demand directly dictates mitochondrial vulnerability to apoptosis. Glycolysis, lactate metabolism, glutamine utilization, mtROS, and the Bcl-2 family are integrated as nodes of a “metabolic rheostat” for apoptosis, where oncogenic signaling tunes glycolytic and mitochondrial fluxes to keep ROS and MOMP below the lethal threshold. The extrinsic apoptotic pathway is activated by extracellular death ligands binding to their cell-surface receptors, such as Fas ligand (FasL) and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL). This binding triggers the formation of the death-inducing signaling complex (DISC), leading to caspase-8 activation and initiation of downstream apoptotic cascades [78,79]. Importantly, metabolism intersects with this pathway in multiple ways. Adequate glycolytic activity and mitochondrial function are necessary to provide ATP for caspase activation and execution of apoptosis. Studies show that metabolic stress, such as glucose deprivation, can impair DISC assembly and reduce caspase-8 activation, enabling cancer cells to resist death signals from TRAIL or FasL [78,79]. Furthermore, lipid metabolism also plays a major role. The clustering of death receptors within cholesterol- and sphingolipid-rich microdomains, known as lipid rafts, is critical for DISC formation and efficient signal transduction. Changes in membrane lipid composition can disrupt receptor clustering and weaken apoptotic signaling. EMILIN2, an extracellular matrix protein, can activate DR4 and DR5 within lipid rafts to induce extrinsic apoptosis, highlighting the link between extracellular cues, lipid organization, and death receptor activation [80]. Tumors often acquire strategies to suppress apoptotic pathways, necessitating alternative mechanisms of cellular demise. In recent years, non-apoptotic cell death modalities have gained prominence for their roles in tumor suppression and therapy resistance. Necroptosis, ferroptosis, pyroptosis, autophagic cell death, and cuproptosis represent distinct processes with unique molecular drivers, yet they are all influenced by metabolic reprogramming. Their dual nature—as both tumor-suppressive and tumor-promoting under different contexts—adds complexity to cancer biology and therapy design. Necroptosis is a caspase-independent, regulated form of necrosis that serves as a backup death pathway when apoptosis is blocked. It is primarily governed by receptor-interacting serine/threonine-protein kinase 1 (RIPK1) and 3 (RIPK3), along with the executioner protein mixed lineage kinase domain-like pseudokinase (MLKL). Necroptosis can act as a tumor suppressor by eliminating apoptosis-resistant cells [81]. However, necroptotic cell lysis releases damage-associated molecular patterns (DAMPs) that drive a pro-inflammatory TME, potentially supporting angiogenesis, immune suppression, and metastasis [82]. Importantly, metabolic reprogramming and mtROS strongly influence the functional outcome of necroptosis. Specifically, enhanced glutaminolysis lead to succinate accumulation, which induces mtROS overproduction by reverse electron transport (RET). ROS causes the oxidation of cysteine residues in RIPK1, subsequently activating RIPK1 autophosphorylation on serine residue 161, which triggers functional necrosome formation [83,84]. This suggests that necroptosis is not only a backup but a redox-modulated switch whose outcome is shaped by the same metabolic perturbations that are targeted by cancer therapies. Ferroptosis is characterized by iron-dependent lipid peroxidation and catastrophic plasma membrane damage [85]. In particular, superoxide anions that originate from respiratory chain electron leakage or NADPH oxidase further fuel the Fe2+-mediated Fenton reaction when cytoplasmic Fe2+ levels are overloaded. This reaction produces hydroxyl radicals (•OH)—one of the most reactive ROS—which initiates the oxidation of polyunsaturated fatty acids (PUFAs) [86]. PUFAs in membrane phospholipids, particularly phosphatidylethanolamine species generated during lipid remodeling, are prime substrates for peroxidation and a key execution point in ferroptosis, while monounsaturated fatty acids (MUFAs) confer resistance to ferroptosis. Enzymes like ACSL4, LPCAT3, and ALOX facilitate the PUFA-CoA synthesis and membrane incorporation, leading to massive PL-PUFA-OOH accumulation and subsequent irreversible membrane damage [87]. Beyond directly driving lipid peroxidation, ROS also deplete intracellular antioxidant systems such as GPX4 [88]. GSH homeostasis is critical for GPX4 function. GPX4 utilizes GSH as a reducing substrate to reduce lipid peroxides to their corresponding lipid alcohol forms, consequently arresting ongoing oxidative damage [89]. The system xc– transporter, with SLC7A11 as its core functional subunit, mediates the uptake of extracellular cystine, which serves as the rate-limiting substrate for GSH biosynthesis [90]. Accordingly, glutamine metabolism exerts a dual role in ferroptosis regulation: GLS1-mediated glutaminolysis promotes GSH synthesis and inhibits ferroptosis via the SLC7A11/GPX4 axis; conversely, GLS2, rather than GLS1, elevates lipid ROS levels by promoting the conversion of glutamate to α-KG, thereby facilitating ferroptosis [91,92]. Thus, redox chemistry, iron, lipid and amino acid metabolism are related biological processes that control aspects of ferroptosis [93]. Pyroptosis is an inflammatory form of programmed cell death mediated by the gasdermin family of pore-forming proteins. It is typically initiated when caspase-1 or caspase-4/5/11 cleaves gasdermins, releasing their N-terminal pore-forming domains. This leads to plasma membrane rupture, cell swelling, and the release of pro-inflammatory cytokines such as IL-1β and IL-18, along with DAMPs [94,95]. Hypoxia signaling can modulate inflammasome assembly and gasdermin activation, while mitochondrial dysfunction amplify pyroptotic cascades. Lactate induces inflammasome complex formation by AARS2-dependent lysine lactylation of NLRP3 [96]. In colitis-associated cancer, FAO fuels acetyl-CoA production, which increases α-tubulin acetylation to support NLRP3 inflammasome assembly [97]. On the other hand, ROS-mediated chemical modification serves as an essential connection between redox imbalance and inflammasome signaling. For example, cofilin-1 is a negative regulator of the NLRP3, which can be oxidized by ROS and dissociated from NLRP3, thereby promoting inflammasome assembly [98]. Intriguingly, mtROS directly facilitates GSDMD pore formation by enhancing the palmitoylation of GSDMD Cys191. This finding challenges the notion that cleavage is the only initiator of GSDMD activation [99]. Accordingly, pyroptosis can be viewed as a “metabolic-inflammatory checkpoint”. When metabolic stress and redox imbalance converge on the inflammasome-gasdermin axis, they convert cellular demise into an immunostimulatory signal, reshaping the TME [100]. Chronic or dysregulated pyroptosis can promote tumor progression by sustaining a pro-tumor inflammatory microenvironment. Thus, pyroptosis in cancer is context-dependent, and the interplay between tumor metabolism, ROS dynamics, and inflammasome activity shapes whether pyroptosis functions as an ally or adversary in cancer progression. Autophagic cell death (ACD) is a caspase-independent form of regulated cell death that arises when excessive or dysregulated autophagy leads to massive self-digestion and degradation of essential cellular components. Morphologically, it is characterized by abundant autophagosomes and heightened lysosomal activity. In melanoma cells under arginine deprivation, an initially protective autophagic response progressed to massive autophagy and cell death. ACD has also been observed in apoptosis-resistant tumors—such as Bax/Bak-deficient cells—where it becomes the dominant death route. NAD+ synthesis inhibitors induced ACD in myeloma and leukemia models [101]. Some natural alkaloids promote ACD and suppress tumor growth via ROS-dependent activation of AMPK or ERK1/2 signaling, and this antitumor effect can be rescued by autophagy inhibitors [102,103]. Many cancers avoid lethal autophagy by maintaining it at a “Goldilocks” level—enough to cope with stress but insufficient to trigger self-destruction. Low activation of pro-death regulators such as JNK can blunt lethal autophagy and promote tumorigenesis [104,105]. Cuproptosis is caused by copper-induced mitochondrial dysfunction [106]. Excessive mitochondrial Cu+ directly interacts with lipoylated DLAT to induce its aggregation, meanwhile destabilizing the Fe–S cluster proteins. DLAT aggregation-triggered proteotoxic stress and impaired TCA cycle enzyme function distinguish cuproptosis from other forms of cell death. This disruption of mitochondrial function can lead to the accumulation of ROS; however, ROS elimination cannot fully suppress copper-induced cell death, indicating that ROS may act as a secondary rather than an initiatory factor in cuproptosis. GSH and other endogenous copper-chelating metabolites effectively mitigate the cytotoxicity of elesclomol–Cu [107,108]. Since cuproptosis is heavily reliant on mitochondrial respiration, GAPDH or ARID1A deficiency sensitizes cancer cells to copper treatment by inhibiting aerobic glycolysis [109,110]. A more innovative framing of cuproptosis would emphasize that it represents a copper-dependent “metabolic catastrophe” mechanism, where copper overload directly targets the core energy-producing machinery of the cell, opening a new therapeutic window for cancers that are resistant to classical apoptosis- or ferroptosis-based regimens. Cell death pathways feed back on metabolism by modulating key metabolic enzymes, energy-, and redox-sensitive signaling pathways to orchestrate metabolic adaptation under therapeutic and nutrient stress. For exmaple, RIPK3 activation during necroptosis is tightly connected with mitochondrial metabolism. Specifically, RIPK3 directly binds and allosterically potentiates the activity of multiple metabolic enzymes, such as pyruvate dehydrogenase complex (PDC), glutamate–ammonia ligase (GLUL), and glutamate dehydrogenase1 (GLUD1) to enhance aerobic respiration [83]. In comparison, elevated glycolysis serves as an intrinsic adaptive mechanism during cuproptosis. Blocking glycolytic pathway by targeting LDH-A reverses cuproptosis resistance [111]. Mechanistically, high levels of copper inhibits the ubiquitin-mediated degradation of HIF-1α, thereby driving glycolysis [112]. Similarly, pyroptosis-derived IL-1β acts in a paracrine manner to initiate the NF-κB pathway, which further transcriptionally upregulates HIF-1α and reinforces the reciprocal metabolic-inflammatory loop [113]. Tumor cells under ferroptotic stress also increase glycolysis and suppress mitochondrial OXPHOS to reduce ROS production and maintain ATP supply. Additionally, the Nrf2 signaling cascade promotes the transcription of SLC7A11, GPX4, and FTH1, coordinating antioxidant defense and metabolic rewiring. To counteract lipid peroxidation, cancer cells boost SREBP1-SCD1-mediated MUFA synthesis and downregulate key enzymes such as FADS1/2 and ELOVL5 to restrain PUFA generation, attenuating ferroptosis sensitivity [[114], [115], [116], [117]]. Therapeutic interventions targeting metabolism or cell death must therefore consider these dynamic interactions to avoid unintended compensation that leads to treatment resistance. For instance, inhibiting glycolysis may sensitize cells to apoptosis initially, but metabolic rewiring through alternative pathways or autophagy activation can restore survival. Similarly, blocking one death pathway may enhance reliance on another, necessitating combination strategies [118]. In summary, the feedback loops between metabolic reprogramming and cell death pathways form a complex, adaptive network that governs tumor fate. These dynamic relationships influence apoptosis, necroptosis, ferroptosis, and other regulated death modes. Understanding them is critical for designing therapies that disrupt both metabolic and cell death programs, preventing cancer from evading treatment. Autophagy-mediated regulation of cell death thresholds Autophagy plays a critical role in regulating the sensitivity of tumor cells to various forms of cell death. By acting as a cellular quality control and metabolic support system, autophagy adjusts the threshold at which cells commit to death, thereby influencing tumor survival and therapeutic response. Experimental studies in multiple cancer types confirm that autophagy can either raise or lower this threshold depending on its level and duration of activity. This regulatory capacity makes autophagy a central modulator within the triangular relationship linking metabolism, autophagy, and cell death pathways [119,120]. Cancer cells display remarkable adaptability by dynamically balancing autophagy with diverse regulated cell death programs, including apoptosis, necroptosis, ferroptosis, and pyroptosis. This plasticity enables tumors to escape one death pathway while exploiting another to sustain survival. At basal levels, mitophagy limits the release of pro-apoptotic factors like cytochrome c and reduces ROS accumulation, stabilizing mitochondrial integrity and delaying intrinsic apoptotic activation. Research using mitophagy inducers such as magnolol shows that mitophagy can suppress apoptosis and promote tumor survival, while mitophagy inhibition sensitizes cancer cells to apoptotic death. This maintenance reduces cellular stress and delays the activation of death pathways, effectively raising the threshold for apoptosis and other forms of programmed cell death [119,120]. Crosstalk between autophagy and necroptosis in cancer is bidirectional and highly dependent on the metabolic and stress context. In many cases, autophagy delays or suppresses necroptosis by maintaining ATP levels, removing damaged organelles, and limiting ROS accumulation. By preserving mitochondrial function, autophagy prevents the activation of RIPK1–RIPK3 signaling that initiates necroptosis. For example, in breast cancer MCF-7 cells exposed to DNA-damaging agents, autophagy clearance of damaged mitochondria delays both apoptotic and necroptotic death, prolonging tumor cell survival under stress [121,122]. Similarly, in ATP depletion models, autophagy is activated first as a survival response; only when it fails to restore energy balance does necroptosis ensue. Conversely, impaired autophagy can sensitize cancer cells to necroptosis. Loss of core autophagy genes such as ATG5, ATG7, or Beclin 1 increases metabolic stress and ROS, both of which can promote necroptotic signaling. In Atg5-deficient mouse embryonic fibroblasts, amino acid starvation markedly enhances necroptotic death. Pharmacological inhibition of autophagy in some cancer models similarly increases necroptosis, highlighting a therapeutic opportunity in apoptosis-resistant tumors [123,124]. Beyond indirect effects, autophagic components can directly regulate necroptotic machinery. The autophagy receptor p62/SQSTM1 binds RIPK1, facilitating necrosome localization to autophagosomes and efficient MLKL phosphorylation. Loss of p62 not only disrupts necroptosis but can also shift the mode of death toward apoptosis, demonstrating that autophagic scaffolds influence the choice between programmed death pathways [125,126]. Likewise, crosstalk between autophagy and ferroptosis in cancer is multifaceted and highly context dependent, with autophagy capable of either promoting or suppressing ferroptotic death. The best-characterized pro-ferroptotic mechanism is ferritinophagy, the selective degradation of ferritin via the autophagy cargo receptor NCOA4, which liberates stored iron and increases the labile iron pool. Elevated iron accelerates lipid peroxidation, a biochemical hallmark of ferroptosis. In human cancer cell lines and mouse tumor models, activation of ferritinophagy by ferroptosis inducers such as erastin leads to robust ferroptotic death, whereas genetic silencing of ATG5 or NCOA4 blocks this effect [127,128]. Autophagy also promotes ferroptosis through other selective degradation processes. Clockophagy, in which the circadian protein ARNTL is targeted for autophagic degradation, sensitizes tumor cells to ferroptosis and impairs tumor growth in vivo [127,128]. GPX4 degradation via autophagy is another pathway, exemplified in bladder cancer cells where the ferroptosis inducer Fin56 requires autophagy-mediated GPX4 turnover; inhibiting autophagy at early or late stages protects against this form of death. Additional processes such as lipophagy (mobilizing lipid droplets to provide peroxidizable substrates) further sensitize cancer cells to ferroptosis [129]. The regulatory relationship between mitophagy and ferroptosis remains controversial. Mitophagy often plays a cytoprotective role, delaying ferroptosis. Lipid peroxidation-induced endoplasmic reticulum stress upregulates Parkin transcription in an ATF4-dependent manner to facilitate mitophagy. By reducing lipid peroxidation products and mtROS, mitophagy exerts an inhibitory effect on ferroptosis. Mitophagy-deficient tumors are deprived of this anti-ferroptotic defense mechanism, resulting in robust lipid peroxidation and profound ferroptotic cell death upon exposure to erastin, RSL3, cysteine deprivation, radiotherapy or immunotherapy [130]. In BNIP3/NIX double knockout HeLa cells, the mitophagy is entirely blocked, whereas bulk autophagy remains intact. The deficiency in mitophagy led to an increase in mtROS, rendering it more susceptible to ferroptotic stimuli than wild-type cells [131]. These findings indicate that mitophagy confers protection against ferroptosis via eliminating dysfunctional mitochondria and excessive mtROS. However, overwhelming mitophagy sensitize cells to ferroptosis. Inhibition of myoferlin, SIRT3, or O-GlcNAcylation triggers mitochondria fragmentation and enhanced mitophagy, which provides an extra source of labile iron, culminating with heightened vulnerability to ferroptosis [[132], [133], [134]]. MitoQ or MitoTEMPO eliminates mitochondrial superoxide, suppresses mitophagy, and thereby alleviates RSL3/erastin-induced ferroptosis [135]. Accordingly, the intensity of mitophagy differs across contexts, which in turn exert divergent impacts on ferroptosis. Similarly, crosstalk between autophagy and pyroptosis in cancer is increasingly recognized as an important, yet context-dependent, regulatory axis in tumor biology. Autophagy can act as a negative regulator of pyroptosis by degrading inflammasome components and upstream activators, thereby reducing caspase-1 activation and limiting the release of pro-inflammatory cytokines. This suppressive effect helps prevent excessive inflammation that could destabilize the TME. For instance, loss of autophagy proteins such as ATG16L1 or ATG7 leads to heightened inflammasome activation and increased cytokine release in macrophages. In breast and colon cancer cells, autophagic clearance of damaged mitochondria lowers ROS and mitochondrial DNA levels—both potent inflammasome triggers—thereby reducing pyroptotic death and dampening immune activation [136,137]. In contrast, autophagy can also promote pyroptosis under certain conditions, especially when it sustains ROS production or supports metabolic stress signals that amplify inflammasome activation. In ovarian cancer cells, nobiletin induces ROS-mediated pyroptosis through autophagy regulation [138], while in melanoma, inhibition of eEF-2K sensitizes cells to pyroptosis via autophagy-linked mechanisms [139]. Autophagy may further enhance pyroptosis by facilitating antigen presentation and modulating immune checkpoint proteins such as PD-L1, which can indirectly strengthen immune-driven inflammatory cell death [140]. In summary, autophagy serves as a gatekeeper of cell death, modulating the threshold for apoptotic and non-apoptotic pathways through metabolic support and quality control. Its effects are context-dependent—protecting tumor cells when moderate, but potentially triggering cell death when excessive. This regulation reinforces the triangular axis connecting autophagy, metabolism, and cell death, and highlights autophagy's pivotal role in tumor survival, resistance to therapy, and, in some cases, therapy-induced death. Autophagy and metabolism in therapeutic resistance Autophagy-mediated drug resistance: mechanisms and evidence One of the major challenges in cancer treatment is the development of resistance to chemotherapy, targeted therapies, and radiotherapy [141,142]. Although further research is needed to define the precise context, we can postulate that autophagy is a crucial mechanism contributing to therapy resistance (Table 1) [143]. By sustaining metabolic balance, removing damaged cellular components, and modulating cell death pathways, autophagy enables tumor cells to survive therapeutic stress that would otherwise be lethal. Critically, autophagy converts metabolic and oxidative stresses that are designed to kill into signals that reinforce tumor resilience, thereby acting as a central node in therapeutic resistance rather than a passive bystander [144]. Table 1Summary of autophagy mediated drug resistance mechanisms in cancer therapy.Table 1Drug/TreatmentCancer typeKey molecule/Pathway involvedDOICDDP and VP16small cell lung cancerMETTL3/DCP2/Pink1-Parkin/mitophagyhttps://doi.org/10.1186/s13046-023-02638-9CDDPNSCLCE2F1/autophagy/ALDH1A1https://doi.org/10.1186/s13046-025-03506-4CDDPNSCLCmTORC1/autophagyhttps://doi.org/10.1038/s41467-020-18504-7OsimertinibNSCLCNCOA4/ferritinophagyhttps://doi.org/10.1038/s41467-024-48433-8Anti-PD-1NSCLCNDP52/autophagy/MHC-I degradationhttps://doi.org/10.1038/s41401-025-01541-7Anlotinib combined with anti-PD-1NSCLCAKT/mTOR/autophagyhttps://doi.org/10.1136/jitc-2024-0108125-FU/OxaliplatinCRCIL-6/JAK2/BECN1/autophagyhttps://doi.org/10.1038/s41467-021-23923-1CDDPCRCSNHG14/miR-186/ATG14/autophagyhttps://doi.org/10.1016/j.biopha.2019.109580IrinotecanCRCSOX2-βcatenin/Beclin 1/autophagyhttps://doi.org/10.1038/s41419-021-03733-5IFN-γ and TNFCRCATG16L1/autophagyhttps://doi.org/10.1038/s41467-023-41618-7BevacizumabCRCHistone lactylation/RUBCNL/BECN1/autophagyhttps://doi.org/10.1080/15548627.2023.2249762ApatinibCRCIRE1α/autophagyhttps://doi.org/10.1016/j.canlet.2018.05.04CDDPHCCBCAT1/mTOR/autophagyhttps://doi.org/10.1038/s41419-021-03456-7CDDPHCCOPN/integrin αvβ3/FoxO3a/autophagyhttps://doi.org/10.1016/j.canlet.2016.09.033LenvatinibHCCSTX6/Beclin 1/VTI1A/VAMP3/autophagyhttps://doi.org/10.1038/s41388-025-03371-7SorafenibHCCCDK9/SIRT1/FOXO3/BNIP3/PINK1/PRKN/mitophagyhttps://doi.org/10.1080/15548627.2021.2007027Anti-PD-1HCCCDK5/PD-L1/HSC70/autophagyhttps://doi.org/10.1136/jitc-2023-007529CDDPGastric CancercircCUL2/miR-142-3p/ROCK2/autophagyhttps://doi.org/10.1186/s12943-020-01270-xCDDPGastric CancerTRIB3/FOXO3/FOXM1/autophagyhttps://doi.org/10.1016/j.phymed.2024.156161TamoxifenBreast cancerH19/SAHH/DNMT3B/autophagyhttps://doi.org/10.1186/s13045-019-0747-0Anti-HER2 therapiesHER2 breast cancerGSDMB/Rab7/LC3B/autophagyhttps://doi.org/10.1186/s13046-022-02497-wPI3K/AKT inhibitorsTNBCAutophagyhttps://doi.org/10.1186/s12967-022-03462-zPaclitaxelTNBCUBC12/TRIM25/TFEB/autophagyhttps://doi.org/10.1186/s13046-024-03085-wDoxorubicinTNBCANXA2/HSF1/ATG7/autophagyhttps://doi.org/10.1080/15548627.2024.2305063Anti-PD-L1 combined with paclitaxelTNBCATAD3A/PINK1/mitophagyhttps://doi.org/10.1038/s41422-022-00766-zCDDPCervical cancerMAP4K4/SOX6/PI3K-Akt-mTOR/autophagyhttps://doi.org/10.1038/s41419-021-04474-1Anti-PD-1Cervical cancerNANOG/LC3B/p-EGFRhttps://doi.org/10.1080/15548627.2020.1805214GefitinibHNSCCIL-6/STAT3/autophagyhttps://doi.org/10.1016/j.canlet.2024.216612anti-PD-1HNSCCLIMP-2/autophagy/GSK3β/β-cateninhttps://doi.org/10.1038/s41368-023-00229-0CDDPHNSCCRAB3B/autophagyhttps://doi.org/10.1038/s41419-022-04627-wCDDPEsophageal cancerTAOK3/KMT2C/autophagyhttps://doi.org/10.1002/advs.202300864PaclitaxelEsophageal cancerSIRT1/autophagyhttps://doi.org/10.1016/j.lfs.2024.122530CDDPOSCCHIST3H2A and HIST3H2B/chromatin remodeling/autophagyhttps://doi.org/10.1038/s41419-024-06975-1CDDPOSCCcircAP1M2/miR-1249-3p/ATG9A/autophagyhttps://doi.org/10.1002/jcp.31116ApalutamideProstate cancerATF3/CAPN2/FOXO1/ATG5/autophagyhttps://doi.org/10.1186/s12967-024-05335-zDocetaxelProstate cancerPrLZ/LKB1/AMPK/autophagyhttps://doi.org/10.7150/thno.20356CDDPBladder cancerCAB39/LKB1/AMPK/LC3/autophagyhttps://doi.org/10.1016/j.freeradbiomed.2023.09.017GemcitabinePDACH3K27 acetylation/CSNK2A1/autophagyhttps://doi.org/10.1016/j.canlet.2024.216640anti-PD1/anti-CTLA4PDACNBR1/autophagy/MHC-Ihttps://doi.org/10.1038/s41586-020-2229-5TemozolomideGBMCANX/MEK/ERK/BNIP3/mitophagyhttps://doi.org/10.7150/thno.105591Anti-PD-L1GBMOMA1/HSPA9/mitophagy/cGAS/PD-L1https://doi.org/10.1136/jitc-2023-008718TemozolomideGliomaZDHHCs/PI3K/AKT/autophagyhttps://doi.org/10.1016/j.omtn.2022.04.030SunitinibRCCIGFL2-AS1/hnRNPC/TP53INP2/autophagyhttps://doi.org/10.1158/0008-5472.CAN-21-3432SunitinibRCCSNHG1/PTBP1/ATG7/autophagyhttps://doi.org/10.1038/s41420-024-02021-3TemsirolimusRCCCDK7https://doi.org/10.1016/j.canlet.2019.12.005CDDP and PARP inhibitorsOvarian cancerlncRNA LOC730101/BECN1/VPS34/autophagyhttps://doi.org/10.1038/s41419-024-07278-1CDDPOvarian cancerBRCA1/autophagyhttps://doi.org/10.1002/cam4.1975-FUMelanomaNOD2/TYMS/PLK1/autophagyhttps://doi.org/10.1038/s41419-024-07104-8Anti-PD-1MelanomaULK1/autophagyhttps://doi.org/10.1158/1541-7786.MCR-22-0684BRAFi and MEKiMelanomaSEC61/ERK/ATF4/autophagyhttps://doi.org/10.1158/2159-8290.CD-18-0348CDDPOsteosarcomaSestrin2 PERK/eIF2α/CHOP/mTOR/autophagyhttps://doi.org/10.3389/fcell.2021.722960MethotrexateOsteosarcomaFOXM1/HMMR/ATG7/autophagyhttps://doi.org/10.3724/abbs.2024084AdriamycinAMLCELF1/ATG5/autophagyhttps://doi.org/10.1038/s41419-025-07926-0CytarabineAMLSDF-1α/CXCR4/SIRT1/autophagyhttps://doi.org/10.1016/j.canlet.2018.03.024FLT3 kinase inhibitorsAMLAKT/mTORC1/ULK1/autophagyhttps://doi.org/10.1038/s41375-022-01678-yTyrosine kinase inhibitorCMLULK1/autophagyhttps://doi.org/10.1126/scitranslmed.abd5016Tyrosine kinase inhibitorCMLCirc_0009910/miR-34a-5p/ULK1/autophagyhttps://doi.org/10.1016/j.lfs.2020.117255BortezomibMultiple myelomaENO1/YWHAZ/mitophagyhttps://doi.org/10.1186/s12929-024-01101-xBortezomibMultiple myelomaUSP12/HMGB1/autophagyhttps://doi.org/10.1038/s41388-021-02167-9 Summary of autophagy mediated drug resistance mechanisms in cancer therapy. Autophagy-mediated resistance often begins with its role as a context-dependent redox-metabolic gatekeeper. By sensing oxidative stress through AMPK-ULK1 and p62/Keap1/Nrf2 signaling, it selectively clears damaged mitochondria, fine-tunes ROS flux, and stabilizes bioenergetic homeostasis to raise the threshold for therapy-induced cell death [145]. This adaptive circuitry enables malignant cells to survive otherwise lethal oxidative and metabolic insults, as demonstrated in various cancer models where autophagy neutralizes ROS bursts from agents like carnosol and artemisinin, directly fueling treatment resistance. Pharmacological or genetic (ATG5/7 knockdown) disruption consistently reverses this tolerance and restores ROS-driven apoptosis, confirming a causal rather than correlative role [146]. However, therapeutic exploitation of this gatekeeper mechanism requires precision, combining autophagy inhibitors with ROS-inducing therapies only in contexts where tumors have become autophagy-dependent, while accounting for tumor-type heterogeneity, dynamic redox thresholds, and the narrow therapeutic window of lysosomotropic agents. Evidence from multiple studies supports the role of autophagy in drug resistance across different tumor types. In CRC, IL-6-enhanced autophagic activity has been linked to resistance against chemotherapy. Mechanistically, IL-6 induces JAK2–BECN1 interaction, leading to Y333 phosphorylation-mediated activation of BECN1. More important, BECN1 Y333 phosphorylation can be a predictive marker for poor cancer prognosis and chemotherapy resistance [147]. In renal cell carcinoma (RCC), IGFL2-AS1 drives sunitinib resistance via TP53INP2-triggered autophagy [148]. Similarly, in glioblastoma, the CANX/MAPK/BNIP3 mitophagy pathway protects against temozolomide-induced cell death [149]. In immunotherapies, autophagy can attenuate anti-tumor responses by modulating antigen presentation and regulating inflammatory cell death pathways, thereby weakening immune-mediated cytotoxicity [150]. For example, NDP52 serves as a critical autophagy receptor that drives MHC-I degradation, which in turn compromises the antitumor effect of PD-1 inhibitors [151]. This convergence across chemo-, targeted, and immune therapies highlights that autophagy-mediated resistance is not pathway-specific but lies at the core of a broad-spectrum survival network centered on metabolic and redox homeostasis. Such cross-talk reflects a learned metabolic escape strategy, in which autophagy converts the intended disruption of one metabolic node into fuel for alternative pathways, thereby undermining the rationale of single-target inhibitors. In summary, autophagy-mediated drug resistance reflects a multifaceted survival strategy in cancer. Through its metabolic support, mitochondrial protection, and suppression of cell death pathways—including the FOXO3a–PUMA apoptosis axis and p62–Keap1–NRF2 metabolic escape pathway—autophagy reduces the efficacy of diverse anticancer treatments. These mechanisms highlight the need for therapeutic approaches that consider autophagy not only as a survival pathway but also as a key contributor to resistance, making its modulation a promising avenue for improving treatment outcomes. Metabolic alterations driving resistance to chemotherapy and targeted therapy Metabolic reprogramming is a defining feature of cancer cells and plays a decisive role in shaping their response to therapy. While chemotherapy and targeted drugs are designed to disrupt cell growth and induce death, many tumors adapt by rewiring their metabolic pathways. These alterations maintain energy production and biosynthesis, contributing to therapeutic resistance (Table 2). Table 2Key metabolic molecules and pathways driving resistance to chemotherapy and targeted therapy in cancer.Table 2Drug/TreatmentMolecular mechanism of resistanceKey metabolic cuesDOICDDPLactate/NBS1 K388 lactylation/homologous recombination -mediated DNA repair.Lactatehttps://doi.org/10.1038/s41586-024-07620-9NiraparibLactate/histone H4K12la/RAD23ALactatehttps://doi.org/10.1186/s12943-025-02295-wDocetaxelEIF4A3/circARHGAP29/IGF2BP2/LDHA transcription.LDH-Ahttps://doi.org/10.1158/0008-5472.CAN-21-2988LenvatinibACYP1/HSP90/MYC/LDHALDH-Ahttps://doi.org/10.1016/j.drup.2023.1009765-FUMETTL3/HIF-1α/LDHALDH-Ahttps://doi.org/10.7150/thno.73746ICISRSF10/MYB/GLUT1, HK1, LDHA/Lactate/H3K18la/SRSF10GLUT1, HK1, LDH-Ahttps://doi.org/10.1002/cac2.12607GemcitabineZDHHC9/LDHA palmitoylationLDH-Ahttps://doi.org/10.1016/j.canlet.2024.216696PemetrexedAKR1B10/LDHA/lactate/histone H4K12la/CCNB1LDH-Ahttps://doi.org/10.1186/s12967-023-04403-0ICISAA/LDHA/STAT3/PD-L1LDH-Ahttps://doi.org/10.1038/s41467-024-46118-wGemcitabineGRP78/HIF-1α/LDHALDH-Ahttps://doi.org/10.1158/1078-0432.CCR-23-1143EGFR-TKIEGFR C797S mutation/AKT/HIF-1α/PDK1PDK1https://doi.org/10.1038/s12276-024-01221-2GemcitabineUPP1/PDK1/2/AKTPDK1/2https://doi.org/10.7150/ijbs.83774PI3K inhibitorMAPK4/PDK1/AKTPDK1https://doi.org/10.1371/journal.pbio.3002227CDDPCIgG/PDK1/AKTPDK1https://doi.org/10.1111/odi.13583ICIFTO/PDK1/AKT/STAT3/PD-L1PDK1https://doi.org/10.1016/j.jare.2024.12.0265-FUGLIPR1/PI3K/PDK1/ROCK1PDK1https://doi.org/10.1038/s41417-022-00490-1GemcitabineSTAMBP/E2F1/PDK1PDK1https://doi.org/10.1038/s41419-024-07048-zPaclitaxelPRMT6/G6PD/GSTP1G6PDhttps://doi.org/10.1016/j.bcp.2020.114092PI3K inhibitorPI3Kγ/AKT/NRF2/G6PD/PPPG6PDhttps://doi.org/10.1182/blood.2023022202CDDPPRMT6 methylated R324 of G6PD to enhancing its activity.G6PDhttps://doi.org/10.1016/j.apsb.2022.05.019OxaliplatinMETTL3/TRIM21/G6PD axis in driving oxaliplatin resistanceG6PDhttps://doi.org/10.1038/s41418-024-01406-2DoxorubicinG6PD/DOX-induced oxidative stress resistanceG6PDhttps://doi.org/10.7150/ijbs.65555TrastuzumabBMAL1/CLOCK/PER1/HK2 glycolysis/trastuzumab resistance.HK2https://doi.org/10.1158/0008-5472.CAN-21-1820TrametinibBCKDK/BCLAF1/MYC/HK2/lactate/ABC transporters/Trametinib resistanceHK2https://doi.org/10.1038/s41418-025-01531-6CDDPATF3/Nrf2/Keap1/xCT/ferroptosisSLC7A11https://doi.org/10.1186/s11658-021-00271-yEnzalutamideAndrogen receptor/SLC7A11/ferroptosis resistanceSLC7A11https://doi.org/10.1158/0008-5472.CAN-23-0285SorafenibYAP/TAZ/ATF4/SLC7A11/ferroptosis resistanceSLC7A11https://doi.org/10.15252/emmm.202114351CDDPPFKFB3/SLC7A11/dephosphorylation/ferroptosis resistanceSLC7A11https://doi.org/10.1016/j.intimp.2025.114914CDDPlow SLC7A11-AS1/p38MAPK/ATP7A/GSHSLC7A11https://doi.org/10.1016/j.freeradbiomed.2022.03.02CDDPGSH/ferroptosis resistanceGSHhttps://doi.org/10.1016/j.redox.2021.101977TemozolomideNRF2/ABCC1/GSH/ferroptosis resistanceGSHhttps://doi.org/10.1038/s41419-022-05044-9DocetaxelDecreasing the level of FTH1P8/ferroptosisGSHhttps://doi.org/10.1016/j.biopha.2024.117472EnzalutamidePHGDH/ferroptosis resistancePHGDHhttps://doi.org/10.1016/j.drup.2023.100985TemozolomideLRSAM1 ubiquitinates and degrades SLC40A1/ferroptosisSLC40A1https://doi.org/10.1038/s41419-024-06902-4Sorafenibloss of SLC27A5/NRF2/GSR/ferroptosis resistanceGSRhttps://doi.org/10.1038/s41419-023-05558-wDoxorubicinZFP64 histone lactylation/GCH1/lipid peroxidation inhibition/ferroptosis resistanceGCH1https://doi.org/10.1186/s12967-025-06246-3EGFR-TKIIGF2BP3/COX6B2/nicotinamide metabolism/OXPHOS/EGFR-TKI acquired resistanceCOX6B2https://doi.org/10.1158/0008-5472.CAN-22-3059IbrutinibTCF4/EGR1/PDP1/OXPHOSPDP1https://doi.org/10.1182/blood.2023020142ChemotherapyZDHHC21/AK2/OXPHOSAK2https://doi.org/10.1182/blood.2022019056AbemaciclibAbemaciclib-resistant HR-positive, HER2-negative metastatic breast cancer cell line upregulated mediators of the OXPHOS and responded to OXPHOS inhibitors.OXPHOShttps://doi.org/10.1158/0008-5472.CAN-23-0705PalbociclibTreatment by the OXPHOS inhibitor IACS-010759 strongly inhibits tumour growth in multiple endocrine and palbociclib resistant PDXOXPHOShttps://doi.org/10.1038/s41467-023-40022-5ICIHypoxic TME inhibits T-cell function through OXPHOS, thereby leading to immunotherapy resistance.OXPHOShttps://doi.org/10.1136/jitc-2023-008226ICIHigher levels of OXPHOS in tumor- and peripheral blood-derived CD8+ T cells correlated with ICI resistance in melanoma patientsOXPHOShttps://doi.org/10.1084/jem.20202084EGFR-TKISmall molecule compound OPB inhibits OXPHOS resulting in decrease in metabolic activity and reduction in tumor size.OXPHOShttps://doi.org/10.1016/j.redox.2018.1010765-FUOXPHOS blockade by metformin enhances 5-FU chemosensitivity upon NIT2 lossOXPHOShttps://doi.org/10.1126/scitranslmed.ado8333DoxorubicinC210 disrupted TRAP1's interaction with OXPHOS-associated client proteins. C210, overcomes DOX resistance by targeting TRAP1.OXPHOShttps://doi.org/10.1016/j.drup.2025.101226TrastuzumabCDC42/NF-κB/GLS1 Tumor-derived GLS1 microvesicles drove M2 macrophage polarization.GLS1https://doi.org/10.1002/cac2.124595-FUTargeting both mTORC2-ccGAS and glutaminase provides a promising strategy to eliminate quiescent resistant cancer cells.GLS1https://doi.org/10.1038/s41556-024-01473-0CDDPNRF2/GLS1/ferroptosis resistanceGLS1https://doi.org/10.1158/1078-0432.CCR-22-2747DocetaxelIL-1β/SOD2/ROS/ICAM1/docetaxel resistantSOD2https://doi.org/10.1172/jci.insight.157285EGFR-TKITargeting NRF2/GPX4/SOD2 pathway is a potential strategy for overcoming resistance to EGFR-TKIs.SOD2https://doi.org/10.1038/s41401-020-0443-1ChemotherapyHIF-2α/SOD2/mtROS/PDI/GRP78 UPRER/chemoresistanceSOD2https://doi.org/10.1038/s41418-022-00963-8Chemotherapyp300/H4K12la/GCLC/ferroptosis resistanceGCLChttps://doi.org/10.1038/s41419-025-07498-zChemotherapyLIN28B/GSH metabolic enzymes/ROSGCLC, G6PDhttps://doi.org/10.1016/j.canlet.2025.217572SorafenibFASN/HIF1α/SLC7A11/Sorafenib resistanceFASNhttps://doi.org/10.1186/s13046-022-02567-zICIRiplet/K48-linked polyubiquitination of FASN/free fatty acids/STAT3/CD8 T cell exhaustionFASNhttps://doi.org/10.1126/sciimmunol.ado3485ChemotherapyPRMT1/PHGDH methylation/FASN/palmitate/lipogenesis/chemoresistanceFASN, PHGDHhttps://doi.org/10.1158/0008-5472.CAN-23-2266OxaliplatinPOU2F1/ALDOA/oxaliplatin resistanceALDOAhttps://doi.org/10.1038/s41388-021-02148-y Key metabolic molecules and pathways driving resistance to chemotherapy and targeted therapy in cancer. In breast cancer, upregulation of glycolytic enzymes and glucose transporters confers resistance to cisplatin (CDDP), paclitaxel, tamoxifen, and doxorubicin; targeting glycolysis with inhibitors such as 2-deoxy-d-glucose (2-DG) or metformin can resensitize tumors [152]. This metabolic shift provides rapid ATP production and reductive metabolites. In non-small cell lung cancer (NSCLC), PKM2 enhances glycolysis, which reduces drug-triggered ROS accumulation and protects cancer cells from chemotherapy-mediated toxicity [153]. Alternatively, lactate and other glycolytic metabolites confer drug resistance through PTM mechanisms. Lactylation of NBS1 at lysine 388 promotes homologous recombination-mediated DNA repair and drives chemoresistance [154]. Moreover, the acidification of the TME can impair drug penetration and weaken immune cell activity, further reducing therapeutic efficacy. Increased lactate levels in TME promoted M2 macrophage polarization and inhibiting CD8+ T cell activity [155]. The link between lactate accumulation with the immune suppressive microenvironment robustly propose monocarboxylate transporter 1/4 (MCT1/4) inhibition as a direct therapeutic strategy to sensitize immunotherapy. Mitochondrial metabolism also contributes to resistance. OXPHOS is the primary mode of energy production in many cancers, especially those exposed to chemo- and radiotherapy [156]. Melanoma cells with BRAF mutations evade MAPK inhibitor therapy by enhancing mitochondrial OXPHOS [157]. IGF2BP3 stabilizes COX6B2 mRNA in an m6A-dependent manner, which remodels OXPHOS and drives EGFR-TKI acquired resistance [158]. Moderately elevated mtROS does not trigger oxidative cell death; instead, it promotes chemoresistance by stabilizing HIF and maintaining stemness of cancer stem cells (CSCs) [159,160]. Mechanistically, mitochondrial translational activator of cytochrome oxidase subunit 1 (TACO1) facilitates the mitochondrial cytochrome c oxidase I (MTCO1) translation, which augments OXPHOS and increases mtROS accumulation, ultimately fueling cancer stemness and cisplatin resistance [161]. While CSCs are heavily reliant on OXPHOS, CSCs can dynamically switch between glycolysis and OXPHOS in response to therapy-induced stress. When OXPHOS is inhibited, they often upregulate glycolytic pathways alongside activation of AMPK signaling, which collectively promote survival and resistance to apoptosis. In some cases, cancer cells adopt a hybrid metabolic phenotype, leveraging both glycolysis and OXPHOS concurrently. This hybrid state contributes to resistance against single-pathway targeted therapies and enhances metastatic potential [162]. Amino acid metabolism provides another layer of protection. Glutamine, in particular, serves as a critical fuel for resistant tumor cells. Enhanced glutamine metabolism is a hallmark of Nrf2-driven metabolic reprogramming, thereby inducing gemcitabine resistance in KRAS-mutant pancreatic cancer [163]. Glutamine-based antioxidant support (e.g., NADPH, GSH) emerges as a discrete redox-adaptive strategy that tumors use to neutralize oxidative damage and prevent the activation of ROS-sensitive death pathways. In CDDP-resistant head and neck squamous cell carcinomas (HNSCC), inhibition of GLS1 restores chemosensitivity by impairing proliferation and reducing GSH synthesis [164]. On the other hand, direct crosstalk between glutamine metabolism and oncogenic signaling contributes to drug resistance. The GLS1/c-Myc feedback loop promotes malignant progression and therapeutic resistance in multiple cancers [[165], [166], [167]]. Moreover, several metabolic enzymes are emerging as novel prognostic biomarkers or druggable targets to modulate glutamine metabolism and reverse drug resistance. For example, ILF3 phase separation upregulates GLUD1 expression at the posttranscriptional level, which potentiates mitochondrial glutamine catabolism and ultimately fuels EGFR-TKI resistance in NSCLC [168]. In hepatocellular carcinoma (HCC), OGDHL suppression triggers reductive carboxylation of glutamine-derived α-KG through retrograde TCA cycling. The resultant accumulation of NADPH and GSH sustains cellular antioxidant capacity and compromises the efficacy of sorafenib [169]. Lipid metabolism is also reshaped during resistance. These adaptations serve dual roles—as structural buffers and as regulators of redox-dependent cell death. Tumor hypoxia induced by anti-angiogenic drugs initiates FAO reprogramming and augments free fatty acid (FFA) uptake, facilitating cancer cell proliferation. CPT1 inhibition significantly abrogates FFA-mediated cell proliferation and sensitizes anti-angiogenic therapeutic efficacy [170]. Lipid remodeling not only supplies membranes for rapidly dividing cells but also influences susceptibility to death pathways such as ferroptosis. Elevated fatty acid synthesis and lipid droplet formation protect resistant tumors against lipid peroxidation and oxidative stress [171]. DGAT1 catalyzes the biosynthesis of triglycerides from fatty acids and their sequestration in lipid droplets. Pharmacological inhibition of DGAT1 triggers ferroptosis and enhances the efficacy of immune checkpoint inhibition (ICI) therapy [172]. These metabolic alterations are tightly interwoven with autophagy. By recycling intracellular components, autophagy supports glycolysis, OXPHOS, and amino acid metabolism, reinforcing the survival of resistant cells. For instance, ferritinophagy serves as a crucial adaptive mechanism underlying RAS-mutant cancer cells' resistance to MAPK inhibitors. Ferritinophagy is activated by MAPK inhibition to increase iron availability, which is essential to support the production of iron-sulfur cluster (ISC) proteins and mitochondrial health of cancer cells, leading to therapy resistance [173]. Mitophagy, by maintaining mitochondrial quality, preserves OXPHOS efficiency during therapeutic stress. Simultaneously, mitophagy/ROS negative feedback maintains ROS homeostasis and consumes mitochondrial CPT1 to repress FAO. This drives the sequestration of acyl-CoA within lipid droplets, thereby protecting cells against therapy-induced oxidative damage [174]. This coordination illustrates how metabolic plasticity and autophagy together create a formidable barrier against chemotherapy and targeted therapy [175]. From an integrative perspective, autophagy-metabolism-drug-resistance must be viewed as a dynamic feedback loop, where therapeutic stress repeatedly resets the metabolic and redox landscape to favor survival rather than death. Such integration emphasizes that resistance is not driven by isolated pathways but by a coordinated autophagy-metabolism-redox network, which must be targeted in parallel rather than in isolation. Emerging therapeutic strategies Pharmacological modulators of autophagy and metabolism in cancer Pharmacological interventions targeting autophagy and metabolism are rapidly gaining attention as promising strategies to overcome therapeutic resistance in cancer. By directly modulating these interconnected pathways, such approaches aim to weaken tumor survival mechanisms while enhancing the efficacy of standard treatments. However, while much focus has been placed on autophagy-dependent cell survival and death, it is increasingly evident that integrating knowledge of other regulated cell death modalities—such as apoptosis, ferroptosis, necroptosis, and pyroptosis—is essential to developing effective targeted therapeutics. Understanding how metabolic and autophagic modulation intersects with these death pathways provides a more complete framework for exploiting tumor vulnerabilities. Autophagy modulators are among the most actively studied agents. Chloroquine (CQ) and hydroxychloroquine (HCQ), which block lysosomal degradation, remain the most widely used inhibitors in clinical trials (Table 3). Several clinical studies have tested these agents, especially in combination with chemotherapy or radiotherapy, showing tumor cell sensitization and reduced survival. For example, a phase I trial combining HCQ with palbociclib and letrozole in metastatic breast cancer demonstrated safety and tumor responses linked to autophagy marker modulation [176]. Although their efficacy as monotherapies is limited, their combination with chemotherapy or radiotherapy has shown the ability to reduce tumor survival by preventing nutrient recycling and dampening DNA repair responses [177]. Table 3Recent clinical trials targeting autophagy in cancer.Table 3Intervention/TreatmentNCT NumberPhaseStatusMechanism of ActionYearHCQ/Nelfinavir/BevacizumabNCT06971744Phase 2Not yet recruitingHCQ inhibits autophagy, Nelfinavir inhibits HIV protease, Bevacizumab targets VEGF.2025-2027HCQ + palbociclibNCT05953350Phase 1b/2RecruitingHCQ inhibits autophagy; Palbociclib blocks CDK4/6 to halt cell cycle progression.2023-2024Intra-arterial HCQ in Lipiodol + transarterial embolization followed by oral HCQNCT05842174Phase 1Not yet recruitingHCQ inhibits autophagy while TACE delivers chemotherapy and embolization for liver tumors.2025-2030Phase 2TACE + axitinib + HCQNCT04873895Phase 1bTerminatedTACE combined with axitinib inhibits angiogenesis, HCQ inhibits autophagy.2022-2024Trametinib + HCQNCT04566133Phase 2TerminatedTrametinib inhibits MEK1/2, while HCQ blocks autophagic flux, synergizing anti-cancer effects.2022.5-2022.12HCQ is used one day before chemotherapy and radiotherapy, 400-600 mg, oral tablet, once. During chemotherapy and radiotherapy, HCQ maintenance dose is 200-400 mg daily.NCT06389201N/ARecruitingHCQ inhibits autophagy, potentiating chemo- and radiotherapy efficacy.2024-2034Ulixertinib + HCQNCT05221320Phase 2TerminatedUlixertinib inhibits EGFR/ERK signaling, while HCQ inhibits autophagy to enhance response.2024.5-2024.7Abemaciclib + HCQNCT04523857Phase 2RecruitingAbemaciclib inhibits CDK4/6, and HCQ inhibits autophagy to promote cancer cell death.2021-2028HCQ + Sacituzumab GovitecanNCT06328387Phase 1RecruitingHCQ inhibits autophagy, while Sacituzumab Govitecan targets Trop-2 for targeted therapy.2024-2026HCQ + Trastuzumab DeruxtecanPhase 2DigoxinNCT05507775N/ACompletedInhibits Na+/K + -ATPase to disrupt tumor cell signaling and induce autophagic stress.2022-2024Ezurpimtrostat + Atezolizumab-BevacizumabNCT05448677Phase 2TerminatedEzurpimtrostat targets autophagy, Atezolizumab blocks PD-L1, and Bevacizumab inhibits VEGF.2022-2024Temozolomide+HaloperidolNCT06218524Phase 2Not yet recruitingTemozolomide induces DNA damage, Haloperidol enhances autophagy inhibition for synergy.2024-20281HCQ, Metformin, Sirolimus2HCQ, Metformin, Sirolimus, Dasatanib3HCQ, Metformin, Sirolimus, Nelfinavir4HCQ, Metformin, Sirolimus, Nelfinavir, Dasatinib5HCQ, Metformin, Sirolimus, Nelfinavir6HCQ, Metformin, Sirolimus, Nelfinavir, DasatinibNCT05036226Phase 1RecruitingN/A2022-2026Phase 2HCQ Sulfate/LY3214996NCT04386057Phase 2CompletedHCQ inhibits autophagy, LY3214996 targets MET for tumor progression inhibition.2020-2024Trametinib + HCQNCT03979651Not ApplicableCompletedTrametinib inhibits MEK1/2, HCQ blocks autophagic flux, enhancing anti-cancer effects.2019-2023HCQ SulfateNCT06408298Phase 1Not yet recruitingInhibits autophagy, potentiating cancer cell death and enhancing therapy.2025-2028Atezolizumab + HCQ + CobimetinibNCT04214418Phase 1TerminatedAtezolizumab blocks PD-L1, Cobimetinib inhibits MEK1/2, HCQ targets autophagic flux.2020-2022Phase 2HCQ + Trametinib + Dabrafenib mesylateNCT04527549Phase 2TerminatedHCQ inhibits autophagy, Trametinib and Dabrafenib target MAPK/ERK pathways for synergy.2021-20241HCQ + Palbociclib2Avelumab + Palbociclib3Avelumab4HCQNCT04841148Phase 2RecruitingN/A2021-2028HCQ + Carfilzomib + DexamethasoneNCT04163107Phase 1CompletedHCQ inhibits autophagy; Carfilzomib inhibits proteasome; Dexamethasone modulates inflammation.2020-20211Dabrafenib + Trametinib + HCQ2Trametinib + HCQNCT04201457Phase 1Active, not recruitingDabrafenib and Trametinib inhibit BRAF and MEK1/2, while HCQ inhibits autophagy.2020-2030Phase 2Gemcitabine + HCQ + Nab-paclitaxel + ParicalcitolNCT04524702Phase 2Active, not recruitingCombines chemotherapy with autophagy inhibition and vitamin D modulation for synergy.2020-2024HCQNCT04911816Phase 1RecruitingInhibits autophagy, enhancing the effects of chemotherapy and radiotherapy.2021-2030Phase 2 Recent clinical trials targeting autophagy in cancer. 1HCQ, Metformin, Sirolimus HCQ, Metformin, Sirolimus 2HCQ, Metformin, Sirolimus, Dasatanib HCQ, Metformin, Sirolimus, Dasatanib 3HCQ, Metformin, Sirolimus, Nelfinavir HCQ, Metformin, Sirolimus, Nelfinavir 4HCQ, Metformin, Sirolimus, Nelfinavir, Dasatinib HCQ, Metformin, Sirolimus, Nelfinavir, Dasatinib 5HCQ, Metformin, Sirolimus, Nelfinavir HCQ, Metformin, Sirolimus, Nelfinavir 6HCQ, Metformin, Sirolimus, Nelfinavir, Dasatinib HCQ, Metformin, Sirolimus, Nelfinavir, Dasatinib 1HCQ + Palbociclib HCQ + Palbociclib 2Avelumab + Palbociclib Avelumab + Palbociclib 3Avelumab Avelumab 4HCQ HCQ 1Dabrafenib + Trametinib + HCQ Dabrafenib + Trametinib + HCQ 2Trametinib + HCQ Trametinib + HCQ Beyond these classical agents, more selective inhibitors are under development. Compounds targeting upstream autophagy initiation steps, such as ULK1 (e.g., SBI-0206965, ULK101, DCC-3116), VPS34 (e.g., SAR405), and ATG4 (e.g., S130) provide more precise control over autophagy while limiting off-target effects [178]. Surprisingly, some promising targeted agents, such as SHP2 allosteric inhibitors IACS-13909, show the unexpected off-target activity of blocking autophagic flux in an SHP2-independent manner [179]. Although most of these inhibitors remain in preclinical or early clinical testing, recent studies have revealed that several United States Food and Drug Administration (FDA)-approved agents exhibit selective and robust autophagy-inhibiting effects; the representative drugs include eltrombopag (EO) and apilimod. EO specifically suppresses activity of TFEB—a central transcriptional regulator of autophagy—at the genomic scale, decreasing autophagy in a dose-dependent manner [180]. Phosphatidylinositol-3-phosphate 5-kinase (PIKfyve) inactivation by apilimod induces lysosomal dysfunction and attenuated autophagic flux with much greater potency than CQ treatment in PDAC [181]. Interestingly, such modulation can trigger secondary cell death responses, including apoptosis or ferroptosis, depending on the metabolic state of the cell and redox balance, indicating that autophagy inhibition may function as a sensitizer rather than a direct death inducer [182,183]. This nuanced behavior highlights that the therapeutic value of autophagy modulators may lie less in global suppression and more in shifting the redox and metabolic tipping point toward lethal cell-death programs. On the metabolic front, inhibitors of glycolysis, glutamine metabolism, and mitochondrial respiration are emerging as valuable tools. Recently, several clinical trials were conducted to assess its efficacy (Table 4). Agents such as 2-DG and lonidamine impair glycolytic flux, depriving tumor cells of rapid ATP production and biosynthetic intermediates. Preclinical studies show that 2-DG and lonidamine enhance the cytotoxic effects of treatments such as photodynamic therapy in breast cancer cell lines [184,185]. Furthermore, GLS inhibitors like CB-839 limit glutamine utilization, weakening the redox buffering capacity of resistant tumors and reducing nucleotide synthesis. CB-839 has demonstrated activity in several hematologic and solid tumors, with enhanced efficacy in melanoma models when combined with immunotherapies [186,187]. Mitochondrial complex I inhibitors, including metformin and IACS-010759, disrupt OXPHOS, creating an energy deficit that can sensitize tumors to cytotoxic therapies. Metformin has been studied in diverse cancers, while IACS-010759 is under evaluation in clinical trials, with preclinical data showing robust anti-tumor effects [188]. For example, IACS-010759 induces ROS-dependent tumor cell death through pathways involving protein phosphatase 2A activation and CMA [189]. Other complex I inhibitors like efavirenz cause peroxynitrite formation and mitochondrial protein oxidation leading to cell injury. In addition to these agents, phenformin, a more potent biguanide analog of metformin has shown strong preclinical activity in targeting metabolic vulnerabilities of NSCLC [190]. Mechanistically, phenformin suppresses OXPHOS but with greater potency than metformin, leading to pronounced ATP depletion and ROS accumulation. Innovative formulations, such as phenformin-loaded micelles, further enhance delivery to tumor cells and CSCs, improving efficacy while reducing systemic toxicity. Although clinical trials with metformin added to standard therapy have not improved survival in unresectable stage III NSCLC, ongoing efforts to clinically evaluate phenformin could better define its role in exploiting mitochondrial dysfunction in refractory lung cancers [191,192]. A critical takeaway is that these metabolic agents do not simply kill tumor cells by energy depletion but by reconfiguring the mitochondrial–redox landscape to push cells across the threshold from adaptive stress into lethal oxidative collapse. Table 4Recent clinical trials targeting metabolic pathways in cancer.Table 4Intervention/TreatmentNCT NumberPhaseStatusMechanism of ActionYearDapagliflozin + Standard of Care ChemotherapyNCT05521984Phase 1RecruitingSGLT2 inhibitor that reduces glucose reabsorption, enhancing chemotherapy efficacy.2023-2026Ketoconazole and/or PosaconazoleNCT03763396Early Phase 1Not yet recruitingInhibits hexokinase 2 (HK2) to reduce glucose metabolism in cancer cells.2024-2027KetoconazoleNCT04869449Early Phase 1N/AInhibits hexokinase 2 (HK2), blocking glucose metabolism in cancer cells.2022Arginine-restricted diet + conventional anti-tumor treatmentNCT07038044Phase 1Not yet recruitingRestricts arginine to starve tumor cells and enhance conventional therapies.2025-2027DapagliflozinNCT04542291Phase 1CompletedInhibits SGLT2, reducing glucose reabsorption and altering tumor metabolism.2021-2022DRP-104+ DurvalumabNCT06027086Phase 1RecruitingGlutamine antagonist that reduces glutamine-dependent tumor growth.2024-2033Phase 2DRP-104+ atezolizumabNCT04471415Phase 1TerminatedGlutamine antagonist that inhibits tumor growth by targeting glutamine metabolism.2020-2023Phase 2Gemcitabine + L-glutamine + Nab-paclitaxelNCT04634539Phase 1CompletedCombines chemotherapy with glutamine supplementation to enhance tumor cytotoxicity.2021-202510 g glutamine+5 g Maltodextrin for 28 days after surgery with tolerable oral intake or enteral feedingNCT06027242N/ARecruitingSupports post-surgery recovery with glutamine and maltodextrin for metabolic balance.2023-2026Sapanisertib + CB-839NCT04250545Phase 1Active, not recruitingAnti-GLS1 inhibitor targeting glutaminolysis in tumor cells.2020-2026JZP341NCT05631327Phase 1TerminatedLong-acting asparaginase to deplete asparagine and inhibit tumor growth.2022-2024CB-839NCT03944902Phase 1TerminatedAnti-GLS1 inhibitor targeting glutamine metabolism to impair tumor cell growth.2021-2022CB-839+TalazoparibNCT04824937Phase 2Unknown statusInhibits GLS1 to target glutamine metabolism and enhance DNA damage response.2021.7-2021.121IACS-62742bevacizumab + paclitaxel + IACS-6274NCT05039801Phase 1RecruitingInhibits GLS1 to block glutamine metabolism and enhance chemotherapy efficacy.2021-20261Anti-PD-1+ Metformin2Anti-PD-1+ Rosiglitazone3Anti-PD-1NCT04114136Phase 2RecruitingMetformin inhibits mitochondrial complex I, enhancing immune response in cancer.2020-2032Metformin + Radiation IMRT + TemozolomideNCT04945148Phase 2RecruitingMetformin is an oral inhibitor of mitochondrial complex I2024-2028GamitrinibNCT04827810Phase 1Recruitingmitochondrial-targeted Hsp90 inhibitor2021-2026PioglitazoneNCT05013255Phase 2RecruitingPPARγ agonist that alters lipid and glucose metabolism to affect cancer cell growth.2021-2026CPI-613NCT05325281Phase 1Active, not recruitingCPI-613 is a mitochondrial metabolism inhibitor2022-2025CPI-613 + HydroxychloroquineNCT04593758Phase 1CompletedTargets mitochondrial metabolism and autophagy for synergistic anti-cancer effect.2021-2023Phase 2CPI-613+ gemcitabine + CDDPNCT04203160Phase 1CompletedCombines mitochondrial metabolism inhibition with chemotherapy for enhanced efficacy.2020-2024Phase 2CPI-613+ BendamustineNCT04217317Phase 2TerminatedDisrupts mitochondrial function and enhances chemotherapy-induced cancer cell death.2020-2024CPI-613+5-FU + HCQ + GemcitabineNCT05733000Phase 2RecruitingSynergistic inhibition of mitochondrial metabolism and autophagy with chemotherapy.2023-2030IACS-010759NCT03291938Phase 1CompletedOXPHOS inhibitor2017-2020IACS-010759NCT02882321Phase 1TerminatedOXPHOS inhibitor2016-2022Gemcitabine + Nab paclitaxel + IM156NCT05497778Phase 1RecruitingLixumistat (IM156) is a novel biguanide mitochondrial protein complex 1 inhibitor ofOXPHOS (OXPHOS) with anti-tumor activity.2022-2027IM156NCT03272256Phase 1CompletedInhibits mitochondrial OXPHOS to reduce tumor cell energy production.2017-2020SulfasalazineNCT06134388Phase 3RecruitingIntact SSZ, but not its metabolites 5-aminosalicylic acid and sulfapyridine, competitively inhibits xCT.2023-2026SulfasalazineNCT05580861Phase 1RecruitingInhibits xCT, disrupting cysteine/glutamate exchange and GSH metabolism in tumors.2023-2026Phase 2Gabapentin + Sulfasalazine+Memantine + Temozolomide + RadiotherapyNCT05664464Phase 1RecruitingCombination therapy targeting metabolic and cell survival pathways.2023-2026Phase 2LeflunomideNCT06540937Phase 2RecruitingInhibits pyrimidine synthesis, reducing cancer cell proliferation.2020-2024Leflunomide + DecitabineNCT06923488Phase 1RecruitingCombines pyrimidine synthesis inhibition with epigenetic modulation to target cancer cells.2025-2028Phase 2Leflunomide + Cholestyramine + Steroid TherapyNCT05443425Phase 1RecruitingModulates immune response and lipid metabolism in cancer cells.2023-2027AZD3965NCT01791595Phase 1CompletedMCT1 inhibitor that disrupts lactate transport and tumor cell metabolism.2013-2020TVB-2640+ EnzalutamideNCT05743621Phase 1RecruitingInhibits FASN to block lipid synthesis in tumor cells.2023-2026TVB-2640+ BevacizumabNCT05118776Phase 3Active, not recruitingInhibits FASN to block lipid synthesis in tumor cells.2022-2025BMS-986205+ Nivolumab + TemozolomideNCT04047706Phase 1Active, not recruitingInhibits IDO1 to alter tumor metabolism and enhance immune checkpoint therapy.2019-2027Epacadostat + PembrolizumabNCT03322540Phase 2CompletedInhibits IDO1 to restore T cell function and enhance immune response.2017-2020IO102-IO103NCT05280314Phase 2RecruitingIO102-IO103 is a combination of an indoleamine 2,3-dioxygenase 1 (IDO1) peptide (IO102) and a programmed death-ligand 1 (PD-L1) peptide (IO103), emulsified with an adjuvant (Montanide ISA 51 VG).2023-2027Epacadostat + radiation + CAPOX chemotherapyNCT03516708Phase 1RecruitingCombines IDO1 inhibition with radiation and chemotherapy to enhance tumor control.2020-2030Phase 2SM-88NCT04229004Phase 3CompletedTargets metabolic pathways to disrupt cancer cell growth and survival.2020-2025 Recent clinical trials targeting metabolic pathways in cancer. 1IACS-6274 IACS-6274 2bevacizumab + paclitaxel + IACS-6274 bevacizumab + paclitaxel + IACS-6274 1Anti-PD-1+ Metformin Anti-PD-1+ Metformin 2Anti-PD-1+ Rosiglitazone Anti-PD-1+ Rosiglitazone 3Anti-PD-1 Anti-PD-1 Beyond autophagy-dependent mechanisms, attention has turned to regulated non-autophagic forms of cell death such as ferroptosis, pyroptosis, necroptosis, and apoptosis evoked through metabolic stress (Table 5). Drugs that deplete GSH or inhibit GPX4 enhance lipid peroxidation, pushing tumors toward ferroptotic death. However, autophagy-mediated ferroptosis escape is a known resistance mechanism, suggesting that blocking autophagy could theoretically enhance ferroptosis induced by GPX4 inhibitors [193]. Targeting ferroptosis alongside autophagy inhibition prevents tumor defenses, thereby amplifying cytotoxicity in therapy-resistant cancers [194]. Similarly, necroptosis can arise in contexts where apoptosis is suppressed. Metabolic stress induced by OXPHOS inhibition or glycolytic blockade can promote necroptotic signaling through mtROS accumulation. For example, inhibition of mitochondrial complex I in melanoma cells leads to mitochondrial depolarization, increased ROS, and necroptosis involving RIPK1 and MLKL, with necroptosis inhibitors reducing cell death [195]. Additionally, certain therapies trigger pyroptosis, particularly following mitochondrial dysfunction and ROS generation. Molecular targeted therapies against KRAS-, EGFR-, or ALK-driven lung cancer activate caspase-3, which cleaves GSDME, inducing pyroptosis alongside apoptosis [196]. The ferroptosis inducer RSL3 also triggers cleavage of GSDMD/E in multiple cancer cell lines, with involvement of mitochondrial pathways suggested but requiring further study [197]. Recent findings demonstrate a antagonistic relationship between ferroptosis and pyroptosis, with 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR) playing a critical role in this interaction [198], highlighting the crosstalk between metabolism, mitochondrial status, and regulated cell death modalities. Targeted metabolic disruption can therefore regulate which form of regulated cell death predominates—apoptosis, necroptosis, pyroptosis, or ferroptosis—depending on the tumor genotype and microenvironment [8]. These findings support the concept that metabolic and mitochondrial states modulate the balance among different forms of regulated cell death in cancer therapy. Table 5Recent clinical trials targeting cell death pathways in cancer.Table 5Intervention/TreatmentNCT NumberPhaseStatusMechanism of ActionYearTQB3107NCT06413953Phase 1RecruitingInduces apoptosis through activation of intrinsic and extrinsic pathways.2024-2026Carbon Nanoparticle-Loaded Iron [CNSI–Fe(II)]NCT06048367Phase 1CompletedInduces ferroptosis via lipid peroxidation and iron overload.2022-2025Temozolomide + HaloperidolNCT06218524Phase 2Not yet recruitingEnhances ferroptosis by modulating oxidative stress and mitochondrial dysfunction.2024-2028BGB-24714NCT05381909Phase 1Active, not recruitingInhibits apoptosis pathways by modulating tumor cell survival signaling.2022-2026Papaverine + Radiation TherapyNCT06834126Phase 1RecruitingPPV is an enzyme inhibitor, and it may stop the growth of tumor cells by blocking some of the enzymes needed for cell growth.2025-2028Papaverine + ChemoradiationNCT05136846Phase 1RecruitingPapaverine targets mitochondrial metabolism to decrease the cancer growth process2021-2026ONC206NCT04732065Phase 1RecruitingDRD2 antagonist and ClpP agonist, promoting cell death through mitochondrial stress.2021-2027PIN + anti-PD1NCT06883149Phase 1RecruitingInduces pyroptosis by activating inflammasome and caspase-1 signaling.2025-2030 Recent clinical trials targeting cell death pathways in cancer. In summary, pharmacological modulators of autophagy and metabolism represent a dynamic and evolving frontier in cancer therapy. Extensive preclinical evidence supports their potential, particularly in rational combination regimens. By targeting the adaptive machinery that sustains tumor survival, these agents hold the potential to break therapeutic resistance. However, it should be noted that most findings are derived from in vitro or mouse models, which generally fail to replicate the metabolic intricacy of the TME. More realistic models, such as patient-derived organoids (PDOs) and 3D systems, should be used to validate therapies targeting the metabolic-autophagic axis. Combination therapies targeting metabolic and autophagic pathways Given the adaptability of tumor cells, single-agent therapies often fall short in achieving durable responses. Cancer cells can compensate for the inhibition of one survival pathway by activating others, particularly within the tightly linked networks of metabolism and autophagy. This has led to a growing interest in therapy paradigms that combine chemo-, radio-, and targeted therapies alongside disrupting the autophagy–metabolism circuit, aiming to limit tumor plasticity and overcome resistance. Emerging studies reveal strong autophagy induction following compromising glycolysis or mitochondrial ETC in multiple tumor models [199,200]; in parallel, autophagy-deficient tumor cells can be more vulnerable to metabolic drugs, which deprive tumor cells of essential nutrients and energy, while autophagy inhibition prevents the recycling of intracellular substrates that would otherwise offset these shortages. For example, pharmacological or genetic inhibition of autophagy sensitizes esophageal squamous cell carcinoma (ESCC) cells to metformin-induced apoptotic cell death [201]. In addition, co-targeting glycolysis and autophagy with 2-DG plus CQ in PTEN/TP53-loss prostate cancer drives robust apoptosis and tumor regression in mouse models [202]. 2-DG induces energy stress and CQ blocks autophagic recycling pathways, amplifying cytotoxic effects. Moreover, dual inhibition of GLS and autophagy has demonstrated synergistic effects by limiting external nutrient use while preventing intracellular recycling in ovarian cancer models [203]. In TNBC models, the metabolic inhibitor BMS-754807 synergized with carboplatin to increase cytotoxicity, and the addition of HCQ further strengthens the tumor reduction. [204]. Furthermore, perturbation of FAO boosts starvation-induced Atg7-deleted cell death, suggesting that a combination of autophagy inhibition and lipid metabolism reprogramming could be a promising therapeutic strategy for LKB1-deficient lung cancer [205]. Likewise, small molecule MK8722 suppresses late-stage autophagy through disturbing FASN-regulated rewiring of lipid metabolism, leading to the death of epithelial ovarian cancer cells [206]. These results highlight the potential for dual targeting of metabolic and autophagic programs in cancer treatment. A key conceptual advance here is that the synergy comes less from individual drug potency and more from coordinated disruption of interconnected metabolic and autophagic survival nodes. Targeted therapies—agents designed to inhibit key oncogenic drivers—have demonstrated substantial efficacy in certain cancers, yet resistance develops through metabolic reprogramming and activation of autophagy. Therefore, understanding how these targeted agents interface with cellular metabolism and the autophagic machinery provides a critical rationale for dual targeting strategies [207]. Mitochondria-targeted drugs interrupt mitochondrial metabolism and stimulate mitophagy in both KRAS wild-type and mutant colon cancer cells, linking mitochondrial function modulation to autophagic processes [208]. Cytosolic acetyl-coenzyme A (AcCoA) is a signaling metabolite that connects metabolism to mitophagy through its receptor NLRX1. The AcCoA-NLRX1 axis is a foundation of KRAS-inhibitor-driven mitophagy and contributes to drug resistance [209]. PINK1-mediated mitophagy enhances OXPHOS and supports tumor cells' survival under MAPK inhibitor treatment [210]. Similarly, blocking autophagy may abrogate BRAF inhibitor resistance caused by increased mitochondrial biogenesis and elevated OXPHOS rates [211]. Combined MEK inhibition and autophagy blockade amplify ferroptosis in LKB1-mutant KRAS-driven lung tumors by impairing mitochondrial respiration and increasing lipid peroxidation [212]. In a recent study, autophagy defects render thyroid carcinoma cells more susceptible to BRAFV600E inhibitors; this can be explained by the finding that mitochondrial respiration cannot support survival upon glycolytic blockade in autophagy-deficient cells [213]. In addition, inactivation of PIKfyve impedes PDAC cells' proliferation by impairing autophagy and lysosomal function while upregulating de novo lipid synthesis. Dual targeting of PIKfyve and KRAS-MAPK promotes tumor regression in numerous PDAC preclinical models, indicating that disrupting lipid metabolism and autophagy drives synthetic lethality plus KRAS-MAPK-targeted therapies for PDAC [214]. Glutamine metabolism also plays a critical role in KRAS-directed therapy resistance. Autophagy-dependent GSH synthesis facilitates MRTX1133 resistance by restricting apoptosis, while blocking autophagy potentiates MRTX1133 efficacy in KRASG12D-mutated PDAC [215]. Glutamine restriction suppresses the MEK inhibitor trametinib-induced autophagy, thereby promoting pyroptosis in KRAS-mutant NSCLC cells [216]. Likewise, the combination of glutamine transporter inhibitor V9302 and EGFR-TKI synergistically triggers apoptosis in NSCLC cells by reducing autophagic flux [217]. Additionally, increased OXPHOS is a hallmark in EGFR-TKI adaptive-resistant NSCLC cells. Mechanically, NCOA4-mediated ferritinophagy supports the synthesis of ISC proteins of ETC and OXPHOS [218], suggesting that ferritinophagy could be a druggable metabolic vulnerability to be exploited. Moreover, some metabolites such as lactate accumulate in the hypoxic TME and act as a bridge linking glycolysis and autophagy through lactylation. Histone lactylation upregulates the transcription of Pacer, facilitating autophagosome maturation through interacting with Beclin 1, which is crucial for cancer cells' survival in oxygen-deprived environments. In bevacizumab-resistant PDOs, simultaneous inhibition of histone lactylation and autophagy with bevacizumab treatment demonstrated remarkable treatment efficacy [48]. Beyond EGFR-, KRAS-, BRAF-, MEK-, and VEGFR-inhibitors, agents modulating PI3K-AKT signaling show similar interactions with autophagic flux and mitochondrial metabolism. For example, PI3K inhibition limits glycolysis but initiates protective autophagy, which provides lipid metabolites to sustain OXPHOS in a PLA2-dependent manner [219]. More important, PI3K inhibitor-resistant breast cancer shows enhanced sensitivity to various metabolic drugs (e.g., metformin) due to mTORC1-driven autophagy inhibition, highlighting autophagy deficiency as a metabolic sensitizing mechanism [200]. These examples illustrate that autophagy-metabolism crosstalk acts as a cytoprotective mechanism activated by kinase pathway inhibition, and its concurrent blockade can improve responses to targeted therapies across multiple oncogene-driven cancers. However, the dual role of autophagy in cancer progression suggests careful consideration of timing and context for such combination treatments. Integrating these strategies—metabolic drugs and/or autophagy modulators—into chemotherapy offers a new dimension of chemosensitization by exacerbating replicative and metabolic stress while preventing autophagy-mediated survival. For instance, mTORC1-mediated suppression of autophagy renders CDDP-resistant tumor cells more susceptible to 2-DG [72]. The glycolytic pathway is activated in paclitaxel-resistant cervical and breast cancer cell lines. Blockade of glycolysis by 2-DG or inhibiting phosphorylation of HK2 on Thr473 augments the paclitaxel response by decreasing autophagy flux [220,221]. Mitophagy sustains mitochondrial homeostasis and fuels malignant progression under chemotherapeutic stress. PGC1α enhances mitochondrial biogenesis during clusterin-induced mitophagy; combining PGC1α inhibitor with CDDP is more effective compared to monotherapy [222]. Notably, the crosstalk between metabolism and autophagy is also governed by epigenetic mechanisms. Histone H3 lysine18 lactylation (H3K18la) upregulates HNRNPC expression, promoting autophagy and glycolysis to establish a self-reinforcing histone lactylation-HNRNPC positive feedback loop, thereby driving gemcitabine resistance [223]. In addition, some agents can simultaneously target autophagy and metabolism to sensitize chemotherapy. Dehydrocostus lactone accelerates the ubiquitination and degradation of ACLY, resulting in lipid depletion, while impeding cytoprotective autophagy by down-regulating LAMP1/2 expressions, ultimately potentiating the cytotoxicity of 5-FU in patient-derived xenograft models [224]. Recent studies also reveal the value of the targeted degradation strategy for some key pro-survival molecules by triggering selective autophagy. Co-administration of GPX4-autophagy-targeting chimera (GPX4-AUTAC) and chemotherapy drugs shows a synergistic anti-tumor efficacy against breast cancer [225]. Moreover, polynucleotide kinase/phosphatase (PNKP) inhibition robustly increases ferritinophagy and intracellular iron levels, modulating the expression of key ferroptosis regulators, which is beneficial for ferroptosis induction in doxorubicin treatment for TNBC [226]. Withaferin A (WA) facilitates selective autophagic degradation of acetyl-CoA carboxylase 1 (ACC1), consequently disrupting de novo lipogenesis and lipid droplet accumulation. WA primes cancer cells for gemcitabine and prolongs survival in cholangiocarcinoma preclinical models [227]. These findings underscore a wider therapeutic landscape in which co-targeting autophagic and metabolic pathways can boost chemotherapy potency. Radiotherapy provides another setting where such combinations are effective. Radiation induces metabolic stress and autophagy as a protective response. In glioblastoma, NRBF2 reprograms the metabolic state of tumor cells—ATP production and the oxygen consumption rate increase—by activating autophagy, thus causing radioresistance [228]. Studies demonstrate that inhibiting mitochondrial respiration reduces hypoxia-induced radioresistance. Combining OXPHOS inhibitors such as metformin and atovaquone with radiation has been shown in preclinical models to improve radiosensitivity, leading to better tumor control and survival [229,230]. Meanwhile, autophagy inhibition further prevents the clearance of damaged mitochondria and ROS, amplifying radiation-induced cell death [231]. Although direct clinical evidence for this combinatorial approach remains limited, preclinical data support that blocking autophagy alongside metabolic inhibition (e.g., targeting OXPHOS, GLS) during radiation enhances radiosensitivity and promotes tumor cell death, particularly in therapy-resistant tumors, by exploiting the autophagy–metabolism linkage [231,232]. Despite the promise, challenges remain in balancing efficacy with safety. Both autophagy and metabolic pathways are critical for normal tissue homeostasis, raising concerns about toxicity. Strategies to mitigate these risks include intermittent dosing schedules, tumor-specific drug delivery systems, and biomarker-guided patient selection. For example, predictive markers such as autophagic flux indicators or metabolic profiling signatures may help identify patients most likely to benefit. Intermittent inhibition schedules can reduce adverse effects while preserving therapeutic impact. Furthermore, integrating autophagy and metabolic modulation into precision oncology frameworks—guided by genomic, proteomic, and metabolomic profiles—can enable more rational design of combination regimens with targeted therapies. Such biomarker-driven stratification will be key to improving efficacy while minimizing systemic toxicity in future trials. Biomarkers for predicting response to autophagy and metabolism-targeted therapies The effectiveness of therapies that target autophagy and metabolism varies widely among cancer patients, underscoring the need for reliable biomarkers to guide clinical use. Because autophagy and metabolic rewiring are highly dynamic and context-dependent, biomarkers can help identify tumors that rely most heavily on these pathways, enabling more precise and effective treatment strategies. One class of biomarkers involves measuring autophagic activity directly. Markers such as LC3-II accumulation, p62 degradation, and autophagosome-to-lysosome flux provide insights into baseline autophagy levels in tumors. However, LC3-II on its own cannot discriminate between physiological, basal autophagy that maintains cellular homeostasis and the pathological or therapy-induced autophagy relevant to tumor biology. Because its accumulation can result from both enhanced autophagosome formation or impaired degradation—making it difficult to determine whether autophagy is truly activated or blocked without additional assays [233,234]. Moreover, LC3 is expressed ubiquitously across tissues and can be influenced by non-autophagic processes such as endocytosis or certain stress responses, which further complicates interpretation [235]. Adding further complexity, autophagy is a continuous and dynamic process that fluctuates with nutrient availability, oxygen tension, and stress signaling—both under normal physiological conditions and in pathological states such as cancer. This inherent dynamism makes snapshot measurements of LC3-II or p62 unreliable proxies for real-time autophagic flux. Clinical studies also show inconsistent LC3-II changes in patient samples during autophagy modulation [236]. Collectively, these limitations highlight that while LC3-based assays are useful research tools, they currently lack the specificity, reproducibility, and clinical standardization required for use as reliable biomarkers in patient stratification. Consequently, combining multiple markers alongside functional assays improves specificity and helps differentiate basal from pathological or therapy-induced autophagy [237]. In addition, assessing autophagic flux using lysosomal inhibitors (e.g., bafilomycin A1) or live-cell imaging with tandem fluorescent LC3 reporters is also necessary for accurate evaluation [238]. Beyond LC3 and p62, broader panels of autophagy-related biomarkers such as ULK1, Beclin 1, Vps34, and ATG5 can provide deeper insight into distinct stages of autophagy regulation and flux. Furthermore, regulatory proteins such as PINK1, Parkin, NBR1, LAMP2A, and Hsc70 govern selective autophagy and CMA, linking mitophagy and proteostasis with tumor metabolism and survival [239,240]. Recognizing this complexity, autophagy-related signatures combining multiple biomarkers have been developed for prognosis and therapy response prediction. Combined 4E-BP1T37/46 phosphorylation—indicating high mTORC1 activity—and p62 accumulation not merely predict poor overall survival in breast cancer patients but also stratify patients for metabolic drugs selection [200]. Overall, autophagy-related proteins serve as promising indicators for cancer prognosis and therapy stratification but require further clinical validation to optimize patient selection for autophagy and metabolism-targeted therapies. Metabolic signatures also serve as critical predictors. Tumors exhibiting high glycolytic activity, as measured by ^18F-FDG PET imaging or the expression of glycolytic enzymes such as HK2 and LDH-A, often show resistance to cytotoxic therapies but may respond better to glycolysis inhibitors. HK2 is frequently upregulated in highly glycolytic cancers, and its inhibition—via agents like 2-DG—has reduced tumor proliferation and improved immunotherapy outcomes in TNBC and lung adenocarcinoma models [241,242]. Broadly, meta-analyses confirm that elevated HK2 expression correlates with worse overall and progression-free survival in various solid tumors, reinforcing its value as a prognostic biomarker and therapeutic target [243,244]. Similarly, overexpression of GLS or increased glutamine uptake has been correlated with resistance to platinum-based chemotherapy and tyrosine kinase inhibitors, suggesting a role for GLS inhibitors in these patients [245,246]. Circulating biomarkers provide a non-invasive avenue for patient monitoring. Serum lactate, glutamine levels, and lipid peroxidation products can reflect tumor metabolic activity and sensitivity to metabolic or autophagy-targeted agents. Circulating tumor DNA (ctDNA) offers a highly sensitive marker for tumor burden, stage, and therapy response due to its short half-life, allowing early detection of progression or resistance [247]. Furthermore, the predictive value of these biomarkers is closely linked to the underlying metabolic state of the tumor. For instance, while baseline levels of TCA cycle intermediates and amino acids could theoretically indicate sensitivity to agents like GLS inhibitors, their interpretation is complex and highly context-dependent. In multiple myeloma, sensitivity to the GLS inhibitor correlates with specific metabolic shifts, such as decreased nucleotide levels, rather than a simple baseline measurement. Resistance to such inhibitors often arises from compensatory pathways, such as the upregulation of pyruvate carboxylase, which sustains TCA cycle activity independently of glutamine [248]. Similarly, in TNBC, the availability of exogenous pyruvate can serve as an alternative anaplerotic substrate to replenish TCA intermediates like fumarate, thereby reducing drug sensitivity [249]. In glioblastoma stem cells, the therapeutic effect is linked more to intracellular glutamate depletion triggering an amino acid deprivation response than to direct changes in TCA cycle intermediates [250]. Thus, circulating levels metabolites of could offer a more dynamic and comprehensive picture of a tumor's metabolic plasticity and its likelihood of responding to or developing resistance to metabolic therapies [248,249]. In summary, biomarkers that reflect autophagic flux, metabolic rewiring, and sensitivity to cell death pathways offer powerful tools for guiding the use of autophagy- and metabolism-targeted therapies. Their integration into precision medicine approaches holds promise for selecting the right patients, optimizing treatment combinations, and ultimately overcoming therapeutic resistance in cancer. Challenges and future perspectives Despite remarkable progress in understanding the interplay between autophagy, metabolism, and cell death in cancer, several challenges hinder the full exploitation of this knowledge for therapeutic benefit. The complexity and context-dependent nature of these pathways demand more nuanced approaches to effectively target them in patients. One major challenge arises from the dual and often contradictory roles of autophagy in cancer. While autophagy can suppress tumor initiation by maintaining cellular quality control, it also enables established tumors to survive under stress [251]. For example, genetically engineered mouse models of lung adenocarcinoma and PDAC driven by mutant RAS or BRAF mutations show that autophagy suppresses early benign tumor formation but supports the growth of advanced malignancies [211,252]. Similarly, in K-ras-driven NSCLC models, deletion of autophagy genes reduces tumor growth, confirming that some tumors are highly dependent on autophagy for survival and metabolic adaptation [253]. This paradox complicates therapeutic strategies, as blanket inhibition of autophagy may have unintended consequences, including toxicity to normal tissues that depend on autophagy for homeostasis. Systemic inhibition in preclinical models has been shown to impair normal tissue homeostasis, underscoring the need for context-specific targeting. Future research must refine our understanding of when and where autophagy acts as a friend or foe, guiding more precise and context-specific interventions. Metabolic plasticity poses another significant barrier. Cancer cells dynamically shift between glycolysis, OXPHOS, and alternative nutrient sources to evade metabolic stress and therapy. RAS-mutant cancers, for instance, use autophagy to maintain metabolic flexibility under nutrient deprivation, enabling resistance to therapies targeting single pathways. This adaptability often leads to resistance against metabolic inhibitors [105]. Overcoming this challenge will require combination therapies that target multiple metabolic pathways and adaptive mechanisms simultaneously. Studies combining autophagy inhibitors (e.g., CQ) with chemotherapy or metabolic inhibitors in preclinical models have shown sensitization of tumors and enhanced therapeutic efficacy. However, designing such regimens demands careful consideration of potential side effects, given the shared metabolic processes in normal cells. A further complication is the TME, which influences autophagy and metabolism through nutrient availability, hypoxia, and interactions with stromal and immune cells [254]. Hypoxia, for example, can induce autophagy that protects tumor cells from immune attack, but under certain treatment contexts, such as HCQ combined with chemotherapy, it can also modulate immune cell polarization to enhance anti-tumor immunity. The heterogeneity within tumors means that different regions or cell populations may respond variably to therapy. Advanced tools that integrate single-cell and spatial analyses will be critical to unravel this complexity and to identify effective targets [254]. Biomarker development remains essential but challenging. Reliable, dynamic biomarkers are needed to predict which tumors depend on autophagy or specific metabolic pathways, to monitor treatment response, and to detect emerging resistance. The heterogeneity of tumors and the transient nature of autophagic and metabolic states require biomarker strategies that capture temporal and spatial variability. Primary research supports the utility of markers such as LC3B, for predicting therapy response, but assay standardization and validation remain obstacles [255]. Furthermore, the standardization of autophagy measurements is crucial for biomarker validation. The measurement of autophagic flux should rely on more comprehensive and objective experimental methods, such as live-cell imaging with tandem fluorescent reporters, rather than measuring static levels of certain proteins alone (e.g., LC3-II). Looking ahead, integration of multidisciplinary approaches holds great promise. Advances in imaging, single-cell technologies, and computational modeling can provide comprehensive insights into the triangular axis of autophagy, metabolism, and cell death. These tools will enable the identification of novel vulnerabilities and the design of personalized treatment strategies. Finally, translating these insights into safe and effective therapies will require robust preclinical models that faithfully recapitulate human tumor biology and microenvironmental complexity. PDOs, genetically engineered mouse models, and in vivo imaging have already proven valuable for studying therapy responses and resistance mechanisms. In summary, while the autophagy–metabolism–cell death axis offers exciting opportunities for cancer therapy, addressing these challenges through refined experimental models, biomarker development, and personalized approaches will be critical. Continued collaborative efforts across basic and clinical research are essential to transform these insights into durable patient benefit. Conclusions In conclusion, this review underscores the central role of the triangular axis connecting metabolism, autophagy, and cell death in supporting tumor survival and driving therapeutic resistance. The dynamic interplay among these processes allows cancer cells to adapt to environmental and treatment-induced stresses, maintain energy and biosynthetic demands, and evade cell death. Targeting this interconnected network, particularly through dual or combinatorial strategies that disrupt both metabolic reprogramming and autophagy while modulating cell death pathways, represents a promising approach to overcome resistance. Continued research into the molecular regulation of this axis, coupled with the development of precise biomarkers and innovative therapeutic strategies, will be critical to translating these insights into more effective cancer treatments. Consent to participate Not applicable. Ethics approval and consent to participate Not applicable. Consent for publication All authors have given consent for publication. Availability of data and material Not applicable. Funding This work was supported by the National Natural Science Foundation of China (Grant No. 82170974), the Hunan Provincial Natural Science Foundation of China (Grant No.2025JJ50539), and the Central South University Research Program of Advanced Interdisciplinary Studies (Grant No. 2023QYJC038). CRediT authorship contribution statement Wen-Dong Wan: Conceptualization, Visualization, Writing – original draft, Writing – review & editing. Muhammad Tufail: Conceptualization, Visualization, Writing – review & editing. Can-Hua Jiang: Writing – review & editing. Ning Li: Supervision, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments Not applicable. Abbreviations 2-deoxy-d-glucose AMP-Activated Protein Kinase Adenosine Triphosphate Autophagy-Related Gene/Protein Autophagy-Related Gene 16 Like 1 Beclin 1 B-cell lymphoma 2 BCL2/adenovirus E1B 19 kDa Interacting Protein 3 Chaperone-Mediated Autophagy Chloroquine Circulating Tumor DNA Fatty Acid Oxidation Fatty Acid Synthase serine hydroxymethyltransferase 2 Glucose Transporter 1 Glutathione Peroxidase 4 Hexokinase 2 Kirsten Rat Sarcoma Viral Oncogene Homolog Microtubule-Associated Proteins 1A/1B Light Chain 3 Lactate Dehydrogenase A Mitogen-Activated Protein Kinase Kinase Monocarboxylate Transporter 1/4 Mechanistic Target of Rapamycin Mechanistic Target of Rapamycin Complex 1 Nicotinamide Adenine Dinucleotide Phosphate (reduced form) Nuclear Receptor Coactivator 4 Nuclear Factor Erythroid 2–Related Factor 2 Oxidative Phosphorylation Pyruvate Dehydrogenase Kinase Positron Emission Tomography Phosphoinositide 3-Kinase Pyruvate Kinase M2 p53 Upregulated Modulator of Apoptosis Reactive Oxygen Species Tricarboxylic Acid Cycle Triple-Negative Breast Cancer Unc-51 Like Autophagy Activating Kinase 1 Voltage-Dependent Anion Channel Vacuolar Protein Sorting 34 Focal Adhesion Kinase Family-Interacting Protein of 200 kDa Hypoxia-Inducible Factor 1-alpha PTEN-Induced Kinase 1/Parkin Solute Carrier Family 7 Member 11 Programmed Death-Ligand 1 Cancer-Associated Fibroblasts Data availability No data was used for the research described in the article. References Associated Data Data Availability Statement Not applicable. No data was used for the research described in the article.