1. Introduction

Ferroptosis is an iron-dependent and lipid peroxidation-driven form of regulated cell death, which is distinct from traditional cell death pathways such as apoptosis, necrosis, and autophagy^[1]. Operationally, it proceeds without the caspase activation and chromatin condensation hallmarks of apoptosis, lacks the massive cellular swelling and early plasma membrane rupture typical of necrosis, and occurs independently of the double-membrane vesicles characteristic of autophagy. Instead, the key features of ferroptosis include the generation of reactive oxygen species (ROS), elevated intracellular ferrous iron levels, inhibition of antioxidant pathways, and accumulation of lipid peroxides on cellular membranes. Morphologically, it is uniquely characterized by shrunken mitochondria with increased membrane density and reduced or absent cristae^[2]. In recent years, ferroptosis has garnered significant attention due to its broad involvement in multiple pathophysiological processes, playing important roles in neurodegenerative diseases, ischemia-reperfusion injury, fibrosis, and more^[3]. In the treatment of malignancies, inducing ferroptosis has emerged as a promising strategy to overcome resistance to conventional therapies, including chemotherapy, radiotherapy, and immunotherapy^[4,5].

Metabolic reprogramming is a fundamental hallmark of cancer^[6]. To meet the demands of rapid proliferation, tumor cells undergo metabolic remodeling to acquire energy and biosynthetic precursors. However, this hypermetabolic state also confers specific vulnerabilities^[7]. Mounting evidence indicates that metabolic reprogramming is a central driver of the ferroptosis regulatory network, where the flux through carbohydrate, lipid, and amino acid pathways directly determines cellular susceptibility to lipid peroxidation^[8,9].

Among various nutrients, profound alterations in amino acid metabolism provide multi-dimensional support for tumor cells spanning from energy supply to oxidative stress^[10-12]. Amino acids are not only the fundamental blocks for protein synthesis but also serve as crucial metabolic intermediates. By supplying carbon to the tricarboxylic acid (TCA) cycle and nitrogen for nucleotide synthesis, they drive the continuous proliferation and biomass accumulation of tumor cells^[10,13,14]. This metabolic dependency is particularly prominent in maintaining redox homeostasis^[11]. Multiple studies have highlighted that tumor cells enhance the metabolism of glutamine, cysteine, and glycine to promote the synthesis of antioxidant molecules like glutathione (GSH)^[10,11,15]. This mechanism allows tumor cells to neutralize ROS generated during metabolism, thereby sustaining survival under high oxidative stress. Concurrently, specific amino acids (e.g., glutamine) are involved in lipid synthesis and regulate oxidative phosphorylation to meet the energy demands of specific tumor cell populations, such as leukemia stem cells^[14,16].

Tumor cells often exhibit “addiction” to specific amino acids due to the upregulated expression of amino acid transporters or enhanced activity of key catabolic enzymes^[17]. Consequently, targeted disruption of amino acid supply or utilization represents a powerful strategy to sensitize cancer cells to ferroptosis. In light of this, this review systematically explores the regulatory landscape of amino acid metabolism in ferroptosis, elucidating the molecular mechanisms of ferroptotic evasion and the potential for leveraging these metabolic dependencies to enhance ferroptosis-based research and precision strategies.

2. The Canonical Redox Axis

2.1 Cysteine

Long before the formal conceptualization of ferroptosis, the absolute cellular dependency on specific amino acids for survival and oxidative defense was firmly established. In 1955, Harry Eagle demonstrated that cysteine is strictly essential for the proliferation of mouse L fibroblasts and human HeLa cells^[18,19]. Building on this foundational observation, Shiro Bannai et al. later revealed that cultured cells exhibit a unique vulnerability to cystine depletion compared to other amino acids, laying the groundwork for the cystine-dependent ferroptosis paradigm^[20].

Cysteine dictates the cellular threshold for lipid peroxidation and serves as the central node of the anti-ferroptotic regulatory network. As the rate-limiting precursor for GSH biosynthesis, cysteine determines the availability of the essential reductive cofactor for glutathione peroxidase 4 (GPX4)^[4,21]. GPX4 relies on GSH to reduce toxic lipid peroxides on cell membranes into non-toxic alcohols, thereby preventing ferroptosis^[22,23]. Intracellular cysteine is primarily sourced via the cystine/glutamate antiporter system xc- (an SLC7A11/SLC3A2 heterodimer), which imports extracellular cystine for subsequent cytosolic reduction. Consequently, the SLC7A11-cysteine-GSH-GPX4 axis constitutes the backbone of the cellular antioxidant defense system. In this system, the availability of cysteine is crucial for cell survival^[24,25].

The profound reliance on this axis exposes a targetable metabolic vulnerability. Pharmacological inhibition of system xc- by classical ferroptosis inducers, such as erastin and imidazole ketone erastin, precipitates a rapid depletion of the intracellular GSH pool. This upstream substrate starvation collapses the antioxidant system and triggers ferroptosis, even when GPX4 remains functional^[26]. Beyond pharmacological intervention, the tumor suppressor protein p53, interferon-γ secreted by immune cells, and radiotherapy can downregulate SLC7A11 expression, thereby inducing ferroptosis in tumor cells through deprivation of cysteine^[27-29]. In contrast, drug-resistant cancer cells often remodel their metabolism to fortify this defense. For instance, the circadian regulator ARNTL2 directly enhances SLC7A11 transcription, sustaining high-capacity cysteine uptake. This metabolic adaptation establishes a robust shield against chemotherapeutic stress, as demonstrated in both in vitro and in vivo models of colorectal cancer^[30]. Similarly, the stem cell factor SOX2 directly binds and activates the SLC7A11 promoter to confer ferroptosis resistance in lung cancer stem cells, which can be abrogated by SOX2 oxidation at Cys265^[31]. Furthermore, lipid metabolism exhibits strict crosstalk with this axis. Inhibition of the lipid enzyme SCD1 blocks the AKT-NRF2 pathway to transcriptionally downregulate SLC7A11, priming lung adenocarcinoma for ferroptosis^[32]. Post-transcriptionally, RNA-binding proteins exert context-dependent control: RBMS1 promotes SLC7A11 translation by bridging its UTRs in non-small cell lung cancer (NSCLC)^[33], whereas RBMS2 destabilizes SLC7A11 mRNA to promote ferroptosis in colorectal cancer^[34]. At the post-translational level, the lysosomal protein LAPTM4B suppresses the NEDD4L/ZRANB1-mediated ubiquitin-proteasomal degradation of SLC7A11 to counteract ferroptosis in NSCLC^[35]. Concurrently, the E3 ligase ZNRF2 tunes the ubiquitination of the auxiliary subunit SLC3A2 at K147, driving its plasma membrane translocation to maximize cystine import in lung adenocarcinoma, which can be therapeutically blocked by the synthetic peptide K147^[36].

Under extreme pressure from extracellular cysteine scarcity, cells exhibit complex compensatory logic. When external uptake is blocked, cells activate the GCN2-ATF4 signaling axis to drive the transsulfuration pathway, converting methionine to cysteine to sustain basic cell growth^[37,38]. Concurrently, inhibition of cysteinyl-tRNA synthetase in the protein translation machinery leads to the accumulation of uncharged tRNA, triggering metabolic reprogramming. This includes upregulation of key enzymes like cystathionine β-synthase (CBS) to induce endogenous cysteine synthesis, thereby antagonizing ferroptosis^[39]. Furthermore, broader amino acid starvation (such as arginase-induced arginine depletion) activates the GCN2-eIF2α pathway to suppress global protein translation. This translational arrest functions as a metabolic shunt, redirecting the critically limited intracellular cysteine pool away from protein synthesis and exclusively toward GSH biosynthesis to neutralize lipid peroxides^[40,41]. Importantly, the restriction of non-cysteine amino acids in the tumor microenvironment (TME) also acts as a potent ferroptosis sensitizer. Such restriction triggers the GCN2-dependent integrated stress response and ATF4-driven mitochondrial respiration, which increases ROS leakage and drastically lowers the threshold for lipid peroxidation when GSH is depleted^[42]. Beyond canonical GSH production, cancer cells could exhibit atypical metabolic adaptations to cysteine scarcity. Under cystine starvation, the catalytic subunit of glutamate-cysteine ligase catalytic subunit promiscuously utilizes alternative amino acids to synthesize non-canonical γ-glutamyl peptides. This atypical peptide generation buffers toxic glutamate accumulation, powerfully protecting NRF2-hyperactive NSCLC from ferroptosis^[43].

Cysteine metabolism displays high spatial compartmentalization and functional diversity. For example, CHAC1-mediated GSH degradation may preferentially maintain mitochondrial cysteine levels to support iron-sulfur cluster synthesis, although such preservation of mitochondrial function may paradoxically promote lipid peroxidation under specific conditions in NSCLC^[44]. Moreover, cysteine serves as a crucial precursor for hydrogen sulfide (H₂S), which modulates antioxidant gene expression via protein S-sulfhydration^[38,45].

The high cellular dependence on cysteine metabolism provides a clear point for clinical intervention. In non-malignant pathologies, such as diabetic wound healing, therapeutic efforts focus on mitigating ferroptosis by correcting cysteine deficiencies within hyperglycemic microenvironments^[46]. Conversely, in pancreatic ductal adenocarcinoma, synthetic biology approaches have engineered tumor-colonizing bacteria that secrete cysteinase to locally deplete cysteine, thereby dismantling the anti-ferroptotic shield^[47]. Crucially, the intersection of epigenetic modulation and amino acid metabolism constitutes a targetable vulnerability, as exemplified by the dual function of histone deacetylase (HDAC) inhibitors. In non-malignant tissues like neurons, specific HDAC inhibitors suppress ferroptosis by preserving redox homeostasis and upregulating neuroprotective cascades^[48]. Conversely, in acute myeloid leukemia, HDAC inhibition transcriptionally represses SLC7A11, choking off cystine import and precipitating GSH depletion^[49]. Nevertheless, safely balancing systemic amino acid requirements and identifying robust biomarkers for cysteine deprivation remain critical imperatives for future clinical translation^[23,27].

2.2 Glycine

Glycine, as one of the three precursor amino acids for GSH synthesis, primarily exerts an inhibitory effect on ferroptosis. However, it may exhibit dual regulatory characteristics under specific pathological contexts^[50,51].

Glycine mainly inhibits ferroptosis by participating in GSH biosynthesis. Together with cysteine and glutamate, glycine is used to synthesize GSH via the sequential actions of glutamate-cysteine ligase and glutathione synthetase. This process is crucial for clearing lipid peroxides during ferroptosis^[52,53]. Therefore, glycine abundance directly limits the de novo synthesis rate of GSH. Depletion of intracellular glycine leads to decreased GSH reserves and, consequently, heightened sensitivity to ferroptosis.

Within the regulatory network of this core pathway, the serine-glycine conversion and its upstream transcription factors play pivotal roles. Serine hydroxymethyltransferase (SHMT) catalyzes the conversion of serine to glycine and 5,10-methylenetetrahydrofolate (5,10-CH₂-THF). This reaction drives one-carbon metabolism, generating reduced coenzyme NAD(P)H to maintain cellular redox balance^[54-56]. This reaction is the primary pathway for glycine synthesis. Previous research indicated that in hepatoblastoma models, the LATS2/YAP1 pathway activates ATF4, which further upregulates the expression of phosphoserine aminotransferase 1, a key enzyme in serine synthesis. This increases the synthesis of glycine and cysteine, thereby enhancing cellular antioxidant capacity and inhibiting ferroptosis^[51].

Glycine uptake and exogenous supply are also important peripheral factors influencing ferroptosis. Dietary restriction of serine and glycine can effectively reduce the cellular capacity for GSH regeneration and exhibits significant synergistic effects with ferroptosis inducers to suppress tumor growth^[50,57]. Furthermore, the impact of glycine on plasma membrane integrity has garnered attention. Although glycine is widely recognized for its ability to inhibit plasma membrane rupture mediated by NINJ1, recent evidence suggests that glycine cannot block ferroptosis-induced membrane damage^[58]. This phenomenon reveals that the membrane damage mechanism in ferroptosis may possess unique properties independent of NINJ1 polymerization or may involve a glycine-insensitive, non-canonical pore formation process.

Beyond canonical GSH biosynthesis, the metabolic fate of glycine extends to profound epigenetic regulation. Although the effect, dominated by the GSH pathway, is anti-ferroptotic, prolonged or high-dose glycine exposure might instead promote ferroptosis in certain pathological states by inducing methylation reactions (e.g., GPX4 promoter methylation), which could downregulate the expression of antioxidant enzymes^[57]. This mechanistic duality dictates that therapeutic interventions targeting amino acid metabolism must be rigorously calibrated to the precise tissue context and metabolic dosage.

2.3 GPX4-independent compensatory defense system

While the canonical SLC7A11-GSH-GPX4 axis is a primary barrier against lipid peroxidation, cancer cells can employ robust, GPX4-independent compensatory mechanisms to survive when GPX4 is deficient or when canonical amino acid supplies are compromised.

The most prominent parallel defense system is mediated by ferroptosis suppressor protein 1 (FSP1). Localized primarily at the plasma membrane, FSP1 acts as an oxidoreductase that uses NAD(P)H to reduce ubiquinone (Coenzyme Q10, CoQ10) to ubiquinol (CoQ10H₂). Ubiquinol then functions as a potent lipophilic radical-trapping antioxidant, which directly halts the propagation of lipid peroxides independent of the GSH pool^[59,60].

In addition to the FSP1-CoQ10 axis, other organelle-specific and metabolic pathways provide supplementary layers of defense. The GTP cyclohydrolase 1 (GCH1) pathway synthesizes tetrahydrobiopterin (BH4), another endogenous lipophilic antioxidant that protects cellular membranes from oxidative damage^[61]. Furthermore, within the mitochondria, dihydroorotate dehydrogenase (DHODH) operates in parallel to mitochondrial GPX4 to reduce CoQ to CoQH₂, specifically protecting the inner mitochondrial membrane from ferroptotic damage^[62].

Crucially, the operation of these GPX4-independent surveillance systems is intrinsically coupled to amino acid metabolism. The generation of GTP, the obligate precursor for GCH1-mediated BH4 synthesis, demands a continuous influx of glutamine, glycine, and aspartate^[63]. In parallel, DHODH-driven protection relies on de novo pyrimidine biosynthesis, which is fueled by glutamine and aspartate^[64]. Consequently, even when circumventing the canonical cysteine-GSH axis, tumors remain metabolically tethered to the continuous flux of these specific amino acids to sustain alternative anti-ferroptotic shields.

3. The Mitochondrial Axis

3.1 Glutamine and glutamate

Glutamine metabolism exerts a profound, dichotomous influence on ferroptosis regulation. Following cellular uptake via the SLC1A5 transporter, the metabolic fate of this conditionally essential amino acid dictates whether it drives lethal lipid peroxidation or fortifies antioxidant defenses. In pro-ferroptotic contexts, especially during cystine deprivation, glutaminolysis acts as an indispensable driver. Glutamine is converted by glutaminase (GLS) to glutamate, which is subsequently transaminated to α-ketoglutarate (α-KG). The entry of α-KG into the mitochondrial TCA cycle fuels oxidative phosphorylation. However, this elevated mitochondrial respiration concomitantly generates substantial electron leakage, culminating in the burst accumulation of lipid-ROS that executes ferroptosis^[65-68].

Conversely, glutamine-derived glutamate functions as a critical precursor for GSH biosynthesis, equipping cells with the antioxidant capacity to neutralize lipid peroxides^[69-72]. This metabolic bifurcation is precisely governed by complex regulatory networks and subcellular contexts. For instance, the tumor suppressor p53 transcriptionally upregulates GLS2, directing glutamine flux toward α-KG and thereby sensitizing cells to mitochondria-dependent ferroptosis^[73]. Notably, the requirement for mitochondrial metabolism varies by the ferroptotic trigger: while essential for cystine-deprivation-induced ferroptosis, its significance diminishes when ferroptosis is initiated by direct GPX4 inhibition, highlighting distinct metabolic dependencies^[67,74].

Beyond the core pathway, upstream signaling and peripheral metabolic networks further reshape glutamine’s impact on ferroptosis. In hepatocellular carcinoma (HCC), loss of the tyrosine catabolic enzyme HPD forces cells into a state of strong glutamine addiction. This is mediated by activation of the AMPK/mTOR/p70S6K axis, which upregulates GLS activity to support rapid tumor proliferation, simultaneously exposing a metabolic vulnerability to ferroptosis^[75]. In head and neck squamous cell carcinoma, blocking glutamine metabolism triggers autophagy and nutrient sensing, leading to compensatory accumulation of polyunsaturated fatty acids and, consequently, sensitizing cancer cells to lipid peroxidation^[76]. Additionally, inhibiting GLS in lung adenocarcinoma can ameliorate nutrient competition in the TME, activating the anti-tumor immunity of CD8⁺ T cells^[77]. Consequently, deciphering the contextual fate of glutamine is paramount for designing ferroptosis-targeted therapeutic interventions.

3.2 Aspartate and asparagine

Aspartate and its derivative, asparagine, primarily inhibit ferroptosis but exhibit dual regulatory characteristics under specific metabolic backgrounds. Aspartate mainly relies on the following three metabolic pathways to regulate ferroptosis: pyrimidine biosynthesis, urea cycle shunting, and transaminase-mediated TCA cycle anaplerosis.

Aspartate inhibits ferroptosis through a key mechanism involving its participation in pyrimidine biosynthesis to maintain mitochondrial antioxidant potential. Within the inner mitochondrial membrane, aspartate serves as a precursor for pyrimidine synthesis. Its metabolic product, dihydroorotate, is oxidized to orotate by DHODH, a process where CoQ is reduced to CoQH₂^[62]. CoQH₂ is a potent lipophilic antioxidant capable of directly scavenging lipid peroxyl radicals. Therefore, the aspartate-DHODH axis constitutes an important ferroptosis defense system independent of GPX4 and FSP1, specifically responsible for clearing lipid peroxides within mitochondria. Furthermore, the dramatic consumption of N-carbamoyl-L-aspartate upon GPX4 system impairment confirms the compensatory protective role of this pathway under ferroptotic stress^[62].

Beyond its role as a precursor for pyrimidine synthesis, aspartate can also inhibit ferroptosis through metabolic shunting into the urea cycle. Hu et al. revealed that in NSCLC, Argininosuccinate Synthase 1, a key urea cycle enzyme, promotes the catabolism of glutamine in the cytosol. This process channels glutamine-derived nitrogen into the urea cycle via aspartate, thereby preventing excessive glutamine entry into the TCA cycle and subsequently reducing ROS production^[78].

Aspartate aminotransferase (GOT) primarily coordinates the interconversion between aspartate and TCA cycle intermediates (such as oxaloacetate and α-KG), playing an important role in maintaining mitochondrial membrane potential and redox homeostasis^[78,79]. In certain tumor types, glutamine is converted via GOT to α-KG, which enters the TCA cycle, driving ROS production by enhancing oxidative phosphorylation and thereby promoting ferroptosis^[79].

4. The Epigenetic and Translational Axis

4.1 Methionine

As an essential sulfur-containing amino acid, methionine exerts a context-dependent effect on ferroptosis, capable of both defending against it by maintaining antioxidant systems and, under specific metabolic backgrounds, exacerbating oxidative stress.

In pathways inhibiting ferroptosis, methionine primarily functions via the transsulfuration pathway and the one-carbon metabolism pathway. Methionine can be converted to homocysteine and subsequently to cysteine via the transsulfuration pathway. This metabolic flow is a crucial supply line for maintaining the intracellular GSH pool and GPX4 activity, especially when extracellular cystine uptake is impaired^[3,80]. Besides, methionine can be converted to S-adenosylmethionine (SAM) by the key metabolic enzyme methionine adenosyltransferase 2A, participating in epigenetic regulation to inhibit ferroptosis. SAM mediates histone H3K4me3 modification, which upregulates the expression of ACSL3, promoting the production of monounsaturated fatty acids to resist lipid peroxidation in gastric cancer^[81]. Methionine transmembrane transport is also an important factor influencing ferroptosis sensitivity. SLC43A2-mediated methionine uptake activates the NF-κB signaling pathway, thereby promoting the transcription of both SLC43A2 and GPX4. Concurrently, elevated methionine levels increase intracellular GSH content, forming a positive feedback loop^[82].

However, the regulation of ferroptosis by methionine metabolism is not solely inhibitory. Recent studies have revealed its potential to promote a pro-oxidant environment under specific conditions. Xia et al. demonstrated that SAM, produced from methionine metabolism, is an essential methyl donor for CoQ10 synthesis, which plays a central role in lipid-ROS accumulation induced by cystine deprivation^[83]. This finding implies that, in specific cellular contexts, depriving methionine might prevent lipid peroxidation by blocking ubiquinone synthesis.

Intervention strategies targeting methionine metabolism show great potential for clinical translation, but their efficacy is significantly influenced by the duration and frequency of intervention. Although long-term methionine deprivation may induce ferroptosis resistance by inhibiting CHAC1 protein synthesis, studies indicate that intermittent methionine starvation can significantly enhance tumor sensitivity to ferroptosis inducers and immunotherapy by stimulating CHAC1 transcription and accelerating GSH degradation^[84]. Meanwhile, in glioma, methionine deprivation inhibits tumor proliferation and epithelial-mesenchymal transition via non-coding RNA networks such as TP53TG1/miR-96-5p/STK17B^[85], further supporting the application value of dietary interventions in precision oncology. Additionally, while methionine shows preliminary value as a salivary biomarker for early cancer screening, the precise correlation between fluctuations in salivary amino acid concentrations and intratumoral metabolic states requires validation through larger-scale clinical data^[86].

4.2 Serine

Serine functions as a potent endogenous suppressor of ferroptosis. Positioned at the nexus of glycolysis and one-carbon metabolism, it transcends its canonical role in macromolecular biosynthesis to serve as the central metabolic reservoir for antioxidant networks. By fueling these defense systems, serine dictates cellular redox homeostasis and robustly fortifies tumor cells against lethal lipid peroxidation^[11,87].

Serine inhibits ferroptosis by driving the one-carbon metabolism unit to generate NADPH and GSH. Firstly, serine is converted to glycine by SHMT. This process accompanies the conversion of tetrahydrofolate to methylenetetrahydrofolate, ultimately driving the folate cycle to produce reduced NADPH, which is a key cellular reducing power^[88]. NADPH serves as the cofactor for glutathione reductase to regenerate GSH and as the electron donor for the FSP1-CoQ10 antioxidant system, directly determining the cell’s capacity to clear lipid peroxides^[11]. Secondly, serine is the direct carbon skeleton donor for cysteine synthesis, replenishing the cysteine pool via the transsulfuration pathway and thereby ensuring GSH synthesis^[87].

Within this metabolic network, key enzymes CBS and phosphoglycerate dehydrogenase (PHGDH) play decisive roles. CBS is responsible for condensing serine and homocysteine to form cystathionine, the rate-limiting step of the transsulfuration pathway. Erdélyi et al. confirmed that under pressure from cysteine scarcity or ferroptosis induction, tumor cells heavily rely on CBS to utilize serine for maintaining cysteine levels. Additionally, CBS can drive persulfidation reactions, a specific post-translational modification mechanism that also significantly inhibits lipid peroxidation in breast cancer^[89]. PHGDH is the rate-limiting enzyme for de novo serine synthesis, catalyzing the conversion of D-3-phosphoglycerate to 3-phosphohydroxypyruvate, the first critical step in serine biosynthesis^[90,91]. In podocyte models of diabetic kidney disease, PHGDH deficiency exacerbates cellular damage (e.g., cytoskeleton disruption) and significantly promotes ferroptosis^[92].

Moreover, serine metabolism indirectly regulates ferroptosis via “non-canonical” pathways such as epigenetic mechanisms. Serine is a major donor of one-carbon units, supporting SAM synthesis. SAM, as a universal methyl donor, directly influences the methylation status of histones and DNA, thereby modulating the transcriptional expression of ferroptosis-related genes^[93]. This mechanism indicates that serine determines cell fate not only by directly providing antioxidant molecules but also by reshaping the epigenetic landscape of the genome.

4.3 Selenocysteine (Sec)

As a critical determinant of tumor ferroptosis susceptibility, Sec biosynthesis and incorporation fundamentally subvert canonical translation. Sec is uniquely synthesized in situ on its cognate tRNA[Ser]Sec via a sequential enzymatic cascade^[94]. Its targeted integration into the anti-ferroptotic enzyme GPX4 requires reprogramming the UGA stop codon. This is orchestrated by the 3′-UTR SECIS element, which recruits SBP2 and eEFSec to circumvent translation termination and guide Sec-tRNA into the ribosomal A-site^[95]. Evolutionary selection shapes that GPX4 strictly relies on a Sec residue, rather than a canonical cysteine, at its catalytic center. This requirement is essential for GPX4 to maintain its catalytic activity and resist irreversible inactivation when confronted with severe lipid hydroperoxide stress^[96]. As demonstrated in engineered in vitro neurons and in vivo mouse models, GPX4 variants harboring a canonical cysteine are highly vulnerable to hydroperoxide-induced overoxidation, underscoring the absolute necessity of Sec for survival and central nervous system integrity^[96]. To sustain the supply of this highly energy-intensive amino acid, tumor cells orchestrate multidimensional metabolic reprogramming and Sec-specific uptake mechanisms. Furthermore, selenium functions beyond acting as a mere translational building block; it can drive a robust adaptive transcriptional program via transcription factors such as TFAP2c and Sp1, which significantly upregulate the broader selenome, including GPX4, to confer enhanced transcriptional resistance against ferroptotic stress^[97].

Therapeutically, the systemic dependency on exogenous selenium-containing amino acids exposes a critical metabolic vulnerability. In MYCN-amplified neuroblastoma, tumor cells are profoundly dependent on LRP8 receptor-mediated uptake of selenoprotein P to maintain Sec homeostasis. Therefore, genetic or pharmacological blockade of this influx precipitates robust ferroptosis^[98]. Intracellularly, peroxiredoxin 6 (PRDX6) functions as a pivotal selenium receptor and chaperone, driving a highly efficient, SCLY-independent delivery system by covalently binding to selenoglutathione (GS-Se-SG). Consequently, PRDX6 is critical for sustaining GPX4 biosynthesis and anti-ferroptotic defense^[99,100]. Furthermore, the incorporation of Sec during selenoprotein translation is deeply subjected to crosstalk regulation by other metabolic pathways. In HCC, the mevalonate pathway not only promotes CoQ10 synthesis but also directly facilitates Sec-tRNA modification. This crosstalk ensures the successful translation of Sec-dependent antioxidant enzymes, an adaptation that can be therapeutically dismantled using statins^[101].

Recent paradigms have further expanded the metabolic repertoire of selenium-containing amino acids beyond classical selenoprotein synthesis. Notably, selenium-mediated antioxidant defense can bypass the rate-limiting translation of Sec-dependent enzymes. The Sec metabolic intermediate, hydrogen selenide, can directly and rapidly reduce ubiquinone within mitochondria via sulfide quinone oxidoreductase, establishing a GPX4-independent shield against lipid peroxidation^[102]. Conversely, the incorporation of Sec does not universally confer ferroptosis resistance. The Sec residue within thioredoxin reductase 1 actively promotes the E3 ubiquitin ligase-mediated degradation of the KEAP-NRF2 axis. This cascade ultimately suppresses GPX4 expression, thereby exerting a paradoxical but potent pro-ferroptotic effect across both in vitro and in vivo models^[103].

5. Emerging Amino Acids and Multi-level Crosstalk

5.1 Arginine

Arginine primarily acts as a “promoting factor” in ferroptosis regulation, but exhibits complex dual regulatory characteristics under specific conditions of metabolic deprivation or within distinct microenvironmental contexts^[104,105]. As a semi-essential amino acid, arginine is not only a fundamental material for protein synthesis, but its catabolic products also serve as significant disruptors of intracellular redox balance, capable of directly driving lipid peroxidation through multiple metabolic branches.

In terms of core metabolic pathways, arginine promotes ferroptosis mainly via three routes: “polyamine metabolism”, the “urea cycle branch”, and “nitric oxide metabolism”. Firstly, intracellular arginine is converted to ornithine, which enters the polyamine synthesis pathway. During the catabolism of polyamines (such as spermidine and spermine), significant amounts of hydrogen peroxide (H₂O₂) are produced as byproducts. This high level of H₂O₂ directly provides the oxidant required for the Fenton reaction, thereby inducing lipid peroxidation^[105]. Further research has found that this sensitization effect is modulated by KEAP1 status. SAT1 is the rate-limiting enzyme for polyamine catabolism. In tumor cells with a wild-type KEAP1 background, KRAS inhibitors activate the JNK/c-Jun/SAT1 axis, accelerating the conversion of arginine to polyamines and inducing ferroptosis. However, this metabolic sensitivity disappears in cells with KEAP1 mutations^[104]. Secondly, metabolic intermediates generated from arginine in the urea cycle can also disrupt the antioxidant system. Fumarate, generated from arginine via argininosuccinate lyase, is electrophilic and can conjugate with GSH to form succinicGSH, directly depleting the intracellular antioxidant reserve and accelerating ferroptosis^[106].

L-Arginine serves as the key precursor for nitric oxide (NO). Under catalysis by inducible nitric oxide synthase or oxidation by ROS, L-Arginine releases NO, which is rapidly converted to reactive nitrogen species (RNS)^[107]. RNS not only directly oxidize and deplete GSH, but also downregulate SLC7A11 expression to restrict cystine uptake and inhibit glutathione reductase, thereby blocking GSH regeneration. Furthermore, NO can exacerbate mitochondrial damage by inhibiting the activity of aconitase activity^[108,109].

The availability of arginine and its transport regulation constitute peripheral factors influencing ferroptosis. In colorectal cancer studies, Lin et al. found that due to the loss of ornithine transcarbamylase, colorectal cancer cells exhibit strong arginine dependence^[110]. Upon deprivation of exogenous arginine, cells activate the AMPK-p53-p21 pathway and enter a reversible “quiescent state”. Notably, these quiescent cells show heightened sensitivity to ferroptosis inducers, providing a rationale for the clinical application of “arginine-restricted diets” combined with ferroptosis inducers^[110]. Conversely, in tumors like liver cancer, where arginine concentration is abnormally elevated, arginine acts as an oncogenic signaling molecule. Mossmann et al. confirmed that arginine directly binds the RNA-binding protein RBM39, driving oncogenic metabolic reprogramming by regulating gene expression (e.g., ASNS) rather than inducing cell death^[111].

Furthermore, the heterogeneity of arginine metabolism within the TME is a key determinant of ferroptosis sensitivity. Mbah et al. suggested that arginine accessibility in the TME not only affects cancer cell survival but also modulates the ferroptosis response and anti-tumor activity of infiltrating T cells^[112]. In esophageal squamous cell carcinoma, dysregulation of the ISCU-p53 axis leads to upregulated expression of arginase 1 in macrophages. This not only promotes M2 polarization but also enhances macrophage resistance to ferroptosis via metabolic remodeling, ultimately contributing to resistance against PD-1 inhibitors^[113]. Given that arginine deprivation therapies (e.g., ADI-PEG20) have entered clinical trials for various arginine auxotrophic tumors, future research needs to define how to precisely trigger ferroptosis in tumor cells via arginine metabolic pathways without damaging immune cell function^[114].

5.2 Tryptophan

Tryptophan metabolism orchestrates ferroptotic resistance through a sophisticated defense network comprising direct radical-scavenging, receptor-mediated signaling, and co-factor biosynthesis.

Mechanistically, metabolites derived from the kynurenine and serotonin pathways, specifically 3-hydroxyanthranilic acid (3-HA) and 5-hydroxytryptamine (5-HT), function as potent radical-trapping antioxidants. These molecules possess unique chemical moieties enabling them to directly donate hydrogen atoms to lipid peroxyl radicals, thereby terminating the autocatalytic peroxidation chain^[115]. This antioxidant protection holds significant clinical relevance for intestinal health. Multiple studies indicated that tryptophan derivatives produced by the gut microbiota (e.g., indoles, 5-HT) can protect the intestinal barrier by mitigating oxidative stress, preventing cell damage associated with inflammatory bowel disease^[57,116]. The efficacy of this metabolic defense largely depends on the regulation of key enzymes. Among them, kynureninase promotes 3-HA generation to support cell survival, whereas monoamine oxidase A degrades 5-HT, weakening cellular resistance to ferroptosis^[115].

Beyond direct antioxidation, the IL4i/tryptophan/aryl hydrocarbon receptor (AhR) axis constitutes a robust compensatory defense. The L-amino acid oxidase IL4i1 synergizes with IDO1 to convert tryptophan into indole-3-pyruvate, which dual-functions as a lipid-ROS scavenger and an endogenous ligand for the AhR. AhR activation transcriptionally upregulates a cohort of antioxidant genes, fortifying cellular tolerance to ferroptotic stress^[117].

Furthermore, BH4, which is derived from tryptophan metabolism, serves as another potent antioxidant, assisting cancer cells in combating ferroptosis under oxidative stress^[118].

5.3 Branched-chain amino acids (BCAA)

BCAAs (including leucine, isoleucine, and valine) exert bidirectional regulatory effects on ferroptosis. On one hand, BCAAs function as important carbon and nitrogen sources for glutamate, playing an inhibitory role by maintaining antioxidant system activity. On the other hand, abnormal accumulation of BCAAs can promote ferroptosis by altering lipid metabolism and signal transduction.

BCAAs exert their ferroptosis-inhibitory function primarily via the “branched-chain amino acid aminotransferases (BCAT)/glutamate/GSH” metabolic axis, a process dependent on BCAT1/BCAT2^[119,120]. BCAT2 catalyzes the transamination of BCAAs to α-KG, generating branched-chain α-keto acids and glutamate. In liver, pancreatic, and prostate cancers, high expression of BCAT2 establishes resistance to ferroptosis in cancer cells^[119,121]. Ferroptosis inducers such as sulfasalazine, erastin, and sorafenib can inhibit BCAT2 transcription by activating the AMPK/SREBP1 signaling pathway, leading to reduced intracellular glutamate levels and GSH depletion^[119].

Conversely, impaired BCAA catabolism and the consequent intracellular accumulation of these amino acids actively promote ferroptosis. In cisplatin-induced chronic kidney disease models, miR-429-3p specifically inhibits key enzymes in the downstream BCAA catabolism pathway (e.g., ABAT, ALDH6A1) in renal proximal tubules, leading to abnormal intracellular BCAA accumulation^[121]. This BCAA buildup does not act via the glutamate pathway but rather directly drives ferroptosis by upregulating pro-ferroptotic genes ACSL4 and TFRC^[121]. Additionally, high concentrations of BCAAs can promote oxidative stress by activating the Akt-mTOR pathway, thereby damaging mitochondrial function^[120].

5.4 Alanine

Alanine and its isomers act as potent metabolic suppressors of ferroptosis. Their protective mechanisms are multifaceted, encompassing the maintenance of mitochondrial redox homeostasis and the epigenetic silencing of pro-ferroptotic signaling cascades.

Alanine metabolism precisely regulates ferroptosis sensitivity through multiple pathways. In neuroblastoma, the OGT/FOXC1 signaling axis drives high expression of glutamate pyruvate transaminase 2, which converts glutamate and pyruvate to alanine and α-KG. This conversion not only provides anaplerotic substrates for the TCA cycle but, more critically, optimizes intracellular redox balance, conferring resistance to ferroptosis in neuroblastoma cells^[122]. Consistently, endogenous alanine derivatives, such as γ-glutamyl-L-alanine, serve key defensive roles against oxidative damage, with their cellular depletion acting as a metabolic hallmark of heightened susceptibility to mitochondrial lipid peroxidation^[123]. Furthermore, alanine participates in epigenetic regulation. Hong et al. found that lactyltransferase AARS1 can induce cellular ferroptosis in diabetic nephropathy through lactylation modification, and β-alanine, as a competitive substrate, can effectively block this pathway, demonstrating significant clinical intervention potential^[124].

5.5 Sarcosine

Sarcosine exhibits context-dependent effects in regulating ferroptosis and oxidative stress. Multiple studies have shown that decreased sarcosine levels correlate with exacerbated oxidative stress and uncontrolled inflammation. Sarcosine supplementation can activate the macrophage GCN2 signaling axis to alleviate tissue damage, holding clinical significance in conditions like sarcopenia and diabetic immune stress^[125,126]. Conversely, Shan et al. identified sarcosine as a potent ferroptosis inducer. It triggers mitochondrial ROS burst via the PDK4/PDHA1 pathway and meanwhile downregulates the iron export protein SLC40A1 to accumulate intracellular iron, thereby enhancing chemosensitivity in lung adenocarcinoma^[127].

6. Challenges and Prospects

Although the central role of amino acid metabolism in regulating tumor ferroptosis is well-established (Figure 1 and Figure 2), translation from basic research to clinical therapies faces formidable challenges^[128]. Current clinical efforts targeting these pathways highlight both the therapeutic potential and the inherent complexities of this approach (Table 1).

Display Full Size

Figure 1. The regulatory landscape of amino acid metabolism in tumor ferroptosis. Created in BioRender.com. GCL: glutamate-cysteine ligase; GSS: glutathione synthetase; GPX4: glutathione peroxidase 4; GLS: glutaminase; BCAT: branched-chain amino acid aminotransferases; SHMT: serine hydroxymethyltransferase; GPT2: glutamate pyruvate transaminase 2; DHODH: dihydroorotate dehydrogenase; AhR: aryl hydrocarbon receptor; CoQ: coenzyme Q; THF: tetrahydrofolate; TCA: tricarboxylic acid; ROS: reactive oxygen species; BCAA: branched-chain amino acids.

Display Full Size

Figure 2. Context-dependent stress signaling pathways activated by specific amino acid deprivation. Created in BioRender.com. GCLC: glutamate-cysteine ligase catalytic subunit; GCN: general control nonderepressible; ATF: activating transcription factor; AMPK: AMP-activated protein kinase; GSH: glutathione.

A primary obstacle in the clinic is the high plasticity and compensatory capacity of tumor metabolic networks. For instance, pharmacological blockade of a single amino acid uptake pathway (e.g., System xc-) frequently triggers bypass mechanisms like the transsulfuration pathway, thereby diminishing the efficacy of monotherapies. Secondly, the regulation of ferroptosis by amino acids is highly context-dependent and spatiotemporally heterogeneous, which is dictated by the tissue of origin, specific genetic mutations, and the local TME. The same amino acid can exert opposite effects in different tumor types or even at different stages of the same tumor. Additionally, nutrient competition within the TME presents a non-negligible obstacle. Amino acid deprivation might induce ferroptosis in cancer cells while simultaneously impairing the function of infiltrating immune cells like T cells.

To overcome these translational barriers, future research must leverage single-cell and spatial omics technologies to comprehensively map amino acid metabolic profiles within the TME and identify robust biomarkers of ferroptosis susceptibility. Therapeutically, developing precise delivery systems, such as engineered bacteria or tumor-targeted nanocarriers, holds promise for maximizing ferroptosis and minimizing systemic side effects. Concurrently, combining amino acid deprivation therapies or inhibitors targeting non-canonical pathways (e.g., targeting DHODH or FSP1) with immune checkpoint inhibitors to reshape the metabolic state of the TME will be a crucial direction for overcoming drug resistance. Furthermore, the clinical value of non-invasive strategies like intermittent dietary intervention in ferroptosis modulation still requires validation through large-scale clinical trials.

7. Conclusion

Amino acid metabolism constitutes a core regulatory network of ferroptosis, which exhibits pathway dependency and circumstance-specific effects. Targeting key metabolic nodes can synergize with ferroptosis inducers to effectively kill tumor cells by disrupting redox homeostasis. Future studies should integrate spatiotemporal metabolic heterogeneity to develop precise strategies for clinical translation.

Authors contribution

Bi G, Zhan C: Conceptualization, funding acquisition, writing review & editing.

Shan G, Bian Y, Ren S: Data curation, writing-original draft, writing review & editing.

Conflicts of interest

The authors declare no conflicts of interest.

Ethical approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Availability of data and materials

Not applicable.

Funding

This work was supported by the National Natural Science Foundation of China (Grant No. 82403100 to Guoshu Bi and Grant No. 82473184 to Cheng Zhan), the Natural Science Foundation of Shanghai (Grant No. 24ZR1409900 to Guoshu Bi), the Fellowship from the China Postdoctoral Science Foundation (Grant No. 2024M760555 to Guoshu Bi), and the 18th Special Fund (In-Station) from China Postdoctoral Science Foundation (Grant No. 2025T180549 to Guoshu Bi).

Copyright

© The Author(s) 2026.