Despite sufficient oxygen availability, malignant cells frequently maintain continuous lactate extrusion via aerobic glycolysis. In neoplastic populations overexpressing SLC7A11, maintaining cystine uptake necessitates substantial glucose consumption to support reduced nicotinamide adenine dinucleotide phosphate (NADPH) biosynthesis via the pentose phosphate pathway. Although this adaptation sustains glutathione (GSH) synthesis, it inadvertently establishes a strict glucose dependency^[6]. Consequently, intense intercellular glucose competition within the TME limits NADPH regeneration rates. This microenvironmental deprivation, combined with metabolic acidosis, destabilizes redox equilibrium, facilitating the lipid peroxidation cascade when reactive oxygen species (ROS) accumulation surpasses antioxidant thresholds^[7]. Evidence indicates that restricting glucose transport or inhibiting key glycolytic enzymes compromises tumor cells’ redox buffering capacity, thereby lowering the threshold for ferroptosis induction^[8].

Lipid metabolic rewiring functions as a primary determinant of ferroptotic vulnerability. While neoplastic cells frequently upregulate de novo fatty acid synthesis, polyunsaturated fatty acids (PUFAs) (the principal substrates for oxidation) determine cellular ferroptosis sensitivity based on their integration rate into cellular phospholipids^[9]. Acyl-CoA synthetase long-chain family member 4 (ACSL4) regulates the esterification of PUFAs into membrane phospholipids, indicating that its relative expression alters the probability of ferroptotic lipid damage^[10]. Conversely, the incorporation of exogenous monounsaturated fatty acids (MUFAs) or the upregulation of stearoyl-CoA desaturase 1 (SCD1) modulates membrane lipidomic profiles, elevating the MUFA/PUFA ratio and establishing a biochemical resistance mechanism that impedes lipid peroxide propagation^[11].

Disruption of cellular iron homeostasis is a core driver and key permissive factor for ferroptosis. Tumor cells frequently exhibit elevated iron reliance, increasing iron influx via transferrin receptor (TfR1) upregulation while restricting ferritin-mediated sequestration. This imbalance generates an expanded intracellular labile iron pool (Fe²⁺)^[12]. The availability of free Fe²⁺ catalyzes lipid peroxidation via Fenton chemistry. In scenarios where endogenous antioxidant defenses are compromised, the accumulation of these oxidized lipids results in structural membrane destabilization, facilitating ferroptotic cell death^[13].

Ferroptosis represents a regulated mechanism distinct from apoptosis, necroptosis, and autophagy regarding morphological, biochemical, and genetic criteria^[14]. Its pathway relies on the accumulation of iron-dependent lipid peroxides exceeding cellular tolerance limits - a process subject to inhibition by specific iron chelators or radical-trapping antioxidants^[15]. Morphologically, it involves condensed mitochondrial membrane densities and diminished cristae; biochemically, it requires iron-dependent ROS propagation and specific oxidized phospholipids^[3]. The cellular response to this process is mediated by parallel antioxidant networks.

There are three core ferroptosis defense pathways within cells:

First, the System X_c^--GSH-Glutathione Peroxidase 4 (GPX4) axis, where the cystine/glutamate antiporter regulates cystine import for GSH biosynthesis. GPX4 utilizes GSH to reduce phospholipid hydroperoxides to corresponding lipid alcohols, functioning as the primary mechanism for maintaining membrane integrity^[16].

Second, the ferroptosis suppressor protein 1 (FSP1) - coenzyme Q10 (CoQ10) system, which operates independently in various membrane compartments. FSP1 interfaces with membrane-embedded CoQ10 to generate an extramitochondrial antioxidant rheostat, scavenging lipid peroxyl radicals^[17].

Third, the guanosine triphosphate (GTP) cyclohydrolase 1 (GCH1)-tetrahydrobiopterin (BH4) signaling node, which synthesizes endogenous antioxidants [BH4 and dihydrobiopterin (BH2)] and specifically shields di-polyunsaturated fatty acid (di-PUFA)-tailed phospholipids from oxidative degradation, thus modifying the lipidome^[18].

The dependency on these respective networks is highly lineage-specific and subject to microenvironmental pressure. This heterogeneity accounts for the differential sensitivities between neoplastic and immune populations under identical metabolic stress, highlighting the necessity for calibrated therapeutic targeting.

Importantly, the immunological consequence of ferroptosis depends strongly on the cellular compartment in which it occurs. Ferroptosis of tumor cells may promote immunogenic signaling or generate macrophage-modulating cues, whereas ferroptosis of macrophages, myeloid-derived suppressor cells (MDSCs), or effector lymphocytes can produce distinct and sometimes opposing immune consequences^[19]. Therefore, the cellular source of ferroptosis is specified below when interpreting its effects on the TME.

Initially, early-stage ferroptotic cancer cells release specific DAMPs capable of initiating immunogenic responses^[20]. However, this effect is frequently counteracted as surviving neoplastic populations develop adaptive resistance through SLC7A11 up-regulation, NRF2 activation, or lipidomic restructuring, thereby maintaining tumor viability under persistent therapeutic pressure^[21].

Additionally, the ferroptotic process generates secondary signaling cues. The release of oxidized phospholipids and metabolic intermediates like extracellular glutamate establishes a paracrine gradient that conditionally activates or suppresses adjacent leukocytes depending on localized concentrations^[22]. In the context of the TME, these chemical outputs may alter inflammatory, immunoregulatory, and tissue-remodeling programs in infiltrating TAMs in a concentration- and context-dependent manner, complicating the assumption that ferroptosis acts uniformly as a tumor-suppressive event.

The abundance and activity of core ferroptosis regulators, including GPX4 and ACSL4, may influence macrophage survival, oxidative-stress responses, and inflammatory functions. Their effects should be interpreted according to the molecular and functional readouts measured in each experimental model, rather than as evidence of conversion into fixed or mutually exclusive macrophage phenotypes.

Tumor cells undergoing ferroptotic disruption continually extrude a distinct sequence of DAMPs and oxidized lipid mediators. However, assuming these components function solely as immunological activators ignores their capacity to inadvertently drive immunosuppression depending on receptor interaction and microenvironmental conditions. In this section, ferroptosis occurs primarily in tumor cells, whereas TAMs act as recipient cells that sense ferroptosis-associated extracellular signals. Depending on the molecular cue and microenvironmental context, the resulting macrophage response may be immunostimulatory, immunosuppressive, or context-dependent.

Histotype-specific differences in lipid metabolism, glycolytic activity, antioxidant capacity, and macrophage composition may contribute to heterogeneous ferroptotic responses across tumor types. For example, Hepatocellular Carcinoma (HCC) and Pancreatic Ductal Adenocarcinoma (PDAC) exhibit distinct metabolic and immunometabolic features^[39,40]. However, these differences should not be interpreted as establishing a fixed hierarchy of ferroptosis sensitivity among cancer types. Rather, they provide a biological rationale for investigating whether tumor-specific metabolic states influence ferroptosis susceptibility and TAM responses. Direct comparative studies across histotypes are still required.

Concurrently, the derivation pattern of TAM populations further dictates mechanistic susceptibility. In Glioblastoma (GBM), the TAM compartment involves dual ontogeny: sessile brain-resident microglia integrating via Lysyl oxidase (LOX) signals, countered by peripheral monocytes recruited via tumor-derived CCN1^[2]. The differing hypoxic exposure histories and cytokine proximities associated with these differing ontogenies strictly restrict how each subset responds to induced metabolic stress. Similarly, breast carcinoma models suggest that spatially localized glycolytic variation correlates with regional differences in the density of macrophages expressing M2-associated markers^[41].

First, sex hormones and receptor signaling establish organ-specific baseline immune landscapes. In urological malignancies, cell-type-specific expression and function of androgen and estrogen receptors drive divergent immune microenvironment phenotypes, classifying bladder cancer as immune-hot and prostate cancer as immune-cold. These hormone-driven differences directly shape TAM polarization states and overall responsiveness to immunotherapy^[42]. Such baseline immune variation may further modify the activity of the metabolism-ferroptosis-TAM axis across organ sites.

Second, crosstalk between tumor metabolic reprogramming and CD8⁺ T-cell function contributes to tissue- and tumor-specific immune profiles. Tumor metabolic rewiring can impair CD8⁺ T-cell effector function through nutrient competition, the accumulation of immunosuppressive metabolites, and metabolic reprogramming of stromal and myeloid populations, including macrophages. Accordingly, targeting selected tumor metabolic adaptations may help alleviate immunosuppression and restore CD8⁺ T-cell activity^[43]. The relative contribution of individual metabolic pathways varies across tumor types; for example, lipid metabolic abnormalities have been implicated in hepatocellular carcinoma, whereas glycolytic alterations and nutrient competition have been described in pancreatic ductal adenocarcinoma^[43]. These observations suggest that tissue-specific metabolic-immune environments may also influence TAM functional states and, potentially, the local response to ferroptosis-associated stress.

Third, therapeutic regimens can dynamically remodel intratumoral immune-cell subsets in a context-dependent manner. In non-small cell lung cancer, neoadjuvant chemoimmunotherapy was associated with coordinated changes in bystander CD8⁺ and conventional CD4⁺ T-cell populations, some of which correlated with major pathological response^[44]. Changes in HIF-1α expression also showed an association with changes in conventional CD4⁺ T-cell abundance, although the underlying mechanism remains to be established^[44]. Such therapy-induced changes in the T cell compartment alter local cytokine signaling and TAM-T cell crosstalk, which further adjusts the output of the core metabolism-ferroptosis-TAM regulatory axis.

Accordingly, the efficacy of generalized pro-ferroptotic strategies is unlikely to be uniform across solid tumors. Future clinical translation will require prospective evaluation of tumor metabolic features, ferroptosis susceptibility, and TAM functional-state composition to determine whether these variables can support patient stratification. At present, such stratification remains investigational rather than clinically established. Clinical implementation therefore requires pretreatment stratification that considers the tumor-specific lipidomic profile together with the local composition and spatial distribution of TAM functional states. Calibrating ferroptosis induction alongside agents targeting discrete TAM signaling parameters remains a rigorous prerequisite for circumventing the functional limitations of existing monotherapies and securing durable combinatory responses^[45] [Table 1].

Current ferroptosis inducers (FINs) typically utilize three mechanisms: (1) System X_c^- antagonism, (2) direct GPX4 inhibition, and (3) indirect orchestrators (e.g., Fenton reaction manipulation). While in vitro applications demonstrate efficient cytolysis, their clinical translation is challenged by critical limitations^[19]. First, FINs display substantive pharmacokinetic restrictions, specifically regarding the metabolic instability and aqueous solubility limitations of agents such as erastin^[19]; Second, the non-specific induction of lipid peroxidation corresponds to documented systemic toxicities in the central nervous system and renal parenchyma^[47]; Third, neoplastic adaptation to FIN exposure frequently involves upregulation of compensatory mechanisms including the FSP1-CoQ10 system or SCD1 overexpression combined with reduced PUFA integration, progressively reducing susceptibility over successive treatment cycles^[17,19].

Against this backdrop, natural product-derived ferroptosis modulators have attracted preclinical interest, although their pharmacological specificity and safety advantages remain unproven. For example, aqueous-soluble components of sporoderm-removed Ganoderma lucidum spore powder (A-GSP) promoted ferroptosis in oral squamous cell carcinoma cells, as indicated by Fe²⁺ accumulation, GSH depletion, increased lipid peroxidation, reduced GPX4 expression, and rescue by ferroptosis inhibitors, and also suppressed xenograft tumor growth without observable adverse reactions in mice^[48]. Nevertheless, this complex extract remains far from clinically actionable because its active constituents and direct molecular targets are undefined, formal target-engagement and systemic pharmacokinetic/pharmacodynamic data are lacking, and its safety in immune cells and immunocompetent models has not been established.

The simultaneous application of FINs with immune checkpoint inhibitors (ICIs) seeks to link adaptive immunity activation with cytolytic ferroptosis. Mechanistically, ICI-activated CD8⁺ T cells release IFN-γ, which suppresses SLC3A2 and solute carrier family 7 member 11 (SLC7A11) expression in tumor cells, thereby increasing tumor-cell lipid peroxidation and ferroptotic susceptibility^[49]. In this setting, CD8⁺ T cells act as the upstream immune effector, whereas tumor cells are the principal ferroptotic compartment, resulting in an antitumor effect. Preclinical melanoma models demonstrate numerical tumor suppression compared to monotherapies. However, this synergy assumes that ferroptosis-derived DAMPs act preferentially to sustain anti-tumor immunity without promoting compensatory immunoregulatory TAM programs^[49].

Beyond combination with exogenous immune checkpoint inhibitors, ferroptosis-inducing platforms may themselves modulate tumor-cell immune checkpoint expression. A hydrogen peroxide-responsive gallium-releasing nanoplatform, Ga@MnO₂@Alb, was reported to induce tumor-cell ferroptosis while suppressing CD47 and programmed cell death ligand 1 (PD-L1) expression through a mitochondria-AMP-activated protein kinase (AMPK)-c-MYC signaling axis^[50]. Downregulation of the innate immune checkpoint CD47 enhanced macrophage-mediated phagocytosis of tumor cells, whereas reduced PD-L1 expression increased T-cell cytotoxicity. These findings suggest that ferroptosis-associated immune checkpoint modulation may reshape the tumor immune microenvironment by simultaneously relieving macrophage- and T-cell-directed inhibitory signals. However, the reported macrophage effect primarily involved enhanced phagocytic function rather than a directly demonstrated transition toward a stable M1-like or M2-like state. Whether such checkpoint modulation induces durable TAM functional-state remodeling remains to be established. Checkpoint-associated immune escape is further complicated by intertumoral and spatial heterogeneity in checkpoint expression. In melanoma, for example, PD-L1 expression has shown associations with response to programmed cell death protein 1 (PD-1) blockade under selected assessment conditions, but its prognostic and predictive value remains controversial^[51]. Therefore, PD-L1 expression alone is unlikely to capture the broader immune context determining the efficacy of ferroptosis-checkpoint combinations.

Furthermore, within untreated niches, TAMs displaying M2-associated immunoregulatory features can transfer metabolites such as taurine, which subsequently limits oxidative collapse via liver X receptor alpha (LXRα)/SCD1 activation^[46]. Establishing an effective FIN/ICI response requires specific disruption of this pathway to shift macrophage activity toward pro-inflammatory and tumor-restrictive programs rather than toward a discrete M1 phenotype^[52]. By contrast, ferroptotic stress occurring within activated CD8⁺ T cells or Natural killer (NK) cells themselves can impair their survival and cytotoxic function, producing an immunosuppressive consequence rather than enhancing tumor-cell killing^[22]. Current developmental strategies propose targeted nanomedicine platforms to isolate these variables^[33,53]. Early evidence involving attenuated tumoral SLC3A2 expression in patients responsive to nivolumab provides initial correlations^[49], yet achieving reliable clinical modulation remains contingent upon rigorous patient stratification.

Recent studies have explored two principal approaches to modulating TAM functional states through ferroptosis. The first involves manipulating ferroptosis-related pathways in tumor cells or macrophages. In nasopharyngeal carcinoma models, ACSL4 overexpression or erastin treatment increased lipid peroxidation and tumor-cell ferroptosis and was accompanied by a shift from M2-associated toward M1-associated macrophage features; ferrostatin-1 reversed these effects, supporting a ferroptosis-dependent component^[29]. In gastric cancer, tumor-derived exosomal calcium/calmodulin-dependent protein kinase II alpha (CaMK2A) promoted zinc finger DHHC-type palmitoyltransferase 3 (ZDHHC3)-dependent GPX4 palmitoylation and stabilization, thereby suppressing macrophage ferroptosis and favoring an M2-like immunoregulatory program. Conversely, macrophage-specific GPX4 deletion restrained tumor growth and enhanced the response to PD-1/PD-L1 blockade in murine models^[54]. The second approach employs multifunctional delivery systems to coordinate tumor-cell ferroptosis with TAM remodeling. For example, an Fe³⁺- and erastin-containing nanomedicine induced ferroptotic stress, promoted M1-associated TAM changes, reduced fibroblast activation and collagen deposition, and improved nanoparticle penetration in pancreatic cancer models^[55].

Collectively, these approaches may simultaneously enhance tumor-cell killing, alleviate myeloid immunosuppression, improve intratumoral drug penetration, and increase responsiveness to immunotherapy. Nevertheless, most of the available evidence remains preclinical. Moreover, multifunctional interventions often alter iron availability, oxidative stress, stromal organization, and immune signaling concurrently, making it difficult to determine whether the observed TAM remodeling is directly attributable to ferroptosis. Excessive ferroptosis may also damage beneficial macrophage subsets or other immune cells, whereas compensatory antioxidant responses, systemic toxicity, and formulation complexity may limit therapeutic efficacy and clinical translation. Importantly, ferroptosis does not uniformly promote antitumor macrophage activity. In KRAS^G12D pancreatic cancer models, ferroptotic tumor cells released exosomal KRAS^G12D, which was internalized by macrophages through RAGE/AGER and induced a STAT3-fatty acid oxidation-dependent M2-like pro-tumor program^[35]. Thus, the therapeutic outcome of ferroptosis-guided TAM modulation is likely to depend on the targeted cell population, tumor genotype, ferroptotic signals released, treatment intensity, and local immune context.

Translating these combined paradigms entails confronting the aforementioned dual-effect paradox: while inducing ferroptosis in tumor cells to exert tumoricidal activity, such approaches must spare tumor-infiltrating immune cells within the TME to preserve overall antitumor immunity^[36]. Variations in baseline tumor lipid composition signify that lipid-enriched models (e.g., HCC) may demonstrate sensitivity, while tumors lacking sufficient PUFA densities (e.g., specific gliomas) inherently resist treatment due to substrate limitation alone^[56]. Resolving these issues may require targeted delivery strategies together with candidate biomarkers reflecting ferroptosis susceptibility, such as ferritin-related iron status, 4-hydroxynonenal, or the ACSL4/GPX4 expression balance^[56]. However, these markers should currently be regarded as hypothesis-generating or exploratory indicators rather than validated clinical stratification tools, and their predictive value requires prospective validation in well-defined patient cohorts. At present, the reliance on these proposed biomarkers requires rigorous multivariable prospective validation before stratifying clinical cohorts.

As a core component of the combined ferroptosis-immunotherapy paradigms, TAM polarization-targeted strategies also face translational constraints in safety and efficacy. Several focused improvements merit further development. Future regimens can shift from broad M1/M2 modulation toward subtype-specific functional reprogramming to improve therapeutic precision^[5]. Pairing ferroptosis inducers with inhibitors of TAM fatty acid oxidation may block paracrine M2-like polarization and enhance combinatorial efficacy^[35]. TAM-targeted nano-delivery systems also represent a viable direction to reduce off-target immune impairment.