Tumor cell ferroptosis in antitumor immunity and immunotherapy efficacy

Abstract

Cancer immunotherapy, particularly immune checkpoint blockade (ICB), has represented a major advancement for the treatment of various cancers. Though ICB has been an effective tool in the clinical management of patients, a significant subset of individuals do not respond clinically. This observation has motivated deeper investigations into the mechanisms of efficacy versus resistance to ICB. Of these, tumor cell-intrinsic mechanisms, such as the activation of immune regulatory oncogene pathways or immune selection for antigen-loss variants, have been a promising focus of study. Recent evidence has emerged that ferroptosis, a tumor cell-intrinsic cell death pathway, plays an important role in CD8⁺ T cell-mediated tumor cell killing, and that some resistant tumors have lost sensitivity to this process. Ferroptosis is an iron-dependent mechanism of cell death that is mediated by the peroxidation of lipids. Tumor cell-intrinsic regulators govern ferroptosis sensitivity and, in turn, can influence the success of immune-mediated tumor control. Here, we discuss the growing understanding of regulation of ferroptosis and examine the complex interplay between tumor cell-intrinsic ferroptosis pathways and the host immune response, with an emphasis on how ferroptosis might be leveraged to potentiate immunotherapy efficacy.

Introduction

Immunotherapy has transformed cancer treatment. Of these, T cell-based strategies such as immune checkpoint blockade (ICB), which restores the efficacy of host antitumor T cells, or the use of adoptively transferred genetically engineered T cells (eg, chimeric antigen receptor (CAR) T cells and T-cell receptor (TCR)-engineered T cells) have reached success to the point of Food and Drug Administration approval in multiple cancer types.

However, despite these successes, many patients exhibit primary resistance, and a subset develops secondary resistance. CAR-T therapy is highly effective in hematologic malignancies but shows limited efficacy in solid tumors, highlighting the need to better understand mechanisms driving therapeutic response and resistance.¹

The cancer-immunity cycle illustrates key steps governing an anti-tumor immune response.² Tumor antigens are captured by dendritic cells (DCs), whose activation—partly via the cGAS-STING pathway—induces type I interferons and cytokines. Activated DCs prime tumor-specific T cells in lymph nodes, which differentiate into effector and memory cells and migrate to tumors guided by chemokines such as CXCL9 and CXCL10 (C-X-C motif chemokine ligands 9 and 10). In the tumor microenvironment (TME), CD8⁺ T cells recognize tumor cells via TCR-major histocompatibility complex (MHC) I interactions and can mediate cytotoxicity.³

Importantly, many solid tumors have supported the majority of this process up to the point of containing activated tumor antigen-specific T cells in the TME, but without the final step of tumor cell killing. This phenotype has been termed the “T cell-inflamed tumor microenvironment” and is at least partially explained by the presence of inhibitory processes that keep those T cells in check. This includes the programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) axis, the blockade of which has led to a major cancer immunotherapy approach, with activity in a broad range of human cancer types. But it has become clear that some tumor cells escape immune attack through additional tumor cell-intrinsic mechanisms.

Mechanisms of T cell-mediated tumor cell death

Whether occurring endogenously after checkpoint blockade or via adoptive transfer of tumor-specific T cells, the final effector step is tumor cell killing. Prior work has described two major pathways of T cell-mediated killing: the perforin-dependent cytotoxic granule pathway and the engagement of death receptors that trigger target cell apoptosis.

In the cytotoxic granule pathway, perforin forms pores in tumor cell membranes, allowing granzymes to induce caspase-dependent apoptosis with minimal inflammation. Activated CD8⁺ T cells also express death ligands such as FasL and TRAIL, which trigger apoptotic cascades on binding their respective receptors.⁴

Though these apoptotic pathways are well-established, they do not fully account for all forms of tumor cell death driven by CD8⁺ T cells, and there is significant evidence suggesting alternative pathways of T cell-based tumor cell killing. Adoptive transfer of T cells lacking perforin, Fas ligand (FasL), or tumor necrosis factor-α into tumor-bearing mice can still mediate tumor rejection in vivo.⁵ Additionally, in a genetically engineered melanoma mouse model, tumor eradication has been found to require 5 days of sustained contact between T cells and tumor cells,⁶ which would not be consistent with a rapid apoptosis mechanism. In contrast, interferon (IFN)-γ-deficient T cells very poorly control solid tumors in vivo,⁵ highlighting a critical role for IFN-γ in contributing to tumor cell death. Supporting this notion, in patients with melanoma treated with checkpoint blockade, acquired resistance has been observed in patients whose tumors have mutations in the IFN-γ signaling pathway.¹

Aside from apoptosis, recent studies have identified ferroptosis as a potential alternative pathway of T cell-induced tumor cell death.⁷ Ferroptosis is a regulated, iron-dependent form of non-apoptotic cell death marked by the accumulation of lipid peroxides and oxidative damage to cell membranes.^{8 9} This pathway has emerged as an important mechanism through which CD8⁺ T cells, particularly via IFN-γ signaling, can sensitize tumor cells to oxidative stress, offering new insight into how T cells can eliminate cancer cells beyond classical apoptotic routes.

Mechanisms involved in ferroptosis

Ferroptosis, a distinct form of cell death, was first discovered and defined in 2012.¹⁰ It was observed that erastin, originally developed as a strategy to treat RAS-transformed cells, induced cell death which was different from apoptosis, pyroptosis, or necrosis, but via a novel, iron-dependent mechanism. Ferroptosis occurs when phospholipids undergo oxidative damage.^{11 12} Some metabolic pathways can prevent ferroptosis by blocking the accumulation of oxidized phospholipids, thereby protecting cells. Conversely, metabolic processes that promote the production of polyunsaturated phospholipids and reactive oxygen species can drive ferroptosis by facilitating lipid peroxidation. To date, lipid metabolism, reactive oxygen species (ROS) biology, and iron regulation have emerged as the three major metabolic processes that influence ferroptosis.¹³

Lipid metabolism

Polyunsaturated fatty acids (PUFAs) are long-chain fatty acids with multiple double bonds, making them highly susceptible to oxidative damage from ROS. PUFA oxidation drives lipid peroxidation, a key trigger of ferroptosis. However, free PUFAs alone are insufficient.^{14 15} The incorporation of intracellular PUFAs into specific phospholipid pools within the cell membrane is essential for PUFAs to promote ferroptosis.^{16 17} Cell membranes contain a wide variety of phospholipids, including phosphatidylethanolamine (PE), phosphatidylcholine (PC), phosphatidylserine (PS), and phosphatidylinositol (PI).^{18 19}

Although peroxidation occurs across various PUFA-containing phospholipid classes, peroxidized molecular species are predominantly found in PUFA PE, with PUFA PC and other PUFA phospholipids (such as PS, PI, cardiolipin, and ether-linked lipids) also contributing substantially. PUFA PEs—especially those carrying arachidonic acid (AA) (20:4) or adrenic acid (AdA) (22:4)—are key oxidation substrates in ferroptosis.¹⁷ Samovich et al²⁰ demonstrated that 15 lipoxygenase strongly oxidizes di PUFA PE species, producing hydroperoxy-PEs that appear early in ferroptosis.²⁰ Lipidomic analysis in GPX4-inhibited (RSL3-treated) cells revealed significant accumulation of oxidized PE, including ether-linked PE, more so than PC or other lipid classes.²¹

Genes regulating PUFA-phospholipid abundance modulate ferroptotic sensitivity. The acyl-CoA synthetase long-chain family member 4 (ACSL4)– lysophosphatidylcholine acyltransferase 3 (LPCAT3) axis is central: ACSL4 activates AA, AdA, and other PUFAs into coenzyme A (CoA) derivatives, which LPCAT3 incorporates into PUFA-containing phospholipids. Loss of ACSL4 or LPCAT3 reduces peroxidation-prone PUFA-phospholipids, increasing resistance to ferroptosis.^{22 23}

Peroxisomal lipid metabolism also influences ferroptosis via ether phospholipids (ePLs), especially plasmalogens. Depletion of peroxisomal genes (PEX10, PEX3, PEX12) decreases polyunsaturated ePLs, conferring resistance to GPX4 inhibition-induced ferroptosis.^{24 25} Similarly, knockdown of Decr2 in B16 melanoma cells reduces ePLs and promotes resistance, highlighting the critical role of PUFA-phospholipid metabolism in ferroptotic cell death.²⁶

Redox homeostasis

Antioxidants play a critical role in counteracting free radicals and neutralizing oxidative stress. Central to this protective mechanism is the glutathione (GSH)/glutathione peroxidase 4 (GPX4) axis. System Xc–, composed of the subunits solute carrier family 7 member 11 (SLC7A11) and solute carrier family 3 member 2 (SLC3A2), facilitates the import of cystine, a precursor essential for GSH synthesis. GPX4 then uses GSH as a reducing cofactor to convert cytotoxic lipid hydroperoxides (L-OOH) into non-toxic lipid alcohols, during which GSH is oxidized to glutathione disulfide.27,29 Cancer cells deficient in SLC7A11 or SLC3A2 exhibit impaired cystine uptake, GSH depletion, and subsequent GPX4 inactivation, rendering them highly sensitive to ferroptosis.³⁰ Similarly, genetic or pharmacological inhibition of GPX4 sensitizes tumor cells to ferroptotic cell death.³¹ This vulnerability is evident in multiple cancer types, including triple-negative breast cancer,^{32 33} colorectal,³⁴ ovarian,³⁵ and drug-resistant cancers,³⁶ where GPX4 knockdown or knockout results in lipid ROS accumulation and iron overload, which are hallmarks of ferroptosis. Small molecules such as erastin disrupt this protective axis by inhibiting system Xc–, while compounds like RSL3 directly inhibit GPX4, promoting ROS accumulation and triggering ferroptosis.^{37 38}

In addition to the cysteine–GSH axis, the mevalonate pathway also plays a critical role in GPX4 maturation during ferroptosis. A key feature of GPX4 is its reliance on selenocysteine, an amino acid essential for its enzymatic activity in detoxifying lipid peroxides. The incorporation of selenocysteine into GPX4 depends on a specialized transfer RNA (tRNA), which must be modified by isopentenyl pyrophosphate—a product of the mevalonate pathway.^{38 39}

Beyond the canonical GPX4 axis, alternative ROS-detoxifying systems have been identified. These include the FSP1–CoQ10–NADPH pathway at the plasma membrane, the mitochondrial DHODH–CoQ10 system, and the GCH1–BH4 axis—all of which act independently of GPX4 to suppress lipid peroxidation and ferroptosis. These pathways provide redundant layers of protection against ROS-induced damage, especially in cells with compromised GPX4 activity.^{40 41}

Iron metabolism

Cellular pathways for iron uptake, storage, and export tightly regulate the labile iron pool. Excess free iron catalyzes the generation of ROS via the Fenton reaction, triggering lipid peroxidation and ultimately leading to ferroptotic cell death.⁴² Key regulators include transferrin receptor 1-mediated iron import, ferritin-mediated storage, and ferroportin-mediated export.⁴³ Dysregulation of ferritinophagy, where ferritin is transported to lysosomes and degraded—mediated by NCOA4—releases stored iron into the cell’s labile iron pool, increasing free iron levels and promoting ferroptosis.^{44 45}

Tumor cell-intrinsic regulators of ferroptosis and antitumor immunity

Tumor cell-intrinsic resistance to ferroptosis is emerging as a key mechanism of immune evasion. Tumor cells that escape ferroptosis have been shown to resist ICB therapy. Suppressing negative regulators of ferroptosis can therefore dramatically improve immunotherapy efficacy in model systems (figure 1).

Figure 1. Core molecular pathways regulating ferroptosis induction and resistance overview of ferroptosis execution and key regulatory nodes. Extracellular Fe3+ binds to transferrin and is internalized via the TFRC. Intracellular Fe3+ is then reduced to Fe2+, which can generate ROS and drive lipid peroxidation. Cystine is imported via the SLC7A11/SLC3A2 antiporter, then reduced to cysteine, a crucial precursor for GSH synthesis. GSH is a critical cofactor for GPX4, which reduces toxic L-OOH (eg, PUFA-PEs) to inert lipid alcohols, thereby preventing ferroptosis. This antioxidant pathway is negatively regulated by CHAC1 (which degrades GSH) and modulated by BIN1. ACSL4 and DECR2 play roles in the synthesis of susceptible phospholipid species. When GSH is depleted or GPX4 activity is compromised, L-OOH accumulate, leading to ferroptosis. Key transcriptional regulators such as NRF2 (which upregulates genes like SLC3A2, GPX4, and FTH to enhance ferroptosis resistance) and IRF8/P53 (which suppresses SLC7A11 expression, promoting ferroptosis) are also highlighted. The receptor TYRO3 has been implicated in modulating ferroptosis. Red arrows indicate promotion or activation, while blunt-ended lines indicate inhibition. AA, arachidonic acid; AdA, adrenic acid; DECR2, 2,4-dienoyl-CoA reductase 2; GPX4, glutathione peroxidase 4; GSH, glutathione; L-OOH, lipid hydroperoxides; PE, phosphatidylethanolamine; PUFA, polyunsaturated fatty acids; ROS, reactive oxygen species; SLC7A11, solute carrier family 7 member 11; TFRC, transferrin receptor.

Deletion of tumor cell-intrinsic ferroptosis genes promotes resistance to immunotherapy

ACSL4, which stands for Acyl-CoA synthetase long chain family member 4, contributes to the synthesis of cellular PUFA lipids, especially influencing the incorporation of AA and AdA into membrane phospholipids. Knockout of ACSL4 in cells rendered them resistant to drug-induced ferroptosis. Notably, cell death by ferroptosis induced by CD8⁺ T cell-secreted IFN-γ in combination with AA was reduced in ACSL4 knockout cells. ACSL4 was found to be upregulated in response to IFN-γ stimulation through the STAT1-IRF1 pathway. Clinically, tumors exhibiting higher ACSL4 expression showed increased infiltration by CD8⁺ T cells, stronger immune activity signatures, and better responses to ICB therapy. ACSL4-deficient Yumm5.2, MC38, and B16F10 tumors grew more aggressively in immune-competent mice and had fewer CD8⁺ T cells in the TME.⁴⁶ This observation suggests a connection between ferroptosis within tumor cells and enhanced immune activation.

Our recent work has identified Decr2 (2,4-dienoyl-CoA reductase 2), a peroxisomal enzyme involved in the metabolism of PUFAs, as a regulator of tumor cell ferroptosis and a critical determinant of ICB efficacy. A genome-wide CRISPR screen identified DECR2, which, when deleted, resulted in diminished tumor cell killing by CD8⁺ T cells. It was found that DECR2 knockout (or knockdown) tumor cells showed reduced synthesis of PUFA-ePLs, which are indispensable substrates for lipid peroxidation during ferroptosis. Knockdown or genetic ablation of DECR2 led to a marked reduction in ferroptosis, as evidenced by decreased lipid reactive oxygen species and impaired execution of ferroptotic cell death. Functionally, DECR2 deficiency rendered tumors resistant to anti-PD-L1 therapy in vivo in mice. Interestingly, DECR2-deficient tumors showed markedly decreased expansion of tumor-infiltrating CD8⁺ T lymphocytes following anti-PD-L1 therapy, suggesting that tumor cell ferroptosis contributes to driving immune response reinvigoration. Consistent with these mouse model data, transcriptomic analyses from human melanoma cohorts revealed that higher DECR2 expression correlated with improved clinical responses to anti-PD-1 therapy, whereas deep and shallow deletions of the DECR2 gene were associated with poorer patient outcomes treated with anti-cytotoxic T-lymphocyte associated protein 4. Collectively, these findings position DECR2 as a key metabolic checkpoint that modulates ferroptosis susceptibility and shapes the tumor immune microenvironment, thereby influencing the therapeutic success of ICB.²⁶

A similar trend has been observed with BIN1. The loss of BIN1 increases GSH levels in tumor cells, leading to resistance to ferroptosis. Knockout of BIN1 in tumor cells resulted in aggressive tumor growth and reduced CD8⁺ T-cell infiltration in mouse models. Consistent with preclinical data, low BIN1 expression was associated with advanced clinical disease stages and decreased CD8⁺ T-cell infiltration.⁴⁷ This observation further supports the notion that genes regulating ferroptosis are crucial for the host immune response.

An additional factor, IRF8, can induce ferroptosis cell death in tumor cells by regulating p53 expression. Tumor cells lacking IRF8 were more resistant to RSL3-induced ferroptosis. More importantly, cells expressing a dominant-negative IRF8 mutant were less sensitive to antigen-specific CD8⁺ T cell-induced killing, due to their decreased sensitivity to T cell-induced ferroptosis. Consistent with in vitro results, IRF8-deficient tumors grew more aggressively in immunocompetent mice. Restoring IRF8 expression in these tumors reversed the phenotype and led to diminished tumor growth in mice. Clinically, the level of IRF8 expression in melanoma cancer cells correlated with the outcomes of immunotherapy. Patients with melanoma who responded to treatment showed higher IRF8 expression in tumors compared with non-responders.⁴⁸

Recent studies have identified a glutathione-degrading enzyme, CHAC1, as another critical mediator of ferroptosis and antitumor immunity. The loss or downregulation of CHAC1 in tumor cells led to the accumulation of intracellular glutathione, thereby sustaining GPX4 activity and preventing lipid peroxidation and ferroptotic cell death. CHAC1-deficient B16F10 cells were resistant to CD8⁺ T cell-mediated killing in vitro. In vivo, CHAC1-deficient B16F10 tumors exhibited resistance to ICB, characterized by reduced CD8⁺ T-cell infiltration, impaired IFN-γ-induced oxidative stress, and a shift toward an immunosuppressive TME. These findings suggest that CHAC1 is an important link to T cell-driven ferroptosis, and that its loss represents a novel mechanism of immune evasion. In clinical samples, higher CHAC1 expression in tumors correlated with improved patient survival and a greater likelihood of response to checkpoint blockade immunotherapy.⁴⁹

Ferroptosis and antigen presentation

Emerging data suggest that ferroptosis may impact tumor cell immunogenicity by modulating the release and presentation of antigens. Early-stage ferroptotic cell death has been shown to augment immune responses both in vitro and in vivo. For example, cancer cells were treated with RSL3 for 1 hour, and the ferroptosis-dying cells were subsequently engulfed by bone marrow-derived DCs. Maturation markers such as CD80, CD86, and MHC II on the surface of CD11c⁺ cells were upregulated. Moreover, early ferroptotic MCA205 cells (after 3 hours of RSL3 treatment) were used to immunize mice. These mice showed partial protection against tumor rechallenge, consistent with increased immune priming. This observed immunogenicity was partly attributed to the release of danger-associated molecular patterns (DAMPs), including ATP and high-mobility group box 1, which can promote the activation of DCs (figure 2). These preclinical findings suggest that harnessing ferroptosis could have clinical potential to enhance antitumor immunity or improve the efficacy of immunotherapies.⁵⁰

Figure 2. CD8⁺ T cell-mediated tumor ferroptosis and its impact on the immune environment. CD8⁺ T cells release IFN-γ, which induces cancer cells into ferroptosis by lowering antioxidant defenses (Slc7a11/GPX4) and increasing fatty acid oxidation (Acsl4/Decr2). The dying cell then releases signals (DAMPs and exosomes) that activate dendritic cells and influence macrophage state. DC, dendritic cell; DAMP, danger-associated molecular pattern; DECR2, 2,4-dienoyl-CoA reductase 2; ePL, ether phospholipid; GPX4, glutathione peroxidase 4; IFN, interferon; PUFA, polyunsaturated fatty acid.

In our own work with DECR2-deficient tumors, we observed impaired expansion of tumor-infiltrating lymphocytes (TILs) following anti-PD-L1 treatment compared with control tumors, consistent with a feed-forward mechanism reinvigorating the host immune response when ferroptosis is enabled.²⁶

However, some conflicting evidence exists. Other studies have reported that, despite DAMP release, ferroptotic cancer cells did not display immunogenic cell death (ICD). In that study, vaccination of mice with ferroptotic cancer cells failed to protect against tumor growth.⁵¹ These discrepancies suggest the possibility of a context-dependent nature of ferroptosis and immunogenicity. Whether ferroptosis-induced cell death strictly qualifies as ICD remains a subject of ongoing debate. It is not yet clear whether immune activation is a direct consequence of ferroptosis—specifically through lipid peroxidation and iron-dependent mechanisms—versus a combination of death processes. Nevertheless, several studies have shown that ferroptosis-sensitive cells can effectively prime T-cell responses, potentially linking ferroptotic signaling in tumor cells and antitumor immunity. IFN-γ secreted by activated CD8⁺ T cells can downregulate SLC7A11 and SLC3A2, which are key anti-ferroptosis genes that affect cystine uptake and glutathione synthesis, leading to decreased GPX4 activity and increased lipid peroxidation.7 Further, IFN-γ signaling upregulates lipid metabolism genes, including ACSL4 and DECR2, thereby further promoting ferroptotic cell death.^{26 46} These findings suggest a positive immune feedback loop where IFN-γ enhances tumor ferroptosis, which in turn amplifies CD8⁺ T-cell activation to catalyze additional tumor cell killing (figure 2).

Suppressing negative regulators of ferroptosis can improve immunotherapy efficacy

In preclinical models, combining cyst(e)ine deprivation (via cyst(e)inase) with ICB synergistically enhanced CD8⁺ T cell-mediated tumor clearance through ferroptosis induction. Moreover, tumors lacking SLC7A11 exhibited significantly increased sensitivity to anti-PD-L1 therapy and combination treatments. In B16F10 melanoma models treated with radiotherapy plus anti-PD-L1 therapy, SLC7A11-deficient tumors achieved complete long-term regression in 56% of mice, and cured mice were resistant to tumor rechallenge, indicating durable immunologic memory.⁵²

A study by Jiang et al found that tumors with high TYRO3 levels showed significant resistance to ICB in both mouse models and human patients treated with anti-PD-1/PD-L1 therapy. At a molecular level, TYRO3 overexpression increased the expression of genes that prevent ferroptosis, such as SLC7A11, SLC3A2, and GPX4. At the same time, it decreased the expression of genes that promote ferroptosis, like Slc5a1 and Tfrc. These changes suggest that TYRO3 helps tumor cells survive by inhibiting their ability to undergo ferroptotic death. Further studies confirmed that TYRO3 suppressed ferroptotic death in response to anti-PD-L1 therapy. This study also revealed that tumors with high TYRO3 expression had an altered immune environment, specifically a reduced ratio of M1 to M2 macrophages. Therapeutically, TYRO3 inhibitor LDC1267 triggered ferroptotic tumor cell death, and combining this inhibitor with anti-PD-1 therapy in mouse models augmented therapeutic efficacy. These results suggest that targeting TYRO3 could be a way to overcome resistance to immunotherapy by making tumor cells more susceptible to ferroptosis.⁵⁰

Collectively, these findings indicate a pivotal role of ferroptosis regulators in shaping tumor immune responses and determining the efficacy of immune checkpoint therapies. Targeting ferroptosis pathways holds promise for overcoming resistance and improving the success of cancer immunotherapy.

Potential roles for ferroptosis in immune cells

Ferroptosis in immune cells within the TME may also influence the outcome of tumor immunotherapy (figure 3). Excessive ferroptosis in CD8⁺ T cells, DCs, or natural killer (NK) cells would impair their viability, weakening antitumor immunity. Conversely, controlled induction of ferroptosis in immunosuppressive cells, such as regulatory T cell (Treg) cells and tumor-associated macrophages, could enhance therapeutic efficacy. Any interventions aiming to promote tumor cell ferroptosis in vivo must take into account potential effects on immune cells as well.

Figure 3. Effects of ferroptosis on immune cell function in the TME. Ferroptosis, driven by iron, metabolic changes, or oxidative stress, can reshape the tumor microenvironment by altering immune cell activity indirectly. However, ferroptosis also can occur within immune cells, altering their function or survival. NK, natural killer; TME, tumor microenvironment; Treg, regulatory T cell.

CD8 T cells

CD8⁺ T cells appear to be vulnerable to ferroptosis. T cell-specific GPX4 knockout mice had fewer T cells in the periphery, due to reduced survival and expansion. T-cell population expansion was diminished after infections with viruses and parasites. The decrease in peripheral T-cell numbers was attributed to the increased sensitivity of GPX4-deficient T cells to ferroptotic cell death, leading to diminished survival.⁵³ In the lipid-rich TME, CD8⁺ T cells were shown to be vulnerable to ferroptotic cell death. Tumor-infiltrating CD8⁺ T cells (TILs) demonstrate elevated lipid uptake and storage, including increased accumulation of oxidized low-density lipoprotein (OxLDL). Exhausted CD8⁺ TILs showed increased CD36 expression, which mediates OxLDL uptake. Depletion of CD36 can enhance CD8⁺ T-cell function by reducing ferroptotic cell death, thereby contributing to decreased tumor growth.^{54 55} Although generating ferroptosis-resistant CD8⁺ T cells seems a promising strategy, the ferroptosis-related genes may have functions beyond regulating ferroptosis, potentially complicating therapeutic outcomes. For example, overexpression of ferroptosis suppressor genes such as FSP1 and GPX4 can decrease CD8⁺ T-cell ferroptosis without affecting their function. In contrast, deletion of the pro-ferroptotic gene ACSL4 in CD8⁺ T cells protects T cells from ferroptotic cell death, but the cells showed reduced competitive expansion compared with wild-type CD8⁺ T cells in vivo.⁵⁶ Therefore, careful modulation—rather than blanket inhibition—of ferroptosis pathways in T cells may be necessary to balance survival with effective immune responses.

NK cells

NK cells can mediate tumor surveillance and elimination, but their infiltration and numbers are reduced in the gastric cancer TME. This loss is driven by indoleamine 2,3-dioxygenase metabolites, particularly L-kynurenine (L-KYN), which induce NK cell death. L-KYN-induced death was rescued by Fer-1 but not by inhibitors of necrosis, apoptosis, or pyroptosis, indicating ferroptosis as the primary mechanism. L-KYN-resistant NK cells exhibited higher GPX4 levels, highlighting its protective role. GPX4 overexpression enhanced NK cell survival without affecting functional gene expression (IFN-γ, perforin, granzyme B), cytotoxicity, or migration. Adoptive transfer of GPX4-overexpressing NK cells into tumor-bearing mice improved tumor control, correlating with increased NK cell accumulation in tumors.⁵⁷

Cancer-associated fibroblasts (CAFs), a major component of the tumor stroma, were also found to inhibit NK cell activity by inducing ferroptosis. Specifically, CAFs secreted iron and FSTL1 into the TME, leading to upregulation of NCOA2 in NK cells and resulting in iron overload, which induced ferroptotic cell death in NK cells. Blocking FSTL1 with antibodies and reversing iron overload with DFO restored NK cell function in vitro. These findings suggest that improving NK cell survival by targeting ferroptosis could enhance immunotherapy efficacy in gastric cancer.⁵⁸

Regulatory T cells

Tregs play a crucial role in suppressing antitumor T-cell responses. Clinically targeting and eliminating Treg cells could provide an immunotherapeutic benefit. GPX4 deletion in Tregs led to increased sensitivity to ferroptosis. Those ferroptotic Treg cells could promote Th17 responses and enhance CD8⁺ T cell-mediated antitumor immunity. Notably, Treg-specific deletion of GPX4 in mice led to improved tumor growth control in the B16 and MC38 tumor models in vivo. These findings suggest that inducing ferroptosis in Tregs may be a promising strategy for combating immunosuppression and enhancing the effectiveness of cancer immunotherapies.⁵⁹

Dendritic cells

DCs have been reported to be susceptible to ferroptosis induced by RSL3, but not by Erastin. Functionally impaired ferroptotic DCs showed a reduced ability to activate CD8⁺ T cells to secrete IFN-γ. Ferroptosis in DCs also was influenced by PPARG expression. Death of DCs by ferroptosis resulted in poorer outcomes for tumor immunotherapy.⁶⁰

Macrophages

The relative abundance of inflammatory versus anti-inflammatory macrophages in the TME influences tumor growth and antitumor immune responses. Exosomes released from ferroptotic breast cancer cells have been reported to inhibit anti-inflammatory macrophage polarization. Mechanistically, macrophages engulfed exosomes from ferroptotic tumor cells, which led to a significant decrease in anti-inflammatory macrophage markers compared with macrophages exposed to exosomes from PMA-induced cell death. When macrophages were co-cultured with exosomes from ferroptotic cells, breast cancer cell migration and invasion were inhibited.⁶¹

The pancreatic cancer (PDAC) TME is highly oxidative and immunosuppressive, contributing to reduced immunotherapy efficacy. A recent study showed that PDAC cells undergoing H₂O₂-induced ferroptosis release exosomes containing KRAS^G12D, which are taken up by macrophages and promote an M2-like, immunosuppressive phenotype. In mouse models, local injection of ferroptosis inhibitors reduced tumor growth when PANC-1 cells were co-injected with macrophages. Clinically, higher KRAS^G12D levels in macrophages correlate with poorer patient survival. These findings suggest that ferroptotic cell-derived exosomes can modulate macrophage behavior, depending on the tumor type and exosome content (figures 2 and 3).⁶²

Neutrophils

Tumor-infiltrating neutrophils (TINs) can be immunosuppressive and contribute to immune evasion and tumor metastasis. In TINs, Acod1 is highly expressed, producing itaconate to activate Nrf2 and protect cells from oxidative death.^{63 64} In Acod1 knockout mice, tumor growth was reduced, and neutrophil number was decreased due to greater ferroptosis. Notably, when Acod1-deficient mice were treated with ICB, tumor control was significantly improved, suggesting that promoting ferroptosis in TINs may enhance the efficacy of immunotherapy.⁶³

This result parallels findings in tumor-associated macrophages, where Nrf2 activation—often initiated by early progenitor dysregulation in the bone marrow—serves as a deterministic survival switch.^{65 66} By preventing their own ferroptosis, these “pre-programmed” myeloid cells persist within the TME as a stable, long-lived immunosuppressive barrier that collectively undermines the efficacy of immunotherapy.

MDSCs

Myeloid-derived suppressor cells (MDSCs) comprise various immature myeloid cells that can suppress immune responses within the TME. In head and neck squamous cell carcinoma (HNSCC) patients, higher GPX4 expression has been associated with poorer patient survival and was inversely correlated with calreticulin, an immune-modulating signal. Treatment with RSL3 in HNSCC mouse models reduces tumor burden by decreasing immunosuppressive MDSCs and increasing CD8⁺ T cells within the TME.⁶⁷ In colon cancer, MDSCs upregulated expression of ASAH2, which suppresses ferroptosis. NC06, a drug targeting ASAH2, induced ferroptosis in these MDSCs, reducing both tumor burden and MDSC levels in vivo.⁶⁸ Notably, the PMN-MDSC subset undergoes ferroptosis in the TME; however, these ferroptotic PMN-MDSCs remain immunosuppressive and can inhibit T-cell activity. In CT26 and LCC mouse tumor models, the ferroptosis inhibitor Lip-1 in combination with anti-PD-1 therapy resulted in a significant reduction in tumor size.⁶⁹ While inducing ferroptosis in MDSCs can reduce their numbers and potentially enhance antitumor immunity, ferroptotic MDSCs themselves may suppress the proliferation of immune cells. Thus, strategies targeting ferroptosis in MDSCs might need to be carefully tailored to the specific tumor type.⁷⁰

Strategies for targeting ferroptosis in cancer immunotherapy

Inducing cancer cell death is a major goal of cancer therapies. Traditionally, apoptosis has been studied as the main pathway to achieve this and to influence the TME. However, facilitating cancer cell ferroptosis may provide a new strategy to not only kill cancer cells but also reshape the TME and affect immune responses (table 1).

Table 1. Potential therapeutic strategies targeting ferroptosis in cancer immunotherapy.

Target	Agent(s)	Mechanism	Stage	Immunological impact /therapeutic relevance	Limitations
SLC7A11 (System Xc⁻)	Erastin	Inhibits cystine import	In vitro	Proof-of-concept ferroptosis induction	Metabolic instability and poor solubility limit in vivo use
	IKE	Potent inhibition of cystine uptake	Preclinical (in vivo)	Suppresses tumor growth	Limited clinical translation
	Sulfasalazine	Blocks cystine transport	Clinical	Potential to sensitize tumors to immunotherapy	Modest clinical efficacy
GPX4	RSL3, ML210, ML162	Covalent inhibition of GPX4 leading to lipid peroxide accumulation	Preclinical	Robust ferroptosis induction in cancer cells	On-target toxicity due to ubiquitous GPX4 expression
	N6F11	Targeted GPX4 degradation	Preclinical	Improved tumor-selective ferroptosis	Early-stage development
PGAM1	PGAM1 inhibitors	Downregulation of LCN2 promoting ferroptosis	Preclinical	Enhances CD8+ T-cell infiltration and anti-tumor immunity	Limited validation
PI3K/HDAC	BEBT-908	Activates p53 and IFN-γ–STAT1 signaling to induce immunogenic ferroptosis	Preclinical	Augments CD8+ T-cell responses and sensitizes tumors to anti-PD-1 therapy	Safety and efficacy not yet established

GPX4, glutathione peroxidase 4; GSH, glutathione; HDAC, histone deacetylase; IFN, interferon; IKE, imidazole ketone erastin; LCN2, lipocalin 2; PD-1, programmed cell death protein 1; PGAM1, phosphoglycerate mutase 1; PI3K, phosphoinositide 3-kinase; PMN-MDSC, polymorphonucler myeloid-derived suppressor cells; SLC7A11, solute carrier family 7 member 11.

Targeting SLC7A11 could reduce cysteine uptake and induce tumor cell ferroptosis. Additionally, SLC7A11 suppression can sensitize tumors to immune checkpoint inhibitors by reshaping the TME and overcoming immunotherapy resistance.^{7 46} Several drugs have been developed to target SLC7A11 and induce tumor cell ferroptosis, including erastin, imidazole ketone erastin (IKE), and sulfasalazine (SASP). Erastin is widely used in vitro but not in vivo due to its metabolic instability and poor solubility, which preclude its use in preclinical and clinical studies. IKE, alone or in combination with other drugs, has been shown to reduce tumor growth in various tumor models. SASP induces ferroptosis in mouse tumors; however, in human clinical studies, outcomes have been suboptimal.

GPX4 is highly expressed in different types of cancer cells. Targeting GPX4 to induce cancer cell ferroptotic death has become an attractive therapeutic strategy. RSL3, ML210, and ML162 bind to the active site of GPX4, and this inhibitory effect has been tested in cell lines in vitro. RSL3 and ML210 have also been investigated in mouse tumor models in vivo. One challenging problem is that GPX4 is expressed in most tissues, so targeting GPX4 could have off-tumor side effects. The cell-specific GPX4 degrader N6F11 has been shown to induce tumor cell-specific ferroptosis and may represent a better treatment in the future.⁷¹ Though most GPX4-targeting therapies are still in preclinical development, they offer a novel strategy to overcome drug resistance and to target aggressive cancer types selectively.⁶⁹

Beyond these targets, inhibition of phosphoglycerate mutase 1 (PGAM1) has been shown to downregulate LCN2, trigger ferroptosis, and increase CD8⁺ T-cell infiltration in hepatocellular carcinoma. When combined with anti-PD-1 therapy, PGAM1 inhibition significantly enhanced tumor suppression.⁷²

Furthermore, BEBT-908, a dual phosphoinositide 3-kinase/histone deacetylase inhibitor, was shown to induce immunogenic ferroptosis via activation of p53 and IFN-γ-signal transducer and activator of transcription 1 (STAT1) signaling. This led to elevated MHC-I expression and enhanced CD8⁺ T-cell responses, further sensitizing tumors to anti-PD-1 treatment and producing a combinatorial anticancer effect.⁷³

Challenges and future directions

Several studies have demonstrated that a ferroptosis-related gene signature is correlated with the efficacy of cancer immunotherapy. However, identifying a specific and universal marker of ferroptosis across different human cancer types remains a significant challenge. Ferroptosis is regulated by multiple pathways. These include iron metabolism, lipid peroxidation, antioxidant defenses, and metabolic rewiring. These pathways can vary widely between tumor types and even among individual tumor cells within the same cancer. Many genes involved in ferroptosis also have roles in other metabolism pathways, such as lipid, iron, and cystine metabolism. The TME also varies between different cancer types. Even within the same type of cancer, different patient samples may have different ferroptotic susceptibility and marker expression. As a result, no single biomarker has yet proven reliable or specific enough to universally indicate ferroptosis in all types of cancer. This underscores the need for combinatorial marker panels and context-specific validation to accurately assess ferroptosis in clinical settings. Nonetheless, expression profiles of ferroptosis-related genes may be potential biomarkers for predicting immunotherapy outcomes. Tumors with low SLC7A11 or SLC3A2 expression, or high ACSL4 and Decr2 expression, exhibit greater sensitivity to ferroptosis and improved clinical responses to ICB. Ultimately, the capacity of tumor cells to undergo ferroptosis may represent a strategy for patient selection.

Developing drugs that target ferroptosis may become a promising future therapeutic strategy for tumors. Ferroptosis-driven tumor cell death could modulate the TME, and combining these drugs with immunotherapy has the potential to control tumor growth. However, challenges remain. First, ferroptosis-inducing drugs may harm healthy cells and tissues, as many ferroptosis proteins are not tumor-specific. The primary aim is to selectively target and kill tumor cells without harming normal cells. Tumor cell-targeted degraders may overcome this limitation. Second, while ferroptotic tumor death may enhance antitumor immunity, partly by increasing CD8⁺ T cells in the TME, certain immune cells—such as T cells, macrophages, and DCs—are also susceptible to ferroptosis. Targeting these cells may impair their antitumor functions and potentially promote tumor growth. This dual impact underscores the importance of understanding and balancing ferroptosis induction to optimize tumor cell death while preserving immune responses. Achieving this balance may require precise targeting and combination strategies, or careful sequencing of therapies, as research in ferroptosis-based cancer therapies advances.