The state and expression levels of critical regulatory proteins in the cell determine how the cell responds to signals that trigger ferroptosis (Figure 2).

Cells have their own protective mechanisms against this type of death, and the key role is played by GPX4. GPX4 eliminates lipid peroxide-induced damage caused by lipid peroxides and protect the integrity of the cell membrane, thereby preventing ferroptosis (17). Because GPX4 is the core regulator of ferroptosis, it is crucial for maintaining cell survival and balance. It has been found that in gastric cancer (GC), a higher level of GPX4 promotes tumor formation by stimulating the mammalian target of rapamycin (mTOR) signaling pathway (25). In pancreatic cancer, high mobility group at-hook 2 (HMGA2) enhances transcription by directly binding to the GPX4 promoter, promoting GPX4 protein synthesis through the mTORC1-4EBP1-S6K pathway. Forming an HMGA2-GPX4 positive feedback loop that confers ferroptosis tolerance (26). Similarly, in esophageal squamous cell carcinoma (ESCC), the upregulation of heat shock protein 27 (Hsp27) increases the expression of GPX4, thereby inhibiting ferroptosis in cancer stem cells (CSCs), and is associated with a poor prognosis for patients (27). R et al. found that in hepatocellular carcinoma (HCC) (28), nuclear-enriched GPX4 inhibits grainyhead-like 3 (GRHL3) transcriptionally, thereby weakening its control over the PTEN/PI3K/AKT pathway and promoting tumor metastasis. Therefore, the expression level of GPX4 in digestive tract tumor tissues is significantly increased, and it confers ferroptosis resistance through multiple mechanisms, becoming a key driver of tumor progression and metastasis.

SLC7A11, as a functional subunit of System xc^-, can directly affect the sensitivity of GSH synthesis and ferroptosis. Downregulated SLC7A11 expression reduces GSH levels, increasing cellular susceptibility to ferroptosis. Studies have shown that in CRC, N6-methyladenosine (m6A) demethylase fat mass and obesity-associated protein (FTO) can upregulate the expression of SLC7A11 through the M6A-YTHDF2-dependent pathway, thereby protecting CRC cells from ferroptosis and promoting tumorigenesis. In addition, inhibiting FTO or using the novel FTO inhibitor mupirocin can induce ferroptosis in CRC cells, enhance their sensitivity to ferroptosis inducers, such as erastin and RAS selective lethal 3 (RSL3), and inhibit tumor growth (29). In addition, lysophosphatidylcholine acyltransferase 2 (LPCAT2) also induces ferroptosis in CRC cells by inhibiting the nuclear translocation of protein arginine methyltransferase 1 (PRMT1) and suppressing the expression of SLC7A11 (30). In pancreatic cancer (31), low expression of nudc domain-containing protein 2 (NudCL2) is associated with a poor prognosis in patients. Its absence can upregulate SLC7A11, promoting cell migration and invasion, while inhibition of SLC7A11 can reverse this phenotype, indicating that NudCL2 can inhibit epithelial-mesenchymal transition (EMT) by regulating the transcriptional activity of SLC7A11. Highlighting SLC7A11 as a potential target for inhibiting tumor growth.

Delivering iron into the cytoplasm, ferric iron from 1 (FTH1) is the key iron metabolism protein (32). Umathum et al. discovered that in CRC, NaB can induce ferritin autophagy by upregulating NCOA4 and degrading FTH1, thereby promoting intracellular Fe²⁺ accumulation and ferroptosis, which inhibits the growth of transplanted tumors (33). Studies have shown that in Kirsten rat sarcoma viral oncogene homolog (KRAS) mutant pancreatic ductal adenocarcinoma (PDAC), the expression of FTH1 contributes to cell survival and tumorigenesis. This suggests that FTH1 plays a significant role in supporting malignant progression. However, the regulation of miRNA-5000-3p can counteract FTH1’s effects, leading to metabolic disorders and worsening disease progression (34). FTH1 in GC is a key molecule downstream of NSUN5. The NSUN5-FTH1 axis is activated by blocking erastin-induced iron death, promoting SGC7901 cell growth and tumor development in vivo (35). These studies reveal that FTH1 has multiple roles in tumor metabolism and cell fate. It relates to both ferroptosis susceptibility and metabolic reprogramming.

In conclusion, GPX4, SLC7A11, FTH1, and other key regulatory proteins form an interconnected network that can not only jointly regulate the ferroptosis process in digestive tract cancer but also serve as potential biomarkers to predict treatment response.

Non-coding RNAs (ncRNAs)—including microRNA (miRNA), long non-coding RNA (lncRNA), and circular RNA (circRNA)—play essential roles (36). Play an essential role in regulating the expression of key genes involved in ferroptosis at the epigenetic and post-transcriptional regulatory levels, thereby affecting the sensitivity of digestive tract tumor cells to ferroptosis (Figure 3).

As a key post-transcriptional regulator, miRNA has been demonstrated by several studies to have a significant impact on regulating ferroptosis. MiR-103a-3p is an oncogenic miRNA associated with GC development (37); miR-103a-3p influences the ferroptosis process in GC cells by modulating GSH levels. Additionally, the local anesthetic bupivacaine can inhibit the proliferation of GC cells by inducing ferroptosis through the miR-489-3p/SLC7A11 axis. Similarly, miR-375 can mediate the occurrence of Helicobacter pylori-related GC by inhibiting the JAK2-STAT3 signaling pathway (38), and can trigger ferroptosis by targeting SLC7A11 (39). However, notably, in CRC, the expression levels of miR-509-5p and miR-15a-3p are significantly decreased compared with normal colorectal cells. The overexpression of miR-509-5p and miR-15a-3p can promote iron cell apoptosis in CRC cells by targeting SLC7A11 and GPX4, respectively (40) (Table 1).

LncRNA, which is typically more than 200 nucleotides in length and does not encode proteins, has yielded significant research achievements in the study of various cancers (41). Recently, in the study of ESCC, it was found that the LncRNA SNHG7 was significantly upregulated in ESCC cells and tissues. By regulating the expression of p15 and p16, it could promote the proliferation ability of ESCC cells and inhibit their apoptosis (27) (Table 1).

Circular RNA, as an emerging non-coding RNA, is also involved in ferroptosis-related processes (42). Studies have confirmed that in GC cells, the upregulation of circ_0000190 and circ_0008035 expression can inhibit tumor cell proliferation and migration by inducing ferroptosis (43). The latter inhibits apoptosis and ferroptosis of GC cells directly by regulating the miR-599/EIF4A1 axis (44). However, in ESCC, circ_0120816, as a miRNA sponge of miR-1305, can not only promote the development of ESCC but also directly target the key enzyme thioredoxin reductase 1 (TXNRD1) for GSH synthesis, thereby exerting its anti-cancer effect through miR-1305. This indicates that targeting circ_0120816 may regulate the initiation mechanism of ferroptosis by influencing GSH synthesis (45) (Table 1).

In conclusion, ncRNA plays a crucial yet complex role in regulating ferroptosis in digestive system tumors. Existing studies have initially revealed the action pathways and mechanisms of some ncRNAs; however, the overall interaction network between ncRNAs and ferroptosis remains far from being fully elucidated. Therefore, in the future, developing more efficient technical platforms to comprehensively depict the ncRNAs’ regulatory network and verify it in combination with clinical samples will be the key direction for technological breakthroughs.

TME coordinately regulates tumor cell ferroptosis susceptibility through three core mechanisms (46): hypoxia, cancer-associated fibroblasts (CAFs), and immune cell infiltration (Figure 2). Under hypoxia, HIF-1α is stabilized by low COMMD10 expression, promoting the transcription of ceruloplasmin and SLC7A11 and inhibiting ferroptosis in HCC. Concurrently, the lncRNA CBSLR reduces cystathionine β-Synthase (CBS) mRNA stability in a YTHDF2-dependent manner, decreasing ACSL4 protein levels and weakening the synthesis of pro-ferroptotic phosphatidylethanolamine, conferring chemotherapy resistance in GC (35). A study demonstrated that CAFs inhibit ferroptosis in pancreatic cancer through two mechanisms: primarily by secreting exosomes with miR-522 that directly target and inhibit arachidonate 15-lipoxygenase (ALOX15); additionally, through the TGF-β/SMAD3/ATF4 signaling pathway, which facilitates cysteine secretion to increase GSH synthesis (47). In the immune microenvironment, the key immune cells, CD8⁺ T cells, secrete interferon-γ (IFN-γ). IFN-γ not only downregulates SLC7A11 but also upregulates ACSL4. Notably, this regulatory pattern—via the apolipoprotein L3 (APOL3)-lactate dehydrogenase A (LDHA) axis—reduces intracellular lactic acid levels and enhances CD8⁺ T cell cytotoxicity, which may eventually lead to their direct triggering of ferroptosis in target tumor cells. Furthermore, studies have shown that Rg3 can further enhance the activity of CD8⁺ T cells by regulating the circFOXP1-miR-4477a-PD-L1 axis and cooperatively induce ferroptosis and apoptosis in gallbladder cancer (GBC) cells (48).

However, the impact of TME does not always promote ferroptosis. For instance, immunosuppressive cell populations, such as M2-type macrophages, can protect tumor cells from the threat of ferroptosis by secreting antioxidants or directly blocking the ferroptosis signaling pathway (49). TME hypoxia, CAF metabolic changes, and complex immune cell interactions collectively determine whether ferroptosis occurs in tumor cells. Understanding the interaction between these molecules has enabled us to discover many new ways to develop new therapies, such as designing treatments for key components in TME.

Ferroptosis can not only directly eliminate tumor cells but also exert its effects by regulating tumor-related pathways and the microenvironment. Therefore, ferroptosis inducers can prevent EMT by directly eliminating epithelial-phenotype cancer cells and inhibiting key EMT transcription factors. They can also reverse the already formed transformation state (50). Thereby effectively limiting local invasion and distant metastasis of tumors.

Studies have shown that the expression level of sirtuin 3 (SIRT3) in GBC tissues is significantly lower than that in adjacent normal tissues (51), and this decline in expression is associated with a poor prognosis for patients. The mechanism of action of SIRT3 lies in its ability to inhibit the AKT signaling pathway. On the one hand, it alleviates the inhibitory effect of AKT on ACSL4, thereby promoting ferroptosis. On the other hand, it also blocks AKT-mediated EMT. Therefore, SIRT3 can ultimately inhibit the growth and metastasis of GBC (51). It is worth noting that HIF-1α also plays a key role in connecting EMT and ferroptosis interactions. Specifically, EMT may increase the sensitivity of cells to ferroptosis by enhancing the accumulation of PUFAs and iron. However, the antioxidant capacity provided by HIF-1α can partially alleviate this situation. This balance between EMT and ferroptosis is not fixed, and it will be affected by factors such as TME and cell heterogeneity. Suggesting that targeting HIF-1α or its regulatory pathways is a viable therapeutic strategy (52).

Clinical studies confirm that inducing ferroptosis significantly impairs CSC self-renewal capacity. As CSCs play critical roles in tumor recurrence, metastasis, and drug resistance, they are key therapeutic targets (53). Ferroptosis induces mitochondrial D-lactate dehydrogenase (LDHD) inactivation via lipid reactive oxygen species (ROS) amplification, leading to D-lactic acid accumulation. D-lactic acid downregulates xCT/GPX4 and increases Fe²⁺ levels, further accelerating ferroptosis. Concurrently, it oxidizes NANOG/OCT4, remodels the epigenome, degrades membrane lipid rafts, and reverses EMT, stripping CSCs of their self-renewal and tumorigenic capabilities (54).

Interestingly, CSCs are both sensitive and resistant to ferroptosis. CSCs exhibit responsiveness to ferroptosis, with their fate determined by the transient equilibrium among three dimensions. These dimensions encompass antioxidant capacity, free iron pool capacity, and membrane polyunsaturated fatty acid content. When two or more dimensions shift toward a pro-oxidative state, such as reduced DHODH activity due to POLQ or E2F4 inhibition, D-lactic acid accumulation triggered by CDK7-YAP-LDHD pathway disruption, or CD8⁺ IFN-γ silencing of SLC7A11, CSCs lose self-renewal capacity and become susceptible to ferroptosis. Conversely, enhancing axial effects through M2 TGF-β, CAF-cysteine, and stearoyl-CoA desaturase 1 (SCD1) upregulation protects CSCs from ROS damage and maintains stem cell properties (55). Previous studies have demonstrated that we can induce CSCs to be more susceptible to ferroptosis by modulating key signaling pathways, including CD44/NRF2, YAP/TAZ, and autophagy (56). Specifically, inducing ferroptosis disrupts the fundamental processes that maintain the characteristics of CSCs. It disrupts redox homeostasis by inhibiting NRF2-mediated antioxidant responses. Inhibit invasion and metastasis characteristics by down-regulating EMT-related transcription factors, and by altering the activity of iron-dependent epigenetic enzymes, which hinders the epigenetic reprogramming essential for CSC self-renewal (57).

Animal studies have shown that ferroptosis is highly effective in inhibiting the metastasis of various digestive tract cancers to distant organs such as the peritoneum and liver. Especially in GC, the down-regulation of NCOA4, the core receptor for ferritin phagocytosis, has been found to be a key driver of peritoneal metastasis. Decreased expression of NCOA4 will block the release of intracellular iron, thereby restraining the Fenton reaction and ROS generation. Due to the weakened ferroptosis, GC cells acquire anti-apoptotic ability and are more prone to peritoneal dissemination (58). These findings suggest that inducing ferroptosis may be a promising therapeutic strategy for preventing and treating metastatic tumors.

To survive, cancer cells often evade ferroptosis induction through complex escape mechanisms. These escape mechanisms are the key biological basis for tumors to maintain survival, mainly including the upregulation of the antioxidant system, the remodeling of lipid metabolism, the enhancement of iron uptake capacity, and the alteration of the TME (Figure 4).

Tumor cells typically upregulate key antioxidant components, such as GPX4 and SLC7A11, to prevent ferroptosis. Merkel et al. believe that mitochondrial-targeted ROS scavengers can inhibit ferroptosis driven by mitochondrial damage by reducing mitochondrial mROS production (59). Specific tumors, such as PDAC, rely on the system xc^-. This dependence promotes cystine uptake, synchronously synthesizes GSH and CoA, and establishes a synergistic mechanism to resist ferroptosis (60).

Lipid metabolism remodeling—converting saturated fatty acids (SFAs) to monounsaturated fatty acids (MUFAs)—is a key ferroptosis escape mechanism. This conversion is mediated by SCD1. SCD1 can strengthen the cell membrane by producing MUFA. It can reduce the toxicity of ferroptosis inducers in several ways, such as reducing the number of easily oxidized PUFA substrates, reducing the accumulation of ROS, and reducing the expression of GPX4. It has been found that if SCD1 is excessively active, it often indicates a poor prognosis. If SCD1 is inhibited by genetic methods or drugs, the PUFA-dependent ferroptosis process can be restarted (61). Consistent with this, Qin et al. (62) further verified in GC that targeting SCD1 via inhibiting ubiquitin-specific protease 7 (USP7) promotes SCD1 degradation, thereby inducing ferroptosis and suppressing tumor growth. This highlights the congenitally pro-survival effect of SCD1 in digestive tract malignancies, even in different TMEs.

Tumor cells often upregulate the expression of CD44 to enhance their resistance to iron-dependent lipid peroxidation. CD44 is not only important in cell adhesion and migration but also helps cells absorb more iron. Cells start related channels by increasing CD44, allowing iron to enter from outside the cell so that the concentration of free iron in the cell becomes higher (63). Moreover, CD44-mediated iron uptake may drive cancer progression by affecting iron-dependent epigenetic features, such as the demethylation of H3K9me2. This will further enhance the expression of EMT-related genes and ultimately strengthen the stemness characteristics of tumor CSCs.

To resist the attack of ferroptosis inducers and immune cells, tumor cells have developed various mechanisms to inhibit ferroptosis in the TME, among which CAFs provide them with a relatively safe environment. CAFs achieve this protective effect through several ways, including secretion of ferroptosis inhibitors, regulation of iron metabolism, and direct interaction with immunosuppressive cells (64). In addition, some cancer cell membrane-derived vesicles, such as CCM-FSS and CHM-ABI, have also been confirmed to be important inhibitory strategies. Specifically, CCM-FSS refers to the nanoparticle composed of a ferroptosis-sensitizing cascade agent (FSS) encapsulated by a cancer cell membrane (CCM) and is abbreviated as CCM-FSS; CHM-ABI refers to the nanoparticle composed of a dual inhibitor of CXCR4 and NOX4 (ABI) encapsulated by a cancer cell-CAF hybrid membrane (CHM) and is abbreviated as CHM-ABI. These vesicles can not only directly inhibit the ferroptosis of tumor cells themselves through multiple mechanisms but also reprogram CAFs to a quiescent state. Therapeutic strategies based on these vesicles have demonstrated efficacy in orthotopic CRC, colorectal cancer liver metastases (CLM), and humanized immune patient-derived xenograft (PDX) models. This therapy not only effectively inhibits tumor growth but also enhances the anti-tumor immune response, and it shows extremely low toxicity, providing a novel strategy for CRC liver metastases (65).

The emergence of immunotherapy represents a significant breakthrough in cancer treatment. It fights cancer by using our own immune system. Immunotherapy leverages the immune system to combat cancer, and its synergy with ferroptosis is supported by their close association with the TME (72). Because ferroptosis is closely related to the TME, the cancer cells killed by ferroptosis not only affect immune cells, but also non-cancer cells related to immune responses are affected in the immune microenvironment. This connection is essential for stopping the spread of cancer. Therefore, it provides robust support for therapeutic modalities in conjunction with immunotherapy. Studies have found that in esophageal cancer, ferroptosis combined with immunotherapy can produce a synergistic effect. CD8⁺ T cells promote ferroptosis in tumors, and inducing ferroptosis can trigger ICD.

Ferroptosis dynamically regulates ICD through three stages (73): early mild lipid peroxidation and mitochondrial DNA (mtDNA) release activate dendritic cells (DCs) and enhance antigen recognition (74); mid-stage massive adenosine triphosphate (ATP) release further recruits T cells, but excessive peroxidation damages the extracellular matrix (ECM) and impedes infiltration (75); and terminal stage membrane rupture massively discharges immunosuppressive factors like TGF-β and adenosine while inducing ferroptosis in T/Natural Killer Cells (NK cells), thereby counteracting antitumor immunity (76). Therefore, a phased precision intervention combining low-dose inducers, ACSL4 agonists, moderate-potency GPX4 inhibitors, and low-dose liproxstatin-1 can maximize the ICD effect and enhance the efficacy of immunotherapy in gastrointestinal tumors. In digestive tract malignancies, ferroptosis activates the DC-cGAS-STING-CD8⁺ T cell axis by triggering classic ICD events. The IFN-γ secreted by CD8⁺ T cells then secondarily upregulates SLC7A11 and ACSL4, further amplifying the sensitivity of cancer cells to ferroptosis. Forming a metabolism-immunity positive feedback loop that deprives CSCs of immune evasion ability and leads to their sustained elimination (77).

Targeting ferroptosis may overcome ICB resistance (78). In GC, tumor-associated neutrophils (TANs) play a subtle role. They can induce ferroptosis but do not clear the debris; instead, they release oxidized lipids. TANs release oxidized lipids, impairing anti-tumor immune responses, weakening the immune response that should have attacked the tumor. However, a liposome encapsulating both ferroptosis inhibitor Liproxstatin-1 and modified photosensitizer Icy7, denoted LLI. LLI reshapes the immune landscape through a dual effect—first, it prevents ferroptosis in TANs, and second, it induces ICD in these neutrophils through ceramide accumulation. Notably, this ceramide-mediated ICD reshapes the tumor immune microenvironment, thereby enhancing the efficacy of anti-PD-1 therapy (79). Alternatively, Cui et al. proposed a different mechanism involving L-kyurine (L-KYN). In the immune microenvironment of GC, the indoleamine 2,3-dioxygenase (IDO) enzyme produces L-KYN. They proposed that this L-KYN might prompt NK cells to move towards ferroptosis, but the key lies in avoiding the aromatic hydrocarbon receptor (AHR) pathway. This process will eventually eliminate a large number of NK cells, depleting them to the point where they are no longer able to effectively combat tumors. However, high expression of the ferroptosis protective factor GPX4 shields NK cells from ferroptosis. Thus, genetic engineering to enhance GPX4 expression in NK cells may prolong their survival and activity in the harsh TME, opening new avenues for NK cell-based immunotherapy (76).

Consistent with this, in HCC. Zheng et al. revealed a dual mechanism that inhibits ferroptosis and promotes immune escape, thereby facilitating HCC progression. They found that simultaneously targeting the phosphoglycerate Mutase 1 (PGAM1) enzyme could not only activate ferroptosis but also significantly enhance the effect of anti-PD-1 therapy (80). Turning to GBC, Ye et al. focused on a natural compound, ginsenoside Rg3. This substance not only halts iron-induced cell death but also activates CD8⁺ T cells, making them more effective against the cancer. At the same time, Rg3 triggers iron death in GBC cells through a specific molecular pathway (circFOXP1-miR-4477a-PD-L1). It’s like hitting multiple targets at once to fight the tumor (48). In pancreatic cancer research, Li and his team highlighted the pivotal role of the MCP-GPX4/HMGB1 axis in linking ferroptosis to the tumor immune microenvironment. Their research indicates that targeting monocyte chemoattractant protein (MCP) has a dual effect: it can not only induce ferroptosis with “immunogenicity”, but also initiate anti-tumor immune responses by activating M1-type macrophages. The immunogenicity of ferroptosis exhibits stage-dependent characteristics. Its advantage lies in the early exposure of cell membrane transferrin receptors/ATP upon event initiation, activating the dendritic cell-STING-CD8⁺ T cell feedback loop. Its disadvantage involves terminal rupture, releasing transforming growth factor-β, adenosine, and LPO, leading to NK cell/CD8⁺ T cell exhaustion, as well as the recruitment of regulatory T cells. This discovery addresses the issue of ferroptosis inducers potentially increasing immunosuppressive cell numbers, identifying MCP as a promising new target for pancreatic cancer immunotherapy (81).

Ferroptosis-generated heme/ATP activates the DC-STING-IFN-γ pathway, reversing the TME from a TGF-β/adenosine-driven suppressive state to an inflammatory immune-permissive state by blocking SLC7A11 depletion of Treg/MDSC GSH reserves and eliminating immunosuppressive neutrophils. This enhances the efficacy of ICB in gastrointestinal cancers. These examples demonstrate that there are interesting interactions between them. Ferroptosis seems to regulate the sensitivity of TME by affecting the existence, vitality, and metabolic state of immune cells. Conversely, the characteristics of the immune microenvironment, including cell types and their metabolic states, also seem to affect the sensitivity of cancer cells to ferroptosis. This subtle relationship between ferroptosis and the immune system not only provides valuable scientific insights but also opens up an exciting new avenue for optimizing immunotherapy for digestive tract cancer.