Lipid peroxidation is a trigger of ferroptosis and is caused by a complicated process of lipid metabolism. In the process of ferroptosis, polyunsaturated fatty acids (PUFAs), especially AA and adrenic acid (AdA), are sensitive to be oxidized to lipid peroxides by ROS, resulting in damage to the lipid bilayer and thus promoting ferroptosis. ACSL4 and lysophosphatidylcholine acyltransferase 3 (LPCAT3) are required for biosynthesis and modification of PUFAs in cell membranes and act as important drivers of ferroptosis. In the ferroptotic process, ACSL4 catalyzes the ligation of AA and AdA with coenzyme A to form lipid derivatives AA-CoA or AdA-CoA (Yuan et al., 2016b; Doll et al., 2017; Kagan et al., 2017). These products can be re-esterified into membrane phosphatidylethanolamine (PE) by LPCAT3 to produce AA-PE or AdA-PE (Dixon et al., 2015; Doll et al., 2017). However, loss function of ACLS4 could induce a substantial change of long-chain PUFAs tails to short-chain and monounsaturated fatty acids (MUFAs) tails in phospholipids (Kim et al., 2001; Doll et al., 2017). These MUFAs, including exogenously supplemented and endogenously produced stearoyl-CoA desaturase 1 (SCD1), can be converted to acyl-coenzyme A esters, which can bind to the membrane phospholipids, thereby protecting cancer cells from ferroptotic cell death (Chen et al., 2021). Thus, depletion of ACSL4 or LPCAT3 could inhibit ferroptosis by reducing the accumulation of lipid peroxidation substrates in cells.

Moreover, the accumulation of iron can promote the activity of arachidonate lipoxygenases (ALOXs), which are a class of non-heme iron-dependent enzymes, and induce ferroptosis. ALOXs, including ALOXE3, ALOX5, ALOX12, ALOX12B, ALOX15, and ALOX15B, can directly oxygenate PUFAs to generate various hydroperoxy PUFA derivatives, such as lipid hydroperoxides (LOOHs), malondialdehyde (MDA), and 4-hydroxynonenal (4HNE), thus inducing ferroptosis (Kagan et al., 2017). Moreover, ALOXs also play important physiological roles in the immune system by producing pro-inflammatory and anti-inflammatory molecules that directly modulate (neuro) inflammatory processes and the TME (Friedmann Angeli et al., 2019; Proneth and Conrad, 2019). Besides the context-dependent role of ALOXs in ferroptosis, recent findings suggest that the cytochrome P450 oxidoreductase (POR) also plays a role in promoting lipid peroxidation in cancer cells in an ALOX-independent manner (Zou et al., 2020). Prostaglandin-endoperoxide synthase 2 (PTGS2/COX2) is not involved in the oxidation of phospholipids, but can oxidize lysophospholipids. PTGS2 is usually upregulated during ferroptosis and considered as a biomarker, rather than as a driver, of ferroptosis (Yang et al., 2014). However, the inhibitor of PTGS has little effect on the lethality of erastin or RSL3, but PTGS2 may modulate ferroptosis in neural cells after traumatic brain injury (Yang et al., 2014; Yang et al., 2016; Xiao et al., 2019). The specific role of PTGS2 in ferroptosis of cancer cells remains to be elucidated.

The antioxidant enzyme GPX4 can covert GSH into oxidized glutathione (GSSG) and directly reduce toxic phospholipid hydroperoxides (PL-OOH) to nontoxic phospholipid alcohols (PLOH), thus acting as a central repressor of ferroptosis in cancer cells (Yang et al., 2014). Inactivation of GPX4-dependent antioxidant defense leads to the accumulation of the lipid peroxides, which is a trigger of ferroptosis. Selenium and GSH are necessary components for the expression and activity of GPX4 in ferroptosis (Ingold et al., 2018; Ursini and Maiorino, 2020). Selenium can boost the anti-ferroptosis activity of GPX4 by inserting selenocysteine into GPX4 residue at Sec 46 (U46) (Ingold et al., 2018). Meanwhile, selenium-supplementation has been demonstrated to upregulate GPX4 expression by coordinated activation of the transcription factor AP-2 gamma (TFAP2C) and specificity protein 1 (SP1), which protects against ferroptosis-related cerebral hemorrhage (Alim et al., 2019). The antioxidant GSH is composed of three amino acids, cysteine, glycine, and glutamic acid; cysteine is considered to be the main rate-limiting factor in de novo synthesis of GSH. In mammalian cells, cystine (the oxidized form of cysteine) is imported into cells by system Xc⁻ for subsequent GLC (glutamate-cysteine ligase)-mediated GSH production. As a transporter that facilitates the exchange of cystine and glutamate across the plasma membrane, system Xc⁻ is a heterodimeric protein complex consists of two subunits: solute carrier family 7 member 11(SLC7A11; also known as xCT) and solute carrier family 3 member 2 (SLC3A2, also known as CD98 or 4F2). The expression of SLC7A11 is promoted by nuclear factor erythroid 2 like 2 (NFE2L2/NRF2) and downregulated by the tumor suppressor genes, such as TP53, BAP1, and BECN1 (Jiang et al., 2015; Song et al., 2018a; Zhang et al., 2018). This dual regulation of SLC7A11 expression and activity creates a fine-tuning mechanism to modulate cell antioxidant capacity and lipid ROS during ferroptosis (Chen et al., 2021). Inhibition system Xc⁻ (with erastin, sulfasalazine, or sorafenib) or GPX4 (with RSL3, ML162, ML210, FIN56, or FINO2) can produce ROS in the absence of GSH depletion and trigger ferroptosis. Similarly, depletion of SLC7A11 or GPX4 promotes lipoxygenase-mediated lipid peroxidation and leads to ferroptosis in specific cells or organs (Sato et al., 2005; Friedmann Angeli et al., 2014). In addition, the endogenous antioxidant coenzyme Q10 (CoQ10) produced by the mevalonate (MVA) pathway is converted to a reduced form to suppress lipid peroxidation, thereby protecting cells from undergoing ferroptosis in a GPX4-independent manner (Mullen et al., 2016). Tetrahydrobiopterin (BH₄) and ESCRT (endosomal sorting complexes required for transport)-III membrane repair systems also show a context-dependent antioxidant property (independent of the Cysteine/GSH/GPX4 axis) during ferroptosis (Dai et al., 2020c; Kraft et al., 2020). Furthermore, the ferroptosis suppressor protein 1 (FSP1/AIFM2) controls the generation of reduced CoQ10, and can also inhibit ferroptosis in cancer cells by activating the ESCRT-III membrane repair system (Dai et al., 2020d). This indicates that the antioxidant system could prevent ferroptosis through mutual synergy or complementation.

Pancreatic ductal adenocarcinoma (PDAC) is a one of the deadliest tumors that is resistant to most therapies. In recent years, an increasing number of researchers have investigated the role of ferroptosis in the treatment, initiation, and progression of pancreatic cancer. Daniel et al. found that cytosolic aspartate aminotransaminase (GOT1) inhibition represses mitochondrial metabolism and enhances labile iron availability through autophagy, which accelerates pancreatic cancer cell death by ferroptosis (Kremer et al., 2021). Furthermore, specific and conditional depletion of pancreatic Slc7a11 could induce pancreatic tumor ferroptosis and inhibit pancreatic tumorigenesis in mice (Badgley et al., 2020). However, ferroptosis may play oncogenic roles in pancreatic cancer in addition to suppressing tumor progression. During Kras-driven PDAC in mice, ferroptotic cancer cells promote the release of damage-associated molecular patterns (DAMPs) such as 8-hydroxy-2ʹ-deoxyguanosine (8-OHG) and KRAS-G12D protein, leading to macrophage polarization and subsequent pancreatic tumor growth (Dai et al., 2020a; Dai et al., 2020b). The dual role of ferroptosis in pancreatic cancer needs to be considered when developing new cancer therapeutic approaches based on ferroptosis induction.

HCC is the most major form of liver cancer diagnoses and deaths (McGlynn et al., 2021). Clinically, sorafenib is one of the six approved systemic therapies for patients with advanced HCC (Llovet et al., 2021), and inducing ferroptosis of cells is an important mechanism in sorafenib treating HCC. In terms of mechanism, sorafenib blocks the import of cysteine by inhibiting SLC7A11, the subunit of cystine/glutamate antiporter system Xc⁻, thereby triggering ferroptosis and ultimately killing HCC cells (Louandre et al., 2013). The effect of sorafenib on HCC cells is closely related to ferroptosis process. GSH depletion through cysteine deprivation or cysteinase inhibition can enhance the susceptibility of HCC cells to sorafenib-induced ferroptosis, which provides a new sight in therapeutic combinations for advanced HCC (Li Y. et al., 2021). Moreover, research also reported that iron chelator deferoxamine (DFX) protects HCC cells from the sorafenib-induced ferroptotic cell death (Louandre et al., 2013). The status of retinoblastoma (Rb) protein modulates the tumorigenesis of liver and the response of human liver cancer cells to sorafenib (Mayhew et al., 2007; Louandre et al., 2015). HCC cells with Rb-negative status promote the occurrence of ferroptosis and cell death upon exposure to sorafenib (Louandre et al., 2015).

It has been reported that several proteins, such as NRF2, DAZAP1, metallothionein-1G (MT-1G), Branched-chain amino acid aminotransferase 2 (BCAT2), and CDGSH iron-sulfur domain 1 (CISD1), inhibit ferroptosis in HCC cells through different ways. Sun et al. revealed that activation of the p62-Keap1-NRF2 pathway prevents HCC cells from ferroptosis, and the Ras/Raf/MEK pathway is an important target for ferroptosis in the treatment of HCC (Sun et al., 2016b; Nie et al., 2018). RNA-binding protein DAZAP1 maintains the SLC7A11 mRNA stability to negatively regulate ferroptosis in HCC cells (Wang Q. et al., 2021). MT-1G knockdown increases GSH depletion and lipid peroxidation, thereby enhancing sorafenib-induced ferroptosis (Sun et al., 2016a). The aminotransferase BCAT2 protects HCC cells from system Xc⁻ inhibitor-induced ferroptosis by mediating the metabolism of sulfur amino acid (Wang K. et al., 2021). Inhibition of CISD1, an iron-containing outer mitochondrial membrane protein, contributes to erastin-induced ferroptosis in HCC cells (Yuan et al., 2016a). In addition, inhibition of sigma 1 receptor (S1R) by haloperidol also promotes erastin- and sorafenib-induced ferroptosis in HCC cells (Bai et al., 2017).

Breast cancer is one of the most common causes of cancer-related death in women (DeSantis et al., 2019). The triple-negative breast cancer (TNBC) is an aggressive breast cancer subtype with poor prognosis, and lacks effective therapeutic target. Notably, TNBC have been shown to more sensitive to ferroptosis than estrogen receptor (ER) positive breast cancer, suggesting that ferroptosis is a promising treatment for TNBC patients. The drug-tolerant persister breast cancer cells acquire a dependency on GPX4, making cells sensitive to ferroptosis triggered by GPX4 inhibition (Hangauer et al., 2017). Sulfasalazine (SAS) can trigger ferroptosis in breast cancer cells by inactivating GPX4 and system Xc⁻, especially in cells with low ER expression (Yu et al., 2019). It was observed that the tyrosine kinase inhibitors lapatinib and neratinib induce ferroptosis by increasing the level of iron-dependent ROS in breast cancer cells (Ma et al., 2017; Nagpal et al., 2019). Cystine is one of the most essential amino acids in TNBC. Starvation of cystine reduces the level GSH by directly decreasing synthesis and the GCN2-eIF2α-ATF4-CHAC1 pathway, thereby inducing ferroptosis in TNBC cells (Chen et al., 2017). The anti-diabetic drug metformin reduces the stability of xCT by inhibiting its UFMylation, thus inducing ferroptosis and inhibiting the proliferation of breast cancer cells (Yang et al., 2021). In contrast, MUC1-C, a transmembrane protein, can maintain the GSH level and redox balance by forming a complex with xCT and the CD44 variants (CD44v), which inhibits ferroptosis in breast cancer cells (Ishimoto et al., 2011). Another study found that prominin-2 protects breast cancer cells from ferroptotic cell death by promoting the formation of ferritin-containing multivesicular bodies and exosomes that transport ferritin out of the cell (Brown et al., 2019).

Several studies revealed that iron is closely associated with the development of lung cancer, and the disruption of iron homeostasis makes cells more sensitive to ferroptosis (Kuang and Wang, 2019). USP35 is abundant in lung cancer cells and acts as a deubiquitinase to maintain the protein stability of ferroportin, which is a cellular efflux channel for iron (Tang et al., 2021). Depletion of USP35 can promote ferroptotic cell death and sensitivity to cisplatin and paclitaxel chemotherapy in lung cancer cells (Tang et al., 2021). The iron-sulfur cluster biosynthetic enzyme NFS1 is overexpressed in well-differentiated lung adenocarcinomas and protects cells from ferroptosis within high oxygen environment by sustaining the levels of iron-sulfur cofactors (Alvarez et al., 2017). Accumulating evidence showed that system Xc⁻ is increased in several types of cancers, including lung cancer. As a result, suppression of system Xc⁻ by either genetic depletion or pharmacological inhibition to trigger ferroptosis becomes a potential therapy target for lung cancer. Recently, Wang et al. reported that the stem cell factor SOX2 increases the resistance of lung cancer cells to ferroptosis by increasing the expression of SLC7A11. Oxidation of SOX2 at Cys265 can inhibit its transcriptional activity and make lung cancer cells with high SOX2 expression more susceptible to ferroptosis. (Wang X. et al., 2021). In addition to the typical GSH-dependent GPX4 pathway, FSP1 acts as a ferroptosis-resistance factor by reducing CoQ10 to inhibits lipid peroxidation in lung cancer cells (Bersuker et al., 2019).

Recently, researchers have shown that the PUFA biosynthesis pathway is critical in the ferroptosis of gastric cancer. Gastric cancer cells with high expression of elongation of very long-chain fatty acid protein 5 (ELOVL5) and fatty acid desaturase 1 (FADS1) are more sensitive to ferroptosis, while cells that highly expressed stearoyl-CoA Desaturase 1 (SCD1) exhibit ferroptosis resistance (Wang C. et al., 2020; Lee et al., 2020). Similarly, inhibition of the GSH-GPX4 antioxidation system, which directly reduces lipid hydroperoxides to nontoxic lipid alcohols, can also induce ferroptosis in gastric cancer. Apatinib, an approved anti-angiogenic agent for advanced gastric cancer therapy, can induce lipid peroxidation by inhibiting the expression of GPX4 in gastric cancer cells, and thus leading to ferroptosis (Zhao et al., 2021). In addition, Sirtuins 6 (SIRT6) is significantly expressed in sorafenib-resistant gastric cancer cells. Depletion of SIRT6 inactivates the Keap1-NRF2 pathway and downregulates GPX4 expression, and enhances the sensitivity of gastric cancer cells to sorafenib-induced ferroptosis (Cai et al., 2021). Cysteine dioxygenase 1 (CDO1), on the other hand, regulates erastin-induced ferroptosis in gastric cancer cells by limiting the synthesis of GSH (Hao et al., 2017). Increasing CDO1 activity can competitively absorb cysteine and oxidize it to its sulfinic acid, thereby reducing cellular GSH levels and promoting ferroptosis.

Iron is a fundamental element that plays a critical role in a variety of vital biological processes via distinct oxidation states (from −2 to +6). In biological systems, iron is primarily restricted to ferrous (Fe²⁺), ferric (Fe³⁺), and ferryl (Fe⁴⁺) states. Due to its flexible oxidation states, iron is usually employed as an electron transporter in crucial biological reactions, such as DNA synthesis and cellular respiration. As previously mentioned, iron metabolism has been identified as one of the pivotal regulators of ferroptosis. Studies have shown that the excess iron overload could result in ferroptosis by accumulating ROS through the Fenton reaction. In this reaction, hydrogen peroxides react with ferrous iron (Fe²⁺) to produce hydroxyl radicals and ferric iron (Fe³⁺) (Winterbourn, 1995). In turn, the generation of Fe³⁺ in the Fenton reaction can be reduced to Fe²⁺ by the superoxide, a by-product of cellular respiration, to undergo further Fenton reaction. This cycle between oxidized and reduced form of iron, known as Haber-Weiss reaction, enables it to produce massive amounts of harmful ROS, which could attack most cellular proteins, lipids, and nucleic acids (Torti et al., 2018). When the GSH-mediated anti-oxidative system is insufficient to eliminate the increasing iron-dependent lipid peroxides, cells were induced into ferroptosis.

Multiple areas of research have found that cancer cells exhibit a distinct iron metabolism from their non-malignant counterparts (Torti and Torti, 2013; Fonseca-Nunes et al., 2014). Cancer cells, especially cancer stem cells, are iron addicted and uptake more iron into cellular to maintain the activity of iron-dependent proteins and enhance tumor growth and metastasis (Ludwig et al., 2015). The intracellular iron metabolism is tightly regulated by iron regulatory proteins (IRP1 and IRP2), which modulate the cellular Fe²⁺ concentrations by regulating iron import, storage, release, and efflux (Andrews and Schmidt, 2007). Of note, cancer cells likely to express different IRPs to promote iron uptake and repress iron export due to the rapid proliferation rate and high metabolic activities that are commonly associated with malignancy. Down-regulation of IRP2 protects cells against ferroptosis, while depletion of FBXL5, an IRP2 negative regulator, sensitizes HT-1080 cells to erastin-induced ferroptosis (Dixon et al., 2012). Moreover, the ferritin heavy subunit (FTH) produced by cancer cells can activate circulating T cells to secret cytokines such as TNF-α and IFNγ, resulting in increased iron retention in macrophages (Meng et al., 2017). Degradation of FTH by lysosomes can accumulate large amounts of iron (Terman and Kurz, 2013). Increasing FTH levels in fibrosarcoma HT1080 cells by inhibiting lysosomal activity or silencing nuclear receptor activator 4 (NCOA4), a cargo receptor that mediates FTH autophagic degradation and iron release, can inhibit ferroptosis (Gao et al., 2016; Hou et al., 2016). Emerging studies, on the other hand, have revealed that iron metabolism in cancer cells is implicated in the complicated nature of tumor-associated macrophages (TAMs) polarization in the TME. In general, the macrophages with high plasticity can be polarized into two forms: the classical M1 (proinflammatory) and the alternative activated M2 (anti-inflammatory) macrophages (Murray, 2017). The M1 macrophages (activated by microbial agents and/or Th1 cytokines) are highly expressed with iron storage protein ferritin and low expressed with iron exporter ferroportin, thus M1 macrophages exhibit iron-accumulating properties (Agoro et al., 2018). However, M2 macrophages (activated by Th2 cytokines IL-4 or IL-13) display the opposite expression profile, with high expression of ferroportin and low expression of ferritin, which enhanced the release of iron and dysregulated iron metabolism in cancer cells (Recalcati et al., 2010). Changes in iron metabolism-related components involving ferritin, ferroportin, and iron levels could stimulate the M1 and M2 polarization (Liang and Ferrara, 2020). Ferroptotic tumor cells induced by excessive iron overload can release 8-OHG, the main product of oxidative DNA damage, to activate the stimulator of interferon genes protein (STING)-mediated DNA sensor pathway in TAMs, which results in TAMs infiltration and M2 polarization, thus enhancing pancreatic carcinogenesis (Dai et al., 2020b).

Emerging evidence shows that innate immune cells in the TME, such as TAMs and neutrophils, play a critical role in the regulation of iron metabolism by controlling iron availability in cancer cells (Figure 1). A large amount of iron required for daily iron metabolism was retrieved by macrophages from senescent red blood cells and then returned into the circulation (De Domenico et al., 2008). Therefore, cancer cells are iron consumers, while neutrophils and macrophages are iron providers. Within the TME, TAMs are implicated in cancer cell proliferation, angiogenesis, as well as immunosuppression and therapy resistance (Mantovani et al., 2006; Cassetta and Pollard, 2018). Moreover, TAMs could also increase the release of IRPs, such as the paracrine factor lipocalin 2 (Lcn2), to serve as an iron donor and promote the progression of breast cancer cells (Duan et al., 2018). TAMs can therefore engage in iron metabolism modulation by altering their polarization and interact with cancer cells to influence tumor development. Like macrophages, immune cell neutrophils could also generate iron-related proteins Lcn2 and lactoferrin to affect the iron metabolism in the TME (Cronin et al., 2019). With the stimulation of Lcn2, renal cell carcinoma cells demonstrated resistance to erastin-induced ferroptosis (Meier et al., 2021). In turn, Lcn2 is also involved in promoting the chemotaxis of neutrophils and stimulating M2 polarization to promote tumor progression (Sola et al., 2011; Ye et al., 2016). Hepcidin is an iron-regulatory hormone that induces degradation of ferroportin and negatively regulates iron efflux. Proinflammatory cytokines such as tumor necrosis factor (TNF)-α and IFNγ have been shown to induce hepcidin production in macrophages and neutrophils (Sow et al., 2009; Wu et al., 2012; Blanchette-Farra et al., 2018). Meanwhile, it has been demonstrated that interleukin (IL)-6 generated by cancer-associated fibroblasts (CAFs), TAMs, and neutrophils enhances hepcidin expression, thereby boosting the iron concentration in breast cancer cells (Nemeth et al., 2004; Blanchette-Farra et al., 2018). Overall, the crosstalk between tumor-associated innate immunity in the TME and dysregulated iron metabolism in cancer cells is crucial for triggering ferroptosis.

Abnormal iron overload in cancer cells promotes the production of ROS via the Fenton reaction and hence increases oxidative stress. Accumulating ROS oxidizes the PUFAs to lipid peroxides to induce ferroptosis of cancer cells. Amino acid metabolism is involved in the regulation of ferroptosis (Figure 2). To protect from ferroptosis, cancer cells activate their antioxidant capabilities through a variety of mechanisms such as increasing demand for cysteine to synthesize the antioxidant GSH. GSH is a tripeptide composed of cysteine, glutamate, and glycine. Cysteine availability is the main rate-limiting factor in de novo biosynthesis of GSH. The system Xc⁻ functions as a cystine/glutamate antiporter in a sodium-independent manner and consists of SLC7A11 (also known as xCT) and SLC3A2. SLC7A11 is the light chain subunit of system Xc⁻ with 12 transmembrane domains, and mediates the activity of system Xc⁻ to specifically transport cystine and glutamate. SLC3A2, the heavy chain component, is a single transmembrane protein that acts as a chaperone protein to regulate the localization and protein stability of SLC7A11 (Nakamura et al., 1999; Shin et al., 2017). In extracellular environment, cysteine is often found in its oxidized form, cystine, and the cystine transported into cells through the system Xc⁻ will be immediately reduced into cysteine to synthesize GSH. Recently, it has been reported that the limiting of exogenous cystine can promote the ferroptosis of pancreatic cancer cells by enhancing the transaminase enzyme GOT1 inhibition (Kremer et al., 2021). The canonical function of GOT1 is to regulate the malate-aspartate shuttle to generate reduced NADPH, which is used to maintain redox balance and proliferation in PDAC. Inhibition of GOT1 impairs mitochondrial oxidative phosphorylation and iron release in PDAC cells, thereby increasing their susceptibility to ferroptosis. In addition, high extracellular glutamate levels could inhibit system Xc⁻ function and cause ferroptosis. Increasing SLC7A11 expression induced by anti-VEGF treatment in glioma leads to the accumulation of extracellular glutamate, which subsequently promotes the proliferation and activation of regulatory T cells (Tregs) in glioblastoma (Long et al., 2020). Like the iron-carrier protein transferrin, the extracellular amino acid glutamine has been discovered as an inducer of ferroptotic cell death (Gao et al., 2015). Furthermore, the glutamine-driven intracellular metabolic process glutaminolysis, which is highly active in cancer cells, is required for ferroptosis triggered by starvation of full amino acids or of cystine alone (Gao et al., 2015).

With the ability to import cysteine from and export glutamate into the extracellular environment, SLC7A11 can exert effects on the interplay with cancer cells and TME, which in turn influence tumor growth and response to cancer therapies. SLC7A11 has been regarded as a primary TME-redox state homeostasis regulator and controls the prooxidant levels including O₂ ^•- and H₂O₂ (Chaiswing and Oberley, 2010). It has been reported that CAFs released cysteine and GSH can protect ovarian cancer cells from chemoresistance-induced oxidative stress, leading to resistance to platinum-based chemotherapy. However, IFNγ produced by effector CD8⁺ T cells downregulates SLC7A11 and SLC3A2 by reducing the release of cysteine and GSH from CAFs, thereby inhibiting cystine uptake in cancer cells, resulting in the induction of tumor lipid peroxidation and ferroptotic cell death (Wang et al., 2016). The role of effector CD8⁺ T cells in abrogating CAFs-mediated chemoresistance indicates the anti-cancer potential of the interplay between chemotherapy and immunotherapy. Meanwhile, Wang et al. further proved that the combination of cyst(e)inase (an engineered enzyme which depletes extracellular cystine and cysteine) with PD-L1 blockade synergistically inhibit cystine uptake through system Xc⁻ in tumor cells, resulting in intracellular GSH deficiency and thus inducing potent tumor cell ferroptosis (Wang et al., 2019). In summary, T cells activated by immune checkpoint inhibitor (ICI) therapy can cause amino acid metabolic reprogramming in tumor cells, ultimately contributing to ICI-induced anti-tumor T cell immunity. In addition, ferroptotic cancer cells can release some signals, notably DAMPs and eicosanoids, which increase their immunogenicity to induce tumor-specific immune responses, enhancing the efficacy of ICI therapy (Turubanova et al., 2019; Yu et al., 2020). Of note, SLC7A11 deficiency has no effect on the anti-tumor activity of T cells, but inhibits mouse T cells proliferate in vitro (Arensman et al., 2019). Thus, the role of SLC7A11 in immune cells within the TME warrant further investigation. Taken together, the interaction of amino acid metabolism and immune cells within the TME precisely regulates cancer cell ferroptosis and tumor progression.

Lipid metabolism is tightly linked to the cell sensitivity to ferroptosis (Figure 2). Oxidized PUFAs, especially oxidized AA and AdA, have been shown to drive cells to ferroptosis, but exogenous MUFAs, such as exogenous oleic acid (OA) and palmitoleic acid (POA), can protect cancer cells from ferroptosis (Yang et al., 2016; Kagan et al., 2017; Magtanong et al., 2019). Blood and lymphatic vessels are the main sources of intracellular FAs. Meanwhile, the absorption and release of FAs are strictly regulated by the state of the cells and external stimuli. Cancer cells always display aberrant lipid metabolism, which has been regarded as one of the hallmarks of aberrant cell growth and cancer progression. Elevated de novo synthesized lipids and FA oxidation are necessary for cancer cells to produce energy and required for post-translational modification of proteins (Carracedo et al., 2013; Rohrig and Schulze, 2016). The intricate interactions within the TME and adjacent stroma frequently impact the ways how cancer cells utilize lipids. For example, the presence of tumor low oxygen (hypoxia) environment results in the stabilization and activation of hypoxia inducible factors (HIFs), which regulate the transcription of target genes and consequently modulate lipid metabolism (Metallo et al., 2011). Activation of HIF1 in the head and neck tumor cells enhances SIAH2-mediated OGDN2 degradation, resulting in glutamine-derived carbon shunting to FA synthesis (Sun and Denko, 2014). The sterol regulatory element binding proteins (SREBPs), which control cholesterol and FA synthesis and uptake, are upregulated in hypoxic conditions and essential for glioblastoma multiforme cell survival (Lewis et al., 2015). SREBP can control the FA desaturases to introduce the double bonds into newly synthesized FAs and protect cancer cells from lipotoxicity under nutrient stress (Peck and Schulze, 2016). Furthermore, in response to metabolic stress and extracellular acidification, SREBP2 drives acetyl-CoA synthetase 2 (ACSS2) expression enabling the conversion of acetate to acetyl-CoA for lipid synthesis to promote tumor progression (Schug et al., 2015; Kondo et al., 2017). Recently evidence suggests that CAFs may also influence lipid transfer and uptake from the TME by enhancing the expression of FATP1 in breast cancer (Lopes-Coelho et al., 2018). In summary, the TME regulates fatty acid pools in cancer cells, which in turn affects ferroptosis sensitivity.

On the other hand, FAs secreted by cancer cells into the microenvironment could also impact infiltrating immune cell function and phenotype during tumor progression. During the stromal invasion, cancer cells can be impacted by circulating free fatty acids (FFAs) and other lipid molecules, which can substantially affect cell signaling or supply new substrates for cell development. Cancer cell-secreted FAs within the TME may stimulate TAMs to an M2 phenotype, which is marked by an increase in fatty acid oxidation (FAO) (Cook and Hagemann, 2013). The role of FFAs from circulation or within the TME in modulating the CD8⁺ T cell effector functions depends on context. Within hypoxic environments, Tregs in glioblastoma utilize extracellular FFAs to support the suppression of CD8⁺ T cells (Miska et al., 2019). Accumulation of lipid within lipid droplets in tumor-associated DCs reduces the antigen presentation, leading to dysfunction of DCs and poor stimulation of T cell responses, thereby modulating anti-tumor immunity (Herber et al., 2010; Ramakrishnan et al., 2014). Several studies demonstrate that exogenous lipids can disrupt the mTORC1-mediated glycolytic increase, which is necessary for the synthesis of granzyme B and IFNγ, leading to a deleterious impact on melanoma natural killer (NK) cell effector functions and their capacity to respond to stimuli, especially in the context of obesity (Michelet et al., 2018). Moreover, when the availability of glucose in the TME is limited, neutrophils and polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) can utilize FAO to support ROS production and lead to T cell suppression (Hossain et al., 2015; Rice et al., 2018). Therefore, exogenous lipids within the TME can influence the activity of immune cells to promote tumor progression.

As previously stated, lipid peroxidation is a hallmark of ferroptosis and is tightly linked to the process of lipid metabolism. The interplay between ferroptosis and lipid metabolism is important in anti-tumor immunity modulation. Accumulating evidence has revealed that ferroptotic cancer cells can affect the activity of immune cells by releasing various oxidized lipid metabolites. In ferroptotic cancer cells, AA can be metabolized to different eicosanoids and their derivatives, such as prostaglandin E2 (PGE2), 15-hydroperoxy-eicosatetra-enoyl-phosphatidylethanolamine (15-HpETE-PE), 15-HETE, 12-HETE, and 5-HETE, through the function of various pathways and released into the tumor immune microenvironment (Friedmann Angeli et al., 2019). PGE2 has been identified as an immunosuppressive factor that suppresses the anti-tumor functions of NK cells, conventional type 1 dendritic cells (cDC1s), and cytotoxic T cells to alter the TME and promote colon cancer cell proliferation, migration and invasion (Johnson et al., 2020). The pro-ferroptotic lipid 15-HpETE-PE that released by ferroptotic cancer cells will induces immune cells undergo ferroptosis in the TME (Kapralov et al., 2020). Taken together, lipid metabolic abnormalities of ferroptotic cancer cells can impact infiltrating immune cell function and phenotype in the TME, while TME often affects the way cancer cells utilize lipids. The interaction of lipid metabolism and immune cells within the TME builds a regulatory network for modulating ferroptosis sensitivity of cancer cells.

Lactate is a hydroxycarboxylic acid that is used as a fuel source by the heart, brain, and skeletal muscle under physiological conditions (Gladden, 2004). Based on the “lactate shuttle theory,” lactate can transcend compartment barriers and shuttle occur within and among cells, tissues, and organs to deliver oxidative and gluconeogenic substrates as well as in cell signaling (Brooks, 2009; 2018). In 1926, Otto Warburg demonstrated that glucose, as a primary energy source, was heavily ingested in several cancer cells, producing an excessive amount of lactate, even in the presence of oxygen (Warburg et al., 1927). Tumor cells and CAFs in the TME are the main sources of lactate production. Due to the accumulation of extracellular lactate, the extracellular pH is ranging between 5.5 and 6.6 (Webb et al., 2011; Justus et al., 2013). It seems that the acidic environment in tumors is associated with the more aggressive tumor characteristics such as metastasis, angiogenesis, and therapy resistance (Estrella et al., 2013; Garcia-Canaveras et al., 2019). For instance, lactate induces the expression of transforming growth factor-β2 (TGF-β2), a key regulator of tumor cell invasion, in glioma cells (Baumann et al., 2009). Therefore, increasing evidence reveals that lactate is not only a major energy source (Brooks et al., 1991; Bergman et al., 1999) but also a signaling molecule (Brooks, 2002; Hashimoto et al., 2007; Brooks, 2009) that has many regulatory roles in the TME, and particular relevance to oxidative stress resistance (Tauffenberger et al., 2019; Tasdogan et al., 2020) and lipid biosynthesis (Chen et al., 2016). Pucino et al. reported that the accumulation of extracellular lactate contributes to regulating the production of various lipid substances, including acetyl-CoA and citrate, via AMP-activated protein kinase (AMPK) or signal transducer and activator of transcription 3 (STAT3) signaling (Pucino et al., 2019). These finds suggest that lactate may play a role in cancer cell ferroptosis regulation. Recently, Zhao et al. discovered the aberrantly high level of lactate in the HCC extracellular environment could promote ATP production and further deactivate the AMPK signals (Zhao et al., 2020). In addition, it was also confirmed that lactate can induce the formation of MUFAs through the HCAR1/MCT1-SREBP1-SCD1 pathway, which contributes to the resistance of liver cancer cells to ferroptosis induced by oxidative stress (Zhao et al., 2020) (Figure 3). Therefore, the role of lactate receptor HCAR1 and transporter MCT1 in the maintenance of redox homeostasis has been found, suggesting that HCAR1 may be a potential therapeutic target for ferroptosis in HCC (Zhao et al., 2020). In addition, lactate in the TME can impair the immune system and promote tumor development by triggering and recruiting immunosuppressive related cells and molecules (de la Cruz-Lopez et al., 2019; Ivashkiv, 2020). It has been reported that Treg cells can use lactate to maintain the strong suppression of effector T cells within the TME (Watson et al., 2021). Tumor cell-derived lactate can polarize TAM into M2 type, activate the expression of VEGF and arginase 1 (ARG1) through the HIF1-α signaling pathway in macrophages, and assist TAM to promote tumor growth (de la Cruz-López et al., 2019; Zhang et al., 2020). Overall, high lactate concentration promotes acidification of the TME and suppresses the immune system, thereby contributing to immune evasion and tumor growth, invasion, and migration. The definitive information regarding the lactate in regulating ferroptosis of cancer cells is limited. Considering the role of lactate in immune suppression, the hypothesis that lactate modulates the sensitivity to ferroptosis through the immune system within the TME deserves consideration in future studies.

Recently, Wang et al. reported that ferroptosis induction-based treatment can be combined with ICIs to enhance the potency of killing cancer cells (Wang et al., 2019). ICIs (e.g., anti-PD-1 and anti-PD-L1 antibodies) can activate cytotoxic T cells to exert anti-tumor immunity, thereby inducing ferroptosis in cancer cells. It has been reported that ICI anti-PD-L1 antibodies can potentiate ferroptosis activators (e.g., erastin and RSL3) induced tumor growth inhibition in vitro and in vivo (Wang et al., 2019). Specifically, anti-PD-L1 antibodies stimulate CD8⁺ T cells to release cytokine IFNγ to activate the JAK-STAT1 pathway in cancer cells, thereby reducing the expressions of SLC7A11 and SLC3A2, and ultimately increasing the sensitivity of cancer cells to ferroptosis (Wang et al., 2019). In melanoma patients, an increase in ICIs efficacy is consistently associated with a decrease in SLC3A2 expression (Wang et al., 2019). Furthermore, depletion of STAT1 in tumor cells disrupts the IFNγ-mediated SLC7A11 suppression and reverses RSL3-induced ferroptosis. Notably, STAT1 can be activated by a variety of ligands, which means that other cytokines may have similar effects to that of IFNγ in triggering ferroptosis. Meanwhile, the released DAMPs play a dual role in anti-tumor immunity. As mentioned above, ferroptotic cancer cells release HMGB1 to promote the inflammatory responses of macrophages through binding to AGER (Wen et al., 2019). Therefore, blocking the HMGB1-AGER pathway with small molecules or genetic methods could limit the inflammatory response mediated by ferroptosis (Chen et al., 2021). These DAMPs, oxidized lipid mediators, and cytokines released into the extracellular formed a complex regulatory network between ferroptotic cancer cells and the immune system. As a result, immunotherapy combined with ferroptosis induction is a promising method that synergistically promotes anti-tumor effect.

Radiotherapy can directly cause oxidative damage to lipid membrane, leading to the accumulation of toxic lipid peroxidation and triggering ferroptosis in cancer cells. Meanwhile, recent studies suggest that radiotherapy can enhance the anti-tumor effect by synergizing with ferroptosis triggers and immunotherapy. Mechanically, radiotherapy can activate Ataxia-Telangiectasia mutated gene (ATM) to produce IFNγ and synergizes with immunotherapy-activated CD8⁺ T cells to suppress SLC7A11 expression in cancer cells. In addition, radiotherapy also upregulates ACSL4 expression in cancer cells, thereby increasing lipid peroxidation and inducing ferroptosis (Lei et al., 2020). Li et al. reported that mitochondrial DNA damage can activate the STING1/TMEM173-dependent DNA sensing pathway and lead to ferroptotic cancer cell death (Li C. et al., 2021). Similarly, this or other DNA sensor pathways activated by radiotherapy may similarly enhance the anti-tumor effect of inducing ferroptosis. Thus, the combination of radiotherapy and immunotherapy also increases the sensitivity of cancer cells to ferroptosis, indicating that inducing ferroptosis is a promising anti-cancer strategy.

Nanoparticle-based delivery of ferroptosis inducers can precisely and efficiently kill tumor cells with low toxicity by controlling the release of the drug and improving the pharmacokinetic properties of drugs. For example, due to the poor solubility of ferroptosis inducer withaferin A in water, withaferin A can be formulated in the amphiphilic degradable pH-sensitive nanocarrier to improve the convenience of drug delivery and reduce drug toxicity (Hassannia et al., 2018). Furthermore, nanoparticle-based ferroptosis induction is often combined with other treatments to enhance the anti-tumor effect. Chemodynamic therapy (CDT) is one of the ROS-dependent anti-cancer therapies that generate ROS through the iron-mediated Fenton reaction to kill cancer cells efficiently (Tang et al., 2019). Khalaf et al. used the novel tumor-targeted conjugated polymer nanoparticles (CPNPs) to deliver iron to the targeted tumor cells to trigger ferroptosis-assisted CDT through Fenton reaction that produces ROS (Jasim and Gesquiere, 2019). They conjugated iron-carrying CPNPs with the endothelin 3 (EDN3-CPNP), which specifically targets the melanoma endothelin-B receptor (EDNRB), and showed enhanced melanoma tumor targeting and tumor cell killing effects (Jasim and Gesquiere, 2019). In addition, nanoparticle-based ferroptosis induction can combination with photodynamic therapy (PDT) to markedly enhance the anti-tumor effect of PDT (Zhu et al., 2019). In order to counteract the depressant effect of the hypoxic TME in solid tumors on PDT, researchers constructed a nanoparticle containing the photosensitizer chlorin e6 (Ce6) and the ferroptosis inducer erastin via supramolecular interaction (Zhu et al., 2019). The obtained Ce6-erastin nanodrugs by cancer cells subsequently trigger ferroptosis and increases oxygen concentration, thus promoting the PDT anti-tumor efficacy upon light irradiation (Zhu et al., 2019). It has been reported that nanoparticle-based ferroptosis induction can also collaborate with photothermal therapy (PTT) to construct the nanoparticles named SRF@MPDA-SPIO. This nanoparticle is composed of ferroptosis inducer sorafenib (SRF) and super-paramagnetic iron oxide (SPIO) loaded into the mesopores or onto the surface of mesoporous polydopamine (MPDA) NPs through π–π stacking and/or hydrophobic interaction. In this system, the PTT (provide by MPDA NPs under laser emissions) synergizes with ferroptosis (induced by SPIO and SRF) to develop a promising cancer therapy (Guan et al., 2020). For chemotherapy, the ability of nanoparticle-induced ferroptotic cell death is highly dependent on nanoparticle structure and chemical modifications.