Ferroptosis in cancer therapy: a novel approach to reversing drug resistance

Yang et al.’s study induced ferroptosis with different ferroptosis-inducing compounds (FINs) and found that all FINs inhibited GPX4 directly or indirectly through GSH depletion. They thus concluded that GPX4 is the key regulator of ferroptosis [7]. As a glutathione peroxidase in mammals, GPX4 appears to play a particularly important role in catalyzing the reduction of phospholipid hydroperoxides (PLOOH) into corresponding phospholipid alcohols [16]. GSH is necessary for the normal physiological function of GPX4 [17]. Intracellular GSH synthesis is catalyzed by glutamate–cysteine ligase. For this reason, cysteine, which is taken up as cystine via the cystine/glutamate antiporter (xCT) system, is referred to as the rate-limiting amino acid for the synthesis of GSH [18]. As a result, depriving cells of cysteine by inhibiting xCT system also contributes to the indirect inhibition of GPX4. Consequently, inactivation of GPX4 can lead to the accumulation of PLOOH, thus inducing cell membrane damage and ferroptotic death (Fig. 1). Earlier in 2008, inactivation of GPX4 was demonstrated to trigger redox-regulated cell death and cause neurodegeneration in vivo and ex vivo [19]. In Yang et al.’s study, it was also demonstrated that genetic inhibition of GPX4 can induce tumor cell ferroptosis and inhibit tumor growth in vivo [7]. These findings suggest that the canonical GPX4-regulated pathway plays an important role in tumor biology.

As an iron-dependent form of cell death, ferroptosis is characterized by an increase in the small pool of Fe²⁺, which is referred to as the labile iron pool (LIP). In 1997, it was found that cellular iron uptake was mostly mediated by the binding of serum transferrin to the transferrin receptor and transferrin endocytosis [20]. In 2016, Wen et al. demonstrated that autophagy contributes to ferroptosis by degrading ferritin in fibroblasts and cancer cells. Overexpression of nuclear receptor coactivator 4 increased intracellular LIP by increasing ferritin degradation (namely ferritinophagy) [21]. On the one hand, the increased intracellular LIP can generate free radicals (hydroxyl radicals) through Fenton reaction and participate in the peroxidation of phospholipids to produce PLOOH [22]. On the other hand, the generation of most ROS in cells is iron-catalyzed. The generation of ROS initiates lipid peroxidation and eventually leads to ferroptosis [23]. Cancer cells have been shown to have increased iron requirements for survival in comparison to normal cells [24]. Iron uptake is enhanced and intracellular iron level is increased in rapidly proliferating cancer cells, shedding light on the potential of ferroptosis inducing as a therapeutic target for cancer. Alvarez et al. showed that increasing intracellular LIP by suppressing NFS1 sensitizes lung cancer cells to ferroptosis and reduces the growth of lung tumors in vivo [10]. In addition, Chang et al. found that BAY 11–7085, a well-known IκBα inhibitor, activates heme oxygenase-1 and enhances cancer cell ferroptosis by increasing LIP [25]. Taken together, the modulation of the iron metabolism pathway serves as a therapeutic means to trigger cancer cell ferroptosis.

The specific feature of ferroptosis is increased lipid peroxidation; thus, the metabolism of lipid peroxides is considered to be critical in the process of ferroptosis. Ferroptosis is likely performed by the peroxidation of membrane phospholipids to produce PLOOH and the decomposition of PLOOH to generate 4-hydroxynonenal or malondialdehyde. The products of lipid peroxidation cause membrane instability and permeabilization and eventually lead to cell death [26].

In nonenzymatic lipid peroxidation, polyunsaturated fatty acids (PUFAs) are ligated with coenzyme A (CoA) through the function of acyl-CoA synthetase long chain family member 4 (ACSL4) to produce acyl-CoA. Then, acyl-CoA can be re-esterified in phospholipids through various lysophosphatidylcholine acyltransferases (LPCATs) to produce PL. Therefore, the regulation of ACSL4 and LPCATs can dictate ferroptosis sensitivity [27, 28]. In enzymatic lipid peroxidation, PLOOH production can also be mediated by lipoxygenases (LOX) and cytochrome P450 oxidoreductase (POR). LOXs are nonheme iron-containing enzymes that directly catalyze the deoxygenation of free and esterified PUFAs to generate PLOOH [29]. A previous study demonstrated that overexpression of LOX-5, LOX-12, and LOX-15 sensitizes cells to ferroptosis. LOX inhibitors have also been demonstrated to be effective antioxidants that protect cells from lipid peroxidation [30]. In 2020, Zou et al. identified POR as necessary for ferroptotic cell death in cancer cells by genome-wide, CRISPR–Cas9-mediated suppressor screens [31]. Prior studies suggested that P450 could accept electrons from POR and catalyze the peroxidation of PUFAs [32]. The pro-ferroptotic role of POR has also been demonstrated by genetic depletion across a wide range of lineages and cell states [31].

GPX4 functions in concert with GSH to catalyze the reduction of hydrogen peroxide and organic hydroperoxides to water or the corresponding alcohols, which plays a vital role in protecting against ferroptosis [7, 8]. Direct inhibition of GPX4 leads to ferroptotic cell death. Chen et al. showed that androgen receptor induces resistance to temozolomide treatment in glioblastoma. Androgen receptor ubiquitination induced by the curcumin analog was demonstrated to suppress GPX4, thereby inducing ferroptosis and reversing temozolomide resistance in glioblastoma [33]. Similar results were obtained by Yang et al., who found that highly expressed KIF20A was correlated with oxaliplatin resistance in colorectal cancer. Cellular ferroptosis may be induced by disrupting the KIF20A/NUAK1/PP1β/GPX4 pathway, thus overcoming the resistance of colorectal cancer to oxaliplatin [34]. Moreover, ferroptosis can also be triggered indirectly by blocking the synthesis of GSH, the cofactor of GPX4. Recently, ent-kaurane diterpenoids have been reported to be able to overcome cisplatin resistance by inducing ferroptosis. Further mechanistic studies have revealed that ferroptosis induced by ent-kaurane diterpenoids is caused by targeting peroxiredoxin I/II and depleting GSH [35]. Cysteine is an essential cellular antioxidant and the rate-limiting amino acid for GSH biosynthesis. The xCT system mediates the uptake of cystine, which is rapidly reduced to cysteine, thus affecting GSH synthesis. Inhibition of the xCT system has also been shown to be capable of triggering ferroptosis. A study by Roh et al. revealed that xCT system inhibitors (erastin and sulfasalazine) or genetic silencing of the xCT system could trigger head and neck cancer cell ferroptosis and overcome cisplatin resistance [36]. Similarly, Fu et al. showed that inducing ferroptosis by restraining Nrf2/Keap1/xCT signaling also sensitized cisplatin-resistant cells to cisplatin in gastric cancer [37].

Ferroptosis is defined as an iron-catalyzed form of regulated necrosis [38]. Elevated levels of cellular LIP increase vulnerability to ferroptosis. Du et al. reported that dihydroartemisinin treatment could overcome cisplatin resistance in pancreatic ductal adenocarcinoma by inducing ferroptosis. Further study revealed that tumor cell ferroptosis contributed to catastrophic accumulation of LIP after dihydroartemisinin treatment [38]. Lipocalin 2 (LCN2) is a secreted glycoprotein and may regulate iron homeostasis [39]. A recent study showed that LCN2 overexpression was correlated with 5-fluorouracil resistance in colon cancer. Targeting LCN2 overcame 5-fluorouracil resistance by increasing intracellular iron levels, which in turn led to tumor cell ferroptosis [40]. Divalent metal transporter 1 (DMT1) is another key protein in the regulation of iron homeostasis. Turcu et al. reported that increasing cellular LIP caused by DMT1 inhibition triggered ferroptosis, thus killing breast cancer stem cells and reversing multidrug resistance [41].

Accumulation of PLOOH is the hallmark of ferroptosis. Lipid metabolism is closely related to cell vulnerability to ferroptosis. PLOOH can accumulate in both enzymatic and nonenzymatic ways. Dull et al.’s study revealed that ACSL4 is an essential component for ferroptosis execution. ACSL4 functions in converting long chain-PUFAs to acyl-CoA, subsequently participating in producing PLOOH in an enzymatic way [27]. Ye et al. reported that the activation of ACSL4 by inhibiting ADP ribosylation factor 6 (ARF6) overcame gemcitabine resistance in pancreatic cancer by inducing ferroptosis [48]. In addition, several LOXs have been reported to be able to oxygenate PUFAs directly in a nonenzymatic way, thereby mediating ferroptosis [49]. Zhang et al. showed that arachidonate lipoxygenase 15 (ALOX15), a LOX member, was closely linked with the suppression of ferroptosis in gastric cancer. Downregulation of microRNA-522, which promotes ALOX15 expression, provides novel methods to enhance cisplatin/paclitaxel sensitivity in gastric cancer by inducing ferroptosis [50].

Several tumor cell-intrinsic mechanisms leading to immunotherapy resistance have been identified, including the activated mitogen-activated protein kinase (MAPK) signaling pathway, loss of PTEN expression, expression of the WNT/β-catenin signaling pathway, loss of interferon-gamma signaling pathway, and loss of tumor antigen expression [51]. These changes result in the inability to mount potent antitumor immune responses. Recent studies have proposed that stimulating the adaptive immune system by causing immunogenic cell death may turn the immunologically cold state into a checkpoint blockade responsive state [54, 55]. Interestingly, ferroptosis has been demonstrated to be immunogenic. Efimova et al. described a novel approach to stimulate the adaptive immune system by triggering ferroptosis-dependent immunogenic cell death in preclinical models. Early ferroptotic cells have been reported to release damage-associated molecular patterns (adenosine triphosphate and high-mobility group box 1) and promote the phenotypic maturation of bone marrow-derived dendritic cells [56, 57]. In addition, Luo et al. identified an eat-me signal on the ferroptotic cancer cell surface, 1-steaoryl-2-15-HpETE-sn-glycero-3-phosphatidylethanolamine (SAPE-OOH). Moreover, enrichment of SAPE-OOH facilitated phagocytosis by targeting toll-like receptor 2 on macrophages [58]. Taken together, inducing ferroptosis in cancer cells may elicit a vaccination-like effect to stimulate antitumor immunity, thus overcoming immunotherapy resistance.

Immune suppressor cells within the TME also contribute to resistance to immunotherapy, such as regulatory T cells (Tregs) and tumor-associated macrophages (TAMs) [51].

Quezada et al. reported that the effect of anti-cytotoxic T-lymphocyte antigen 4 (CTLA-4) immunotherapy was related to the ratio of effectors T cells to Tregs in the TME [59]. Tregs exert inhibitory control over infiltrating effectors T cells. In this study, combination with GM-CSF–transduced tumor cell vaccine potentiates the efficacy of CTLA-4 blockade by increasing the ratio of effectors T cells to Tregs. Another study by Oweida et al. also demonstrated that Tregs infiltration was associated with anti-PD-L1 immunotherapy resistance. Targeting Tregs depletion has been shown to restore antitumor immunity [60]. Notably, a recent study showed that GPX4 protects Tregs from ferroptosis. GPX4-deficient Tregs produce interleukin-1βand facilitate the T helper 17 response, which enhances antitumor immunity [61]. Taken together, these results indicate that inducing Tregs ferroptosis by suppressing GPX4 may reverse immunotherapy resistance.

As another subset of cells that seem to affect responses to immunotherapy, TAMs can polarize to two major phenotypes: antitumor M1 (TAM1) and protumor M2 (TAM2). TAM2 is often the dominant subset of TAMs in TME. Zhu et al. demonstrated that reprogramming TAMs by blocking CSF1/CSF1R enhanced the efficacy of immune checkpoint inhibitors in pancreatic cancer [62]. A recent study indicated that TAM1, which leads to higher levels of inducible nitric oxide (NO) synthase (iNOS)/NO•, exerted higher resistance to ferroptosis than TAM2. Regulation of ferroptosis by iNOS/NO• inhibited TAM2 survival without affecting TAM1, thus potentiating antitumor immunity in the TME [63]. Additionally, Jiang et al. showed that high TYRO3 expression was associated with anti-PD-1/PD-L1 immunotherapy resistance in a preclinical mouse model and in patients. TYRO3 inhibited tumor cell ferroptosis and facilitated the polarization of TAM1 to TAM2. Inhibition of TYRO3 also triggered ferroptosis and reprogrammed TAMs, thus resensitizing cancer cells to immunotherapy [53].

Collectively, triggering ferroptosis in the TME may be a tumor cell-extrinsic approach to overcoming immunotherapy resistance by depleting immune suppressor cells.