Ferroptosis, identified in 2012 by Dr. Brent R. Stockwell, is an iron-dependent form of non-apoptotic regulated cell death (Dixon et al., 2012). Ferroptosis is morphologically, biochemically, and genetically distinct from other well-known forms of cell death, including apoptosis, various forms of necrosis, and autophagy (Yagoda et al., 2007; Yang and Stockwell, 2008; Dixon et al., 2012; Fatokun et al., 2014). Apoptosis is a classical regulated cell death executed by dedicated pathways involving key pro-death effector proteins such as BCL2-associated X protein (BAX). However, ferroptosis is initialized through GSH depletion or inactivation of GPX4 activity (Dixon et al., 2012, 2014; Yang et al., 2014). During the ferroptosis, canonical hallmarks of apoptosis were not observed in several cancer cells, such as poly (ADP ribose) polymerase 1 (PARP1) cleavage, cytochrome c releasing from mitochondria, or pro-caspase-3 cleavage (Yagoda et al., 2007; Yang and Stockwell, 2008). Besides, there are special and distinct changes in mitochondrial morphology during the ferroptosis (Xie et al., 2016), mainly the loss of structural integrity, such as smaller than normal mitochondria with condensed mitochondrial membrane densities, reduced or absent mitochondria crista (Yagoda et al., 2007; Dixon et al., 2012) and outer mitochondrial membrane rupture (Friedmann Angeli et al., 2014). Thus, there are no cross-talk in the initial mechanisms or downstream evens between apoptosis and ferroptosis. So far, several groups have demonstrated that, in the early stage of ferroptosis, autophagy contributes to ferroptosis through providing available labile iron via NCOA4-mediated ferritinophagy (Gao et al., 2016; Hou et al., 2016). However, once ferroptosis is induced, especially in the later upstream events, autophagy maybe no longer a key regulator.

The overwhelming iron-dependent accumulation of lipid peroxidation products is considered the main killer in the ferroptotic process. Thus, small-molecule lipophilic antioxidants (e.g., ferrostatin-1 and α-tocopherol) and iron chelators (e.g., deferoxamine) can prevent ferroptotic cell death, whereas inhibitors of the other abovementioned forms of cell death cannot accomplish this feat. It is well-established that system Xc^- (Dixon et al., 2014) and GPX4 (Yang et al., 2014) are key regulators of ferroptosis. Two classical compounds, erastin (Dolma et al., 2003) and (1S, 3R)-RSL3, can induce ferroptosis through the inhibition of system Xc^- and GPX4, respectively. Cell death is crucial for normal development, homeostasis, and prevention of hyperproliferative diseases, such as cancer. Cancer cells can undergo several forms of regulated cell death during tumor development (e.g., apoptosis), and activation of regulated cell death is a primary and promising strategy for cancer therapy. Despite success in clinical cancer treatments, drug resistance to existing chemotherapeutic agents due to genetic alterations remains a problem. Intriguingly, erastin, a ferroptosis inducer, shows the ability to enhance the effectiveness of chemotherapy drugs (e.g., temozolomide, cisplatin, cytarabine/ara-C, and doxorubicin/Adriamycin) in certain cancer cells (Chen et al., 2015; Yu et al., 2015; Roh et al., 2016). Notably, RCCs, B cell-derived lymphomas and a subset of triple-negative breast cancer cell lines acquire a strong dependence on GPX4 and system Xc^- (Conrad et al., 2016). These results indicate that inducing ferroptosis maybe a new therapeutic anticancer strategy in tumors, especially tumors containing the two-abovementioned metabolic signatures. p53 is a crucial tumor suppressor and is mutant in most cancer cells. Recently, the tumor-suppressive activity of p53 was partially attributed to its ability to induce ferroptosis via inhibiting system Xc^- (Jiang et al., 2015); thus, ferroptosis may play a crucial role in tumorigenesis. In addition, some clinical drugs approved by the FDA [e.g., sorafenib (Dixon et al., 2014), sulfasalazine and ART] can also induce ferroptosis in several cancer cells, supporting the feasibility of using ferroptosis in preclinical and clinical settings. Taken together, ferroptosis shows an intimate association with tumors and cancer therapy, although it’s exact function in cancer remains unelucidated. In this review, we discuss the role of ferroptosis in tumorigenesis and cancer therapy, providing an improved understanding of ferroptotic mechanism-based therapeutic approaches.

Ferroptosis is initiated by the loss of GPX4 activity, which is mediated by two distinct mechanisms (Figure 1). The first mechanism is the inhibition of system Xc^- (e.g., mediated by erastin), which indirectly inhibits GPX4 (Dixon et al., 2012, 2014). System Xc^- is a cystine/glutamate antiporter that imports extracellular cystine in exchange for intracellular glutamate (Bridges et al., 2012). Cysteine, a reduced form of cystine, is a precursor for the synthesis of GSH. By using GSH as an essential cofactor, GPX4 exerts its phospholipid peroxidase activity to catalyze the reduction of lipid peroxides. Thus, the inhibition of system Xc^- by small molecules causes GSH depletion and subsequently inactivates GPX4, ultimately leading to the accumulation of lethal lipid peroxides and the initiation of ferroptosis (Dixon et al., 2012, 2014; Yang et al., 2014). Notably, inhibition of γ-GCS (e.g., mediated by buthionine sulfoximine), which is the rate-limiting enzyme for the synthesis of GSH, is also sufficient to initiate ferroptosis (Yang et al., 2014). Another molecular mechanism of ferroptosis is the direct inhibition of GPX4 through the loss of its activity or promoting its degradation. The classical ferroptotic inducer (1S, 3R)-RSL3, can inhibit GPX4 enzymatic activity by covalently targeting its active site selenocysteine in an irreversible manner (Yang et al., 2014, 2016). FIN56, a new and specific ferrotosis inducer, is involved in decreasing GPX4 abundance. However, the detailed mechanism remains unclear, maybe related to the enzymatic activity of acetyl-CoA carboxylase (Shimada et al., 2016). Furthermore, genetic inhibition of GPX4 by siRNA was also sufficient to induce the accumulation of lipid ROS and ferroptotic cell death (Yang et al., 2014). In general, GPX4 is a central regulator of ferroptosis-triggering mechanisms. Given that the iron-dependent accumulation of lipid ROS kills cells undergoing ferroptosis, both iron metabolism and lipid peroxidation are two critical processes involved in the mechanism of ferroptosis. Therefore, the following contexts were mainly carried out in these two aspects (Figure 1).

Cancer cells can undergo several regulated forms of cell death during tumor development, including apoptosis (Kerr et al., 1994), autophagy (Kondo et al., 2005; Levine, 2006), and necrosis (Amaravadi and Thompson, 2007). Unsurprisingly, ferroptosis also plays a role in the development of cancer and maybe a beneficial strategy for anticancer treatment.

Different lines of evidence suggest that ferroptosis plays a crucial role in the suppression of tumorigenesis. Specifically, GPX4 knockdown using siRNAs decreases the level of GPX4 protein, a central mediator of ferroptosis, inducing renal cell carcinoma cell death with accompanying lipid ROS generation (Yang et al., 2014). This process can be rescued by the iron chelator DFO and the antioxidant vitamin E. Moreover, inhibition of a member of cystine/glutamate antiporter, SLC7A11, can induce ferroptosis (Dixon et al., 2012). SLC7A11 is highly expressed in human tumors (Huang et al., 2005; Liu X. X. et al., 2011), and its overexpression rescues cancer cells from ferroptosis. Both the pharmacological and genetic inhibition of SLC7A11 induce ferroptotic cell death and enhance cisplatin cytotoxicity in cisplatin-resistant head and neck cancer (HNC) cells in vitro and in vivo (Roh et al., 2016). In addition, a recent study showed that p53 inhibits tumor growth partially by repressing the expression of SLC7A11 and subsequently inducing ferroptosis in addition to traditional p53-mediated functions, such as cell cycle arrest, senescence, and apoptosis (Jiang et al., 2015). Notably, p53^3KR, an acetylation-defective mutant that fails to induce cell cycle arrest, senescence and apoptosis, fully retains its ability to regulate SLC7A11 expression and induce ferroptosis upon ROS-induced stress (Jiang et al., 2015). p53^3KR mice did not succumb to the early-onset spontaneous thymic lymphoma formation that is commonly observed in p53-null mice (Li et al., 2012). Consistently, unlike p53^-/-XRCC4^-/- mice, p53^3KR/3KRXRCC4^-/- mice did not succumb to pro-B cell lymphomas (Li et al., 2016). This suggests that p53^3KR retains its tumor-suppressive activity and that uncanonical functions of p53 play crucial roles in the suppression of tumorigenesis. Taken together, these findings indicate that ferroptosis serves as a critical barrier to cancer development.

Accumulating evidence has shown that the activation of ferroptosis contributes to anticancer treatment for several forms of human cancer (Table 1). Erastin, a RAS-selective lethal compound (Dolma et al., 2003) that triggers ferroptosis by directly inhibiting cystine/glutamate antiporter system Xc^- activity, displays synthetic lethality in engineered cells. However, erastin does not show selective lethality in RAS-mutated cancer cell lines over RAS wild-type counterparts in a large, diverse panel of cell lines. Analysis of erastin sensitivity data from 117 cell lines and a 60-cancer cell line panel (NCI 60) harvested from eight diverse tissues (leukemia, lung, colon, CNS, melanoma, ovarian, kidney, and breast) revealed that diffuse, large B cell lymphomas and RCCs are particularly sensitive to ferroptosis (Yang et al., 2014). Moreover, low-cytotoxic doses of erastin remarkably enhance the anticancer activities of 2 first-line chemotherapy drugs, cytarabine and doxorubicin, in HL60 cells (Yu et al., 2015). Given that AML has a high recurrence rate, inducing ferroptosis may overcome drug resistance in AML cells.

Sorafenib, a multikinase inhibitor, is clinically approved for the treatment of renal cell carcinoma and other indications. Recently, sorafenib was identified as a ferroptosis inducer by inhibiting system Xc^- function and triggering subsequent ferroptotic cell death (Dixon et al., 2014). However, acquired resistance to sorafenib has been found in HCC patients, which results in poor prognosis. Previous studies show that NRF2 (Sun et al., 2016) and MT-1G (Houessinon et al., 2016), two negative regulators of ferroptosis that function by blocking GSH depletion-mediated lipid peroxidation in HCC cells, play a central role in protecting HCC cells against ferroptosis and result in sorafenib resistance. Importantly, the genetic and pharmacological inhibition of NRF2 or MT-1G in HCC cells enhances the anticancer activity of sorafenib in vitro and in tumor xenograft models.

In addition, numerous molecules, such as RSL3, RSL5, ART (Eling et al., 2015), DHA (Lin et al., 2016), and a series of small-molecule inducers (FINs) (Weiwer et al., 2012; Yang et al., 2012), have been shown to introduce ferroptotic cell death in cancer cells. Co-treatment with the ferroptosis inhibitor ferrostatin-1 can block ART-induced lipid peroxidation and cell death and increase long-term cell survival and proliferation.

As discussed above, ferroptosis participates in cancer development and treatment response, contributing to tumor suppression activity. Although many experiments in vitro have shown that ferroptosis can be detected via several measuring methods, such as cell viability, iron level and ROS level, it would be harder to prove the presence of ferroptosis in vivo. So far, PTGS2, a gene encoding cyclooxygenase-2 (COX-2), is the unique widely used marker for erastin or RSL-3-induced ferroptosis in tumor (Yang et al., 2014).

Activation of regulated cell death is an anticancer treatment strategy. Despite success in clinical cancer treatment, drug resistance to existing chemotherapeutic agents due to genetic alterations remains a problem (Pommier et al., 2004; Liu L. et al., 2011; Sui et al., 2013). Considering the clinical limitations and the discovery of ferroptosis (Dixon et al., 2012), activation of a regulated non-apoptotic form of cell death, ferroptosis, may provide new drug targets.

As described above, compelling evidence suggests that the activation of ferroptosis is an effective, novel pathway for cancer intervention. In principle, ferroptosis activation could be achieved by (i) the identification of small molecules as ferroptosis inducers using systematic screening and structural basis or by (ii) the inhibition of key molecules related to ferroptosis, such as the cystine/glutamate antiporter SLC7A11. We will comment here on the avenues in which progress can be made in cancer treatment by ferroptosis activation.

The data described herein clearly show that ferroptosis is an iron-dependent form of non-apoptotic cell death and an ideal target for biochemical studies. ROS-induced ferroptosis contributes to tumor growth suppression and chemotherapy sensitivity. Because the activation of ferroptosis is widely recognized as a new target of drug discovery, an increasing number of small molecules have been identified to induce ferroptosis directly or indirectly by targeting iron metabolism and lipid peroxidation, and other studies have focused on the inhibition of key molecules related to ferroptosis. Meanwhile, whether compounds targeting ferroptotic cell death regulators have high specificity and minimal side effects in preclinical and clinical settings and what kinds of cancer prefer to be targeted by ferroptosis induction remain to be elucidated.

In summary, an improved understanding of the ferroptosis mechanism and the role of ferroptosis in cancer will create new opportunities for diagnosis and therapeutic intervention.

BY and JC conceived, designed the conception of review article, and made the amendments of the paper. BL conducted the paper. XC, MY, and QH collected the related research articles.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer RP and handling editor declared their shared affiliation.