Iron overload is a prerequisite of ferroptosis. The erastin-induced ferroptosis was inhibited by deferoxamine (DFO, an iron chelator), evidenced by increased cell viability and decreased lipid ROS production in HT-1080 cells. In contrast, the erastin-induced ferroptosis was triggered by incubation with three different exogenous iron supplements (11). In intestinal ischemia/reperfusion-induced acute lung injury model of C57BL/6 mice, the injection of Fe (15 mg/kg) aggravated lung injury and pulmonary edema, while the injection of ferrostatin-1 (Fer-1, 5 mg/kg) rescued this injury (17). Iron transport involves import, storage, and export (18, 19). Circulating iron exists in the form of ferric iron (Fe³⁺) by binding to transferrin (Tf). Fe³⁺ enters the endosome through membrane protein transferrin receptor 1 (TfR1). Then, Fe³⁺ is reduced to ferrous iron (Fe²⁺) by the iron reductase activity of the six-transmembrane epithelial antigen of the prostate 3. The divalent metal transporter 1 (DMT1, also known as SLC11A2) releases Fe²⁺ from the endosome into the cytoplasm. While part of the Fe²⁺ in the cytoplasm is stored as ferritin, part is oxidized to Fe³⁺ and transported outside the cell by the membrane protein iron transporter ferroportin (FPN, an iron efflux pump, also known as SLC11A3), and the rest is stored in the labile iron pool (LIP) of the cytoplasm or mitochondria (20). The iron in LIP spontaneously undergoes redox reactions, namely, Fenton and Harber Weiss reactions, to generate ROS, which in turn leads to lipid peroxidation (21). Moreover, iron and iron derivatives, such as heme or [Fe-S] clusters, also affect ferroptosis as they act on the active centers of ROS producing enzymes, such as lipoxygenase (LOX), cytochrome P450, NADPH oxidase and so on (22). Therefore, iron overload is essential for ferroptosis. Maintaining the LIP within a relatively narrow concentration range is crucial for preventing ferroptosis. Using DFO could prevent cell death caused by erastin and RSL3 (11, 23).

In this process, the core negative regulators of ferroptosis are ferritin heavy chain (FTH) (24–27) and FPN (28), and the core positive regulators of ferroptosis are Tf (29–31), TfR1 (24, 26, 31–33), and DMT1 (24, 34, 35). These iron homeostasis proteins involved in iron uptake, storage, utilization, and efflux from cells are regulated by IREs/IRPs (36, 37). In 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced PD mice models, apoferritin inhibited ferroptosis by downregulating the iron importers DMT1 and FSP1, and conversely upregulating long-chain acyl-CoA synthetase 4 (38). The lipopolysaccharide then increased the expression of nuclear receptor co-activator 4, which directly interacted with ferritin and degraded ferritin in a ferritin phagocytosis-dependent manner. It then released a large amount of iron (39). In addition, heme oxygenase 1 (HO-1) mediates the release of free iron from heme, resulting in the accumulation of Fe²⁺ in LIP, which also exacerbates ferroptosis (40, 41).

Fe²⁺ in LIP can spontaneously undergo redox reactions to produce ROS, including both Fenton and Harber-Weiss reactions. The chemical equations are as follows (42):

The electron transport system of mitochondria is the primary source of H₂O₂ and active oxygen (O₂^•–) (43). ROS, generated by the above mechanism, causes damage to the biomembrane in two ways. One is an enzyme-independent way, that is: Firstly, the hydroxyl radical (OH^•) combines with polyunsaturated fatty acids (PUFA) on the biomembrane to generate the phospholipid radical (PL^•). Secondly, PL^• reacts with O₂ generating a phospholipid peroxyl radical (PLOO^•). Thirdly, PLOO^• reacts with PUFA to generate phospholipid hydroperoxide (PLOOH) and PL^•, which can react again with O₂, forming a vicious circle. The other is the enzymatic way, that is, PUFA generates PLOOH under the action of LOXs. However, the detailed mechanism of PLOOH resulting in ferroptotic lipid peroxidative cell death remains obscure. Continued oxidation and consumption of PUFA may alter the structure of lipid pores, ultimately leading to compromised membrane integrity. In addition, PLOOH may be decomposed into active toxic aldehydes, such as 4-hydroxy-2-nonenal or malondialdehyde (MDA), causing cytotoxic effects (44). The hallmark of ferroptosis is the iron-dependent accumulation of lipid hydroperoxides to cell-lethal levels, especially peroxidized phosphatidylethanolamine (PEox). However, only a few studies detected and quantified these directly. In heart transplantation mice models, hydroperoxy-arachidonoyl-phosphatidylethanolamine (HOO-C20:4/C18:0-PE) was elevated and the resulting ferroptosis triggered early inflammation by recruiting neutrophils. In IRI mice models, the abundance of several hydroxyeicosatetraenoic acids (HETE) (such as 5-HETE, 11-HETE, 12-HETE, and 15-HETE) and epoxyeicosatrienoic acid species were increased (45). Besides, in RSL3-induced ferroptosis in H9C2 cardiomyocytes, via LC/MS, three significant species of hydroperoxy-PE were found to be up-regulated, namely, PE(36:4)-OOH, PE(38:4)-OOH, and PE(40:4)-OOH (46). Sparvero et al. used gas cluster ion beam secondary ion mass spectrometry imaging with a 70 keV (H₂O) (_n) (+) (n > 28000) cluster ion beam to visualize them at the single-cell and subcellular levels (47). The ferroptosis inhibitor, Fer-1 inhibited ferroptosis by reducing PLOOH.

The cyst(e)ine-GSH-GPX4 pathway is considered the canonical pathway for restricting ferroptosis. System Xc^–, a heterodimeric 12-pass transmembrane cystine–glutamate anti-porter, consists of a xCT light chain (also known as SLC7A11) that mediates cystine transport specificity, and a 4F2 heavy chain (also known as SLC3A2) (48), plays an important role in this process. SLC7A11 transports extracellular cystine, which is rapidly reduced to cysteine (Cys) by an NADPH-consuming reduction process. Cys participates in the production of GSH, a fundamental component of GPX4 (49). GPX4 is the primary inhibitor of ferroptosis. It can prevent lipid peroxidation by reducing PLOOH to non-toxic phospholipid alcohols (49). Cardiac impairments were ameliorated in GPX4 Tg mice and exacerbated in GPX4 heterodeletion mice. In cultured cardiomyocytes, GPX4 overexpression prevented DOX-induced ferroptosis (50). Surprisingly, a recent study showed that in non-small-cell lung cancer cell lines, cystine starvation induces an unexpected accumulation of γ-glutamyl-peptides under the influence of glutamate-cysteine ligase catalytic subunit, which limits the accumulation of glutamate, thereby protecting against ferroptosis (51). In addition, methionine can be used as one of the sources of intracellular cystine through the trans-sulfuration pathway (49). GPX4 is a selenoenzyme and its biosynthesis relies on the co-translational incorporation of selenocysteine (49). Selenium augments GPX4 and other genes by enhancing adaptive transcription factors TFAP2c and Sp1 to protect the cells from ferroptosis (52).

The FSP1-CoQ10-NADPH pathway exists as a GPX4-independent one. FSP1 catalyzes the transformation of CoQ10 into ubiquinol, which is an excellent radical-trapping antioxidant in phospholipids and lipoproteins (53–55). Furthermore, the pathway can reduce oxidized α-tocopheryl radical to its non-radical form, increasing antioxidant capacity (56). The MDM2-MDMX complex is a negative regulator of FSP1. It changes the activity of PPARα, resulting in a decrease in the level of FSP1 protein and an increase in the level of CoQ10 (57). MiR-4443, whose target gene is METLL3, inhibited FSP1-mediated ferroptosis induced by cisplatin treatment in vitro and enhanced tumor growth in vivo (58).

The GTP-GCH1-BH4 pathway is also not dependent on GPX4. BH4/dihydrobiopterin synthesis by GCH1-expressing cells caused lipid remodeling, suppressing ferroptosis by selectively preventing depletion of phospholipids with two polyunsaturated fatty acyl tails (59). Using a co-culture model system, iNOS/NO (^•) in M1 macrophages has been confirmed to inhibit the effect of NO (^•) on epithelial cells by inhibiting phospholipid peroxidation, especially the generation of 15-HpETE-PE signal that promotes ferroptosis. It is an intercellular mechanism that distantly prevents the ferroptosis of epithelial cells stimulated by Pseudomonas aeruginosa (60).

Nuclear factor (erythroid-derived 2)-like 2 signaling is implicated in many molecular aspects of ferroptosis, including glutathione homeostasis, mitochondrial function, and lipid metabolism (61, 62). The Nrf2-Focad-Fak signaling pathway is closely related to ferroptosis caused by Cys deprivation. In non-small-cell lung carcinoma, brusatol (an Nrf2 inhibitor) was added based on ferroptosis inducer erastin or RSL3. The therapeutic effect based on ferroptosis was better than single treatment in vivo and in vitro (63). In immunocompetent mice and humanized mice, ZVI-NP, a dual-functional nanomedicine, enhanced the degradation of Nrf2 by GSK3/β-TrCP through AMP-activated protein kinase (AMPK)/rapamycin activation, leading to cancer-specific ferroptosis of lung cancer cells (64).

The antioxidant system of the heart is very complex. In addition to the above three major systems, the antioxidant system of the heart also includes some other elements that inhibit ferroptosis. O₂^•–, OH^•, OH^–, H₂O₂, PL^•, PLOO^•, PLOOH, ROS, etc. play important roles in the occurrence and development of ferroptosis. Superoxide dismutase (SOD) and superoxide reductases can reduce O₂^•– to H₂O₂. Catalase catalyzes H₂O₂ to water and O₂. Water-soluble ascorbic acid (vitamin C), lipid-soluble vitamin E or α-tocopherol (α-TOH), and lipoic acid can reduce lipid hydroperoxide production and peroxyl radicals. Besides, ascorbic acid increases the vitamin E content by reducing vitamin E semiquinone (65). The thioredoxin system, consisting of the thioredoxin (Trx) and thioredoxin reductase (TrxR), is also an important antioxidant system (66). Ferroptocide causes an accumulation of lipid peroxidation by inhibiting this system, thereby inducing ferroptosis (67). TrxR and NADPH reduce the active site disulfide in Trx. Under the combination of Trx and TrxR, peroxides, including lipid hydroperoxides and H₂O₂ were observed to be reduced effectively. In addition, there are several crosstalks between these antioxidants. For example, the thioredoxin system promotes the regeneration of certain antioxidants. It reduces ascorbyl free radical to ascorbic acid and turns GSSG to GSH. The thioredoxin system increases the content of ascorbic acid by reducing dehydroascorbic acid (65).

We know that ferroptosis is an iron-dependent lipid peroxidation process, and iron is essential in both the occurrence and development of this process. It is generally believed that excessive iron can aggravate DIC. More than 20 years ago, research indicated that iron overload aggravated DIC (80). DOX reduced the viability of H9C2 cardiomyocytes, while ferric ammonium citrate aggravated it in a concentration-dependent manner (81). Male Sprague Dawley rats fed with iron-rich chow showed significantly higher DOX cardiotoxicity, accompanied with a significant weight loss and severe myocyte injury as evidenced through electron microscopy and light microscopy. However, feeding an iron-rich meal alone did not result in any cardiotoxicity (82). Paradoxically, in cultured H9C2 cardiomyocytes and male C57BL/6 mice, researchers concluded that pretreatment with dextran-iron (125–1,000 μg/mL) in combination with DOX did not potentiate DIC and even prevented some aspects of it (83). Therefore, the regulation of DIC by iron is a complicated process, which may involve the balance between iron dosage, protection, and damage.

The HFE gene encodes HFE protein, which binds to TfR1 and facilitates the uptake of iron-bound to Tf. The elevation of iron concentration in the heart was much more accentuated in DOX-treated HFE^–/– mice. Mutations in the HFE gene led to iron overload in cardiomyocytes, which increased the susceptibility of cardiomyocytes to ferroptosis and exacerbated DIC (84). One study concluded that the mutation status of HFE RS1799945 H63D could be used as one of the critical markers to identify patients at high risk for AIC (85). Among survivors of high-risk acute lymphoblastic leukemia in children, patients with the C282Y mutation of the HFE gene had a more severe DIC, which was reflected in higher levels of cardiac troponin-T, lower left ventricular quality and thickness, and worsened left ventricular function in echocardiography (86). These studies on the iron-regulated gene HFE support the view that iron plays a pivotal role in DIC.

Proteins associated with iron transport are considered significant indicators of cellular iron homeostasis and are mainly involved in cellular iron uptake (TfR1) and storage (ferritin). As for TfR1, most studies believe that DOX could elevate its expression (87–89). A study believed that TfR1 is critical for the DOX-induced increase in iron uptake. After incubation with the specific anti-TfR antibody (12 μg/ml), DOX-induced increase of ⁵⁵Fe uptake in bovine aortic endothelial (BAEC) cells was reversed (88). The mechanism of increased TfR1 could be because DOX inhibited the expression of miR-7-5p by increasing the METTL14-mediated expression of KCNQ1OT1m6A, thus reducing the degradation of TfR1 (89). In AC16 cells, METTL14 knockdown, KCNQ1OT1 silencing, and miR-7-5p mimic attenuated the DOX-induced increase of Fe²⁺ and lipid ROS, and reduced DOX-induced decrease in mtDNA and MMP levels. In METTL14 shRNA mice models, DOX-induced increase in the levels of MDA and 4-HNE were also alleviated (89). However, one study suggested that TfR1 was reduced in heart lysates extracted from DOX-treated mice that received a single i.p. DOX dose (20 mg/kg) as observed in Western-blot and RT-PCR analysis (90). As for ferritin, DOX elevated it, and this change was accompanied by an increase in the level of iron bound to it (90). After being treated with 5 and 10 μM DOX for 24 h, the content of ferritin, especially ferritin heavy chain (FHC), in H9C2 cells increased, and was ROS-dependent. The increase in DOX-induced ferritin was reversed after clearing ROS using N-acetylcysteine (a known ROS scavenger) in H9C2 cells. Interestingly, the increased ferritin induced by DOX protected H9C2 cells from iron toxicity, as demonstrated by increased cell viability in 500, 750, and 1,000 μg/ml FAC measured by MTT assays (91). Similarly, ferritin mRNA and protein levels were also increased in an adult cell line of cardiomyocytes (HL-1) exposed to 5 μM DOX (92). In addition, elevated intracellular iron levels were also associated with high mobility group box 1 (HMGB1)-mediated heme degradation. Compared with the DOX group, the ferroptosis-related indexes (PTGS2, MDA, and 4-HNE) and cardiac injury-related indexes (Anp, Bnp, and Myh7) of the rat heart in the DOX + shHMGB1 group were significantly decreased (73). Therefore, the effect of DOX on iron homeostasis regulatory proteins is a complex process, possibly related to cell lines, drug dosage, and imbalances in protection and injury.

At present, many studies have shown that DOX can act on IRPs, but the results are numerous. Some studies believed that DOX irreversibly inactivated IRP1 and IRP2. This study believed that the secondary alcohol doxorubicinol (DOXol) and certain products of DOX metabolism, converted cytoplasmic aconitase to the cluster-free IRP1 by removing iron from its catalytic [Fe-S] clusters. This eventually produced a null protein, which could not adapt the levels of TfR1 and ferritin, and IRP2 was inactivated only by DOX related ROS (93). Some scholars thought H₂O₂ activated IRP1, leading to the upregulation of TfR1 and cellular iron accumulation, which was considered an important molecular mechanism in DIC (94). One study demonstrated that the IRP1 activity of BAEC cells increased in a dose-dependent manner after being treated with 0.5 μM DOX (88). However, one study found that even exposing DOX-sensitive GLC4 cells to 12.5 μM DOX for 24 h did not change IRP activity. It can be seen that the effect of DOX on cellular IRP activity may be related to cell line and cell resistance (95). However, Juliana et al. believed that the effect of DOX on cardiomyocyte IRP activity was closely related to DOX concentration and incubation time. After 6 h of DOX administration, total IRP1 did not change significantly, but active IRP1 and active IRP2 decreased in a concentration-dependent manner (1 μM DOX, 5 μM DOX, 10 μM DOX, and 20 μM DOX). IRP2, in particular, dropped by more than 50% at 20 μM DOX. After 24 h of administration, while the total IRP1 level decreased, the active IRP1 and active IRP2 did not significantly reduce but even increased. Interestingly, they thought that the free radical scavengers, DOXol, and cis-aconitate had little effect on IRP-RNA-binding activity in SK-Mel-28 melanoma cells and cardiomyocytes. The Fe and Cu complexes of anthracyclines altered iron metabolism in cardiomyocytes (96). Gianfranca suggested that DOX had differential effects on the two IPRs, as evidenced by reduced IRP2 activity and unchanged IRP1 activity. The effect of DOX on IRP2 resulted in an upregulation of ferritin gene expression and a decrease in TfR1 expression. Surprisingly, this change reduced the iron content in LIP and protected cardiomyocytes from ferroptosis induced by DOX (97). Furthermore, one study concluded that DOX at low concentrations (≈1 μM) activated IRP1 in cardiomyocytes, while at higher concentrations (>5 μM), it irreversibly inactivated IRP1 in BAEC cells (88). Besides, according to one study, DOX can directly interact with IREs. DOX intercalated double-stranded RNA by recognizing the IREs hairpins located in the 50-UTR of ferritin mRNAs, thereby changing the tertiary structure of the RNA drastically altering the effectiveness of the IREs/IRPs interaction (98). Anyway, DOX could modify the expression of genes involved in iron metabolism by inactivating IRPs binding to IREs (99).

Under normal physiological conditions, Nrf2 binds to Keap1 in the cytoplasm, and then Nrf2 is degraded by the ubiquitin-proteasome system. In the case of oxidative stress, Nrf2 dissociates from Keap1 and translocates to the nucleus, where it binds to promoter regions (AREs), activates the transcription of a series of downstream genes, and exerts physiological functions (100). Notably, many of the genes associated with ferroptosis are target genes for Nrf2, but DIC-related studies mainly focus on GPX4 and HO-1 genes. Nrf2 up-regulates the expression of GPX4 and has an anti-ferroptosis effect (101–103). Nrf2 can be methylated by the protein arginine methyltransferase 4 (PRMT4), leading to its nuclear restriction and consequently decreased GPX4 expression. PRMT4 aggravated the expression of ferroptosis markers (ROS, MDA, NCO4, and Fe²⁺) in the DOX-induced primary neonatal rat ventricular myocytes and C57BL/6 J mice cardiotoxicity models, and this influence can be mitigated by PRMT4 knockout (103). However, studies have also shown that Nrf2-mediated activation of HO-1 promotes ferroptosis. HO-1 mediates the release of Fe²⁺ from heme, which accumulates in cardiomyocytes and induces ferroptosis (74, 104). Through Nrf2^+/+ mice, Nrf2^–/–mice, Znpp (an HO-1 inhibitor), Hemin (an HO-1 activator), DOX has been proven to increase HO-1 by affecting Nrf2 and accelerating the degradation of heme, leading to an increase in non-heme iron and myocardial ferroptosis. An increase in intestinal iron absorption did not accompany this increase of iron content. In this study, DOX was confirmed to be able to upregulate hepatic Hamp1 mRNA to increase hepcidin and reduce FPN degradation (74). In addition, Nrf2 can be activated by deacetylation of SIRT1, and Fisetin activated Nrf2 by up-regulating the expression of SIRT1, leading to the up-regulation of HO-1, FTH1, TfR1, and FPN, and exerting an anti-ferroptosis effect. After being transfected with SIRT1 and Nrf2 siRNA, the anti-ferroptosis effect of Fisetin was abolished in H9C2 cells (105).

The morphological changes of ferroptosis under the electron microscope were mainly observed in the mitochondria (11). The metabolic activity of mitochondria drives ferroptosis. Mitochondria are the main source of cellular ROS. When electrons are transferred to O₂, some escape from the ETC and react directly with O₂ to form O₂^•–, a precursor to many other ROS such as OH^• and H₂O₂ (106). Energy stress inhibits ferroptosis by activating AMPK, while AMPK inactivation promotes ferroptosis (107). ETC complex inhibitors can inhibit ferroptosis, indicating that mitochondria play an important role in ferroptosis, and possibly through the activation of AMPK (108). The mitochondrial TCA cycle is involved in ferroptosis induced by Cys deprivation, for which glutaminolysis is essential. However, mitochondria are not essential for ferroptosis induced by GPX4 inhibition (108).

Doxorubicin can cause cell iron metabolism disorders, but there is much debate about where iron metabolic disorders occur. Some studies have suggested that the site where the DIC occurs is the mitochondria. Compared with cytoplasm, DOX and iron preferentially accumulated in mitochondria (109), especially in mitochondrial cardiolipin (110–112). After incubation with 10 μM DOX, the accumulation of DOX in the mitochondria of neonatal rat cardiomyocytes increased significantly compared with the cytoplasm, and DOX caused a significant increase in mitochondrial iron levels detected by ⁵⁵Fe colorimetric measurement of mitochondrial non-heme iron (109). In the presence of Fe²⁺, DOX induced the activation of the mitochondrial permeability transition pore (113). Tadokoro et al. believed that ferroptosis was triggered by GPX4 deficiency in mitochondria for the following reasons: (1) DOX-induced lipid peroxidation occurred on mitochondria rather than other organelles; and (2) even though the cell viability was improved and lipid peroxidation indexes (MitoPeDPP, MDA) were reduced in both isolated neonatal rat ventricular cardiomyocytes cells transfected with Ad-cytoGPx4-FLAG and Ad-mitoGPx4-FLAG, electron microscopy revealed that mitochondrial GPX4 was almost exclusively localized to mitochondria during this process. In contrast, cytoplasmic GPX4 was transferred to mitochondria (50). Mitochondria-2,2,6,6-tetramethylpiperidin-N-oxyl, a mitochondrial superoxide scavenger, abolished DOX-induced lipid peroxidation and cardiac ferroptosis in DOX-treated mice models. In contrast, the non-mitochondrial targeted version only mildly rescued the DOX-induced effects (74). Dexrazoxane (DXZ) reduced iron in mitochondria. The poor impact of DFO in treating DIC compared with DXZ may be attributed to its inability to penetrate mitochondria and chelate iron in mitochondria specifically (109, 114). The most likely target of free radicals produced by ANTs through redox reactions was cardiolipin, a major phospholipid component of the inner mitochondrial membrane, known to be susceptible to peroxidative injury with abundant PUFA. Iron overload can aggravate the damage of ANTs to the mitochondria of cardiomyocytes (115).

The mechanism of DOX causing myocardial ferroptosis may be related to its effect on mitochondrial iron regulation-related proteins. Compared with wild-type mice, cardiotoxicity was more pronounced in FtMt^–/– mice injected intraperitoneally with a single dose (15 mg/kg of body weight) of DOX, as manifested by higher mortality, more morphological changes (incomplete cristae, condensation, and fragmentation of most myofibril), more severe lipid peroxidation, and worse cardiac function (ATP and BNP) (116). In mice, neonatal cardiomyocytes, and H9C2 cardiomyoblasts, DOX led to the reduction of frataxin, a nuclear-encoded mitochondrial protein involved in maintaining mitochondrial iron homeostasis through the ubiquitin-proteasome system. In addition, the mitochondrial iron export protein ABC protein-B8 (ABCB8) is essential for maintaining mitochondrial iron homeostasis. The depletion of ABCB8 led to compromised systolic and diastolic functions, a significant accumulation of ⁵⁵Fe in the mitochondria, and higher lipid peroxidation levels both in vivo and in vitro (87, 109, 117). The levels of ABCB8 in explanted hearts from patients with end-stage cardiomyopathy were significantly reduced (117). Intriguingly, one study concluded that there was no effect on ABCB8 in the myocardium of the DIC mice models, and the silencing of ABCB8 did not increase the iron content in cultured cardiomyocytes after 30 h of exposure to 2 μM DOX (50).

On the contrary, another study showed that DOX affected more cytosolic than mitochondrial iron metabolism in murine hearts and human HeLa cells, as manifested by alterations in proteins associated with cytoplasmic iron transport proteins (ferritin, TfR1, and hepcidin). In contrast, mitochondrial iron-related proteins (aconitase, succinate-dehydrogenase, and frataxin) appear to be unaffected (90). In addition, Kwok thought that the mechanism of DOX-induced cardiomyocyte ferroptosis was related to lysosomes. It may be that ANTs act on lysosomes to inhibit ferritin degradation, thereby affecting the iron-dependent life activities of cells. However, this study was conducted in SK cells, not cardiomyocytes, so the conclusion is debatable (118). Interestingly, some studies believed that DOX affected neither total cellular Fe content nor total cellular ferritin protein levels (119). Furthermore, the myocardial iron content was not statistically different between the saline and the DOX groups (82). In the presence of NADPH-cytochrome p450 reductase, ANTs underwent redox cycling to generate superoxide, which mediated a slow reductive release of iron from ferritin. More the cardiotoxic ANTs, more the rapid and extensive iron release it led to (120).