Targeting ferroptosis opens new avenues for the development of novel therapeutics

1Ferroptosis is a unique form of iron-dependent regulated cell death originally identified by screening RSL (RAS-selective lethal) compounds. The key characteristic of ferroptosis is an extensive accumulation of lipid peroxides, making ferroptosis genetically, morphologically, and biochemically distinct from other types of cell death. In the decade since ferroptosis was first reported, increasing evidence has emerged suggesting that ferroptosis may play a pivotal role in many biological processes such as tumor suppression and immunity, indicating that ferroptosis is important for maintaining health by regulating metabolism and redox homeostasis. Recently, a plethora of studies have shown that ferroptosis also plays a critical role in a variety of pathophysiological processes such as ischemic organ injury, stroke, cardiac myopathy, and neurodegenerative diseases; ferroptosis has also been implicated in many oncogenic pathways, suggesting it might serve as a target for novel cancer therapeutics.

2In mammalian cells, ferroptosis is regulated mainly by iron homeostasis, lipid metabolism, and glutathione-dependent redox balance. Thanks to rapid progress in the study of ferroptosis, numerous efforts have made to identify potent, druggable ferroptosis modulators for use in clinical applications, opening new avenues for developing novel treatment strategies to target many ferroptosis-related diseases, including cancer and heart injury. With respect to cancer, novel ferroptosis agonists have shown promise in several cancer types. On the other hand, antagonists of ferroptosis have been shown to help alleviate ferroptosis-related diseases such as ischemia/reperfusion (I/R)-induced damage, neurodegenerative diseases, and inflammatory diseases.–

3Here, we provide a comprehensive update of recent efforts to target ferroptosis for treating a variety of relevant diseases, and we discuss the safety and efficacy of ferroptosis agonists and antagonists, as well as their pharmacokinetics profiles, their experimental stage in the drug discovery process, and opportunities for further clinical development. We also review the biological roles, molecular mechanisms, and clinical implications of ferroptosis, and we discuss insights into ferroptosis-focused translational research, as well as current knowledge gaps and opportunities. Finally, we discuss future research orientations that will lead to the clinical implementation of these novel ferroptosis modulators for treating a diverse range of diseases.

4Although the term “ferroptosis” was first coined by Dixon et al. in 2012 (Fig. 1a), a similar form of neuronal cell death triggered by cysteine depletion called “oxytosis” was reported by Murphy et al. back in 1989; this neuronal cell death was induced by the excitotoxin glutamate by inhibiting SLC7A11 (solute carrier family 7 member 11), a component of the cystine-glutamate antiporter system Xc− known to play a role in ferroptosis. Therefore, oxytosis and ferroptosis have been suggested to share several key characteristics, including their gene expression patterns, high activity of lipoxygenases, and high accumulation of reactive oxygen species (ROS). By 2003, a distinct form of erastin-induced cell death in RAS-expressing cancer cells was receiving considerable attention; this form of cell death was not inhibited by caspase inhibitors but could be suppressed by treating the cells with iron-chelating agents. Yang et al. subsequently found that RAS-selective lethal small molecule 3 (RSL3) could also trigger this iron-dependent form of cell death. Based on these characteristics, Dixon et al. named this type of cell death ferroptosis. Morphologically, cells undergoing ferroptosis generally have shrunken mitochondria with increased membrane density and reduced—or absent—mitochondrial cristae. Biochemically, ferroptosis is characterized by increased oxidative stress and depleted antioxidative defense. Although ferroptosis is generally well-accepted as a tightly regulated cellular metabolic process, its precise mechanisms such as how excessive phospholipid (PL) peroxidation drives ferroptosis and the tissue- and disease-specific epilipidome signatures remain unknown.

5The mechanisms that regulate ferroptosis have begun to emerge (Fig. 1a). For example, Yang et al. found that the glutathione/glutathione peroxidase 4 (GSH/GPX4) pathway serves as an important regulator of ferroptosis. Soon after, other groups identified ACSL4 (acyl-CoA synthetase long-chain family member 4) as a predominant mediator of ferroptosis., In addition, selenium was shown to protect against ferroptosis. In 2019, two groups independently reported that the FSP1 (ferroptosis suppressor protein 1)-CoQ10 (coenzyme Q10)-NAD(P)H pathway inhibits ferroptosis via a GPX4-independent mechanism., In addition, the GCH1 (GTP cyclohydrolase 1)-BH4 (tetrahydrobiopterin) pathway was reported as an additional GPX4-independent modulator of ferroptosis. Recently, Mao et al. identified the DHODH (dihydroorotate dehydrogenase)-CoQ10 axis, which is located at the mitochondrial inner membrane, as an important pathway that protects against ferroptosis. Even more recently, Mishima et al. reported that vitamin K is a potent FSP1-dependent inhibitor of ferroptosis by functionally screening vitamin compounds in Gpx4-deficient mouse embryonic fibroblasts.

6A growing body of evidence indicates that ferroptosis is regulated by a complex network involving iron homeostasis, lipid metabolism, and the oxidative-reductive system (Fig. 1b). For example, cellular ferroptosis is driven directly by an accumulation of toxic PL peroxides in membranes, which are produced primarily by the free iron‒induced Fenton reaction and oxidized PL-containing polyunsaturated fatty acids (PL-PUFAs).

7Because iron chelators such as deferoxamine (DFO) can block ferroptosis, ferroptosis was originally defined as iron-dependent. In contrast, iron overload has been shown to sensitize numerous cells types to ferroptosis.,– Therefore, maintaining iron homeostasis is critical in order to protect cells from ferroptosis. The intracellular labile iron pool (LIP) is regulated by the uptake, export, storage, and utilization of iron (Fig. 1b). Cellular iron uptake is regulated primarily by the transferrin/transferrin receptor (TF/TFR) system, which has been shown to play a role in ferroptosis. Recently, we reported that TF protects against ferroptosis-induced liver damage via SLC39A14 (solute carrier family 39, member 4)-mediated non-TF-bound iron (NTBI). Excess intracellular iron is sequestered largely by the principal iron-storage protein ferritin (FTH), which protects cells from ferroptosis. Consistent with this protective role, we found that mice lacking Fth in cardiomyocytes develop ferroptosis-induced cardiomyopathy. Notably, ferritinophagy—an NCOA4 (nuclear receptor coactivator 4)-mediated autophagic degradation of ferritin in the lysosome—has been shown to induce ferroptosis by releasing free iron from ferritin., Moreover, the iron exporter ferroportin (FPN) has also been shown to regulate cellular sensitivity to ferroptosis in vitro.

8The metabolism of lipids—particularly PL-PUFAs—critically regulates ferroptosis (Fig. 1b). Using a combination of genome-wide CRISPR screening and microarray analysis of ferroptosis-resistant cell lines, Doll et al. found that the enzyme ACSL4 is a major mediator of ferroptosis. ACSL4 catalyzes the conversion of free PUFAs to acyl-CoA derivatives (PUFA-CoAs), which can be further catalyzed by lysophosphatidylcholine acyltransferase 3 (LPCAT3) to produce PL-PUFAs. PL-PUFAs participate in the biosynthesis of cellular membranes and are highly susceptible to peroxidation due to their bis-allylic hydrogen atoms. Interestingly, loss of either ACSL4 or LPCAT3 increases cellular insusceptibility to ferroptosis by reducing the production of substrates for PL peroxidation.,, In addition, inhibiting lipoxygenases (LOXs), which are iron-containing enzymes, suppresses erastin-induced ferroptosis via iron- and oxygen-triggered free radical chain reactions, independent of their enzymatic activity.

9With respect to oxidative and reductive reactions, several major pathways are involved in protecting against ferroptosis (Fig. 1b). First, the well-characterized system Xc–-GSH-GPX4 axis serves as a GPX4-dependent mechanism for scavenging PL peroxides via system Xc–-mediated GSH synthesis., Moreover, inhibiting either component of system Xc– (i.e., SLC7A11 or SLC3A2) induces ferroptosis by disrupting cystine uptake, thereby limiting GSH synthesis. In this pathway, GPX4 serves as an essential regulator of ferroptosis. Second, in 2019 an FSP1-dependent metabolic pathway was shown to protect against ferroptosis via an GPX4-independent process., Specifically, two groups simultaneously reported that FSP1 traps lipid peroxyl radicals to suppress ferroptosis by reducing CoQ10 levels at the plasma membrane,, a process that is independent of the system Xc–-GSH-GPX4 pathway. Third, Mao et al. showed that mitochondrial inner membrane-located DHODH can protect against ferroptosis by reducing CoQ to form ubiquinol, an antioxidant that inhibits ferroptosis; thus, inhibiting DHODH in cancer cells potently activates ferroptosis in parallel with mitochondrial GPX4, independent of cytosolic GPX4 and FSP1. Finally, Liang et al. recently identified the PL-modifying enzymes MBOAT1 and MBOAT2 as novel sex hormone‒dependent ferroptosis inhibitors. Mechanistically, the authors showed that MBOAT1/2 suppress ferroptosis by remodeling cellular PLs to protect cells from ferroptosis independent of GPX4, providing novel ferroptosis-targeted therapeutic strategies to sensitize either estrogen receptor (ER) antagonists in ER-positive breast cancer and androgen receptor (AR) antagonists in AR-positive prostate cancer, respectively.

10Intracellular accumulation of iron and PL peroxides are two central biochemical events that occur during ferroptosis, and these processes have also been implicated in a wide range of diseases (Tables 1 and 2). Indeed, a plethora of studies suggest a putative causal connection between ferroptosis and the pathophysiological processes underlying many conditions and diseases, including cancer, neurodegeneration, I/R-induced tissue injury, acute kidney damage, and hematologic diseases (Fig. 2).

11Apoptosis was long considered the principal form of cell death in various cancer types; however, apoptosis-based cancer therapies have limited clinical benefit. Therefore, identifying novel cancer therapeutics that target pathways other than apoptosis is now an urgent unmet clinical need. Indeed, the notion of ferroptosis originated from lethal screening of compounds in RAS-mutated cancer cells. In support of this notion, ferroptosis has been functionally validated in cancer types with increased PL peroxidation, including hepatocellular carcinoma (HCC), pancreatic ductal adenocarcinoma, triple-negative breast cancer (TNBC),, and renal cell carcinoma., Thus, these specific cancer types may be more sensitive to ferroptosis-inducing compounds.

12It is important to note that ferroptosis has also been closely linked to resistance to various cancer treatments. For example, several studies found that cancer cells with a high-mesenchymal state are more resistant to a variety of cancer treatments, but are particularly susceptible to ferroptosis-inducing compounds., In addition, priming cancer cells with ferroptosis-inducing compounds can sensitize the cancer cells to subsequent immunotherapies. Taken together, these findings highlight the promise of targeting ferroptosis as a novel strategy for treating various forms of cancer.

13Iron deposition and lipid peroxidation are common pathophysiological features in various neurological diseases. Notably, the disease group known as neurodegeneration with brain iron accumulation (NBIA) refers to neurodegenerative disorders associated with altered iron metabolism in the lesion area and includes Alzheimer’s disease (AD), Parkinson’s disease (PD), and Huntington’s disease (HD). In addition, ablating or inactivating GPX4 promotes neuronal damage and neurodegeneration., In experimental models of degenerative brain conditions, both ferroptosis inhibitors and iron chelators were shown to improve outcome and prognosis.–

14With respect to AD, several studies have shown that dysregulated iron metabolism is linked to ROS production, mitochondrial dysfunction, and neurodegeneration. Iron deposition in the brain, accompanied by a decrease in endogenous antioxidant capacity, has been associated with the pathogenesis of AD, and iron levels in the brain have been correlated with disease progression. Furthermore, accumulated iron interacts with the Aβ peptide and tau protein via the formation of a peptide-hemin complex, possibly implicating ferroptosis.

15PD is characterized by reduced dopaminergic neurons selectively in the substantia nigra and the presence of Lewy bodies. In the substantia nigra pars compacta of affected patients, iron was shown to accumulate in glia cells and dopaminergic neurons, with the level of iron correlated with disease severity. Moreover, reports suggest that the increased oxidative stress present in PD may be attributed to decreased GSH levels in the substantia nigra. In addition, the brain consumes high amounts of oxygen, making this organ more sensitive to lipid peroxidation.

16HD, a hereditary neurodegenerative disease due to an abnormally high number of CAG repeats in the huntingtin (HTT) gene, has also been linked to ferroptosis in animal models. Excess iron remains a major cause of oxidative stress in neurons, which directly induces ferroptosis during the pathogenesis of HD. Interestingly, similar to PD, reduced levels of GSH have also been observed in HD. Together, these findings suggest that ferroptosis may be involved in the pathogenesis of HD.

17Ferroptosis was originally identified as the major pathogenic driver in several conditions related to organ injury. In 2014, mice lacking Gpx4 were shown to develop spontaneous acute renal failure and hepatic I/R-induced damage via ferroptosis. Subsequently, the pathogenic role of ferroptosis in the progression of I/R-induced acute damage in the lungs, heart, and intestinal tract has been validated in animal models.,, We previously systematically reviewed the varied role of ferroptosis in liver disease, and we refer the reader to this review. Notably, tissue-specific Gpx4 knockout mice present with tissue damage accompanied by massive death of photoreceptor cells and endothelial cells.

18Acute kidney injury (AKI) refers to sudden-onset kidney failure and/or kidney damage characterized by a rapid loss of the kidney’s excretory function caused by massive levels of cell death and inflammation. In mouse kidney tubular cells, Gpx4 has been reported to prevent AKI by blocking ferroptosis. Around the same time, Linkermann et al. independently found that ferroptosis plays a critical role in synchronizing kidney tubular cell death in both severe I/R-induced injury and oxalate crystal‒induced models of AKI; moreover, the authors showed that SRS 16–86—a novel third-generation ferroptosis-specific inhibitor—potently protected against AKI. Recently, Li et al. reported that ferroptosis is the primary form of cell death causing folic acid‒induced AKI in mice.

19In the heart, we first identified ferroptosis as the major pathogenic mechanism in both doxorubicin- and I/R-induced cardiomyopathy, and we showed that targeting ferroptosis significantly alleviated heart injury in mouse models. In addition, inhibiting glutaminolysis protected against I/R-induced ferroptotic heart damage in vitro, indicating that glutaminolysis plays an essential role in regulating ferroptosis during heart injury. Recently, Jacobs et al. suggested the possible presence of ferroptosis in the myocardium of a patient with COVID-19‒induced myocarditis. Using the antibody E06 to stain oxidized phosphatidylcholine, they found that a ferroptosis “signature” may be specific to injured cardiomyocytes in COVID-19‒induced myocarditis, as this signature was not present in either myocarditis of unknown etiology or in COVID-19 patients who presented without myocarditis. Recently, we summarized the role of ferroptosis in regulating cardiovascular disease, and we refer to the reader to this review.

20Previous studies using animal models of brain injury showed features resembling ferroptosis, including increased lipid peroxidation, increased intracellular iron levels, and decreased GSH levels., A subsequent study also showed that inhibiting ferroptosis prevented the death of primary oligodendrocytes. Notably, organotypic hippocampal slice cultures with ferroptosis-specific inhibitors were shown to prevent neuronal death and decrease hemoglobin-induced iron accumulation, suggesting a pathogenic role of ferroptosis in intracerebral hemorrhage. In addition, neurons obtained from an ischemic stroke mouse model were shown to have significantly decreased levels of GSH and increased lipid peroxidation, suggesting neuronal ferroptosis. In addition, carvacrol (the major monoterpenic phenol found in some essential oils and known to reduce oxidative stress and apoptosis) has been shown to help protect hippocampal neurons from ferroptosis by upregulating Gpx4 expression in a gerbil model of cerebral ischemia. In spinal cord injury, recent studies suggest that ferroptosis contributes to secondary injury, and blocking ferroptosis may help repair traumatic spinal cord injury.

21The liver serves as a central metabolic organ and is highly susceptible to various toxic metabolites. We previously showed that high dietary iron directly leads to ferroptosis-induced liver injury, and this damage was aggravated by either knocking out Slc7a11 expression or increasing non-TF-bound iron (NTBI). Ferroptosis was recently linked to various chemical-induced forms of liver injury such as alcoholic liver disease, methionine/choline-deficient‒induced non-alcoholic steatohepatitis (NASH), and arsenic-induced NASH. In addition, both ferroptosis and a concomitant accumulation of lipid ROS were also observed in CCl4-induced liver fibrosis, acetaminophen-induced liver damage, and I/R-induced liver injury.

22Critically ill patients with multiorgan dysfunction syndrome (MODS) were shown to present with a ferroptosis signature that includes increased plasma levels of MDA (malondialdehyde, a lipid peroxidation degradation product) and catalytic iron. In addition, a ferroptosis signature was shown in tissue samples obtained from a patient with COVID-19‒induced MODS (including cardiac and renal failure). Using an experimental model of MODS, Van Coillie et al. showed that excess iron can trigger ferroptosis in multiple organs; moreover, they showed that the novel ferroptosis inhibitor UAMC-3203 may be a viable therapeutic agent for the clinical treatment of MODS.

23Ferroptosis has also been reported to directly modulate the host’s immune response and inflammation during inflammatory and infectious diseases. Moreover, the functions of various immune cell types such as T cells, B cells, and macrophages can be affected by ferroptosis. For example, Gpx4-deficient CD4+ and CD8+ T cells undergo ferroptosis due to accumulated lipid peroxides. Thus, these ferroptotic T cells are unable to proliferate and have reduced function against the acute lymphoblastic choriomeningitis virus and Leishmania major parasites, suggesting that Gpx4 is required for maintaining T cell‒mediated immunity by protecting the cells against ferroptosis. With respect to B cells, various B cell populations have been shown to respond differently to ferroptosis induction. Similarly, in bone marrow cells inducible nitric oxide synthase (iNOS) is more abundant in M1 macrophages than in M2 macrophages, and M1 macrophages tend to be more resistant to iron deposits and ferroptosis. Interestingly, a recent study found that Gpx4-dependent ferroptosis in neutrophils drives the pathogenesis of systemic lupus erythematosus (SLE), suggesting that targeting ferroptosis-induced neutropenia may be a potential strategy for treating autoimmune diseases such as SLE.

24The aerobic gram-negative bacterium Pseudomonas aeruginosa can infect both immunocompetent and immunocompromised hosts. Interestingly, this pathogen expresses lipoxygenase (which oxidizes AA-PE to 15-hydroperoxide AA-PE in host cells) and drives ferroptosis in human bronchial epithelial cells, as well as in clinically isolated cells from patients with persistent lower respiratory tract infection. Thus, targeting ferroptosis may be a viable strategy for managing P. aeruginosa infection.

25The bacterium Mycobacterium tuberculosis is one of the major pathogens that causes tuberculosis, an infectious disease that affects the lungs, bone tissue, brain, and spine. Interestingly, M. tuberculosis has been shown to trigger ferroptosis in macrophages, accompanied by decreased GPX4 expression, increased iron content, and elevated levels of membrane lipid peroxidation. Similar results were also obtained in mice acutely infected with M. tuberculosis. Given these findings, ferroptosis may also be a promising therapeutic target in patients with tuberculosis and other infectious diseases.

26Notably, ferroptosis has also been closely linked to sterile inflammation, a process that occurs in the absence of pathogens and is commonly triggered by release of the intracellular contents from damaged and/or necrotic cells. Sterile inflammation is often triggered by acute conditions such as I/R-induced injury, trauma, oxalate crystal‒induced inflammation, and chronic inflammatory diseases such as atherosclerosis. For instance, ferroptosis increases the recruitment of neutrophils to damaged heart tissue under ischemic conditions, and ferroptosis-specific inhibitors can alleviate this inflammatory damage. In addition, blocking ferroptosis was shown to significantly reduce atherosclerosis in ApoE−/− mice fed a high-fat diet.

27Given that iron metabolism, oxidative-reductive pathways, and lipid metabolism coordinately control ferroptosis, altering these three pathways using genetic and pharmacological approaches has been shown to affect ferroptosis (Table 3). Below, we summarize the bioactive pharmacological agents that affect ferroptosis by targeting core components in these three ferroptosis-regulating processes (see also Table 4).

28Iron is essential to many physiological processes, including electron delivery, oxygen transport, and DNA biosynthesis. Iron is stored and transported as Fe3+, while Fe2+ serves as an electron donor to catalyze the Fenton reaction. During iron overload conditions, excess iron is involved in the generation of free radicals, lipid peroxidation, and DNA damage, which in turn promotes ferroptosis. At the cellular level, transferrin (TF) mediates the cellular uptake of Fe3+ via endocytosis of iron-loaded TF-bound transferrin receptor 1 (TFR1), followed by release from the endosome,, in which Fe3+ is then converted to Fe2+ via the metalloreductase STEAP3. When the binding capacity of TF-2Fe3+ is saturated, ferrireductases reduce Fe3+ to Fe2+, which can be transported to the LIP via NTBI transporters such as DMT1 (divalent metal transporter 1) and SLC39A14. The LIP can also increase due to the degradation of either heme or hemoglobin, which are captured by hemopexin and haptoglobin, respectively. Upon binding to CD163 (a monocyte/macrophage-specific scavenger receptor) and LRP1 (LDL receptor related protein 1), heme-hemopexin and hemoglobin-haptoglobin complexes are internalized via endosomes, after which the rate-limiting enzyme heme oxygenase-1 (HO-1) catabolizes heme to produce biliverdin, releasing iron into the LIP. Iron is reserved primarily in ferritin, a complex consisting of 24 light chain (FTL) and heavy chain (FTH1) subunits. FTH also has ferroxidase activity, converting Fe2+ to Fe3+ to prevent iron-induced toxicity. Excess iron is exported by FPN, the body’s sole iron exporter, which is regulated primarily by the hepatic peptide hepcidin, the master regulator of iron homeostasis. Thus, together with our recent finding that the newly identified E3 ubiquitin ligase RNF217 mediates the degradation of FPN, these results indicate that systemic iron homeostasis is tightly controlled by the hepcidin-FPN axis, and targeting dysregulated iron metabolism may directly affect ferroptosis (Fig. 3).

29The term “ferroptosis” originated from the rescue effect of the iron chelator DFO. In the clinic, DFO, deferiprone (DFP), and deferasirox (DFX) are commonly used iron chelators. DFO is derived from Streptomyces pilus and is approved for managing iron overload‒related diseases such as hemochromatosis, thalassemia, and sickle cell anemia, as well as other chronic iron overload‒induced conditions such as blood transfusion.– Although DFO has been reported to prevent ferroptosis in several animal disease models, including neurodegeneration, I/R-induced injury, NASH, and hemorrhagic stroke,,,,, DFO can cause toxic side effects such as anemia and edema. DFP, an effective, orally administered hydrophilic iron chelator, was developed as an alternative to DFO. Compared to DFO, DFP is relatively inexpensive and more effective, with lower toxicity, making DFP more suitable for use in iron-overloaded transfusion patients who currently cannot receive any form of chelation therapy. In addition, DFP can cross the blood-brain barrier (BBB) and can transfer iron in DFP-Fe to transferrin., DFP has been shown to improve motor function by decreasing iron content in the brain in patients with Friedreich’s ataxia.

30Similar to DFP, DFX is a highly selective iron chelator showing a high affinity for Fe3+ and a relatively long biological half-life. Studies have shown that DFX treatment can prevent the accumulation of hemosiderin (a complex composed of partially digested ferritin and lysosomes for iron storage) and can ameliorate ferroptosis-related kidney and neuronal damage.,

31Currently, dexrazoxane (DXZ) is the only iron chelator approved by the FDA for protecting against doxorubicin (DOX)-induced cardiotoxicity by chelating DOX-induced mitochondrial iron. Compared to DXZ, the novel orally administered iron chelator CN128 is more potent and has fewer side effects and is currently being studied in a phase II clinical trial for treating β-thalassemia following regular blood transfusion. Finally, ciclopirox olamine, which is currently approved for treating cutaneous fungal infections, has also been used as an iron chelator in various model systems due to its inhibitory activity on iron-dependent ribonucleotide reductase.,

32The TF receptor TFR1 is a glycoprotein that interacts with iron-bound TF in order to mediate cellular iron uptake. Gao et al. reported that TFR1-mediated cellular uptake of TF-bound iron is required for ferroptosis, showing that inhibiting TFR1 using RNA interference (RNAi) can efficiently block ferroptosis. Moreover, Wu et al. showed that knocking down TFR1 inhibits ferroptosis under cystine starvation conditions, whereas upregulating TFR1 significantly activates ferroptosis via the NF2-YAP signaling pathway. Using the TFR1-specific antibody 3F3-FMA, Feng and colleagues showed that TFR1 can serve as a marker for cells undergoing ferroptosis. Given that various cancer cells express relatively high levels of TFR1, antibodies that either neutralize of block TFR1 have been investigated as potential cancer therapies. For example, Horonchik and Wessling-Resnick found that the small-molecule TFR1 inhibitor ferristatin (also known as NSC306711) can inhibit iron uptake by mediating the degradation of TFR1(ref. ). Based on these results, it is reasonable to speculate that TFR1 antibodies may be suitable for treating ferroptosis-induced diseases, warranting further study.

33DMT1 (also known as SLC11A2, DCT1, and NRAMP2) is known for its ability to transport Fe2+ into the duodenum and out of the endosome during the TF cycle, but it can also transport NTBI. DMT1 inhibitors such as ebselen, pyrrolidine dithiocarbamate (PDTC), and benzylisothiourea have been shown to reduce iron-induced damage by potently reducing the DMT1-mediated cellular uptake of NTBI, indicating that DMT1 may be a promising target for regulating ferroptosis to manage ferroptosis-related diseases.

34Two strategies have been suggested to modulate ferroptosis by targeting iron export, namely hepcidin agonists such as minihepcidins and FPN inhibitors such as VIT-2763 (ref. ). Similarly, small molecules that directly upregulate hepcidin expression may affect ferroptosis by promoting intracellular iron accumulation. In addition, endogenous inducers of hepcidin such as the cytokine oncostatin M (encoded by the OSM gene) may promote ferroptosis. Conversely, clinical trials involving hepcidin antagonists such as PRS-080, NOX-H94, and LY2787106 may suppress ferroptosis by reducing intracellular iron content. Thus, further studies are needed in order to ascertain the functional role of these FPN/hepcidin regulators in modulating ferroptosis under pathological conditions.

35Under iron-deficient conditions, NCOA4 binds to iron-loaded ferritin, thus promoting lysosomal ferritin degradation and releasing iron into the LIP, a process known as ferritinophagy. Ferritinophagy has been shown to initiate ferroptosis via iron overload and lipid peroxidation., JQ1, a thienotriazolodiazepine that inhibits bromodomain proteins such as BRD4, has been reported to potentiate ferroptosis via ferritinophagy in breast cancer cells. Additionally, DpdtC (2,2′-di-pyridylketone dithiocarbamate) mobilizes iron by inducing ferritinophagy and can increase ferritinophagy-induced ROS production in MGC-803 cells, a human gastric carcinoma cell line. MMRi62, a small molecule compound initially identified as an inducer of apoptosis, is shown to potently trigger ferroptotic cell death in pancreatic ductal adenocarcinoma cells by increasing lysosomal ferritinophagy. On the other hand, Fang et al. recently found that a novel compound called 9a potently suppresses ferritinophagy-induced ferroptosis by competitively binding to NCOA4 and perturbing the interaction between NCOA4 and FTH1.

36Macrophages in the spleen and Kupffer cells (also known as stellate macrophages) in the liver maintain iron homeostasis by recycling iron obtained from senescent erythrocytes and damaged cells. Macrophage-mediated iron recycling is processed primarily in the spleen, and recycling iron is returned to the storage pool for utilization in various processes such as heme biosynthesis and Fe/S cluster formations in the mitochondria. The enzyme HO-1, encoded by the HMOX1 gene, catalyzes heme to produce carbon monoxide, biliverdin, and free iron; biliverdin and free iron can be used to generate bilirubin and sequestered by ferritin, respectively. Upregulation of HMOX1 expression has been shown to play a cytoprotective role and to increase the resistance of HCC cells to ferroptosis mediated by the p62-KEAP1 (Kelch-like ECH-associated protein 1)-NRF2 (nuclear factor erythroid-2-related factor 2) pathway. On the other hand, high HMOX1 expression can also be toxic due to high levels of ferrous iron, which in turn accelerates the Fenton reaction, particularly in the context of insufficient levels of free radical scavengers. Together, these reports support the notion that HMOX1 expression has a dose-dependent differential role. Targeting HO-1 has been proposed as a viable strategy for treating many diseases and conditions, including cardiovascular disease and inflammation. Notably, we previously showed that inhibiting HO-1 prevents ferroptosis-induced cardiomyopathy in mice. To date, many HO-1 agonists and antagonists have been proposed for use in various disease models. For example, Vreman et al. found that metalloporphyrins—synthetic heme analogs used to treat jaundice in newborn infants—can inhibit HO-1. However, several metalloporphyrins can induce phototoxicity and/or off-target adverse effects; given this poor safety profile, azalanstat was subsequently developed as a safer HO-1 inhibitor; however, studies are needed to test its efficacy in ferroptosis-related diseases.

37Under physiological conditions, ROS production is controlled by a coordinated network of antioxidative pathways. Among the various cellular antioxidative defense systems, the system Xc–-GSH-GPX4 metabolic pathway plays a key role in regulating ferroptosis. Intracellular cysteine is taken up primarily in the form of cystine via system Xc– or converted from methionine via the transsulfuration pathway, or is transported directly by alanine/serine/cysteine transporters (known as system ASC)., In addition, the KEAP1/NRF2 antioxidative signaling pathway, the glutaminolysis pathway, the FSP1-CoQ10-NAD(P)H pathway, and the recently identified DHODH-mediated pathway have all been recognized as central mediators of ferroptosis by directly regulating reductive-oxidative pathways. Below, we summarize these ferroptosis modulators that target the reductive-oxidative pathways (see Fig. 4).

38Extracellular cystine imported by system Xc– is subsequently reduced to cysteine, the rate-limiting precursor of GSH,, an essential cofactor for the antioxidant enzyme GPX4, which quenches PL hydroperoxides. GSH is produced from cysteine, glutamate, and glycine exclusively in the cytosol via glutamate-cysteine ligase and glutathione synthetase. Newly synthesized cytosolic GSH is then actively transported into the mitochondria via the recently identified GSH importer SLC25A39. Moreover, studies have shown that inhibiting either system Xc– or GPX4 potently induces ferroptosis via cysteine deprivation, driving GSH depletion and ultimately impairing the cell’s antioxidant defense mechanism.

39Erastin is a first-generation ferroptosis inducer that directly suppresses system Xc–1. Although the pro-ferroptotic effects of erastin have been reported in many types of malignant cancer cells in vitro, its further development has been limited due to its low efficacy in vivo. To optimize the chemical properties of erastin, Yang et al. developed a more effective erastin derivative, piperazine erastin, by inserting a piperazine group in the aniline ring of erastin, yielding improved drug-like properties such as increased water solubility and metabolic stability. Piperazine erastin was subsequently shown to reduce tumor growth in an HT-1080 xenograft mouse model, with no apparent toxicity. However, piperazine erastin showed limited activity against established tumors, perhaps due to its moderate potency. To overcome this limitation, imidazole ketone erastin (IKE) was developed and shown to have high (nanomolar) potency and high metabolic stability. The tumor-suppressing efficacy of IKE was then demonstrated in SUDHL6 (diffuse large B cell lymphoma) cell xenografts in mice. Capitalizing on the fact that IKE is soluble under acidic aqueous conditions, nanoparticles were used as an IKE delivery system to further improve its therapeutic index, with reduced toxicity compared to free IKE. Previous studies found that erastin was able to increase the sensitivity of various cancer cell lines to chemotherapy drugs, including doxorubicin, actinomycin D, cisplatin, temozolomide, and cytarabine,– providing new motivation for exploring the feasibility of combination therapies. Recently, we solved the high-resolution structure of erastin-bound human system Xc– and showed that IKE is a highly potent inducer of ferroptosis, providing a structural basis for designing more effective ferroptosis modulators.

40The anti-rheumatic drug sulfasalazine (SAS) is clinically approved for treating inflammatory arthritis and inflammatory bowel disease. SAS has been reported to suppress the growth of lymphoma cells both in vitro and in vivo through the activation of ferroptosis by suppressing system Xc–(ref. ). Moreover, a lot of studies have shown that SAS has anti-tumor activity in various xenograft tumor models, including glioblastoma, prostate cancer, small cell lung cancer, pancreatic cancer, and TNBC. Based on its excellent safety profile in animal studies, several phase I and phase II clinical studies have been initiated. However, various doses of SAS failed to produce a clinical response in malignant glioma, and side effects were reported, including anorexia, gastrointestinal toxicity, and hematological toxicity. In addition, long-term high-dose (8–12 g/day) treatment with SAS can induce several adverse side effects and should be avoided; therefore, combination therapy using SAS together with conventional radiotherapy and/or chemotherapy has been proposed as a promising therapeutic strategy.– To mitigate its toxicity, studies involving SAS focus primarily on sensitizing agents and applications using nanoparticles, a novel strategy designed to increase the effectiveness of low-dose SAS.

41Sorafenib (SRF), a small molecular kinase inhibitor, is approved for treating various solid tumors, has well-characterized clinical efficacy and tolerability, and remains the only drug approved to treat advanced HCC. SRF has been shown to trigger ferroptosis in cancer cells by suppressing system Xc–., However, the relatively limited clinical benefits of SRF and the emergence of drug resistance are major hurdles hampering its further development. On the other hand, several studies found that combination therapies can increase SRF’s anti-tumor activity,, as well as GSH starvation‒based nanoscience for cancer therapy., Notably, however, Zheng et al. recently reported that SRF failed to increase ferroptosis in various cancer cell lines. Thus, the precise role of SRF in ferroptosis, and its clinical value in the context of cancer, remains controversial and poorly understood.

42Interestingly, a subset of FDA-approved drugs such as the muscle relaxant lanperisone and the analgesic acetaminophen (N-acetyl-p-aminophenol, or APAP) have been shown to activate ferroptosis by functionally suppressing system Xc–. For example, lanperisone has been shown to effectively target RAS-mutated cancers in vivo without causing overt toxicity. APAP is widely used as an analgesic and antipyretic, but can cause dose-dependent hepatotoxicity. Interestingly, Yamada et al. recently showed that APAP-induced hepatotoxicity is attributed to GSH depletion‒induced ferroptosis and proposed this as the predominant mechanism, providing new insights into the potential use of APAP in triggering ferroptosis in order to inhibit tumor growth. Indeed, several studies have shown a synergistic therapeutic effect of combining erastin and APAP to treat melanoma and lung cancer xenografts.,

43GSH has a central role in the protection of cells from oxidative damage and toxic reactive species. Therefore, targeting pathways involved in GSH synthesis has been studied extensively as a major strategy for treating ferroptosis-related diseases.

44Auranofin (AUR) is a gold (I)‒containing compound used to treat rheumatic arthritis. Recent studies have shown that AUR has therapeutic potential for other diseases and conditions such as cancer, metabolic disease, and infectious and inflammatory diseases. Indeed, we provided the first report that AUR can be used in vivo as to induce ferroptosis in mice by suppressing the activity of thioredoxin reductase, providing compelling evidence supporting its further clinical testing—either alone or combined with other therapies—for treating ferroptosis-resistant disease conditions such as cancer. To expand the potential clinical applications of AUR, several studies have investigated repurposing AUR; for instance, AUR has been repurposed as an anticancer drug in TNBC cells by inhibiting GSH biosynthesis. Recently, we found that malic enzyme 1 (Me1) regulates ferroptosis in hepatic I/R-induced injury, and its inactivation further decreased the production of NADPH and failed to restore GSH synthesis, suggesting that targeting Me1 may be a viable strategy for treating ferroptosis-related diseases by altering GSH synthesis.

45Buthionine sulfoximine (BSO) activates ferroptosis by inhibiting glutamate-cysteine ligase (GCL). Similar to erastin, BSO has also been used to deplete GSH and induce ferroptosis in several cancer cell lines., Based on early clinical trials, BSO is considered safe but has limited therapeutic benefit for treating refractory malignancies. To improve its clinical efficacy, researchers attempted to identify BSO-sensitive cancer types and patients, revealing that low GSH levels can serve as a marker for BSO susceptibility.

461,3-bis-(2-chloroethyl)-1-nitrosourea (BCNU, also known as carmustine) is a selective glutathione disulfide reductase (GSR) inhibitor and is used as a chemotherapy drug for treating brain cancer and lymphoma. Based in its mode of action, BCNU is believed to induce ferroptosis by directly inhibiting GSH synthesis. Recently, a combination of BCNU and sorafenib was shown to significantly suppress the in vivo growth of liver cancer by promoting ferroptosis. In addition, both BCNU and AUR have been shown to cause cell death in oxidant-enriched tumorigenic endothelial (EOMA) cells, possibly by activating ferroptosis. These findings may therefore increase the clinical application of BCNU for the treatment of HCC and endothelial cell tumors.

47N-acetylcysteine (NAC) is clinically approved to treat APAP overdose. As an antioxidant, NAC has been shown to inhibit ferroptosis by targeting cysteine metabolism. In addition to protecting the liver from APAP-induced ferroptosis, NAC has also been clinically shown to improve neurodegeneration-related symptoms by increasing cysteine levels and facilitating the synthesis of γ-glutamyl-cysteine and GSH. Due to its poor bioavailability, however, NAC must be administered in fairly high doses and requires a long treatment time in patients with severe APAP overdose, thereby increasing the risk of an anaphylactoid reaction (i.e., nonimmunologic anaphylaxis), fluid overload, and high fluid osmolarity. To overcome these issues, N-acetylcysteine amide (NACA), a modified form of NAC with increased membrane permeability and bioavailability (67% compared to only 15% for NAC), has been developed. To date, NACA has been shown to have antioxidant activity in several preclinical models, but is yet to be approved for clinical use.

48The essential antioxidant enzyme GPX4 is not only responsible for maintaining redox homeostasis, but is also recognized as the “ferroptosis gatekeeper” by transforming lipid ROS into lipid alcohols. Direct inactivation of GPX4 has been shown to drive ferroptosis independent of intracellular cysteine and GSH levels. Thus, numerous studies have focused on identifying novel activators of ferroptosis by targeting GPX4 in order to develop new cancer therapeutics.

49As a covalent inhibitor of GPX4, RSL3 was originally identified as a pro-ferroptosis compound through chemical screening. Although sharing the more common features of ferroptosis, RSL3-mediated ferroptosis does not include decreased levels of cellular GSH. GPX4 was subsequently identified as the primary target of RSL3 using an unbiased, affinity-based chemoproteomics approach. Moreover, genetically knocking down and overexpressing GPX4 were shown to cause sensitization and resistance to RSL3, respectively, providing further evidence that GPX4 is the target of RSL3. Since these previous studies, RSL3 has been widely used as a ferroptosis agonist in vitro in a wide range of cancer cell types, including pancreatic cancer, adrenocortical carcinoma, fibrosarcoma, and breast cancer cells. In addition, RSL3 may also serve as a chemosensitizer for cisplatin, doxorubicin, and actinomycin D by activating ferroptosis in various cancer cells, including neuroblastoma, osteosarcoma, lung cancer, and rhabdomyosarcoma cell lines. In mouse xenograft models, RSL3 is well-tolerated, with no overt toxicity or body weight loss observed even at the relatively high dose of 400 mg/kg.,, As an experimental tool, RSL3 has been used extensively to study the role of ferroptosis in various disease models, including AKI, I/R-induced injury, and neurodegenerative disease.–

50Similar to RSL3, ML162 was identified as a GPX4 inhibitor, and subsequent studies showed that ML162 can directly bind GPX4 and inhibit its activity. However, both RSL3 and ML162 have relatively poor selectivity and pharmacokinetics, as they covalently bind to GPX4 via a reactive alkyl chloride moiety. To develop a new GPX4 inhibitor containing a different moiety, Eaton et al. used masked nitrile-oxide electrophiles to create ML210, a more selective covalent suppressor of GPX4 with improved pharmacokinetics and comparable activity. However, given that the in vivo efficacy of all GPX4 inhibitors developed to date remains limited, efforts have focused on developing more efficacious GPX4-selective inhibitors.

51Based on the structural optimization of CIL56 (caspase-independent lethal 56), FIN56 was developed as a highly selective inducer of ferroptosis that promotes GPX4 degradation as well as the GPX4-independent activation of squalene synthase (SQS). FIN56 potently induces ferroptosis in all TNBC cell lines tested to date, suggesting that FIN56 may be an effective additional pro-ferroptosis anticancer drug. In contrast, the endoperoxide-containing 1,2-dioxolane FINO2 was shown to potently and selectively induce ferroptosis in engineered cancer cells by directly oxidizing iron and indirectly inactivating GPX4. In addition, the naturally occurring C28 steroidal lactone withaferin A (WA), one of the best-studied withanolides, has a diverse range of pharmacological activities, including anti-inflammatory, anti-tumor, and antioxidant properties. In both high-risk neuroblastoma cell lines and neuroblastoma xenografts, WA has been shown to dose-dependently induce ferroptosis via the dual mechanisms of inactivating GPX4 and activating the NRF2 pathway. Thus, natural products may serve as an additional resource for the discovery of new GPX4 inhibitors.

52Recently, Wu et al. identified creatine kinase B (CKB)-enhanced GPX4 as a novel mechanism for regulating ferroptosis, accounting for the resistance of HCC cells to ferroptosis by stabilized GPX4 via CKB phosphorylated GPX4 at residue S104. Mutations in either CKB or GPX4 significantly increase the ferroptosis-mediated cell death of HCC by inhibiting this phosphorylation. Importantly, the authors showed a positive correlation between phosphorylated GPX4 and HCC aggressiveness in clinical samples, suggesting a potential new strategy for treating HCC by blocking CKB/GPX4-mediated ferroptosis.

53Although GPX4 inhibitors potently inhibit cells growth in vitro, all currently identified GPX4 inhibitors have limited prospects for further clinical development due to their poor pharmacokinetics and specificity. Therefore, more studies are needed in order to develop GPX4-specific inhibitors with improved pharmacological properties.

54The essential micronutrient selenium (Se) is required for the production of selenocysteine, which serves as the active site for GPX4. Ionic selenite (SeO32-) is commonly used to deliver Se to cultured cells. For example, treating cultured neurons with SeO32- was shown to increase GPX4 transcription and inhibit ferroptosis induced by either hemin or homocysteic acid. In addition, systemically treating mice with Tat SelPep, a selenocysteine-containing peptide that can cross the BBB, was shown to improve functional recovery following hemorrhagic and ischemic stroke by blocking ferroptosis. Together, these studies suggest that Se supplementation may help protect against ferroptosis-related tissue damage and disease.

55The neurotransmitter dopamine has many physiological roles, particularly in controlling various functions in the central nervous system, including movement, memory, motivation, mood, and attention. Insufficient levels of dopamine production are known to contribute to the progression of PD, and dopamine-based therapies such as levodopa and dopamine receptor agonists have been used clinically to treat PD and various cardiovascular conditions. Interestingly, non-oxidative dopamine was reported to potently inhibit erastin-induced ferroptosis in both cancerous and non-cancerous cells by stabilizing GPX4, suggesting that dopamine may be a promising candidate drug for alleviating ferroptosis-related tissue injury and disease, as well as certain neurodegenerative diseases.

56Using a combination of computational prediction and experimental validation, Li et al. discovered eight new potential GPX4 activators and found that compound 1d4 was the most potent allosteric activator of GPX4, significantly increasing GPX4 activity by 50% when applied at 20 μM in a cell-free assay and 61 μM in cell extracts; moreover, they found that compound 1d4 potently inhibited ferroptosis when applied to HT-1080 fibrosarcoma cells.

57The diterpenoid triepoxide curculigoside is one of the primary bioactive phenolic compounds isolated from Curculigo orchioides Gaertn (commonly known as “Kali Musli”), an ancient medicinal plant well known for its immunomodulatory and rejuvenating effects. Recently, Wang et al. reported that curculigoside inhibits ferroptosis in IEC-6 cells (a cell line used to model intestinal iron transport) by upregulating GPX4 expression. In mice with dextran sulfate sodium (DSS)‒induced ulcerative colitis, curculigoside was shown to protect against disease progression by reducing ferroptosis via GPX4 induction, suggesting that activating GPX4 is a viable therapeutic option for ulcerative colitis. Another plant-derived bioactive compound, puerarin, is an isoflavone glycoside isolated from Pueraria lobata (also known by its Chinese name, Gegen) with various pharmacological effects, including antioxidant, anticancer, and anti-inflammation properties, in addition to alleviating pain, promoting bone formation, attenuating insulin resistance, and exerting cardiac and neuronal protective effects. However, its molecular target and mechanism remain unknown. Interestingly, the anti-ferroptosis activity of puerarin in heart disease has been linked to the induction of FTH1 and GPX4 in rats. Given that the majority of bioactive natural compounds have relatively low bioavailability, future studies should focus on developing rationally designed combination therapies and/or their incorporation in nanoparticles.

58The p62-KEAP1-NRF2 signaling pathway plays a regulatory role in response to oxidative stress, environmental insult, and toxic chemicals. In response to oxidative stress, p62 activates NRF2 by directly binding to its ubiquitin ligase adapter KEAP1; as a result, NRF2 translocates to the nucleus and regulates cellular redox homeostasis by modulating the expression of target genes, including several genes encoding enzymes involved in both GSH synthesis and iron homeostasis. Thus, NRF2 signaling is important for regulating ferroptosis.

59The polar hydrophilic alkaloid trigonelline can be extracted from many plant species, including coffee beans and fenugreek seeds, and has been shown to significantly sensitize cancer cells to ferroptosis inducers by inhibiting NRF2.,, In addition, sitagliptin, a selective inhibitor of dipeptidyl peptidase 4 (DPP4) used to treat type 2 diabetes, was shown recently to inhibit ROS production, inflammation, and excessive autophagy by promoting Nrf2 translocation to the nucleus in mouse models of acute lung injury induced by severe pancreatitis. In addition, a modest dose of WA was shown to induce ferroptosis and increase the LIP by directly targeting KEAP1, mediating the NRF2-mediated upregulation of HO-1.

60Other reductive-oxidative pathways, including the transsulfuration, FSP1-CoQ10-NAD(P)H, glutaminolysis, and DHODH-mediated pathways, have also been linked to ferroptosis and are discussed in detail below.

61When cysteine availability is restricted, the transsulfuration pathway is activated in order to increase the synthesis of cysteine from methionine. The enzyme cystathionine beta-synthase (CBS) activates the first step in this pathway, followed by the formation of cystathionine into cysteine by the enzyme cystathionine gamma-lyase (CTH). Sustained activation of the reverse transsulfuration pathway has been implicated in the resistance of ovarian cancer cells to erastin-induced ferroptosis. CH004, a pharmacological inhibitor of CBS, has been identified as a novel stimulator of ferroptosis in liver cancer both in vitro and in vivo, providing experimental evidence to support the therapeutic potential of targeting the transsulfuration pathway in liver cancer. Similarly, the CTH inhibitor propargylglycine has been shown to sensitize NSC-34 cells to both erastin- and RSL3-induced ferroptosis. In mice, propargylglycine was also shown to increase susceptibility to APAP-induced liver damage and mortality, possibly by inducing ferroptosis.

62The FSP1 gene was initially identified as a pro-apoptotic gene, but has also been identified as an anti-ferroptosis gene via FSP1’s oxidoreductase activity, which reduces CoQ10 to ubiquinol (CoQ10H2)., Moreover, the FSP1 inhibitor iFSP1 was shown to selectively activate ferroptosis in GPX4-knockout FSP1-overexpressing cells and was identified by screening 10,000 drug-like compounds; iFSP1 was also shown to potently increase the sensitivity of cancer cells to RSL3-induced ferroptosis. Recently, Yoshioka et al. identified the novel compound called NPD4928 as an FSP1 inhibitor; the authors then showed that NPD4928 increased the cytotoxicity of various cell types in response to GPX4 antagonists, indicating that NPD4928 and GPX4 inhibitors may work synergistically to treat cancer by inducing ferroptosis. Most recently, Hendricks et al. developed another FSP1 inhibitor called FSEN1, which they found sensitized various cancer cells to ferroptosis. Together, these findings suggest that pharmacologically inhibiting the FSP1-CoQ10-NAD(P)H pathway may provide a plausible strategy for sensitizing cancer cells to ferroptosis-resistant therapeutic agents. On the other hand, Fang et al. recently identified the diphenylbutene derivative compound 3f as a ferroptosis inhibitor, and they showed that this compound can protect against ischemic stroke in rats by increasing FSP1 protein levels, suggesting that novel agents that upregulate FSP1 may be used to treat ferroptosis-related diseases.

63Cellular uptake of the amino acid glutamine (Gln) is mediated primarily by specific transporters such as SLC38A1 and SLC1A5. Following its cellular uptake, intracellular Gln is converted by glutaminase (GLS) to produce glutamate, which can be further catalyzed to α-ketoglutarate either by glutamate dehydrogenase (GLUD1)‒mediated deamination or by transaminase-mediated transamination. Given that this reaction promotes ferroptosis by causing an accumulation of lipid peroxidation, blocking the glutaminolysis pathway has been proposed as a potential therapeutic approach for treating ferroptosis-induced tissue damage.

64Inhibiting glutaminolysis using L-g-glutamyl-p-nitroanilide (GPNA, an inhibitor of the SLC38A1/SLC1A5 complex), compound 968 (an inhibitor of GLS), or amino-oxyacetic acid (AOA, an inhibitor of pan-transaminases) has been shown to block cystine deprivation‒induced ferroptosis in mouse embryonic fibroblasts, melanoma cells, as well as head and neck cancer cells.,, In addition, compound 968 was shown to potently prevent I/R-induced heart damage in an ex vivo model.

65Recent studies showed that low-concentration paclitaxel (PTX, a mitotic chemotherapeutic drug) suppressed cancer cell proliferation by promoting the production of lactate and changing the pH of the tumor microenvironment by downregulating glutaminolysis-related genes such as GLS, SLC7A11, and SLC1A5., In addition, low-dose PTX was shown to trigger ferroptosis by decreasing SLC7A11 expression by upregulating the well-characterized transcription factor p53(ref. ). Moreover, a growing body of evidence suggests that p53 regulates ferroptosis either by transcriptional or posttranslational mechanisms such as modulating the downstream expression of SLC7A11, SAT1 (encoding spermidine/spermine N1-acetyltransferase 1), GLS2 (encoding glutaminase 2), and CDKN1A/p21 (encoding cyclin dependent kinase inhibitor 1A), or by directly inhibiting DPP4 activity in colorectal cancer cells.

66The flavin-dependent mitochondrial enzyme DHODH catalyzes the de novo synthesis of pyrimidine. Originally, targeting DHODH was shown clinically to improve autoimmune diseases such as multiple sclerosis and rheumatoid arthritis. Although functional studies characterized DHODH as a potential target for treating cancer, several potent DHODH suppressors such as brequinar, leflunomide, and teriflunomide were clinically assessed but failed to receive FDA approval. Importantly, however, brequinar was recently shown to act synergistically with the FDA-approved drug SAS to suppress tumor growth by potently inducing ferroptosis, suggesting that DHODH may indeed serve as a viable target for treating cancer. Interestingly, DHODH inhibitors have also been developed as antiviral agents to act against cytomegalovirus, Ebola, influenza, and SARS-CoV-2(refs.,). Whether ferroptosis is involved in the multifaceted activities of DHODH inhibitors remains an open question; nevertheless, targeting DHODH may have promise as a novel therapeutic approach for cancer and certain viral infections.

67Lipid metabolism has an important regulatory role in the balance between cell survival and cell death. In ferroptosis, the majority of oxygenated PL species are upregulated, resulting in damage to PL-containing cell membranes. Four types of PLs—namely, the double and triple oxygenated arachidonic acid (AA)- and adrenic acid (AdA)-containing phosphatidylethanolamine (PE) species (C18:0/C20:4 and C18:0/C22:4)—were previously identified as the most susceptible substrates for lipid peroxidation. In addition to PE, other PLs can also be oxidized upon the induction of ferroptosis. Because lipid metabolic pathways are important for regulating lipid peroxidation, targeting these pathways may provide a novel strategy for treating ferroptosis-related diseases (Fig. 5).

68A haploid genetic screen in KBM7 cells (a chronic myeloid leukemia cell line) revealed that the enzymes ACSL4 and LPCAT3 serve as ferroptosis regulators. ACSL4 and LPCAT3 catalyze the addition of CoA to the long-chain polyunsaturated bonds of AA and AdA, respectively, thereby promoting the esterification of PUFAs to membrane PLs., PEs containing acylated AA or AdA (AA/AdA-CoA) can be subsequently oxidized by LOXs to form PL hydroperoxides (PE-AA/AdA-OOH), eventually leading to the membrane deposition of lipid ROS and driving ferroptosis. A number of inhibitors that target lipid metabolism have been shown to suppress ferroptosis, including LOX inhibitors, radical-trapping antioxidants (RTAs, which scavenge lipid hydroperoxyl radicals), ACSL4 inhibitors, and deuterated PUFAs/MUFAs (by decreasing PUFA-containing PLs).

69LOXs are a class of non-heme iron-containing dioxygenases that can regulate ferroptosis by mediating PL oxidation. A total of 6 LOX isoforms (ALOX5, ALOX12, ALOX12B, ALOX15, ALOX15B, and ALOXE3) are expressed in humans, and a total of 7 LOX isoforms (Alox5, Alox12, Alox12b, Alox15, Alox15b, Aloxe3, and Alox12e) are expressed in mice., ALOX5/12 inhibitors, including AA-861, cinnamyl-3,4-dihydroxy-alpha-cyanocinnamate (CDC),, the pan-LOX inhibitor nordihydroguaiaretic acid (NDGA), and BWA4C, have been either shown or suggested to suppress ferroptosis. In an effort to re-evaluate the selectivity profile of CDC for LOXs, Pergola et al. identified CDC as an ALOX5-specific antagonist with an IC50 value in the low nanomolar range (9–25 nM) measured in cell-free assays. Zileuton, another widely used ALOX5-selective inhibitor, was the first FDA-approved orally administered drug for treating asthma. Zileuton was also shown to exert a neuroprotective effect by blocking ALOX5-mediated glutamate toxicity and ferroptosis in HT22 cells (a mouse hippocampal cell line). With respect to ALOX12, ML355 is a potent and selective inhibitor that has been shown to reduce ALOX12-induced thrombosis and protect human pancreatic islets from ALOX12-triggered inflammatory injury., Recently, Zhang et al. reported that the compound IMA-1, which disrupts the interaction between ALOX12 and ACC1 (acetyl-CoA carboxylase 1), can reduce NASH progression in mice and Cynomolgus macaques, suggesting that the ALOX12-ACC1 complex may be a clinically viable target for the management of NASH.

70ALOX15 was recently reported to play an essential role in exacerbating I/R-induced cerebral injury and ischemia-induced myocardial injury by driving the peroxidation of PL-PUFAs and inducing ferroptosis, providing compelling evidence supporting the viability of targeting ALOX15 to treat I/R-induced injury in these tissues. Interestingly, Walters et al. functionally confirmed that ALOX15 is involved in inducing oxidative stress in human spermatozoa, supporting the notion of increased ALOX15 expression in sperm cells and suggesting that ALOX15 may be a potential target for the treatment of oxidative stress‒induced male infertility. Indeed, the ALOX15-selective inhibitor PD146176 was shown to protect against oxidative damage by decreasing the production of membrane PL peroxidation and 4-hydroxynonenal (4-HNE) in spermatozoa. Moreover, the ALOX15-specific inhibitor ML351 was recently shown to reduce ferroptosis in cardiac I/R-induced injury.

71Taken together, various LOX inhibitors have been shown to effectively inhibit ferroptosis. In addition, the enzyme POR (cytochrome P450 oxidoreductase) was recently shown to promote PL peroxidation in a LOX-independent manner, suggesting that the regulation of enzymes other than LOXs may also contribute to lipid peroxidation by mediating the oxygenation of membrane PLs.

72One of the principal features of ferroptosis, namely PL peroxidation initiated by the formation and propagation of lipid radicals, can be blocked by RTAs, a class of lipid chain‒breaking antioxidants. Among the various RTAs identified to date, the now-widely used ferrostatin-1 (Fer-1) was originally identified as a potent ferroptosis-specific inhibitor in erastin- and RSL3-treated HT-1080 cells. By scavenging free radicals, Fer-1 reduces ferroptosis-induced tissue injury and prolongs survival in mouse models of disease.,, However, Fer-1 has not been pursued in drug development, given its metabolic instability and poor pharmacokinetics. Currently, Fer-1 is used as a research tool for studying various ferroptosis-related processes. Compared to Fer-1, second-generation (SRS11-92) and third-generation (SRS16-86) RTAs have significantly higher activity but still contain an ester moiety., To develop an optimized set of RTAs, Devisscher et al. developed a series of compounds in which the ester was replaced with a sulfonamide, providing improved water solubility and better metabolic and kinetic properties. For example, they found that compound UAMC-3203 was more stable and more potent than Fer-1 (with a significantly lower IC50 compared to Fer-1). After intravenous injection, the terminal plasma half-life of UAMC-3203 is ~3 and 5 h in mice and rats, respectively, suggesting that this compound has relatively good pharmacokinetics.

73Liproxstatin-1 (Lip-1), a spiroquinoxalinamine derivative, was identified by high-throughput screening and shown to inhibit ferroptosis in vivo with better pharmacological properties than Fer-1. However, Lip-1 also potently inhibits the enzymatic activity of CYP2D6 (with an IC50 of 4.1 μM), a member of the cytochrome P450 superfamily of drug-oxidizing enzymes, indicating that Lip-1 may not be suitable for use in clinical trials.

74In addition to inhibiting ferroptosis, potent RTAs such as Fer-1 and Lip-1 can insert into membrane PLs via arylamines. Notably, the catalytic activity of both Fer-1 and Lip-1 requires a relatively high temperature. Using three different cellular models of ferroptosis, Zilka et al. found that tetrahydronapthyridinols (THNs) inhibit ferroptosis induced by RSL3, cystine deprivation, and GPX4 inactivation. Compared to both Fer-1 and Lip-1, lipophilic THNs have similar potency, while hydrophilic THNs are ineffective at inhibiting ferroptosis.

75Diarylamines are RTAs commonly used to reduce the autoxidation of petroleum-derived products. The efficacy of this class of RTAs depends largely on the kinetics and stability of the rapid hydrogen atom transfer to one-electron oxidation by peroxidic species. Two diarylamine derivatives—phenothiazine and phenoxazine—have been shown to suppress ferroptosis in mouse embryonic fibroblasts. Interestingly, each phenoxazine molecule can trap >2 peroxyl radicals at moderate temperatures, suggesting new strategies for the development of more potent RTA-based ferroptosis inhibitors. Indeed, Mishima et al. showed that promethazine is more potent at protecting kidney function compared to Fer-1(ref. ), suggesting that promethazine may be a viable tool for studying the role of ferroptosis in various disease models. Notably, edaravone—an RTA clinically approved for treating acute ischemic stroke and amyotrophic lateral sclerosis—has been shown to protect against ferroptosis under various pathological conditions. In addition, Zilka et al. found that copper(II)-diacetyl-bis(N4-methylthiosemicarbazone) (CuATSM), a candidate drug for treating ALS and PD, suppresses ferroptosis via its RTA activity. Recently, a group of reduced forms of vitamin K (VKH2), including menaquinone and phylloquinone, was shown to have potent anti-ferroptosis properties, providing new mechanistic insights into the anti-ferroptotic activity of alternative RTAs.

76Recently, Wu et al. and Barayeu et al. independently reported that endogenous sulfane sulfur (S0) species/hydropersulfides (H2S or RSSH) are potent RTAs. By supplying sulfur for S0 biosynthesis, cysteine can inhibit ferroptosis via a GPX4-independent mechanism. Genetic manipulation of enzymes involved in S0 biosynthesis clearly implicate S0 as playing a role in regulating ferroptosis by scavenging ROS and suppressing lipid peroxides. Both endogenous and exogenous S0 have been shown to provide cellular protection against ferroptosis., Together, these results suggest that regulating ferroptosis by targeting S0 warrants further clinical study.

77ACSL4 is a unique and important isozyme involved in the metabolism of PL-PUFAs such as AA, and deleting or inhibiting ACSL4 blocks ferroptosis by preventing the incorporation of PL-PUFAs into cell membranes., Thiazolidinediones (TZDs) such as rosiglitazone, pioglitazone, and troglitazone were originally identified as agonists of peroxisome proliferator-activated receptor γ (PPARγ) and are approved to treat adult type 2 diabetes. Interestingly, TZDs showed effects on suppressing ferroptosis by selectively inhibiting ACSL4, in breast cancer cell lines and reduce mortality due to acute renal failure in kidney-specific Gpx4 knockout mice.

78Similar to ACSL4 inhibitors, deuterated PUFAs and MUFAs can also inhibit ACSL4 by decreasing PL-PUFAs. PUFAs deuterated at the bis-allylic position (D-PUFAs) have been shown to inhibit ferroptosis in cells and animal models of PD and Friedreich’s ataxia.,– Notably, exogenous lipid supplementation can modulate both apoptosis and ferroptosis. Magtanong et al. showed that exogenous MUFAs can also specifically reduce the accumulation of lipid ROS in the plasma membrane and can displace PUFAs from their cellular location, suppressing ferroptosis in an ACSL3-dependent manner.

79Molecules that affect the composition of intracellular lipids may also regulate ferroptosis. For example, sulfosuccinimidyl oleate (SSO), which irreversibly inhibits the fatty acid receptor CD36, has been shown to inhibit ferroptosis in CD8+ T cells by blocking the uptake of oxidized lipids. Moreover, inhibiting stearoyl-CoA desaturase-1 (SCD1) with CVT-11127 or inhibiting acetyl-CoA carboxylase (ACC) with CP-640186 can affect the production of MUFAs, reducing the growth of lung cancer cells. Thus, Batchuluun et al. proposed that other potent ACC inhibitors should be examined for their ability to affect ferroptosis.

80Recently, Minami et al. found that loss of CDKN2A expression remodels the lipidome of glioblastoma cells, and patient-derived CDKN2A-deficient glioblastoma cells have higher levels of lipid peroxidation, thereby sensitizing the cells to ferroptosis in response to GPX4 inhibition. These results suggest potential new therapeutic strategies in which targeting cellular lipidome remodeling can induce ferroptosis in glioblastoma cells, particularly in cells lacking CDKN2A expression.

81The mevalonate pathway regulates ferroptosis by altering GPX4 production by affecting the maturation of selenocysteine tRNA. The small molecule FIN56 potently induces ferroptosis via two mechanisms, namely by reducing GPX4 levels and by decreasing CoQ10 activity by targeting the enzyme SQS, which acts downstream of HMG-CoA (3-hydroxy-3-methylglutaryl-CoA) reductase in the mevalonate pathway.,

82Statins are well-known inhibitors of HMG-CoA reductase (HMGCR) and are commonly used as cholesterol-lowering medications to prevent and/or treat atherosclerosis. Statins have also been shown to increase FIN56-induced cellular ferroptosis and selectively kill cancer cells in a high-mesenchymal state. However, the concentration of statins needed to increase ferroptosis in cancer cells is 100–500 times the recommended dose for decreasing cholesterol, and high-dose statins often cause severe side effects such as myalgia, rhabdomyolysis, and hepatotoxicity., To mitigate these side effects, it may be possible to deliver the statins in nanoparticles, leading to a high concentration specifically in tumor cells.

83Although numerous ferroptosis-targeting agents have been studied, their properties often make them poorly suited for use as therapeutic compounds; indeed, their low solubility, high metabolic clearance, low cellular permeability, and systemic toxicity limit their clinical development. Therefore, multi-target ferroptosis modulators and nanomaterial technologies have been investigated in an attempt to target ferroptosis more effectively in vivo.

84Artemisinins, which are derived from sweet wormwood (Artemisia annua) extracts, are commonly used as effective antimalarial drugs. The semi-synthetic derivative and major active metabolite of artemisinin, dihydroartemisinin (DHA), has been shown to increase ferroptosis in lung cancer cells by inactivating the PRIM2/SLC7A11 axis, as well as in head and neck cancers and glioblastoma cell lines by inhibiting GPX4. In addition, DHA plays a role in ferritinophagy, thereby inducing free iron‒induced ferroptosis.

85Glycyrrhizin, a natural antioxidant extracted from the glycyrrhiza root, was shown to have protective effects on the liver via its antioxidative, anti-inflammatory, antifibrotic, and antiviral properties.,, Wang et al. showed that glycyrrhizin can prevent ferroptosis in cells and in an animal model of acute liver failure by regulating iron levels, the GSH/GPX4 pathway, and the NRF2/HO-1/HMGB1 pathway.

86The flavonoid baicalein (in Chinese: Huangqin) has long been used as a popular antibacterial and antiviral agent. Near the end of the last century, Afanas’ev et al. reported that baicalein exerts its protective role against oxidative damage by inhibiting the iron-catalyzed Fenton reaction. In addition, baicalein is a specific antagonist of ALOX12 and was shown to protect against I/R-induced heart injury in mice. Subsequently, Leyen et al. proposed that baicalein may also protect against I/R-induced brain injury by inhibiting ALOX12/15 (ref. ). Recently, baicalein was shown to prevent ferroptosis in erastin-treated PANC1 cells (a human pancreatic cancer cell line) by reducing ferrous iron, inhibiting GSH consumption, and inhibiting GPX4 degradation, thereby suppressing lipid peroxidation. These results suggest that baicalein may prevent ferroptosis-associated tissue damage.

87Compared to small-molecule compounds, ferroptosis-inducing nanoparticles have been suggested to have a better preclinical profile. Importantly, nanoparticles can be loaded with chemotherapeutic drugs or tumor-selective molecules, making them ideally suited for developing nanoparticle-based cancer therapies as well as combination therapies.

88Many nanoparticles have been designed to increase ferroptosis by triggering the Fenton reaction in targeted tumor cells (Fig. 6). For instance, the magnetic nanoparticle FeGd-HN@Pt@LF/RGD2, the PEGylated single-atom Fe-containing nanocatalyst PSAF, the nanoprobes UCNP@GA-FeIII and UCNP@LP(Azo-CA4), and the nano-platforms HA@MOF and PB@FePt-Ha-g-PEG can effectively deliver Fe2+ to cancer cells in order to accelerate the Fenton reaction and produce lethal amounts of ROS, specifically triggering ferroptosis. Interestingly, Zhang et al. found that the FePt@MoS2 nanocomposite efficiently induces ferroptosis by releasing >30% of loaded Fe2+ into the tumor microenvironment within 72 h. In addition, GOD-Fe3O4@DMSNs, in which the nanoparticle is loaded with a natural glucose oxidase, has been shown to exhaust glucose in tumor cells and generate excessive levels of H2O2, thereby accelerating the Fenton reaction catalyzed by Fe3O4(ref. ). In the relatively acidic tumor microenvironment, several nanoparticles can be reduced in order to release two molecules such as iron and doxorubicin., For example, FePt-NP2 can deliver iron oxide in order to enhance the chemotherapeutic effects of cisplatin specifically in cancer cells. Moreover, Shen et al. designed FeGd-HN@Pt@LF/RGD2, a cisplatin-loaded Fe3O4/Gd2O3 hybrid nanoparticle conjugated to lactoferrin (LF) and an RGD (Arg-Gly-Asp) dimer in order to potently induce ferroptosis specifically in brain tumor cells. Thanks to its small size (6.6 nm) and LF receptor‒mediated transcytosis, FeGd-HN@Pt@LF/RGD2 nanoparticles can cross the BBB and deliver cisplatin together with reactants (e.g., Fe2+, Fe3+, and H2O2) to brain tumor cells, significantly slowing the growth of orthotopic brain tumors.

89An alternative nanoparticle-based strategy for triggering ferroptosis is to deplete GSH and/or inhibit GPX4 (Fig. 6). For example, arginine-rich manganese silicate nanobubbles have been shown to activate ferroptosis by depleting GSH and subsequently inactivating GPX4. SRF@FeIIITA is a rationally designed nanoparticle spontaneously formed by a Fe3+ and tannic acid (TA) network‒like corona onto a nanocore containing the kinase inhibitor sorafenib (SRF). In the acidic lysosomal environment, SRF@FeIIITA releases SRF, thereby inhibiting the production of GSH and inducing ferroptosis. Another SRF-based nanoparticle, FaPEG-MnMSN@SFB, has been validated to effectively activate ferroptosis and suppress tumor growth via the dual GSH-depleting effects of MnMSN (manganese-loaded mesoporous silica nanozyme) and SRF. Moreover, SRF@MPDA-SPIO nanoparticles were generated using SRF and ultrasmall superparamagnetic iron oxide (SPIO) nanoparticles, making the particles visible on MRI and potently inducing ferroptosis by supplying iron. In addition, the biomimetic magnetic nanoparticle Fe3O4-SAS@PLT was generated using SAS-loaded mesoporous magnetic nanoparticles (Fe3O4) and a platelet (PLT) membrane camouflage, thereby activating ferroptosis by blocking the system Xc− pathway. Notably, Fe3O4-SAS@PLT‒induced ferroptosis significantly increased PD-1 (programmed cell death 1)‒based cancer immunotherapy and long-term tumor elimination in mice transplanted with 4T1 metastatic breast cancer cells. Recently, Li et al. reported the development of the multifunctional nanosheet system Au/Cu-TCPP(Fe)@RSL3-PEG-iRGD, which triggers ferroptosis in tumor cells by simultaneously blocking the GSH/GPX4 and FSP1/CoQ10H2 pathways via its nanocatalytic activity, as well as synergizing with RSL3 to further inactivate GPX4 in tumor cells.

90Interestingly, nanoparticles can also be designed to directly deliver molecules that trigger ferroptosis (Fig. 6). For example, αMSH-PEG-C’ are ultrasmall (<10 nm in diameter) PEG-coated silica nanoparticles conjugated to MC1-R (melanocortin-1 receptor) in order to target alpha melanocyte‒stimulating hormone (αMSH) peptides; this nanoparticle was then shown to activate ferroptosis in a variety of starved cancer cells, as well as in mice xenografted with 786-O and HT-1080 cells. In addition, MON-p53, an encapsulated metal-organic network (MON) containing a plasmid expressing the tumor suppressor p53, is designed to treat cancer by switching apoptosis to ferroptosis, and long-term treatment (i.e., for 75 days) was shown to have potent anticancer activity in tumor-bearing mice. Similarly, RSL3 micelles—nanoparticles composed of an AA-conjugated amphiphilic copolymer encapsuling RSL3—have been shown to increase the efficacy of RSL3 in vivo, providing a novel approach to overcome multidrug resistance and improve cancer treatment. Moreover, a nanodrug consisting of nanoparticle ferritin‒bound erastin and rapamycin (NFER) with significantly improved drug-like properties, was proposed to increase ferroptosis by decreasing GPX4 activity as well as reducing lipid peroxidation levels. Interestingly, Sal-AuNP, a gold nanoparticle loaded with salinomycin, has been shown to induce ferroptosis in breast cancer stem cells by causing oxidative stress, mitochondrial dysfunction, and lipid oxidation. Recently, a hypercarbon-centered gold(I) cluster prodrug was engineered and then shown to accelerate ferroptosis in EJ cells (a human bladder cancer cell line) by inhibiting thioredoxin reductase (TrxR) activity.

91To maximize the efficacy of cancer therapeutics, nanoparticle-based combination therapies have been widely studied in various preclinical models. For example, CSO-SS-Cy7-Hex/SPION/Srfn and SPFeN have been studied as photothermal ferrotherapies for use in cancer, combining photosensitizers with ferroptosis-inducing molecules that can exert synergistic anticancer effects. Another photosensitizer—chlorin e6 (Ce6)—has been used in several delivery systems, including Ce6-erastin nanoparticles, SRF@Hb-Ce6, Ce6@MOF, and HAS-Ce6-IrO2. Liu et al. reported excellent chemodynamic/photodynamic synergy using mCMSN, a biodegradable cancer cell membrane‒coated mesoporous copper/manganese silicate nanosphere that increases ferroptosis via laser irradiation‒triggered ROS production and the GSH-activated Fenton reaction. To achieve a synergistic ferroptosis and immunomodulatory effect in treating cancer, Zhang et al. engineered a biomimetic magnetosome (Pa-M/Ti-NC) consisting of an Fe3O4 magnetic nanocluster (NC) as the core, pre-engineered leukocyte membranes loaded with TGF-β inhibitor (Ti), and a PD-1 antibody (Pa) anchored to the membrane surface; they found that this magnetosome potently promoted ferroptosis in cancer cells via cooperative activity between the Pa and Ti components, causing reprogramming of the immunogenic microenvironment. Interestingly, NMIL-100@Gox@C was designed to activate ferroptosis and induce cancer starvation; specifically, the glucose oxidase (Gox) enzyme catalyzes glucose to generate sufficient levels of H2O2 in order to induce ferroptosis, simultaneously causing starvation of the cancer cells via the Gox-mediated consumption of glucose. Yang et al. developed an HLCaP nanoreactor that encapsulates lipoxidase and hemin in biodegradable PLGA (poly(lactic-co-glycolic acid)), which releases its contents in the acidic tumor microenvironment, synergistically inducing ferroptosis using the PUFA substrates produced in the tumor debris using radiofrequency ablation; importantly, these so-called tumor debris‒fueled nanoreactors have been shown to prevent tumor recurrence and metastasis. Similarly, the mesoporous carbon nanoparticle FeCO-Dox@MCN is designed to exploit the combined effects of chemotherapy, photothermal therapy, and gas therapy in order to increase carbon monoxide‒induced ferroptosis.

92By contrast, relatively few studies have focused on ferroptosis-suppressing nanoparticles (Fig. 6). Here, we discuss two nanoparticles that have been reported to inhibit ferroptosis. The first is a carboxyl-modified polystyrene nanoparticle (CPS) that effectively inhibits ferroptosis by reducing cellular ROS mediated by TFEB (transcription factor EB). The second is DEF-HCC-PEG, a deferoxamine-binding nanoparticle that protects cells against both ferroptosis and senescence. Importantly, ferroptosis is a rapidly developing field, with new mechanisms underlying ferroptosis being identified; thus, additional studies are clearly helpful for accelerating the development of clinical applications using nanoparticles to target ferroptosis.

93Over the past decade, a rapidly growing body of evidence has shown that ferroptosis plays a key role in a wide range of diseases and conditions, including cancer, neurodegenerative disease, tissue/organ injury, inflammation, and inflammatory diseases. Despite many obstacles along the way, it is now widely accepted that targeting ferroptosis holds great promise for promoting the development and clinical translation of ferroptosis-based therapeutic strategies.

94To date, nearly all in vivo ferroptosis-related studies were based on preclinical animal models, and several challenges limit their translation to clinical practice.

95As discussed in this review, numerous molecules—including experimental agents, naturally derived compounds, nanoparticles, and clinically approved drugs—have been shown to modulate ferroptosis, revealing promising new strategies for the development of ferroptosis-based therapies. However, with respect to small molecule compounds, only ~15% of proteins in the human proteome are estimated to be druggable. In addition, poor pharmacokinetics remains a major bottleneck for the further development of both ferroptosis antagonists such as Fer-1 and Lip-1 and ferroptosis agonists such as RSL3 and ML210. To develop novel bioactive ferroptosis-targeted drug candidates, researchers have studied natural products derived from microorganisms and plants, particularly traditional herbs. Furthermore, high-throughput screening, artificial intelligence, and other new technologies may also be used to identify more effective drug candidates to target ferroptosis.

96Another challenge to ferroptosis-based drug discovery is the rational design of combination strategies to treat various ferroptosis-related diseases. It is generally well accepted that ferroptosis plays a pathogenic role in various diseases. In addition, many existing therapies can trigger apoptosis or other types of cell death. For conditions such as cancer, tissue injury, and neurodegenerative disease, various forms of cell death can occur simultaneously. Therefore, developing ferroptosis modulators that can act synergistically with existing therapies may serve as a promising strategy for treating these diseases, particularly drug-resistant cancer. Combining ferroptosis agonists with other anticancer therapies may also increase ferroptosis, as well as other types of cell death, thereby eradicating cancer more effectively.

97In addition to conventional ferroptosis inhibitors and inducers, new technologies such as proteolysis-targeting chimeras (PROTACs), RNA-based therapies, gene editing, and peptide and protein drugs have been applied to develop ferroptosis-targeting drugs.

98Targeted protein degradation (TPD) is a new and challenging therapeutic modality that can degrade “undrugged” targets and other difficult protein targets such as proteins with a broad and shallow active pocket and/or “smooth” surfaces. Three major classifications of protein degraders have been developed, including PROTACs, monomeric targeted protein degraders, and molecular glues (MGs). PROTACs are heterobifunctional small molecules consisting of two ligands connected via a linker. The ligands are designed to bind to an E3 ubiquitin ligase and a target of interest. This chemically induced close proximity between the protein and E3 ligase results in the target protein’s ubiquitylation and subsequent degradation via the ubiquitin-proteasome system (UPS). Recently, TPD has gained considerable attention due to encouraging results obtained from clinical trials involving ARV-110 and ARV-471, two PROTACs that target and degrade the androgen receptor (AR) and estrogen receptor (ER), respectively. Recently, several groups have attempted to develop PROTAC-based ferroptosis modulators. For example, Luo et al. developed the PROTAC dGPX4 to target GPX4 and showed in vivo that it potently induces ferroptosis selectively in cancer cells with no apparent side effects, providing a promising novel strategy for the development of clinically applicable GPX4-targeted drugs. Similar to PROTACs, photodegradation-targeting chimeras (PDTACs), lysosome-targeting chimeras (LYTACs), and autophagy-targeting chimeras (AUTACs) have also been developed for use in TPD-based therapies. For example, Liu et al. reported the targeted photodegradation of GPX4 using a PDTAC conjugated with the photosensitizer verteporfin upon red-light irradiation, and showed that this PDTAC can induce immunogenic ferroptosis and may improve the efficacy of cancer immunotherapy.

99Unlike PROTACs, monomeric targeted protein degraders are designed to either directly bind to their target protein or induce their target protein’s degradation. Therefore, monomeric degraders have a comparatively small molecular weight and can easily cross the BBB. Fulvestrant, a selective ER degrader, has been approved for treating ER-positive breast cancer, and some monomeric degraders have been tested in clinical trials. MGs are relatively simple small molecules that increase the interaction between E3 ligase and the protein substrate in order to induce the UPS-induced degradation of target proteins. Thanks to their low molecular weight, MGs are believed to have more favorable pharmacokinetics than other compounds. Thalidomide was developed as a MG to bring various targets in proximity to the E3 ubiquitin ligase CUL4-RBX1-DDB1-CRBN for degradation. Lenalidomide, a thalidomide derivative, has been approved to treat multiple myeloma by inducing the degradation of oncoproteins. Thus, monomeric degraders and MGs represent promising new strategies for targeted protein degradation. However, to the best of our knowledge no reports have yet been published regarding the use of monomeric targeted protein degraders or MG-based degraders to target ferroptosis-related proteins.

100AI, which was first defined by John McCarthy back in 1956 as the “science and engineering of making intelligent machines”, primarily involves computer science, mathematics, and neuroscience, but can also be used for drug discovery. Given that the traditional process of drug discovery is time-consuming and labor-intensive, it is now widely accepted that AI—particularly machine learning—can significantly increase prediction accuracy and accelerate the drug discovery process. Deep learning, which combines machine learning and AI, can serve as a particularly cost-effective platform for discovering new drugs, as it directly enables rapid machine-based decision-making using an artificial neural network. Today, AI-based technologies have been applied to virtually every step in the drug discovery process, including target identification, drug design, drug screening, chemical synthesis, and prediction of the drug’s properties and mode of action., In addition, AI can be used to quickly identify lead compounds extracted from plants and microorganisms. Recently, AI was successfully used in cancer research and precision medicine. With respect to ferroptosis, Wu et al. used a combination of bioinformatics, network pharmacology, and AI to show that ferroptosis and the TGF-β signaling pathway may contribute to the protective effects of celastrol, a plant-derived triterpene, against type 2 diabetes. It is therefore reasonable to speculate that AI will significantly accelerate the discovery of new ferroptosis-targeted drugs.

101In addition to protein targets, a new line in the field of drug discovery focuses on RNA-based therapeutics, particularly for the management of genetic diseases. RNA-based therapeutics now include the use of messenger RNA (mRNA), RNAi, single-stranded antisense oligonucleotides, aptamers, ribozymes, and CRISPR-Cas endonuclease‒mediated gene editing. For example, mRNA-based therapies have been successfully used to develop personalized cancer vaccines, as well as vaccines against infectious diseases. With respect to RNAi, four RNAi-based therapies have been approved for use in patients, including patisiran for hereditary transthyretin amyloidosis, givosiran for acute hepatic porphyria, lumasiran for primary hyperoxaluria type 1, and inclisiran for heterozygous familial hypercholesterolemia and atherosclerotic cardiovascular disease. These four approved RNAi-based compounds are based on small-interfering RNA (siRNA) technology, which decrease the level of specific mRNAs in order to reduce the corresponding protein levels. Currently, >50 clinical trials are investigating siRNA-based compounds. Unfortunately, however, no clinical trials are investigating ferroptosis-targeting agents using RNA-based technologies. Nevertheless, although the inability to precisely deliver RNA-based tools remains a major barrier to developing RNA therapies, new RNA-based modalities—particularly CRISPR/Cas9-based gene editing—will likely gain considerable traction in the near future.

102Peptide-based drugs represent a unique pharmaceutical niche between small-molecule compounds and biologicals. Compared to chemical compounds, peptide-based drugs are more efficient, safer, and better tolerated. This category of drugs has several other advantages as well, including high selectivity and reduced tissue accumulation. Biological drugs such as monoclonal antibodies bind with high specificity to their target proteins. Recently, bispecific antibodies—which simultaneously target multiple antigens and/or epitopes using two specific antibodies—have attracted considerable attention and may represent the next generation of antibody-targeted T cell‒based cancer immunotherapies. Furthermore, antibody-drug conjugates have high specificity and are highly efficient at delivering toxic small molecules in order to specifically target cancer cells.

103Ferroptosis is now generally accepted as playing a key pathogenic role in a wide range of diseases and conditions. As summarized in this review, a large amount of compelling evidence supports the notion that targeting ferroptosis can provide promising new options for the treatment of many ferroptosis-related diseases. As our understanding of ferroptosis-regulating pathways—including iron metabolism, lipid metabolism, and oxidative-reductive pathways—increases to reveal new druggable targets, ferroptosis modulators are likely to provide new opportunities to develop treatments for many currently incurable diseases. Although further studies are clearly needed, the novel technologies summarized in this review will facilitate the development of safer, more effective ferroptosis-targeted drugs.