Ferroptosis Regulated by Hypoxia in Cells

Ferroptosis is an oxidative damage-related, iron-dependent regulated cell death with intracellular lipid peroxide accumulation, which is associated with many physiological and pathological processes. It exhibits unique features that are morphologically, biochemically, and immunologically distinct from other regulated cell death forms. Ferroptosis is regulated by iron metabolism, lipid metabolism, anti-oxidant defense systems, as well as various signal pathways. Hypoxia, which is found in a group of physiological and pathological conditions, can affect multiple cellular functions by activation of the hypoxia-inducible factor (HIF) signaling and other mechanisms. Emerging evidence demonstrated that hypoxia regulates ferroptosis in certain cell types and conditions. In this review, we summarize the basic mechanisms and regulations of ferroptosis and hypoxia, as well as the regulation of ferroptosis by hypoxia in physiological and pathological conditions, which may contribute to the numerous diseases therapies.


Introduction
Ferroptosis is a novel characterized oxidative damage-related regulated cell death, which is mainly driven by iron accumulation, lipid peroxidation, and subsequent plasma membrane rupture [1,2]. As a newly discovered form of regulated cell death, ferroptosis is morphologically, chemically, and genetically different from apoptosis, autophagy, necroptosis, and pyroptosis, and is regulated by its unique metabolism and regulatory mechanism [1,3]. With the increasing studies in recent years, more and more evidence demonstrated that ferroptosis may play a significant role in multiple physiological and pathological processes, such as embryonic development, cancer, neurodegeneration disease, and organ disorders caused by drug-induced toxicity or ischemia/reperfusion injury (IRI) [4][5][6][7][8][9]. In cancer, ferroptosis is reported to play an essential role in suppressing spontaneous tumorigenesis in mouse models [10,11]; many cancer cells resistant to conventional therapies were found to be susceptible to ferroptosis [12,13], suggesting that the induction of ferroptosis in cancer cells is a promising anti-tumor strategy. In neurodegeneration disease and the injuries of many organs such as the heart, brain, kidney, and liver, ferroptosis was identified as an important form of cell death that results in organ damage [7][8][9], suggesting that inhibiting ferroptosis could be an important way to limit organ damage in disorders. It seems that ferroptosis is regulated in context-dependent mechanisms, which may be important for developing therapeutic strategies for these diseases by targeting ferroptosis.
Low levels of oxygen in tissues, called hypoxia, can affect multiple cellular functions across a variety of cell types, and is associated with both physiological and pathological conditions, such as high altitude, solid tumors, and ischemia of organs [14][15][16]. Hypoxia is caused by many potential reasons in humans, such as insufficient blood flow to a specific area, decrease in hemoglobin levels, or treatment with chemical compounds [17,18]. Under hypoxic conditions, cells activate the hypoxia signaling pathway, which is mainly mediated by the hypoxia-inducible factors (HIFs), to induce a series of cellular responses to help cells to adapt to or escape from the hypoxic environment [19][20][21]. Interestingly, hypoxia is often  Excess intracellular Fe 2+ can be stored by ferritin, a protein complex composed of ferritin light chain (FTL) and ferritin heavy chain 1 (FTH1), to prevent Fe 2+ from being oxidized [27]. Nuclear receptor coactivator 4 (NCOA4) binds to ferritin and then transports iron-bound ferritin to the autophagosome for lysosomal degradation (also known as ferritinophagy), leading to the release of Fe 2+ [33]. Thus, increasing the expression of FTL and FTH1, or the inhibition of NCOA4, will reduce the intracellular free Fe 2+ and inhibit ferroptosis [27,33]. In addition to ferritin, the overexpression of ferritin mitochondrial (FTMT), another iron-storage protein in mitochondria, leads to the decrease in intracellular free Fe 2+ and inhibits ferroptosis [34].
The intracellular Fe 2+ can be exported by iron-efflux protein ferroportin (FPN1, also known as solute carrier family 40 member 1, SLC40A1) and then be reoxidized to Fe 3+ [35]. The internalization and degradation of FPN1 are induced by hepcidin (encoded by the Hamp1 gene), a peptide hormone regulated by the BMP/Smad pathway [36]. The decrease in FPN1 or elevation of hepcidin levels releases the free Fe 2+ to promote ferroptosis [27].
Additionally, iron was used for biogenesis of some mitochondrial proteins with ironsulfur cluster, such as NFS1, CISD1, and CISD2, leading to the decrease in the available iron levels to inhibit ferroptosis [37][38][39].
PUFA production for subsequent lipid peroxidation is supported by upstream lipid synthesis and metabolism pathways. PUFA-phospholipids, especially arachidonic acid (AA) and adrenic acid (AdA), are the main substrates of lipid peroxidation in ferroptosis [40,41]. Acyl-CoA synthetase long-chain family member 4 (ACSL4) catalyzes the ligation reaction of CoA with AdA/AA to form CoA-AdA/AA intermediate, which subsequently undergoes esterification with membrane phosphatidylethanolamine to form PE-AdA/AA by lysophosphatidylcholine acyltransferase 3 (LPCAT3). The loss of either ACSL4 or LP-CAT3 results in the resistance to ferroptosis [49]. Meanwhile, monounsaturated fatty acids (MUFAs) that lack the bis-allylic positions readily for peroxidation, inhibit ferroptosis by competing with PUFAs. The inhibition of ferroptosis by MUFAs relies on ACSL3, which catalyzes the ligation reaction of CoA with MUFAs, or stearoyl-CoA desaturase 1 (SCD1), the rate-limiting enzyme in MUFA production [50,51].
The processes of lipid synthesis, absorption, storage, and release are also involved in ferroptosis. Long-chain PUFAs, including AA and AdA, are synthesized from dietary essential fatty acids in cells by a series of enzymatic reactions involving the elongation of very long-chain fatty acid protein (ELOVL) and fatty acid desaturase (FADS). The silencing or inhibition of ELOVL5, FADS1, and FADS2 was shown to inhibit ferroptosis [52,53]. β-oxidation, the process in which fatty acids are broken down to produce acetyl-CoA, is generally believed to negatively regulate ferroptosis by decreasing the availability of unesterified PUFAs [41]. Fatty acids are absorbed into cells through various fatty acid transport proteins, including fatty acid translocase (FAT/CD36), fatty acid transport proteins (FATPs), and fatty acid-binding proteins (FABPs) [54]. CD36 and FATP2 mediate the absorption of AA and AdA to promote ferroptosis in certain cells [55,56]. In addition to lipid uptake, FABPs also facilitate the transport of lipids to specific compartments in the cell, such as to the lipid droplet for lipid storage [57]. FABP3, FABP4, and FABP7 were reported to inhibit ferroptosis through enhancing fatty acid uptake and lipid storage with lipid droplet formation [58,59]. Moreover, increased lipid storage by tumor protein D52 (TPD52) represses lipid peroxidation and ferroptosis, whereas degradation of lipid droplets by autophagy (known as lipophagy) enhances free fatty acids production, and increases lipid peroxidation and ferroptosis [60]. Hypoxia-inducible, lipid droplet-associated protein (HILPDA), can increase PUFAs incorporation into triacylglycerols and phospholipids through binding to and inhibiting adipose triglyceride lipase, responsible for breaking down triacylglycerols [61]. During the continuous membrane remodeling, PUFAs are released from membrane phospholipids through hydrolysis catalyzed by phospholipase A2 (PLA 2 ). PLA2G6 (also known as iPLA 2 β), a Ca 2+ -independent PLA 2 , was reported to inhibit ferroptosis by cleaving oxidized hydroperoxy-arachidonoyl (C20:4)-or adrenoyl (C22:4)-phosphatidylethanolamine (Hp-PE) from membrane phospholipids [62].

Antioxidant Defense System
Ferroptosis is caused by lipid peroxidation, which is inhibited by the antioxidant defense system in cells. Glutathione peroxidase 4 (GPX4) is a selenocysteine-containing enzyme that plays a central role in ferroptosis inhibition, which inhibits lipid peroxidation by directly reducing toxic phospholipid hydroperoxides (PL-OOH) to non-toxic phospholipid alcohols (PL-OH) by oxidizing glutathione (GSH) [63,64]. GPX4 dysfunction always results in the uncontrolled accumulation of lipid peroxides and ferroptosis. A group of smallmolecule compounds, including RSL3, ML162, ML210, FIN56, and FINO 2 , can directly or indirectly inhibit GPX4 activity as well as promote the GPX4 protein degradation, to induce ferroptosis [63][64][65][66]. GPX4 is transcriptionally upregulated by the transcription factor AP-2 gamma (TFAP2C) and specificity protein 1 (SP1), which leads to ferroptosis inhibition [67]. As a selenocysteine-containing protein, GPX4 levels are regulated by the selenocysteine tRNA. Selenocysteine tRNA is positively regulated by isopentenyl pyrophosphate (IPP), which is a product of the mevalonate pathway for lipid synthesis [68]. Furthermore, GPX4 is also regulated by SCD1/FADS2, leading to the ferroptosis inhibition [69]. It seems that GPX4 is also regulated by lipid metabolism.
The reduction reaction catalyzed by GPX4 requires GSH, which is synthesized from glutamate, cysteine, and glycine [1,2,6]. Thus, ferroptosis is also regulated by the amino acid metabolism involved in GSH synthesis. System Xc − , a glutamic acid/cystine antiporter in the plasma membrane that is constituted by the transport subunit solute carrier family 7 member 11 (SLC7A11) and regulatory subunit solute carrier family 3 member 2 (SLC3A2), is responsible for the cystine uptake and plays a crucial role in ferroptosis repression [70]. Several small-molecule compounds (e.g., erastin, sulfasalazine, and sorafenib), have been shown to be the inhibitors of system Xc − mediated cystine uptake, which decreases the GSH synthesis to trigger ferroptosis [65,70]. Beclin-1, an autophagy-associated protein, interacts with SLC7A11 when phosphorylated at Ser90/93/96 by AMPK, leading to the inhibition of system Xc − to induce ferroptosis [71][72][73][74]. The cystine uptaken by system Xc − is further reduced to cysteine by GSH and/or thioredoxin reductase 1 (TXNRD1) and is used for GSH synthesis by glutamate cysteine ligase (GCL) and glutathione synthetase (GSS). Inhibition of GCL or GSS, or activation of ChaC glutathione specific gamma-glutamylcyclotransferase 1 (CHAC1), an enzyme that catalyzes the GSH degradation, leading to ferroptosis induction [24,65,75,76].
Moreover, many proteins, especially transcription factors, regulate ferroptosis by directly or indirectly modulating a group of the above-mentioned genes or proteins. For example, nuclear factor erythroid 2-related factor 2 (NRF2, also known as nuclear factor erythroid-derived 2-like 2, NFE2L2), a master regulator of oxidative stress signaling, inhibits ferroptosis by transcriptionally regulating the expression of FPN1, HO-1, FTL, and FTH1 involved in iron metabolism as well as SLC7A11, TXNRD1, GSS, GCLC, GCLM, CHAC1, GPX4, FSP1, and NQO1 (quinone oxidoreductase 1) involved in antioxidant systems [84]. Tumor suppressor p53, a transcription factor that regulates the expression of stress response genes and mediates a variety of anti-proliferative processes, plays a dual role in ferroptosis [85]. p53 induces ferroptosis through transcriptionally regulating SLC7A11 for the anti-oxidant defense, PTGS2, ALOX12 for lipid metabolism, FDXR for iron Cells 2023, 12, 1050 6 of 22 metabolism, and GLS2 (glutaminase 2, an enzyme for glutaminolysis that is involved in GSH synthesis and lipid synthesis) for both anti-oxidant system and lipid metabolism [85]. p53 also transcriptionally induces SAT1 (Spermidine/spermine N1-acetyltransferase 1) to elevate ALOX15 levels, leading to lipid peroxidation and ferroptosis [85,86]. On the contrary, p53 was also found to inhibit ferroptosis by transcriptionally inducing p21 to enhance GSH levels and Parkin to induce mitophagy, as well as by directly binding to the dipeptidyl peptidase DPP4 to inhibit NOX-mediated lipid peroxidation [85]. Interestingly, HIFs, acting as a transcription factor for hypoxia, also play a complex role in ferroptosis, which will be reviewed later.

Hypoxia and Hypoxia-Induced Factors
Hypoxia refers to a limited oxygen level in tissues, which is caused by a variety of mechanisms in both physiological and pathological conditions. For instance, hypoxia in tumors can be caused by the abnormal disorganization of tumor vasculature (perfusion hypoxia), long oxygen diffusion distances (>70 µm) from blood vessels to tumor cells (diffusion hypoxia), or a reduced oxygen transport capacity resulting from chemotherapy-induced anemia (anemic hypoxia) [17,18]. These hypoxia conditions lead to the proliferation, angiogenesis, metastasis, metabolism reprogramming, stemness, immune evasion, and therapy resistance of tumor cells, and ultimately contribute to tumor progression [14,20,87,88]. During surgical operation or drug treatment in some organs, ischemia/reperfusion (I/R) results in the hypoxia/reoxygenation (H/R) of cells because of the changes in oxygen supply from the bloodstream, which finally leads to cell death and injury [16,89]. Moreover, hypoxic placenta resulting from the dysregulation of angiogenic factors, induces oxidative stress, lipid peroxidation, endothelial dysfunction, and peripheral vasoconstriction, leading to abnormal trophoblastic invasion and preeclampsia, which is a serious and distinct type of pregnancy-induced hypertension [13]. Generally, hypoxia is an important environmental factor in many physiological and pathological conditions.
Through transcriptional regulation of different genes by HIFs and other mechanisms, hypoxia regulates many different cell processes, (1) to change the hypoxic environment (e.g., inducing angiogenesis), (2) to reprogram the cellular metabolism to adapt to hypoxia (e.g., increase in glycolysis), (3) to let cells to move to a new place without hypoxia (e.g., promoting cell migration and invasion), and eventually resulting in cell death or survival [19,21]. As an oxidative-related and iron-dependent regulated cell death, ferroptosis is an important form of cell death involved in the hypoxia-induced cell response.

Hypoxia and HIFs Inhibit Ferroptosis in Tumor Cells
Hypoxia is often found during tumor progression, which is always accompanied by decreased cell death. Many studies reported that HIF-1 can inhibit the many kinds of regulated cell death, such as apoptosis and autophagy [99,100]. In addition to apoptosis and autophagy, ferroptosis was also found to be inhibited in tumors ( Figure 2). In clear cell renal cell carcinoma (ccRCC), both HIF-1α and HIF-2α are stabilized by VHL loss, leading to the decrease in ferroptosis sensitivity to erastin or BSO via fatty acid β-oxidation and mitochondrial ATP-synthesis [101]. A recent study on clear cell renal cell carcinoma showed that the iron sulfur cluster assembly 2 (ISCA2), a component of the late mitochondrial iron sulfur cluster assembly complex, was induced by hypoxia (1% O 2 ) to promote HIF-1α/HIF-2α translation, leading to the inhibition of iron overload and erastin-or RSL3-induced ferroptosis [102].
Therefore, ISCA2 inhibition induces ferroptosis and reduces tumor growth in vivo. In malignant mesothelioma under 1% O 2 hypoxia, HIF-1 induces Carbonic anhydrase 9 (CA9) expression, which in turn reduces catalytic Fe 2+ by downregulating TFRC and upregulating FTL and FTH1 to inhibit erastin-induced ferroptosis [103]. In human fibrosarcoma HT-1080 cells and non-small cell lung cancer Calu-1 cells, HIF-1α stabilized by 1% O 2 hypoxia enhances fatty acid uptake and lipid storage by transcriptionally elevating FABP3 and FABP7 expression, and finally inhibits RSL3-induced ferroptosis [59]. In gastric cancer cells under the 1% O 2 hypoxia condition, HIF-1α transactivates the long noncoding RNA CBSLR, which in turn forms a complex with YTH N6-methyladenosine RNA binding protein 2 (YTHDF2) and CBS mRNA to reduce CBS mRNA stability in an N6-methyladenosine (m 6 A) modification-dependent manner, thus reducing the ACSL4 methylation, degrading ACSL4 via the ubiquitination-proteasome pathway, and eventually protecting gastric cancer cells from ferroptosis in a hypoxic tumor microenvironment [104]. Another study on the peritoneal metastasis of gastric cancer showed that 1% O 2 hypoxia activates HIF-1α to transcriptionally induce lncRNA-PMAN expression via binding to the HRE sequence in its promotor, which recruits the embryonic lethal abnormal vision like RNA binding protein 1 (ELAVL1) to cytoplasm to stabilize SLC7A11 mRNA, thus leading to the inhibition of erastinor RSL3-induced ferroptosis [105]. The SLC7A11 mRNA is also stabilized by blocking its N6methyladenosine (m 6 A) modification and YTHDF2-dependent degradation by suppressing methyltransferase like 14 (METTL14), which is a central component of the m6A methylated transferase complex, in a HIF-1α-dependent manner, to inhibit ferroptosis in hepatocellular carcinoma [106]. Furthermore, it was reported that hypoxia (1% O 2 ) induces HIF-1α through activating PI3K to phosphorylate AKT, then upregulating SLC7A11 expression in glioma cells to suppress sulfasalazine-induced ferroptosis [107]. The core circadian clock gene period 1 (PER1) can bind to and degrades HIF-1α; while HIF-1α can bind to the PER1 promotor and reduce PER1 transcription, which forms a negative feedback loop [108]. In oral squamous cell carcinoma, PER1 expression is often decreased, which in turn induces HIF-1α to inhibit ferroptosis via promoting GPX4 and SLC7A11 expression and inhibiting TFRC expression [108]. A recent study showed that HIF-1α stabilized by myriocin, an inhibitor of the de novo synthesis of sphingolipid, could promote the expression of PDK1 (3-phosphoinositide-dependent protein kinase-1, an enzyme involved in glycolysis) and BNIP3 (BCL2/adenovirus E1B 19 kDa protein-interacting protein 3, a member of the apoptotic Bcl-2 protein family) and alter the intracellular levels of glucose metabolites, to inhibit erastin-or glutamate-induced ferroptosis in HT-1080 human fibrosarcoma cells, but not in GES-1 human gastric epithelial cells and SK-Hep-1 human hepatoma cells [109]. Taken together, HIFs, especially HIF-1α, play a crucial role in suppressing ferroptosis by hypoxia. Therefore, ISCA2 inhibition induces ferroptosis and reduces tumor growth in vivo. In malignant mesothelioma under 1% O2 hypoxia, HIF-1 induces Carbonic anhydrase 9 (CA9) expression, which in turn reduces catalytic Fe 2+ by downregulating TFRC and upregulating FTL and FTH1 to inhibit erastin-induced ferroptosis [103]. In human fibrosarcoma HT-1080 cells and non-small cell lung cancer Calu-1 cells, HIF-1α stabilized by 1% O2 hypoxia enhances fatty acid uptake and lipid storage by transcriptionally elevating FABP3 and FABP7 expression, and finally inhibits RSL3-induced ferroptosis [59]. In gastric cancer cells under the 1% O2 hypoxia condition, HIF-1α transactivates the long noncoding RNA CBSLR, which in turn forms a complex with YTH N6-methyladenosine RNA binding protein 2 (YTHDF2) and CBS mRNA to reduce CBS mRNA stability in an N6-methyladenosine (m 6 A) modification-dependent manner, thus reducing the ACSL4 methylation, degrading ACSL4 via the ubiquitination-proteasome pathway, and eventually protecting gastric cancer cells from ferroptosis in a hypoxic tumor microenvironment [104]. Another study on the peritoneal metastasis of gastric cancer showed that 1% O2 hypoxia activates In addition to the HIFs, hypoxia also inhibits ferroptosis of tumor cells through other mechanisms. In HT-1080 fibrosarcoma cells, 1% O 2 hypoxia increases FTH1 expression in an NCOA4-independently manner to inhibit RSL3-induced ferroptosis [110]. In non-small cell lung cancers, 1% O 2 hypoxia induces the intracellular expression of angiopoietin-like 4 (ANGPTL4) and its extracellular secretion by exosomes to neighboring normoxic cells, both of which increase the expression of GPX4, SLC7A11, FTL, and FTH1, to inhibit the RSL3-or irradiation-induced ferroptosis [111].

Hypoxia and HIFs Inhibit Ferroptosis in Normal Cells
Hypoxia and HIFs suppresses ferroptosis not only in tumor cells but also in normal cells. In primary human macrophages, hypoxia (1% O 2 ) reduces NCOA4 expression to inhibit ferritinophagy, resulting in the increases in FTMT levels to reduce RSL3-induced ferroptosis [110]. A following study showed that hypoxia (1% O 2 ) stabilized HIF-2α to promote the transcription of FTMT, and activated thrombin to cleave FTMT from a 27 kDa precursor to a 22 kDa mature form, both of which resulted in increased FTMT to attenuate erastin-or RSL3-induced ferroptosis [112]. In the receptor activator of nuclear factor Kappa B ligand (RANKL)-induced differentiation of osteoclasts, hypoxia inhibits RANKL-induced ferritinophagy via inhibiting autophagosome formation by HIF-1α, thereby protecting osteoclasts from ferroptosis [113]. In rat embryonic cardiomyoblast H9C2 cells, 1% O 2 hypoxia upregulates SENP1 to promote the deSUMOylation of HIF-1α and ACSL4, leading to HIF-1α stabilization and decreased ACSL4 protein levels, in turn protecting cells against erastin-induced ferroptosis [114]. In pulmonary artery smooth muscle cells and a pulmonary arterial hypertension rat model, hypoxia (3% O 2 for cells and 10% O 2 for mice, respectively) induces the deubiquitinase OTU domain-containing ubiquitin aldehydebinding protein 1 (OTUB1), to stabilize SLC7A11, and eventually inhibiting erastin-induced ferroptosis [115]. It was reported recently that hypoxic-ischemic conditions induce HSPB1 in the hippocampus tissues to repress the hypoxia (anoxia)-induced ferroptosis of neuronal cells and hypoxic-ischemic brain damage via promoting expression of glucose-6-phosphate dehydrogenase, GPX4, and SLC7A11, as well as augmenting TFRC levels [116]. HIF-1α stabilized by myriocin was also found to inhibit erastin-induced ferroptosis via promoting PDK1 and BNIP3 expression as well as altering the intracellular glucose metabolism in HT22 mouse hippocampal neuronal cells and PC-12 rat neural cells [109]. Another recent study in cardiomyocytes showed that miR-210-3p is enriched in the hypoxia (1% O 2 )-conditioned cardiac microvascular endothelial cells-derived exosomes, which inhibits TFRC expression by directly interacting with TFRC mRNA, attenuating erastin-induced myocardial cell ferroptosis [117].

Hypoxia and HIFs Promote Ferroptosis in Tumor Cells
Ferroptosis is not always inhibited by hypoxia and HIFs. In tumor cells, hypoxia and HIFs also induces ferroptosis in some conditions ( Figure 3). Interestingly, HIF-2α, but not HIF1α, mainly mediates the induction of ferroptosis in tumor cells. In renal clear-cell carcinomas, HIF-2α activates HILPDA, which selectively enriches PUFA to promote lipid peroxidation and GPX4 inhibitor (RSL3, ML210, or ML162)-induced ferroptosis [61]. In glioblastoma (GBM), HIF-2α induced by the prolyl hydroxylase (PHD) inhibitor roxadustat, upregulated ferroptosis regulatory genes such as ACSL4, PTGS2, and CHAC1, to enhance lipid peroxidation and erastin-induced ferroptosis, leading to the suppression of GBM cell growth in vitro and in vivo [118]. Another report showed that HIF-2α upregulated genes involved in lipid and iron metabolism in both colorectal cancer (CRC) cells and colon tumors in mice, which result in the cells being susceptible to ferroptosis induced by dimethyl fumarate [119]. It was also reported that lysine (K)-specific demethylase 4A (KDM4A), which was highly expressed in cervical cancer tissue and could be induced under cobalt chloride mimicking hypoxia conditions, reduces the H3K9me3 level in the HIF-1α promoter region to elevate HIF-1α transcription, leading to the increased expression of TFRC and DMT1 via activating the HRE sequence in their promoters, and finally resulting in the increase in erastin-induced ferroptosis of cervical cancer cells [120]. In addition to HIFs, hypoxia also induced E2F transcription factor 7(E2F7) to increase the transcription of splicing factor quaking (QKI), which promotes the biogenesis of circBCAR3, a circular RNA highly expressed in the esophageal cancer cells [121]. CircBCAR3 binds to miR-27a-3p by the competitive endogenous RNA mechanism to upregulate transportin-1 (TNPO1), a binding partner of CA9, and leading to the inhibition of CA9 to promote ferroptosis. Cells 2023, 12, x FOR PEER REVIEW 11 of 23 In retinal pigment epithelium cells, stabilized by the 3%O2 hypoxia condition and PHD inhibitor dimethyloxalylglycine (DMOG) treatment, HIF aggravates sodium iodateinduced ferroptosis by upregulating superoxide dismutase (SOD) to execute a peroxidative rather than antioxidative role, as well as increasing the iron importers ZIP8 and ZIP14 to enhance iron import [132]. Moreover, HIF-2α was found to be upregulated by myostatin in response to cigarette smoke exposure, leading to ferroptosis in myotubes and chronic obstructive pulmonary disease-related skeletal muscle dysfunction [133].
As an environment featuring hypobaric hypoxia, high altitude (HA) exposure increased ferroptosis sensitivity in adipose tissue with elevated levels of iron, ROS, MDA, and 4-HNE, as well as GSH depletion [134]. Acute high-altitude hypoxia exposure in mice leads to cerebral formaldehyde accumulation to induce neuronal ferroptosis [135]. A recent study suggested that hypobaric hypoxia at high-altitude induces both apoptosis and ferroptosis via the JNK signaling pathway by depleting keratin 18 (Krt18) and elevating JNK3 (MAPK10), as well as increasing ASCL4 for ferroptosis [136]. An integrative comparison of the oviduct epithelial cells between yak at high altitude and bovine suggested that the mitophagy-animal pathway and HIF-1 signaling pathway may also be involved in high-altitude-induced ferroptosis [137].

Hypoxia and HIFs Induce Ferroptosis in Normal Cells
Ferroptosis can also be induced in normal cells under hypoxia conditions. It was reported that chronic intermittent hypoxia (CIH) induced ferroptosis in the hippocampus, lung, liver, and cardiomyocytes by downregulating NRF2 and GPX4, as well as upregulating ACSL4, leading to cognitive impairment and injury of brain, lung liver, and heart [122][123][124][125]. In the brain, hypoxic-ischemic induces ferroptosis with the increased TFRC expression, and decreased expression of SLC7A11, TRX-1, and GPX4 [126][127][128]. A recent study showed that hypoxia and ischemia induce chromobox7 (CBX7) in neural progenitor cells, which promote ferroptosis through suppressing the NRF2/HO-1 signaling pathway [128]. In microglial BV-2 cells, hypoxia down-regulates GPX4 and SLC7A11 to induce ferroptosis, which can be reversed by the wild bitter melon extract [129].
It was also found that the hypoxia (1% O 2 ) increased miR-30b-5p to repress SLC7A11 and Pax3 (a transcription factor that promotes the expression of SLC7A11 and the iron exporter FPN1) levels, to induce ferroptosis of trophoblasts, leading to preeclampsia [130]. Additionally, this hypoxia-induced ferroptosis of trophoblasts in the placenta can be inhibited by the phospholipase PLA2G6, which catalyzes the hydrolysis of oxidized PUFAs from membrane phospholipids to attenuate ferroptosis [62]. On the contrary, miR-2115-3p interacts with the mRNA of glutamic-oxaloacetic transaminase1 (GOT1) to repress its expression, increasing GPX4 levels and decreasing ACSL4 and TFRC levels, and eventually leading to inhibiting the hypoxia-promoted ferroptosis in a preeclampsia model [131].
In retinal pigment epithelium cells, stabilized by the 3%O 2 hypoxia condition and PHD inhibitor dimethyloxalylglycine (DMOG) treatment, HIF aggravates sodium iodateinduced ferroptosis by upregulating superoxide dismutase (SOD) to execute a peroxidative rather than antioxidative role, as well as increasing the iron importers ZIP8 and ZIP14 to enhance iron import [132]. Moreover, HIF-2α was found to be upregulated by myostatin in response to cigarette smoke exposure, leading to ferroptosis in myotubes and chronic obstructive pulmonary disease-related skeletal muscle dysfunction [133].
As an environment featuring hypobaric hypoxia, high altitude (HA) exposure increased ferroptosis sensitivity in adipose tissue with elevated levels of iron, ROS, MDA, and 4-HNE, as well as GSH depletion [134]. Acute high-altitude hypoxia exposure in mice leads to cerebral formaldehyde accumulation to induce neuronal ferroptosis [135]. A recent study suggested that hypobaric hypoxia at high-altitude induces both apoptosis and ferroptosis via the JNK signaling pathway by depleting keratin 18 (Krt18) and elevating JNK3 (MAPK10), as well as increasing ASCL4 for ferroptosis [136]. An integrative comparison of the oviduct epithelial cells between yak at high altitude and bovine suggested that the mitophagy-animal pathway and HIF-1 signaling pathway may also be involved in high-altitude-induced ferroptosis [137].

Hypoxia/Reoxygenation (H/R) Promotes Ferroptosis in Normal Cells
I/R in organs, which leads to the H/R of cells, often results in injury because of cell ferroptosis. Many studies reported ferroptosis induction in IRI, especially in the heart. A study on myocardial IRI reported that H/R represses NRF2 to reduce FPN1, resulting in the increase in Fe 2+ and erastin-induced ferroptosis in H9C2 cardiomyocytes [138]. H/R also induces PTGS through HIF-1α to induce ferroptosis in H9C2 cardiomyocytes [139] and a rat model of coronary microembolization (CME)-induced myocardial injury [140]. H/R treatment decreased SMAD7 expression and increased Hamp1 expression to promote erastininduced ferroptosis in H9C2 cardiomyocytes [141]. Atorvastatin blocked the HIF-1α/COX-2 axis and the SMAD7/hepcidin pathway to inhibit CME-or erastin-induced ferroptosis of cardiomyocytes [140,141]. It is reported that H/R treatment induces USP7 to activate p53, leading to the increased TFRC levels to promote ferroptosis of H9C2 cardiomyocytes, as well as in a rat model of myocardial IRI [142]. DNA (cytosine-5)-methyltransferase 1 (DNMT-1) was found to augment the H/R-induced ferroptosis of H9C2 cardiomyocytes by enhancing NCOA4-mediated ferritinophagy by elevating DNA methylation in the NCOA4 promoter [143]. DMT1 expression is significantly induced by H/R treatment to promote the ferroptosis of myocardial cells isolated from mouse models for acute myocardial infarction (AMI) and cardiomyocyte hypoxia injury, which could be inhibited by miR-23a-3p carried by the exosome from human umbilical cord blood-derived mesenchymal stem cells [144]. The transcription factor forkhead box C1 (FOXC1) transcriptionally elevates the expression of ELAVL1, which binds to and stabilizes Beclin-1 mRNA, to promote the ferroptosis of cultured myocardial cells exposed to H/R and mouse myocardial I/R model [145]. A recent study demonstrated that HO-1 is upregulated in response to hypoxia (0.5% O 2 ) and H/R to degrade heme, thereby resulting in iron overload and ferroptosis in the endoplasmic reticulum (ER) of cardiomyocytes [146]. Moreover, ferroptosis is also induced by activating endoplasmic reticulum stress in H/R-treated H9C2 cardiomyocytes and cardiomyocytes in the myocardial IRI model [147].
In addition to the heart, ferroptosis was also found in the IRI of other organs. A study on lung IRI showed that H/R induces lung epithelial cell ferroptosis through enhancing ACSL4 expression [148]. In intestinal I/R-induced acute lung injury (ALI), HIF-1 activation by H/R increases the mRNA levels of TF, PTGS2, and ACSL4 to promote erastin-or H/R-induced ferroptosis of mouse lung epithelial (MLE)-2 cells, which can be blocked by isoliquiritin apioside or by enhancing iASPP (inhibitor of apoptosis-stimulating protein of p53)/NRF2 [149,150]. A study on IRI in the intestine also demonstrated that special protein 1 (Sp1), a crucial transcription factor, was increased in the hypoxic condition (1% O 2 ) or ischemia to promote ACSL4 expression, thus enhancing H/R-induced ferroptosis of cells and I/R-induced ferroptosis in mouse intestine [151]. H/R induces the expression of transmembrane member 16A (TMEM16A), a component of the hepatocyte Ca 2+ -activated chloride channel, which interacted with GPX4 to induce its ubiquitination and degradation, thereby enhancing the erastin-or RSL3-induced ferroptosis of hepatocytes, contributing to the I/R-induced liver injury [152]. In IRI-induced acute kidney injury (AKI), H/R induces the ferroptosis of human proximal tubular epithelial cells (HK-2) through upregulating ACSL4 and COX-2, as well as down-regulating GPX4 and FTH1 in the kidney tissues [153]. Either decreasing ACSL4 by blocking serine/arginine splicing factor 1 (SRSF1) with lncRNA TUG1 carried by urine-derived stem cells (USCs)-derived exosomes (USC-Exo), or inhibiting the activation of the c-Jun NH2-terminal kinases (JNK) pathway by inositol requiring enzyme 1 (IRE1), a proximal ER stress sensor, can inhibit the H/R-induced ferroptosis of HK-2 cells [153,154]. Additionally, ubiquitin-specific protease 11 (USP11) is elevated in neuronal cells after H/R and in the spinal cord in mice with IRI, which promotes erastin-induced autophagy-dependent ferroptosis by stabilizing Beclin-1 [155].
In most studies on IRI, the cells were treated by H/R or in organs with I/R. It is difficult to determine whether the promotion of ferroptosis resulted from hypoxia/ischemia or reoxygenation/reperfusion. As the reoxygenation process may increase oxidative stress, it is often considered the reason for H/R-induced ferroptosis. However, studies that employed both hypoxia and H/R treatment found that hypoxia alone could promote ferroptosis, which could not be further elevated by reoxygenation in some cases [146,151]. Another study focused on the reoxygenation process showed that reoxygenation did not alter Nrf2 or HIF-1α activity [156], whereas these two pathways have been reported to be involved in the H/R treatment in several studies [138][139][140][141]149,150]. These studies suggested that hypoxia/ischemia at least partially contributes to promoting ferroptosis. Therefore, we also reviewed the ferroptosis promoted by H/R in this part, although we cannot attribute all the effects on ferroptosis to hypoxia.

Other Mechanisms Involved in Ferroptosis Regulated by Hypoxia
Besides the genes that directly mediate the regulation of ferroptosis by hypoxia and HIFs, some genes that are not induced by hypoxia, modulate the ferroptosis regulated by hypoxia through different mechanisms.
For example, a study on tumor recurrence found that hypoxia during treating cancer cells with tyrosine kinase inhibitors or cisplatin, which is confirmed by CA9 elevating as a hypoxia marker, induced SCD1 expression in cancer cells, as well as FABP4 in tumor endothelial cells (TECs) and adipocytes in the tumor microenvironment (TME), to inhibit ferroptosis and promote tumor recurrence [58]. Mechanistically, SCD1 catalyzes the fatty acid desaturation to produce monounsaturated fatty acids (MUFA), leading to ferroptosis inhibition in cancer cells, while FABP4 in the TME sustains lipid droplet (LD) formation and promotes cancer cell survival under hypoxia-induced ferroptosis. Additionally, diabetes generates NOX2 in an AMPK-dependent manner to enhance oxidative stress, which led to promoting H/R-induced ferroptosis in H9C2 cardiomyocytes [157]. MiR-124-3p, which is enriched in the HO-1-modified bone marrow mesenchymal stem cells-derived exosomes, reduces STEAP3 by directly binding to its mRNA, to attenuate the H/R-induced ferroptosis of IAR20 (normal rat hepatocyte cell line) and LO2 (human fetal hepatocyte cell line) by changing iron homeostasis [158]. In acute spinal cord injury (ASCI), lncGm36569 carried by mesenchymal stem cells-derived exosomes (MSCs-exo) can act as a competitive RNA of miR-5627-5p to induce FSP1 expression, thereby attenuating hypoxia-induced ferroptosis neuronal cell and neuronal dysfunction [159].
There are also some drugs regulating hypoxia-induced ferroptosis through the mechanisms independent of hypoxia. In myocardial IRI, dexmedetomidine activates Nrf2 via AMPK/GSK-3β (AMP-activated protein kinase/Glycogen synthase kinase 3β) pathway, to inhibit H/R induced-ferroptosis of cardiomyocytes and to protect hearts from IRI [160]. It was also reported that icariin and irisin protect against myocardial and lung injury, respectively, by suppressing H/R-induced ferroptosis of cardiomyocytes and lung epithelial cells via activating Nrf2/HO-1 signaling [161,162]. Additionally, dimethyl fumarate, a therapeutic agent for relapsing-remitting multiple sclerosis, inhibits H/R-induced ferroptosis in alpha mouse liver cells and a mouse liver IRI model by activating NRF2 pathway [163]. Melatonin, an antioxidant that regulates the sleep-wake cycle, was reported to inhibit RSL3-, erastin-, or H/R-induced ferroptosis through modulating NRF2, AKT, and GPX4, in hypoxic-ischemic brain damage, as well as through upregulating NRF2 and downregulating SLC7A11 in mouse tubular epithelial cells and a mouse acute kidney injury (AKI) model [164,165]. Moreover, gastrodin, one of the main functional substances of functional food raw material Gastrodia elata BI, inhibited hypoxia-or erastin-induced ferroptosis in hippocampal neurons by activating NRF2 and GPX4, as well as inhibiting KEAP1 and COX-2 [166]. In hypoxic-ischemic brain damage, glycyrrhizin (GL), an HMGB1 inhibitor, can suppress the RSL3-or oxygen-glucose-deprivation-induced neuronal ferroptosis via the HMGB1/GPX4 pathway [167]. A recent study reported that Xingnaojing, a traditional Chinese medicine, inhibits ferroptosis in hypoxia-treated SH-SY5Y neuroblastoma cells and middle cerebral artery occlusion (MCAO)-induced cerebral ischemia rats, via upregulating HO-1, FPN, GPX4 and downregulating TFRC, DMT1, and COX-2 [168]. Additionally, lidocaine, a local anesthetic, was found to attenuate the H/R-induced ferroptosis of lung epithelial cells in lung IRI via the p38 MAPK pathway [169].
Although they are not dependent on hypoxia clearly, all these mechanisms are also involved in the regulation of ferroptosis by hypoxia.

Conclusions and Perspectives
The effect of hypoxia and HIFs on ferroptosis is highly context-dependent, especially for determining cell death or survival. According to the studies we reviewed above, hypoxia has a dual role in both inhibiting and promoting ferroptosis in a context-dependent manner. Hypoxia's effect on ferroptosis seems to mainly depend on the cell type: hypoxia usually inhibits ferroptosis in cancer cells, while in normal cells, hypoxia often induces or promotes ferroptosis. However, there is some exception: hypoxia was reported to promote ferroptosis in certain cancer cells; and in several normal cells, hypoxia was found to inhibit ferroptosis. These suggested that the regulation of ferroptosis by hypoxia may depend on the expression of other genes involved in ferroptosis, which will be differently expressed in tumor and normal cells. Studying the cancer cells and normal cells simultaneously may be helpful to understand the different regulations. However, most studies only focus on the regulation of ferroptosis by hypoxia in either tumor or normal cells. Only a few studies simultaneously compare them. Fuhrmann et al. found that hypoxia protects both primary human macrophages and HT-1080 fibrosarcoma cells from RSL3-induced ferroptosis, through increasing FTH1 and FTMT in macrophages, and through increasing only FTH1 but not FTMT in HT-1080 cells that basally express FTH and low levels of FTMT [112]. Another study by Liu and colleagues reported that the sphingolipid synthesis inhibitor myriocin stabilized HIF1α to decrease erastin-induced ferroptosis in two normal cell lines (mouse hippocampal neuronal cell HT22 and rat neural cell PC-12) and a tumor cell line (human fibrosarcoma cell HT-1080), but not in another normal cell line (human gastric epithelial cell GES-1) and tumor cell line (human hepatoma cell SK-Hep-1) [109]. However, they did not further investigate the different regulations of ferroptosis in these cell lines. As these two studies only investigated two tumor cell lines from different tumor types and compared them with the normal cells from different tissues and organs, it is hard to find out whether the difference of ferroptosis regulation by hypoxia is dependent on some genes differently expressed in tumor and normal cells.
Interestingly, hypoxia also has a dual effect on another form of regulated cell death, apoptosis, which is apparently dependent on the oxygen concentration: oxygen levels in the range of 0-0.5% (anoxia) in cells induce apoptosis, whereas oxygen levels in the range of 1-3% (hypoxia) in cells do not lead to apoptosis but protect cells for survival [170,171]. From the studies we reviewed, hypoxia-regulated ferroptosis does not appear to be related to the oxygen concentration. A potential reason for this difference is the oxidative stress induced by hypoxia. Oxidative stress always induces ferroptosis, whereas apoptosis is only induced in severe oxidative stress, and mild oxidative stress leads to cell cycle arrest and survival [172,173]. Thus, different oxygen concentrations lead to different levels of oxidative stress, resulting in the controversial effects on apoptosis, as well as similar effects on ferroptosis. It is also worth to notice that most above-mentioned studies on ferroptosis used oxygen concentration at 1-3% for the hypoxic conditions, which could induce not apoptosis but only ferroptosis. Hypoxia-regulated ferroptosis in these studies may be independent of apoptosis. Only a few studies on IRI used <1% O 2 or anoxic conditions to mimic I/R, leading to ferroptosis in normal cells, which may also induce apoptosis.
Mechanistically, hypoxia inhibits ferroptosis mainly by activating HIF-1α, whereas it promotes ferroptosis mainly by activating HIF-2α, in both normal and cancer cells. Studies also reported that HIF-1α mediates the hypoxia-induced ferroptosis in normal cells in certain conditions, and HIF-2α mediates the inhibition effect of hypoxia on ferroptosis in some cancer cells. The roles of HIF-1α and HIF-2α in ferroptosis appear to be different from previous studies, which suggests that HIF-1α plays a key role in the initial response to acute and intense hypoxia, whereas HIF-2α drives the hypoxic response during chronic and mild hypoxic exposure [91,92,174]. Although HIF-2α still mediates some effects of chronic hypoxia on ferroptosis, it is also involved in the acute effects of hypoxia. Moreover, noncanonical mechanisms modulating HIFs were also found in the regulation of ferroptosis by hypoxia. It was also reported that HIF-1α and HIF-2α have different targets: HIF-1α regulates most of the glycolytic-related genes, whereas HIF-2α selectively regulates many iron-regulatory genes [175,176]. Consistently, the abovementioned studies showed that HIF-2α inhibits ferroptosis through its direct targets involved in iron metabolism, such as FTMT, in both tumor and normal cells [102,112]. However, it is also reported that HIF-1α regulates iron metabolism directly through transcriptionally regulation of its targets, such as TFRC and DMT1 [120], as well as indirectly through some other targets, such as CA9 [103]. Fe 2+ is a cofactor for PHDs and FIH-1, which regulates HIFs under normoxia. Therefore, the regulation of iron metabolism by HIF-1α and HIF-2α seems to be very complex, and cannot be summarized here. In addition to HIFs, NRF2, the master regulator of oxidative stress signaling, also plays an important role in hypoxia-induced ferroptosis. Many studies reported that NRF mediates the regulation of ferroptosis induced by hypoxia or H/R [124,125,128,138], while another group of studies reported that targeting NRF2 could inhibit hypoxia-or H/R-induced ferroptosis [149,[160][161][162][163][164][165][166].
Hypoxia-regulated ferroptosis is involved in many diseases, such as cancer and IRI of the organs. Thus, understanding the mechanism for hypoxia to regulate ferroptosis may contribute to developing the therapy for these diseases. For instance, as hypoxia is a biomarker of solid tumors, targeting the hypoxic area is often used as a strategy to deliver drugs to tumor cells, including drugs to induce ferroptosis [177][178][179]. Since hypoxia inhibits ferroptosis in many tumor cells, the delivery of ferroptosis inducers to the hypoxic area may limit their effect on cancer cells. From understanding the mechanism involved in the inhibition of ferroptosis caused by hypoxia, avoiding the ferroptosis inhibition caused by hypoxia could be a new strategy for treating tumors with ferroptosis inducers [180,181]. Considering that ferroptosis is the major reason for IRI, many drugs have been developed to inhibit hypoxia-induced ferroptosis based on its mechanism [129,140,141,150,[160][161][162][163][164][165][166][167][168][169]. Intriguingly, hypoxia is a symptom of the Coronavirus Disease-2019 (COVID-19); HIF-1α plays an important role in the early phase of SAR-CoV-2 infection and is also associated with secondary organ damage [182,183]. On the other hand, some clinical features in COVID-19 pathobiology, such as erythropoiesis suppression and anemia, were found to result from the dysregulation of iron homeostasis and abnormal ferroptosis [183]. However, few studies reported a direct connection between hypoxia and ferroptosis in COVID-19. Further investigation of the effect and mechanism regulating ferroptosis caused by SAR-CoV-2 infection-induced hypoxia could provide more strategies for treating COVID-19 or the post-acute sequelae of SARS-CoV-2 infection through ferroptosis inhibition.
Taken together, hypoxia regulates ferroptosis through a complex network in a contextdependent manner, leading to the inhibition or promotion of ferroptosis in different cell types and conditions, which is mediated by HIFs, NRF2 signaling, and other mechanisms. The regulation of ferroptosis by hypoxia is involved in many diseases, such as cancer and the IRI of the organs. Therefore, elucidating its mechanism could be helpful to invent novel therapeutic strategies for these diseases.

Conflicts of Interest:
The authors declare no conflict of interest.