Hypoxia Enhances HIF1α Transcription Activity by Upregulating KDM4A and Mediating H3K9me3, Thus Inducing Ferroptosis Resistance in Cervical Cancer Cells

Objective Cervical cancer (CC) is a prevalent cancer in women. Hypoxia plays a critical role in CC cell ferroptosis resistance. This study explored the mechanism of hypoxia in CC cell ferroptosis resistance by regulating HIF1α/KDM4A/H3K9me3. Methods Cultured SiHa and Hela cells were exposed to CoCl2 and treated with Erastin. Cell viability was detected by MTT assay, and concentrations of iron ion, MDA and GSH were determined using corresponding kits. Expressions of KDM4A, HIF1α, TfR1, DMT1, and H3k9me3 were detected by RT-qPCR, Western blot, and ChIP assay. The correlation of KDM4A and HIF1α was analyzed on Oncomine, UALCAN, and Starbase. CC cells were co-transfected with shKDM4A or/and pcDNA3.1-HIF1α. Iron uptake and release were assessed using the isotopic tracer method. The binding relationship between HIF1α and HRE sequence was verified by dual-luciferase assay. Results Cell viability and GSH were decreased while iron concentration, MDA, KDM4A, and HIF1α levels were increased in hypoxia conditions. The 2-h hypoxia induced ferroptosis resistance. KDM4A and HIF1α were highly-expressed in CC tissues and positively correlated with each other. KDM4A knockdown attenuated cell resistance to Erastin, increased H3K9me3 level in the HIF1α promoter region, and downregulated HIF1α transcription and translation. H3K9me3 level was increased in the HIF1α promoter after hypoxia. HIF1α overexpression abrogated the function of KDM4A knockdown on ferroptosis in hypoxia conditions. Iron uptake/release and TfR1/DMT1 levels were increased after hypoxia. Hypoxia activated HRE sequence in TfR1 and DMT1 promoters. Conclusion Hypoxia upregulated KDM4A, enhanced HIF1α transcription, and activated HRE sequence in TfR1 and DMT1 promoters via H3K9me3, thus inducing ferroptosis resistance in CC cells.


Introduction
As one of the most common types of malignancies in women, cervical cancer (CC) is linked to high incidence and mortality second only to breast cancer [1,2]. China sees particularly high mortality and incidence, with about 75000 females diagnosed with CC each year, of whom 40000 die from it [3]. Females at lower ages are more susceptible to CC [4]. The public regards CC as one of the dominant public health problems and a substantial burden of disease especially in developing and underdeveloped countries due to restrained medical conditions [5]. Therefore, it is of para-mount importance to develop novel therapeutic strategies for CC.
Ferroptosis is a new type of iron-dependent programmed cell death that varies from apoptosis, necrosis, and autophagy [6]. Increased free iron, lipid peroxidation, and glutathione depletion are distinguishable characteristics of ferroptosis [7,8]. It's important to note that ferroptosis has unique functions in cancer immunotherapy [9]. For example, ATF2 increases CC cell survival by attenuating ferroptosis [10]. Ferroptosis inducers, including small molecule compounds (such as Erastin) or drugs (such as sulfasalazine, sorafenib, and artesunate) produce inhibitory effects on tumor growth by provoking cell death [11,12]. Conversely, tumor cell survival can be suppressed by hypoxia-induced factor 1α (HIF1α), a subunit of HIF-1 and a heterodimeric transcription factor, showing close association with the local suppression of anti-tumor immune responses [13][14][15]. It has been reported that HIF1α plays a key part in CC [16]. Hypoxia response element (HRE) is a significant regulatory sequence mediating hypoxic responses that interact with HIF1 in the nucleus [17]. Transferrin receptor 1 (TfR1) and divalent metal transporter 1 (DMT1) are both target genes of HIF1 [18]. Therefore, it's reasonable to assume that HIF1α might function in CC with the engagement of HRE, TfR1, and DMT1.
Lysine (K)-specific demethylase 4A (KDM4A/JMJD2A) is a member of the Jumonji domain 2 histone demethylases family [19] that catalyses histone demethylation from lysine residues, transcriptionally regulates gene expressions in tumor cells, and serves as a potential therapeutic target for cancers [20][21][22]. For example, Yan et al. have reported that JMJD2A was upregulated in human CC cells and cervical epithelial cancer tissues, which further impeded cancer cell apoptosis, and JMJD2A upregulation was closely correlated with poor overall and disease-free survival rate [23]. Nonetheless, the present understanding of KDM4A in CC as a carcinogenic protein is limited. Hypoxic areas are common birthplaces of solid tumors [24]. However, the functional mechanism of KDM4A in CC cell ferroptosis under hypoxic conditions has not been described. Histone H3 lysine 9 trimethylation (H3K9me3) is a pivotal epigenetic mechanism suppressing gene expressions [25]. Once KDM4A is depleted or deactivated, H3K9me3 accumulates at the HIF1α site, leading to HIF1α downregulation and stability decline [26]. Loss of SUV39H1 manipulates the H3K9me3 status of the DPP4 gene promoter to raise its expression, thus contributing to ferroptosis [27]. Thereby, this study outlines the functional mechanism by which KDM4A regulates CC cell ferroptosis under hypoxia by mediating HIF1α and HRE in TfR1 and DMT1 promoters through H3K9me3.

Cell
Culture. Hela cell line was provided by the Institute of Basic Medical Sciences Chinese Academy of Medical Sciences (Beijing, China) and cultured in DMEM (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS). SiHa cell line was procured from Shanghai Institutes for Biological Sciences (Shanghai, China) and maintained in DMEM high-glucose complete medium (Gibco) contain-ing 10% FBS. The cells were cultured in an incubator at constant temperature of 37°C with 5% CO 2 in the air. Upon 80% confluence, cells were detached with 0.025% trypsin (Gibco) and passaged.
2.3. Cell Treatment. The hypoxia and normoxia treatment procedures were as follows. Firstly, 100 μL SiHa and Hela cells were seeded in 96-well plates (5 × 10 4 cells/mL). According to the different culture conditions, the cells were allocated to the normoxia group (Normal) (cells cultured at the constant temperature at 37°C with 5% CO 2 ) and hypoxia group (Hypoxia) (cells cultured in the medium containing 100 μmol/L CoCl 2 in an incubator with constant temperature at 37°C with 5% CO 2 for 0.5, 1, 2, 4 and 12 h).
2.5. MTT Assay. Cell activity was detected using 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) method. Cells in each well were added with 200 μL serum-free medium and 20 μL MTT reagent (5 mg/mL), cultured at 37°C with 5% CO 2 for 4 h, and centrifuged at 1000 r/ min for 5 min to discard the supernatant. Then, cells in each well were added with 150 μL dimethyl sulfoxide (DMSO) and shaken for 10 min. The optical density at 490 nm was measured using a microplate reader. At the same time, the blank group (containing culture medium only) was set up for correction. Cell activity = experimental group -blank group/control group -blank group.
2.6. GSH and Iron Ion Concentration Determination. Glutathione (GSH) and iron ion concentration were determined using Glutathione Colorimetric detection kit (ARBOR ASSAYS, Michigan, USA) and Iron Colorimetric Assay kit (Applygen, Beijing, China) as instructed by the manufacturer's protocols.
2.7. MDA Determination. Proteins from cells or tissues were extracted using lysate, and the protein concentration was determined using the bicinchoninic acid (BCA) method. Thiobarbituric acid reaction was performed using lipid Finally, the counts per minute (cpm) were measured using a liquid scintillation counter (Perkinelmer, 1450), and the remaining 200 μL lysate was used for protein quantification using the BCA method. The iron content per μg protein was estimated. Iron release was measured as mentioned. After 3 PBS washes, cells were incubated with 1 μmol/L 55 FeCl3 (1 mL sucrose incubation solution +16 μL 62.5 μmol/L 55 FeCl3) at 37°C for 30 min, and washed with cold PBS 2 times. Then, the cells were added with 500 μL 1% SDS solution for 10 min to lyse cells, and 300 μL lysate was added into 2 mL scintillator. The cpm was measured using a liquid scintillation counter (Perkinelmer, 1450). The release of 55 Fe (%) = supernatant cpm value/(supernatant cpm value + cell lysate cpm value) ×100%.       The data were expressed as mean ± standard deviation. An independent t test was to compare data between 2 groups, and one-way analysis of variance (ANOVA) was adopted to compare data among multi-groups. Tukey's multiple comparison test was utilized as the post hoc test. The p value was obtained by a bilateral test, where p <0.05 indicated statistical significance.      8 Stem Cells International thus promoting tumorigenesis and progression [31]. However, the epigenetic mechanism of ferroptosis is still elusive. We speculated that hypoxia-induced ferroptosis resistance might be in close relation to the epigenetic mechanism. KDM4A (JMJD2A) can catalyze the histone to remove methyl from lysine residues, and KDM4A can promote the growth of CC cells and inhibit their apoptosis [23]. By searching the Oncomine database, we found that KDM4A was highly expressed in cervical squamous cell carcinoma compared with that in the normal cervical epithelial tissues (Figure 2(a), p< 0.05). Subsequently, SiHa and Hela cells were treated with hypoxia for 0.5-12 h. RT-qPCR and Western blot showed that the mRNA and protein levels of KDM4A were upregulated with the extension of hypoxia treatment (Figure 2(b)-2(c), all p <0.05).

Hypoxia Mediated KDM4A Inducing Ferroptosis
Resistance in CC Cells. We speculated that hypoxia treatment might induce ferroptosis resistance via regulating epigenetic gene KDM4A. Various shRNAs of KDM4A were used to transfect SiHa and Hela cell lines to identify the role of KDM4A in ferroptosis resistance, and the transfection efficiency was shown in Figure 3(a). shKDM4A-1 (named shKDM4A) with the highest interference efficiency was selected for subsequent experimentation. Following transfection, SiHa and Hela cell lines were treated with normoxia or hypoxia for 2 h and co-incubated with Erastin. KDM4A mRNA level was partially increased in hypoxia condition relative to normoxic condition, that is, in the Hypo-2 h + Erastin + shNC group relative to the Normal + Erastin + shNC group or in the Hypo-2 h + Erastin + shKDM4A group relative to the Normal + Erastin + shKDM4A group (Figure 3

KDM4A Mediated H3K9me3
Enhancing the Transcriptional Activity of HIF1α. H3K9me3 is an important epigenetic mechanism inhibiting gene expression and is mediated by KDM4A to regulate in various tumor cells [32]. HIF1α is a composition of HIF-1, a key transcription regulator that mediates tumor cells to accommodate to hypoxic tumor microenvironment. We speculated the same regulatory mechanism in CC cells. By searching the CC dataset on the UALCAN website, we found that HIF1α was highly expressed in cervical squamous cell carcinoma tissues compared with that in normal cervical epithelial tissues (Figure 4(a)). Analysis of ENCORI pan cancer coexpression demonstrated a positive correlation between KDM4A mRNA and HIF1α mRNA level in CC tissues (Figure 4(b)). Subsequently, SiHa and Hela cell lines were (G-I) The intracellular iron concentration, MDA and GSH levels were measured using colorimetry method. The experiments were conducted in 3 replicates independently. The data were expressed as mean ± standard deviation. The independent t test was employed for comparisons between two groups. * p < 0:05. 11 Stem Cells International iron concentration. To explore the mechanism of HIF1α in iron metabolism in CC cells, we observed the alterations of iron uptake and release in CC cells using the isotope 55 Fe tracer method. Compared with the normoxia group, the hypoxia group had significantly enhanced ability of iron uptake and weakened ability of iron release (Figure 6(a), 6(b), p <0.05). RT-qPCR and Western blot showed that elevated mRNA and protein levels of TfR1 and DMT1 in the Hypo-2 h group compared with those in the normal group ( Figure 6(c)-6(f), all p <0.05). To determine whether hypoxia conditions affected the interactions among HIF1α and DMT1 and TfR1, luciferase activity was detected to evaluate their HRE functions. The different HRE fragments of DMT1 and TfR1 were cloned into luciferase reporter vectors and co-transfected into CC cells with pcDNA3.1-HIF1α plasmid. Following 2-h hypoxia, the luciferase activities of DMT1-HRE-Luc and TfR1-HRE-Luc were significantly stimulated ( Figure 6(g)-6(h), p < 0.05).

Discussion
CC is a growing health problem around the world [33]. Ferroptosis is a newly-discovered type of programmed cell death, and CC treatment can be achieved by targeting cancer cell ferroptosis [34]. Yet hypoxia is adverse to cancer therapy [35]. The present study demonstrated that hypoxia upregulated KDM4A and downregulated H3K9me3 in the HIF1α promoter region to enhance HIF1α transcription activity and activate HRE sequence in TfR1 and DMT1 promoters, thus inducing ferroptosis resistance in CC cells.
Uncontrolled multiplication of solid tumors leads to an evident hypoxic condition status [36]. Existing research supports the idea of hypoxia being a negative influencing factor in CC treatment and reveals that hypoxia induces resistance to cisplatin in CC cells [37]. In addition to this, hypoxia also induces resistance to apoptosis [38]. Conversely, the combination of silencing GSK-3β and CDK1 counteracts the apoptosis resistance induced by hypoxia [39]. By performing the TUNEL assay to assess CC cell apoptosis after hypoxia treatment, we pointed out that hypoxia could induce apoptosis resistance in CC cells (Supplementary Figure 1). Ferroptosis is recently found as a form of programmed cell death characterized by iron-dependent lipid peroxide accumulation [40]. As reported earlier, hypoxia can induce resistance to ferroptosis in pancreatic cancer cells [41]. Ferroptosis of liver cancer cells can be impeded by hypoxia [42]. Our results unraveled that hypoxia for 0-2 h had no predominant impact on CC cell viability, Fe 2+ , MDA, and GSH levels. On the other hand, CC cells pre-treated with hypoxia for 2 h had decreased ferroptosis level in comparison to those pre-treated with 2h normoxia after Erastin treatment, indicative of enhanced ferroptosis resistance in SiHa and Hela cells by 2-h pretreatment with hypoxia.
Increased expression of KDM4A has been documented in various types of cancers [26,43], and KDM4A serves as an oncogene in CC [32]. Our result showed the KDM4A expression was gradually increased in squamous cell carcinoma of the cervix and in SiHa and Hela cells treated with hypoxia for 0.5-12 h with the prolongation of hypoxia time. More than that, KDM4A protein levels were increased in RKO cells exposed to hypoxia [32]. And our results elicited that hypoxia upregulated KDM4A in CC cells. KDM4A depletion was reported to promote cell ferroptosis in osteosarcoma [44]. Therefore, we knocked KDM4A down in CC cells and then observed that KDM4A knockdown inhibited hypoxia-induced ferroptosis resistance in CC cells. Overall, KDM4A facilitated resistance to ferroptosis in CC cells.
H3K9me3 represents a posttranscriptional modification in regulating various biological processes [45]. There is ample evidence that KDM4A regulates HIF1α expressions via H3K9me3 [32]. Our result showed that HIF1α was highly expressed in squamous cell carcinoma of cervix exposed to hypoxia, which, not coincidentally, complied with a previous study that HIF1α expression was increased in CC cells under hypoxia [46]. Moreover, our result showed a positive correlation between KDM4A mRNA expression and HIF1α mRNA expression. KDM4A and HIF1α were downregulated whereas H3K9me3 was upregulated in SiHa and Hela cells after transfection with shKDM4A. KDM4A inactivation stimulates the H3K9me3 level [47]. Besides, H3K9me3 was accumulated at the HIF1α promoter after silencing KDM4A. KDM4A demethylates H3K9me3 and H3K9me3 accumulation leads to decreased levels of HIF1α [48,49]. Collectively, KDM4A was upregulated in CC cells exposed to hypoxia and enhanced HIF1α transcriptional activity by downregulating H3K9me3.
HIF1α accumulation is implicated in the development of malignant tumors [50]. Our result showed that HIF1α levels were increased while H3K9me3 level in the HIF1α promoter region was decreased in CC cells under hypoxic conditions. Increased expression of HIF1α can enhance the hypoxic viability of glioma cells [51]. Accordingly, cell viability was increased, whereas iron concentration and MDA content were decreased and GSH level was increased in CC cells transfected with shKDM4A and pcDNA3.1-HIF1α under hypoxic conditions, suggesting that HIF1α overexpression counteracted the effect of KDM4A knockdown on suppressing ferroptosis resistance in CC cells. Our results elicited that hypoxia upregulated HIF1α expression and induced ferroptosis resistance in CC cells by upregulating KDM4A. Moreover, HIF1α binds to HRE, an important regulatory sequence mediating cell hypoxic response to activate downstream gene expressions [17]. HIF-1/HRE axis plays a principal role in hypoxia-induced iron uptake [29]. Our results showed that iron uptake and release capacities were increased in CC cells under hypoxic conditions. TfR1 and DMT1 participate in iron transportation and act as target genes of HIF1 [18]. Our result showed that TfR1 and DMT1 were increased in CC cells under hypoxic conditions. After transfection with pcDNA3.1-HIF1α, the luciferase activities of DMT1-HRE-Luc and TfR1-HRE-Luc were elevated after 2-h hypoxia treatment. Increasing evidence documented that hypoxia promotes iron uptake by regulating TfR1 and DMT1 expressions [30,52]. Altogether, our results clarified that hypoxia upregulated the expressions of iron transporter proteins by activating the HIF-1/HRE pathway. 12 Stem Cells International In conclusion, hypoxia-upregulated KMD4A stimulated HIF1α transcriptional activity and activated HRE sequence in TfR1 and DMT1 promoters, thus inducing resistance to ferroptosis in CC cells. KDM4A is perceived as a carcinogenic protein in CC, yet the specific functional mechanism remains elusive. Hypoxia is a basic and widespread stimulator engaged in various physiological and pathological pro-cesses and stimulation of cells, tissues, organs, and physiological systems in the human body [53,54]. There is no doubt that hypoxia and cancer occurrence and development are closely linked to each other [55]. Hypoxia provokes complex cellular changes, in which HIF-1 activation plays a dominant role [56]. Modern studies concerning hypoxia see HIF-1 as a critical nuclear transcription factor with 13 Stem Cells International regulatory properties in expressions of hypoxia target genes and intracellular oxygen environment [57,58]. By purifying HIF1 protein from Hela cells, Semenza et al. delineated that HIF1 is a dimer structure consisting of HIF1α and HIF1β subunits, and the former is regulated by oxygen [59,60]. Hundreds of target genes of HIF1 have been identified in every aspect of tumorigenesis and development, such as angiogenesis, lymphangiogenesis, cell invasion and metastasis, and tumor metabolism [61]. In recent years, HIF1 has also been found in the regulation of tumor stem cells [62]. Our results indicated a regulatory relationship between KDM4A and HIF-1. Further exploration into their interactions could reveal the regulatory mechanism of KDM4A in tumor cells.
Ferroptosis is a pattern of cell death discovered recently in close relation to the imbalance of cellular redox homeostasis [63,64]. An existing study has manifested that tumor cells need higher iron levels than normal cells to facilitate growth, which makes cancer cells more susceptible to ferroptosis [65]. Scholars have put forward the strategy of treating tumors with ferroptosis since its discovery. This study, with ferroptosis resistance as the breakthrough point, unveiled that HIF1α affected ferroptosis resistance in CC cells. This study also implied that hypoxia induced alterations in iron metabolism in CC cells, including elevation in iron uptake and release, yet the mechanism of changes regarding iron release requires further investigation. HIF plays a transcriptional and regulatory role in multiple aspects of tumorigenesis. However, except HIF-1, the engagement of HIF-2 in the regulatory mechanism of hypoxia in CC remains to be addressed in future studies.

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
The authors declare that they have no competing interests.