HIF2α Exerts Carcinogenesis in Clear Cell Renal Cell Carcinoma by Targeting High Expression of NUDT1 to Inhibit Oxidative Stress

Background: Emerging evidence highlights the important roles of HIF2α in the development and progression of clear cell renal cell carcinoma (ccRCC). Recently, oxidative stress has been shown to play a vital role in an increasing number of tumor types. However, the relationship between the two factors in ccRCC is still unclear. The aim of this work is to study the role of oxidative stress in ccRCC and its relationship with HIF2α. Methods: Molecular screening and bioinformatics analysis in ccRCC based on data from the TCGA database. Regulated pathways were investigated by qRT-PCR, immunoblot analysis, luciferase reporter and chromatin immunoprecipitation (ChIP) assays. A series of functional analyses were conducted in cell lines and xenograft models. Results: By screening three independent oxidative stress-related gene sets and the whole transcriptome sequencing data obtained after HIF2α knockdown, NUDT1 was discovered as a bridge molecule mediating the interrelationship between HIF2α and oxidative stress. Bioinformatics and experimental studies have found that NUDT1 is upregulated in ccRCC and has signicant prognostic implications. Mechanistically, HIF2α directly increases NUDT1 expression by binding to HIF2α response elements in the NUDT1 promoter region. Reducing the expression of NUDT1 can signicantly increase the level of oxidative stress in ccRCC cells, resulting in the inhibition of the carcinogenic effect of HIF2α. Conclusions: Our research systematically identify the regulatory mechanisms of HIF2α and oxidative stress in ccRCC for the rst time. It provides a new understanding of ccRCC and can help us creating new strategies for its treatment. by western blotting and qPCR. (e) ChIP experiment results of potential HIF2 binding sites in the NUDT1 promoter are based on the HIF2α binding sequence. (f) The results of the luciferase assay were obtained according to the materials and methods described previously. The truncation of the promoter showed that HIF2α bound to the NUDT1 promoter 1 region (-2143 to -2139), which is important for HIF2α to regulate NUDT1.


Introduction
Renal cell carcinoma (RCC) currently accounts for approximately 3% of the world's cancer diagnoses (1).
Clear cell renal cell carcinoma (ccRCC) is the most common type of renal cell carcinoma, accounting for approximately 85%. In 90% of ccRCC, the pathogenesis is characterized by constitutive activation of hypoxia-inducible factor (HIF) due to the lack of von Hippel-Lindau (VHL) tumor suppressor, which is required for oxygen (O 2 )-dependent inhibition of hypoxia-inducible factors (HIF) signaling (2). The process of HIF regulation is coordinated by two HIFα subunits (HIF1α and HIF2α), which form a complex to direct the transcriptional activation of hundreds of genes promoting adaptation to hypoxia that are associated with tumor development (3). Current research shows that HIF1α plays a role as a tumor suppressor gene, while HIF2α is one of the most important oncogenes, directly promoting the progression of ccRCC by activating its target genes (4)(5)(6).
Oxidative stress is de ned as a kind of cellular state in which the reactive oxygen species (ROS) level is increased and eclipses the antioxidant defense mechanisms in cells. Oxidative stress is characterized by an imbalance in the production of cellular oxidants and the removal of their byproducts (7,8). Regulation of intracellular ROS levels is essential for cell homeostasis, as different ROS levels can induce different biological responses. At low and medium levels, ROS can serve as a signaling molecule to maintain cell proliferation and differentiation. At high levels, ROS can affect the structure and function of critical cellular molecules such as proteins, lipids, and DNA, which in uence the growth, mutation of the cells and chromosome instability. These changes can ultimately yield neoplasm formation (9)(10)(11). Previous studies had con rmed that the HIF family is inextricably linked to oxidative stress. The expression of HIF1α in cells is associated with ROS-induced carcinogenesis in several human tumors (12). In addition, HIF2α is potentially linked to oxidative stress, which can regulate erythropoietin and the mitochondrial matrix protein superoxide dismutase 2 (SOD2) (13)(14)(15). However, the speci c mechanism and role between HIF2α and oxidative stress are still unclear.
Nucleoside diphosphate linked moiety X-type motif 1 (NUDT1) is an 18kD nudix pyrophosphatase with substrate speci city for 8-oxo-dGTP and 8-oxo-GTP (16,17). NUDT1-dependent increases have been observed in tumors with elevated oxidation levels and a tendency for DNA damage (18,19). NUDT1 is essential for RAS/ROS-related transformation and proliferation in tumorigenic cells (20). It can reduce the effect of oxidative stress on cells (especially tumor cells), and it also has a DNA maintenance function, which protects tumor cells from genomic DNA breaks and related negative impacts on viability (21)(22)(23). Furthermore, numerous studies have shown that high expression of NUDT1 in human lung tumors (24), colorectal tumors (25), esophageal squamous cell carcinomas (26) and glioblastomas (27) is associated with greater malignancy and poor prognosis. However, it has never been studied in ccRCC.
In this study, we identi ed the effects of oxidative stress on the progression of ccRCC and analyzed the role and association between HIF2α and NUDT1 during the progression of ccRCC. It provides new strategies for the treatment of ccRCC.
Human ccRCC tissue samples All samples were obtained from patients with ccRCC who underwent partial or radical nephrectomy from 2016 to 2018 at Union Hospital of Huazhong University of Science and Technology (Wuhan, China). The excised tissues were immediately frozen in liquid nitrogen for subsequent experiments. No patients who underwent surgery received any antitumor treatment before surgery. The samples were obtained with written informed consent from each patient. This study was approved by the Huazhong University of Science and Technology Committee.
Immunohistochemistry and immuno uorescence staining Four-micrometer sections prepared from para n-embedded C4-2 tissues were used for immunohistochemical staining. The sections were sequentially depara nized, rehydrated, and incubated for antigen retrieval. Blocking was performed with fetal bovine serum after the sections were incubated with 3% H 2 O 2 at room temperature for 15 min. Then, the sections were incubated with primary antibodies overnight at 4°C. Immunodetection was performed with an HRP-conjugated secondary antibody for 1 h.
DAB was used to visualize the immune complexes. Finally, nuclei were counterstained with hematoxylin.
For immuno uorescent staining, cells were xed in 4% paraformaldehyde, permeated with 0.5% Triton X-100, and blocked with 5% goat serum in PBS for 1 hour at 37°C, which was followed by incubation with primary antibodies. Finally, Alexa Fluor 488-conjugated Donkey Anti-Rabbit IgG (H+L) (Abclonal, AS035) was used as the secondary antibody to generate a visible signal. DAPI was used to stain nuclei.

RNA Extraction, cDNA Synthesis and qPCR
Total RNA was extracted from tissues with TRIzol reagent (Thermo, Massachusetts, USA) according to the instruction manual, and 1 μg of enriched tissue or cell RNA was used in reverse transcription. A NanoDrop 2000 spectrophotometer (NanoDrop Technologies, Wilmington, USA) was used to measure the purity and concentration of the RNA solution. The housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control to standardize the variability in expression levels. SYBR Green mix (Thermo, Massachusetts, USA) was used for qPCR analysis. The speci c gene primer sequences are as follows:

Transfection assay
The following HIF2α-targeted short hairpin RNAs and NUDT1-targeted short hairpin RNAs were designed and synthesized by Genechem Co. Ltd (Shanghai, China). The constructed lentivirus was used to infect A498 or 786-0 cells according to the manufacturer's protocol. The plasmids for overexpression of NUDT1, and negative controls were purchased from Genechem Co. Ltd (Shanghai, China).

Western blotting assays
For western blotting assays, the tissues and cells were lysed in RIPA protein lysis buffer (Beyotime Institute of Biotechnology, Haimen, China) containing protease inhibitor cocktail and PMSF.
Subsequently, the protein concentrations were measured using a BCA kit (Beyotime Institute of Biotechnology) according to the manufacturer's instructions. A total of 40 μg of protein was subjected to SDS-PAGE, which was then separated by gel electrophoresis and transferred to polyvinylidene uoride (PVDF) membranes. Five percent non-fat dried skim milk was used to block the PVDF membranes for 1 hour at room temperature. Then, the membranes were incubated with primary antibodies overnight at 4°C. Subsequently, the membranes were washed and incubated in blocking buffer with secondary antibodies for 2 hours at room temperature. Finally, ChemiDoc-XRs+ (Bio-Rad Laboratories, Inc., Hercules, CA, USA) was used to visualize the proteins.

Cell viability assays
For cell viability assays, A498 and 786-0 cells were transfected with sh-NC and sh-NUDT1, respectively.
Then, the cells were seeded in 96-well plates at a density of 2x10 3 /well. The proliferation rate of cells was determined using the CCK8 method according to the manufacturer's instructions. Cell viability was assessed at 0, 24, 48, 72 and 96 hours after treatment.
Colony formation assays A498 and 786-0 cells transfected with shNUDT1 or the control were plated in 6-well plates at 1000 cells per well. After two weeks, the cells were xed with methanol and then were stained with 0.05% crystal violet to enable visualization of the viable colonies (> 50 cells/colonies). The colony forming ability was evaluated from the staining results.
Wound healing assays Cells were seeded in 6-well plates. The cells were wound in a straight line by a 10 µl pipette tip through the monolayer when the cells reached 70-80% con uence. Subsequently, the cells were washed with PBS to remove impurities and then were maintained at 37°C. Images were captured at 0, 12, and 24 hours post wounding.

Transwell assays
For migration and invasion assays, cells were incubated in serum-free medium for 24 hours. Migration and invasion assays were performed using uncoated and Matrigel™-coated Transwell® inserts according to the manufacturer's instructions. Cells were seeded in the top chamber of the insert and then were allowed to invade through the Matrigel. After incubation for 24 hours, cells invading the lower surface of the membrane insert were xed with 100% methanol. Then, the cells were stained with 0.05% crystal violet, and 10 regions were randomly selected for counting.

Measurement of intracellular ROS levels
Cells were stained using a Cell Active Oxygen Detection Kit (Deep Red Fluorescence, Abcam) according to the manufacturer's instructions and then were photographed using a uorescence microscope.

Cellular MDA measurement
The method for measuring cell malondialdehyde (MDA) was based on a malondialdehyde (MDA) assay kit (Nanjing Jiancheng Biotechnology Research Institute, Nanjing, China). Intracellular MDA expression was measured using a Multi-Mode Microplate Reader (SpectramMax M5) at 532 nm.
Chromatin immunoprecipitation assay and promoter analysis (ChIP) Chromatin immunoprecipitation assay was conducted by SimpleChIP® Enzymatic Chromatin IP Kit (Agarose Beads) #9002 obtain from CST, which was performed according to the manufacturer's manual.
The plasmid with the truncated NUDT1 promoter region was constructed by Tianyi Huiyuan, China. Cell lysates were pretreated with normal rabbit IgG and protein A-agarose. An anti-NUDT1 antibody (2.0 µg) was added to the cell lysate, and they were incubated together at 4 °C overnight. IgG was used as the negative control. The speci c primer sets used to amplify the target sequence within the human NUDT1 promoter were designed as follows: Control Forward 5'-CACCATTGCTAAACCACCCA-3'

Luciferase assays
The 3' UTR of NUDT1 was constructed into RiboBio (RiboBio, Guangzhou, China), and promoter regions of NUDT1 were constructed by Tianyi Huiyuan, China. Cells were placed in 24-well plates, and complimentary DNA was transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. pRL-TK was used as an internal control. The luciferase activity was measured using a double luciferase detection reagent (Promega), and it was performed according to the instructions.

In vivo tumor implantation
A total of 2×10 6 cells were injected subcutaneously into 6-week-old male nude mice purchased from Vital River Laboratory Animal Technology Co., Ltd. To evaluate the metastatic potential of tumor cells, a nude mouse tail vein metastasis model was used. Tumor growth was measured with a digital caliper every 4 days for 7 weeks. Tumor weight was measured when mice were sacri ced on day 49 after cell implantation. Immunohistochemical staining was conducted using a standard procedure as previously described.

Bioinformatics analysis
Three independent oxidative stress pathway-related gene sets from the Oncomine database (https://www.oncomine.org) were used for screening. Gene set enrichment analysis (GSEA) was used to assess pathways of enrichment in the gene set. The mRNA levels of genes in normal kidney tissue and ccRCC tissue and relevant information (clinical stage, gender, age, survival time, etc.) of ccRCC patients were obtained from the database of TCGA-KIRC (http://www.cbioportal.org/public-porta).

Statistical analysis
Data for at least three independent experiments are expressed as the mean ± SEM. All statistical analyses were performed using t-tests or ANOVA with Excel 2016 (Microsoft) and SPSS Statistics 22.0 (IBM SPSS, Chicago, IL) software. Linear correlation analysis was used to determine the correlation between gene expression levels. Receiver operator characteristic (ROC) curve and area under the curve (AUC) were plotted to detect the best cut-off point and then obtained the highest overall accuracy rate, thereby clarifying different clinical classi cations. Pearson correlation coe cient was used to assess the correlation between two factors. p<0.05 was considered statistically signi cant.

Results
HIF2α is related to oxidative stress and its downstream molecule NUDT1 possesses the potential of novel biomarker in ccRCC In order to determine whether HIF2α is associated with oxidative stress, gene set enrichment analysis (GSEA) of HIF2α was performed in renal cancer. The results indicated that HIF2α was signi cantly enriched in the oxidative stress pathway in renal cancer ( Figure. 1a). In order to con rm the nding, stable knockdown of HIF2α by lentivirus transfection were completed in A498 and 786-0 cell lines. The western blotting result of cell lines demonstrated that the expression of acknowledged markers of oxidative stress (HO-1, CAT and SOD2) were signi cantly decreased in the cells with HIF2α knocked down ( Figure. 1b) (12,28). TCGA-KIRC database analysis results showed that the expression of HO-1 and SOD2 was positively correlated with HIF2α (Figure. S1a). Based on the results of GSEA, western blotting and heatmap visualization, we further con rmed that the biological function of HIF2α in ccRCC was related to oxidative stress pathways. To explore genes downstream of HIF2α that affect oxidative stress, bioinformatics analysis was conducted using the sequencing results of HIF-2a knockdown in renal cancer and oxidative stress pathway-related ccRCC gene datasets (Beroukhim, JONE, GUMZ), which involved the target genes of oxidative stress-related pathway in renal cancer. The overlap of the above gene sets indicated that NUDT1 and SOD2 were dysregulated in ccRCC ( Figure. 1c). Furthermore, the analysis of the database from TCGA demonstrated that both NUDT1 and SOD2 were overexpressed in ccRCC ( Figure. 1d) (Figure. S1b). The Kaplan-Meier survival analysis showed that ccRCC patients with higher expressions of NUDT1 or SOD2 had a poorer survival while the NUDT1 expression possess more signi cant differences than SOD2. (Figure. 1e). Moreover, ROC curve analysis also showed that NUDT1 had a larger AUC than SOD2, which meant that there was a better diagnostic value for NUDT1( Figure. 1f). In summary, NUDT1 was the focus of subsequent research.
To con rm the importance of NUDT1 in ccRCC, further bioinformatics analysis strategies were adopted. First, the high expression of NUDT1 in ccRCC was con rmed by using four independent datasets from the Oncomine database (Figure. S1c). At the same time, the expression of NUDT1 was strongly correlated with the clinicopathological parameters of ccRCC (Table 1), and its expression gradually increased with the increase of tumor stage and grade (Figure. S1d). Additionally, analysis of Kaplan-Meier curves by subgroup both for overall and disease-free survival showed that high expression of NUDT1 was highly correlated with poor prognosis in patients with ccRCC (Figured. S2-S3). Univariate and multivariate analyses indicated that NUDT1 is a potential independent prognostic marker for ccRCC ( Table 2-3). In addition, ccRCC tissues were analyzed to assess the protein and mRNA levels of NUDT1. The results indicated that the levels of NUDT1 protein and mRNA detected in ccRCC tissues were signi cantly higher than those in normal tissues ( Figure. 1g-i). Compared with the control group, NUDT1 was also highly expressed at the protein and mRNA levels in a ccRCC cell line ( Figure. 1j). In general, NUDT1 is highly expressed in ccRCC and associates with poor prognosis of ccRCC.    The GSEA analysis focusing on NUDT1 expression in ccRCC was performed and the results showed that NUDT1 was highly related to renal cell carcinoma, cell metabolism, and energy production ( Figure. Figure. 3b-d). After mapping the transcripts to a gene ontology (GO) database, we found that NUDT1 played a key role in cell development and antioxidant activity pathways ( Figure. 3e). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed to further analyze the effect of NUDT1 in ccRCC. The results showed that the differentially expressed genes were mainly enriched in basic processes of cell activity, which indicated that NUDTI played a vitral role in ccRCC cells (Fig. 3f-g).
Therefore, we conclude that NUDT1 is highly correlated with oxidative stress and involved in key biological processes in ccRhCC.

Nudt1 Inhibits Oxidative Stress In Ccrcc
CcRCC is a special tumor type that exhibits a signi cant change in cellular redox balance (29). As we discussed above, the results indicated the functions of NUDT1 in ccRCC cells, but the mechanism of NUDT1 in ccRCC remains unclear. To explore the relationship between NUDT1 and oxidative stress in ccRCC, bioinformatics analyses were performed. We analyzed the database from TCGA-KIRC by GSEA.
The results indicated that NUDT1 participated in mitochondrial formation and division in ccRCC, further demonstrating the effect of NUDT1 on cellular oxidative stress (Fig. 4a). To prove the connection between NUDT1 and oxidative stress, we generated a correlation heatmap and linear correlation curves between NUDT1 and the most critical ROS level-related molecules HO-1 and SOD2(30, 31) based on data from the database from TCGA-KIRC. The results showed that NUDT1 was signi cantly positively correlated with molecules related to oxidative stress ( Figure. 4b-c). Subsequently, we detected the expression of oxidative stress markers HO-1, CAT, and SOD2 in ccRCC cell lines with NUDT1 stable knock down or overexpression. The results indicated that HO-1, CAT, and SOD2 protein levels were signi cantly downregulated in stably silenced NUDT1 cell lines, and these protein levels were signi cantly upregulated in cell lines stably overexpressing NUDT1 ( Figure.  HIF2α regulates NUDT1 expression by directly binding to its promoter region Above, we veri ed that NUDT1 is an important downstream target of HIF2α. Based on the premise that HIF2α is an important oncogenic gene of ccRCC, we speculate that NUDT1 be directly regulated by HIF2α transcription. We analyzed HIF2α-related pathways by GSEA. The results showed that HIF2α was involved in the process of cellular oxidative stress ( Figure. 1a) and nucleoside metabolism ( Figure. 6a), which is consistent with the function of NUDT1 in cells. Therefore, the speci c mechanism of HIF2α regulating NUDT1 has become the focus of our attention. HIF2α, as one of the most important oncogenes in ccRCC, predominantly acts as a transcription factor that directly promotes the expression of downstream genes.
Transcriptome sequencing after HIF2α knockdown, western blotting and qPCR showed that the expression of NUDT1 at the protein and RNA levels decreased with HIF2α silencing, indicating that there is a positive regulatory relationship between HIF2α and NUDT1 ( Figure. 6b-d). Therefore, we rst explored the regulatory mechanism of HIF2α on NUDT1 from the perspective of direct transcription. Based on the HIF2α binding sequence (32), we identi ed three potential binding sites in the 3000 bp region upstream of the NUDT1 promoter: sites 1, 2 and 3 ( Figure. S5). Chromatin immunoprecipitation (ChIP) assays were used to verify the binding of HIF2α to these potential sites. We found that HIF2α binds to all predicted potential hypoxic response elements (HREs) of NUDT1 in A498 and 786-0 cell lines ( Figure. 6e, Figure. S5). A luciferase reporter gene assay was introduced to further illustrate the speci c roles of the sites. The truncated plasmids were constructed based on the binding sites of the NUDT1 promoter region HIF2α as site1, site2 and site3, respectively. We found that the reduction of HIF2α silencing on luciferase activity was signi cantly reversed after site 1 was excised, while removal of sites 2 or 3 had no signi cant signi cance, which suggested that site 1 was the main site of HIF2α regulation of NUDT1 ( Figure. 6f).
Finally, we conclude that HIF2α acts as a transcription factor to directly increase NUDT1 expression by binding to HIF2α response elements in the NUDT1 promoter region.

Knockdown Of Nudt1 Suppresses Ccrcc Progression In Vivo
Encouraged by the in vitro results, in vivo experiments based on NUDT1 were carried out. A498 cells with NUDT1 stably silenced (shNUDT1-1 and shNUDT1-2) were injected subcutaneously into the axilla of immune-de cient mice to generate a xenograft tumor model. The mice were observed for 7 weeks, and the tumor sizes were measured every week. Consistent with the in vitro experiments, knockdown of NUDT1 inhibited tumor growth based on tumor weight and size ( Figure. 7a-c). Meanwhile, an immune de ciency mouse tail vein metastasis model was employed to assess the metastatic ability of the tumor cells. After 7 weeks of observation, we observed that knockdown of NUDT1 signi cantly reduced ccRCC liver metastasis (Figure. 7d). Therefore, to further evaluate the effect of NUDT1 on ccRCC metastasis, uorescence images were obtained of living mice. The results showed that the whole-body uorescence intensity of mice was signi cantly reduced after NUDT1 silencing, which indicated that silencing NUDT1 could signi cantly reduce ccRCC metastasis ( Figure. 7e).
In vitro experiments con rmed that silencing NUDT1 could signi cantly promote oxidative stress in ccRCC. Therefore, we also evaluated changes in oxidative stress levels in vivo. Through immunohistochemistry (IHC), we found that after silencing NUDT1, the oxidative stress markers HO-1, CAT, and SOD2 were signi cantly reduced, and Ki67, a measure of tumor malignancy, was also signi cantly reduced (Figure. 7f). The above results demonstrated that silencing NUDT1 in vivo can also signi cantly promoted oxidative stress and inhibited the renal tumor progression.
Based on this evidence, we propose a model in which HIF2α acts as a transcription factor to directly increases NUDT1 expression by binding to the HIF2α response element in the NUDT1 promoter. NUDT1 can reduce the level of oxidative stress and balance the mitochondrial metabolism in tumor cells, thereby promoting ccRCC progression. When HIF2α is knocked down, it can target the reduction of NUDT1 expression, leading to increased levels of ROS and oxidative stress in tumor cells, thereby inhibiting the occurrence and progression of ccRCC ( Figure. 7g).

Discussion
Recently, an increasing number of studies have shown that oxidative stress plays an important role in ccRCC. Widespread activation of HIF2α is an important feature of ccRCC. Relevant literature has con rmed that HIF2α affects the progression of oxidative stress (33), but the speci c mechanism linking HIF2α and oxidative stress has not yet been clari ed. In this study, we identi ed a novel pathway by which HIF2α regulates oxidative stress levels in tumor cells through NUDT1. Mechanistic investigations have shown that a large amount of ROS is produced in highly metabolized malignant tumors, leading to the destruction of cell structure and inhibiting tumor progression (34). HIF2α can directly regulate the expression of NUDT1 to eliminate the effect of ROS on tumor cells, leading to the progression of ccRCC.
Oxidative stress is characterized as an imbalance between the production of cellular oxidants and the process of removing their byproducts (8). Reactive oxygen species (ROS) can be used as indicators of oxidative stress, which is mainly produced in the mitochondrial electron transport chain during cell metabolism and plays a vital role in normal cell signaling pathways such as proliferation and apoptosis (35,36). However, excessive ROS levels can cause structural damage in cells (34). Highly metabolically active cancer cells can actually produce more ROS than normal cells, and these metabolically active cells consequently showed more evidence of DNA damage and buffer system engagement (7). Therefore, ccRCC needs some "means" of reducing ROS damage while promoting cell proliferation and survival (37).
Increased levels of ROS in tumor cells result in the accumulation of high levels of 8-oxo-dGTP in their nucleotide pools (38)(39)(40). NUDT1 can eliminate excessive 8-oxo-dGTP to avoid oxidative damage of tumor cell nucleic acid (41). Recent studies have also shown that NUDT1 can reduce the level of ROS induced by oncogenic RAS, and plays an important role in oncogenic RAS-mediated transformation and proliferation (20,42,43). Overexpression of NUDT1 may have several protective functions for cancer cells by hydrolyzing 8-oxo-dGTP or reducing ROS levels. Based on the protective properties of NUDT1 against oxidative stress in cancer cells, NUDT1 inhibitors have been developed as potential anticancer drugs (38,44). HIF2α is a common transcription factor, and its role of HIF2α in ccRCC has been con rmed. The relationship between HIF2α and oxidative stress has been mentioned in the current research (45,46), but its mechanistic link is still unclear. In this study, we not only revealed a new pathway by which HIF2α can regulate oxidative stress but also demonstrated a new mechanism by which it can regulate oxidative stress in ccRCC. Typically, the activation of HIF2α in ccRCC cells directly promotes the expression of NUDT1, thereby reducing the level of oxidative stress in the cells to eliminate ROS effects. In summary, this promotion process may be one of the important mechanisms by which HIF2α eliminates the effects of ROS and promotes the ccRCC process.
As an important oncogene in ccRCC, HIF2α has been widely studied and treated in ccRCC. Currently, most of the research has focused on the role of HIF2α in angiogenesis in ccRCC. The rst-line drug sunitinib that is used to treat ccRCC also targets downstream molecules of HIF2α, speci cally vascular endothelial growth factor receptor (VEGFR) and platelet-derived growth factor receptors (PDGFR) (47,48). However, 10-20% of advanced RCC patients are inherently unresponsive to sunitinib treatment, and most of the remaining patients eventually develop resistance and tumor progression after 6-15 months of therapy (49). Thus, it is urgent to further clarify the mechanism of HIF2α cancer promotion and nd new strategies for the treatment of ccRCC. According to recent research, oxidative stress plays a key role in tumorigenesis and development. A large amount of evidence proves that the pro-cancer effect of HIF2α on ccRCC is not only related to angiogenesis but that oxidative stress is also a factor that cannot be ignored (33). Nevertheless, there have been few reports that have assessed targeted oxidative stress therapy for ccRCC. This study con rmed that NUDT1 is a key molecule by which HIF2α regulates oxidative stress in ccRCC. Based on these results, targeted NUDT1 therapy may play an important role in oxidative stress in ccRCC. Therefore, we propose a new drug combination strategy: a combination therapy of anti-angiogenesis and NUDT1 targeting inhibitors.

Conclusion
Through comprehensive analysis, this study identi ed a new pathway by which HIF2α can directly regulate NUDT1 expression in ccRCC. NUDT1 can eliminate the killing effect of oxidative stress on ccRCC by reducing ROS. In addition, this study provides new insights into the role of HIF2α in regulating the ccRCC process. These ndings provide us with a strategy for overcoming cellular oxidative stress defense in the treatment of ccRCC and present certain guiding signi cance for the development of new drugs and combined drug treatment strategies for ccRCC therapy.

Availability of Data and Materials
The analyzed datasets generated during the current study are available from the corresponding author on reasonable request.   of the transwell assay of the migration and invasion of NUDT1-overexpressing cell lines. **** P <0.0001, *** P <0.001, ** P <0.01, and * P <0.05.  knockdown and overexpressing cell lines was measured. Experiments were repeated 3 times, as described in the Materials and Methods section. **** P <0.0001, *** P <0.001, ** P <0.01, and * P <0.05.   promoter are based on the HIF2α binding sequence. (f) The results of the luciferase assay were obtained according to the materials and methods described previously. The truncation of the promoter showed that HIF2α bound to the NUDT1 promoter 1 region (-2143 to -2139), which is important for HIF2α to regulate NUDT1.

Figure 7
Knockdown of NUDT1 suppresses ccRCC progression in vivo. (a-b) A498 cells and control cells transfected with shNUDT1-1 and shNUDT1-2 were injected subcutaneously into nude mice in the three groups. The tumor size and weight of mice in each group was measured after the seventh week. Data are expressed as the mean ± SEM from tumors of each group. **** P <0.0001, *** P <0.001, ** P <0.01, and * P <0.05 (c) The tumor volume of each group was measured every week. This graph is drawn based on the relationship between the number of weeks after tumor cell implantation and tumor size (mm3). Data are expressed as the mean ± SEM from tumors of each group. **** P <0.0001, *** P <0.001, ** P <0.01, and * P <0.05. (d) H&E staining of liver tissue in the NUDT1 knockdown group and control group.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.