Long noncoding RNA MIR210HG is induced by hypoxia-inducible factor 1α and promotes cervical cancer progression.

Increasing evidence has indicated that long noncoding RNAs (lncRNAs) play essential roles in various types of cancer, especially the ability of tumor cells to adapt to hypoxia conditions. However, only a few of them have been experimentally validated in cervical squamous cell carcinoma (CSCC). In the current study, we identified a hypoxia-induced lncRNA MIR210HG was excessively expressed in CSCC tissues and regulated by human papillomavirus (HPV) type 16 E6 and E7 via hypoxia-inducible factor 1α (HIF-1α). Functional assays revealed the role of MIR210HG in promoting proliferation, migration and invasion of CSCC cells in vitro under normoxia as well as hypoxia conditions. Meanwhile, stable MIR210HG silencing dramatically repressed tumor growth and pulmonary metastasis in vivo. Mechanistically, the depletion of MIR210HG or HIF-1α decreased each other's expression level, while silencing MIR210HG or HIF-1α respectively downregulated the expression levels of phosphoglycerate kinase 1 (PGK1), one of key metabolic enzymes in the glycolysis pathway. Furthermore, decreased expression of PGK1 by HIF-1α knockdown was reversed through the overexpression of MIR210HG. Also, we demonstrated HIF-1α can activate the transcription of MIR210HG via binding its promoter. Taken together, these results expand our understanding of the cancer-associated functions of hypoxia-induced lncRNAs, and highlight MIR210HG forms a feedback loop with HIF-1α contributing to cervical carcinogenesis, with potential implications for therapeutic targeting.


Introduction
Cervical cancer is one of the main causes of death in females. Although screening and vaccination programs have been expanded, the number of new cases of cervical cancer has continued to increase, which means that cervical cancer is still a major public health concern [1]. Mortality rates in low-income developing countries and regions are vastly different from those in developed countries, with an 18-fold difference in mortality and 85% of deaths occurring in underdeveloped countries due to limited treatment options and economic and cultural factors [2]. At present, the conventional treatment of cervical cancer includes radiotherapy, chemotherapy and surgery, but patients at advanced stages are prone to developing radiotherapy and chemotherapy resistance [3]. Therefore, it is urgently necessary to identify new therapeutic targets for cervical cancer.
Hypoxia is a hallmark of the solid tumor microenvironment. It induces genomic instability, which in turn helps cancer cells respond adaptively to meet the needs of carcinogenesis, cancer progression and relapse [4]. Considering that hundreds of protein-coding genes are transactivated by HIF-1 under hypoxic conditions, whether lncRNAs could be responsible for hypoxia and their regulatory functions are far from clear [5]. Choudhry et al. [6] performed the integrated genomic analyses of both non-coding and coding transcripts in hypoxic cells, and indicated that noncoding RNA was also hypoxia-inducible. Since then, dysregulated lncRNAs targeted by HIF-1 in several types of cancers have been investigated [7,8].
However, there are still some unknown lncRNAs which may be the potential molecules in response to hypoxia and their functional mechanism in tumorigenesis need to be further elucidated.
Interestingly, increasing evidence suggests that lncRNAs are closely implicated in regulating HIF-1α activity. For example, Xiang et al. [9] proposed that c-Myc-mediated repression of lncRNA IDH1-AS1 sustains activation of the Warburg effect by HIF-1α under normoxic conditions. LncRNA FEZF1-AS1 promotes pancreatic cancer cell proliferation and invasion, through miR-142/HIF-1α axis under hypoxic condition while through miR-133a/EGFR axis under normoxic condition [10]. Furthermore, Hua et al. [11] found that lncRNA AC020978 promote the nuclear translocation of PKM2 and regulate PKM2-enhanced HIF-1α transcription activity. However, few studies have explored the link between hypoxia and MIR210HG in CSCC.
In the current study, we identi ed a hypoxia-inducible lncRNA, MIR210HG, which contributes to cervical cancer progression. Further study demonstrated that HPV16 E6 and E7 could regulate the expression level of MIR210HG by modulating transcription factor HIF-1α. Meanwhile, we found that HIF-1α knockdown could inhibit the transcription of MIR210HG and vice versa, silencing HIF-1α or MIR210HG could respectively downregulate PGK1 expression. In addition, HIF-1α can activate the transcription of MIR210HG via binding its promoter. Taken together, these ndings indicate the functions of MIR210HG in CSCC progression and uncover the positive feedback loop of HIF-1α/MIR210HG which subsequently regulate the expression of PGK1 in CSCC, and this axis may serve as a potential target for cancer therapy.

Tissue specimens
Total of twenty-one CSCC and eighteen normal cervical epithelial (NCE) tissues used in this study were collected at Nanfang Hospital from September 2019 to September 2020. Informed consent was obtained from all patients. No patients had been treated with radiotherapy or chemotherapy before surgery and all specimens preserved in sample protector for RNA/DNA (Takara, Japan) at −80°C. This study was approved by the Ethics Committee of Southern Medical University.

Cell culture
The human cervical squamous cancer cell lines CaSki, SiHa and HEK293T cells were purchased from the Cell Bank of Chinese Academy of Sciences (Shanghai, China). All the cell lines were grown in Dulbecco's modi ed Eagle's medium (DMEM; 4.5 g/L D-glucose) supplemented with 10% (v/v) fetal bovine serum (FBS) and incubated at 37°C in the humidi ed incubator with 5% CO 2 .

RNA isolation, cDNA synthesis and qRT-PCR
Total RNA was extracted from the cells or tissues using RNAiso Plus reagent (Takara) following the manufacture's protocol. The concentration and purity of extracted RNA were assessed by NanoDrop 2000 (Thermo). The reverse transcription reactions were performed by using a PrimeScript RT reagent kit (RR047A, TaKaRa). qRT-PCRs were performed with a SYBR Premix Ex Taq kit (RR420A, TaKaRa) on a ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA). Gene expression levels relative to β-actin were calculated by 2 −ΔΔCT method. All the primer sequences used in this study are listed in Table S1.

CCK-8 proliferation assay
After transfection, CaSki and SiHa cells in the logarithmic growth phase were seeded into 96-well plates at a density of 3 × 10 3 cells per well and incubated for 1.5 h at 37 °C and 5% CO 2 before the optical density (OD) at 450 nm was detected. Proliferation was measured by the Cell Counting Kit-8 (CCK-8, Dojindo Molecular Technologies, Kumamoto, Japan) every 24 h after transfection for 4 days.

Transwell assay
For migration and invasion assays, transwell assay was conducted using 24-well transwell chamber precoated with or without Matrigel (BD). Treated CaSki and SiHa cells (4 × 10 4 /well) in serum-free DMEM medium were added to the upper chamber and 20% fetal bovine serum medium was added in the lower chamber. Cells were cultured at 37 °C with 5% CO 2 for 24 h and then those on the lower surface were xed with methanol and stained with 0.1% crystal violet. Cells were counted in ve randomly selected areas under microscope eld.

Western blot
Total cell lysates were prepared using immunoprecipitation assay lysis buffer containing protease and phosphorylase inhibitors and clari ed by centrifugation (12 000 g for 15 min at 4°C). The protein concentration was detected by a BCA Protein Assay Kit (Beyotime). Proteins were mixed with loading buffer, then cell lysates were separated by sodium dodecyl sulphate SDS-PAGE and transferred onto polyvinylidene di uoride membranes, which were subsequently blocked in 5% skimmed milk for 2 h. Afterwards, primary antibodies against E6 (ab70, Abcam), E7 (ab20191, Abcam), HIF-1α (36169S, CST), PGK1 (17811-1-AP, Proteintech) were incubated overnight at 4°C. The next day, the membrane was incubated with second antibody at room temperature for 2 h, washed with TBST and then developed with the ECL system and normalized to the gray value of β-actin.

Isolation of nuclear and cytoplasmic RNA
Cells were partitioned into nuclear and cytoplasmic fractions using the PARIS™ Kit (Invitrogen) according to the manufacturer's instructions. RNA isolated from each of the fractions was analyzed by qRT-PCR to determine the expression levels of MIR210HG, nuclear control transcript (MALAT1) and cytoplasmic control transcript (GAPDH).

Luciferase reporter assay
The length of MIR210HG promoter region containing two kilobase (kb) was constructed into pGL3-based vectors. To determine the effect of HIF-1α on MIR210HG promoter, pGL3-based construct containing MIR210HG promoter sequences and renilla luciferase reporter plasmid were cotransfected into HEK293T cells with or without HIF-1α knockdown. Cultured under normoxia or 250 µM CoCl 2 24 h after transfection, re y and renilla luciferase activity were measured by dual-luciferase reporter assay kits (Beyotime). The ratio of re y luciferase to renilla activity was calculated for each sample.

Xenografted tumor and pulmonary metastasis model in vivo
All animal experiments were approved by the Institutional Committee on Animal Care and Use of Southern Medical University. The female nude mice between 4 and 5 weeks were purchased from Beijing HFK Bioscience Co., Ltd. SiHa cells transfected with the negative control vector or the sh-MIR210HG vector were paired, 1 × 10 7 cells were inoculated subcutaneously into either side of ank of the same female nude mouse and 1 × 10 6 cells were injected intravenously into tail. The tumor volumes were measured every week and calculated as Length × Width 2 × 0.5. After 5 weeks, the mice were euthanized after injection and the tumors were xed for hematoxylin-eosin (HE) and immunohistochemistry (IHC) staining.

Statistical analysis
For the cell functional analyses, results are presented as mean ± standard deviation (SD). The comparison of means between two groups was conducted using Student's t test, while one-way analysis of variance was used for multiple comparisons. Correlation was calculated according to Pearson correlation. P < 0.05 is considered signi cant. All the experiments were repeated at least three times.

MIR210HG was a hypoxia-inducible and aberrantly upregulated lncRNA in CSCC
To uncover the function of noncoding transcripts under physiological conditions such as hypoxia, we explored the gene expression pro les of cervical cancer cell in response to hypoxic pathway inhibition from Gene Expression Omnibus (GEO) database (GSE120675). The results provided the expected hypoxic coding signature and four overlapping lncRNAs ( Figure 1A). Among them, MIR210HG was one of notably overexpressed lncRNAs in HPV-positive CSCC compared to HPV-negative normal controls from our previous studies (unpublished data). qRT-PCR analysis found that MIR210HG was up-regulated in 21 CSCC tissues as compared to 18 normal tissues ( Figure 1B). In the TCGA database, MIR210HG was highly expressed in the cervical cancer tissues ( Figure 1C). The Kaplan-Meier survival analysis revealed that patients with high MIR210HG expression levels had a shorter overall survival time ( Figure 1D). Cobalt (II) chloride (CoCl 2 ), as a prolyl hydroxylase (PHD) inhibitor, is a commonly used hypoxia mimetic agent in establishing experimental hypoxia. Interestingly, MIR210HG was signi cantly upregulated exposed to hypoxia mimetic CoCl 2 in a dose-and duration-dependent manner by qRT-PCR analysis ( Figure 1E and   1F). In summary, we identi ed MIR210HG as a hypoxia inducible lncRNA related with CSCC.

MIR210HG promotes the proliferation, migration and invasion of CSCC cells in vitro
To explore the biological function of MIR210HG in CSCC cells, we performed CCK-8 and transwell assays to evaluate the effect of MIR210HG on the proliferation, migration and invasion of CSCC cells. siRNA or MIR210HG overexpression pcDNA3.1 vector were used to establish loss-of-function and gain-of-function cell models in CaSki and SiHa cells. The knockdown and over-expression e ciency of MIR210HG in CSCC cells was checked by qRT-PCR assay (Figure 2A and 3A). As a result, MIR210HG knockdown signi cantly suppressed CSCC cell proliferation ( Figure 2B and 2C), whereas ectopic overexpression of MIR210HG promoted the proliferation of CSCC cell ( Figure 3B and 3C) under normoxia or treatment of CoCl 2 . We also performed transwell assays under normoxic and hypoxic conditions. Consistently, downregulation of MIR210HG signi cantly inhibited the migration and invasion capacities of CSCC cells under hypoxic conditions ( Figure 2D). On the contrary, upregulation of MIR210HG remarkably enhanced the migration and invasion capacities of CSCC cells ( Figure 3D). Taken together, these results strongly suggested that MIR210HG promotes the proliferation, migration and invasion of CSCC cells.

MIR210HG could be modulated by HPV16 E6/E7 through HIF-1α
Our previous data indicated that MIR210HG was one of remarkably overexpressed lncRNAs in HPVpositive CSCC compared to HPV-negative normal cervical epithelial tissues, we next conducted qRT-PCR analysis to determine whether HPV16 E6 and E7 could affect the expression of MIR210HG. As expected, MIR210HG was downregulated after HPV16 E6 or E7 silencing, while upregulated after HPV16 E6 or E7 overexpression ( Figure 4A and 4B). Considering that HPV16 E6 and E7 oncoproteins have been demonstrated to induce the expression of HIF-1α in CC cells [12] and MIR210HG was a hypoxia-inducible lncRNA, we speculated whether E6 and E7 could regulate MIR210HG via HIF-1α. Further studies revealed that HIF-1α protein level was suppressed by E6 or E7 repression ( Figure 4C), and MIR210HG upregulation by HPV16 E6 or E7 overexpression could be abolished by HIF-1α depletion ( Figure 4D). Consequently, these ndings indicated HPV16 E6 and E7 might enhance the expression of MIR210HG partially by modulating HIF-1α.

MIR210HG forms a positive feedback loop with HIF-1α to promote the expression of PGK1
To elucidate the potential underlying mechanisms of MIR210HG in tumorigenesis, Gene Set Enrichment analysis (GSEA) was performed in cervical cancer cohort from TCGA database. Surprisingly, GSEA analysis revealed that high MIR210HG expression was positively correlated with "GLYCOLYSIS" and "HYPOXIA" pathway ( Figure 5A), and we con rmed that MIR210HG knockdown dramatically inhibited mRNA expression level of HIF-1α and its downstream gene PGK1 under normoxia or treatment of CoCl 2 ( Figure 5B). However, HIF-1α reduction could in turn diminish the RNA level of MIR210HG and PGK1, suggesting a positive feedback loop between MIR210HG and HIF-1α ( Figure 5C). Subsequently, we observed that depletion of MIR210HG expression could notably decrease HIF-1α and PGK1 protein level, which was in line with the qRT-PCR results ( Figure 5D). Moreover, the expression of HIF-1α in CaSki and SiHa cells was signi cantly elevated in accordance with upregulated PGK1 expression during hypoxia condition ( Figure 5E). qPCR analysis of nuclear and cytoplasmic lncRNA showed that MIR210HG was mainly detected in the nucleus of CaSki and SiHa cells, indicating that MIR210HG might exert its functions in nucleus ( Figure 5F). To further validate the ability of HIF-1α to transactivate MIR210HG, the promoter region of MIR210HG were inserted into re y luciferase reporter plasmid. Luciferase assay showed that HIF-1α knockdown signi cantly suppressed the luciferase activity in both normoxia and chemical hypoxia conditions. Additionally, the expression of MIR210HG is positively correlated with HIF-1α in CSCC tissues ( Figure 5H). Collectively, our results strongly suggested that the existence positive feedback loop between MIR210HG and HIF-1α, and HIF-1α activates the transcription of MIR210HG via binding its promoter.

MIR210HG promotes CSCC cells migration and invasion via PGK1
Considering that MIR210HG forms a positive feedback loop with HIF-1α to affect expression of PGK1, we next examined whether MIR210HG or HIF-1α synergistically regulate the expression of PGK1. The results showed that co-transfection of si-HIF1α and OE-MIR210HG greatly rescued the mRNA and protein expression of PGK1 in CSCC cells caused by only knockdown of HIF-1α under normoxia or chemical hypoxia conditions ( Figure 6A-6C). Then, further experiments were carried out to verify whether regulation work of MIR210HG on cell migration and invasion was related with PGK1. The e ciency of silencing PGK1 was validated by western blot ( Figure 6D). As expected, the results of transwell assay indicated that suppression of PGK1 apparently abrogated the enhanced effect of MIR210HG upregulation on cell migration and invasion ( Figure 6E). Taken together, PGK1 was a crucial downstream target of MIR210HG in the regulation of migration and invasion in CSCC cells.

Silencing of MIR210HG impairs xenograft tumor growth and pulmonary metastasis in vivo
To further investigate the oncogenic role of MIR210HG, SiHa cells stably transfected with sh-MIR210HG or NC were used to establish xenograft model and pulmonary metastasis model in vivo. The infection e cacy was con rmed in Figure 7A. After cell injection for 5 weeks, tumor volume and weight were signi cantly reduced in sh-MIR210HG group in contrast to that in sh-NC group ( Figure 7B and 7C). Meanwhile, compared with sh-NC group, sh-MIR210HG transformed cells had the milder capacity for pulmonary metastasis (Figure 7E and 7F). Immunohistochemical analysis further revealed a signi cant downregulation of Ki-67, HIF-1α and PGK1 in tumor tissues of sh-MIR210HG group compared with that in sh-NC group ( Figure 7D). These results supported the function of MIR210HG in promoting proliferation and metastatic capacity in vivo.

Discussion
Although the complex biological functions of lncRNAs are still largely unknown, accumulating evidence has shown that lncRNAs contribute to the initiation and development of various cancers by acting as oncogenic or tumor suppressive regulators [13]. MIR210HG has been reported to play oncogenic functions in cervical cancer [14], non-small cell lung cancer (NSCLC) [15], hepatocellular carcinoma [16] and so on. Majority of these researches revealed that MIR210HG could act as a competing endogenous RNA (ceRNA) in tumorigenesis. For example, Wang et al. [14] demonstrated that it promoted CC progression through regulating the miR-503-5p/TRAF4 axis. However, the role and mechanism of MIR210HG in cervical cancer have not fully been elucidated.
Adaptation of cancer cells to a hypoxic tumor microenvironment is of great importance for their malignant growth and distant metastasis [17]. Based on bioinformatics analysis using GEO database, four hypoxia-related lncRNAs (MIR210HG, HLA-DQB1, DARS-AS1 and UPK1A-AS1) in cervical cancer were screened and DARS-AS1 has been identi ed as a hypoxia-induced lncRNA [18]. Interestingly, MIR210HG was remarkably upregulated lncRNA in HPV-positive CSCC tissues, when compared to HPV-negative normal cervical epithelial tissues from our previous microarray data. Further investigation showed that MIR210HG was signi cantly upregulated in samples derived from cells subjected to dose-and timedependent hypoxia treatment. And gain-and loss-of-function studies strongly suggested that MIR210HG could play a stimulative role in the tumor migration, invasion and metastasis. Recent studies also indicated the oncogenic function of MIR210HG in human cancers, which was consistent with our ndings. For instance, MIR210HG was found to be highly expressed in NSCLC which promoted proliferation and migration of NSCLC cells by inhibiting CACNA2D2 via binding to DNMT1 [19]. Similarly, Bu et al. [20] suggested that MIR210HG is associated with NSCLC cell progression through regulating the miR-874/STAT3 axis. Meanwhile, Li et al. [21] reported that MIR210HG sponge miR-1226-3p to facilitate the invasion and metastasis of breast cancer cells by regulating mucin-1c and EMT pathway. Thus, our results provided important evidence that MIR210HG was highly expressed in CC and may be used as a potential biomarker.
Persistent infections of high-risk human papillomavirus (HR-HPV) is regarded as the most signi cant risk factor for cervical carcinogenesis. Until now, several lncRNAs including the metastasis associated lung adenocarcinoma transcript 1 (MALAT-1), thymopoietin pseudogene 2 (TMPOP2) and small nucleolar RNA host gene 12 (SNHG12) have been identi ed to be regulated by oncoproteins E6 or E7 [22][23][24]. However, the relationship between MIR210HG and HPV viral proteins has not been investigated. To explore whether the expression of MIR210HG could be induced by HPV 16 E6 or E7, we transfected HPV16 E6-or E7encoding plasmids and siRNAs targeting E6 or E7 into SiHa and Caski cells and found that both E6 and E7 stimulated MIR210HG expression, indicating the involvement of viral proteins in the regulation of MIR210HG gene in CC cells. Furthermore, the mechanism by which viral proteins promote the expression of MIR210HG has not been clari ed. Subsequently, we performed the analysis of the promoter region of MIR210HG using the JASPAR core database and found the presence of putative hypoxia response elements. Interestingly, previous studies showed that E6 and E7 can induce HIF-1α protein accumulation [12]. Considering that MIR210HG was induced by hypoxia as aforementioned, we further provided the rst evidence that hypoxia-induced MIR210HG is upregulated by HPV16 E6/E7 via HIF-1α. Mechanically, luciferase reporter assays demonstrated that HIF-1α might bind directly to the MIR210HG promoter region and activate its transcription. This nding not only shed new light on the orchestrated interactions between HIF-1α and MIR210HG, but also identi ed its upstream TF. Even so, further experiment such as chromatin immunoprecipitation assay (ChIP) are needed in the furture.
In addition, silencing MIR210HG or HIF-1α respectively decreased the mutual expression level, implying that a positive feedback loop between MIR210HG and HIF-1α. Given its importance in sensing hypoxic tension, it is not surprising that cells utilize feedback mechanisms to precisely control HIF-1α signals. A recent study showed that HIF-1α binds the promoter region of LINC00511 to active its transcription and LINC00511 indirectly regulates the expression of HIF-1α through sponging miR-153-5p, forming a positive feedback loop of HIF-1α/LINC00511/miR-153-5p in CRC [25]. And lncRNA HITT was demonstrated to form a regulatory loop with HIF-1α to modulate angiogenesis and tumor growth [26]. Our ndings also showed that either MIR210HG or HIF-1α knockdown could attenuate the expression level of PGK1, a HIF-1α downstream target which acts as the rst ATP-producing enzyme in glycolysis. Ectopic overexpression of MIR210HG could reverse the reduced PGK1 expression caused by HIF-1α silencing. Many studies suggested that PGK1 is highly expressed in various cancers, such as breast cancer [27], liver cancer [28], and colon cancer [29]. In particular, PGK1 has also been proved to be transactivated by some transcription factors including HIF-1α [30,31]. Nevertheless, it remains unknown that the roles of PGK1 in MIR210HG-mediated cancer progression. Our data con rmed that PGK1 depletion notably reversed the migration and invasion capacities in MIR210HG overexpression CSCC cells.

Conclusions
In conclusion, our ndings indicate that MIR210HG functions as an oncogenic lncRNA in CSCC and its high levels are associated with tumor progression and unfavourable prognosis of patients. The present study provides the rst evidence that hypoxia-inducible MIR210HG is induced by HPV16 E6/E7 via transcription factor HIF-1α, and might act as a tumor promoter in CSCC by enhancing the expression of PGK1. Investigation of the molecular mechanism showed that existence of a positive feedback loop between MIR210HG and HIF-1α, while HIF-1α could directly bind to the MIR210HG promoter region and activate its transcription (Figure 8). Therefore, MIR210HG has now emerged as a novel prognostic biomarker and a potential new target for CC treatment.

AVAILABILITY OF DATA AND METERIALS
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

CONSENT FOR PUBLICATION
Written informed consent was obtained from the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal.

CONFLICTS OF INTEREST
The authors declare that they have no competing interests AUTHOR CONTRIBUTIONS JYZ and XLH conceived and designed the study. XLH, XTH and JNZ performed the experiments and wrote the manuscript. XLH, XTH, JNZ, LJW, JL and XX performed the data collection, statistical analysis and data interpretation. LJW, JL and XX contributed to clinical tissue specimens' enrollment and collection. JYZ and XLH performed manuscript revision. All authors read and approved the nal manuscript.          A schematic depicture of the mechanism underlying the role of MIR210HG in CSCC. Our results indicate that MIR210HG could be modulated by HPV16 E6/E7 through HIF-1α. Moreover, hypoxia-induced MIR210HG forms a positive feedback loop with HIF-1α to promote expression of PGK1, simultaneously HIF-1α binds the promoter of MIR210HG and activate it expression at transcriptional level. As a consequence, CSCC cells migration, invasion and glycolytic metabolism capabilities are enhanced.