1 INTRODUCTION

As a most common malignant gynecological tumor, cervical cancer has been reported as a life-threatening reproductive health problem for females all over the world, especially among women between the ages of 20 and 39.1, 2 Although the HPV vaccine has been considered a good method to protect women from cervical cancer, approximately one-half of adolescent girls have not been fully vaccinated.3 Owing to no typical symptoms in an early stage, the majority of patients with cervical cancer have developed invasive cancer at the time of diagnosis.4 Therefore, there is a great significance to explore the molecular mechanism underlying cervical cancer tumorigenesis for seeking alternative therapeutic strategies.

Currently, some studies discovered that more than 90% of the human genome was transcribed as non-coding RNAs.5 As a novel family of non-coding RNAs, circular RNAs (circRNAs) are characterized by a closed continuous loop without a 5′ cap or a 3′ poly-A tail,6 presenting higher stability than linear RNAs. CircRNAs also possess remarkable abundance, evolutionary conservation, and are highly tissue-specific, emphasizing the importance of circRNAs in human diseases, containing cancer.7, 8 Accumulating evidence indicates that circRNAs are widely aberrantly expressed in many cancers, and most of them have been confirmed to be involved in tumor development and progression.9, 10 In terms of cervical cancer, Tang et al. displayed that has_circ_0000515 could boost the malignant behaviors of cervical cancer through binding to miR-326.11 Moreover, Cai et al. confirmed that has_circ_0000263 could exert an oncogenic role by promoting proliferation and migration in cervical cancer.12 circ_0023404, a typical circRNA originating from ring finger protein 121 gene (RNF121), has been confirmed to be upregulated in non-small cell lung cancer and osteoarthritis.13, 14 Furthermore, relevant research indicated that the dysregulation of circ_0023404 could participate in the modulation of cell growth and metastasis in cervical cancer.13 Nevertheless, the precise function and potential mechanism of circ_0023404 in cervical cancer are far from being addressed.

In recent years, numerous studies are focusing on the circRNA–microRNA (miRNA)–mRNA regulatory mechanism in human cancers, which propose circRNAs can derepress target mRNAs expression by competitively binding to miRNAs.15-17 Here, bioinformatics analysis discovered that circ_0023404 has some binding sites with miR-636. Apart from that, it has been proved that miR-636 exhibits the tumor suppressor in hepatocellular carcinoma and non-small cell lung carcinoma.18, 19 Furthermore, miR-636 has been pointed out to inhibit cell survival in cervical cancer.20 Hence, the aim of this study is to explore whether the regulatory role of circ_0023404 on cervical cancer progression could be mediated by sponging miR-636.

2 MATERIALS AND METHODS 2.1 Clinical samples and cell culture

Fifty-one cervical cancer tissue samples and paired adjacent non-tumor tissues were collected from patients with cervical cancer at Dalian Municipal Women and Children's Medical Center (Group) in this research, which had acquired the approval of the Ethics Committee of Dalian Municipal Women and Children's Medical Center (Group) (Approval no. dlfy-201906002). Meanwhile, written informed consent was provided by all patients.

Human cervical epithelial cells (HcerEpic, ScienCell, Carlsbad, California) and cervical cancer cell lines (C33A, HeLa, SiHa, American Type Culture Collection, Manassas, Virginia) were respectively cultured in Cervical Epithelial Cell Growth Supplement (ScienCell) and Eagle's Minimum Essential Medium (Invitrogen, Paisley Scotland, UK). And cervical cancer cell line Caski (American Type Culture Collection) were grown in Roswell Park Memorial Institute-1640 medium (RPMI; Invitrogen). All cells were maintained in a humidified atmosphere with 5% CO2 at 37°C, and the above medium was supplemented with 10% fetal bovine serum (FBS; PAN Biotech, Aidenbach, Germany).

2.2 Real-time quantitative polymerase chain reaction After the treatment with TRIzol reagent (Invitrogen), the isolated total RNAs from clinical specimens and cell lines were reversely transcribed to cDNA using PrimeScript RT Reagent Kit (TaKaRa, Tokyo, Japan). After that, real-time quantitative polymerase chain reaction (RT-qPCR) reaction was carried out on a ABI 7900 thermocycler (Applied Biosystems, Rotkreuz, Switzerland), along with a SYBR Green PCR Kit (Takara), followed by normalization with GAPDH for circ_0023404, linear RNF121, cytochrome P450 2S1 (CYP2S1), and U6 for miR-636. And then, the relative expression of genes was calculated by the 2–ΔΔCt method, and the primers for the target genes were shown as below: circ_0023404: 5′-CTGGTGCAGTGGAAGCAGAG-3′ (sense), 5′-CGACCCTCC ATTGCTCTTCT-3′ (antisense); RNF121 mRNA: 5′-TGGTCCTCATCCTCATCGCA-3′ (sense), 5′-GGGTCACCATATTGTAGGAGCG-3′ (antisense); miR-636: 5′-TGTGCTTGCTCGTCCC-3′ (sense), 5′-GAACATGTCTGCGTATCTC-3′ (antisense); CYP2S1: 5′-AGGCGTTCCTGCCCTTCTCC-3′ (sense), 5′-CAGTGGGACGGACTTGCAGC-3′ (antisense); U6: 5′-CTCGCTTCGGCAGCACA-3′ (sense), 5′-AACGCTTCACGAATTTGCGT-3′ (antisense); GAPDH: 5′-GGTCACCAGGGCTGCTTT-3′ (sense), 5′-GGAAGATGGTGATGGGATT-3′ (antisense). 2.3 Ribonuclease R treatment

In this study, to evaluate circRNA resistance to a specific ribonuclease R (RNase R), total RNAs (2 μg) were incubated for 60 min at 37°C with or without 4 U/μg of RNase R (Qiagen, Valencia, California). After purification with an RNeasy MinElute Cleaning Kit (Qiagen), RT-qPCR assay was used to detect the expression levels of circ_0023404 and linear RNF121.

2.4 Cell transfection

For circ_0023404 downregulation, C33A and HeLa cells were transfected with 40 nM of circ_0023404 small interference RNA (si-circ_0023404, RiboBio, Guangzhou, China) or the negative control (si-NC, RiboBio), severally. Also, miR-636 upregulation or knockdown was achieved by 40 nM of miR-636 mimic (miR-636, RiboBio) or miR-636 inhibitor (anti-miR636, RiboBio), and their negative controls (miR-NC, anti-NC, RiboBio). Meanwhile, CYP2S1 overexpression vector was constructed by inserting the sequence of CYP2S1 into pcDNA3.0 (Invitrogen, a negative control), followed by transfection into C33A and HeLa cells with 50 ng vector. In this study, Lipofectamine 3000 reagent (Invitrogen) was applied for all transfection, in line with the operation manual. After transfection for 48 hr, the obtained cells were applied for the subsequent assays.

2.5 Cell viability assay

After incubation for 24 hr in 96-well plates, transfected C33A and HeLa cells were added with 20 μL of 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT, Sigma-Aldrich, St. Louis, Missouri) for 4 hr. The formed formazan crystals were dissolved in 150 μL of dimethyl sulfoxide (DMSO, Sigma-Aldrich). At length, the absorbance was recorded at 490 nm, based on the supplier's direction of a microplate reader (Bio-Tek Instruments, Winooski, Vermont).

2.6 5-Ethynyl-2′-deoxyuridine assay

Briefly, transfected C33A and HeLa cells (4 × 103 cells/well) were introduced in 96-well plates, followed by addition with 50 μM 5-ethynyl-2′-deoxyuridine (EdU) (RiboBio) at 37°C. After incubation for 2 hr, the 4% formaldehyde solution was used to fix the cells, which then was treated with 0.5% Triton X-100 for 30 min. After that, the treated cells were stained using Apollo and DAPI, and the proliferation-positive cells were photographed and counted under a fluorescence microscope (Nikon, Tokyo, Japan).

2.7 Colony formation assay

In general, transfected cells were trypsinized and uniformly dispersed into 6-well plates at a concentration of 500 cells/well. After incubation for 10 days, the treated cells were fixed with 3.7% methanol (Sigma-Aldrich) for 20 min before being stained with 0.1% crystal violet (Sigma-Aldrich) for 30 min. After that, the number of colonies was analyzed using a light microscope (Nikon, magnification ×40), as per the operation manual.

2.8 Transwell assay

The assessment of migration and invasion was performed using a transwell chamber (Corning Costar, Corning, New York). In short, transfected cells (5 × 104 cells/well) in FBS-free medium were introduced into the upper chamber for migration assay, whereas 1 × 105 transfected cells were located on the pre-coated matrigel (BD Biosciences, Heidelberg, Germany) chamber for invasion assay. Meanwhile, the bottom counterparts were filled with the medium supplemented 10% FBS. Twenty-four hours later, 0.1% crystal violet (Sigma-Aldrich) was applied to stain the cells in the bottom chamber, followed by detection with a light microscope (Nikon, magnification ×100).

2.9 Cell apoptosis assay

In brief, transfected cells were re-suspended in binding buffer, followed by stain with 5 μL Annexin (V-fluorescein isothiocyanate) V-FITC (Bender Med System, Vienna, Austria) and 10 μL propidium iodide (PI, Bender Med System). Fifteen minutes later, the FACSCalibur (BD Biosciences) was implemented to analyze the apoptosis rate, based on the manufacturer's instructions.

2.10 Western blot assay

After treatment with RIPA lysis buffer (Beyotime, Shanghai, China), the extracted protein samples from tissues and cells were quantified using a BCA kit (Pierce, Rockford, Illinois). Following blockage with 5% fat-free milk, the samples were subjected to 10% SDS-PAGE and transferred to PVDF membranes (Millipore, Molsheim, France). After that, the primary antibodies: cyclin D1 (ab40754, 1:1000, Abcam, Cambridge, Massachusetts), matrix metallopeptidase 9 (MMP9, ab76003, 1:1000, Abcam), Bcl-2-associated X protein (Bax, ab32503, 1:1000, Abcam), CYP2S1 (ab69650, 1:1000, Abcam), and GAPDH (1:1000, ab181602, Abcam) were used to hybridize the membranes at 4°C overnight, which was then were incubated with secondary antibody (ab6721, 1:10,000, Abcam) for 2 hr at the room at room temperature. Finally, the detection of the protein bands was conducted based on the user's guidebook of ECL reagent (Invitrogen) and Quantity One software (Bio-Rad).

2.11 Dual-luciferase reporter assay

In this assay, we first predicted the binding sites between miR-636 and circ_0023404 or CYP2S1 3′ untranslated region (3′ UTR) by using the online software Circinteractome (https://circinteractome.nia.nih.gov/) and targetscan (http://www.targetscan.org). Then, a dual-luciferase reporter assay was carried out to verify the predication in C33A and HeLa cells. In short, the pGL3 Basic reporter vector (Promega, Fitchburg, Wisconsin) was applied for the establishment of wild-type (WT) luciferase reporter vectors (circ_0023404 WT or CYP2S1 3′ UTR WT) by introduction the sequences of circ_0023404 or CYP2S1 3′ UTR possessing miR-636 binding sites. Also, the mutant-type (MUT) vectors (circ_0023404 MUT or CYP2S1 3′ UTR MUT) were acquired via mutating the binding sites of miR-636. After that, C33A and HeLa cells were transfected with the above vectors together with miR-NC or miR-636, followed by the analysis of luciferase activities using a dual-luciferase reporter assay kit (Promega).

2.12 RNA pull-down assay

According to the user's guidebook of Lipofectamine RNAiMax (Invitrogen), C33A and HeLa cells were transfected with biotinylated (Bio)-miR-NC or Bio-miR-636 (50 nM), followed by incubation for 48 hr. After sonicating, the obtained cell lysates were treated with magnetic beads (Sigma-Aldrich) for 3 hr. At last, the samples were subjected to purification and RT-qPCR analysis of circ_0023404.

2.13 Statistical analysis

GraphPad Prism7 (GraphPad Prism software, San Diego, California) was used to analyze all data, which was presented as the mean ± standard deviation (SD). Statistical significance was assessed by Student's t-test for two groups and one-way analysis of variance (ANOVA) with Tukey's tests for multiple groups. The significance level was set at P < 0.05.

3 RESULTS 3.1 circ_0023404 was elevated in cervical cancer tissues and cells

First of all, to check the functional role of circ_0023404 in cervical cancer, its expression pattern was determined by RT-qPCR assay. Data displayed that circ_0023404 expression in 51 cervical tumor tissues was significantly higher than that in paired-paracancer tissues (Figure 1A). Furthermore, we noticed that circ_0023404 level in all four cervical cancer cell lines (C33A, HeLa, Caski, and SiHa) compared with the normal cervical cell line HcerEpic. Meanwhile, C33A and HeLa cell lines presented the highest expression of circ_0023404 (Figure 1B), so those two cell lines were selected for the following experiments. Subsequently, to prove the circular characteristics of circ_0023404, RT-qPCR assay were performed using random primers and oligo (dT)18 primers. As presented in Figure 1C,D, circ_0023404 level was markedly reduced in C33A and HeLa cells when using the oligo (dT)18 primers, whereas RNF121 mRNA level was not, suggesting that circ_0023404 had no poly-A tail. In addition, we further verified the stability of circ_0023404 by the treatment of RNase R. Results indicated that circ_0023404 was resistant to RNase R, while linear RNA RNF121 could be digested by RNase R in C33A and HeLa cells (Figure 1E,F). Together, these data suggested that circ_0023404, a circular and stable transcript, might be involved with cervical cancer progression.

circ_0023404 was elevated in cervical cancer tissues and cells. (A) RT-qPCR assay was used to measure the expression level of circ_0023404 in cervical tumor tissues (n = 51) and adjacent non-tumor tissues (n = 51). (B) Relative expression of circ_0023404 was detected in HcerEpic cells and cervical cancer cell lines (C33A, HeLa, Caski, and SiHa) by RT-qPCR assay. (C and D) Expression levels of circ_0023404 and RNF121 mRNA in C33A and HeLa cells were analyzed by RT-qPCR assay after normalized with random primers and oligo (dT)18 primers. (E and F) circ_0023404 and RNF121 mRNA were determined by RT-qPCR assay in C33A and HeLa cells treated with or without RNase R. **P < 0.01, ***P < 0.001

3.2 circ_0023404 knockdown repressed proliferation, migration, invasion, and promoted apoptosis of cervical cancer cells in vitro

In view of the high expression of circ_0023404 in cervical cancer, we knocked down circ_0023404 expression in C33A and HeLa cells. And transfection efficiency of si-circ_0023404 was examined and illustrated in Figure 2A. After that, the results of MTT assay and EDU assay suggested that cell proliferative ability was inhibited due to the deficiency of circ_0023404 in C33A and HeLa cells compared with their respective controls (Figure 2B–E). Also, we found that the silencing of circ_0023404 could reduce the number of colonies in C33A and HeLa cells relative to the negative control (Figure 2F). Furthermore, we conducted transwell assays to assess the migration and invasion of cervical cancer cells. As shown in Figure 2G,H, the abilities of migration and invasion were evidently after the transfection of si-circ_0023404 in C33A and HeLa cells when compared with the cells transfected with si-NC. Apart from that, enhanced apoptosis rated was observed caused by circ_0023404 downregulation in C33A and HeLa cells (Figure 2I). In addition, to further explore the effects of circ_0023404 on cervical cancer progression, we determined the protein levels of cyclin D1 (a promoting factor for cell growth), MMP-9 (a pro-migration/invasion factor), and Bax (a pro-apoptosis factor). As presented in Figure 2J,K, the knockdown of circ_0023404 led to a substantial decline in cyclin D1 and MMP-9, and an obvious improvement in Bax level in C33A and HeLa cells. Collectively, these results suggested that the downregulation of circ_0023404 could mitigate the progression of cervical cancer in vitro.

The effects of circ_0023404 downregulation on proliferation, migration, invasion, and apoptosis of cervical cancer cells. C33A and HeLa cells were transfected with si-NC and si-circ_0023404. (A) The expression level of circ_0023404 was determined in transfected C33A and HeLa cells by RT-qPCR assay. (B and C) Cell viability was assessed in transfected C33A and HeLa cells by MTT assay. (D and E) Cell proliferation was determined in transfected C33A and HeLa cells by EDU assay. (F) The number of colonies was detected in transfected C33A and HeLa cells by cell colony formation assay. (G and H) Migration and invasion were examined in transfected C33A and HeLa cells by transwell assay. (I) Apoptosis rate was tested in transfected C33A and HeLa cells by flow cytometry assay. (J and K) Protein levels of cyclin D1, MMP-9, and Bax was tested in transfected C33A and HeLa cells by western blot assay. ***P < 0.001

3.3 circ_0023404 could directly binding to miR-636

Previous studies have discovered that miR-636 can take part in the regulation of cell growth and metastasis of various cancers.18, 19 Interestingly, we found that miR-636 expression was decreased in cervical tumor tissues versus adjacent non-tumor tissues (Figure 3A), which is contrary to the expression of circ_0023404. Meanwhile, we also verified that the trend of miR-636 expression at the cervical cancer cellular level was in line with that in tissues (Figure 3B). Therefore, we further explored the relationship between circ_0023404 and miR-636 using the online software circinteractome. The bioinformatics analysis revealed that there were some binding sites between circ_0023404 and miR-636 (Figure 3C), as proved by a dual-luciferase reporter assay. As shown in Figure 3D,E, the overexpression of miR-636 efficiency reduced the luciferase activity of circ_0023404 WT reporter vector but not that of circ_0023404 MUT reporter vector in C33A and HeLa cells. Moreover, RT-qPCR analysis indicated that the expression level of miR-636 was promoted caused by circ_0023404 downregulation in C33A and HeLa cells (Figure 3F). To further confirm the interaction between circ_0023404 and miR-636 in cervical cancer cells, we performed RNA pull-down assay. As presented in Figure 3G, that the enrichment of circ_0023404 in the captured fraction was much higher in the Bio-miR-636 group than in the Bio-miR-NC group. The above results demonstrated that circ_0023404 could directly bind to miR-636 in cervical cancer cells.

circ_0023404 directly interacted with miR-636. (A) miR-636 level was determined in 51 cervical tumor tissues and matched adjacent non-tumor tissues by RT-qPCR assay. (B) miR-636 level was assessed in HcerEpic cells and cervical cancer cell lines (C33A, HeLa, Caski, and SiHa) by RT-qPCR assay. (C) The sequences of circ_0023404 containing the miR-636 binding sites or mutant binding sites were shown. (D and E) A dual-luciferase reporter assay was performed to verify the binding relationship in C33A and HeLa cells. (F) miR-636 level was determined in si-NC or si-circ_0023404-transfected C33A and HeLa cells by RT-qPCR assay. (G) RNA pull-down assay was carried out to confirm the interaction between miR-636 and circ_0023404 in C33A and HeLa cells. **P < 0.01, ***P < 0.001

3.4 miR-636 upregulation could block the malignant behaviors of cervical cancer cells in vitro

Furthermore, we further explored the effects of miR-636 on cervical cancer progression by using gain-of-function assays. First of all, the overexpression efficiency of miR-636 mimic was determined and presented in Figure 4A. Subsequently, functional analysis displayed that cell proliferative ability was obviously restrained on account of miR-636 upregulation in C33A and HeLa cells (Figure 4B–E). Likewise, the colony number in C33A and HeLa cells transfected miR-636 mimic was declined relative to the cells transfected with miR-NC (Figure 4F). Apart from that, the forced expression of miR-636 could repress the abilities of migration and invasion in C33A and HeLa cells compared with the negative controls (Figure 4G,H). In addition, flow cytometry assay exhibited that the apoptosis rate was distinctly accelerated after miR-636 overexpression in C33A and HeLa cells (Figure 4I). Besides, miR-636 upregulation resulted in a significant decrease in cyclin D1 and MMP-9 levels, and a striking elevation in Bax level in C33A and HeLa cells (Figure 4J,K), further verifying the effects of miR-636 on proliferation, migration, invasion, and apoptosis in cervical cancer cells. In a word, miR-636 upregulation could impede cervical cancer progression in vitro.

Overexpression of miR-636 suppressed proliferation, migration, invasion, and induced apoptosis of cervical cancer cells. C33A and HeLa cells were transfected with miR-NC or miR-636. (A) miR-636 level was detected in transfected C33A and HeLa cells by RT-qPCR assay. (B and C) MTT assay was used to assess cell viability in transfected C33A and HeLa cells. (D and E) EDU assay was performed to test cell proliferation in transfected C33A and HeLa cells. (F) Cell colony formation assay was conducted to examine colonies number in transfected C33A and HeLa cells. (G and H) Transwell assay was carried out to measure migration and invasion in transfected C33A and HeLa cells. (I) Flow cytometry assay was executed to assess apoptosis rate in transfected C33A and HeLa cells. (J and K) Western blot assay was implemented to determine the protein levels of cyclin D1, MMP-9, and Bax in transfected C33A and HeLa cells. ***P < 0.001

3.5 circ_0023404 silencing could repress the malignant behaviors of cervical cancer cells by interacting with miR-636 in vitro

In order to provide further mechanistic insight into the link between miR-636 and circ_0023404 on tumor progression in cervical cancer cells, we performed rescue assays. The knockdown efficiency of anti-miR-636 was assessed and shown in Figure 5A. Functionally, cell proliferative ability impaired by circ_0023404 deficiency was relieved after the co-transfection of anti-miR-636 in C33A and HeLa cells (Figure 5B–E). Analogously, the downregulation of miR-636 could overturn si-circ_0023404-mediated decrease in colony number in C33A and HeLa cells (Figure 5F,G). Meanwhile, transwell assay displayed that miR-636 inhibitor could abrogate the negative action of circ_0023404 deletion on migration and invasion in C33A and HeLa cells (Figure 5H–K). What's more, the promoting effect of apoptosis rate caused by circ_0023404 knockdown was reversed by miR-636 downregulation in C33A and HeLa cells (Figure 5L,M). After that, western blot assay results exhibited that circ_0023404 deficiency-triggered reduction in cyclin D1 and MMP-9 levels, and improvement in Bax level was abolished by miR-636 knockdown in C33A and HeLa cells (Figure 5N,O). Collectively, these results discovered that the miR-636 downregulation could partly reverse the repression effect of circ_0023404 silencing on cervical cancer progression in vitro.

Downregulation of miR-636 abolished the effects of circ_0023404 silencing on proliferation, migration, invasion, and apoptosis in cervical cancer cells. (A) miR-636 level was determined in anti-NC or anti-miR-636-transfected C33A and HeLa cells by RT-qPCR assay. (B–O) C33A and HeLa cells were transfected with si-NC, si-circ_0023404, si-circ_0023404 + anti-NC, and or si-circ_0023404 + anti-miR-636. (B and C) The detection of cell viability was performed by MTT assay in transfected C33A and HeLa cells. (D and E) The assessment of cell proliferation was conducted by EDU assay in transfected C33A and HeLa cells. (F and G) The analysis of colonies number was carried out by cell colony formation assay in transfected C33A and HeLa cells. (H-K) The measurement of migration and invasion was implemented by transwell assay in transfected C33A and HeLa cells. (L and M) The detection of apoptosis rate was performed by flow cytometry assay in transfected C33A and HeLa cells. (N and O) The determination of cyclin D1, MMP-9, and Bax protein levels were conducted by western blot assay in transfected C33A and HeLa cells. **P < 0.01, ***P < 0.001

3.6 CYP2S1 is a direct target of miR-636

It has been reported that CYP2S1 serve as a carcinogenic factor in diverse cancer, such as thyroid cancer and liver cancer.21, 22 In this study, RT-qPCR assay and western blot assay discovered that the upregulation of CYP2S1 in cervical tumor tissues and cells when compared with their respective controls (Figure 6A–D). Notably, bioinformatics analysis suggested that CYP2S1 3′ UTR has some binding sites with miR-636 (Figure 6E). And the following dual-luciferase reporter assay displayed that the luciferase activity in C33A and HeLa cells transfected with CYP2S1 3′ UTR WT and miR-636 was notably declined relative to that in cells transfected with CYP2S1 3′ UTR WT and miR-NC, whereas there was no evident impact in the cells with CYP2S1 3′ UTR MUT (Figure 6F,G). Furthermore, we noticed that the protein level of CYP2S1 was enhanced due to the downregulation of miR-636 in C33A and HeLa cells (Figure 6H). Intriguingly, western blot results suggested that the co-transfection of anti-miR-636 could counteract the inhibitory effect of circ_0023404 silencing on CYP2S1 protein level in C33A and HeLa cells (Figure 6I,J), implying that circ_0023404 could regulate CYP2S1 expression by sponging miR-636. Overall, these data suggested that CYP2S1 acted as a target of miR-636 in cervical cancer cells.

CYP2S1 was a target of miR-636. (A and B) CYP2S1 expression was determined in cervical tumor tissues and matched adjacent non-tumor tissues by RT-qPCR assay and western blot assay. (C and D) CYP2S1 expression was assessed in HcerEpic cells and cervical cancer cell lines (C33A, HeLa, Caski, and SiHa) by RT-qPCR assay and western blot assay. (E) Schematic of a putative target sequence for miR-636 in CYP2S1 3′ UTR and mutated miR-636-binding sites. (F and G) The binding relationship was verified by a dual-luciferase reporter assay in C33A and HeLa cells. (H) CYP2S1 level was detected in anti-NC or anti-miR-636-transfected C33A and HeLa cells by western blot assay. (I and J) CYP2S1 level was measured in C33A and HeLa cells transfected with si-NC, si-circ_0023404, si-circ_0023404 + anti-NC, and or si-circ_0023404 + anti-miR-636 by western blot assay. ***P < 0.001

3.7 CYP2S1 upregulation could partly reverse the effects of circ_0023404 knockdown on proliferation, migration, invasion, and apoptosis in cervical cancer cells

Based on the above findings, we inferred that circ_0023404 could exert its carcinogenic role partially by the circ_0023404/miR-636/CYP2S1 axis in cervical cancer. To verify the assumption, we further tunneled whether the regulatory role of circ_0023404 on cervical cancer development could be mediated by CYP2S1. And the overexpression efficiency of pcDNA3.0-CYP2S1 was detected and exhibited in Figure 7A. Functional analysis indicated that the deficiency of circ_0023404 could hinder cell proliferative ability in C33A and HeLa cells, while the re-introduction of pcDNA3.0-CYP2S1 strikingly counteracted the effects (Figure 7B–G). Simultaneously, the overexpression of CYP2S1 could attenuate the suppressive action of circ_0023404 downregulation on migration and invasion in C33A and HeLa cells (Figure 7H–K). Also, the increased apoptosis rate caused by si-circ_0023404 was remarkably abrogated by CYP2S1 upregulation in C33A and HeLa cells (Figure 7L,M). That was to say, the forced expression of CYP2S1 could overturn the effects of circ_0023404 knockdown on proliferation, migration, invasion, and apoptosis in C33A and HeLa cells, accompanied with higher cyclin D1 and MMP-9, and lower Bax level (Figure 7N,O). These mentioned results unraveled that circ_0023404 silencing might dampen cervical cancer progression by regulating miR-636/CYP2S1 axis.

CYP2S1 upregulation could partly reverse the effects of circ_0023404 knockdown on proliferation, migration, invasion, and apoptosis in cervical cancer cells. (A) CYP2S1 protein level was detected in pcDNA3.0 or pcDNA3.0-CYP2S1-transfected C33A and HeLa cells by western blot assay. (B–O) C33A and HeLa cells were transfected with si-NC, si-circ_0023404, si-circ_0023404 + pcDNA3.0, and si-circ_0023404 + pcDNA3.0-CYP2S1. (B and C) Cell viability was measured by MTT assay. (D and E) Cell proliferation was assessed by EDU assay. (F and G) The number of colonies was examined by cell colony formation assay. (H–K) Migration and invasion were tested by transwell assay. (L and M) Apoptosis rate was monitored by flow cytometry assay. (N and O) Protein levels of cyclin D1, MMP-9, and Bax were detected by western blot assay. **P < 0.01, ***P < 0.001

4 DISCUSSION

Over the past several years, plenty of circRNAs have been gradually identified from various animal genomes owing to the development of next-generation sequencing.23 Unlike linear RNAs, the stable structure of circRNAs has highlighted the underlying promising biomarkers in human cancer.24, 25 In practice, an extensive body of recent research has demonstrated that circRNAs function as essential regulators in the tumorigenesis and development of different cancers, including cervical cancer.26, 27 In this article, circ_0023404, located on chromosome 11, was identified as a typical circRNA and presented a high expression in cervical cancer tissues and cells, consistent with former report.28 Furthermore, functional analysis manifested that the downregulation of circ_0023404 dampened cell proliferative ability and metastasis in cervical cancer in vitro. That is, circ_0023404 exerted a carcinogenic action in cervical cancer both in vitro and in vivo.

It has been widely believed that circRNAs can perform their biological function through serving as miRNAs sponges.29, 30 Meanwhile, miR-636 has been described to participate in the regulation of tumor cell growth and metastasis in bladder cancer and hepatocellular carcinoma.19, 31 In this article, the downregulation of miR-636 in cervical cancer tissues and cells was found, in accordance with the earlier literature.20 Therefore, using the online software circinteractome, we are the first to identify that the binding relationship between miR-636 and circ_0023404 in cervical cancer. And the dysregulation of miR-636 could suppress cell growth and metastasis of cervical cancer cells. Further, the present research demonstrated that circ_0023404 silencing repressed cervical cancer cell development by increasing miR-636. Apart from that, circ_0023404 was also reported to accelerate proliferation and metastasis of cervical cancer and non-small cell lung via sponging various miRNAs.13, 28, 32 These data discovered that circ_0023404 takes part in complex regulatory networks and provided cell type-specific modulation of cell function in multiple tumors.

CYP2S1 is a somewhat atypical cytochrome P450 enzyme, which has been reported to be involved in the metabolism of toxic and carcinogenic compounds.33 Some studies have exhibited that CYP2S1 is highly expressed in breast cancer and thyroid cancer.21, 34 The current article proved the upregulation of CYP2S1 in cervical cancer tissues and cells for the first time. Canonically, miRNAs could regulate tumorigenesis by inducing the degradation of their target mRNAs. In this study, our results first displayed that miR-636 directly targeted CYP2S1 and interacted with CYP2S1. More importantly, the present findings exhibited that circ_0023404 could regulate CYP2S1 expression by targeting miR-636 in cervical cancer cells, verifying the circ_0023404/miR-636/CYP2S1 axis. In addition, the overexpression of CYP2S1 could partly reverse the suppression effect of circ_0023404 downregulation on the development of cervical cancer cells, further supporting the regulatory role of circ_0023404 on cervical cancer progression was mediated partially via the miR-636/CYP2S1 axis.

Taken together, our findings are the first to reveal the regulatory role of circ_0023404/miR-636/CYP2S1 axis in the regulation of cervical cancer progression (Figure 8), providing a new potential therapeutic target for cervical cancer.

circ_0023404 promotes the malignant behaviors of cervical cancer cells by targeting the miR-636/CYP2S1 axis

CONFLICT OF INTEREST

All authors declare no conflict of interest.

ETHICS APPROVAL STATEMENT

The study was approved by the Ethics Committee of Dalian Maternal and Child Health Care Hospital. Meanwhile, written informed consent was provided by all patients.