CircPSMC3 alleviates the symptoms of PCOS by sponging miR‐296‐3p and regulating PTEN expression

Abstract Polycystic ovary syndrome (PCOS), the most common female endocrine disease that causes anovulatory infertility, still lacks promising strategy for the accurate diagnosis and effective therapeutics of PCOS attributed to its unclear aetiology. In this study, we determined the abnormal reduction in circPSMC3 expression by comparing the ovarian tissue samples of PCOS patients and normal individuals. The symptom relief caused by up‐regulation of circPSMC3 in PCOS model mice suggested the potential for further study. In vitro functional experiments confirmed that circPSMC3 can inhibit cell proliferation and promote apoptosis by blocking the cell cycle in human‐like granular tumour cell lines. Mechanism study revealed that circPSMC3 may play its role through sponging miR‐296‐3p to regulate PTEN expression. Collectively, we preliminarily characterized the role and possible insights of circPSMC3/miR‐296‐3p/PTEN axis in the proliferation and apoptosis of KGN cells. We hope that this work provides some original and valuable information for the research of circRNAs in PCOS, not only to better understand the pathogenesis but also to help provide new clues for seeking for the future therapeutic target of PCOS.

physiological features of GCs and overcomes the above problems, has been successfully employed to explore the precise functions and associated molecular mechanisms of GCs many times. 8 Circular RNAs (circRNAs) are a type of non-coding RNAs (ncRNAs) that lack canonical 5' cap and 3' poly A tail attributed to their covalent closed-loop structures. 9 As the competitive endogenous RNAs, circRNAs regulate the expression of multiple genes at the transcriptional and post-transcriptional levels by sponging downstream mi-croRNA(miRNA). 10 Mounting evidences showed that the differential expressions of circRNAs were present in a variety of clinical illnesses such as cancers, CNS diseases, cardiovascular diseases, and some endocrine dysfunctions. [11][12][13][14] Recently, some reports aiming to probe into the differential expression profiles of circRNAs in GCs of PCOS patients found that 4 circRNAs were up-regulated whereas 23 were down-regulated. 15 These circRNAs, which were abnormally expressed primarily in inflammation, proliferation and VEGF signalling pathways, indicated potential function in the occurrence and development of PCOS.
Although a series of differentially expressed circRNAs have been widely discovered in various diseases, little is known about their biogenesis processes and underlying mechanisms. 16 It has been learned that the regulatory function of circRNAs mainly depends on its directly bound downstream miRNAs. 17 MicroRNAs (miRNAs) are a type of conservative small regulatory non-coding RNAs that can suppress the expression of target genes. 18 As miRNAs are involved in key biological processes relied by cell life such as cell proliferation, apoptosis, development and differentiation, they are considered as vital regulators of gene expression. [19][20][21][22] Several reports have shown that miRNAs are mainly regulated by various lncRNAs and circRNAs to further exerting regulatory effects. 23,24 A previous research in mouse ovarian provides evidence that circEGFR could regulate the proliferation of GCs by sponging miR-125a-3p to regulate Fyn. 25 It also reported that 3 cir-cRNAs in cumulus cells of PCOS patients mediate the post-transcriptional regulation of multiple genes by acting as sponges for miRNAs. 26 In this study, we determined the abnormal reduction in circPSMC3 expression by comparing the ovarian tissue samples of PCOS patients and normal individuals. The symptom relief caused by up-regulation of circPSMC3 in PCOS model mice suggested the potential for further study. In vitro functional experiments confirmed that circPSMC3 can inhibit cell proliferation and promote apoptosis by blocking the cell cycle in human-like granular tumour cell lines. Mechanism study revealed that circPSMC3 may play its role through miR-296-3p/PTEN axis. We hope that our work contributes to the valuable insights into the cir-cRNAs in the pathogenesis and treatment of PCOS by exploring the biological functions and molecular mechanisms involved in circPSMC3.

| Study approval
Ethical approvals were obtained from the Ethics Committee of the Renmin Hospital of Wuhan University. All collected samples were obtained subsequent to receive full written informed consent from the patients. Relevant animal experiments were conducted complied with the institutional ethical guidelines. Reproduction and Embryology. Non-PCOS women with regular menstrual cycles and normal ovarian morphology were identified as the control group. As described before, the human granulosa cells for subsequent RNA analyses were recovered and purified from the follicular fluid. 27

| Animal experiments
Sixteen 3-week-old female C57BL/6 mice were housed in our animal centre for 1 week after purchased from Shanghai Animal Centre. The PCOS mouse model was established by subcutaneous injection of 60 mg/kg dehydroepiandrosterone (DHEA, Aladdin, Shanghai, China) every day for 3-6 weeks. Then, these mice were divided into two groups of 8 each, which received a subcapsular injection into the ovary of 5 × 10 8 PFU/mL lentivirus carrying mock vector or circ-PSMC3 vector (GeneChem, Shanghai, China). The ovarian tissues required for subsequent experiments were taken from mice killed after 2 weeks.

| Serum insulin assay
After fasting the mice for 12 hours, 2 g/kg of glucose was intraperitoneally injected. Then, the orbital venous blood of above mice was extracted at 0, 30, 60, 90 and 120 minutes, respectively. Finally, the concentration of serum insulin was detected by mouse insulin ELISA kit (Westang, Shanghai, China), calculating the area under the curve (AUC) for further evaluation.

| H&E staining
According to standard H&E staining protocol, mouse ovarian tissues were stained with haematoxylin and eosin solution (Beyotime, Shanghai, China) after fixed in formalin, paraffin-embedded and sliced at 5 µm in turn. The images were captured and analysed using a fluorescence microscope (Olympus, Tokyo, Japan).

| Quantitative reverse-transcription-polymerase chain reaction (qRT-PCR)
The expression levels of circPSMC3, miR-296-3p and PTEN in the experimental groups were determined by qRT-PCR. In strict accordance with the manufacturer's instructions, total RNAs were ex- Before calculation, we normalized the circRNA and miRNA expression levels by GAPDH or U6, respectively. Signaling and GAPDH (#ab181602, Abcam) overnight. Subsequently, the membranes were then conjunct with secondary antibodies (1:5000) at 25°C for 40 minutes. Finally, every protein band was visualized by enhanced chemiluminescence detection kit (Sangon Biotech, Shanghai, China) and related data were quantified by Image Lab Software.

| Cell counting kit-8 (CCK-8) assay
The KGN cells were inoculated into 96 wells with the density of 4000 cells per well and incubated until cell attachment. Then, these seeded cells were cultured for 1, 2, 3, 4 and 5 days after transfection, respectively. Following treatment, 10 μL of the Cell Counting Kit-8 (CCK-8) solution (KeyGen, Nanjing, China) were supplied to every well. Finally, a microplate spectrophotometer (BioTek, VT, USA) was employed to detect the OD value of each well at 450 nmol\L.

| EdU assay
The effects on the proliferation of KGN cells were measured by EdU assay using the BeyoClick™ EdU-555 Kit (Beyotime Biotechnology, Shanghai, China) in accordance with manufacturer's protocols.
Transfected KGN cells were seeded and cultured for 2 day. After fixed with 4% paraformaldehyde, cells were added with 5-ethynyl-2'-deoxyuridine (EdU, 50 μmol\L each) and incubated at 37°C for 2 hours. These cells were subsequently stained with Apollo Dye Solution and DAPI. Finally, the EdU-positive cells were visualized and analysed under the fluorescence microscope (Olympus, TKY, Japan).

| RNA pull-down assay
To pull down the miR-296-3p through circPSMC3, the cell lysates transfected with miR-296-3p mimics were incubated with biotinlabelled probes of circ-PSMC3-WT or circ-PSMC3-MUT which were previously bound to magnetic beads. Similarly, to pull down the miR-296-3p through PTEN, the cell lysates transfected with miR-296-3p mimics were incubated with biotin-labelled probes of PTEN-WT or PTEN-MUT which were previously bound to magnetic beads. After the pull-down product was extracted, reverse transcription was carried out by qRT-PCR.

| Immunohistochemistry
After fixed in 4% paraformaldehyde, embedded in paraffin and sliced successively, the mouse model samples were incubated with the corresponding primary antibodies against PTEN (Catalogue#9552, Cell Signaling technology) at 4℃ for 24 hours. The images were captured and analysed using an optic microscope (Olympus, Tokyo, Japan).

| Statistical analysis
The results of each experiment were repeated three times, and the experimental data were represented as means ± standard deviation (SD). The analyses were mainly interpreted using the software statistical SPSS version 19.0 (IBM, Chicago, USA). The statistical differences between control group and experimental group were calculated through the unpaired two-tailed t test. Other statistical differences among multi-groups were assessed by utilizing one-way ANOVA or two-way ANOVA. A P < 0.05 was determined statistically significant.

| CircPSMC3 is markedly down-regulated in PCOS patients and alleviates the symptoms in PCOS mice
To preliminarily explore the possible relevance between circPSMC3 and PCOS, we determined the mRNA levels of circPSMC3 in ovarian granulosa cells from 22 PCOS patients and 22 normal individuals through qRT-PCR. The results showed that circPSMC3 expressed exceptionally low in PCOS specimens compared to the normal control group ( Figure 1A). We overexpressed circPSMC3 in the PCOS mice model to further characterize the in vivo effect of circPSMC3 on this disease ( Figure 1B). In the serum insulin release test, we found that circPSMC3 overexpression could effectively alleviate the increase in insulin release at different time points caused by sugar treatment (Figure 1C,D). Further analysing the H&E-stained sections of ovarian tissues in these PCOS mice, the overexpression group showed a more regular tissue structure, a thicker granular cell layer and a larger number of benign granular cells compared with the control group ( Figure 1E-G). These results indicated that circPSMC3 might play a potential role in the biological process of PCOS, which deserves further research.

| CircPSMC3 suppresses the viability and proliferation of KGN cells
The roles of circPSMC3 on viability and proliferation of KGN cells (human granulosa-like tumour cell line) were determined. We de-

| Prediction and verification of the interactions between circPSMC3, miR-296-3p and PTEN
Previous studies have reported that the realization of physiological function of circRNAs requires the presence of various binding miR-NAs to regulate downstream proteins. In consequence, we resorted to the bioinformatic analysis (https://circi ntera ctome.nia.nih.gov/index. html) to predict the direct binding target of circPSMC3. We found that the seed region of miR-296-3p has the sequence complementary to circPSMC3 ( Figure 5A). In order to verify the website prediction, we conducted the luciferase activity assay and RNA pull-down assay. The results showed that miR-296-3p could significantly reduce the luciferase activity of circ-PSMC3-WT-transfected cells instead of MUTtransfected cells ( Figure 5B). In addition, the ability of circ-PSMC3-WT to pull down miR-296-3p was also offset by mutations ( Figure 5C).
Bioinformatic tools (http://www.targe tscan.org/vert_72/) also forecasted Phosphatase and Tensin Homolog (PTEN) as a potential target for miR-296-3p ( Figure 5E). Similarly, we confirmed the direct interaction between miR-296-3p and PTEN with the help of luciferase assay and RNA pull-down assay ( Figure 5F,G). Moreover, the differences in the expression levels of miR-296-3p and PTEN between PCOS patients and healthy control further confirmed the clinical relevance of this predicted axis ( Figure 5D,H). Collectively, we forecasted and verified the physical basis of the circPSMC3/miR-296-3P/PTEN axis.

| CircPSMC3 suppresses the proliferation of KGN cells by sponging miR-296-3p to regulate PTEN
We next determined the biological correlations and biological effects of the circPSMC3/miR-296-3P/PTEN axis in the KGN cells. The cell apoptosis ability was detected by flow cytometer using Annexin V/PI staining in KGN cells transfected with circPSMC3 vector or mock vector. N = 3 for each experiment, and the data were expressed as mean ± standard deviation (SD). *P < 0.05 the proliferation of KGN cells. CircPSMC3 knockdown has been found to be beneficial for the proliferative capacity in our previous study, whereas down-regulating miR-296-3p could dramatically offset this effect ( Figure 6G). These facts indicated that circPSMC3 suppresses the proliferation of KGN cells by sponging miR-296-3p to regulate PTEN.

| The relief of PCOS symptoms by circPSMC3 is achieved by up-regulating the expression of PTEN in vivo
To further clarify the relationship between circPSMC3 and PTEN in the progression of PCOS, we quantified the PTEN levels in the ovarian tissues of model mice. We found that circPSMC3 overexpression greatly increased the number of PTEN-positive cells ( Figure 7A,B). Regardless of the PTEN mRNA level or protein level, 8 circPSMC3 overexpressing mice were significantly higher than the control group ( Figure 7C,D). Obviously, the relief of PCOS symptoms by circPSMC3 is achieved by up-regulating the expression of PTEN in vivo.

| D ISCUSS I ON
Serving as the major endocrine disease-causing anovulatory infertility, PCOS plagues 5%-20% of women of childbearing age in the whole world. 28 However, searching the ideal biomarker for the accurate diagnosis and effective treatment of PCOS has been a formidable challenge because its pathogenesis has not been fully elucidated. Over the past decade, circRNAs were regarded as novel and vital regulators for the expression of genes as their altered expressions had been strongly associated with a variety of diseases. [29][30][31][32] Accumulating evidences suggest that circRNAs could be used as the excellent indicators of diagnosis and prognosis. For instance, the expression of circ_0074026 is significantly increased (F, H) Cell cycle analysis of COV434 or SVOG cells. N = 3 for each experiment, and the data were expressed as mean ± standard deviation (SD). *P < 0.05 for patients with glioma, and this abnormal physiological phenomena indicates unfavourable prognosis. 33 Circ_0001178 was observed to be highly expressed in colorectal cancer, which facilitated the process of metastasis and invasion of cancer cells in turn. 34 Moreover, circ_0080425 expression in diabetic nephropathy tissues is significantly decreased and it markedly inhibits cell proliferation and fibrosis. 35 Unfortunately, there is little research on the potential role and underlying molecular mechanism of circRNAs in the development of PCOS. Overexpression of circ_0023942 was reported to inhibit the proliferation of KGN and COV434 cells by regulating CDK4. 36 Knockdown of hsa_circ_0118530 could reduce the damage of KGN cells by sponging miR-136. 37  CircPSMC3, a circular RNA derived from the PSMC3 gene, was originally discovered as a tumour-suppressor factor for gastric cancer. 38 Later studies found that circPSMC3 inhibited cell proliferation of nasopharyngeal carcinoma and prostate cancer by down-regulating ROCK1 and DGCR8, respectively. 39,40 Recently, circPSMC3 was reported to suppress the invasion and migration of NSCLC cells by regulating the miR-182-5p/NME2 signalling pathway. 41 Similar to most circRNAs, the major point of focus in circPSMC3 was also on tumour diseases. The disorder of cell proliferation and apoptosis is not only a patent for cancer, but also widely exists in other diseases, such as PCOS. As far as we know, our work is the only study to explore the role of circPSMC3 in non-tumour diseases, which may provide some meaningful reference for the research of circPSMC3 in a wider range of diseases.
Whether it is miR-296-5p (miRNA-296 from the 5' end) or miR-296-3p (miRNA-296 from the 3' end), previous studies principally focused on their expression and role in human cancer. For instance, miR-296-3p was proved to increase tumour cell resistance to natural killer cells as an oncogene in prostate cancer, while exerted inhibitory effects on tumour cell migration and invasion in many types of malignancies such as choroidal melanoma, glioblastoma and non-small-cell lung carcinoma. [42][43][44][45] Similar to its homolog, miR-296-5p has also been studied in the occurrence and development of various human cancers including oesophageal cancer, glioma and breast cancer. [46][47][48] Except for miR-296, many other miRNAs have been studied for their precise function and associated molecular mechanism in PCOS. MiR-3940-5p, miR-140 and miR-155 were found to promote GC proliferation while miR-135a and miR-30d-5p induced apoptosis in GCs. [49][50][51][52][53] We predicted and confirmed that miR-296-3p was involved in regulating the proliferation of KGN  The protein expressions of PTEN were significantly up-regulated in PCOS mice after circ-PSMC3 lentivirus injection (n = 8 individuals per group). N = 3 for each experiment, and the data were expressed as mean ± standard deviation (SD). *P < 0.05 PTEN, as a well-known negative regulator of the PI3K/AKT pathway, has been studied as a tumour-suppressor gene repeatedly. 54 However, increasing evidence suggested that this signalling pathway is equally important for the initiation and progression of PCOS, which may affect insulin resistance and hyperandrogenism. 55 Previous studies reported that miR-17-92 could promote cell proliferation and reduce the ratio of differentiated cells in bovine GCs by down-regulating PTEN, suggesting that up-regulating PTEN might be a strategy to suppress GCs. 56 Consistent with it, our results showed that up-regulating PTEN by circPSMC3 suppressed the proliferation of KGN cells. Nevertheless, what puzzles us is that miR-200b and miR-200c have been reported to reduce KGN cells proliferation by down-regulating PTEN, which is the exact opposite of our conclusion. 57 We speculate that some Our research still remains some limitations and weaknesses.
Firstly, we used KGN cells instead of GCs in this study. As mentioned above, compared to the original GCs or SVOG cells, the KGN cell line does not seem to be the best choice for studying PCOS.
We will search for the basis of the differences in the circPSMC/ miR-29-3p/PTEN axis in these three cell lines in the follow-up work. Additionally, since the focus of this work is on circRNAs, we have not studied the function and mechanism of miR-296-3p and PTEN in detail in vivo experiments. In the future, we plan to explore the potential functions and molecular mechanisms of these two in PCOS.
In conclusion, our finding reveals that circPSMC3 helps to alleviate the symptoms in PCOS mice via sponging miR-296-3p to up-regulate PTEN expression level. We hope that our work contributes to the valuable insights into the better understanding of circRNAs in the occurrence and development of PCOS by exploring the biological functions and molecular mechanisms involved in circPSMC3.

CO N FLI C T O F I NTE R E S T
No potential conflict of interest was reported by the authors.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon reasonable request.