Identification of a novel circ_0018289/miR-183-5p/TMED5 regulatory network in cervical cancer development

Background Circular RNAs (circRNAs) are increasingly implicated in regulating human carcinogenesis. Previous work showed the oncogenic activity of circ_0018289 in cervical cancer. However, the molecular basis underlying the modulation of circ_0018289 in cervical carcinogenesis is still not fully understood. Methods The levels of circ_0018289, microRNA (miR)-183-5p, and transmembrane p24 trafficking protein 5 (TMED5) were measured by quantitative real-time polymerase chain reaction (qRT-PCR) or western blot assay. Ribonuclease (RNase) R and subcellular localization assays were used to characterize circ_0018289. Cell proliferation was detected by the Cell Counting Kit-8 (CCK-8) and 5-ethynyl-2′-deoxyuridine (Edu) assays. Cell apoptosis and tube formation were assessed by flow cytometry and tube formation assays, respectively. A dual-luciferase reporter assay was performed to confirm the direct relationship between miR-183-5p and circ_0018289 or TMED5. The role of circ_0018289 in tumor growth was gauged by mouse xenograft experiments. Results Circ_0018289 was overexpressed in cervical cancer tissues and cells. Circ_0018289 silencing impeded cell proliferation, enhanced cell apoptosis, and suppressed angiogenesis in vitro, as well as diminished tumor growth in vivo. Mechanistically, circ_0018289 targeted and regulated miR-183-5p by binding to miR-183-5p, and circ_0018289 regulated cervical cancer development and angiogenesis partially through miR-183-5p. Moreover, TMED5 was directly targeted and inhibited by miR-183-5p through the perfect complementary sites in TMED5 3′UTR, and TMED5 knockdown phenocopied miR-183-5p overexpression in suppressing cervical cancer development and angiogenesis. Furthermore, circ_0018289 induced TMED5 expression by competitively binding to shared miR-183-5p. Conclusion Our observations identified the circ_0018289/miR-183-5p/TMED5 regulatory network as a novel molecular basis underlying the modulation of cervical carcinogenesis. Supplementary Information The online version contains supplementary material available at 10.1186/s12957-021-02350-y.


Introduction
Among women, cervical cancer ranks fourth for cancer mortality throughout the world [1]. Early-stage cervical cancer is curable, and more advanced tumors cannot be successfully treated despite the use of therapeutic methods that combine surgery and radiochemotherapy [2,3]. Because the molecular determinants of cervical cancer are still not thoroughly understood, developing more effective therapies remains difficult. Hence, there is an urgent need to know the molecular basis of cervical carcinogenesis.
Covalently closed circular RNAs (circRNAs) are naturally occurring RNA biomolecules without 3′ polyadenylate tails or 5′ caps [4]. Insights into the potential functions of circRNAs on gene expression by inhibiting microRNA (miRNA) activity highlight their essential involvement in the biological basis of cancers, including cervical cancer [5][6][7]. For instance, Song et al. illuminated that circ_101996, an overexpressed circRNA in cervical cancer, worked as a strong promoter of the disease via miR-8075-dependent regulation of TPX2 microtubule nucleation factor [8]. Tang et al. uncovered that circ_0000515 contributed to cervical carcinogenesis by binding to miR-326 to induce ETS transcription factor ELK1 [9]. As for circ_0018289, generated from the exons of synaptotagmin 15 (SYT15) mRNA, it was demonstrated to be upregulated in cervical cancer, highlighting its potential as a promising biomarker for the outcome of patients [10]. Previous work also showed the oncogenic activity of circ_0018289 in cervical cancer depending on the modulation of miR-497 [11]. Nevertheless, our understanding of the molecular basis underlying the modulation of circ_0018289 in cervical carcinogenesis has remained incomplete.

Human samples and cells
Fresh-frozen samples of cervical cancer (n = 50) and matched normal cervical tissues (n = 50) were obtained with written informed consent from individuals undergoing cervical resection in the Second Affiliated Hospital of Chongqing Medical University. Samples were analyzed for the expression levels of circ_0018289, miR-183-5p, and TMED5. Approval to collect human samples was granted by the Ethical Committee of the Second Affiliated Hospital of Chongqing Medical University.

Plasmids
Human TMED5 (Accession: NM_001167830.2) coding sequence (lacking the 3'UTR region) incorporated with BamH I and EcoR V sites in two terminals and the scrambled control sequence were synthesized by BGI (Shenzhen, China) and individually inserted into the pcDNA3.1 vector (Invitrogen, Wesel, Germany) opened with BamH I and EcoR V sites to generate TMED5 overexpression plasmid (pcDNA-TMED5) and negative control (pcDNA). The fragments of human circ_0018289 and TMED5 3′UTR encompassing the target sequence for miR-183-5p or mutated target sites were synthesized by BGI and individually cloned into the 3′UTR of Renilla luciferase in a psiCHECK-2 vector (Promega, Milano, Italy) opened with Xho I and Pme I sites to generate luciferase reporter constructs.

Transient transfection of cells
SiHa and HeLa cells were plated at 5 × 10 5 cells/well in 24-well cell culture dishes 18 h before transfection with synthetic circ_0018289-selected siRNA (si-circ_0018289) or control shRNA (si-NC) at 100 nM, mature miR-183-5p mimic or mimic control (miR-NC mimic) at 50 nM, or the antisense oligonucleotide of the mature miR-183-5p sequence designed to silence miRNA (anti-miR-183-5p) or nontarget control sequence (anti-miR-NC) at 50 nM using the HiPerFect transfection reagent as per the manufacturing recommendations (Qiagen, Tokyo, Japan). The details of all oligonucleotides (Ribobio, Guangzhou, China) were in Supplement Table 1. For TMED5 overexpression, 5 × 10 5 SiHa and HeLa cells seeded in each well of 24-well cell culture dishes were transiently transfected with pcDNA-TMED5 or pcDNA plasmid at a final dose of 200 ng using Lipofectamine 3000 as recommended by the manufacturers (Invitrogen). Transfected cells were collected for further assays 2 days post-transfection.

Quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA, inclusive of the small RNA fraction, was prepared from homogenized tissues and cultured cells with Trizol reagent as recommended by the manufacturers (Invitrogen). For circ_0018289 and mRNA quantification, cDNA was randomly primed with ReverTra Ace RT Kit (Toyobo, Tokyo, Japan) in a 25-μL reaction containing 500 ng of extracted RNA. qRT-PCR with iQ SYBR Green (Bio-Rad, Munich, Germany) and amplification primers (shown in Supplementary Table 1) was done on a PCR machine (Rotorgene 6000, Qiagen) with the following conditions: after a denaturation time of 10 min at 95 °C, 40 cycles at 95 °C for 20 s and at 60 °C for 1 min. For miR-183-5p quantification, cDNA preparation and qRT-PCR were carried out as above, using miScript RT Kit and SYBR Green Kit (all from Qiagen), respectively. Fold changes in gene and miRNA expression were determined by the 2 −ΔΔCt method, normalizing the results to normal controls and expression of β-actin or U6. ΔCt was calculated by subtracting the Ct values of β-actin or U6 from the Ct values of the gene of interest. ΔΔCt was then calculated by subtracting ΔCt value of the control from ΔCt of the sample.

Ribonuclease (RNase) R assay
This experiment was performed by incubating 3 μg of extracted RNA with 10 U of RNase R (Epicenter Biotechnologies, Madison, WI, USA) at 37 °C for 20 min. After being purified by the Trizol RNA Purification Kit (Invitrogen), RNA was analyzed by qRT-PCR analysis for quantification of circ_0018289 and SYT15 mRNA levels as above.

Subcellular localization assay
For the preparation of nuclear and cytoplasmic RNA of SiHa and HeLa cells, a Cytoplasmic and Nuclear RNA Purification Kit was used as described by the manufacturers (Norgen Biotek, Thorold, ON, Canada). For localization analysis, U6 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as the nuclear and cytoplasmic control, respectively. The isolated RNA was analyzed by qRT-PCR analysis for assessment of circ_0018289, U6, and GAPDH levels.

Cell proliferation assay
For the evaluation of cell proliferation, 5-ethynyl-2′deoxyuridine (Edu) labeling and Cell Counting Kit-8 (CCK-8) assays were performed. Briefly, for Edu assay, transfected cells were incubated with Edu solution (50 μM, Ribobio) for 2 h, and subsequently, the Edulabeled cells were stained with 1 × Apollo567 (Ribobio) for 30 min. After being staining with 4′,6-diamidino-2-phenylindole (DAPI, Invitrogen) for nuclei staining, the cells were observed under a BZ-8000 fluorescence microscope (Keyence, Osaka, Japan). The ratio of Edu-positive nuclei (red) to total nuclei (blue) was determined as the proliferation rate of transfected cells in 10 randomly captured fields per well. For CCK-8 assay, ~ 2500 transfected cells were plated into each well of 96-well cell culture dishes and incubated in complete medium at 37 °C. The cells were evaluated for potential growth every 24 h using the CCK-8 solution (10 μL per well) as per the manufacturing guidance (Dojindo, Kumamoto, Japan). After addition of CCK-8 solution, a 2-h incubation was allowed at 37 °C. A 96-well spectrometer (BMG Labtech, Ortenberg, Germany) was used to measure the absorption with a 450 nm filter.

Flow cytometry for cell apoptosis
Following the wash in cold phosphate-buffered saline (Solarbio, Beijing, China), transfected cells (5 × 10 5 ) were incubated with propidium iodide (PI, 50 mg/mL, Invitrogen) and Annexin V-fluorescein isothiocyanate (10 μL, Annexin V-FITC, BD Pharmingen, Heidelberg, Germany) as per the accompanying protocols. Following a 15-min incubation in the dark, data were analyzed within 1 h using an Epics XL-MCL flow cytometer (Beckman Coulter, Fullerton, CA, USA) with FlowJo 9.1 software (Tree Star, Ashland, OR, USA). Apoptotic cells were defined as the population that was positive for Annexin.

Tube formation assay
Transfected cells (1 × 10 5 cells per well) were seeded in 24-well cell culture dishes and incubated in complete medium at 37 °C. When the cells reached approximately 70% confluence, the complete medium was replaced by the serum-free EMEM medium. Following a 48-h culture at 37 °C, the conditional medium was harvested by centrifugalization. Subsequently, about 2 × 10 4 HUVECs were suspended in 200 μL of the conditional medium and then plated in 24-well cell culture dishes precoated with Matrigel Basement Membrane Matrix (BD Biosciences, Cowley, UK). Images of tube morphology were photographed using an inverted 100 × microscope (Keyence), and the tube formation was monitored every 6 h. The formed tubes were determined by counting the number of meshes using ImageJ software.

Dual-luciferase reporter assay
Luciferase reporter constructs (200 ng) were transfected using Lipofectamine 3000 into SiHa and HeLa cells seed in 24-well cell culture dishes (5 × 10 5 cells/well) together with synthetic miR-183-5p mimic or control mimic at 50 nM. Luciferase activity was gauged after 2 days using a 10 μL of cell lysates with the Luc-Pair miR Luciferase Assay Kit (GeneCopoeia, Rockville, MD, USA) on a luminometer (Promega) as per the accompanying guidance.

Generation of stable circ_0018289 silencing cell line
Lentiviruses expressing shRNA-circ_0018289 (sh-circ_0018289) or scrambled shRNA (sh-NC) were provided by Geneseed (Guangzhou, China) and used to infect SiHa cells in media containing polybrene (8 μg/mL, Solarbio). A stable cell line was selected using 2 mg/mL of puromycin for over 7 days.

Mouse xenograft experiments
All experiments with mice conformed to the protocols approved by the Animal Use Committee of the Second Affiliated Hospital of Chongqing Medical University, and all mice were cared following National Institutes of Health guidelines. For the formation of xenograft tumors, 4-to 5-week-old female BALB/c nude mice (Gempharmatech Biotechnology Co., Ltd., Jiangsu, China) were implanted subcutaneously with sh-NC-or sh-circ_0018289-transduced SiHa cells (5 × 10 6 cells per mouse) in 200 μL of cell culture medium (n = 6 mice per group). Tumor growth was monitored weekly by caliper measurements and tumor volume was determined by the equation D × d 2 /2, where D was the longest diameter of the tumor and d was the shortest diameter. Mice were sacrificed at day 35 after cell implantation with CO 2 overdose and the tumors were harvested for weight and expression analysis. Proliferation of tumors was evaluated with paraffin embedded tumor sections (4 μm) by immunohistochemistry using a monoclonal antibody against Ki67 (ab16667, Abcam, dilution 1:200), Vectastain ABC Kit (Vector Laboratories, Burlingame, CA, USA), and 3,3′-diaminobenzidine (DAB) substrate (Vector Laboratories) as described [15].

Statistical analysis
For descriptive statistics, data were presented as mean ± standard derivation from at least 3 independent biological replicates (performed in quadruplicate). Statistical analysis was done on the Prism 7 software (Graph-Pad, San Diego, CA, USA). P values were determined by Student's two-tailed t-test (comparison in two groups), Mann-Whitney U test (comparison in human samples), or one-way analysis of variance with Tukey-Kramer multiple comparison test (comparison in multiple groups), with P < 0.05 accepting significance. The Pearson's correlation coefficients were used to determine the correlations among circ_0018289, miR-183-5p, and TMED5 expression levels in human tumors.

Circ_0018289 was overexpressed in cervical cancer
To evaluate the role of circ_0018289 in cervical cancer, we firstly analyzed its expression in 50 primary tumor samples by qRT-PCR analysis. Circ_0018289 was markedly upregulated in tumor samples compared with paired normal tissues (Fig. 1A). In agreement with tumor samples, cancer cells showed higher levels of circ_0018289 compared to ectocervical Ect1/E6E7 cells (Fig. 1B). We then determined the stability of circ_0018289 in SiHa and HeLa cells by RNase R assay. Circ_0018289, rather than the corresponding linear SYT15 mRNA, was resistant to RNase R (Fig. 1C). Furthermore, circ_0018289 was mainly present in the cytoplasm of SiHa and HeLa cells, which was confirmed by subcellular localization assay (Fig. 1D).

Circ_0018289 silencing impeded cell proliferation and enhanced cell apoptosis in vitro
To elucidate the functional action of circ_0018289 in cervical carcinogenesis, we used siRNA-circ_0018289 (si-circ_0018289) in SiHa and HeLa cells to knock down its expression ( Fig. 2A). As expected, si-circ_0018289 transfection inhibited circ_0018289 expression by > twofold (Fig. 2A). Analyses of CCK-8 and Edu showed that circ_0018289 loss of function remarkably impeded cell proliferation (Fig. 2B, C). Conversely, silencing endogenous circ_0018289 promoted cell apoptosis (Fig. 2D). Additionally, western blot analysis revealed that circ_0018289 knockdown increased pro-apoptotic protein Bax level and reduced anti-apoptotic protein Bcl-2 expression in both cell lines (Fig. 2E), reinforcing that circ_0018289 loss of function enhanced apoptosis.

Circ_0018289 silencing suppressed angiogenesis in vitro
We next examined the influence of silencing endogenous circ_0018289 on angiogenesis in vitro. We firstly transfected SiHa and HeLa cells with si-circ_0018289 or nontarget-shRNA (si-NC) and then treated human HUVECs with the conditional medium of transfected cells. Strikingly, the incubation of conditional medium of cells with circ_0018289 loss of function led to an inhibition of tube formation of HUVECs (Fig. 2F), demonstrating that circ_0018289 silencing suppressed angiogenesis in vitro. Additionally, circ_0018289 knockdown decreased the levels of angiogenesis inducers VEGFA and FGF2 in the two cancer cell lines (Fig. 2G).

Circ_0018289 directly targeted miR-183-5p
To further elucidate the mechanism by which circ_0018289 regulated cervical carcinogenesis and angiogenesis, we identified its targeted miRNAs. Computer algorithm circInteractome predicted that circ_0018289 harbored a region that was partially complementary to miR-183-5p (Fig. 3A). To verify this finding, we cloned circ_0018289 segment containing the seed sequence into a luciferase vector and cotransfected the reporter into SiHa and HeLa cells with miR-183-5p mimic. The transfection efficiency of miR-183-5p mimic was examined by qRT-PCR (Fig. 3B). Remarkably, the reporter construct with the binding sites for miR-183-5p was repressed by miR-183-5p overexpression (Fig. 3C). We then generated a mutant in the binding sequence (Fig. 3A) and tested it. As expected, the mutant encompassing a mutated binding region was refractory to repression of miR-183-5p (Fig. 3C). Moreover, we observed a clear augmentation (> twofold) in the level of the endogenous miR-183-5p in circ_0018289-silencing cells (Fig. 3D). Contrary to the level of circ_0018289, miR-183-5p was markedly underexpressed in tumor samples and cancer cells compared with their counterparts (Fig. 3E, F). Intriguingly, there existed an inverse correlation between miR-183-5p and circ_0018289 levels in tumor samples (Fig. 3G). These data together indicated that circ_0018289 targeted miR-183-5p by directly binding to miR-183-5p.

TMED5 was directly targeted and inhibited by miR-183-5p
To identify downstream targets of miR-183-5p, we used computational prediction software starBase. Interestingly, a putative binding region for miR-183-5p was found in the 3'UTR of TMED5 (Fig. 5A), which was shown to be associated with the prognosis of cervical cancer patients using the GEPIA database (http:// gepia. cancer-pku. cn/, Fig. 5B). To establish the direct relationship between TMED5 and miR-183-5p, we generated TMED5 3′UTR wild-type (TMED5 3′UTR-WT) and mutant (TMED5 3′UTR-MUT) reporter constructs and analyzed them by dual-luciferase reporter assays. Overexpression of miR-183-5p by miR-183-5p mimic transfection remarkably repressed reporter gene expression of TMED5 3′UTR-WT, and the mutation abrogated the suppression of miR-183-5p (Fig. 5C). Moreover, endogenous TMED5 mRNA and protein levels were strongly promoted by miR-183-5p depletion and inhibited as a result of miR-183-5p overexpression in the two cancer cell lines (Fig. 5D, E). Analysis of TMED5 expression in cervical cancer showed that TMED5 mRNA and protein levels were significantly promoted in tumor samples and cancer cells compared with the corresponding normal controls (Fig. 5F-I). Importantly, TMED5 mRNA level inversely correlated with miR-183-5p expression and positively correlated with circ_0018289 expression in tumor samples (Fig. 5J, K). All these data established the notion that miR-183-5p regulated TMED5 expression by the perfect binding sites in the 3′UTR.

MiR-183-5p-mediated inhibition of TMED5 impacted cervical cancer development and angiogenesis in vitro
In order to demonstrate whether inhibition of TMED5 by miR-183-5p was responsible for the regulatory effects of miR-183-5p on cervical carcinogenesis and angiogenesis, we adopted a rescue experiment by increasing TMED5 expression in SiHa and HeLa cells. The transfection efficiency of TMED5 overexpression plasmid (pcDNA-TMED5) was gauged by qRT-PCR and western blot assays (Fig. 6A, B). As would be expected, elevated expression of TMED5 significantly abrogated miR-183-5p overexpression-mediated anti-proliferation (Fig. 6C, D) and pro-apoptosis (Fig. 6E, F) effects in the two cancer cell lines. Furthermore, elevated expression of TMED5 reversed the suppression of tube formation of miR-183-5p overexpression in HUVECs (Fig. 6G). Additionally, elevated expression of TMED5 abrogated miR-183-5p overexpression-mediated inhibition on VEGFA and FGF2 levels in both SiHa and HeLa cell line (Fig. 6H).

Circ_0018289 regulated TMED5 expression by targeting miR-183-5p
Next, we asked whether circ_0018289 could operate as a regulator of TMED5. Notably, circ_0018289 loss of function resulted in reduced levels of TMED5 mRNA and protein in SiHa and HeLa cells (Fig. 7A, B). However, this effect was remarkably abolished by miR-183-5p depletion (Fig. 7A, B).

Circ_0018289 silencing diminished tumor growth in vivo
To determine whether circ_0018289 silencing possessed tumor-inhibitory effects, we reduced circ_0018289 expression with a lentiviral delivery system (sh-circ_0018289) in SiHa cells and then implanted the cells subcutaneously in the nude mice. By contrast, transduction with sh-circ_0018289 led to markedly decreased volume and weight of the tumors (Fig. 8A, B). Moreover, the tumors formed by sh-circ_0018289-transduced SiHa cells exhibited lower levels of circ_0018289 compared with the controls (Fig. 8C). qRT-PCR and western blot analyses of the tumors showed that circ_0018289 silencing resulted in increased expression of miR-183-5p (Fig. 8D) and reduced levels of TMED5 mRNA and protein (Fig. 8E, F). Additionally, tumor-inhibitory result of circ_0018289 silencing was also confirmed by staining for cell proliferation using the cell-cycle marker Ki67 (Fig. 8G). These results together suggested that the inhibition of tumor growth might be due to downregulation of circ_0018289 and TMED5 and upregulation of miR-183-5p.

Discussion
CircRNAs are increasingly implicated in regulating human carcinogenesis [16][17][18]. Previous work uncovered the overexpression and tumor-promoting activity of circ_0018289 in cervical cancer [10,11]. We here extended these findings by identifying a novel circ_0018289/miR-183-5p/TMED5 regulatory network in cervical cancer development. Taken together, we proposed that circ_0018289 might be a potent oncogene in cervical tumorigenesis. As previously reported for other circRNAs [19,20], circ_0018289 was resistant to RNase R due to the back-spliced formation and the lack of 3′ and 5′ ends [4]. Additionally, circ_0018289 was mainly present in the cytoplasm, providing the possibility for its relationship with mature miRNAs [21]. Emerging evidence has suggested that alterations in miRNA expression might prove crucial in affecting cervical tumorigenesis [22][23][24]. Here, we focused on miR-183-5p because of the conflicting roles of miR-183-5p in human carcinogenesis [25][26][27]. These contradictory conclusions might partially due to the different types of tumors in these reports, where miR-183-5p facilitated the development of breast cancer [26] and hepatocellular carcinoma [27] and impeded endometrial tumorigenesis [25]. Here, we first elucidated that circ_0018289 directly targeted miR-183-5p, which was established as a strong suppressor in cervical cancer [12,13]. Our data also showed the regulation of circ_0018289 in cervical cancer development and angiogenesis through miR-183-5p.