A Novel Circular RNA circTADA2A Promotes Proliferation And Metastasis Of Ovarian Cancer Through Sponging miR-203


 BackgroundCircular RNAs (circRNAs), a new type of non-coding RNA, have been demonstrated to play critical roles in the progression of various of malignant cancers. In the present study, we identified a circRNA termed as circTADA2A which was hypothesized that may be significantly up-regulated in OC tissues and cell lines.ResultsThe results revealed that ectopic expression of circTADA2A promoted the cell proliferation, migration, invasion and colony formation ability of OC cells. Constantly, silencing of circTADA2A inhibited those of OC cells. Furthermore, we identified that circTADA2A was able to target miR-203 in OC cell. MiR-203 was able to reverse the oncogenic effect of circTADA2A on proliferation, migration and metastasis of OC cells through targeting SMAD1.ConclusionsWe reported that circTADA2A served as a competing endogenous RNA (ceRNA) to sponge miR-203 and blocked its regulation of SMAD1. These findings provide insights into OC progression and also potential new targets for diagnose or treatment of OC.

CircRNAs primarily sponge miRNAs to exert an impact on not only the gene expression but also biological functions [10,12,14].
In the current research work, we gured out an up-regulated circular RNA hsa_circ_0043278 (circTADA2A) in OC, together with nding its up-regulated expression level in both OC tissues and cell lines. The loss and gain of function studies were carried out for the determination of the function of circTADA2A in tumorigenesis of OC. Mechanistically, we not only predicted but also veri ed the target miRNA of circTADA2A, which is miR-203, in addition to further elucidating the underlying mechanism of circTADA2A in OC. To summarize, circTADA2A works as an oncogene in OC by sponging miIR-203, besides being likely to constitute a potential therapeutic target for OC.

Results
The expression of circTADA2A was upregulated in OC tissues and cells Aimed at exploring the role of circRNAs in the progression of OC, we rst performed RNA sequencing. Figure 1A showed the most signi cantly up-regulated and down-regulated circRNAs, among which we selected has_circRNA_102049 termed as circTADA2A. As we know, circRNAs have much more stability for their looped structure. Accordingly, for the purpose of con rming the stability of circTADA2A, RNase R was employed in the experiments. As the results suggested, following the RNase R treatment, the linear forms of TADA2A showed an evident decline while no signi cant change was observed in the level of circTADA2A, indicating the stability of circTADA2A ( Figure 1B). Also, OC cells were treated with Actinomycin D in order to inhibit the transcription after which the half-life of circTADA2A was assessed ( Figure 1C). The results shed light on a longer half-life of circTADA2A as compared with that of TADA2A mRNA. Besides that, a speci c probe was designed and made to carry out the FISH experiment. Figure 1D demonstrates that circTADA2A located primarily in cytoplasma. We detected the expression level of circTADA2A with the use of qPCR. The results illustrated that circTADA2A was remarkably up-regulated in the OC tissues and cells ( Figure 1E, F).

CircTADA2A promoted the proliferation and migration of OC cells
Owing to the modi ed expression in OC tissues and cell lines, we speculated that circTADA2A plays quite a crucial role in the progression of OC. Thus, the gain and loss of function studies were carried out with the use of overexpressing and shRNA vectors for circTADA2A. The qPCR analysis veri ed the overexpressing or knock down e cient of the overexpressing and shRNA vectors for circTADA2A, correspondingly ( Figure 2A). Subsequently, as the MTT results suggested, the circTADA2A overexpression promoted cell proliferation while the circTADA2A knock down inhibited that of OC cells ( Figure 2B). Colony formation was further carried out. As expected, the circTADA2A overexpression augmented the colony formation capability, while silencing of circTADA2A lowered it in OC cells ( Figure 2C). Moreover, transwell assay and wound healing were carried out for the evaluation of the invasion and migration ability of OC cells. As the results suggested, circTADA2A overexpression augmented the invaded cell number in transwell assay ( Figure 2D) as well as the migrated distance in wound healing assay ( Figure  2E). Contrastingly, the silencing of circTADA2A lowered that capability of OC cells as compared with the shRNA control group.

CircTADA2A exerts its function by sponging miR-203
It is well known that circRNAs sponge miRNAs for regulating the downstream gene expression for the purpose of participating in regulating the cell physiology mechanism. We carried out the bioinformatics analysis software circinteractome, aimed at predicting the potential miRNA target of circTADA2A, followed by selecting miR-203. Figure 3A demonstrated the complementary base sequences between circTADA2A and miR-203. The luciferase activity assay was carried out for the veri cation of the binding of circTADA2A and miR-203. We discovered the fact that miR-203 lowered the luciferase activities of the wild-type reporter for circTADA2A but not that of the mutant-type reporter, con rming miR-203 as a sponge target of circTADA2A ( Figure 3B). Subsequent to that, we carried out the circTADA2A overexpressing and shRNA vector transfection. Interestingly, circTADA2A overexpression lowered the level of miR-203 while the circTADA2A knock down augmented the same ( Figure 3C). In addition, Pearson analysis was used for evaluating the relationship between circTADA2A and miR-203; as the results suggested, there was a signi cant negative relationship existing between circTADA2A and miR-203 ( Figure 3D). The RNA pull-down was further performed for the purpose of con rming the binding of circTADA2A and miR-203. By means of the results, we gured it out that the enrichment of miR-203 was substantially higher in the circTADA2A probe pull down product in comparison with that of the control probe ( Figure 3E). Accordingly, more circTADA2A was captured by the biotin-labelled miR-203 groups as compared with the control probe ( Figure 3F). These ndings suggested miR-203 as a sponge targets of circTADA2A in OC cells.

MiR-203 reversed the function of circTADA2A in OC cells
Being a target of circTADA2A, we predicted that miR-203 had been involvement in the biological mechanism of OC cells as well. We carried out the miR-203 and circTADA2A overexpressing vector transfection. Co-transfection with miR-203 and circTADA2A overexpressing vector apparently augmented the level of miR-203 in comparison with circTADA2A overexpressing group ( Figure 4A). By means of the function studies, we, at rst, discovered that miR-203 reversed not only the cell proliferation (Figure 4 B), but also the colony formation ability ( Figure 4C) invitation ( Figure 4D) as well as migration ( Figure 4E, F) of OC cells. Together with that, circTADA2A reversed these functions of miR-203 partially, further indicating the relationship existing between circTADA2A and miR-203.

MiR-203 target SMAD1 in OC cells
MiRDB, DIANA and Targetscan databases were used to predict the targets of miR-203, among which SMAD1 was selected as a potential one for its existence in all the three databases ( Figure 5 A). Figure 5 B showed the target region between miR-203 and SMAD1. Luciferase activity assay in both SKOV-3 and OVCAR-3 cells were carried out. As indicated by ndings, miR-203 evidently down-regulated the luciferase activity co-transfected with pGL3-3'UTR of SMAD1, but not the pGL3-3'UTR-mut ( Figure 5 C). RNA-pull down experiments were carried out, the results indicated that miR-203 directly target SMAD1 ( Figure 5D). In addition, qPCR and western blot analysis further con rmed that miR-203 overexpression decreased the level of SMAD1 while knock down of miR-203 increased that of SMAD1 ( Figure 5 E, F). Pearson analysis revealed a negative relationship between miR-203 and SMAD1 ( Figure 5G) which further veri ed the target relation between miR-230 and SMAD1. In order to con rm the ceRNA relation between circTADA2A and SMAD1, we evaluated whether circTADA2A involved in the regulation of SMAD1 expression. The results veri ed our speculation. CircTADA2A notably promoted SMAD1 expression ( Figure 5H) and Pearson analysis indicated a positive relation between circTADA2A and SMAD1 which further indicated their relationship ( Figure 5I).
CircTADA2A regulated AKT/GSK3β signal pathway and EMT proteins expression As SMAD1 was con rmed to be the target of miR-203, we focused on the downstream gene which may be regulated by SMAD1. Western blot was carried out to evaluate the speci c protein expression. As gure 6 A and B revealed. CircTADA2A signi cantly promoted expression level of SMAD1, N-cadherin, Vimentin, MMP-9 along with the phosphorylation of AKT and GSK3β. Meantime, circTADA2A inhibited the level of E-cadherin. These regulatory effect of circTADA2A was reversed by both miR-203 expression and knock down of SAMD1 indicating the target relation between circTADA2A and miR-203 along with the ceRNA relationship between circTADA2A and SMAD2.

CircTADA2A promoted the growth of OC cells in vivo
Thereafter, we carried out the in vivo study, aimed at further investigating the biological function of circTADA2A in the progression of OC. Stable SKOV-3 cell line was established with the transfection of sh-circTADA2A or circTADA2A vectors along with their negative control, correspondingly. Both the size and weight of the tumours in the circTADA2A were observed as higher in comparison with the control tumours. Contrarily, the circTADA2A deletion lowered the tumour size and weight of tumours in comparison with the control group ( Figure 7A-C). IHC assay, detecting the expression of ki67, the proliferation biomarker, was also carried out. In accordance with the expectations, circTADA2A overexpression promoted the expression of ki67 and SMAD2 while the knock down of circTADA2A inhibited that of ki67 and SMAD2 ( Figure 7D).

Discussion
CircRNAs have been fully documented as playing critical roles in the cellular functions, for instance, proliferation, apoptosis, differentiation, and metabolism [6,11]. In the current research work, we gured out an innovative circRNA, termed as circTADA2A. The stability of circTADA2A was con rmed by its stable expression subjected to the Rnase R digestion. To the best of our understanding, this is the rst study dealing with the circTADA2A emphasizing its modi ed expression as well as biological impact. The up-regulated expression and stability of circTADA2A make it a potential biomarker as a diagnose and therapeutic target.
It is well known that circRNAs mainly act as the sponges of miRNAs, thus block their regulatory effect on the downstream gene expression. Among them, miR-203 has been extensively investigated and proved to play crucial role in the cellular differentiation and development. Several research works shed light on the anti-cancer function of miR-203 in OC. It was con rmed to inhibit proliferation and metastasis of OC cells, meantime, increase apoptosis of OC cells [15][16][17]. Recently, miRNAs were found to be sponged by circRNAs and lncRNAs that further block their regulatory effects on cell process. For instance, CircAGFG1 sponges miR-203 to promote EMT and metastasis of non-small-cell lung cancer by upregulating ZNF281 expression [18]. HCP5 is a SMAD3-responsive long non-coding RNA that promotes lung adenocarcinoma metastasis via miR-203/SNAI axis [19]. However, the relationship between cicRNA and miR-203 in OC has not yet been dicussed. In the present study, we predicted and veri ed that miR-203 was sponged by circTADA2A which could block its anti-cancer role. To further elucidate the mechanism of miR-203, we predicted the target genes through three prediction software including Targetscan, miRDB and DIANA and get 564 genes exist in all 3 datasets. Among them, SMAD1 was selected and veri ed as the target of miR-203.
SMA and mother against decapentaplegic (MAD)-related proteins (SMADs) are intracellular components of TGF-ß signalling pathway. containing eight members as SMAD1-Smad8 respectively [20][21][22][23]. SMAD1 is a critical inducer of the EMT process. SMAD1/Akt/GSK3β signalling pathway regulated by snail altered Nanog status during EMT [24]. Another EMT transcription factor Twist1 also plays the role as a downstream factor of SMAD1 [25]. In the present research, we predicted and con rmed that miR-203 target SMAD1 in SKOV-3 and OVCAR-3 cell. Due to the critical role of SMAD1 in EMT, we thus evaluated the EMT proteins such as E-cadherin, Vimentin, MMP-9 and N-cadherin and found that circTADA2A protmoted SMAD1 expression along with N-cadherin, Vimentin, MMP-9 expression. This effect of circTADA2A can be reversed by both miR-203 overexpression and knock down of SMAD1 indicating the ceRNA relation between circTADA2A and SMAD1. However, whether SMAD1 directly regulated EMT proteins expression or via other key molecular remain not elucidated. SMAD1 was reported to be involved in the AKT signal pathway. For instance, SMAD-PI3K-Akt-mTOR pathway mediates BMP-7 polarization of monocytes into M2 macrophages. It was gured out that SMAD1 was capable of promoting the phosphorylation of AKT and GSK3β. We noticed that AKT was also demonstrated to regulate SMAD1 expression. AKT is involved in the cell proliferation especially in cancer cells [26,27]. Whether there exist a negative loop regulation mechanism between AKT and SMAD1 remain unclear which is on our further exploration plan.
In the current research work, circTADA2A was identi ed as a novel circRNA which play a oncogenetic role in OC. MiR-203 was determined to be the target miRNA of circTADA2A. Interestingly, a recent study also gured out the effect of circTADA2A and miR-203a in osteosarcoma progression. CircTADA2A promote osteosarcoma progression through sponging miR-230a and alter CREB3 expression [28]. Aimed at extending the understanding of circTADA2A and miR-203 in the develop of human cancer, we predicted the target gene of miR-203, besides nding SMAD1 as a potential one which have a close association with the EMT in human cancers and has not reported before. CircTADA2A/miR-203/SMAD1 signal axis was gure out which is likely to be involved in the regulation of cell phenotype including proliferation and metastasis of OC cells.

Conclusions
To conclude, our ndings shed light on the fact that circTADA2A promotes the malignant behaviour of OC by means of sponging miR-203, which indicates its potentiality to be a diagnostic or therapeutic target of OC.

Methods And Materials
Patients A total number of 25 paired OC and the adjacent tissues were collected from The People's Hospital of Linan City between May 2016 and Nov 2016. The samples were stored in liquid nitrogen immediately after surgery. This study has been proved by the ethical committee of The People's Hospital of Linan City and written informed consent was obtained from each patient.

RNA-seq analysis
The total RNAs were extracted from 3 OC tissue and 3 normal tissue by TRIzol (Invitrogen). The tissue samples were then sent to Vazyme Biotech (Nanjing, China) for the sequencing analysis. Overexpressing vector for CircTADA2A and pre-miR-203, control vector, short hairpin RNA (shRNA) targeting circTADA2A and its negative control were purchased from Ribobio (Guangzhou, China). Cells were seeded on 6-well plates at the density of 8×10 5 , followed by getting cultured in RPMI-1640 without FBS at a temperature of 37°C for 12 h prior to the transfection. Cell transfection and co-transfection were carried out with the use of Lipofectamine 2000 (Invitrogen, CA, USA) in accordance with the manufacturer's instructions. 6 h following the transfection, the cell culture medium was replaced with RPMI-1640 medium, which was supplemented with 10% FBS. MTT 24 h following the transfection, cells were collected and resuspended in culture medium, followed by seeding on 96-well plates at a density of 1x10 4 cells/well. Moreover, 10 µl CCK-8 solution was added to each well and incubated for another 2 hours at a temperature of 37°C. Absorbance at 480 nm was detected using a microplate Reader (Biorad, CA, USA). Each sample was performed for a minimum of 3 times.
Transwell assay Subsequent to the transfection, 1×10 5 of cells suspended in 200μL RPMI-1640 medium were seeded onto the upper transwell chamber (Corning, NY, USA). Following 24 hours of incubation at 37°C, the cells on the surfaces of lower chambers were xed using 20% methanol followed by staining with 1% crystal violet (Beyotime, Shanghai, China). Eventually, the cell colonies were both photographed and counted.
Wound healing assay Subsequent to the transfection, the cells were seeded onto 6 well plates followed by culturing in serum free medium for a period of 24 hours. Thereafter, cell monolayers were wounded using a 10μl pipette tip.
After that, cells were replaced with fresh medium and cultured at 37°C. The pictures were taken 48 after the wounding for the purpose of determining the wound-closing procedure.

Colony-formation assay
Following the transfection, the cells were seeded in 12 well plates at the density of 100/well. Subsequent to the incubation for 2 weeks, cells were xed in 10% formaldehyde, together with staining with 1% crystal violet (Beyotime, Shanghai, China). Images were photographed under a microscope (Leica, Germany), besides counting the colonies that contained more than 50 cells.
Xenograft model SKOV-3 cells were harvested and resuspended in RPMI-1640 medium. The nude mice were injected with a total number of 3 × 10 6 /100 μL cells subcutaneously at the posterior ank. In addition, tumour size was monitored through the measurement of the length (L) and width (W) using callipers every 3 days. Following a period of 28 days, the tumours were excised out from the sacri ced mice and weighed.
Real-time PCR Total RNA was extracted with the use of TRIzol reagent (Invitrogen, CA, USA) in accordance with the manufacturer's instructions. Afterwards, RNA was reverse transcribed to cDNA with the help of a PrimeScript RT Reagent kit (Takara, Dalian, China). The quantitative real-time PCR (qRT-PCR) analysis was carried out with the help of a SYBR-Green PCR Master Mix (Thermo Fisher, MA, USA) in a 7900HT PCR System (ABI, CA, USA). The use of GAPDH and U6 was made as internal controls for the mRNA and miRNA analysis, correspondingly. The relative expression levels were assessed with the use of the 2−ΔΔCt methodology. All of the reactions were carried out in triplicate.

Western blot
The total protein was extracted from cells with the help of the RIPA lysis buffer (Beyotime, Shanghai, China). The protein concentration was detected in a nanodrop system (Thermo Scienti c, MA, USA). 40ug protein was separated by 10% SDS-PAGE, followed by transferring to polyvinylidene di uoride membranes (Millipore, CA, USA). Subsequently, the blots were blocked using 5% non-fat milk for a period of 2 hours at the room temperature followed by the incubation at 4°C overnight with primary antibodies. Thereafter, the incubation of membranes was performed with a corresponding horseradish peroxidaseconjugated secondary antibody for 4 hours at the room temperature. The bands were detected with the use of an enhanced chemiluminescence solution (Pierce, MA, USA), together with imaging using the FluorChem imaging system (Alpha Innotech, San Leandro, CA, USA). The use of GAPDH was made as the internal control. All of the reactions were carried out in triplicate.

Fluorescence in situ hybridization (FISH)
Alexa Fluor 555-labeled circTADA2A probes were designed and synthesized by RiboBio (Guangzhou, China). FISH experiment was carried out using a uorescent in Situ Hybridization Kit (RiboBio, Guangzhou, China). 1×10 5 cells were seeded onto the autoclaved glass slides and cultured for a period of 24 hours. Subsequent to xing with 4% paraformaldehyde for 20 minutes, followed by the permeabilization with 0.5% Triton X-100 for 10 minutes, the cells were cultured at 37℃ overnight. Eventually, the incubation of slides was carried out with DAPI for the purpose of staining the cell nuclear and observed under a uorescence microscope (Leica, Wetzlar, Germany).

RNA pull down
The biotin labelled circTADA2A as well as miR-203 probes along with theri control probe were synthesized by Sangon Biotech (Shanghai, China). In addition, the probe-coated beads were generated through the coincubation of the probe with the streptavidin-coated beads (Invitrogen, CA, USA) at 25℃ for 2 hours. Cells were gathered and incubated with speci c probes overnight at 4℃. Following that, the beads were eluted and the complex was puri ed with TRIzol (Takara, Dalian, China). Afterwards, the abundance of both circTADA2A and miR-203 was analysed by the qRT-PCR.

IHC
The tumour tissues were xed in 4% paraformaldehyde for 24 hours, followed by dehydrating in a graded alcohol series and embedding in para n, followed by cutting into 5μm sections. The sections were depara nised, rehydrated with a graded alcohol series and then incubated in 96℃ with 0.01 mol/l sodium citrate buffer for the antigen retrieval. Following the incubation in 5% H 2 O 2 for a period of 2 hours, the sections were incubated using primary antibodies including ki67 and SMAD1 (Abcam, England) overnight at 4℃. Immunostaining was carried out with the use of streptavidin-peroxidase and diaminobenzidinef (DAB) following the manufacturer's instructions (Beyotime, Shanghai, China). Eventually, the sections were not only observed under a uorescence microscope (Leica, Wetzlar, Germany) but also imaged.

Luciferase reporter assay
Cells were transfected with miR-203 mimics or mimic control and co-transfected with pGL3 reporter vectors (Promega, CA, USA) that contained the wild-type (Wt) or mutated (mut) potential binding sequence of circTADA2A as well as SMAD1. 48 hours following the transfection, cells were gathered and analysed with the help of a Dual-Luciferase Reporter Assay kit (Promega, CA, USA). Luciferase activity was detected through the use of a GloMax uorescence reader (Promega, CA, USA). Renilla luciferase activities were put to use as an internal control. Each of the assays was carried out in a minimum of triplicate.

Statistical analysis
All data are presented as the mean s± standard deviation (SD). The statistical analyses were performed using SPSS 20 software (Abbott Laboratories, Chicago, IL, USA