Circular RNA hsa_circRNA_002178 silencing retards breast cancer progression via microRNA‐328‐3p‐mediated inhibition of COL1A1

Abstract Circular RNAs (circRNAs) are a group of non‐coding RNAs implicated in the pathogenesis of cancer progression, which exert their functions via regulation of microRNAs (miRNAs) and genes. The present study uses gain‐ and loss‐of‐function approaches to evaluate the functions of hsa_circRNA_002178 in angiogenesis along with energy metabolism and underlying downstream signals. The expression pattern of hsa_circRNA_002178 in clinical breast cancer tissues and its association with prognosis were characterized at first. Next, the energy metabolism and angiogenesis as well as cell viability were evaluated when the expression of hsa_circRNA_002178 in breast cancer cells was knocked down by siRNA. The interaction between hsa_circRNA_002178 and its downstream miR‐328‐3p was identified, followed by the analysis of their functions in regulation of breast cancer cellular behaviours. The target gene of miR‐328‐3p was predicted and verified, followed by identifying its role in the breast cancer progression. Higher expression of hsa_circRNA_002178 shared an association with worse prognosis in breast cancer. The inhibition of hsa_circRNA_002178 resulted in reductions in cell viability, energy metabolism and tube formation ability. Hsa_circRNA_002178 could competitively bind to miR‐328‐3p and down‐regulated its expression. Restoration of miR‐328‐3p eliminated the tumour‐promoting effects of hsa_circRNA_002178. COL1A1, as a target of miR‐328‐3p, could be up‐regulated by overexpression of hsa_circRNA_002178. In vivo experiments further confirmed the inhibition of tumour growth and inflammation by silencing hsa_circRNA_002178 or up‐regulating miR‐328‐3p. Taken together, hsa_circRNA_002178 is highlighted as a promising target for breast cancer due to the anti‐tumour effects achieved by silencing hsa_circRNA_002178.


| INTRODUC TI ON
Breast cancer ranks the first commonly diagnosed cancer in female population, with an estimated rate of almost 1 in every 4 female cancer cases in 2018. 1 Generally, tumour cells exhibit higher capacities of glucose utilization and oxidative metabolism than their non-tumour equivalents. 2 It greatly varies in the timing and distribution pattern of tumour metastases. Only a small quantity of patients with metastatic breast cancer has been timely diagnosed. 3 Successful neoadjuvant chemotherapy with trastuzumab alone or plus pertuzumab portends even greater benefit for the patients with early-stage HER2-positive breast cancer. 4 Even though the treatment options progressed in the recent decades, heterogeneity of metastatic breast cancer and specific molecular mutations in clinical situations have been identified, and personalized strategies are effective for only a few patients up to now. 5 Strikingly, it is currently acknowledged that circular RNAs (circRNAs) have implications in the pathogenesis of cancer progression, which highlighted their potential as novel biomarkers or therapeutic targets for breast cancer, 6 this brings an attractive focus to seek circRNAs-targeted strategies.
CircRNAs are a large group of non-coding RNAs that exert functions through modulating the activity of microRNAs (miRNAs) and gene expression. 7 A plenty of circRNAs have been identified by RNA-seq data in primary breast cancer, and a few of them have been suggested as important predictors, such as circCNOT2 and CREBBP. 8 Another circRNA hsa_circRNA_001783 has been reported to accelerate breast cancer progression by functioning as a sponge of miR-200c-3p. 9 Additionally, a circRNA circIRAK3 also acts as a promoter of breast cancer metastasis by sponging miR-3607. 10 miRNAs are non-coding RNA molecules that could mediate cell differentiation and survival. Some of them were suggested to be efficacious therapeutic candidates, such as miR-34 and miR-122. 11 Interestingly, miR-362-3p and miR-329 play critical roles as tumour suppressor in the progression of breast cancer via suppression in cellular proliferation and tumour growth. 12 In this study, hsa_circRNA_002178 is determined to be an up-regulated circRNA in breast cancer and a circRNA that could bind to miR-328-3p. Although it has been suggested that miR-328-3p can mediate androgen receptor (AR) in breast cancer cells, 13 the interaction between hsa_circRNA_002178 and miR-328-3p as well as the underlying mechanism is largely unknown. Based on six bioinformatic prediction databases, COL1A1 is identified as a target gene of miR-328-3p. Collagen type I alpha 1 chain (COL1A1), encoding type I collagen, associated with oestrogen/progesterone receptor (ER/PR) status, is proposed as a regulator that facilitates breast cancer metastasis. 14 In this regard, we hypothesized that hsa_cir-cRNA_002178 might regulate the behaviours of breast cancer cells through interacting with in miR-328-3p and COL1A1. Both in vitro and in vivo experiments were designed to characterize the role of hsa_circRNA_002178 and further identify the detailed interactions with miR-328-3p and COL1A1.

| Ethics statement
The study was conducted with the approval of the Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, and written informed consents were obtained from all patients or their guardians. Also, the animal experiments were conducted with the approval of the Animal Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine.

| RNA extraction and quantification
Total RNA was extracted from tissues or cells using the TRIzol kit   were imaged by a Bio-Rad gel imaging system, and protein expression relative to GAPDH was determined by Image J software (NIH).

| Basal oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) measurement
The OCR and ECAR were determined by Seahorse XF24 analyzer (Agilent Technologies) as previously described. 15 The cells were cultured in high-glucose DMEM containing 10% FBS and 5% glutamine.

| High-performance liquid chromatography (HPLC)
Cellular energy metabolism analysis was performed using SHIMADZU Shimadzu ultra-fast liquid chromatography to detect the contents of ATP, ADP, AMP, NADH and NAD. The treated cell samples were

| RNA pull-down
The binding of hsa_circRNA_002178 to miR-328-3p was ana-

| Dual luciferase reporter assay
The miR-328-3p putative binding sites were predicted using RNA22

| Tumour xenografts in nude mice
Breast cancer cells were stably transfected with si-hsa_cir-

| Immunohistochemistry
The excised mouse tumour tissues were fixed in 4% paraformaldehyde and embedded in paraffin, and sliced into 5-μm-thick sections.

| Enzyme-linked immunosorbent assay (ELISA)
Blood was collected from the ocular venous plexus of mice after deep anaesthesia and allowed to stand at 4°C for 1 hour. The serum was collected after centrifugation at 700 g and then stored at −80°C for subsequent use. The levels of inflammatory factors IL-6 and TNF-α were measured by Simple Step ELISA ® Mouse Kits for IL-6 (ab100712) and TNF-α (ab208348) form Abcam, and the OD value was read at 450 nm using an EON spectrophotometer (BioTek Instruments).

| Statistical analysis
All data were processed using SPSS 21.0 statistical software (IBM).
All data were tested for normal distribution and homogeneity.
Measurement data were expressed by mean ± standard deviation.
Data were compared using unpaired t test. Comparison between groups was performed by one-way analysis of variance (ANOVA) with Tukey post hoc test. Data at multiple time points were compared by repeated measurement ANOVA. A value of P < .05 indicates that the difference is statistically significant.

| The differential expression pattern of hsa_ circRNA_002178, miR-328-3p and COL1A1 in breast cancer
Differentially expressed circRNAs were screened from circRNA expression microarray data of breast cancer (GSE10 1123), which showed that hsa_circRNA_002178 expression in breast cancer tissue was notably higher than in the non-tumour breast tissues ( Figure 1A). hsa_circRNA_002178 is also called hsa_cir-cRNA_0000519 in the circBase database (http://www.circb ase. org/). The downstream miRNAs of hsa_circRNA_002178 were then predicted in databases circBank and starBase, and two intersection miRNAs were found in the predicted miRNAs from the two databases, named hsa-miR-1296-5p and hsa-miR-328-3p ( Figure 1B).
It was further suggested that miR-328-3p was underexpressed in breast cancer in the Pan-Cancer Analysis Platform of starBase database ( Figure 1C), suggesting that hsa_circRNA_002178 might function in the progression of breast cancer through regulating miR-328-3p. To further analyse the possible regulatory mechanisms of miR-328-3p, the target genes of miR-328-3p were predicted in RNA22, mirDIP, starBase, miRWalk, miRmap and TargetScan databases. The predicted target genes were compared, and Venn maps of those genes were plotted. As shown in Figure 1D, there were 122 intersection target genes that were highly likely to be regulated by miR-328-3p. In addition, differentially expressed genes were obtained from gene expression microarray data (GSE80754 and GSE10797), and these differentially expressed genes were intersected with the predicted target genes of miR-328-3p. Two intersection genes COL1A1 and PADI2 were identified ( Figure 1E).
COL1A1 was observed to be highly expressed in breast cancer in both GSE80754 ( Figure 1F) and GSE10797 ( Figure 1G), whereas PADI2 was lowly expressed. In addition, the high expression of   Figure 1H). Therefore, we hypothesized that the dysregulation of COL1A1 in breast cancer may be regulated by miR-328-3p.

| High expression of hsa_circRNA_002178 in breast cancer is associated with poor prognosis
To investigate the involvement of hsa_circRNA_002178 in breast experiments demonstrated that hsa_circRNA_002178 was highly expressed in breast cancer tissues and cells and associated with prognosis in breast cancer patients. Additionally, FISH was utilized to identify its subcellular localization of hsa_circRNA_002178, which showed that hsa_circRNA_002178 was localized in the cytoplasm of breast cancer cells ( Figure 2D).

| Silencing of hsa_circRNA_002178 impairs breast cancer cell proliferation, energy metabolism and angiogenesis
Given expression pattern of hsa_circRNA_002178 in breast cancer, the potential function of hsa_circRNA_002178 was examined by silencing hsa_circRNA_002178 in the MDA-MB-231 cells. RT-qPCR data exhibited that the expression of hsa_circRNA_002178 was successfully knocked down after transfection with either si-hsa_circRNA_002178-1 or si-hsa_circRNA_002178-2 (P < .05) ( Figure 3A) Considering si-hsa_circRNA_002178-1 had the highest knockdown F I G U R E 2 Hsa_circRNA_002178 was highly expressed in breast cancer tissues and cells, which was associated with poor prognosis. A, the expression of hsa_ circRNA_002178 normalized to GAPDH in breast cancer and adjacent normal tissues determined by RT-qPCR, n = 70. B, the expression of hsa_circRNA_002178 normalized to GAPDH in breast cancer cell lines and mammary epithelial cell line determined by RT-qPCR. C, survival rates of breast cancer patients with high or low hsa_circRNA_002178 expression by Kaplan-Meier analysis. D, localization of hsa_circRNA_002178 in breast cancer cells (×400). * P < .05 vs adjacent normal tissues or normal cells. The measurement data were expressed as mean ± standard deviation. Data between the two groups were analysed by unpaired t test, and among multiple groups were analysed using one-way ANOVA. The cell experiments were repeated three times

| Hsa_circRNA_002178 binds to miR-328-3p
Prediction results from circBank and starBase databases found that hsa_ circRNA_002178 may bind to miR-328-3p in breast cancer cells. As shown in Figure 4A, complementary sequences between hsa_circRNA_002178 and miR-328-3p were predicted to be potential binding sites. The expression of miR-328-3p in breast cancer tissues was lower than that in the matched adjacent normal tissues determined by RT-qPCR (P < .05; Figure 4B). The results of the RNA pull-down experiment showed that the biotinylated miR-328-3p-enriched hsa_circRNA_002178 content was significantly increased compared with the biotinylated NC-enriched one (P < .05; Figure 4C). The miR-328-3p mimic was transfected into the breast cancer cells expressing hsa_circRNA_002178-wt or hsa_ circRNA_002178-mut for dual luciferase reporter assay. The results showed that only the luciferase activity of the hsa_circRNA_002178-wt was decreased by miR-328-3p mimic transfection, indicating that miR-328-3p specifically bound to hsa_circRNA_002178 (P < .05) ( Figure 4D).

| Hsa_circRNA_002178 facilitates cancer progression in vitro through inhibiting miR-328-3p
To further investigate functional significance of the interaction be- Taken together, the data supported the conclusion that miR-328-3p can eliminate the promotive effect of hsa_circRNA_002178 on energy metabolism and angiogenesis.

| Hsa_circRNA_002178 up-regulates COL1A1 expression through sponging miR-328-3p
To further explore the mechanism by which miR-328-3p neutralized the cancer-promoting effect of hsa_circRNA_002178, a series of assays were conducted to identify the relationship between hsa_circRNA_002178 and miR-328-3p the interaction between miR-328-3p and its target gene of miR-328-3p, COL1A1.
Bioinformatics analysis suggested a miR-328-3p binding site in COL1A1 3′-UTR ( Figure 6A). RT-qPCR and immunohistochemical staining characterized higher expression of COL1A1 mRNA and protein in breast cancer tissues than in adjacent normal tissues ( Figure 6B,C). As shown by the dual luciferase reporter assay, the luciferase activity of COL1A1-wt 3′-UTR was remarkably reduced si-NC si-circ_002178-1 si-circ_002178-2

F I G U R E 4
The interaction between hsa_circRNA_002178 and miR-328-3p. A, bioinformatics analysis of binding sequence between hsa_ circRNA_002178 and miR-328-3p. B, determination of miR-328-3p expression in breast cancer tissues and their matched adjacent normal tissues by RT-qPCR, n = 70. C, RNA pull-down assay for identifying the interaction between hsa_circRNA_002178 and miR-328-3p. D, the luciferase activity of hsa_circRNA_002178-wt or hsa_circRNA_002178-mut in the presence of miR-328-3p; E, determination of miR-328-3p expression in breast cancer cells after silencing of hsa_circRNA_002178 by RT-qPCR. * P < .05 vs the Biotin NC, mimic NC or si-NC group (cells transfected with Biotin NC, mimic NC or si-NC). The measurement data were expressed as mean ± standard deviation. Data between the two groups were analysed by unpaired t test, and among multiple groups were analysed using one-way ANOVA. The experiment was repeated three times

| Hsa_circRNA_002178 silencing hinders tumour growth in vivo through up-regulating miR-328-3p
To investigate how hsa_circRNA_002178 and miR-328-3p af- The measurement data were expressed as mean ± standard deviation. Unpaired t test was conducted for comparison between the two groups. One-way ANOVA was used for comparison among multiple groups. The experiment was repeated three times

| D ISCUSS I ON
Targeted therapy for breast cancer remains big challenges due to high costs and risk of overtreatment. 16   osteosarcoma. 24 Moreover, miR-328-3p exerts as an anti-tumour miR with potential impeding osteosarcoma cell migration through regulation of MMP16. 25 These studies have strongly suggested that miR-328-3p may play important role in cancer development. This study provided evidence that miR-328-3p was lowly expressed in clinical breast cancer tissues. Similarly, decreased miR-328-3p expression was determined in endometrioid endometrial carcinoma. 26 Consistently, transfection with miR-328-3p mimic in breast cancer cells leads to lowered cell motility. 13 miR-328-3p was also involved in the regulation of liver protein metabolism mediated by protein homeostasis and metabolic efficiency. 27 It was also suggested that hsa_circRNA_002178 positively regulated the expression of COL1A1 via sponging miR-328-3p. Recently, it has been reported that circRNAs in mammals are able to act as sponges or competing endogenous RNA (ceRNAs) of miRs thus affecting its transcriptional control. 28 For instance, another circRNA circHIPK3 has been identified as a miR-29b-3p sponge effectively regulate the expression of COL1A1, which is also a target gene of miR-29b-3p. 29 In addition, circCOL3A1-859267 also binds to miR-29c and restored the inhibition of COL1A1 induced by miR-29c. 30 Those findings are partially consistent with our findings that hsa_cir-cRNA_002178 rescued miR-328-3p-induced decrease in mRNA and protein expression of COL1A1. On the contrary, silencing of hsa_circRNA_002178 resulted in reduction of COL1A1 by increasing miR-328-3p. COL1A1 is identified as a hub gene for non-inflammatory breast cancer. 31 Elevated expression of COL1A1 not only correlates with advanced breast cancer and poor prognosis but also related to cancer cell invasion and metastasis. 32 Like the finding in this study, COL1A1 acts as a target gene of miR-196b-5p and underlies the inhibitory effects on breast cancer cell survival and metastasis. 33 Based on the aforementioned findings, we conclude that hsa_circRNA_002178 played inductive role in aggressive malignant behaviours of breast cancer cells through impairing the repression of COL1A1 mediated by miR-328-3p.
Taken together, the findings in our paper suggests a potential therapeutic target hsa_circRNA_002178 for breast cancer treatment.
A molecular network that hsa_circRNA_002178 positively regulates COL1A1 by sponging miR-328-3p is proposed (Figure 8), which might be conducive to the understanding of the pathological process and genetic mechanisms. Despite the model illustrated in this study, the specific metabolic pathways and pro-or anti-inflammatory signals warrant future studies.

ACK N OWLED G EM ENTS
We acknowledge and appreciate our colleagues for their valuable efforts and comments on this paper.

AUTH O R CO NTR I B UTI O N S
TL and PY designed the study. TL, YYY and SL collated the data, carried out data analyses and produced the initial draft of the manuscript. PY and BSH contributed to drafting the manuscript. All authors have read and approved the final submitted manuscript.

F I G U R E 8
The mechanism underlining the role of hsa_circRNA_002178 in the progression of breast cancer with the involvement of miR-328-3p and COL1A1. In breast cancer cells, hsa_circRNA_002178 sponges miR-328-3p that inhibits the expression of COL1A1. The potential miR-328-3p targeting COL1A1 is decreased by hsa_circRNA_002178, the level of COL1A1 is increased accordingly, thereby promoting breast cancer. Inhibition of hsa_circRNA_002178 expression decreased its ability to sponge miR-328-3p and decreased the level of COL1A1, thereby inhibiting breast cancer cell proliferation, energy metabolism, angiogenesis and tumour inflammation. RISC, RNAinduced silencing complex The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.