Inhibition of circular RNA CDR1as increases chemosensitivity of 5‐FU‐resistant BC cells through up‐regulating miR‐7

Abstract This study aims to explore the mechanism of Circular RNA CDR1as implicating in regulating 5‐fluorouracil (5‐FU) chemosensitivity in breast cancer (BC) by competitively inhibiting miR‐7 to regulate CCNE1. Expressions of CDR1as and miR‐7 in 5‐FU‐resistant BC cells were determined by RT‐PCR. CCK‐8, colony formation assay and flow cytometry were applied to measure half maximal inhibitory concentration (IC50), 5‐Fu chemosensitivity and cell apoptosis. Western blot was used to detect the expressions of apoptosis‐related factors. CDR1as was elevated while miR‐7 was inhibited in 5‐FU‐resistant BC cells. Cells transfected with si‐CDR1as or miR‐7 mimic had decreased IC50 and colony formation rate, increased expressions of Bax/Bcl2 and cleaved‐Caspase‐3/Caspase‐3, indicating inhibition of CDR1as and overexpression of miR‐7 enhances the chemosensitity of 5‐FU‐resistant BC cells. Targetscan software indicates a binding site of CDR1as and miR‐7 and that CCNE1 is a target gene of miR‐7. miR‐7 can gather CDR1as in BC cells and can inhibit CCNE1. In comparison to si‐CDR1as group, CCNE1 was increased and chemosensitivity to 5‐Fu was suppressed in si‐CDR1as + miR‐7 inhibitor group. When compared with miR‐7 mimic group, CDR1as + miR‐7 mimic group had increased CCNE1 and decreased chemosensitivity to 5‐Fu. Nude mouse model of BC demonstrated that the growth of xenotransplanted tumour in si‐CDR1as + miR‐7 inhibitor group was faster than that in si‐CDR1as group. The tumour growth in CDR1as + miR‐7 mimic group was faster than that in miR‐7 mimic group. CDR1as may regulate chemosensitivity of 5‐FU‐resistant BC cells by inhibiting miR‐7 to regulate CCNE1.


| INTRODUC TI ON
Breast cancer (BC) is one of the most prevalent and death related cancer in female globally and is a major threat to public health with incidence only secondary to lung cancer. 1,2 Most of BC cases are found in women with age more than 50 years old while nowadays witnesses an increasing trend on incidence happened in younger and women aged 20~34 years and 35~44 years with a respective incidence of 1.9% and 10.5%. 3 Although improvements in early detection and systemic therapy have significantly decreased recurrence and prolonged survival, data reported that 30% of the women diagnosed with early-stage BC in turn progress to metastatic BC, for which therapeutic options are limited. 4,5 With prolonged survival and tumour recurrence, serious problems emerge, such as accumulated drug dosages that approach the upper limit of safety, therapy-related toxicity and drug resistance. 6 Consequently, there is an ever-increasing need for new drugs or combination regimens for the treatment of BC.

5-fluorouracil (5-FU) is well-known for its anti-tumour effect
in numerous tumour, including BC, colon cancer and some skin cancers, which can be paired with other drugs based on individual conditions, cancer types and expected outcomes. 7 The side effect of 5-FU includes inflammation of the mouth, loss of appetite, low blood cell counts, hair loss and skin inflammation. 8 Generally, 5-FU is commonly used to treat BC. Nevertheless, it becomes increasingly ineffective with tumour progression due to chemoresistance. 9 Circular RNA (CircRNA) is a naturally occurring family of noncoding RNAs, which has been commonly detected in viruses, plant and animals. [10][11][12] CircRNAs are proved as potential biomarkers in many diseases including hepatoma carcinoma and other cancers. 13 CDR1as (also known as CiRS-7), one of the thousands of CircRNAs, was recently demonstrated to function as a powerful miR-7 sponge/ inhibitor in developing midbrain of zebrafish, suggesting a novel mechanism for regulating microRNA functions. 14 Evidence supported that ectopic expression of CDR1as induced midbrain brain defects, which was similar with the phenotypes found in the knockdown of miR-7. 15 Although previous study has focused emphasis on the relationship between CDR1as and miR-7, little study was published whether CDR1as can regulate the chemosensitivity to 5-FU in BC cells through targeting miR-7. This study was conducted to investigate the mechanism of CDR1as implicating in regulating 5-FU chemosensitivity in BC by competitively inhibiting miR-7 to regulate CCNE1.

| Cell grouping
Further experiments were conducted on MCF-7 cells and MDA-MB-231 cells, the former has the largest difference with MCF10 CDR1as cells and the latter has the least difference with MCF10 CDR1as cells. The two kinds of cells were induced for drug resistance and grouped into following groups: Blank group, Empty plasmid group, si-CDR1as group, CDR1as group, negative control (NC) group, miR-7 mimic group, miR-7 inhibitor group, si-CDR1as + miR-7 inhibitor and CDR1as + miR-7 mimic group. No treatment on blank group, Empty plasmid group transfecting with empty plasma, si-CDR1as group transfecting with siRNA interference plasmid of CDR1as, CDR1as group transfecting with plasma with overexpression of CDR1as, NC group transfecting with NC sequence of miR-7, miR-7 mimic group transfecting with mimics of miR-7, miR-7a inhibitor group transfecting with inhibitor of miR-7, si-CDR1as+ miR-7 inhibitor group transfecting with siRNA interference plasmid of CDR1as and inhibitor of miR-7, CDR1as + miR-7 mimic group transfecting with plasma with overexpression of CDR1as and mimics of miR-7a. Empty plasma, siRNA interference plasmid, overexpression plasma as well as NC, mimics and inhibitor of miR-7 were all purchased from Shanghai GenePharma Co., Ltd. Cell transfection was conducted based on the introduction of Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Cells in each group were incubated in incubators for 48 h before further experiments.

| Reverse transcript polymerase chain reaction (RT-PCR)
The TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used to isolate total RNA, including microRNA, according to the manufacturer's instructions. One microgram of total mRNA was used to reverse transcribe cDNA with Takara RT-PCR synthesis kit, according to the manufacturer's instructions (TAKARA, Dalian, China).
The cDNA was performed using SYBR Premix Ex Taq II (TAKARA, Dalian, China) on the PikoReal 96 qPCR system (Thermo Scientific, United States). The mRNA level of β-actin was relative to internal control. For miR-7a expression, microRNA was analyzed by using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) with provided RT-U6 and microRNA-specific stem-loop primers, and the expression levels were determined through TaqMan MicroRNA assays with the TaqMan Universal PCR Master Mix (Applied Biosystems), and U6 snRNA was used as the endogenous control. All reactions were preceded on the ABI 7500 real-time PCR system (Applied Biosystems) by standard protocols.
The sequences used for PCR is listed in Table 1. Data were analyzed using 2 −ΔΔCt method. The experiments were conducted for three times to obtain average value.

| CCK-8
Cells in each group were digested and inoculated in 96-well plate at the density of 8 × 10 3 /well. The volume of each well was 200 uL.
When cells were adherent to the wall of each well, 5-Fu of respective concentration of 0, 0.10 umol/L, 0.50 umol/L, 2.50 umol/L, 12.50 umol/L, 20.00 umol/L and 40.00 umol/L were added. Three duplicate wells were set for each concentration. Blank well and control group were also established. The plate was incubated at 37°C with 5% CO 2 for 48 hours before CCK-8 assay was applied. Replace the culture medium with 100 uL of fresh culture medium which contains 10 uL of CCK-8 regent (Beyotime Biotechnology Co., Ltd.).
Then the plate was maintained in an incubator for 2 hours before the microplate reader (Bio-Rad, USA) was utilized to measure the optical density (OD) at the wavelength of 450 nm. Cell survival rate was calculated and cell growth curve was accordingly drawn. The experiment was repeated for three times. Half maximal inhibitory concentration (IC50) was calculated using Probit regression analysis by SPSS software.

| Colony formation assay
Cells in logarithmic phase were digested by pancreatic enzymes and made into single cell suspension. About 200 cells were separately inoculated into a 6 cm dish and incubated in complete culture medium containing 15 nmol/mL of 5-Fu. Then the dishes were cultured in an incubator at 37°C with 5% CO 2 for 2-3 weeks, during which the culture medium will not be replaced. Terminate the cultivation and abandon the culture medium until cell cloning was visible by naked eyes. The dishes were washed in PBS for twice and then added 5 mL of methanol solution for fixation at room temperature for 15 minutes. After that, the fixation solution was absorbed by using a vacuum pump and Giemsa dye (SIGMA, USA) was used for staining.
About 30 minutes later, the staining solution was gradually washed away and the dishes were dried in the air. With the application of an inverted microscope, cell colon number and colon formation rate were calculated by naked eyes (colon formation rate = cell colon number/number of inoculated cells × 100%).

| Flow cytometry (FCM)
AnnexinV/propidiom iodide (PI) double staining method was applied to detect cell apoptosis. Cells after grouping for 48 hours were collected for concentration adjustment to 1 × 10 6 /mL. Then 0.5 mL of cell suspension was added into a centrifuge tube which was then

| Western blot analysis
Total protein of cells in each group was isolated using protein lysis

| RNA pull down assay
The biotinylated miR-7 and mutant miR-7 (miR-7-MUT) which were designed and synthesized by Invitrogen and transfected into MCF-7 cells and 48 hours later, RNA pull down lysate was applied to lyse cells. Take 100 ug samples as input for further experiments.
Streptavidin magnetic beads (S1421S, NEB, Ipswich, MA, USA) was TA B L E 1 Primer sequences for reverse transcript polymerase chain reaction
Samples were added with Trizol to isolated RNA which was bind to the beads. PCR amplification was performed to detect the expression of CDR1as in target genes.

| Double luciferase reporter assay
Using Targetscan database to evaluate the target genes of miR-7, which preliminary determined CCNE1 as a direct target gene of

| Xenotransplanted tumour in nude mouse
Animal experiments were conducted in strict accordance with the  CDR1as was suppressed, while that of CDR1as group was increased (all P < 0.05) ( Figure 1D,E). The cell apoptosis rate in si-CDR1as group was higher than that in Empty plasmid group, while that in CDR1as group was lower than that in Empty plasmid group (all P < 0.05). Measurement on apoptosis related factors is illustrated in Figure 1H Figure 2B). Comparison on IC50 between NC-miRNA group and Blank group showed no significant difference. As illustrated in Figure 2C, the IC50 in miR-7 mimic group was reduced, while that in miR-7 inhibitor group was increased when compared with NC-miRNA group. The results of colon formation assay showed ( Figure 2D,E) that the colon formation rate in miR-7 mimic group was lower than that in NC-miRNA group, while that in miR-7 inhibitor group was higher than that in NC-miRNA group. No significant difference was found between NC-miRNA group and Blank group.

| Overexpression of miR-7 may increase chemosensitivity of 5-FU-resistant BC cells
Comparison on cell apoptosis rate is demonstrated in Figure 2F,G. In contrast to NC-miRNA group, the cell apoptosis rate in miR-7 mimic group was increased while that in miR-7 inhibitor group was inhib-

| CDR1as competitively inhibits miR-7 to regulate CCNE1
Targetscan software indicates the bind site of CDR1as and miR-7 ( Figure 3A). To verify whether CDR1as can directly bind miR-7, we firstly synthesized biotin labeled miR-7 and transfected into MCF-7 cells. Then total RNA which was bind miR-7 was isolated from cell lysis buffer. RT-PCR was applied to determine the expression of CDR1as.
As indicated in Figure 3B, miR-7 can gather CDR1as in BC cells, but not β-actin, which suggested that miR-7 can specially recognize sequences of CDR1as. Targetscan database was used to evaluate the target genes of miR-7 and CCNE1 was selected as a target gene for further analysis ( Figure 3C). Double luciferase reporter assay ( Figure 3D) showed that in wide type, compared with CCNE1-Wt + miR-7 NC group, the luciferase activity of CCNE1-Wt + miR-7 mimics group was suppressed, while that in CDR1as group and miR-7 inhibitor group were increased. In contrast to miR-7 mimic group, the expression of CCNE1 in CDR1as + miR-7 mimic group was elevated (P < 0.05). The expression of CCNE1 in si-CDR1as + miR-7 inhibitor group was increased in comparison to si-CDR1as group, indicating that CDR1as can competitively inhibit miR-7 to regulate CCNE1.

| Inhibition of miR-7 reverses the enhancement on chemosensitivity of 5-FU-resistant BC cells caused by CDR1as silencing
Comparison on 5-FU-resistant BC cells between si-CDR1as group and si-CDR1as + miR-7 inhibitor group, as well as between miR-7 mimic group and CDR1as + miR-7 mimic group showed that cell survival rate in each group was decreased along with the increase concentration of 5-Fu ( Figure 4A). Compared with si-CDR1as group, IC50 ( Figure 4B) and colon formation rate ( Figure 4C,D) in si-CDR1as + miR-7 inhibitor group were increased, while cell apoptosis rate ( Figure 4E,F) and expressions of Bax/Bcl2 and cleaved-Caspase-3/Caspase-3 were decreased ( Figure 4G,H). Similarly, IC50 ( Figure 4B) and colon formation rate ( Figure 4C,D) in CDR1as + miR-7 mimic group were also el-

| Implication of CDR1as in development of BC and its effect on chemosensitivity to 5-Fu
The mice with xenotransplanted tumour were decreased in weight and the tumour sizes were increased. The tumour growth curve illustrated that the tumour growth rate in CDR1as group and miR-7

| D ISCUSS I ON
Widespread and substantial studies have been focused on the implication of circular RNA on cancer progression. [17][18][19] CDR1as, one of the well-known circular RNA, has been suggested to have certain role in certain diseases, including hepatocellular carcinoma. 13 However, less study was found on the involvement of CDR1as in chemosenstivity of BC.
In this study, we firstly CDR1as acts as a strong sponge for miR-7. Studies on in embryonic zebrafish midbrain showed that overexpression of CDR1as can induce developmental defects and regulate insulin transcription and secretion in islet cells through blocking miR-7. 14,21 The mechanism of circRNAs in gene regulation has certain relationship with its function of competing endogenous RNAs or microRNA sponges during biological processes and diseases progression even including carcinogenesis. 22 CDR1as has long been recognized as a tumour oncogene in various genes 23 in addition to its role in myocardial infarction which suggested that CDR1as was elevated in mice with myocardial infarction under hypoxic treatment, and overexpression of CDR1as can promote cell apoptosis, which can be reversed by overexpression of miR-7a. 24 Furthermore, CDR1as also proved to promote hepatocellular carcinoma cell proliferation and invasion and act as a risk factor of hepatic microvascular invasion in hepatocellular carcinoma. 23 As one of the highlights of this study, we further explore the possible mechanism of CDR1as in regulating chemosensitivity of 5-FU-resistant BC cells. The results supported the implication of miR-7 and CCNE1. MiR-7, a putative tumour-suppressor, regulates the expression of several important drivers in multiple types of cancer. 25 Consistent with the result of our study, an investigation on hepatocellular carcinoma showed that miR-7 exerts its tumour suppressive F I G U R E 5 Observation on xenotransplanted tumour size of nude mice in each group. *, compared with Empty plasmid group, P < 0.05, #, compared with si-CDR1as group, P < 0.05, &, compared with miR-7 mimic group, P < 0.05 function though the inhibition of oncogene CCNE1 expression. 26 Since miR-7 interacts with Cdr1as, we examined whether overex-

CO N FLI C T S O F I NTE R E S T
The authors declare no potential conflicts of interest.