Knockdown of hsa_circ_0074298 suppressed the progression of pancreatic cancer by targeting the miR-519/SMOC2 axis

BACKGROUND: Pancreatic cancer (PC) is among the most malignant tumors in digestive system because of its fast progression, metastasis as well as resistance to chemotherapy. Former studies found that circRNAs that differentially expressed were correlated to progression and gemcitabine (GEM)-resistance in PC, and hsa_circ_0074298 aberrant expression played a role in PC progression, but the regulatory mechanism was unclear. METHODS: Rt-qPCR was used to detect hsa_circ_0074298 expression in PC cell lines. The CCK8, colony formation, and Transwell assays were used to evaluate the effect of hsa_circ_0074298 on PC cell migration and proliferation. Bioinformatics and luciferase reporter experiments were used to study the regulatory mechanism. Nude mouse xenografts were generated to examine the effect of hsa_circ_0074298 on tumor growth. RESULTS: The present study showed that hsa_circ_0074298 expression increased significantly in both PC cell lines and PC tissues. Downregulation of hsa_circ_0074298 caused a significant decrease in PC cell proliferation and in nude mouse xenograft growth. Hsa_circ_0074298 silencing also increased chemotherapy sensitivity to GEM. Bioinformatics analysis indicated that hsa_circ_0074298 was a miR-519 sponge and that the SMOC2 gene was a miR-519 target. MiR-519 downregulation or SMOC2 overexpression restored cell proliferation, migration, and chemoresistance after hsa_circ_0074298 silencing. The dual luciferase reporter assay showed that hsa_circ_0074298 interacted with miR-519, and miR-519 binding to the SMOC2 3′-UTR, then suppressed posttranscriptional SMOC2 expression. CONCLUSIONS: Taken together, hsa_circ_0074298 functioned as a tumor promoter through novel miR-519/SMOC2 axis, highlighting its possibility as a potential therapy for PC.

expression played a role in PC progression, but the regulatory mechanism was unclear.
METHODS: Rt-qPCR was used to detect hsa_circ_0074298 expression in PC cell lines.
The CCK8, colony formation, and Transwell assays were used to evaluate the effect of hsa_circ_0074298 on PC cell migration and proliferation. Bioinformatics and luciferase reporter experiments were used to study the regulatory mechanism. Nude mouse xenografts were generated to examine the effect of hsa_circ_0074298 on tumor growth. RESULTS: The present study showed that hsa_circ_0074298 expression increased significantly in both PC cell lines and PC tissues. Downregulation of hsa_circ_0074298 caused a significant decrease in PC cell proliferation and in nude mouse xenograft growth.
Hsa_circ_0074298 silencing also increased chemotherapy sensitivity to GEM.
Bioinformatics analysis indicated that hsa_circ_0074298 was a miR-519 sponge and that the SMOC2 gene was a miR-519 target. MiR-519 downregulation or SMOC2 overexpression restored cell proliferation, migration, and chemoresistance after hsa_circ_0074298 silencing. The dual luciferase reporter assay showed that hsa_circ_0074298 interacted with miR-519, and miR-519 binding to the SMOC2 3′-UTR, then suppressed posttranscriptional SMOC2 expression. CONCLUSIONS: Taken together, hsa_circ_0074298 functioned as a tumor promoter through novel miR-519/SMOC2 axis, highlighting its possibility as a potential therapy for PC. Background 3 Pancreatic cancer (PC) is among the major causes of cancer-relevant mortality in the world. Mortality caused by PCs of various classes accounted for 411,600 deaths in 2015 [1]. Chemotherapy, many times using gemcitabine (GEM), is now a common adjunct treatment for PC. However, resistance to gemcitabine is a main gap for effective chemotherapy of PC [2][3][4], although the molecular mechanism of PC resistance to GEM is still unclear.
Non-coding RNAs (ncRNAs) are RNAs not encoding proteins. It is already known that ncRNA contributes to a large amount of cellular RNAs, making at least 90% of human RNAs [5].
Former studies have illustrated that ncRNAs, like proteins, act as underlying factors in various cellular processes, including cell migration, proliferation, apoptosis, chemoresistance, angiogenesis, and the immune response [6,7]. CircRNAs are considered as essential human disease regulators as well as biomarkers in many kinds of malignancies such like PC. The competing endogenous RNA (ceRNA) mechanism is known as an indispensable mechanism through which circRNAs regulate gene expression.
Previous studies have found that circular RNA circ_0007534 that upregulated is associated with an unfavorable prognosis in pancreatic ductal adenocarcinoma patients, which regulates cell apoptosis, invasion, as well as proliferation by sponging miR-892b and miR-625 [12]. CircRNA_100782 regulates pancreatic carcinoma proliferation via the IL6-STAT3 pathway [13]. Previous studies with microarray analysis involving circRNA expression profiles associated with GEM resistance in PC cells also revealed a family of circRNAs that were differentially expressed, including hsa_circ_0074298. Although hsa_circ_0074298 functions in PC progress, its mechanism of GEM resistance is not clear, so current investigation would identify hsa_circ_0074298 regulatory mechanism. 4

Ethics statement
We collected 30 fresh PC tissues and paired adjacent tissues after obtaining informed consent from patients at the Affiliated Zhongda Hospital, Southeast University, Nanjing.
We snap-froze samples in liquid nitrogen and stored them at -80°C before RNA extraction and fluorescence in situ hybridization ( FISH). The Ethics Committee of the Affiliated Zhongda Hospital of Southeast University approved this research, and the study was carried out in accordance with The Code of Ethics of the World Medical Association .

Cell culture
We obtained PC cell lines (PANC-1, SW1990, AsPC-1, and BxPC-3) as well as normal human pancreatic duct epithelial cells, HPDE from the American Type Culture Collection. The culture medium consisted of fetal bovine serum (FBS; Invitrogen Life Technologies, Carlsbad, USA) of 10% in Dulbecco's Modified Eagle's Medium, as well as penicillin. We cultured cells in an incubator at 37 °C with CO 2 of 5%.

Bioinformatics analyses
We predicted circRNA/miRNA target genes using Circular RNA Interactome, and predicted the interactive relations between miR-519 and SMOC2 through TargetScanHuman.

Transwell migration assay
We analyzed cell migration via Transwell chambers (Corning, Corning, NY, USA) following standard procedures. After incubation for 1 day, we removed cells on the chamber upper surfaces by cotton swabs. We fixed cells locating on the lower surfaces with methanol for ten minutes, followed by Crystal Violet staining. We imaged stained cells and counted them in five fields that randomly selected. In invasion experiments, we coated chamber inserts with 200 mg/mL Matrigel and dried them overnight under sterile conditions.

Cell proliferation assay
Following the CCK-8 assay (Dojindo Laboratories, Kumamoto, Japan), we assessed cell growth of transfected cells in plates with 96 wells at 24, 48, and 72 h. We used a spectrophotometer (Thermo Scientific, Rockford, IL, USA) to detect the absorbance at 450 nm.

Colony formation assay
We transferred BxPC-3 and PANC-1 cells into plates with 6 wells for ten days. Then, we treated the colonies with 10% formaldehyde for half of an hour, followed by staining for 5 min with 0.5% Crystal Violet. Image-Pro Plus 6.0 (https://www.mediacy.com/imageproplus) was used for data analysis.

Cell cycle assay
A total of 2×10 5 cells/mL were diluted with RNase A in 75% ice-cold ethanol overnight, then we stained cells with propidium iodide (PI; 50 mg/mL; MultiSciences Biotech, Hangzhou, China) in the dark for 30 min at 4ºC, followed by measuring with a flow cytometer (FACScan, BD Bioscience, San Jose, CA, USA).

Cell apoptosis assay
Flow cytometry binding buffer (100 μL) was added after harvested cells were washed twice using ice-cold buffer. The mixture containing 5 μL Annexin V/FICC and PI (BD, Franklin Lakes, NJ, USA) of 5 μL was used for staining of cells for 15 min in the dark, followed by another 400 µL binding buffer. We used FACSCalibur flow cytometer (BD Biosciences) to analyze cell apoptosis.

Western blot analysis
We washed cells with precooled phosphate-buffered saline and then lysed them with cell lysis solution (RIPA). We detected the protein concentration via BCA (Thermo Fisher Scientific, Waltham, MA, USA). We then transferred proteins to a polyvinylidene difluoride membrane, and blocked them in TBST (25 mM Tris, 140 mM NaCl, and 0.1% Tween 20, pH 7.5) containing 5% skimmed milk and then incubated them for two hours. We then incubated membranes in primary antibodies against SMOC2 and GAPDH (Abcam, Cambridge, MA, U.S.A.) at 4°C overnight. After washing (3× for 10 min) with TBST, we added secondary antibody and incubated them under room temperature for one hour. We analyzed results using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Dual-luciferase reporter assay
We cloned putative miR-519 binding site in the target gene, SMOC2, and wild-type (WT) or
We maintained all mice and handled them according to the instructions of the Animal Ethics Committee of Affiliated XXXX Hospital, XXXX University Medical School.

Statistical analysis
We conducted statistical analysis via SPSS statistical software (SPSS, Chicago, IL, USA).
We utilized Student's t-test to analyze the data, which are denoted by the mean ± SD. p < 0.05 was regarded as statistical significance.

experiments.
Bioinformatics analysis suggested that hsa_circ_0074298 was derived and cyclized from a portion of the HARS gene exon and was located at chr5:140053489-140058712 ( Figure   1A). Fluorescence in situ hybridization (FISH) detection showed that hsa_circ_0074298 expression increased in PC tissues comparing with their paired non-tumor tissues.
Subcellular localization analysis showed that hsa_circ_0074298 was located predominantly in the cytoplasm ( Figure 1B). We then performed qRT-PCR to determine hsa_circ_0074298 expression in 30 paired tissue specimens and four PC cells such as SW1990, BxPC-3, AsPC-1, and PANC1. We used HPDE cells as a control. Results demonstrated that hsa_circ_0074298 expression increased in PC tissues ( Figure 1C) and cell lines ( Figure   1D). BxPC-3 and PANC-1 cells had the highest hsa_circ_0074298 expressions, which were used in subsequent assays. We conducted Lentivirus-mediated hsa_circ_0074298 silencing of BxPC-3 and PANC-1 cells, and results verified that hsa_circ_0074298 expression decreased significantly ( Figure 1E). Cell cycle distribution analysis showed that the percentage of S-phase cells significantly decreased and the percentage of G2/M-phase cells increased after hsa_circ_0074298 depletion, indicating a cell cycle arrest at the G2/M phase ( Figure 1F). Both the CCK8 assay ( Figure 1G and 1H) and colony formation assay ( Figure 1I and 1J) showed that hsa_circ_0074298 depletion decreased cell proliferation in BxPC-3 and PANC-1 cells. PANC-1 cells transfected with or without hsa_circ_0074298 lentiviral interference vectors were used to assay tumor formation in nude mice xenografts. We determined results by measuring tumor volumes 5 days after grafting, with a vernier caliper, and indicated that hsa_circ_0074298 knockdown resulted in a reduced xenograft volume ( Figure 1K).

hsa_circ_0074298 knockdown decreased migration capabilities and GEM resistance in PC.
Results validated that hsa_circ_0074298 depletion decreased the migration ability in BxPC-3 and PANC-1 cells (Figure 2A Figure 2E and 2F). Overall, the results validated that hsa_circ_0074298 silencing inhibited PC cell migration and chemotherapy resistance, but the mechanism remains to be determined.

MiR-519/SMOC2 was the target of hsa_circ_0074298.
Bioinformatics analysis tool Circular RNA Interactome was utilized to select candidate hsa_circ_0074298 targets. Results validated that hsa_circ_0074298 had 12 more conservative miRNAs ( Figure 3A). We then designed a hsa_circ_0074298 luciferase reporter screening protocol for the miRNAs. The results showed that miR-519 decreased the hsa_circ_0074298 luciferase reporter luciferase activity most by at least 80% ( Figure. 3B). These results indicated that miR-519 had a more conserved binding site for hsa_circ_0074298. The RT-qPCR detection showed that miR-519 expression decreased in PC tissues when comparing with adjacent non-tumor tissues ( Figure 3C). We subsequently conducted a dual-luciferase reporter assay in HEK293T cells. We cloned mutant and wildtype hsa_circ_0074298 sequences to construct the mutant vectors and reporter plasmids, respectively ( Figure. 3D). Although it was found that miR-519 mimics co-transfected with the reporter plasmids decreased luciferase activity, in contrast, miR-519 mimics and mutated vector transfection showed no significant changes in luciferase activity. Hence, these data proved that miR-519 was a direct hsa_circ_0074298 target ( Figure 3E). Next, bioinformatics analysis (http://www.targetscan.org/) showed that SMOC2 was a potential miR-519 target. To confirm that SMOC2 was a miR-519 target, we cloned mutant and wildtype SMOC2 sequences to construct mutant vectors and reporter plasmids, respectively   RT-qPCR analysis validated that hsa_circ_0074298 expression were decreased after hsa_circ_0074298 silence, but treatment with the overexpressing SMOC2 or the miR-519 inhibitor did not affect hsa_circ_0074298 expression ( Figure. 4A and 4B), showing that miR-519 and SMOC2 were hsa_circ_0074298 downstream. RT-qPCR detection also verified that hsa_circ_0074298 silencing promoted miR-519 expression, but SMOC2 overexpression had no effect on hsa_circ_0074298 depletion-induced SMOC2 expression ( Figure 4C [14,15]. Currently, sponging activity is the main function of some circRNAs, so in tumor development, circRNA-miRNA-mRNA interaction networks might be crucially important [16,17].
To further identify the downstream miRNA, bioinformatics analysis was used to show that 12 different miRNAs were the targets of hsa_circ_0074298. Luciferase reporter studies confirmed that hsa_circ_0074298 could interface with miR-519. hsa_circ_0074298 downregulation promoted miR-519 expression, and miR-519 silencing restored the proliferation, migration, and GEM resistance after hsa_circ_0074298 depletion, suggesting that miR-519 had anti-tumor activity. Previous studies also found that upregulation of miR-519 inhibited cancer cell activity, including that of nasopharyngeal carcinoma and gastric cancer [18][19][20]. Together, these results strongly suggested that hsa_circ_0074298 expression promoted the pancreatic cancer progression by sponging miR-519.
Our further bioinformatics analyses showed that miR-519 bound to the SMOC2 3'UTR, and luciferase reporter studies confirmed that miR-519 interacted with the 3'UTR of SMOC2.
Overexpression of SMOC2 restored the proliferation, migration, and GEM resistance after hsa_circ_0074298 depletion. An SMOC-2 named the SPARC-related modular calcium binding 2, which is a member of SPARC family, is highly expressed during wound healing Present investigation showed that hsa_circ_0074298 promoted PC cell proliferation, migration, and chemoresistance by sponging miR-519 and enhancing SMOC2 expression. 14

Conclusions
In conclusion, results verified that hsa_circ_0074298 promoted proliferation in PC cells by possibly activating miR-519/SMOC2 signaling. All our results indicated that hsa_circ_0074298 could be candidate biomarker for PC prognosis and diagnosis, and may extend the use of drugs that target hsa_circ_0074298, which indicates the promising role regarding hsa_circ_0074298 in PC treatment.  Data are presented as the mean ± SD; ***P < 0.001 vs. the untreated group.