Clinical samples and cell culture

GC patients (n = 44, male : female = 1 : 1, 40-70 years old) from First Affiliated Hospital of Kunming Medical University between January 2016 and December 2016 were recruited in this study. All patients did not receive any treatment before surgery. GC tissues and adjacent normal tissues were collected from these patients during radical gastrectomy. The partial samples were snap-frozen in liquid-nitrogen and then stored at −80 °C. In addition, other samples were fixed with formalin and embedded with paraffin for immunohistochemistry (IHC) assay. This study was performed with the acquisition of written informed consents from patients and the approval of Ethics Committee from First Affiliated Hospital of Kunming Medical University.

Human GC cell lines (HGC27 and AGS) were obtained from China Center for Type Culture Collection (Wuhan, China). HGC27 is from undifferentiated GC cells with epithelial morphology, and AGS is from gastric adenocarcinoma cells with epithelial morphology. Human umbilical vein endothelial cells (HUVECs) were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). Human gastric epithelial cells (GES-1) were purchased from Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). HUVECs were maintained in Endothelial Cell Growth Medium (Sigma-Aldrich, St. Louis, MO, USA). GES-1, HGC27 and AGS cells were maintained in RPMI-1640 medium (Thermo Fisher Scientific, Rockford, IL, USA) supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific, Rockford, IL, USA). All cells were cultured in a humidified incubator containing 5% CO2 at 37 °C.

Cell transfection

Short hairpin RNA (shRNA) vector targeting circ_0000620 (sh-circ_0000620: AATTCAAAAAGTGATGAAGAATGATATCCTTCTCGAGAAGGATATCTTCTTCATCAC), shRNA negative control (sh-NC: CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTTG), miR-671-5p mimic or inhibitor (miR-671-5p or anti-miR-671-5p), mimic or inhibitor control (miR-NC or anti-miR-NC), MMP2 overexpression vector (MMP2) and the pcDNA control vector were all obtained from GenePharma Co., ltd (Shanghai, China). The oligonucleotides or plasmids were transfected into HGC27 and AGS cells using Lipofectamine 3000 reagent (Thermo Fisher Scientific, Rockford, IL, USA) following manufacturer’s instruction. Transfection concentrations were 40 nM shRNA vector, 40 nM mimic, 20 nM inhibitor, 2 µg MMP2 or pcDNA vector, respectively.

Quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNAs were isolated from tissues and cells using Trizol reagent (Thermo Fisher Scientific, Rockford, IL, USA). Then, mature miR-671-5p was quantified using TaqMan MicroRNA Assays (Thermo Fisher Scientific, Rockford, IL, USA) with U6 as the internal control. Reverse transcription by M-MLV Reverse Transcriptase (Thermo Fisher Scientific, Rockford, IL, USA) and PCR reaction by SYBR™ Green Master Mix (Thermo Fisher Scientific, Rockford, IL, USA) were used for the determination of circ_0000620, AAGAB and MMP2. The reaction protocols were listed as below: pre-denaturation at 95˚C for 30 s, 40 cycles of denaturation at 95˚C for 5 s and annealing at 60˚C for 30 s. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) functioned as the house-keeping gene to normalize circ_0000620, AAGAB and MMP2 levels. The primer sequences were as follows: 5΄-CTGAATGCCAATGTGTGGTC-3΄ (forward) and 5΄-CTATCAAGGCCCGATTTTTG-3΄ (reverse) for circ_0000620; 5΄-GGCAAAAGCATGGTTACCTGAGG-3΄ (forward) and 5΄-CTCAGGCAACTCCTCTGGACTA-3΄ (reverse) for AAGAB; 5΄-GCCGAGAGGAAGCCCTGGAG-3΄ (forward) and 5΄-CTCAACTGGTGTCGTGGA-3΄ (reverse) for miR-671-5p; 5΄-GATGGCACCCATTTACACCTAC-3΄ (forward) and 5΄-GTCCTTGAAGAAGAAGATCTC-3΄ (reverse) for MMP2; 5΄-GTCAGTGGTGGACCTGACCT-3΄ (forward) and 5΄-CCCTGTTGCTGTAGCCAAAT-3΄ (reverse) for GAPDH; and 5΄-CTCGCTTCGGCAGCACA-3΄ (forward) and 5΄-AACGCTTCACGAATTTGCGT-3΄ (reverse) for U6.

Ribonuclease R (RNase R) assay

The isolated total RNA samples from HGC27 and AGS cells were exposed to 3 U µg−1 RNase R (Epicentre, Madison, WI, USA) for 15 min at 37℃. The expression levels of circ_0000620 and linear AAGAB were determined by qRT-PCR assay. RNA without incubation of RNase R was used as the negative control group (Mock).

Cell Counting Kit-8 (CCK-8) assay

Cell viability was tested using a CCK-8 assay kit (Dojindo Molecular Technologies, Rockville, MD, USA) according to the instructions of manufacturer. Briefly, the transfected HGC27 and AGS cells were seeded into 96-well plates and cultured for 24 h in fresh medium. Then, cells were incubated with CCK-8 solution (10 µL per well) for 3 h and cell absorbance was measured at 450 nm.

Colony formation assay

1000 cells per well were plated into 6-well plates and cultured for 14 days. Subsequently, cells were fixed with 4% formaldehyde for 15 min and stained with 0.1% crystal violet (Sigma-Aldrich, St. Louis, MO, USA) for 10 min. The colonies (containing more than 50 cells) were counted under an optical microscope.

Wound healing assay

Cell migration ability was analyzed by wound healing assay. HGC27 and AGS cells were seeded in 6-well plates and transfected for 24 h, followed by producing two scratches via a sterile pipette tip (200 µL). Whereafter, cells were gently washed using PBS twice and maintained in serum-free medium (SFM) for 24 h. Images of these scratches were captured at 0 and 24 h under a microscope with 40 × magnification. The migration rate (%) was calculated by the formula: (wound width0 h - wound width24 h)/wound width0 h.

Transwell invasion assay

The invasive potential was tested using a matrigel-precoated Transwell Boyden Chamber (Costar, Lowell, MA, USA) containing 8 μm pore size membranes. Cells resuspended in SFM were plated on the upper chamber, and medium with 10% FBS was added into the lower compartment. After 24 h of incubation, cells on the upper surface of membranes were removed. Cells attached to the lower surface of membranes were fixed with methanol for 20 min and stained with 0.1% crystal violet solution (Sigma-Aldrich, St. Louis, MO, USA) for 20 min. The invaded cell images were acquired and counted under the inverted microscope (Olympus, Tokyo, Japan).

Collection of GC cells-conditioned medium (CM)

HGC27 and AGS cells were plated into the 6 well plates and then transfected with or without corresponding miRNA mimic, siRNAs, or plasmids, alone or in combination. At 36 h after transfection, the medium was removed and cells were maintained in SFM for 12 h. Next, the collected CM was centrifuged at 3000 rpm for 10 min to remove cells and at 12,000 rpm for 10 min to eliminate cell debris. Finally, CM was stored in aliquots at −80 °C.

Tube formation assay

Tube formation ability of HUVECs was assessed by In vitro Angiogenesis Assay Kit (Millipore, Bedford, MA, USA) following the protocols of manufacturer. Briefly, HUVECs were seeded into 96-well plates precoated with ECMatrixTM and cultured in SFM or CM for 12 h. Then, the average values of branch points were counted in 10 random views of fields under a phase contrast microscope (CK40; Olympus, Tokyo, Japan).

Western blot assay

The total proteins were extracted using RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific, Rockford, IL, USA) containing protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). Then, protein concentration was measured using Bio-Rad Protein Assay (Bio-Rad, Hercules, CA, USA). Next, 30 µg proteins were separated by SDS-PAGE and electrotransferred onto nitrocellulose membranes (Millipore, Billerica, MA, USA). Subsequently, the membranes were blocked with 5% skim milk followed by incubation with the primary antibodies and the secondary antibody. The antibody information was shown as below: anti-MMP2 (1:1000; ab97779; Abcam, Cambridge, MA, USA), E-cadherin (E-cad, ab133597; 1:2000; Abcam, Cambridge, MA, USA), N-cadherin (N-cad, ab207608; 1:1000; Abcam, Cambridge, MA, USA), and anti-GAPDH (1:5000; ab9485; Abcam, Cambridge, MA, USA) and horseradish peroxidase (HRP)-labeled goat anti-rabbit secondary antibody (1:5000; ab205718; Abcam, Cambridge, MA, USA). Finally, protein signals were detected using Clarity Western ECL Substrate (Bio-Rad, Hercules, CA, USA) and the signal intensity was estimated using Quantity One software Version 4.1.1 (Bio-Rad, Hercules, CA, USA) via gray analysis.

IHC assay

MMP2 protein expression was determined using western blot, with nine pairs of tissues as the experimental samples. The formalin-fixed and paraffin-embedded tissues were suffered from the sequential treatment of antigen retrieval, endogenous peroxidase blockage, nonspecific signal blocking. Then, tissue sections were incubated with anti-MMP2 primary antibody and HRP-conjugated secondary antibody. Then, tissue slices were stained with 3,3’-diaminobezidin (DAB) substrate and counterstained with hematoxylin. The stained tissues were imaged using an inverted microscope (Nikon E-800 M; Nikon Corporation, Japan).

Subcellular fractionation assay

RNA was obtained from the cytoplasm and nucleus of HGC27 and AGS cells using Cytoplasmic & Nuclear RNA Purification Kit based on the guidance of manufacturer (Norgen Biotek, Thorold, ON, Canada). 18 S rRNA and U6 were used as internal controls for cytoplasm and nucleus.

Dual-luciferase reporter assay

The online circinteractome (https://circinteractome.nia.nih.gov/), circbank (http://www.circbank.cn/index.html) and starbase (http://starbase.sysu.edu.cn) were used for predicting the binding sites between targets. The sequence of circ_0000620 or MMP2 3΄ untranslated region (3΄UTR) containing predicted or mutant miR-671-5p binding site was constructed into psiCHECK-2 luciferase vector by Hanbio Biotechnology Co., ltd. (Shanghai, China). The novel plasmids were named as circ_0000620 WT, MMP2 3΄UTR-1 WT and MMP2 3΄UTR-2 WT, circ_0000620 MUT, MMP2 3΄UTR-1 MUT and MMP2 3΄UTR-2 MUT. Then, HGC27 and AGS cells in 24-well plates were co-transfected with 100 ng reporter plasmids and 30 nM miR-671-5p or miR-NC using Lipofectamine 3000 reagent. Forty-eight hours later, luciferase activities were detected using a dual luciferase reporter assay system (Promega, Madison, WI, USA).

RNA immunoprecipitation (RIP) assay

After transfection for 48 h, HGC27 and AGS cells were lysed using RIP lysis buffer (Bio-Rad, Hercules, CA, USA). Subsequently, cell lysates were incubated with Magnetic beads (Bio-Rad, Hercules, CA, USA) containing Argonaute 2 antibody (Ago2; Bio-Rad, Hercules, CA, USA) or Immunoglobulin G antibody (IgG; Bio-Rad, Hercules, CA, USA) for 3 h at 4℃. The RNA levels of circ_0000620, miR-671-5p and MMP2 were detected by qRT-PCR after RNA extraction using TRIzol reagent (Takara, Dalian, China).

RNA pull-down assay

Biotin-labeled wild type miR-671-5p (WT-bio-miR-671-5p), biotin-labeled mutant miR-671-5p (MUT-bio-miR-671-5p) and control probe (bio-miR-NC) were bought from Thermo Fisher Scientific (Rockford, IL, USA). Briefly, HGC27 and AGS cells were harvested and re-suspended in RIPA lysis buffer. Cell lysates and WT-bio-miR-671-5p, MUT-bio-miR-671-5p or bio-miR-NC were incubated for 1 h, followed by addition of streptavidin agarose beads for 1 h. Finally, qRT-PCR was used for measuring the enrichment of circ_0000620 and MMP2.

Mouse xenograft assay

A mouse xenograft model was established by subcutaneous injection with sh-circ_0000620 or sh-NC-transfected HGC27 cells into nude mice (five for each group). Tumor volume (length × width2/2) was measured once a week. The mice were sacrificed after 5 weeks, then the tumor tissues were resected and weighed. All animal experiments were carried out in line with the protocols approved by the Institutional Animal Care and Use Committee of First Affiliated Hospital of Kunming Medical University.

Statistical analysis

All experiments were repeated for three times with the data exhibition as the mean ± standard deviation. The normality of data distribution was analyzed using the Kolmogorov-Smirnov test, and the homogeneity of variances was assessed using the Levene test. Data were analyzed by SPSS 22.0 (SPSS Inc., Chicago, IL, USA). Difference analysis was conducted through Student’s t-test and one-way ANOVA followed by Tukey’s test, with p level set at < 0.05 (as a statistically significant difference). The linear correlation between circ_0000620 and MMP2 was analyzed by Pearson’s correlation coefficient in clinical tissues. Survival analysis was performed by log-rank test. All figures were plotted using GraphPad Prism software (GraphPad Inc., La Jolla, CA, USA).