Written informed consent was collected from all patients with gastric cancer and the present study was approved by the ethics committee of The First Affiliated Hospital of Anhui Medical University (Hefei, China). The patients were aged between 43 and 68 and included 10 males and 7 females. The samples were collected between June and October 2012 by tumor resection. A total of 17 gastric cancer tissues and their matched adjacent normal tissues were obtained from the Department of General Surgery, The First Affiliated Hospital of Anhui Medical University (Hefei, China).

Human AGS and SGC-7901 gastric cancer cells, in addition to the human normal GES-1 gastric mucosal epithelial cell line were obtained from the Cell Bank of Central South University (Changsha, China). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Invitrogen Life Technologies), 100 IU/ml penicillin (Sigma-Aldrich, St. Louis, MO, USA) and 100 µg/ml streptomycin sulfate (Sigma-Aldrich) at 37°C with 5% CO₂.

The Robo1 small interfering RNA (siRNA), Robo1 plasmid, miR-29a mimic, miR-29a inhibitor and scramble miRNA were purchased from Nlunbio (Changsha, China). Cells in the logarithmic growth phase were seeded into 6-well plates. Transfection of cells with these oligonucleotides was performed using Lipofectamine 2000 (Life Technologies, Carlsbad, CA, USA) according to the manufacturer's instructions.

Total RNA was extracted from cells with TRIzol reagent (Life Technologies) according to the manufacturer's instructions. The relative expression of miR-29a was determined by RT-qPCR using the mirVana™ qRT-PCR microRNA Detection kit (Life Technologies) following the manufacturer's instructions. Specific primer sets for miR-29a and U6 (internal reference) were obtained from Life Technologies. Expression of Robo1 mRNA was determined by RT-qPCR using the standard SYBR Green RT-PCR kit (Takara Bio, Inc., Otsu, Japan) following the manufacturer's instructions. The specific primer pairs used were as follows: Robo1, sense 5′-CTTACACCCGTAAAAGTGACGC-3′ and antisense 5′-TGGTCTCTCTAAGACAGTCAGC-3′; GAPDH as an internal control, sense 5′-CTGGGCTACACTGAGCACC-3′ and antisense 5′-AAGTGGTCGTTGAGGGCAATG-3′. The following cycling conditions were used: 95°C for 10 min and 40 cycles of denaturation at 95°C for 15 sec and an annealing/elongation step at 60°C for 60 sec. Reverse transcription was performed at 16°C for 30 min, followed by an incubation step at 42°C for 30 min and enzyme inactivation at 85°C for 5 min. An ABI 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) was used. The relative expression of Robo1 mRNA or miR-29a was quantified using GraphPad Prism software, version 4.0 (GraphPad Software, Inc., La Jolla, CA, USA) and the 2^−ΔΔCt method.

Cell migration and invasive capabilities were determined by performing a Transwell assay (Chemicon International, Inc., Temecula, CA, USA). For the migration assay, cells were plated in the upper chamber with the non-coated membrane. For the invasion assay, the chamber was coated with Matrigel (Becton Dickinson, Franklin Lakes, NJ, USA) and dried overnight, then cells were plated in the upper chamber with the Matrigel-coated membrane. For the two assays, DMEM supplemented with 10% FBS was used as a chemoattractant in the lower chamber. Following 24 h of incubation at 37°C, cells on the lower face of the membrane were stained with 0.4% crystal violet for 20 min and observed under a microscope (CX22; Olympus, Tokyo, Japan).

Total protein was extracted using radioimmunoprecipitation assay solution (Sigma-Aldrich). The protein concentration was determined by the Bradford DC protein assay (Bio-Rad Laboratories, Inc., Hercules, CA, USA). For determination of the protein level, proteins were separated with 10% SDS-PAGE and were blotted onto polyvinylidene difluoride (PVDF; Invitrogen Life Technologies) membranes, which were then incubated in Tris-buffered saline with Tween-20 (Sigma-Aldrich) with 50 g/l skimmed milk at room temperature for 4 h. Subsequently, the PVDF membranes were incubated with mouse anti-Robo1 (cat. no. ab201632; 1:50 dilution; Abcam, Cambridge, MA, USA) and mouse anti-GAPDH antibodies (cat. no. sc-365062; 1:50 dilution; Santa Cruz Biotechnology, Inc., Dallas, TX, USA), respectively, at room temperature for 1 h. Following washing with PBST three times, the PVDF membranes were incubated with peroxidase-conjugated rabbit anti-mouse secondary antibody (cat. no. ab175743; 1:20,000 dilution; Abcam) for 1 h at room temperature. Chemiluminescent detection was performed with an ECL kit (Pierce Chemical Company, Rockford, IL, USA).

TargetScan (Release 6.2, www.targetscan.org) was used to predict the putative target genes of miR-29a.

SGC-7901 cells were cotransfected using Lipofectamine 2000 with the reporter constructs Robo1-3′ untranslated region (UTR)-psi-CHECK2 (containing the 3′-UTR of Robo1 including the miRNA-29a binding sites) or mutant-Robo1-3′UTR-psi-CHECK2 (containing the corresponding mutated sequence of 3′-UTR of Robo1), while miR-29a mimics or scramble miRNA were used as the negative control. Luciferase activity was determined after 48 h using the Dual-Glo Substrate system (Promega Corporation, Madison, WI, USA) and the LD400 Luminometer (Beckman Coulter, Inc., Brea, CA, USA). Data are presented as the ratio of Renilla luciferase to Firefly luciferase.

Data are expressed as the mean ± standard deviation. Differences between two groups were determined using Student's t-test. Analysis was performed using SPSS software, version 15.0 (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

The endogenous expression of miR-29a was determined in human gastric cancer tissues and their adjacent normal tissues using RT-qPCR. As demonstrated in Fig. 1A, the relative expression of miR-29a was frequently downregulated in 77.8% (14/18) of gastric cancer tissues compared with the corresponding adjacent normal tissues. Similarly, it was identified that expression of miR-29a was significantly reduced in AGS and SGC-7901 gastric cancer cell lines, compared with that of the GES-1 human gastric mucosal epithelial cell line (Fig. 1B). It is suggested that downregulation of miR-29a may be associated with the development and progression of gastric cancer.

Based on the results obtained, whether miR-29a serves a role in the regulation of migration and invasion of gastric cancer cells was further investigated. AGS human gastric cancer cells were transfected with miR-29a mimics or the negative control, respectively. As indicated in Fig. 2A, transfection of miR-29a mimics resulted in a significant increase in miR-29a expression, when compared with the negative control transfected with scramble miRNA. In addition, it was further identified that restoration of miR-29a led to reduced migration and invasion of gastric cancer cells, when compared with the control cells (Fig. 2B). Accordingly, it is suggested that miR-29a acts as a tumor suppressor and contributes to inhibition of migration and invasion of gastric cancer cells.

Targetscan prediction software was further used to identify potential targets for miR-29a. Robo1 was identified as a putative target gene for miR-29a, which has been reported to participate in the regulation of cell migration and invasion (13). To further confirm this prediction, the mRNA and protein expression levels of Robo1 were determined in AGS cells transfected with miR-29a mimic. It was observed that the mRNA and protein expression of Robo1 was significantly downregulated following overexpression of miR-29a, compared with the negative control (Fig. 3A and B). Furthermore, the mRNA levels of Robo1 in gastric cancer and adjacent normal tissues were determined with RT-qPCR, then the significance of the miR-29a/Robo1 correlation was assessed in gastric cancer. As presented in Fig. 3C, the mRNA expression of Robo1 was frequently increased in gastric cancer tissues, compared with the matched adjacent normal tissues. Furthermore, as demonstrated in Fig. 3D, a significant inverse correlation was observed between miR-29a and Robo1 mRNA expression. Accordingly, the data suggest that miR-29a negatively regulates Robo1 gene expression and Robo1 is a potential target gene of miR-29a.

To verify whether Robo1 3′-UTR is a target of miR-29a, a luciferase reporter assay was conducted. The vectors containing the wide type of Robo1 3′UTR as well as the mutant type of Robo1 3′UTR were constructed (Fig. 4A). As demonstrated in Fig. 4B, cotransfection of AGS cells with Robo1-3′-UTR/pmirGLO and miR-29a mimics resulted in a significant reduction in the luciferase activity, when compared with the negative control (P<0.05). This inhibitory effect was abrogated by point mutations in the core binding sites of the Robo1 3′-UTR. All of these observations indicated that miR-29a exerts suppressive effects on Robo1 expression via binding to the 3′ UTR of Robo1.

To verify the effect of Robo1 on gastric cancer cell migration and invasion, the expression of Robo1 was observed to be downregulated by Robo1 siRNA. As presented in Fig. 5A and B, the mRNA and protein expression levels of Robo1 were notably downregulated following knockdown of the Robo1 gene in AGS cells. Consistently, downregulation of Robo1 markedly inhibited migration and invasion in AGS cells (Fig. 5C), which resembled the suppressive effects of miR-29a. These observations suggest that miR-29a may inhibit gastric cancer cell migration and invasion via targeting Robo1.