Gastric cancer tissues and normal adjacent tissues were collected from patients who underwent curative resection at the Department of Surgery, the Second Affiliated Hospital of Wenzhou Medical University, Wenzhou, China between 2013 and 2015. All samples were collected after obtaining written informed consent. The tissues were immediately transported into liquid nitrogen in operating theatres and then stored at −80°C. The study protocol was approved by the Ethics Committee of the Second Affiliated Hospital of Wenzhou Medical Univesity, and was carried out according to the Declaration of Helsinki.

The human normal gastric epithelial cell line, GES-1, and the gastric cancer cell lines, SGC-7901, HGC-27, BGC-823 and MKN45, were purchased from the Cell Bank of the Chinese Academy of Sciences. The 293T cells were a kind gift from Wenzhou Medical University. The cells were maintained in DMEM (GES-1, SGC-7901 and 293T), RPMI-1640 medium (BGC-823 and MKN45), or MEM (HGC-27) supplemented with 10% fetal bovine serum (all from HyClone, Salt Lake City, UT, USA), 100 units/ml penicillin and 100 units/ml streptomycin (Sigma, St. Louis, MO, USA). All these cells were cultured in a humidified environment containing 5% CO₂ and maintained at a constant temperature of 37°C.

Total RNA for CMTM3 and β-actin was extracted from the clinical samples using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The concentration of the RNA was measured using a NanoDrop spectrophotometer at 260/280 nm (Thermo Fisher Scientific, Waltham, MA, USA) and the RNA quality was determined via electrophoresis. First-strand cDNA was synthesized PrimeScript RT Reagent kit (Takara, Mountain View, CA, USA) according to the manufacturer's instructions. Briefly, 4 µl of isolated RNA (30 µg) was first mixed with 1 µl of random hexamer primer and 7 µl of RNAse-free H₂O and then incubated at 65°C for 5 min. Subsequently, the microtubes were cooled on ice followed by the addition of 4 µl of reaction buffer, 1 µl of RNase inhibitor, 2 µl of dNTP mix, and 1 µl of reverse transcriptase to each sample. The samples were immediately incubated at 25°C for 5 min and then at 42°C for 60 min. Finally, the reaction was terminated by heating the samples at 70°C for 5 min. The reverse transcription reaction was performed with the final volume of 20 µl per tube. Quantitative PCR was performed using SYBR-Green Ex Taq™ master mix (Takara). The quantitative analysis was carried out using iQ 5 Real-Time PCR System (Bio-Rad, Hercules, CA, USA). The real-time PCR conditions were as follows: 50°C for 2 min, 95°C for 10 min, then 40 cycles at 95°C for 15 sec, and 60°C for 1 min. The primers used for β-actin and CMTM3 were as follows: β-actin forward, 5′-GGCACTCTTCCAGCCTTCC-3′; and reverse, 5′-GAGCCGCCGATCCACAC-3′; and CMTM3 forward, 5′-TCTTGCGTGTGAATCTCTTACC-3′; and reverse, 5′-CAGGATCCACATTGGTGTTACC-3′. Total RNA for miR-135b-5p and U6 was extracted from the clinical samples and cells using the miRNeasy mini kit (Qiagen, Hilden, Germany). The RT-qPCR reactions of miR-135b-5p and U6 were performed according to the manufacturer's instructions of the All-in-One™ miRNA qRT-PCR Detection kit (GeneCopoeia, Rockville, MD, USA). iQ-5 (Bio-Rad) was used to monitor all these RT-qPCR reactions. RNA expression was relative quantified using 2-ΔΔCt method.

Total protein was extracted using RIPA protein lysis buffer (Beyotime, Shanghai, China) with 1% protease inhibitor cocktail and 1 mM phenylmethylsulfonyl fluoride (PMSF). Cell fractions were prepared using a Nuclear and Cytoplasmic Protein Extraction kit (Beyotime) according to the manufacturer's instructions. Generally, 50 µg of protein were used for western blotting. Samples were separated by SDS-PAGE and transferred onto PVDF membranes (Thermo Fisher Scientific, Waltham, MA, USA). After blocking in 5% skim milk, the PVDF membranes were incubated with primary antibodies in blocking buffer overnight at 4°C and then with HRP-conjugated secondary antibody for 2 h. The primary antibodies used were as follows: anti-β-tubulin (1:5,000 dilution, sc-23949; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-CMTM3 (1:1,000 dilution, ab198016; Abcam, Cambridge, UK), anti-caspase 3 [1:1,000 dilution, 9662; Cell Signaling Technology (CST), Danvers, MA, USA], anti-poly(ADP-ribose) polymerase (PARP; 1:1,000 dilution, 9532; CST), anti-caspase 9 (1:1,000 dilution, ab32539; Abcam), anti-cyclin B1 (1:1,000 dilution, ab32053; Abcam), anti-p21 (1:1,000 dilution, ab109520; Abcam). The membranes were then incubated with HRP-conjugated goat anti-mouse IgG (sc-2005; Santa Cruz Biotechnology) or HRP-conjugated goat anti-rabbit IgG (sc-2004; Santa Cruz Biotechnology). Reactive bands were visualized with ECL reagent (Pierce, Rockford, IL, USA) and analyzed. Protein expression was quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

The frozen tumor tissues and adjacent tissues thawed in 4°C prior to fixation, and washed with PBS twice. The tissues were then fixed in 3.7% formalin and embedded in paraffin, and were then cut into 4-µm-thick serial sections. The paraffin-embedded tissue sections were dewaxed, rehydrated and placed in 10 mmol/l citrate buffer (pH 6.0), and heated twice in a microwave oven for 5 min each. The sections were incubated with 3% H₂O₂ for 10 min, washed with PBS, blocked with 10% normal goat serum for 30 min, and then incubated with 4 mg/l purified anti-CMTM3 (ab198016; Abcam) or normal rabbit IgG (ab172730; Abcam) as a control at 4°C overnight. After washing, the sections were stained with the catalyzed signed amplification system kit (Agilent Technologies, Santa Clara, CA, USA) and visualized with a Nikon E800 microscope (Nikon, Tokyo, Japan) and images were acquired.

miR-135b-5p mimics and scramble control mimics were purchased from GeneCopoeia Inc. The target gene of miR-135b-5p, was firstly predicted by TargetScan (http://www.targetscan.org/mamm_31/), miRDB (http://www.mirdb.org/) and miRanda (http://www.microrna.org/), respectively. CMTM3 was then filtered out from the inter-section of above 3 prediction results. Wild-type CMTM3 3′UTR (CMTM3-3′UTR-wt) and miR-135B-5p target site deletion mutation CMTM3 3′UTR (CMTM3-3′UTR-mu) were constructed into the psiCHECK2 plasmid (Promega, Madison, WI, USA). The 293T cells (10⁵ cells) were seeded in 24-well plates prior to transfection. According to the instructions of the manufacturer, mimics and CMTM3 3′UTR reporter plasmids (psiCHECK-CMTM3-3′UTR-wt or psiCHECK-CMTM3-3′UTR-mu) were co-transfected using Lipofectamine^® RNAiMAX (Thermo Fisher Scientific, MA, USA) with a final concentration of 50 nM (mimics) or 200 ng (PrLZ 3′UTR reporter plasmid). After 48 h, the cells were collected and Luciferase activity was detected using the Dual-Glo luciferase assay system (Promega) according to the manufacturer's instructions.

miR-135b-5p inhibitor lentivirus and CMTM3 overexpression lentivirus were purchased from Shanghai R&S Biotechnology Co., Ltd. The SGC-7901 cells were planted into 10 cm dishes (10⁶ cells/dish) 24 h prior to infection. Lentiviral infection was performed at a multiplicity of infection (MOI) of 50. The infection efficiency was deter-mined by counting the number of GFP-positive cells which should be guaranteed to be >90%.

The cells were seeded in 96-well plates in triplicate at densities of 3,000/well. Cell viability was evaluated at the desired time points using CCK8 kits (Dojindo Molecular Technologies, Kumamoto, Japan) according to the manufacturer's instructions. Light absorbance of the solution was measured at 450 nm using a microplate reader (PR4100; Bio-Rad).

Transwell chambers coated with Matrigel (BD Biosciences, San Jose, CA, USA) were used for the analysis of cell invasion. A total of 3×10⁵ SGC-7901 cells in 100 µl serum-free DMEM were seeded on the upper chambers and DMEM with 10% FBS was added to the lower chambers. After 24 h of incubation, the invaded cells in the lower side of the membranes were fixed with methanol and stained with crystal violet (Beyotime). Images were acquired using an inverted microscope. Invaded cells were counted from three different fields. The experiment was repeated three times.

For the analysis of cell apoptosis, the transfected cells were harvested, washed and resuspended in 1 ml of binding buffer. The cells were then stained with 5 µl of Annexin V-APC (BD Biosciences) and 10 µl of propidium iodide (Sigma-Aldrich) in the dark for 15 min at room temperature and analyzed using a flow cytometer (BD Biosciences) equipped with CellQuest software.

Propidium iodide (PI) staining with flow cytometry was used to assess cell cycle distribution. Briefly, the lentivirus-infected SGC-7901 cells were released by trypsinization, collected, washed with cold PBS, and then fixed in 70% ethanol at 4°C overnight. The fixed cells were then suspended in 250 µl of RNase A buffer (100 ng/ml), and labeled with a 2X solution of PI (100 ng/ml) for 30 min at 4°C. Finally, the stained cells were analyzed using a flow cytometer (BD Biosciences) equipped with CellQuest software.

A total of 30 female BALB/c nude mice (6 weeks old, weighing 18-22 g) were purchased from Nanjing Biomedical Research Institute of Nanjing University (Nanjing, China). These BALB/c nude mice were injected with the SGC-7901 cells subcutaneously (5×106 cells/mouse) into the right posterior shoulder area. For the lentiviral infection groups, 10 µl lentivirus were respectively injected into the tumors every 3 days when the tumor volume up had reached up to 50 mm³. The tumor sizes were measured every 3 days using micrometer calipers, and tumor volumes were calculated as follows: Tumor volume = d²×D/2, where d and D represented shortest and the longest diameters, respectively. At 22 days after the first injection of the lentivirus, the mice were euthanized and tissue was collected. The tissues were immediately transported into liquid nitrogen and then stored at −80°C for subsequent pathology and gene expression detection. All animal experiment protocols were approved by the Institutional Animal Care and Use Committee of Wenzhou Medical University.

The results are presented as the means ± standard deviations (SD) of 3 independent experiments. Significant differences in mean values were evaluated by an unpaired t-test. One-way ANOVA with Tukey's post hoc test was used to compare continuous variables among 2 or more groups. Tests of association were conducted using Pearson's χ² test. A value of P<0.05 was considered to indicate a statistically significant difference.

RT-qPCR and IHC were used to examine the expression pattern of CMTM3 and miR-135b-5p in gastric cancer. The results of RT-qPCR revealed that CMTM3 exhibited a lower expression pattern in gastric cancer tissues compared with the adjacent tissues, while the expression of miR-135b-5p was significantly higher in gastric cancer tissues compared with the adjacent tissues (Fig. 1A and B). Moreover, IHC yielded similar results, also showing that CMTM3 was markedly downregulated in gastric cancer tissues (Fig. 1C). We then performed RT-qPCR to verify the differences in the expression of CMTM3 and miR-135b-5p between the normal gastric epithelial cell line, GES-1, and gastric cancer cell lines (HGC-27, SGC-7901, BGC823 and MKN45). Our results indicated that CMTM3 expression was higher in the GES-1 cells than that in the HGC-27, SGC-7901, BGC823 and MKN45 cells (Fig. 1D). By contrast, the miR-135b-5p expression level was significantly lower in the GES-1 cells compared with the HGC-27, SGC-7901, BGC823 and MKN45 cells (Fig. 1E).

To verify the regulatory association between miR-135b-5p and CMTM3, we first analyzed their expression correlation in gastric cancer tissues. Pearson's correlation analysis revealed that CMTM3 expression negatively correlated with miR-135b-5p expression (Fig. 2A). To identify the potential target site of miR449a, we used a combination of three algorithms, TargetScan, miRDB and miRanda. The target site of miR-135b-5p in CMTM3 3′-UTR predicted by using TargetScan, miRDB and miRanda is shown in Fig. 2B. We further confirmed the target association between miR-135b-5p and CMTM3 by Luciferase assay (Fig. 2C). Our results indicated that CMTM3 was specifically targeted by miR-135b-5p.

To verify the effect of CMTM3 on gastric cancer cells in vitro, we needed to select a representative cell line. As shown in Fig. 1D and E, CMTM3 expression decreased most significantly in the SGC-7901 cells compared with the GES-1 cells and other gastric cancer cell lines, and miR-135b-5p expression increased most significantly in the SGC-7901 cells. Thus, the SGC-7901 cells were selected for use in the subsequent experiments. CMTM3 overexpression lentivirus was used to introduce the exogenous expression of CMTM3 in the SGC-7901 cells. The effect of CMTM3 overexpression was validated by western blot analysis (Fig. 3A and B). We then performed cell proliferation assay using the CCK-8 proliferation kit. Our results suggested that CMTM3 inhibited SGC-7901 cell proliferation (Fig. 3C). In addition, flow cytometry was used to assess the function of CMTM3 in SGC-7901 cell apoptosis and cell cycle progression, and the results indicated that CMTM3 overexpression significantly promoted SGC-7901 cell apoptosis (Fig. 3D and E). Furthermore, CMTM3 overexpression markedly suppressed SGC-7901 cell cycle progression (Fig. 3F and G). Transwell invasion assays also indicated that CMTM3 overexpression inhibited the invasiveness of the SGC-7901 cells (Fig. 3H and I).

In order to investigate the function of miR-135b-5p in gastric cancer progression, miR-135b-5p inhibitor lentivirus was used to neutralize miR-135b-5p in the SGC-7901 cells. We found that miR-135b-5p expression was decreased in the SGC-7901 cells following infection with miR-135b-5p inhibitor lentivirus (Fig. 4A), while CMTM3 was significantly upregulated (Fig. 4B and C). Proliferation assay revealed that infection with miR-135b-5p inhibitor inhibited SGC-7901 cell proliferation (Fig. 4D). Flow cytometry was also used to assess the effects of miR-135b-5p inhibitor on SGC-7901 cell apoptosis and cell cycle progression. The results indicated that infection with miR-135b-5p inhibitor significantly promoted SGC-7901 cell apoptosis (Fig. 4E and F). Moreover, infection with miR-135b-5p inhibitor markedly suppressed SGC-7901 cell cycle progression (Fig. 4G and H). Transwell invasion assays also indicated that infection with miR-135b-5p inhibitor suppressed the invasiveness of the SGC-7901 cells (Fig. 4I and J). The above-mentioned results indicated that miR-135b-5p promoted gastric cancer progression by downregulating CMTM3 expression.

In order to confirm the effects of the miR-135b-5p/CMTM3 axis on cell proliferation, apoptosis and invasion, we further examined the indicators of proliferation, apoptosis and invasion by western blot analysis. The results revealed that CMTM3 overexpression increased the level of cleaved caspase 3, caspase 9 and cleaved PARP in the SGC-7901 cells, which indicated that CMTM3 promoted gastric cancer cell apoptosis (Fig. 5B). The expression of p21, which has been reported to be an indicator of cell cycle arrest (21), was significantly increased after CMTM3 was overexpressed in the SGC-7901 cells. However, the expression of cyclin B1, an important regulatory protein involved in mitosis (22), was markedly decreased in the SGC-7901 cells following infection with CMTM3 overexpression lentivirus (Fig. 5B). Moreover, infection with miR-135b-5p inhibitor, which upregulated CMTM3 expression in the SGC-7901 cells, led to a similar effect on the expression of these markers as observed with CMTM3 overexpressoin (Fig. 5A). These results indicated that miR-135b-5p promoted SGC-7901 cell malignant phenotype via targeting CMTM3.

Nude mice with SGC-7901 subcutaneously transplanted tumors were used to verify the function of miR-135b-5p in vivo. miR-135b-5p inhibitor lentivirus was used to neutralize miR-135b-5p in the tumors, and we found that the injection of the miR-135b-5p inhibitor lentivirus markedly suppressed tumor formation in vivo (Fig. 6A–C). We further examined miR-135b-5p and CMTM3 in the transplanted tumors from the nude mice, and the results of RT-qPCR revealed that miR-135b-5p expression was decreased in tumors after the injection of miR-135b-5p inhibitor lentivirus (Fig. 6D). Moreover, IHC revealed that CMTM3 expression was markedly increased in the tumors injected with miR-135b-5p inhibitor lentivirus (Fig. 6E). These results indicated that miR-135b-5p promoted gastric cancer progression in vivo.