All cell lines were purchased from Cell Resource Center of the Chinese Academy of Medical Science (Beijing, China). Gastric epithelium immortalized GES-1 cells were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS), while gastric cancer MKN28, SGC-7901 and AGS cells were cultured in RPMI-1640 medium supplemented with 10% FBS. Cells were incubated in a humidified incubator (37°C, 5% CO₂). CXCL10 was obtained from Sigma-Aldrich (St. Louis, MO, USA).

Antibodies against CXCR3, ERK1/2 and β-actin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibody against phosho-ERK1/2 was obtained from Cell Signaling Technology (Danvers, MA, USA).

Gastric cancer tissues and their corresponding non-cancerous tissues were obtained from the Department of Pathology (n=40), Liaocheng People's Hospital. All specimens were approved by the Committee for Ethical Review of Research in the hospital, and histopathologically confirmed by the pathologist.

Total RNA from the specimens and cancer cell lines were extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions. Reverse transcription PCR was performed with total RNA and First Strand cDNA Synthesis kit (Fermentas, Glen Burnie, MD, USA). Then the cDNA was subjected to real-time PCR with ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Primers used in the real-time PCR are as follows: CXCR3A forward, 5′-ACCCAGCAGCCAGAGCACC-3′ and reverse, 5′-TCA TAGGAAGAGCTGAAGTTCTCCA-3′; CXCR3B forward, 5′-TGCCAGGCCTTTACACAGC-3′ and reverse, 5′-TCGGCGTCATTTAGCACTTG-3′; CXCR3alt forward, 5′-CCAATACAACTTCCCACAGGGGT-3′ and reverse, 5′-GTCTCAGACCAGGATGAATCCCG-3′; β-actin forward, 5′-CATGTACGTTGCTATCCAGGC-3′ and reverse, 5′-CTCCTTAATGTCACGCACG-3′. β-actin expression was used as an internal control. The results were assessed using the 2^−ΔΔCT method.

After washing with cold PBS, cells were lysed in RIPA lysis buffer. Total protein concentration was determined using BCA Protein Assay kit (Pierce Biotechnology, Inc., Rockford, IL, USA). A total of 40 µg protein was resolved by SDS-PAGE and then transferred to PVDF membrane. The membrane was blocked in 5% BSA for 1 h, and incubated with primary antibodies at 4°C overnight. Next, the membrane was incubated with appropriate secondary antibodies and developed by enhanced chemiluminescence (ECL) Plus detection system (Pierce Biotechnology, Inc.).

CXCR3A siRNA (siCXCR3A) and CXCR3B siRNA (siCXCR3B) were purchased from Genchem Biotechnology Co. (Shanghai, China) for transient silence CXCR3A and CXCR3B expression, respectively. A scramble siRNA was purchased as control siRNA (siNC). CXCR3A shRNA (shCXCR3A) was also purchased from Genchem Biotechnology Co. to stably silence CXCR3A expression and a scramble shRNA was used as control shRNA (shNC). Cells were transfected with siRNA or shRNA using Lipofectamine 2000 (Invitrogen). Stable shRNA clone was selected by G418. The knockdown efficiency was determined by real-time PCR.

In vitro invasion and migration assays were performed with 24-well Transwell plates (Corning Incorporated, Corning, NY, USA). The upper filters were coated with 50 µl Matrigel for invasion assay, whereas without Matrigel for migration assay. Cells were resuspended in RPMI-1640 medium at a density of 5×10⁵ cells/ml. A total of 200 µl of cell suspension was added into the upper chambers while 500 µl of RPMI-1640 medium containing 20% FBS was added into the lower chambers. After 12 h, the cells that passed through the membranes were fixed with methanol and stained with crystal blue. Cell number was counted in seven random fields under a light microscope.

Cells were seeded into a 24-well plate at a density of 1×10⁴ cells/well. Further, cells were stimulated with or without CXCL10 and then incubated in the medium for the indicated time. Next, the cells were trypsinized, stained with trypan blue and counted in a hemocytometer.

Cells were seeded in a 96-well plate at a concentration of 1×10³ cells/well. Further, cells were stimulated with or without CXCL10 and then incubated in the medium for the indicated time. Next, CCK-8 was added into the plate and incubated for 2 h. Optical density (OD) was measured by microplate reader (Bio-Rad Model 680) at 490 nm.

After the cell supernatant or tumor tissues in the mice were collected, matrix metalloproteinase (MMP)-13 and IL-6 ELISA kits (Invitrogen) were used to measure the protein level of MMP-13 and IL-6, respectively, according to the manufacturer's instructions.

The male BABL/c nude mice (4-week old) were maintained in the pathogen-free conditions. All procedures were conducted following the Animal Care and Use Committee guidelines of Liaocheng People's Hospital. Cells at a density of 3×10⁵ cells/100 µl were subcutaneously injected in the back of the mice (n=8). The tumor formed in ~3 days. Then the length (L) and width (W) of tumors in mice were measured every week. Tumor volume was calculated with the formula of (L × W²)/2. Eight weeks later, the mice were sacrificed and tumors were lysed to detect MMP-13 and IL-6 expression and ERK1/2 activation. The livers of all mice were fixed in 4% paraformaldehyde and sectioned into slices. Each slice was stained with haematoxylin and eosin (H&E) and then the number of micrometastasis in the livers was observed and counted under a light microscope.

Experiments were performed at least three times. Statistical analysis was performed using SPSS 17.0 (SPSS, Inc., Chicago, IL, USA). Student's t-test was used for comparison between two groups whereas one-way analysis of variance (ANOVA) was used for comparison among multiple groups. P<0.05 was considered as statistically significant.

By western blot analysis, we found that CXCR3 was overexpressed in gastric cancer MKN28, SGC-7901 and AGS cells as compared to gastric epithelium immortalized GES-1 cells (Fig. 1A). The mRNA expression of its three variants in gastric cancer cells were further assessed by real-time PCR. The results showed that CXCR3A expression was upregulated but CXCR3B expression was downregulated in gastric cancer cells. CXCR3alt mRNA expression showed no significant change in any of the detected gastric cancer cells compared with GES-1 cells (Fig. 1B). In addition, CXCR3A expression was increased but CXCR3B expression was decreased in gastric cancer tissues as compared to the corresponding gastric tissues, whereas no significant change was observed for CXCR3alt expression (Fig. 1C).

To investigate the role of CXCR3 in gastric cancer cell invasion and migration in vitro, we stimulated AGS cells with different concentrations of CXCL10 (CXCR3 ligand) to activate CXCR3. Invasion and migration assays showed that CXCL10 promoted the invasion and migration of gastric cancer cells, in a dose-dependent manner (Fig. 2A and B), indicating that activation of CXCR3 may be involved in gastric cancer cell invasion and migration in vitro.

We then focused on the role of CXCR3A and CXCR3B in gastric cancer cells, as they both showed significant changes in gastric cancer tissues and cells. After confirming the specific knockdown efficiency of CXCR3A and CXCR3B in AGS cells (Fig. 3A), the effects of CXCR3A and CXCR3B on cell invasion and migration were detected by in vitro invasion and migration assay. We found that CXCL10 (100 ng/ml) stimulated the invasion and migration in siNC cells. In contrast, knockdown of CXCR3A inhibited CXCL10-induced cell invasion and migration, whereas knockdown of CXCR3B had little effect on the invasion and migration (Fig. 3B and C). These results confirm that it is CXCR3A, but not CXCR3B, that participates in gastric cancer cell invasion and migration in vitro.

We then found that MMP-13 and IL-6 secretion was increased in siNC cells after CXCL10 stimulation for 12 h. However, knockdown of CXCR3A greatly suppressed the CXCL10-induced MMP-13 and IL-6 secretion (Fig. 4A and B). Moreover, after stimulated with CXCL10 for 30 min, ERK1/2 kinases were activated in siNC cells. In contrast, the CXCL10-mediated ERK1/2 activation was markedly attenuated in siCXCR3A cells (Fig. 4C).

Next, we found that CXCL10 treatment dose-dependently promoted the growth of AGS cells (Fig. 5A and B). To test the role of CXCR3A and CXCR3B in the growth of gastric cancer cells in vitro, cells were transfected with siNC, siCXCR3A and siCXCR3B, respectively, and then stimulated with 100 ng/ml CXCL10 for 72 h. The results showed that knockdown of CXCR3A inhibited the CXCL10-mediated growth of AGS cells, whereas knockdown of CXCR3B promoted cell growth in vitro (Fig. 5C and D).

Further, an AGS cell clone that stably silenced the expression of CXCR3A was established and confirmed by real-time PCR (Fig. 6A). Then the shCXCR3A and shNC cells were injected subcutaneously at the back of the mice, respectively. The tumor volume was assessed through measuring the length and width of the tumor in mice each week. The results showed that the tumor size in shCXCR3A group was much smaller than that in shNC group (Fig. 6B). Eight weeks later, the mice were sacrificed, micrometastasis in the liver section was counted under a microscope (Fig. 6C). We found that liver metastasis was observed in all eight mice in the shNC group, whereas it was observed in six mice in the shCXCR3A group. We counted the number of liver micrometastasis in all H&E slices, and found that knockdown of CXCR3A decreased the number of micrometastasis in the liver (Fig. 6D). In addition, we found that the expression of MMP-13 and IL-6 in tumor tissues of shCXCR3A group were significantly decreased (Fig. 7A and B), and the activation of ERK1/2 was greatly inhibited (Fig. 7C), as compared to tumor tissues of shNC group.