The human gastric cancer cell lines SGC7901, GES-1, AGS, MKN45 and MKN28 were bought from the China Center for Type Culture Collection (Wuhan, China). Cells were routinely maintained in RPMI-1640 medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS) and 100 U/ml penicillin/streptomycin at 37°C. The cells were cultured on tissue culture dishes coated with 10 mg/ml fibronectin.

Four- to 6-week-old female/male athymic BALB/c nu/nu mice were obtained from the Central Laboratory of Animal Science at Wuhan University (Wuhan, China) and housed in laminar-flow cabinets under specific pathogen-free conditions, with food and water ad libitum. All experiments on mice were conducted in accordance with the guidelines of National Institutes of Health (NIH; Bethesda, MD, USA) for the care and use of laboratory animals. The study protocol was also approved by Shanghai Medical Experimental Animal Care Committee (China).

Small interfering RNA (siRNA; GeneChem, Shanghai, China) against OPN (GenBank accession no. NM_000582) mediated by lentivirus and its mismatch mutants as negative control were designed using an Internet application system (Invitrogen, Carlsbad, CA, USA). These target sequences tagged with green fluorescent protein (GFP) were annealed and inserted into pGC-LV expression vector (Invitrogen), and then cotrans-fected into 293FT cells with packaging vectors (pHelper 1.0 and pHelper 2.0; Invitrogen) using Lipofectamine 2000 (Invitrogen). After a 48-h incubation period post transfec-tion, viral supernatants were collected, concentrated, and stored at −70°C. MKN45 and MKN28 cells were infected with OPN-specific siRNA lentivirus or control lentivirus with a multiplicity of infection (MOI) from 5 to 20 in the presence of polybrene (10 mg/ml), respectively.

The cells were plated in 96-well plates and maintained at 37°C in a humidified incubator. The cells were harvested by trypsinization every day for 5 or 6 days and counted with a hemocytometer.

Cells were suspended at each chosen time-point. Cells (2×10⁶) were then centrifuged and washed twice with ice-cold phosphate-buffered saline (PBS). Apoptotic cells were detected by flow cytometric analysis using Annexin-V fluorescein and propidium iodide (PI) (Molecular Probes, Invitrogen, Eugene, OR, USA).

Transwell invasion chambers (Millipore Corp., Bedford, MA, USA) consisting of a thin layer of Matrigel basement membrane matrix were used. The upper surface of filters was 6.4 mm in diameter with 8-μm pores. In the upper chambers, 1×10⁵ cells in 200 μl of serum-free medium was added. Then, 500 μl of RPMI-1640 medium containing 20% fetal calf serum was added to the lower chambers as a chemoattractant. After incubation for 24 h at 37°C in a humid atmosphere of 5% CO₂ and 95% air, the cells that had invaded through the membrane and had reached the lower surface of the membrane were fixed with methanol and stained with hematoxylin. Photographs were taken and stained cells were counted under a microscope in five randomly chosen fields.

Cells were cultured in fibronectin-coated 6-well tissue culture plates as confluent monolayer. The near confluent cell monolayer was then carefully scratched using 20 μl sterile pipette tips. Cell debris was removed by washing with PBS. The dishes were incubated in fresh regular culture medium containing hydroxyurea (2 mM) for additional 24 h and photographs were taken under a light microscope.

Proteins were extracted from the cultured cells and then quantitated. Equal amounts of protein (30 /Ag) from different cells were separated by a 10% SDS-polyacrylamide gel, transferred onto nitrocellulose paper (Millipore, Billerica, MA, USA) and then incubated overnight at 4°C with a 1 μg/ml of antibody against target proteins. Protein bands were detected by the enhanced chemiluminescence (ECL) kit (Pierce, Rockford, IL, USA) and exposure to Biomax ML film (Eastman Kodak, Rochester, NY, USA). The primary antibodies against mitogen-activated protein kinase (MEK), phosphorylated MEK (p-MEK), extracellular signal-regulated kinase 1/2 (ERK1/2), p-ERK1/2, Akt, p-Akt (Cell Signaling Technology, Danvers, MA, USA), MMP-2, MM-9, uPA, p65 NF-κB, Caspase-3, lamin B and actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used.

Electrophoretic mobility shift assay (EMSA) was performed using an EMSA Gel-Shift kit (Pierce). Nuclear extracts were prepared and incubated with NF-κB-³²P-labeled oligonucleotide (5′-AGCTATGTG GGTTTTCCCATGAGC-3′) for 30 min at room temperature. Samples were electrophoresed at 100 V and 4°C, transferred to Biodyne nylon membranes (Pierce), and then cross-linked in an ultraviolet cross-linker (Stratagene Inc., La Jolla, CA, USA). Gels were visualized using streptavidin-HRP followed by chemiluminescence detection.

The medium was collected to concentrate using Centricon centrifugal filters devices with YM-30 membranes (Millipore) and the protein concentration was detect using Bio-Rad assay. Protein (10 μg) was loaded onto zymographic sodium dodecyl sulfate gel containing gelatin (1 mg/ml). The gels were washed twice in 2.5% (v/v) Triton X-100 at room temperature before incubation in 50 mM Tris-HCl (pH 7.6), 10 mM CaCl₂, 50 mM NaCl and 0.05% Brij35 overnight at 37°C. The gels were stained with 0.5% Coomassie brilliant blue R-250 for ≥1 h and destained with methanol/acetic acid to reveal zones of gelatinase activity. The gels were analyzed using the Typhoon™ 9400 Scanner (GE Healthcare). Image quantification was performed using ImageQuant ™ TL software (GE Healthcare).

Nude mice were divided into three groups, six mice in each group. To evaluate in vivo tumor growth, 1×10⁷ OPN silencing cells, mock virus-transfected cells or untransfected cells (MKN28) were injected subcu-taneously into the nude mice. Tumor growth was monitored with tumor volume, which was calculated as follows: V (mm³) = width² (mm²) × length (mm)/2. The mice were sacrificed 6 weeks later, and tumors were harvested. The tumor weight was recorded.

Nude mice were injected subcutane-ously with 1×10⁷ OPN silencing cells, mock virus-transfected cells or untransfected cells (MKN28). After 15 days, the tumors were extirpated. Cancer tissue was divided into small pieces, ∼1 mm in diameter, for stomach surgical orthotopic implantation. During surgical implantation, a small area of the gastric serosa was removed, and 1 mm³ tumor fragment per mouse was implanted. Tumor pieces were penetrated and sutured to the gastric wall by a nylon surgical suture. Two months later, all the mice were sacrificed, individual organs were excised, and metastases were checked by hematoxylin-eosin (H&E) staining. The incidence and classification of metastasis were calculated and evaluated independently by two pathologists.

For statistical evaluation of cell culture experiments, mean values and standard deviation of at least triplicate experiments were calculated. Student’s t-test was used to compare the difference of mean values. Difference in the incidence of liver and lung metastasis was analyzed by Fischer’s exact test. P-values <0.05 were defined as statistically significant.

As shown in Fig. 1, the expression levels of OPN in MKN45 and MKN28 cell lines were much higher than that in SGC7901 and GES-1 cells, while the lowest level of OPN among the 5 gastric cancer cells tested was in AGS cells. Therefore, MKN45, MKN28 and AGS cells were selected and used in the following experiments.

OPN expression was inhibited in MKN45 and MKN28 cells which had higher expression levels of OPN, and OPN expression was upregulated in AGS cells which slightly expressed OPN. First, OPN expression was inhibited in MKN45 and MKN28 cells by transfection with OPN-specific siRNA for 24 h. OPN silencing significantly inhibited proliferation and promoted apoptosis in both cell lines compared with the control (P<0.05, OPN silencing relative to untransfected or mock control, Fig. 2A and B). Then, OPN expression was upregulated in AGS cells by treatment with 10 μM exogenous recombinant OPN for 24 h. OPN significantly promoted proliferation and inhibited apoptosis in AGS cells (P<0.05, OPN upregulation relative to control, Fig. 2C). These data show that the OPN status strongly affects proliferation and apoptosis in gastric cancer cells.

Cell invasion and migration were analyzed as described in Materials and methods. OPN silencing significantly inhibited invasion and migration in MKN45 and MKN28 cells (P<0.05, OPN silencing relative to untransfected or mock control, Fig. 3A and B). OPN expression significantly promoted invasion and migration in AGS cells (P<0.05, OPN upregulation relative to control, Fig. 3C). Our findings suggest that the OPN status significantly affects invasion and migration in gastric cancer cells.

MKN28 (parental control), control lentivirus-infected cells (mock control) and OPN-specific siRNA lentivirus-infected cells (si-OPN) were subcutaneously implanted into nude mice, respectively. Tumor growth curves were plotted in the course of the experiment (Fig. 4A), and the weight of tumor was evaluated at 6 weeks post-inoculation (Fig. 4B). Obvious tumor masses could be found in the two control groups of mice inoculated with MKN28 cells (2.76±0.43 g) or mock control cells (2.97±0.52 g), while much smaller tumors were detected in the si-OPN groups (0.45±0.22 g) (P<0.05, si-OPN groups relative to mock control or MKN28 groups). These data suggest that the treatment of OPN siRNA blocked the growth of tumor compared to control groups, the parental or the mock-transfected groups. In addition, compared with the control groups, no significant body weight change or other toxicity was observed in the treatment groups (data not shown).

The effect of OPN silencing on tumor metastasis was further assayed in athymic mice. Autopsy was performed, and the incidence of metastasis in the livers, lungs and other organs was determined by macroscopic and histologic examinations. The incidence of hepatic metastasis in mice of the parental and the mock-transfected groups was 60% (6 of 10) and 70% (7 of 10), respectively. In the si-OPN groups, only 30% (3 of 10) of animals had hepatic metastasis (Fig. 4C). Pulmonary metastases were found in 30% (3 of 10) of the parental groups, 40% (4 of 10) of the mock-transfected groups and 10% (1 of 10) of the si-OPN groups (Fig. 4C). As shown in Fig. 4D, the numbers of hepatic metastatic lesions in the si-OPN groups (4.2±3.9) were greatly decreased compared with those in the mock-transfected (20±3.4) and the parental groups (21.5±3.1). The most lung metastases were observed in the mock-transfected (13.8±2.1) and the parental mice (14.5±2.6). Only a few lung metastases were observed in mice of si-OPN groups (2.5±1.6) (Fig. 4E). Moreover, their metastatic lesions were much smaller than those in the parental or mock control. The difference between the si-OPN groups with the parental or the mock transfected groups was significant (P<0.05).

To explore the possible mechanisms of OPN in progression of gastric cancer, the expression levels of NF-κB and activation of Akt and MEK/ERK1/2 were investigated in gastric cancer cells transfected with siRNA against OPN. Compared with untransfected or mock control, OPN siRNA was demonstrated to induce a significant decrease in the phosphorylation levels of Akt (Ser473), MEK and ERK1/2; however, no significant alteration in p65 NF-κB was observed in the cells transfected with OPN-specific siRNA (Fig. 5A).

To further evaluate the association of OPN with activation of MEK and ERK1/2, MKN28 cells were transfected with OPN siRNA, followed by incubation with OPN. Some of them were pretreated with specific MEK inhibitor, 4-diamino-2, 3-dicyano-1, 4-bis[2-aminophenylthio] butadiene (U0126) (20 μM), and those treated with OPN alone without pretreatment with U0126 were used as control. OPN was demonstrated to significantly increase the phosphorylation of MEK and ERK1/2, and U0126 could inhibit the activation of MEK and ERK1/2 induced by OPN (Fig. 5B).

In Fig. 5C, the si-OPN cells were treated with 10 μM exogenous recombinant OPN. Some of them were pretreated with specific Akt inhibitor 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002) (25 μM), and those treated with OPN alone without pretreatment with LY294002 were used as control. It was found that Akt could be activated in OPN treated cells, and be inhibited in the cells pretreated with LY294002.

To investigate the effect of OPN siRNA on p65 subunit of NF-κB, EMSA of NF-κB was performed. OPN siRNA was demonstrated to inhibit a translocation of p65 into nuclei in MKN28 cells. No significant alteration in NF-κB /DNA binding activity was found in the cells transfected with mock control (Fig. 6A).

To further determine how OPN affects this translocation, MKN28 cells were incubated with recombinant OPN for 90 min, some of the cells were preincubated with U0126 or LY294002 for 30 min. Cell lysates of nucleus were extracted and immunoblotted with antibodies specific to p65. Western blot analysis further confirmed the association of translocation of p65 with the OPN effects, and the effects of OPN on the p65 translocation in MKN28 cells were obviously inhibited by the MEK and Akt inhibitor, respectively (Fig. 6B). Taken together, we suggest these results show that MEK and Akt are partially mediated, OPN-induced the translocation of p65 from cytoplasm to the nuclei in MKN28 cells.

Western blotting was performed to investigate the expression levels of MMP-2, MMP-9 and uPA. The expression levels of MMP-2, MMP-9 and uPA were decreased in MKN28 cells treated with OPN siRNA (Fig. 7A). Then, gelatin zymography was used to investigate the activity of MMP-2 and MMP-9. OPN siRNA-transfected cells were treated with exogenous recombinant OPN (10 μM), cells without OPN treatment were used as control. The levels of both the pro-and active forms of MMP-2 and MMP-9 in OPN-treated cells were significantly enhanced compared with those control cells without OPN treatment (Fig. 7B). OPN siRNA was also demonstrated to induce a significant increase in the cleavage of caspase-3. However, it did not induce a specific alteration in the cleavage levels of caspase-3 in the cells transfected with mock control (Fig. 7C).

Pyrrolidine dithiocarbamate (PDTC), an NF-κB inhibitor, was used to explore the possible mechanisms of OPN in regulating MMP-2, MMP-9, uPA and caspase-3. MKN28 cells were preincubated with PDTC for 30 min followed by incubation with OPN for 90 min. A translocation of p65 into the nuclei and a significant increase in MMP-2, MMP-9 and uPA level were observed in the cells incubated with OPN, even though a significantly decreased expression of MMP-2, MMP-9 and uPA and an inhibited translocation of p65 were detected in cells treated with PDTC. Caspase-3 was found to be inhibited in OPN treated cells, and was activated in the cells pretreated with PDTC (Fig. 8). Together, these results suggest that OPN promotes MMP-2, MMP-9 and uPA expression and caspase-3 inhibition by activation of the NF-κB signaling.