Introduction

Peritoneal metastasis frequently occurs in patients with gastric cancer and is associated with a poor prognosis. Despite advances in chemotherapy, including neoadjuvant intraperitoneal systemic chemotherapy for peritoneal dissemination of gastrointestinal cancers, the results are unsatisfactory, and a standard and effective treatment has not been developed.

The establishment of peritoneal dissemination is a multistep process that includes the detachment of cancer cells from primary cancer, the adhesion and interaction of peritoneal free cancer cells (PFCCs) and mesothelial cells on the peritoneal surface, the penetration of free cancer cells into the submesothelial space, the attachment of cancer cells to the exposed submesothelial basement membrane, and the invasion of the peritoneal-blood barrier finally1. Many molecules are involved in this process. For example, E-cadherin is the key molecule for detachment2. CD44 is considered important for the adhesion and interaction between cancer cells and mesothelial cells3. Integrins, a family of adhesion receptors consisting of 18 α and 8 β subunits that form 24 distinct integrins, mediate cell-cell and cell-extracellular matrix (ECM) connections and are involved in the penetration of cancer cells into the submesothelial basement membrane4, 5.

Recently, c-Met, a receptor tyrosine kinase, and its ligand, hepatocyte growth factor (HGF), have become leading candidates for molecular targeted cancer therapies6 with the potential to block peritoneal metastasis. c-Met and HGF participate in most stages of malignant progression, including adhesion, degradation of ECM, promotion of cell mobility and the interaction between cancer cells and peritoneal mesothelial cells. Thus, integrins are involved with adhesion, invasion, dissemination and metastasis6. Many human cancers, including ovarian cancer7, gastric cancer8, esophageal adenocarcinoma9, and colorectal cancer10, have c-Met overexpression in the primary cancer tissues, which is also associated with poor outcome. Moreover, an effective therapy which targeted c-Met in vivo demonstrated the ability to inhibit peritoneal dissemination and invasion in ovarian cancer xenografts11. The measurement of c-Met expression level in tissues is crucial to predict the treatment response to inhibitors. PFCCs are considered to play a key role in the process of peritoneal metastasis in gastric cancer, and the molecules expressed on the surface of PFCCs usually provide important clues for targeted therapy12. However, only approximately 10% of the malignant cells in the peritoneal cavity or ascites are detected by conventional methods. According to previous reports, immunomagnetic separation can efficiently improve the detection of rare free malignant cells in fluid and blood specimens13, 14 and, thus, has the ability to study c-Met expression on PFCCs.

RNAi has been widely used as a powerful tool in gene function studies and as a potential treatment model for human cancers. miRNAs are members of a class of small regulatory RNAs and are targets of novel anticancer gene therapy by antisense molecules that can inhibit mRNA activity by mRNA cleavage or translational repression15. In this paper, we present a preliminary study on c-Met expression in PFCCs from gastric cancer patients and demonstrate a successful long-term efficient lentiviral miRNA (lenti-miRNA) system for silencing c-Met expression in the SGC7901 human gastric cancer cell line. We also evaluate c-Met as a therapeutic target in the treatment of gastric cancer peritoneal dissemination.

Materials and methods

Immunomagnetic isolation of PFCCs in ascites from gastric cancer patients

PFCCs from ascites of gastric cancer patients were isolated by the magnetic activated cell sorting (MACS) method. Briefly, ascites specimens were collected sterilely at the time of initial diagnosis, as confirmed by biopsy pathology, from 10 primary gastric cancer patients (Table 1). Samples and clinical data were collected after informed consent was obtained. The ascites samples were centrifuged into pellets and resuspended in 10 mL of phosphate-buffered saline (PBS) supplemented with 0.5% bovine serum albumin (BSA). The red blood cells (RBCs) were lysed with fresh lysing buffer with a volume ratio of 1:5. The manufacturer's instructions (Miltenyi Biotec, Germany) for MACS were followed for cancer cell enrichment. Mononuclear cells (MNCs) were recovered by centrifugation (200×g, 10 min) through a Ficoll density gradient according to the manufacture's protocol. Next, 20 μL FcR blocking reagent (Miltenyi Biotec) was added to block unspecific binding of MNCs to the microbeads. Anti-EpCAM conjugated with phycoerythrin (PE) monoclonal antibody (mAb) (Miltenyi Biotec) was incubated with the cells at a dilution of 1:10. Human peritoneal mesothelial cells (HPMCs) were labeled by anti-CD45 conjugated with fluorescein isothiocyanate (FITC) (Beijing CellChip Biotechnology Co, Ltd, China). The cells were incubated with 20 μL anti-PE microbeads for 15 min on ice and then resuspended in 80 μL MACS buffer. The magnetically labeled cells were washed with 1.5 mL MACS buffer and passed through a magnetic separation column (Miltenyi Biotec). The column was removed from the separator, and the positive fraction was eluted with the plunger supplied.

Table 1 PFCCs isolated by MACS and the c-Met/CK20 expression in the ascites of ten gastric cancer patients.
Full size table

Immunocytochemistry (ICC)

Each isolated cell sample underwent routine hematoxylin-eosin (HE) staining to observe cell morphology under a light microscope. Expressions of CK20 and c-Met were evaluated by immunocytochemical staining using streptavidin-biotin-peroxidase method in a cellular smear. First, cells were fixed in 4% paraformaldehyde. Next, primary anti-human CK20 (1:100) and c-Met (1:100) antibodies were incubated with the cells for 2 h at room temperature. The procedure was performed according to the manufacturer's instructions utilizing the horseradish peroxidase (HRP) immunohistochemistry (IHC) kit (Chemicon International Inc, Temecula, CA, USA). Cytoplasmic and/or cytomembrane immunostaining was considered positive. Negative control samples were processed similarly in the absence of primary antibody. The standard for assessment of malignancy was according to cytomorphology. The malignant cells were verified via light microscope criteria based on HE staining, including cellular pleomorphism, an altered nucleus and hyperchromasia.

Cell line and cell culture

The human gastric cancer cell line SGC7901 was obtained from Type Culture Collection of the Chinese Academy of Sciences (Institute of Biochemistry and Cell Biology, Shanghai, China). The tumor cells were cultured in RPMI-1640 media (PAA Laboratories GmbH, Austria) supplemented with 10% fetal calf serum (FCS) and penicillin (100 U/mL)/streptomycin (100 μg/mL) in an atmosphere of 37 °C in 5% CO2.

HPMCs were obtained following informed consent from the omental tissue of patients who underwent abdominal surgery for benign conditions. The HPMCs were isolated according to methods previously reported11. Briefly, the resected omental tissue was incubated in the presence of 0.25% trypsin/0.01% EDTA at 37 °C for 20 min. The detached HPMCs were then cultured in RPMI-1640 media (PAA Laboratories GmbH, Austria) supplemented with 20% FCS, penicillin (100 U/mL)/streptomycin (100 μg/mL), 0.4 mg/L hydrocortisone and 5 mg/L insulin in fibronectin-coated tissue culture flasks at 37 °C in a humid atmosphere of 5% CO2. Cells were used for experiments following 3 rounds of passage.

Recombinant plasmid construction, lentivirus production and transduction

All the procedures were performed according to the BLOCK- iT™ Pol II miR RNAi Expression Vector Kit and the Lentiviral Pol II miR RNAi Expression System (Invitrogen, Carlsbad, CA, USA). Four pairs of miR-155-based c-Met targeting sequences (64 bp) and a negative control sequence were designed using the Invitrogen RNAi Designer (www.invitrogen.com/rnaiexpress). The engineered pre-miRNA sequence structure was based on the murine miR-155 sequence, and synthesized double oligonucleotides were cloned into the pcDNA6.2-GW/EmGFP/miR plasmid (Invitrogen) to produce the recombinant plasmids containing either the c-Met miRNA insert (pCMV-c-Met miRNA-825, pCMV-c-Met miRNA-1872, pCMV-c-Met miRNA-3409, and pCMV-c-Met miRNA-4339) or the negative control (pCMV-c-Met miRNA-neg). Each purified expression plasmid was transfected into the SGC7901 cells. RT-PCR and Western blot analysis were performed to assess the efficiency of c-Met knockdown at 48 h post-transfection.

The lentiviral miRNA system based on the above vectors was generated. The third generation self-inactivating lentivirus vector, pLenti6/V5-DEST containing a CMV-driven EGFP reporter and a SV40 promoter upstream of the cloning sites for high-level expression of miRNA was used. The constructed functional pre-miRNA expression cassette targeting c-Met was transferred into the destination vector (pLenti6/V5-DEST) to generate the lentiviral expression vector (Lenti-c-MetmiRNA) or the negative control vector (Lenti-c-MetmiRNA-neg). The recombinant lentivirus and the control lentivirus were produced by co-transfecting 293FT cells with the transfer vector and the 3 packaging vectors. The virus-containing supernatant was harvested 72 h post-transfection. The prepared SGC7901 cells were transduced with the lentiviral expression vectors (Lenti-c-MetmiRNA or Lenti-c-MetmiRNA-neg). At 48 h after transduction, the knockdown of c-Met was evaluated by Western blot analysis. Stable cell lines were obtained after selection by long-term culture in medium containing 4 μg/mL blasticidin for 12 d.

Proliferation assay

Cell proliferation was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma, USA) assay. Cells were seeded at a density of 5×103 in 96-well plates in Dulbecco's modified Eagle's medium (DMEM) containing 10% FCS in a final volume of 100 μL. After incubation for 18 h, the cells grew to approximately 85% confluency, and then were transduced with the lentivirus (50 μL/well). At 24, 48, 60, and 96 h post-transduction, an MTT assay was performed. Cells were incubated in the presence of MTT (20 μL/well of 5 g/L solution in sterile PBS) for 4 h at 37 °C. The assay was read by determination of the reaction product at 490 nm by using a microplate reader (Multiskan MK3, Thermo Labsystems, Finland).

Invasive assay

The ability of cells to invade through ECM was assessed using the Chemicon QCM 96-well Invasion Assay Kit (ECM555, CHEMICON, International, CA, USA). Cells (3×106) were harvested with trypsin/EDTA, and suspended in PRMI-1640 media with 0.5% FBS. The invasive assay was performed according to the manufacturer's instructions. Briefly, the invasive gastric cancer cells, which migrated through the ECM layer and attached to the bottom of the polycarbonate membrane, were dissociated from the membrane after incubation with the Cell Detachment Solution for 30 min at 37 °C. Next, 50 μL of lysis buffer/CyQuant GR Dye Solution (1:75) was added to each well and incubated for 15 min at room temperature. Finally, 150 μL of the mixture was transferred to a new 96-well plate, and the fluorescence value was detected with a fluorescence plate reader using 480 nm/520 nm filter set. As a negative control, Cell Detachment Solution in the absence of SGC7901 cells was used.

Adhesion of gastric cancer cells to a HPMCs monolayer in vitro

Tumor cells were fluorescently labeled with 5 μmol/mL BCECF/AM (Biotium, Hayward, CA, USA). HPMCs (1.5×104 cells/well) were cultured in 96-well plates, which were pre-coated with 0.1% gelatin, until confluent. SGC7901 cells (3×104) were resuspended in 200 μL medium and added to each well. The plates were incubated for 60 min at 37 °C. The number of adhesive cells was quantified by measuring the fluorescence intensity on a Perkin Elmer plate reader (CA, USA) with 488 nm excitation and 535 nm emission filters.

Adhesion of gastric cancer cells to mice peritoneum in vivo

Six-week-old BALB/c athymic female nude mice (w) were used in this study. The mice were divided into three groups (10 mice per group): SGC7901, Lenti-miRNAc-Met and Lenti-miRNAc-Met-neg. Cultured tumor cells, fluorescently labeled as above, were injected at a concentration of 2×106/mL into the mice peritoneal cavity in a final volume of 1 mL. After 4 h, mice were sacrificed, and the full peritoneum was cut into 1 cm2 pieces and placed in a 24-well culture plate in DMEM containing 10% FCS. Adherent cells were lysed with 1% NP-40, and the fluorescence intensity was measured. All animal experiments were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals of Tongji University.

Reverse transcription-PCR (RT-PCR)

Total RNA was extracted with TRIzol Reagent (Invitrogen). PCR was performed using the RT-PCR Kit (Toyobo, Japan) with 5 μg total RNA to synthesize cDNA. The primers of c-Met were 5′-CCTGCGAAGTGAAGGGTCTCC-3′ (forward primer) and 5′-CTGGCAGCTTTGCACCTGTTT-3′ (reverse primer). GAPDH was used as an internal control. The primers of human GAPDH were 5′-ACCACAGTCCATGCCATCAC-3′ (forward primer) and 5′-TCCACCACCCTGTTGCTGTA-3′ (reverse primer). PCR was performed in 30 μL of reaction mixture. Thermal cycle conditions were conducted with preamplification denaturation at 95 °C for 3 min, followed by 30 cycles of denaturation at 95 °C for 20 s, annealing at 62 °C for 20 s, and extension at 72 °C for 30 s, with a final extension at 72 °C for 3 min. For the semi-quantification, an image of the gel was captured, and the intensity of the bands was quantified by gel analysis system (Bio-Rad, USA).

Western blot analysis

SGC7901 cells (5×106) were collected and lysed in cold PBS with 0.5% NP-40 and a protease inhibitor cocktail (Novogen, Australia). Whole-cell proteins were extracted using the Whole Cell Extraction Kit (Chemicon). Protein concentrations were determined according to the manufacturer's instructions of the BCA (bicinchoninic acid) Protein Assay Kit (Pierce, Rockford, USA). Samples were equalized, and 40 mg of total protein for each sample was loaded into separate lanes on a 12% SDS-PAGE and transferred onto a nitrocellulose membrane (Invitrogen). Western blot analysis was performed using primary antibody for c-Met (1:100 dilution, Santa Cruz Biotechnology, CA, USA). Protein expression was quantified by densitometry and normalized to β-actin (1:100, Santa Cruz) expression using enhanced chemiluminescence (Pierce, Rockford, IL, USA).

Adhesion molecule expression as detected by flow cytometry (FCM)

Tumor cells (3×106) were incubated with primary antibodies (Integrin β1 1.5 μg, E-cadherin 1.5 μg, CD44 0.75 μg, Biolegend, San Diego, CA, USA) for 30 min at 4 °C. PE-labeled secondary antibody was added and incubated for 30 min. Additionally, the primary antibody PE-α3-integrin (BioLegend, San Diego, CA, USA) was incubated with tumor cells for 30 min. Next, cells were resuspended in 500 μL of 1% BSA/PBS. Fluorescence expressions on cells surface were quantitatively analyzed with a FACSCalibur (Becton-Dickinson, San Jose, USA). Cells incubated in the absence of primary antibody were set as a blank, and those incubated in the isotype-specific hamster IgG1-PE were set as a negative control to eliminate non-specific staining.

In vivo studies of peritoneal dissemination model in nude mice

BALB/c female nude mice (5–6 weeks of age) were purchased from the Shanghai Laboratory Animal Center (Shanghai, China). Thirty mice were allocated to three groups (10 mice per group): SGC7901 (group A), Lenti-miRNAc-Met-neg (group B) and Lenti-miRNAc-Met (group C). Suspensions of tumor cells (1×107 cells) in 1 mL RPMI-1640 media were injected into the peritoneal cavity of the mice for incubation. On d 30, all thirty mice were sacrificed.

To evaluate further the treatment effect of Lenti-miRNAc-Met, another 45 nude mice were allocated to three groups as above (15 mice per group). In the PBS group, 4 mL PBS was injected into the peritoneal cavity of each mouse. In the Lenti-miRNAc-Met-neg group, 5×107 copies/4 mL of Lenti-miRNAc-Met-neg was injected in each mouse intraperitoneally (ip). In the Lenti-miRNAc-Met group, 5×107copies/4 mL of Lenti-miRNAc-Met was injected in each mouse ip. Three days postinoculation, 1×107 SGC7901 cells in 1 mL PBS were injected into each mouse ip. On d 30, 10 mice from each group were sacrificed. The remaining five mice in each group were used to evaluate the survival up to d 120. The macroscopic nodules on peritoneal surface were counted, and tumor size of large nodules that exceeded 1.0 cm in diameter was calculated by assuming a spherical shape. The fused nodules were counted as a single nodule. All animal experiments were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals of Tongji University.

Statistical analysis

Statistical analysis was performed using the Statistics Package for the Social Science software (version 11.5; SPSS Inc, Chicago, IL, USA). The comparison between different groups was analyzed by the Independent Samples t Test or the Mann-Whitney Test. Survival curves were obtained using the Kaplan-Meier method and compared by the log-rank test. All the statistical analysis was two sides with significance defined as P<0.05.

Results

Expression of c-Met in the PFCCs

PFCCs were found utilizing MACS in the ascites of 6 out of 10 gastric cancer patients. PFCCs had an increased nucleus: cytoplasm ratio with nuclear hyperchromasia observed by light microscope (Figure 1Aa). CK20-positive expression verified their epithelial origin, notably from gastrointestinal tract (Figure 1Ab). c-Met expression was detected in the PFCCs of 5 patients (Figure 1Ac).

Figure 1
figure 1

Expression of c-Met in the PFCCs and SGC7901 cells. (A) Free malignant cells detected by MACS from the ascites of gastric cancer patients. (a) HE staining (arrows, magnification 200×). (b) CK20 expression in the free gastric cancer cells (immunocytochemical staining, magnification 200×). (c) c-Met expression in the free gastric cancer cells (immunocytochemical staining, magnification 200×). (B) Expression of c-Met mRNA in SGC7901 cells transduced with pCMV-c-MetmiRNA expression plasmids at 48 h. (a) RT-PCR assay. (b) c-Met mRNA in pCMV-c-MetmiRNA-825 transduced cells was significantly reduced by 87.6% when compared with parental SGC7901 cells (bP<0.05, cP<0.01). (C) Expression of c-Met protein in SGC7901 cells transduced with different pCMV-c-MetmiRNA expression plasmids at 48 h. (a) Western blot assay. (b) c-Met protein in pCMV-c-MetmiRNA-825 transduced cells was significantly reduced by 90% when compared with the parental SGC7901 cells (bP<0.05, cP<0.01).

Full size image

Lentivirus-mediated RNAi significantly inhibited c-Met expression in SGC7901 cells

Four recombinant expression plasmids targeting c-Met were constructed. The mRNA and protein expression levels of c-Met in SGC7901 cells were examined by RT-PCR and Western blot analysis at 48 h post-transfection. As shown in Figure 1B and 1C, when compared with the parental SGC7901 cells, the negative control did not affect the expression levels of c-Met, whereas pCMV-c-MetmiRNA-825 resulted in the decrease of both c-Met mRNA and protein levels. Specifically, c-Met mRNA expression was reduced by 80%–83%, while c-Met protein expression was reduced by 88%–90% in pCMV- c-MetmiRNA-825 transfected cells.

The recombinant expression plasmid pCMV-c-MetmiRNA-825 was inserted into pLenti6/V5-DEST destination vector to develop a Lenti-miRNA expression vector. We infected SGC7901 cells with the lentivirus carrying pCMV-c-MetmiRNA-825 or pCMV-c-MetmiRNA-neg. c-Met protein expression decreased by 91.5% in Lenti-miRNAc-Met transfected cells as compared with the parental SGC7901 cells (P<0.05, Figure 2A, 2B) 48 h post-transfection. A stable SGC7901 cell line transduced with lentivirus was successfully generated by blasticidin (4 μg/mL) selection. After continuous culture for 30 d, the lentivirus-transduced cells had a stable inhibition of c-Met expression (EGFP positive expression in cells was 78.6%, Figure 2C).

Figure 2
figure 2

Lentivirus-mediated RNAi suppressed c-Met protein expression in SGC7901 cells. (A) Western blot assay. (B) c-Met protein in Lenti-miRNAc-Met transduced cells was significantly reduced by 91.5% when compared with the parental SGC7901 cells (cP<0.01). (C) The transduction efficiency after continuous culture in the lentivirus-miRNAc-Met stably transduced cell line was 78.6% as measured by FCM.

Full size image

Knockdown of c-Met suppressed cell proliferation in vitro

The effect of knockdown of c-Met expression via Lenti-miRNAc-Met on gastric cancer cell proliferation was determined by MTT assay. As shown in Figure 3, cell proliferation was inhibited in a time-dependent manner in SGC7901 cells transfected with Lenti-miRNAc-Met for 24–96 h post-transduction when compared with the parental cells and negative-control groups.

Figure 3
figure 3

Cell growth was significantly inhibited in SGC7901 cells after transduction with Lenti-miRNAc-Met at various times as measured by MTT assay (cP<0.01).

Full size image

Knockdown of c-Met suppressed cell invasion in vitro

To verify the effect of knockdown of c-Met expression via Lenti-miRNAc-Met on migration ability of SGC7901 cells, an invasion assay was performed. Migration was determined in accordance with the fluorescence value, which represented the invading SGC7901 cells that migrated through the ECM layer. Figure 4 demonstrated that the fluorescence value of the cells transduced with Lenti-miRNAc-Met (23191.21±5183.91) was much lower than the parental cells (45496.13±9408.99) or the negative control cells (42821.36±9513.45), indicating the invasive capacity of the Lenti-miRNAc-Met transduced cells markedly decreased. This result suggested that Lenti-miRNAc-Met attenuated the metastatic potential of gastric cancer cells in vitro.

Figure 4
figure 4

The ability of SGC7901 cells transduced with Lenti-miRNAc-Met to invade through ECM as detected by an invasion assay. (A) SGC7901 cells. (B) Lenti-miRNAc-Met-neg transduced cells. (C) Lenti-miRNAc-Met transduced cells. (D) The invasion to ECM was significantly inhibited in the cells transduced with Lenti-miRNAc-Met when compared with the parental SGC7901 cells according to the fluorescence assay (cP<0.01).

Full size image

Knockdown of c-Met suppressed cell adhesion to HPMCs in vitro and mouse peritoneum in vivo by downregulation of integrin α3β1 and E-cadherin

Cancer cell adhesion to peritoneum is a crucial process and the initial step during peritoneal metastasis. To determine the effect of knockdown of c-Met expression via Lenti-miRNAc-Met on adhesion, fluorescently labeled SGC7901 cells were tested via an adhesion assay. In vitro, the adhesion of the Lenti-miRNAc-Met transduced SGC7901 cells to the monolayer of HPMCs was decreased by 64% (Figure 5A1, P<0.01), and in vivo, adhesion to the full-thickness mice peritoneum decreased by 69% (Figure 5A2, P<0.01), both compared with the parental cells. No statistical significance was observed between the SCG7901 cells and the negative control (P>0.05).

Figure 5
figure 5

Adhesion of SGC7901 cells transduced with Lenti-miRNAc-Met detected by the fluorescence value assay (A) and their expressions of adhesion molecules detected by FCM (B) (bP<0.05, cP<0.01). (A1) Adhesion of SGC7901 cells transduced with Lenti-miRNAc-Met to HPMCs monolayer in vitro was significantly inhibited when compared with parental SGC7901 cells. (A2) Adhesion of SGC7901 cells transduced with Lenti-miRNAc-Met to mice peritoneum in vivo was significantly inhibited when compared with parental SGC7901 cells. (B) The adhesion molecules expressions in SGC7901 cells. Groups: (a) SGC7901 cells; (b) Lenti-miRNAc-Met-neg transduced cells; (c) Lenti-miRNAc-Met transduced cells.

Full size image

To investigate the mechanism of adhesion to the peritoneum, adhesion molecule expression was examined by FCM. As shown in Figure 5B, the expressions of E-cadherin, integrin β1 and integrin α3 were significantly reduced in Lenti-miRNAc-Met transduced SGC7901 cells after 48 h, compared with the SGC7901 cells (Figure 5B, P<0.01). However, CD44 expression was not significantly inhibited (Figure 5B, P>0.05).

Knockdown of c-Met inhibited tumorigenicity of SGC7901 cells in vivo

To observe the inhibition of peritoneal dissemination of SGC7901 cells in vivo after knockdown of c-Met, we inoculated nude mice ip with 1×107 parental, Lenti-miRNAc-Met-neg transduced or Lenti-miRNAc-Met transduced SGC7901 cells. At 30 d postinoculation, numerous tumor nodules of peritoneal dissemination were observed on the peritoneal surface around the radix of mesentery and blood vessels in SGC7901 group and the negative control group, whereas tumor nodules were significantly suppressed in mice in the Lenti-miRNAc-Met inoculated group (P<0.05, Figure 6). The size of large nodules that exceeded 1.0 cm in diameter was also significantly reduced compared with SGC7901 and the negative control inoculated animals (P<0.05, Figure 6A3).

Figure 6
figure 6

Lentivirus-mediated RNAi silencing of c-Met effectively suppresses peritoneal dissemination of gastric cancer in vitro and in vivo. Groups: (a) SGC7901; (b) Lenti-miRNAc-Met-neg; (c) Lenti-miRNAc-Met. (A) Tumor growth in the abdominal cavities of mice was inhibited by suppression of c-Met using Lenti-miRNAc-Met. (A1) Tumor nodules on peritoneal surface and around the radix of mesentery. (A2) The total number of tumor nodules was significantly reduced in Lenti-miRNAc-Met transduced cells when compared with the parental SGC7901 cells (cP<0.01). (A3) The tumor size was significantly reduced in Lenti-miRNAc-Met transduced cells when compared with the parental SGC7901 cells (cP<0.01). (B) Intraperitoneal injection of Lenti-miRNAc-Met significantly suppressed tumor growth in a nude mouse peritoneal dissemination model of gastric cancer. (B1) Tumor nodules on peritoneal surface and around the radix of mesentery were significantly reduced in the Lenti-miRNAc-Met group (c), when compared with the PBS group (a) and the Lenti-miRNAc-Met-neg group (b). (B2) The survival rate of the mice treated with Lenti-miRNAc-Met was higher than those of the mice treated with Lenti-miRNAc-Met-neg or PBS. (B3) The total number of tumor metastatic nodules was significantly reduced in mice treated with Lenti-miRNAc-Met when compared with Lenti-miRNAc-Met-neg or PBS (cP<0.01). (B4) The tumor size was significantly reduced in mice treated with Lenti-miRNAc-Met when compared with Lenti-miRNAc-Met-neg or PBS (cP<0.01).

Full size image

Treatment of peritoneal dissemination by Lenti-miRNAc-Met in nude mice

To further study the efficacy of Lenti-miRNAc-Met as a therapeutic target for peritoneal dissemination in gastric cancer, three days after we injected Lenti-miRNAc-Met and the control lentivirus, we established a peritoneal dissemination model by intraperitoneal incubation with 1×107 SGC7901 cells in nude mice. Peritoneal dissemination was significantly inhibited in the group treated with Lenti-miRNAc-Met as compared with the control group treated with Lenti-miRNAc-Met-neg or PBS (P<0.05, Figure 6B). The mice treated with Lenti-miRNAc-Met-neg or PBS died within 75 d. However, the mice treated with Lenti-miRNAc-Met had prolonged lives up to 116 d (P<0.05, Figure 6B2).

Discussion

The biological function of PFCCs can be considered a decisive factor in tumor peritoneal dissemination and the basis for therapeutic strategy. Isolation of PFCCs is helpful to understand their biological characteristics. However, early detection of rare free cancer cells in the abdominal cavity is difficult with conventional methods. In this study, we successfully isolated PFCCs using the MACS method in the ascites from 6 out of 10 gastric cancer patients, and moreover, c-Met expression was detected in the PFCCs of 5 patients.

The general indication for treatment with kinase and signaling pathway inhibitors is based on the level of expression and mutation status in cancer cells. For example, the effect of targeting human epidermal growth factor receptor-2 (HER-2) therapy in breast cancer depends on the level of HER-2 expression in breast cancer cells. Gastric cancer as a common cancer leading to peritoneal dissemination, the peritoneum is a common site for distant metastasis. Though the previous study has shown the effect of targeting c-Met therapy in ovarian cancer in vivo11, the potential therapeutic efficacy of c-Met for gastric cancer has not been studied. Therefore, the expression of c-Met in PFCCs from gastric cancer ascites suggests that it may have the possibility to develop a treatment targeting c-Met for peritoneal dissemination in gastric cancer patients.

Due to that peritoneal dissemination is characterized by the adhesion and invasion of tumor cells into the peritoneum11, it is an important process that cancer cells detaching from the primary cancer and successfully adhering to mesothelial cells in the dissemination and formation of metastases in gastric cancer. In this study, using lentivirus mediated RNAi technique, the expression of c-Met in both mRNA (80%–83%) and protein levels (88%–90%) were effectively knocked down by artificial engineered pre-miRNA sequences specifically targeting c-Met in human SGC7901 gastric cancer cell line. Moreover, we established the stably transduced Lenti-miRNAc-Met SGC7901 cell line and the control stably transduced cell line. After knockdown of c-Met expression, the adhesion of SGC7901 gastric cancer cells to monolayer HPMC in vitro and to mice peritoneum in vivo were significantly suppressed, compared with the control virus.

Many adhesion molecules are involved in this regulation as the previous reported. We further investigated the mechanism of adhesion. The findings showed that E-cadherin and α3β1 integrin expression were reduced in the Lenti-miRNAc-Met transduced SGC7901 cells, as compared with the negative control and parental cells. However, no close relationship between c-Met and CD44 in our study was found although some CD44 isoforms can promote or increase c-Met activation16. E-cadherin is also involved in the multistep process of peritoneal dissemination in epithelial ovarian carcinoma2. c-Met can remove E-cadherin from cell-cell adhesion sites and regulate the activity of integrins17. In gastric cancer, modulation of E-cadherin by hepatocyte growth factor/c-Met induces aggressiveness of gastric carcinoma18. CD44 can cause very strong cell adhesion to peritoneal mesothelium and an unfavourable prognosis in ovarian cancer3, but we did not find the similar result in gastric cancer in this study. α3β1 integrin plays an essential role in mediating the initial attachment of cancer cells to the peritoneum in the peritoneal implantation of NUCC-4 human gastric cancer cells in athymic mice8. Therefore, c-Met can regulate these molecules in the peritoneal dissemination of gastric cancer. However, in ovarian cancer, targeting c-Met in vivo inhibited peritoneal dissemination and invasion through an α5β1 integrin-dependent mechanism11. So, in human cancers, various integrin combinations of α and β subunits produce polymorphisms with different ligands specificity and function. The molecules involved in peritoneal dissemination may also be different.

After adhesion to peritoneum, cancer cells invade the ECM of the peritoneum and form the metastasis site. c-Met plays an important role in the process of epithelial-mesenchymal interaction, regulation of cell migration, invasion, cell proliferation and survival. Our results indicated the Lenti-miRNAc-Met transduced SGC7901 cells survival and their ability to invade the ECM were significantly inhibited in vitro.

The biological function of HGF-MET axis affects tumor growth and development of metastasis19. The advantage of stable and long-term gene silencing offered by the lentivirus vector system makes it a powerful tool for gene therapy. Based on the peritoneal metastasis model established by SGC7901 gastric cancer cells20, we found that tumor growth (including the number of macroscopic nodules on peritoneal surface and tumor size of large nodules) in mice which received intraperitoneal injections of SGC7901 cells transduced with Lenti-miRNAc-Met was strongly suppressed compared with the Lenti-miRNAc-Met-neg control and parental SGC7901 cell groups. The intraperitoneal administration of Lenti-miRNAc-Met dramatically suppressed the growth of gastric cancer cells on the surface of peritoneum and improved the survival of the mice. Combining these findings in vitro and in vivo, we confirmed the role of c-Met in the growth, proliferation, invasion to ECM, adhesion to peritoneum and peritoneal dissemination in gastric cancer.

In summary, our findings indicate that c-Met is expressed in PFCCs, and it plays an important role in the process of gastric cancer peritoneal dissemination, including tumor growth, adhesion to peritoneum and invasion to ECM. E-cadherin and α3β1 integrin are involved in the regulation of adhesion of SGC7901 cells to mesothelial cells. Therefore, Lentivirus-mediated RNAi targeting c-Met may be a useful tool for gene therapy targeting gastric cancer peritoneal dissemination.

Author contribution

Xiao-lei WANG was the guarantor of integrity of the entire study, responsible for study design, data analysis and manuscript preparation. Xi-mei CHEN provided final approval of the version to be submitted. Jian-ping FANG participated in the immunocytochemical staining experiments. Chang-qin YANG revised the article for important intellectual content.