Abstract
The interaction between gastric carcinoma cells and the peritoneal lining is a key step in peritoneal dissemination. In this study, we examined the roles of the β1 family of integrin receptors in the adhesion of such cells to the peritoneum. The adhesion of several gastric carcinoma cell lines to peritonea excised from mice was inhibited most by an anti-α3 integrin antibody and to a lesser extent by an anti-α2 integrin antibody. In the peritoneal implantation of NUGC-4 human gastric carcinoma cells in athymic mice, treatment of the cells with anti-α2 or anti-α3 integrin antibody reduced the number of disseminated nodules; suppression by the anti-α3 integrin antibody was stronger than that by the anti-α2 integrin antibody. The cDNAs to human α2 and α3 integrins were introduced into K562 leukemic cells, which were positive for the integrin β1 subunit but negative for the α2 or α3 subunit. The α3 integrin-transfected cells adhered to excised peritoneum and to a monolayer of peritoneal mesothelial cells more firmly than did the α2 integrin-transfected cells or the mock transfectant. Reverse transcription-PCR was used to analyze the expression of laminin-5 and laminin-10/11, which have been reported to serve as high-affinity ligands for α3β1 integrin. mRNA for these laminin isoforms was found in mesothelial cells from the diaphragm and parietal peritoneum. These results strongly suggest that α3β1 integrin plays an essential role in mediating the initial attachment of cancer cells to the peritoneum, leading to the formation of peritoneal metastasis.
INTRODUCTION
Peritoneal dissemination frequently occurs in patients with gastric cancer, and it hurts their prognoses. The metastatic dissemination is mediated via exfoliation of carcinoma cells into the peritoneal cavity, followed by adhesion to and invasion through the mesothelium that overlies the contents of the peritoneal cavity. The adhesion of carcinoma cells to the peritoneum is a key step in the initial process of peritoneal dissemination. It has been shown that various cell adhesion molecules expressed on carcinoma cells play crucial roles in their adhesion to the peritoneum. The integrin family of cell adhesion molecules serve as adhesion receptors for extracellular matrix proteins and cellular counterligands. These adhesion receptors are heterodimers of transmembrane glycoproteins (α and β subunits); various combinations of the α and β subunits produce polymorphisms of ligand specificity. A number of reports have demonstrated that alterations in integrin expression profiles in cancer cells are frequently associated with their malignant phenotypes, including invasive and metastatic potentials (reviewed in Refs. 1, 2, 3, 4, 5 ). Integrin receptors expressed on cancer cells contribute directly to the adhesive interactions between metastatic cells and compartments of the host environments. They also contribute to the stimulation of cell growth and motility through intracellular signaling. In the interaction between gastric carcinoma cells and the peritoneum, the β1 integrin family of cell adhesion molecules and CD44 have been suggested to play a crucial role (6 , 7) . The immunohistochemical analysis of specimens of gastric cancer resected from >100 patients demonstrated that the expression of α3β1 integrin was positively correlated with the occurrence of peritoneal and liver metastases and with increased invasiveness of the tumors (8) . Another report showed that α2β1 integrin was significantly expressed on a larger proportion of tumor cells in peritoneal metastases than in primary tumors (9) . These observations suggest a causal relationship between the expression of these integrin receptors on carcinoma cells and their peritoneal metastatic potentials. Thus, we attempted to evaluate the possible role of the β1 integrin family of adhesion molecules, especially the α2β1 and α3β1 integrins in the adhesion of gastric carcinoma cells to the peritoneum. This study presents evidence showing the importance that these integrin receptors expressed on the carcinoma cell surface have in the cells’ interaction with the peritoneum.
MATERIALS AND METHODS
Reagents.
A fluorescent dye, 2′,7′-bis(carboxyethyl)carboxyfluorescein tetraacetoxymethyl ester (BCECF-AM), was purchased from Dojindo Laboratories (Kumamoto, Japan). Thiol-reactive fluorescent probes, 5-chloromethylfluorescein diacetate (CMFDA; Cell Tracker Green) and 5-[((chloromethyl)benzoyl)amino)tetramethylrhodamine (CMTMR; Cell Tracker Orange] were products of Molecular Probes (Eugene, OR). Matrigel and BSA were purchased from BD Biosciences (San Diego, CA) and Sigma (St. Louis, MO), respectively. Restriction endonucleases and modifying enzymes were purchased from TaKaRa (Osaka, Japan), Toyobo (Osaka, Japan), and Life Technologies, Inc. (Rockville, MD). The monoclonal antibodies used in this study were as follows. Anti-α2 integrin (Gi9) and anti-α6 integrin (GoH3) antibodies were purchased from Beckman Coulter, Inc. (Fullerton, CA). Anti-α3 integrin (SM-T1) antibody was prepared in our laboratory (10) . Anti-α4 integrin (SG/73), anti-α5 integrin (KH/33), and anti-β1 integrin (SG/19) antibodies were purchased from Seikagaku Corp. (Tokyo, Japan). Anti-CD44 antibody (G44-26) was purchased from BD Biosciences. A polyclonal antibody against human laminin-5 was prepared in a rabbit in our laboratory (11) . This antibody cross-reacts with murine laminin-5. Rabbit antimouse laminin-1 was purchased from Progen Biotechnik GmbH (Heidelberg, Germany). These polyclonal antibodies were purified on an immobilized protein A column (ImmunoPure; Pierce, Rockford, IL) according to the manufacturer’s instructions. FITC-labeled antimouse IgG and antirat IgG antibodies were purchased from ICN Pharmaceuticals, Inc. (Costa Mesa, CA).
Cell Lines.
Human gastric carcinoma cell lines MKN1, MKN28, and GT3TKB were supplied by the RIKEN Cell Bank (Tsukuba, Japan), and NUGC-4 was supplied by the Health Science Research Resources Bank (Osaka, Japan). Stable transfectant cells expressing α2 integrin (K2F10) and α3 integrin (K3-9) were established by transfection of K562 human erythroblastoid leukemia cells with human α2 and α3 integrin cDNA, respectively, that had been cloned in an Rc/CMV mammalian expression vector (Invitrogen, Carlsbad, CA; Refs. 10 , 12 ). Mock transfectant cells (KR) were also prepared by transfection with an Rc/CMV vector alone. These cells were cultured in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (HyClone, Logan, UT) at 37°C under 5% CO2.
Flow Cytometry.
The expression of integrins was measured by a flow cytometer (FACSCalibur; BD Biosciences) as previously described (10) , using monoclonal antibodies against human integrin α2, α3, α4, α5, α6, and β1 subunits and FITC-labeled secondary antibodies.
Isolation of Peritoneal Mesothelial Cells.
The parietal peritoneum and diaphragm were excised from ICR mice under sterile conditions. Mesothelial cells were isolated essentially as described by Nakashio et al. (7) . Briefly, the resected parietal peritoneum and diaphragm were washed with serum-free RPMI 1640 and incubated with 0.25% trypsin (Life Technologies, Inc.) in 10 mm PBS (pH 7.3) at 37°C for 20 min. After an equal volume of RPMI 1640/10% FBS was added to the cell suspension, the mixture was passed through nylon mesh and centrifuged at 700 rpm for 5 min. The pelleted cells were suspended in RPMI 1640/10% FBS (1 × 106 cells/ml), and an aliquot (0.1 ml) was placed in a 96-well culture plate, which had been coated with 10% Matrigel solution at 4°C for 16 h. These cells were cultured at 37°C for 2–3 days until growth of a monolayer of polygonal mesothelial cells.
Cell Adhesion to Peritoneum.
The excised parietal peritoneum (∼1.6 cm2) was placed in a 24-well culture plate, which had been filled with 1.0 ml of 1% BSA/RPMI 1640. Cells were fluorescently labeled with BCECF-AM (3 μm) at 37°C for 30 min as previously described (13) and were washed twice with 1% BSA/RPMI 1640. A cell suspension (5 × 105 cells/ml in 1% BSA/RPMI 1640; 0.5 ml) was overlaid on the peritoneum in a 24-well plate, and the plate was incubated at 37°C for 40 min. After gentle washing with PBS, the cells adherent to the peritoneum were lysed with 1.0 ml of 1% NP40, and the fluorescence intensity was measured with a fluorescence spectrophotometer (Ex = 490 nm, Em = 520 nm). For the inhibition experiment, the cells were treated with an antibody at 0°C for 30 min before being incubated with the peritoneum.
Cell Adhesion to Mesothelial Cell Monolayer.
A monolayer culture of mesothelial cells in a 96-well culture plate was washed twice with 1% BSA/RPMI 1640. To the monolayer, a suspension of BCECF-labeled cells (1 × 106 cells/ml in 1% BSA/RPMI 1640; 0.1 ml) was added, and the plate was incubated at 37°C for 40 min. After the nonadherent cells were removed by gentle washing three times with warm 1% BSA/RPMI 1640, the adherent cells were lysed with 0.2 ml of 1% NP40, and the fluorescence intensity was measured as described above.
Fluorescence Microscopic Observation.
To examine the adhesion capacity of α3 integrin-positive and -negative cells to the peritoneum specimen, α3 integrin-transfected cells (K3-9) and mock-transfected cells (KR) were separately labeled, each with a distinct fluorescent probe. K3-9 cells and KR cells were suspended in 1% BSA/RPMI 1640 at a density of 1 × 106 cell/ml and incubated with CMTMR (red fluorescence; 5 μm) for K3-9 or with CMFDA (green fluorescence; 5 μm) for KR cells at 37°C for 20 min. The mixture of CMTMR-labeled K3-9 cells and CMFDA-labeled KR cells (5 × 105 cells/ml for each cell type; 0.5 ml) was overlaid on the excised parietal peritoneum in a 24-well plate, incubated at 37°C for 40 min, and washed with PBS. The cells adherent to the peritoneum specimen were observed under a fluorescence microscope (Olympus IX70; Olympus, Tokyo, Japan) equipped with an NIBA filter unit (Ex = 470–490 nm, Em = 515–550 nm) for CMFDA and a WIG filter unit (Ex = 520–550 nm, Em >580 nm) for CMTMR.
Peritoneal Implantation of Gastric Carcinoma Cells.
BALB/c nu/nu mice (5 weeks old, female) were purchased from Sankyo Lab Service (Tokyo, Japan). NUGC-4 cells (5 × 106 cells; 0.5 ml) suspended in RPMI 1640 were incubated with the antibodies against α2 or α3 integrin subunits at 0°C for 20 min and inoculated into the abdominal cavity of BALB/c nu/nu mice. The mice were killed on the 28th day of the study, and disseminated nodules on the mesentery and diaphragm were evaluated. This study was conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals of Hoshi University.
Detection of mRNAs for Laminin Isoforms.
mRNAs for laminin isoforms (laminin-5 and laminin-10/11) were detected by the reverse transcription-PCR method. The total RNA was isolated from peritoneal mesothelial cells by the acid guanidinium-phenol-chloroform method (14) using Trizol (Life Technologies, Inc.). Poly A+ RNA was purified with Oligotex dT30 (TaKaRa) according to the manufacturer’s instructions. cDNA was synthesized by using Moloney murine leukemia virus reverse transcriptase and antisense primers as follows. PCR was performed with a set of primers: 5′-CAGGGGCTGAGTGTGAGCAC-3′ (sense primer for laminin-5 α3 subunit) and 5′-GGAGGATGGAACGTAGGTCC-3′ (antisense primer for laminin-5 α3 subunit; Ref. 15 ); 5′-TGTATCTGTCCACCACGCACTG-3′ (sense primer for laminin-10/11 α5 subunit) and 5′-ACATCTTGAGCCCTGCACGTTC-3′ (antisense primer for laminin-10/11 α5 subunit; Ref. 16 ). The PCR conditions were as follows: 94°C, 1 min; 61°C, 1 min; 72°C, 1 min for 30 cycles. The products were separated by 3% agarose gel electrophoresis in 40 mm Tris-acetate buffer (pH 8.0) containing 1 mm EDTA (1 × TAE), followed by ethidium bromide staining.
RESULTS
Adhesion of Gastric Carcinoma Cells to Peritonea.
We first examined how antibodies against the β1 family of integrin receptors affected the adhesion of several gastric carcinoma cells to the excised parietal peritonea. The adhesion of MKN1 cells was inhibited most by the antibodies against the integrin β1 and α3 subunits and, to a lesser extent, by the antibody against the integrin α2 subunit (Fig. 1A) ⇓ . In contrast, the antibodies against integrin α4, α5, or α6 did not inhibit adhesion. Although a slight inhibition by an anti-CD44 antibody was observed, it was not significant. None of the antibodies completely inhibited the adhesion of these cells to the peritonea. However, the adhesion of MKN1 cells was decreased to 34% in the presence of both anti-β1 integrin subunit and anti-CD44 antibodies. These results suggested that α2β1 and α3β1 integrins play a role in the adhesion of MKN1 cells to the peritoneum, although the adhesion is mediated by multiple interactions with various adhesion molecules. Similar inhibitory effects by the antibodies against the α2 and α3 integrin subunits were observed when three additional gastric carcinoma cell lines were subjected to the adhesion assay (Fig. 1B) ⇓ . The anti-integrin α3 subunit antibody inhibited the adhesion of NUGC-4, GT3TKB, and MKN28 cells more effectively than did the anti-integrin α2 subunit antibody.
Adhesion of gastric carcinoma cells to parietal peritoneum. Fluorescently labeled gastric carcinoma cells were overlaid on excised peritoneum and incubated at 37°C for 40 min. After gentle washing to remove nonadherent cells, the adherent cells were lysed with 1% NP40, and the fluorescence intensity was measured with a fluorescence spectrophotometer (Ex = 490 nm, Em = 520 nm). For the inhibition experiment, the cells were treated with an antibody (100 μg/ml) at 0°C for 30 min before the incubation with peritoneum. The assays were carried out in triplicate, and the error bars indicate the SDs. A, MKN1; B, NUGC-4 (left); GT3TKB (middle); MKN28 (right).
We then measured the expression profiles of the integrin α2 and α3 subunits on carcinoma cells that had been used for the adhesion assay (Fig. 2) ⇓ . In the flow cytometric analysis, all four cell lines examined were found to express both α2 and α3 subunits, although the expression levels of the individual subunits were somewhat different among these cells.
Flow cytometric analysis of the expression of integrin α2 and α3 subunits in gastric carcinoma and transfectant cells. NUGC-4 (A and E), MKN1 (B and F), GT3TKB (C and G), and MKN28 (D and H) cells were treated with monoclonal antibodies against α2 (A–D) or α3 (E–H) integrin subunits and subsequently with FITC-labeled secondary antibodies. The thin lines represent the profiles of control experiments without anti-integrin antibody. K2F10 (α2 integrin transfectant), K3-9 (α3 integrin transfectant), and KR (mock transfectant) cells were assayed for their expression of integrin α2 (I) and α3 (J) subunits.
Suppression of Peritoneal Implantation by Anti-Integrin Antibodies.
Because the antibodies against integrin α2 and α3 subunits suppressed the adhesion of the four cell lines to the peritoneum, we examined the effects of these antibodies on peritoneal implantation after inoculation of NUGC-4 cells into the abdominal cavity of athymic mice. Four weeks after the inoculation, the numbers of disseminated nodules were reduced by the pretreatment of cells with either the anti-α2 or anti-α3 integrin antibody (Table 1) ⇓ , suggesting the involvement of these integrins in the i.p. growth of carcinoma cells. The suppressive effect of the anti-α3 integrin antibody was, however, found to be more prominent than that of the anti-α2 integrin antibody.
Effect of antibodies against α2 and α3 integrin subunits on peritoneal implantation of NUGC-4 human gastric carcinoma cells in athymic mice
Adhesion of α2 and α3 Integrin-Expressing Cells to the Peritoneum.
We next evaluated the adhesion capacities of α2 and α3 integrin-expressing cells, which were established from K562 erythroblastoid-leukemic cells by cDNA transfection. Fig. 2, I and J ⇓ , shows the expression profiles of these integrin molecules in K2F10 cells (α2 integrin-transfectant) and K3-9 cells (α3 integrin-transfectant). Both K2F10 and K3-9 cells adhered to the peritoneum more firmly than did the mock transfectant KR cells (Fig. 3) ⇓ . The increase in the adhesion of K3-9 cells was greater than that of K2F10 cells (a 3.2-fold increase for K3-9 cells versus a 1.7-fold increase for K2F10 cells, when compared with the adhesion of KR cells). The increases in the adhesion of K3-9 and K2F10 cells were inhibited by the addition of corresponding monoclonal antibodies, suggesting that the enhanced adhesion by cDNA transfection was due to the specific function of the integrin molecules expressed on these cells.
Adhesion of α2 and α3 integrin-expressing cells to peritoneum. K2F10 (α2 integrin transfectant), K3-9 (α3 integrin transfectant), and KR (mock transfectant) cells were labeled with a fluorescent dye (BCECF-AM) and incubated with excised peritoneum at 37°C for 40 min. The fluorescence intensity associated with the adherent cells was measured with a fluorescence spectrophotometer. The assays were carried out in triplicate, and the error bars indicate the SDs.
K3-9 cells and KR cells were separately labeled, each with a distinct fluorescence probe. The cells were then mixed together and incubated with a peritoneum specimen. By fluorescence microscopic observation, we were able to distinguish K3-9 cells with red fluorescence from KR cells with green fluorescence present in the same field. As shown in Fig. 4 ⇓ , the proportion of K3-9 cells adherent to the peritoneum was much larger than that of KR cells. By comparing the numbers of the two types of attached cells in the same area of the peritoneum, the possible contribution of heterogeneity in the peritoneal lining to adhesion can be eliminated.
Adhesion of α3 integrin-expressing cells to peritoneum. K3-9 (α3 integrin transfectant) cells were labeled with CMTMR (red fluorescence), and KR (mock transfectant) cells were labeled with CMFDA (green fluorescence). The mixture of these cells was incubated with excised parietal peritoneum at 37°C for 40 min. After the nonadherent cells were removed by gentle washing, the adherent cells were observed under fluorescence microscope. The same field was observed with different filter units: A, CMTMR-labeled K3-9 cells; B, CMFDA-labeled KR cells.
We then assayed the adhesion of K3-9 cells to cultured peritoneal mesothelial cells (Fig. 5) ⇓ . The adhesion of both K3-9 and KR cells to a monolayer of mesothelial cells was increased as compared with their adhesion to Matrigel-coated plates, which were used for the mesothelial cell culture. K3-9 cells, however, adhered to mesothelial cells more firmly than KR cells did. The addition of the anti-α3 integrin antibody decreased the adhesion to the control level, showing that the increased adhesion was α3 integrin dependent.
Adhesion of α3 integrin-expressing cells to peritoneal mesothelial cells. K3-9 (α3 integrin transfectant) and KR (mock transfectant) cells were labeled with a fluorescent dye (BCECF-AM) and incubated with a monolayer of mesothelial cells at 37°C for 40 min. The fluorescence intensity associated with the adherent cells was measured with a fluorescence spectrophotometer. The assays were carried out in triplicate, and the error bars indicate the SDs.
Expression of Ligands for α3β1 Integrin in Mesothelial Cells.
To assess the expression of ligands for the α3β1 integrin in mesothelial cells, we attempted to detect mRNA for laminin-5 and laminin-10/11, which have been reported to serve as high-affinity ligands for this integrin (17, 18, 19, 20) . When the products of the reverse transcription-PCR were separated on agarose gel electrophoresis, the expected PCR product (586 bp) for the murine laminin α3 chain (a component of laminin-5) was obtained from mesothelial cells isolated from both the parietal peritoneum and diaphragm (Fig. 6) ⇓ . A PCR with a set of primers for the laminin α5 chain (a component of laminin-10/11) also gave a band with the expected size (330 bp). The identities of these PCR products were confirmed by restriction endonuclease digestion. When the 586-bp product was digested with MboI, three fragments (245, 193, and 113 bp) were obtained (15) . On the other hand, the digestion of the 330-bp product with ApaI yielded two fragments (223 and 107 bp; Ref. 16 ). These results indicate that mesothelial cells from both the parietal peritoneum and the diaphragm express mRNA for laminin-5 and laminin-10/11, which are most likely to mediate the interaction of α3-expressing carcinoma cells with the peritoneum.
Analysis of mRNA for laminin α3 and α5 chains by the reverse transcription-PCR method. RNA was isolated from murine peritoneal mesothelial cells and subjected to reverse transcription-PCR by the use of a set of primers for laminin α3 chain (Lanes 1 and 2) and laminin α5 chain (Lanes 3 and 4) under the conditions described in Materials and Methods. Lanes 1 and 3, mesothelial cells from parietal peritoneum; Lanes 2 and 4, mesothelial cells from diaphragm.
We then examined the effects of antibodies against laminin isoforms on the adhesion of MKN1 cells to murine peritoneum. In the presence of anti-laminin-5 polyclonal antibody (120 μg/ml), the cell adhesion was decreased to 55% that obtained in the control experiment (Fig. 7) ⇓ . Anti-laminin-1 antibody, however, showed no significant inhibition of the adhesion of MKN1 cells. The inhibition by the anti-laminin-5 antibody was significant but moderate, probably due to the possible contribution of laminin-10/11 to the interaction of carcinoma cells. It is likely that the combination of antibodies against laminin-5 and laminin-10/11 would inhibit cell adhesion more efficiently. This should be examined once anti-laminin-10/11 antibody becomes available, but the results of the present study suggest that laminin-5 plays at least a partial role in the adhesion of α3 integrin-expressing cells to the peritoneum.
Effects of antibodies against laminin isoforms on the adhesion of MKN1 gastric carcinoma cells to parietal peritoneum. Fluorescently labeled cells were overlaid on excised peritoneum that had been treated with antibody against laminin-5 or laminin-1 at 0°C for 30 min. After the cells and peritoneum were incubated at 37°C for 40 min, the nonadherent cells were removed by gentle washing three times with a warm medium. The adherent cells were then lysed with 1% NP40, and the fluorescence intensity was measured with a fluorescence spectrophotometer (Ex = 490 nm, Em = 520 nm). The assays were carried out in triplicate. The error bars indicate the SDs.
DISCUSSION
The attachment of carcinoma cells to the peritoneum is an important step in the initial process of metastatic dissemination of gastric cancer. In this study, we presented evidence that α3β1 integrin plays a crucial role in the adhesion of cancer cells to the peritoneum: (a) inhibition of gastric carcinoma cell adhesion to excised peritoneum by specific monoclonal antibodies (Fig. 1) ⇓ ; (b) suppression of i.p. growth of NUGC-4 cells by pretreatment with anti-α3 integrin antibody (Table 1) ⇓ ; and (c) enhancement of the adhesion of K562 cells to the excised peritoneum and to a monolayer of peritoneal mesothelial cells by cDNA transfection (Figs. 3 ⇓ 4 ⇓ 5 ⇓ ). In peritoneal mesothelial cells, we detected the expression of mRNA for laminin-5 and laminin-10/11, which have recently been identified as high-affinity ligands for α3β1 integrin (Fig. 6) ⇓ . Furthermore, the pretreatment of excised peritoneum with antibody to laminin-5 significantly inhibited the adhesion of carcinoma cells (Fig. 7) ⇓ . With these findings taken together, we can say that α3β1 integrin is a strong candidate as the mediation source in the initial attachment of gastric cancer cells during their peritoneal dissemination processes. These results may provide a biochemical basis for observations in a previous clinical report. In that report, histochemical studies on clinical specimens of gastric cancer associated the expression of α3β1 integrin with peritoneal metastasis (8) . Another report suggested the importance of α2β1 integrin by virtue of its significant expression on a larger proportion of tumor cells in peritoneal metastases than in primary tumors (9) . The present study also showed α2β1 integrin to play a role in adhesion, although to a lesser extent. Taken together, these findings suggest that the attachment of cancer cells to peritoneum is mediated by multivalent interactions with various adhesion molecules, including α2β1 and α3β1 integrins, and that the individual adhesion molecules responsible for the interaction might depend on the type of carcinoma. Nishimura et al. (21) and Kawamura et al. (22) also reported that the expression of α2β1 and/or α3β1 integrin on gastric carcinoma cells was correlated with their peritoneal metastasizing potentials, i.e., highly metastatic cells expressed higher levels of these integrins as compared with the less-metastatic cells. These results, along with those of the present study, strongly suggest that the initial attachment of cancer cells to peritoneal lining mediated by α2β1 and/or α3β1 integrin is a key step toward the subsequent process of metastasis formation. Recently, Ishii et al. (23) demonstrated the inverse correlation between the expression level of α6β4 integrin and the frequency of peritoneal dissemination of gastric cancer from an immunohistochemical study. The finding is of interest because the expression of α6β4 integrin and α3β1 integrin apparently had opposite effects even though these integrins share a common ligand, laminin-5. One possible explanation is that the intracellular signaling events occurring after the ligand is bound to α6β4 might differ from those occurring after binding to α3β1 integrins.
Our previous reports showed that the expression of α3β1 integrin at both the protein and mRNA levels was increased after the oncogenic transformation of fibroblasts by SV40 or polyoma virus (24 , 25) . A number of studies have demonstrated that the increased expression of α3β1 integrin in various tumor cells is correlated with their malignant phenotypes, including invasion and metastasis (26, 27, 28, 29, 30, 31, 32) . Furthermore, the treatment of melanoma and glioma cells with antibodies against α3 integrin was shown to inhibit their ability to migrate and invade (11 , 26 , 33) . These studies suggest that α3 integrin is also involved in the invasion of gastric carcinoma cells into the peritoneum following their attachment to the peritoneal lining and are in agreement with our result that the anti-α3 integrin antibody suppresses the i.p. growth of NUGC-4 cells (Table 1) ⇓ . Our preliminary experiment also showed that the i.p. implantation of α3 integrin-expressing cells (K3-9 cells) in severe combined immunodeficiency mice resulted in a poor survival rate as compared with the implantation of mock transfected cells (KR cells). 3
The integrin-extracellular matrix interaction facilitates the transduction of signals into cells. This leads to regulation of the production of extracellular matrix-degrading enzymes. It has recently been reported that the treatment of invasive mammary carcinoma cells with an anti-α3 integrin antibody reduced the production of matrix metalloproteinase (MMP)-9 (34) . In human glioma and rhabdomyosarcoma cells, anti-α3 integrin antibodies enhanced their invasion through the Matrigel in association with the increased levels of MMP-2 secretion (35 , 36) , suggesting that the antibody acted as an agonist. In addition, laminin-5 was shown to induce MMP-9 from human melanoma cells (11) . Laminin-5 was also found to be specifically cleaved by MMP-2 and MT1-MMP and the modified laminin-5 induced migration of breast epithelial cells (37 , 38) . These observations suggest that the expression of extracellular matrix-degrading enzymes is regulated, at least in part, through the integrin-mediated pathways.
Recently, we found that the Ets family of transcription factors regulated the expression of α3 integrin (39) . The Ets transcription factors were also reported to be involved in tumor metastasis through the promotion of angiogenesis and the expression of MMPs (40, 41, 42) . These observations suggest that α3β1 integrin and MMPs cooperatively promote adhesion and invasion of cancer cells during the metastatic processes. In conclusion, the present study strongly suggests that the up-regulation of α3β1 integrin on gastric carcinoma cells facilitates their adhesion to the peritoneum and the subsequent development of peritoneal dissemination.
Acknowledgments
We thank Miki Ogiwara, Kunie Tanaka, Manami Nishiwaki, and Masashi Takada for their technical assistance.
Footnotes
-
Grant support: Ministry of Education, Culture, Sports, Science and Technology of Japan.
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-
Requests for reprints: Tsutomu Tsuji, Department of Microbiology, Hoshi University School of Pharmacy and Pharmaceutical Sciences, 2-4-41 Ebara, Shinagawa-ku, Tokyo 142-8501, Japan. Phone/Fax: 81-3-5498-5753; E-mail: tsuji{at}hoshi.ac.jp
-
↵3 H. Takatsuki, S. Komatsu, R. Sano, and T. Tsuji, unpublished observation.
- Received January 31, 2004.
- Revision received May 14, 2004.
- Accepted June 24, 2004.
- ©2004 American Association for Cancer Research.