Junctional Adhesion Molecule-C Promotes Metastatic Potential of HT1080 Human Fibrosarcoma*

The junctional adhesion molecule (JAM) family is a key molecule in a process called transendothelial migration or diapedesis. Here, we report implications of JAM-C in cancer metastasis. We first determined the mRNA expression of JAMs in 19 kinds of cancer cell lines. JAM-C was expressed in most of tumors having potent metastatic properties. Especially in murine K-1735 melanoma cell lines, the highly metastatic sublines (M2 and X21) strongly expressed JAM-C when compared with the poorly metastatic ones (C-10 and C23). Next, we investigated the role of JAM-C in cancer metastasis by using human JAM-C (hJAM-C) gene-transfected HT1080 fibrosarcoma cells. In comparison with mock-transfected HT1080 cells, these cells showed a significant increase in the adhesion to various extracellular substrates and the invasion across a Matrigel™-coated membrane. The knockdown of hJAM-C using small interfering RNA resulted in the suppression of both the adhesion and the invasion of HT1080 cells, suggesting that endogenous hJAM-C might be involved in tumor metastasis. Finally, we studied the role of hJAM-C in an in vivo experimental metastatic model. The results showed that the overexpression of hJAM-C in HT1080 cells significantly decreased the life spans of the tumorbearing mice. In contrast, the knockdown of hJAM-C in HT1080 cells suppressed the weight gain of the lungs with metastatic colonies. We conclude that the expression of JAM-C promotes metastasis by enhancing both the adhesion of cancer cells to extracellular matrices and the subsequent invasion.

The junctional adhesion molecule (JAM) family is a key molecule in a process called transendothelial migration or diapedesis. Here, we report implications of JAM-C in cancer metastasis. We first determined the mRNA expression of JAMs in 19 kinds of cancer cell lines. JAM-C was expressed in most of tumors having potent metastatic properties. Especially in murine K-1735 melanoma cell lines, the highly metastatic sublines (M2 and X21) strongly expressed JAM-C when compared with the poorly metastatic ones (C-10 and C23). Next, we investigated the role of JAM-C in cancer metastasis by using human JAM-C (hJAM-C) gene-transfected HT1080 fibrosarcoma cells. In comparison with mock-transfected HT1080 cells, these cells showed a significant increase in the adhesion to various extracellular substrates and the invasion across a Matrigel TM -coated membrane. The knockdown of hJAM-C using small interfering RNA resulted in the suppression of both the adhesion and the invasion of HT1080 cells, suggesting that endogenous hJAM-C might be involved in tumor metastasis. Finally, we studied the role of hJAM-C in an in vivo experimental metastatic model. The results showed that the overexpression of hJAM-C in HT1080 cells significantly decreased the life spans of the tumorbearing mice. In contrast, the knockdown of hJAM-C in HT1080 cells suppressed the weight gain of the lungs with metastatic colonies. We conclude that the expression of JAM-C promotes metastasis by enhancing both the adhesion of cancer cells to extracellular matrices and the subsequent invasion.
Cancer metastasis involves a series of events that include dissociation of malignant cells from a primary site, polarized proteolysis and migration, intravasation into the circulatory system, and adhesion to the vascular endothelium followed by extravasation, invasion, and growth at distant sites (1). Certain cell surface molecules are known to be involved in these processes. For instance, integrins play central roles in regulating cell adhesion, motility, invasion, and angiogenesis (2)(3)(4), and matrix metalloproteinases on tumor cells can degrade the extracellular matrix (ECM) 2 (5). In particular, cell adhesion molecules play key roles in tumor adhesion and invasion, resulting in metastasis (6). When compared with normal tissues, malignant tumors are characterized by disrupted tissue architecture and deranged differentiation. Changes in the expression or function of cell adhesion molecules contribute to tumor progression both by altering the adhesion status of the cell and by affecting cell signaling (7). It has been reported that alteration of cell-cell and/or cell-matrix interactions accounts for the ability of cancerous cells to cross tissue boundaries and to disseminate to distant sites. Many adhesion molecules implicated in tumor metastasis have been identified (6), and certain of them belong to the immunoglobulin superfamily (8).
Recently, the junctional adhesion molecule (JAM) family, a member of the immunoglobulin superfamily, has become a focus of study in relation to cell-cell and/or cell-matrix interactions.
The JAM family, belonging to the larger cortical thymocyte Xenopus molecular family (9), is a glycoprotein. It has two immunoglobulin folds (V H and C2 type) and two extra cysteine residues in its extracellular region, a potential N-glycosylation site (s), and a PDZ protein-interacting domain in its intracellular region (10). JAM molecules are selectively expressed in a variety of human organs (11)(12)(13). In particular, JAMs are strongly expressed in close proximity to tight junction strands of polarized endothelial and epithelial cells and on circulating leukocytes and platelets (14,15).
JAMs are known to not only form homophilic binding but also to interact with certain ligands in a heterophilic manner (16 -20). It has been reported that JAM-A binds to ␤2 integrin LFA-1 (21), that JAM-B interacts with ␣4␤1 integrin (22), and that JAM-C is a counter-receptor for the leukocyte integrin Mac-1 (15). Several cytoplasmic partners associated with the PDZ domain of the JAMs have also been reported (23)(24)(25)(26)(27)(28). These interactions represent the likely determinants for the diverse roles of JAMs such as junction assembly, platelet activation, and leukocyte transmigration (10). In inflammation, the JAMs play an important role on the passage of leukocytes across interendothelial spaces (29,30). Certain adhesion molecules on leukocytes and endothelial cells are known to control sequential steps in leukocyte migration (31). The steps of leukocyte migration are partly similar to those of cancer metastasis in several respects; e.g. both leukocytes and tumor cells need to adhere to the vascular bed and to cross the border of endothelial cells.
In the present study, we investigated certain aspects of the participation of JAMs in cancer metastasis. At first, we determined the expression of JAMs in 14 cell lines and also in five sublines of the K-1735 murine melanoma. Next, we focused on the function of JAM-C in cancer metastasis and used HT1080 human fibrosarcoma cells (HT1080 cells) for metastatic studies. HT1080 cells transfected with human JAM-C (hJAM-C) were used in the experiment of cell adhesion, invasion, and in vivo metastasis for evaluating the functions of hJAM-C. The roles of endogenous hJAM-C were determined by examining the properties of HT1080 cells transfected with small interfering RNA for hJAM-C. This study provides the first evidence that hJAM-C promotes cancer metastasis.
The cells on culture plates were rinsed with phosphate-buffered saline (pH 7.4, PBS). Cell extracts were prepared with lysis buffer composed of 10 mM Tris (pH 7.5), 0.1% SDS, 50 g/ml aprotinin, 200 M leupeptin, 2 mM phenylmethylsulfonyl fluoride, and 100 M pepstatin A. Total protein concentration was measured by using a BCA protein assay reagent kit (Pierce) with bovine serum albumin as a standard, according to the manufacturer's instructions. The cell extracts were subjected to 10 or 15% SDS-PAGE and transferred electrophoretically to polyvinylidene difluoride membranes (Millipore, Billerica, MA). After blocking for 1 h at room temperature with 5% skim milk in Tris-HCl buffered saline containing 0.1% Tween 20 (TTBS, pH 7.4), the membranes were incubated with 0.1 g/ml antibody for 2 h at room temperature. Then, they were incubated with peroxidase-conjugated secondary antibody at a dilution of 1:2000 to 1:10,000 for 1 h at room temperature. Each sample was developed by using chemiluminescent substrate (ECL; Amersham Biosciences).
cDNA Cloning-The hJAM-C sequence was obtained from GenBank TM (accession number AF448478). Total RNA was isolated from human umbilical vein endothelial cells (Bio-Whittaker, Walkersville, MD) by using TRIzol LS reagent and reverse-transcribed into single-stranded cDNA by using the SuperScript first-strand synthesis system (Invitrogen). The fulllength cDNA of hJAM-C was PCR-amplified by using Ex Taq TM (TaKaRa BIO Inc., Shiga, Japan) with specific primers based on the hJAM-C nucleotide sequence that included BamHI and HindIII restriction sites: 5Ј-ACG TAA GCT TAT GGC GCT GAG GCG GCC ACC-3Ј (sense) and 5Ј-ACG TGG ATC CTC AGA TCA CAA ACG ATG ACT-3Ј (antisense). The PCR product was purified by use of a High Pure PCR product purification kit (Roche Diagnostics, Basel, Switzerland). Restriction enzyme digestions of PCR product and pcDNA3.1/Hygro vector (Invitrogen) were carried out at 37°C for 1 h. Both digestion products were purified with the purification kit and then ligated at 16°C for 2 h with T4 DNA ligase (DNA ligation kit, Version 1, TaKaRa). Competent Escherichia coli JM109 cells (TaKaRa) were transformed with ligation mixture, plated on Luria-Bertani (LB) agar plates, and grown overnight at 37°C. Colonies were screened for the insert by restriction enzyme digestion, and DNA sequencing was performed at Macrogen Inc. (Seoul, Korea).
Transfection with hJAM-C-Stable GFP transfectant of HT1080 fibrosarcoma was prepared as described previously (32) and used for transfection with hJAM-C. These cells were transfected with 10 g of pcDNA3.1/Hygro vector containing either full-length hJAM-C or no insert by using Lipofectamine TM 2000 (Invitrogen). Stable transfectant (hJAM-C/HT1080 or Mock/HT1080) was selected with 400 g/ml hygromycin B. The expression of hJAM-C was confirmed by Western blotting as described above. For cell growth assay, hJAM-C/HT1080 or Mock/ HT1080 cells (1.0 ϫ 10 5 cells/dish) were cultured for 24, 48, 72, or 96 h. After staining with 0.5% trypan blue, the numbers of living cells were counted at each time point.
Localization of hJAM-C-Human JAM-C/HT1080 or Mock/ HT1080 cells were fixed with 3.7% paraformaldehyde in PBS. The cells were then blocked with 3% bovine serum albumin in PBS, incubated with 0.1 g/ml anti-JAM-C goat polyclonal antibody for 2 h, and immersed in 1.0 g/ml Alexa Fluor 555labeled anti-goat IgG secondary antibody (Molecular Probes Inc., OR) for 1 h. Localization of hJAM-C was determined by using a LSM510 Meta confocal system (Carl Zeiss Co., Ltd.).
Adhesion Assay-Ninety-six-well flat-bottomed plates were dry-coated with Matrigel TM (3 g/well), fibronectin (0.75 g/well), vitronectin (0.25 g/well), laminin (1.25 g/well), or type IV collagen (1.5 g/well) and blocked with 1% bovine serum albumin for 1 h. Human JAM-C/HT1080 or Mock/ HT1080 cells (5 ϫ 10 4 cells/well) were allowed to adhere to each substrate-coated well at 37°C for 1 h. After washing with PBS, the adherent cells were stained with crystal violet for 10 min. The pigment was extracted with 33% acetic acid for 5 min, and the absorbance was measured at 570 nm as an indicator of adherent cells.
Invasion Assay-The invasion was evaluated as the capacity of hJAM-C/HT1080 cells to pass through a Matrigel-coated transwell insert. BD Matrigel matrix (25 g/well; BD Biosciences) was coated on a fluorescence-blocking micropore membrane of a culture insert (8-m pore size, 0.28 cm 2 , Falcon HTS FluoroBlok TM inserts, BD Biosciences). The culture inserts were set into wells of a 24-transwell plate containing RPMI 1640 medium supplemented with 10% FBS, which had not been heat-immobilized, as a chemoattractant. After the cell density had been adjusted to 5 ϫ 10 5 cells/ml in serum-free medium, 5 ϫ 10 4 cells in 100 l were introduced into each culture insert. After incubation for 4 -5 h at 37°C in a CO 2 incubator, the cells that traversed the Matrigel layer and attached to the lower surface of the filter were counted in five randomized fields under a fluorescence microscope.
Transfection with siRNA-The nucleotide sequences of siRNA with a two-nucleotide overhang (underline) for hJAM-C were 5Ј-CAGGAUGGAGAAAGUUACAAGAAC-CAG-3Ј (sense) and 5Ј-GGUUCUUGUAACUUUCUCCAU-CCUGAU-3Ј (antisense); those for luciferase were 5Ј-UGG-UUUACCUAAAGGUGUACGAAUUAG-3Ј (sense) and 5Ј-AAUUCGUACACCUUUAGGUAAACCAAU-3Ј (antisense); and those for EGFP were 5Ј-GGCUACGUCCAGGAGCGC-ACC-3Ј (sense) and 5Ј-UGCGCUCCUGGACGUAGCCUU-3Ј (antisense). The sequences for hJAM-C, luciferase, and EGFP correspond to nucleotide regions 829 -853, 612-636, and 118 -141, respectively. The siRNAs for luciferase and EGFP were used as a control siRNA. All siRNAs were prepared by Hokkaido System Science Co. (Hokkaido, Japan). Stable GFP transfectant of HT1080 fibrosarcoma (31) was transfected with siRNA by using Lipofectamine 2000 according to the manufacturer's instructions. Forty-eight hours after the addition of siRNA/Lipofectamine 2000 complexes, the knockdown of hJAM-C was evaluated by Western blotting. In all experiments using siRNA, siRNA-transfected HT1080 cells were used at 48 h after the addition of these complexes. Western blotting and cell growth, adhesion, and invasion assays were performed by methods similar to those described above.
Experimental Metastasis-The influence of the overexpression and the knockdown of hJAM-C on tumor metastasis were evaluated in an experimental pulmonary metastatic model. Cells were harvested from culture plates by brief exposure to  0.02% EDTA/PBS. After the cells had been centrifuged and resuspended in the medium, their concentration was adjusted to 5 ϫ 10 6 cells/ml. For evaluating the effect of overexpression of hJAM-C on metastatic potential, hJAM-C/HT1080 or Mock/HT1080 cells (1 ϫ 10 6 cells/mouse) were injected into the tail vein of 4-week-old BALB/c nu/nu male mice (Charles River Japan Inc., Kanagawa, Japan). The life spans of tumor-bearing mice were monitored (n ϭ 10 for mock/ HT1080, n ϭ 11 for hJAM-C/ HT1080). In another three mice in each group, the lungs were dissected, weighed, and photographed for evaluating metastasis at 19 days after implantation.
For evaluating the effect of knockdown of hJAM-C on metastatic potential, hJAM-C siRNA-or EGFP siRNA-treated cells (1 ϫ 10 6 cells/mouse) were injected into the tail vein of 9-week-old BALB/c nu/nu male mice (n ϭ 4). The lungs with tumor metastases were dissected, weighed, and photographed at 38 days after implantation.
Histochemical Analysis-The lungs with metastatic colonies were embedded in optimal cutting temperature compound (Sakura Finetechnochemical Co., Ltd., Tokyo, Japan) and frozen at Ϫ80°C. Six-micrometer lung sections were prepared by using a cryostatic microtome (HM 505E, Microm, Walldorf, Germany), stained with hematoxylin-eosin, and examined histopathologically.
Statistical Analysis-Statistical analysis (mean value, S.D., independent t test) was performed by using Stat View 4.0 (SAS Institute Inc., Cary, NC). Results were expressed as the mean Ϯ S.D.

RESULTS
The amounts of JAM transcripts in cancer cell lines were determined by reverse transcription-PCR. In this experiment, we used 19 kinds of cancer cells lines. Among them, human HT1080 fibrosarcoma, HT-29 colon adenocarcinoma, C8161 melanoma,  SUIT-2 pancreatic carcinoma, AZ-P7a gastric carcinoma, and murine Lewis lung carcinoma, B16BL6 melanoma, Colon26 NL-17 carcinoma, K-1735 M2, and X21 melanoma cell lines were known to have highly metastatic characteristics. The transcripts of JAM-A were similar in amount regardless of the cancer cell lines (Fig. 1). In contrast, the expression levels of JAM-B and JAM-C were different among the various cell lines. The transcripts of JAM-C were detected in highly metastatic cell lines such as HT1080 fibrosarcoma, SUIT-2 pancreatic carcinoma, AZ-P7a gastric carcinoma, Lewis lung carcinoma, B16BL6 melanoma, and Colon26 NL-17 carcinoma (Fig. 1). To assess the relationship between the amounts of JAM transcripts in tumor cells and their metastatic potentials, the expression of JAM members was compared in K-1735 murine melanoma subpopulations having diverse biologic behavior (33). Each clone (C-10, C-23, M2, and X21) was previously established from the parent K-1735 cells. The in vivo metastatic potential of M2 and X21 sublines is very high, as is that of the parent K-1735 cells, whereas that of C-10 and C-23 clones is quite poor. As shown in Fig. 2A, the mRNA level of JAM-C was markedly up-regulated in the parent K-1735 and highly metastatic sublines (M2 and X21) when compared with that in the poorly metastatic sublines (C-10 and C-23). In contrast, the mRNA level of either JAM-A or JAM-B was not much different between the highly metastatic and poorly metastatic cells. Actual expression levels of JAM-C protein in K-1735 sublines determined by Western blotting correlated well with those of JAM-C mRNA (Fig. 2B).
JAM-C functions in cancer metastasis were explored by constructing hJAM-C-overexpressing HT1080 fibrosarcoma cells. HT1080 cells were transfected with hJAM-Cencoded pcDNA3.1/Hygro vector (hJAM-C/HT1080) or empty vector (Mock/HT1080). Overexpression of hJAM-C in hJAM-C/HT1080 cells was confirmed by Western blotting. As shown in Fig. 3A, the amount of hJAM-C protein was very abundant in hJAM-C/HT1080 cells than in Mock/HT1080 cells. Human JAM-C in the transfectant was localized on the extensions of its plasma membrane (Fig. 3B). The forced expression of hJAM-C did not affect the cell growth rate (Fig. 3C).
To evaluate the participation of hJAM-C in the interaction of the cells with ECM, the adhesion of hJAM-C/HT1080 cells to various ECM components was investigated. Human JAM-C expression in HT1080 cells significantly increased their adhesion to Matrigel, fibronectin, and vitronectin and slightly increased their adhesion to type IV collagen and laminin (Fig. 4A). Next, the effect of hJAM-C overexpression on tumor cell invasion was examined. The number of cells that invaded through a Matrigel-coated membrane was compared between hJAM-C/HT1080 and Mock/HT1080 cells. The migration of HT1080 cells was significantly accelerated by the overexpression of hJAM-C, suggesting that hJAM-C expression might be correlated with tumor malignancy (Fig. 4B).
To confirm whether endogenous JAM-C was involved in the invasion, hJAM-C siRNA-transfected HT1080 cells were prepared. The expression of endogenous hJAM-C was observed in both control and luciferase siRNA-transfected HT1080 cells to some extent, and the knockdown of endogenous hJAM-C was confirmed in hJAM-C siRNA-transfected HT1080 cells (Fig.  5A). Transfection with hJAM-C siRNA did not affect the rate of cell growth (Fig. 5B). In an adhesion assay, the knockdown of hJAM-C inhibited the adhesion of hJAM-C siRNA-transfected HT1080 cells to ECM components (Figs. 4A and 5C). Especially, the knockdown of hJAM-C significantly reduced their adhesion to fibronectin and vitronectin. The invasion potential of hJAM-C siRNA-transfected HT1080 cells was significantly less than that of luciferase siRNA-transfected HT1080 cells, indicating that the knockdown of hJAM-C reduced tumor invasiveness (Fig. 5D).
Finally, the influence of hJAM-C expression on the metastatic potential of HT1080 cells was examined in an experimen-tal metastasis model. The formation of pulmonary metastases was observed in both the Mock/HT1080 cell-injected and the hJAM-C/ HT1080 cell-injected group since HT1080 cells are highly metastatic by nature (data not shown). The mean survival times after tumor implantation were 18 Ϯ 2 days for the hJAM-C HT1080 cell-injected group and 30 Ϯ 8 days for the Mock/ HT1080 cell-injected group (Fig.  6A). The life spans of the hJAM-C/ HT1080 cell-injected group were significantly short when compared with those of the Mock/HT1080 cell-injected group (p Ͻ 0.01). Forced expression of hJAM-C caused 40% reduction of the life span of tumor-bearing mice, suggesting that hJAM-C expression promoted the pulmonary metastasis of HT1080 cells. At 19 days after the injection, the hJAM-C/HT1080 cell-injected group showed bloody lungs with metastatic colonies, which might occur due to the increased malignancy of the metastatic cells (Fig. 6B). Most of the mice injected with hJAM-C/ HT1080 cells died before showing obvious differences in the lung weight gain between Mock/HT1080 and hJAM-C/HT1080 cell-injected mice. As shown in Fig. 6C, the results of the hematoxylin-eosinstaining showed the overexpression of hJAM-C-induced expansion of the metastatic tumors. Histochemical data also showed that the knockdown of hJAM-C in HT1080 cells suppressed the pulmonary metastasis of those cells (Fig. 6D). In the case of the injection of EGFP siRNAtransfected cells, the leakages of blood corpuscle cells were observed around the metastases. This phenomenon indicated that the progress of tumor metastasis of the cells transfected with hJAM-C siRNA was slower than that of cells transfected with EGFP siRNA. Furthermore, the knockdown of hJAM-C suppressed the weight gain of lungs with metastatic colonies after the injection of the cells in comparison with that of the cells transfected with EGFP siRNA (p ϭ 0.0570, Fig. 6E).

DISCUSSION
Invasion of metastatic tumor cells to the secondary colonizing organs is often compared with lymphocyte transendothelial migration since several common cell adhesion molecules, such FIGURE 6. Influences of hJAM-C expression on pulmonary metastasis of HT1080 cells. A, the survival times of BALB/c nu/nu mice that had been inoculated with Mock/HT1080 (n ϭ 10, open circle) or hJAM-C/HT1080 (n ϭ 11, closed circle) cells were determined. After intravenous inoculation of these cells, the life spans of tumorbearing mice were monitored. B, the morphology of the lungs bearing Mock/HT1080 or hJAM-C/HT1080 fibrosarcoma was examined. The lungs of each group were dissected and photographed at 19 days after inoculation. Scale bars represent 5 mm. C, histochemical analysis of the lung with metastatic colonies was performed in the hJAM-C overexpression experiment. At 19 days after inoculation, the lung sections were prepared and stained with hematoxylin-eosin. Scale bars represent 500 m. T indicates tumors. D, histochemical analysis of the lung with metastatic colonies was performed in the hJAM-C knockdown experiment. EGFP siRNA-transfected and hJAM-C siRNA-transfected HT1080 cells were injected into the tail vein of BALB/c nu/nu mice. After 38 days after implantation, the lung sections were prepared and stained with hematoxylin-eosin. Scale bars represent 500 m. T indicates tumors. L indicates lymph node. E, the influence of hJAM-C knockdown on pulmonary metastasis of HT1080 cells was examined. EGFP siRNA-transfected (n ϭ 4, open circle) or hJAM-C siRNA-transfected (n ϭ 4, closed circle) HT1080 cells were injected into the tail vein of BALB/c nu/nu mice. After 38 days after implantation, the weights of lungs were measured as an indicator of pulmonary metastasis. Human JAM-C siRNA-transfection tended to decrease the weight gain of lungs with metastatic colonies when compared with mock transfection (p ϭ 0.0570).
as selectins and integrins, are involved in both lymphocyte and tumor cell migration. JAMs are known to participate in the transendothelial migration of lymphocytes (10). Recently, the participation of JAMs in tumor angiogenesis has also been the focus of studies (34,35). Another recent study demonstrated that JAM-C could mediate tumor cellendothelial cell interactions and suggested the involvement of JAM-C in tumor cell metastasis (20).
In the present study, to understand the participation of JAM family molecules in tumor metastasis, we first examined the mRNA levels of the JAM family in various cancer cells derived from human and rodent sources. As a result, we observed cell line-dependent differences in the mRNA level of JAM-C. Interestingly, most of the JAM-C-expressing cell lines in this study have often been used for experimental models of metastasis because of their high malignancy. In murine K-1735 melanoma cell lines, the highly metastatic sublines strongly expressed JAM-C when compared with the poorly metastatic ones. In addition, the expression levels of JAM-A and JAM-B on K-1735 cells did not correlate with the metastatic potential of these cells. These data suggested that JAM-C might be implicated in cancer metastasis.
It has been reported that JAM-C shows only 32 and 36% identity to JAM-A and JAM-B, respectively (13). The intracellular region of JAM-C is slightly longer than that of JAM-A and JAM-B (13). The relative tissue distribution of each JAM shows a distinct pattern, resulting from cell typespecific expression of the individual JAM (9). In this study, the unique expression pattern of JAM-C was observed in various cancer cell lines. These observations may explain the distinctive roles of JAM-C in cancer metastasis. The full open reading frame for hJAM-C displays 84% identity with that for murine JAM-C (mJAM-C) at the DNA level and 86% identity at the amino acid level (37). The tissue distribution of hJAM-C does not overlap entirely with that of the murine homologue (13). We also examined mJAM-C function in the invasion assay, with the result that overexpression of mJAM-C also promoted tumor invasion as well as hJAM-C did (data not shown).
In this study, the roles of JAM-C in cancer metastasis were examined by using hJAM-C/HT1080 cells engineered to express hJAM-C abundantly. When compared with Mock/ HT1080 cells, these cells showed significantly increased adhesiveness to various ECM components and invasiveness across Matrigel-coated membranes. In addition, the experiments using hJAM-C siRNA showed that endogenous hJAM-C would be involved in tumor adhesion and subsequent invasion. It is not clear whether JAM-C expression enhanced the adhesion of HT1080 cells to various ECM components directly or indirectly. In the case of JAM-A, it has been reported that it regulates ␤1 integrin expression and cellular adhesion property (36). It is possible that JAM-C may affect the cell adhesiveness by controlling other adhesion molecules. In addition, JAM-C may function cooperatively with certain proteinases since the degradation of Matrigel requires proteinase activity.
Finally, we demonstrated that hJAM-C promoted pulmonary metastasis in the experimental metastatic model in vivo. We conclude that the expression of JAM-C promotes cancer metastasis by enhancing the adhesion of and subsequent invasion of cancer cells. The present study is the first report that demonstrates that JAM-C expression promotes tumor malignancy and thus provides a new insight into the mechanisms of cancer metastasis.