Syndecan-2 Functions as a Docking Receptor for Pro-matrix Metalloproteinase-7 in Human Colon Cancer Cells*

Although elevated syndecan-2 expression is known to be crucial for the tumorigenic activity in colon carcinoma cells, how syndecan-2 regulates colon cancer is unclear. In human colon adenocarcinoma tissue samples, we found that both mRNA and protein expression of syndecan-2 were increased, compared with the neighboring normal epithelium, suggesting that syndecan-2 plays functional roles in human colon cancer cells. Consistent with this notion, syndecan-2-overexpressing HT-29 colon adenocarcinoma cells showed enhanced migration/invasion, anchorage-independent growth, and primary tumor formation in nude mice, paralleling their morphological changes into highly tumorigenic cells. In addition, our experiments revealed that syndecan-2 enhanced both expression and secretion of matrix metalloproteinase-7 (MMP-7), directly interacted with pro-MMP-7, and potentiated the enzymatic activity of pro-MMP-7 by activating its processing into the active MMP-7. Collectively, these data strongly suggest that syndecan-2 functions as a docking receptor for pro-MMP-7 in colon cancer cells.

The characteristics and functions of cancer cells are critically influenced by the actions of cell adhesion receptors, which mediate interactions of cancer cells with the extracellular matrix (ECM) 3 and the cytoskeleton (1). At different points during carcinogenesis, the cell adhesion receptors regulate various cancer cell functions, including cell growth, differentiation, cell survival, angiogenesis, and inflammation (2,3). Thus, the cancer-specific characteristics and functions of cancer cells are due to the expression and utilization of a distinct set of adhesion receptors that show different expression patterns in normal cells. One group of cancer-related cell adhesion receptors are the syndecans, which are cell surface heparan sulfate proteoglycans known to play diverse roles in cell adhesion and cell communication by serving as co-receptors for both cell signaling and ECM molecules (4). At the plasma membrane, syndecans are capable of transmitting signals from the extracellular environment to the intracellular compartment, thereby regulating adhesion-dependent signal transduction during cell growth (5,6), cell adhesion and migration (6,7), cytoskeleton organization (7,8), and cell differentiation (9). Numerous studies have examined the function of syndecans in various human tumors. Syndecan-1 expression is down-regulated in a great number of squamous cell carcinomas, including uterine cervix, lung, and colorectal cancer (2, 10 -12). However, in other studies, syndecan-1 expression is reportedly up-regulated in prostate, lung, and breast cancers (13)(14)(15). In the case of syndecan-2, it has been reported that, in normal tissues, syndecan-2 is expressed in mesenchymal cells but not in normal epithelial cells. However, we found that syndecan-2 expression is increased in several epithelial-driven colon carcinoma cells, and this up-regulation is necessary for the tumorigenic activity of colon carcinoma cells (16). Therefore, it is likely that altered expression of syndecan-2 allows normal epithelial cells to become tumorigenic. The exact molecular mechanisms underlying syndecan-2-mediated carcinogenesis have not yet been fully elucidated. ECM remodeling is orchestrated by distinct proteolytic systems capable of hydrolyzing a wide spectrum of ECM components (17). The involved proteases include the matrix metalloproteinases (MMPs), a large group of enzymes capable of degrading many of the ECM components (18). Because ECM remodeling is an important factor in carcinogenesis, MMPs have been extensively studied in the context of carcinogenesis (19). One of the smallest known members of the MMP family, MMP-7, was first discovered as an enzyme of the involuting rat uterus and was later found to be an important marker in human cancer progression (17). MMP-7 gene expression has been reported in human cancers of the colon, breast, prostate, stomach, pancreas, kidney, and lung (20). Recent studies have shown that MMP-7 interacts with the specific molecular genetic and signaling pathways involved in colorectal cancer development (21,22); in particular, MMP-7 is activated at an early stage of colorectal tumorigenesis by the ␤-catenin signaling pathway (23).
Studies have shown that the cell surface localization of MMPs, which is tightly regulated in normal cells, is very important for carcinogenesis-related processing. MMP-2 may be localized to the cell surface through interactions with integrin ␣v␤3 (24) or MT1-MMP (24), while the cell surface heparan sulfate proteoglycan, CD44, may dock MMP-7 (25) and  to the cell surface. Therefore, the formation of MMP⅐adhesion receptor complexes appears to be a common pathway through which soluble MMPs are localized to the cell surface.
Because both syndecan-2 and MMP-7 are important regulators in colon carcinogenesis, we herein examined whether syndecan-2 and MMPs might cooperatively regulate tumorigenic activities in human colon cancer. Our results reveal that syndecan-2 expression is up-regulated in colon adenocarcinomas, where it functions as a docking receptor for pro-MMP-7 to regulate tumorigenic activity.
Cell Culture-Colon cancer cell lines HT-29 were purchased from the Korean cell line bank. SW480 cells were generous gift from Dr. Jung H. Park of the Hallym University in Korea. HT-29 cells were maintained in McCoy's 5A and SW480 cells in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum (FBS) and with penicillin (100 units/ml) and streptomycin (10 g/ml, Invitrogen) at 37°C in a humidified 5% CO 2 incubator.
Plasmid Construct-A series of N-terminal truncation syndecan-2 mutants were generated with pairs of degenerate PCR primers and cloned into pGEX-5X-1 vector at the EcoRI/ BamH1 cloning sites. Human MMP-7 cDNA was amplified by RT-PCR using HT29 cell line RNA as a template. The DNA fragment precisely encompassing the full-length pro-MMP-7, pro-domain, or the catalytic domain mutants were cloned into the pET22b(ϩ) vector at the NdeI/BamH1 sites. FLAG-tagged syndecan-2 cDNA was a generous gift from Dr. Y. Yamaguchi (Burnham Institute for Medical Research, San Diego, CA).
Synthesis of Small Interfering RNA Constructs-Oligonucleotides were designed targeting the human MMP-7 RNA, containing a 9-bp hairpin loop. Oligonucleotides were annealed and cloned into the BglII/EcoRI sites of pSUPERretro vector.
Transfection of Syndecan-2 cDNA-For stable transfection, HT-29 cells (3 ϫ 10 6 cells) were trypsinized, harvested, washed twice with cold PBS, and then incubated with 4 g of plasmid DNA on ice for 20 min. Transfection was preceded by electroporation using 0.2 kV, 1 pulse time, and 15 s of length. Clones were selected using G418 (1 mg/ml) for 3 weeks. Syndecan-2 expression was assessed by RT-PCR using a syndecan-2-specific primer. SW480 cells were transfected using Lipofectamine 2000 according to manufacturer's instructions (Invitrogen).
Invasion and Migration Assay-For migration assay, gelatin (10 g/ml) was added to each well of a Transwell plate (Coster, Corning, NY, 8-m pore size), and the membranes were allowed to dry at 25°C for 1 h. The Transwell plates were assembled into a 24-well plate, and the lower chamber was filled with McCoy's 5A media containing 10% FBS, 1% bovine serum albumin, and basis fibroblast growth factor (10 g/ml). Cells (5 ϫ 10 5 ) were added to each upper chamber, and the plate was incubated at 37°C in 5% CO 2 incubator. The cells that had migrated on the lower surface of the filters were stained with 0.6% hematoxylin and 0.5% eosin and counted. For in vitro invasion assay, Transwell plates were coated with gelatin (10 g/ml) on the lower side of the membrane and with Matrigel (3 mg/ml) on the upper side of the membrane.
Cell Proliferation Assay-Cell proliferation was measured by a colorimetric assay using MTT according to the manufacturer's instructions. Briefly, HT-29 cells were harvested with 0.05% trypsin/EDTA and seeded into 48-well plates. After cells were allowed to attach to the plate for 24 h, the medium containing 0.5 mg/ml MTT was added to each plate, and cells were incubated for 1 h. The medium was then removed, and 200 l of dimethyl sulfoxide was added to each plate for half an hour at room temperature. The mean concentration of absorbance at 570 nm in each set of all samples was measured using a 96-well microtiter plate reader (Dynatech, Chantilly, VA).
Flow Cytometry-Cells were washed with PBS and released by the addition of 5% FBS and 0.2% EDTA in PBS through incubation for 30 min at 4°C. After pelletting, cells were resuspended in PBS and counted. Cells were incubated with antisyndecan-2 antibody in 10% FBS in PBS for 1 h at 4°C followed by PBS containing 0.05% Tween 20, washed twice, and incubated with fluorescein isothiocyanate-conjugated anti-mouse in 10% FBS in PBS for 30 min. Syndecan-2 expressions were analyzed by flow cytometry.
Anchorage-independent Growth in Soft Agarose-Each well of a 6-well culture plate was coated with 3 ml of bottom agar mixture (McCoy's 5A/10% FBS/0.6% agar). After the bottom layer had solidified, 2 ml of top agar mixture (McCoy's 5A/10% FBS/0.3% agar) containing 2 ϫ 10 6 cells was added to each well, and the cultures were incubated at 37°C in a 5% CO 2 atmosphere. Every 5 days, normal growth medium was gently layered over the cultures. Colony formation was monitored daily with a light microscope. Colonies in soft agar were photographed with a digital camera after incubation for 3 weeks.
Gelatinolytic Activity Assay-Conditioned media was incubated with 20 g of DQ-gelatin as a substrate in a reaction buffer containing 50 mM Tris-HCl (pH 7.6), 150 mM NaCl, 5 mM CaCl 2 (final 200 l). After 2 h at room temperature, increase in fluorescence at 515 nm due to the DQ-gelatin degradation by the overall gelatinase activities was monitored in triplicate in 96-well plates. Collagenase (0.1 unit/ml) was used as a positive control.
Treatment with NaClO 4 -To inhibit the maturation of the heparan sulfate chain on the cell surface heparan sulfate proteoglycan, HT-29 cells were treated with NaClO 4 (final concentration 25 mM) for 24 h.
Detection of Soluble Syndecan-2 in the Conditioned Media-HT-29 cells were cultured in serum-free media (SFM) for 24 h. Conditioned media (100 l) were collected, briefly centrifuged at 2000 ϫ g to remove any remaining cells, and used for slot blot analysis. For shedding assays, HT-29 cells cultured in 6-well tissue culture plates (6 ϫ 10 5 cells per well) were washed two times with SFM and then cells were incubated with fresh SFM supplemented pro-MMP-7 (0.2 unit/ml) for 2 h 37°C. Conditioned media (100 l) were used for slot blot analysis as described above.
Microarray Analysis-Total RNAs from HT-29 control and HT-29 cells stably overexpressing syndecan-2 were isolated using TRIzol reagent (Invitrogen) following manufacturer's protocol. A reference RNA pool was prepared by way of isolation and mixture of total RNAs 11 cancer cells (27). Reference RNA and target RNA were labeled with Cy3 and Cy5, respectively, during reverse transcription reaction. The labeled cDNAs were co-hybridized in microarrys (Genomictree, Daejeon, Korea) made of oligonucleotides representing 22,000 unique human genes (Sigma-Genosys, St. Louis, MO). Three independent biological replicates were prepared for both cells.
The hybridized slides were washed and then scanned using a GenePix 4000B scanner (Axon Instruments, Foster City, CA), and the data were saved in GenePix Result (GPR) format. Six GPRs were imported into GeneSpring software (Agilent Technologies, Foster City, CA) and "print-tip" normalized using the Lowess function. If Cy5 or Cy3 intensities are not greater than 1.5-fold of background intensities, the spots were removed from further analysis. To select for the genes that show significant difference in expression level between HT-29 control and HT-29 cells stably overexpressing syndecan-2, a two-sample t test was performed. Genes were considered significant if selected at p Ͻ 0.001. A total of 47 genes was selected (Table 1). For heat map presentation of the 47 genes and 6 array samples, a correlation coefficient was used as a gene similarity metric, and a hierarchical clustering algorithm based on the averagelinkage method was employed. The microarray data have been deposited in the ArrayExpress data base (available on-line).
Primary Tumor Formation-HT-29 cells were spun down at 1500 rpm for 5 min, resuspended in the SFM, and then injected into nude mice at a density of 10 ϫ 10 6 cells/300 l. After 4 weeks, the mice were sacrificed and the tumor mass was weighed.
Tissue Sample Preparation-All experiments using the patient tissues were performed with approval from the Internal Review Board of Yonsei University College of Medicine. Tissues from the patients who underwent surgery at the Yonsei University College of Medicine during 1997-1999 were used with permission from the patients. Tissues that contained at least 70% tumor content were stored in liquid nitrogen immediately after surgery.
Immunohistochemistry-Representative core tissue biopsies (2 mm in diameter) were taken from paraffin blocks and arranged in a new recipient block (tissue array block) using a trephine apparatus (Superbiochips Laboratories, Seoul, Korea). 5-m sections from the tissue array block were subjected to immunohistochemical study using the EnVision kit (DAKO, Carpinteria, CA). The deparaffinized sections were boiled for 10 min in citrate buffer (10 mM citric acid, 0.05% Tween 20, pH 6.0) for antigen retrieval, and endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide for 30 min. The sections were preincubated with normal goat serum to prevent nonspecific binding, and then incubated overnight at 4°C with anti-syndecan-2 antibody. The sections was rinsed with PBS and incubated with horseradish peroxidase-labeled polymer for 30 min at room temperature and then washed with PBS. The enzyme reaction was developed with 0.03% 3Ј,3-diaminobenzidine tetrahydrochloride containing 0.006% hydrogen peroxide. As a negative control, an irrelevant antibody replaced the primary antibodies.

Syndecan-2 Is Up-regulated in Human Colon
Adenocarcinoma Tissues-To investigate the potential role of syndecan-2 in colon cancer, we isolated RNA from tissues obtained from 11 colon cancer patients and used RT-PCR to analyze the mRNA expression of syndecan-2 and -4. Gene expression of syndecan-2 was detected in ϳ73% of colon cancer tissue samples, but only 18% of neighboring normal epithelium. In contrast, synde-can-4 mRNA expression was detected in ϳ40% of normal epithelium but actively repressed in colon cancer tissue (Fig. 1A). Consistently, immunohistochemical studies using paraffin sections revealed that syndecan-2 expression was up-regulated in ϳ60% of the examined human colon adenocarcinoma tissues, compared with the neighboring normal epithelium (Fig. 1B). These findings confirm that syndecan-2 expression is up-regulated in human colon cancer cells.
Overexpression of Syndecan-2 Increases the Tumorigenic Potential of Colon Adenocarcinoma Cells-To further investigate the role of syndecan-2 in colon cancer cells, we transfected an expression vector encoding syndecan-2 into HT-29 human colon adenocarcinoma cells, which express very low basal level of syndecan-2 (Fig. 2). HT-29 cells stably overexpressing syndecan-2 (SDC2) showed increased levels of mRNA and cell surface expression ( Fig. 2A). Interestingly, increased syndecan-2 expression was associated with alterations in the epithelial morphology of HT-29 cells, resulting in an aggregated morphology characterized by a lack of clearly discernible cell-cell boundaries ( Fig. 2A). Consistent with these morphological changes, syndecan-2 expression was associated with noticeable increases in cell proliferation ( Syndecan-2 Accelerates Primary Tumor Formation in Vivo-To evaluate the tumorigenic potential of syndecan-2 in vivo, HT-29 or HT-29-SDC2 cells were injected into athymic nude mice and allowed to form tumors over 4 weeks. The mean body weights were not significantly different between the two groups after 4 weeks (data not shown), but the tumor masses were significantly larger in mice injected with HT-29-SDC2 cells ver-sus those injected with the control HT-29 cells (Fig. 3; 0.49 Ϯ 0.14 versus 0.19 Ϯ 0.05, respectively). As expected, syndecan-2 expression was higher in the tumor tissue derived from HT-29-SDC2 cells versus that from control HT-29 cells (Fig. 3D).
Syndecan-2 Expression Leads to Increased MMP-7 Expression in HT-29 Cells-We investigated how syndecan-2 might regulate tumorigenic activity in colon adenocarcinoma cells.
To identify downstream targets of syndecan-2, we used microarray analysis to identify genes that were differently expressed in syndecan-2-overexpressing HT-29-SDC2 cells versus HT-29 controls. After screening of a total of 22,000 human genes, 47 genes showed significant differences in expression levels between the two cell types (p Ͻ 0.001, Fig. 4A and Table 1). The differentially expressed 47 genes represent those with diverse molecular functions, such as "binding" with other proteins, nucleic acids, or lipid, "catalytic activities," including hydrolase, oxidoreductase, or transferases (Fig. 4B). Because MMPs are overexpressed in many cancers, and are  . Out of six cell clones examined, a representative clone was photographed with a phase-contrast microscope attached to a digital camera (right panel). Cells were incubated with an anti-syndecan-2 antibody, and the cell surface expression levels of syndecan-2 were analyzed by flow cytometry. IgG was used as a control (bottom panel). B, cells were split in 48-well plates and incubated for the indicated periods of time. Cell numbers were evaluated with MTT assay, as described under "Materials and Methods." The given data are the mean absorbance Ϯ S.E. from three independent experiments. The asterisk indicates statistically significant increases in proliferation relative to that observed in vector-transfected HT-29 cells (p Ͻ 0.05). C, cells (1 ϫ 10 5 ) were allowed to migrate on gelatin-coated (10 g/ml, migration) Transwell plates or invade through Matrigel-coated (3 g/ml, invasion) Transwell plates. After 18-h incubation, cells were stained with hematoxylin & eosin, and migrated cells were counted. All data are reported as the percentage change in comparison with the vector control cells, which were arbitrarily assigned 100%. The results shown represent the averages of three independent experiments. One or two asterisks indicate statistically significant increases in migration or invasion relative to that of vector control (*, p Ͻ 0.05; **, p Ͻ 0.01). D, cells (2.5 ϫ 10 6 /plate) were seeded in soft agar and allowed to grow for 21 days, and the number of viable colonies was counted (*, p Ͻ 0.05). All data are reported as the percentage change in comparison with the vector control cells, which were arbitrarily assigned 100%. E, SW203 colon adenocarcinoma cells were transfected with syndecan-2 cDNA. After 4 weeks of clonal selection, the expression level of SDC2 was evaluated with RT-PCR, using the indicated primers (top panel). Cells (1 ϫ 10 5 ) were allowed to invade through Matrigelcoated (3 g/ml) Transwell plates as described in C. The results shown represent the averages of three independent experiments (*, p Ͻ 0.05). DECEMBER 18, 2009 • VOLUME 284 • NUMBER 51 capable of mediating invasion and metastasis via ECM degradation (28), we focused on differently expressed members of MMP and tissue inhibitor of matrix metalloproteinase (TIMP) families. We found that one of the highly overexpressed genes in HT-29 following syndecan-2 overexpression was MMP-7 (Table 1). MMP-7 mRNA expression was 3.3-fold higher in HT-29-SDC2 cells (Fig. 4C), and MMP-7 was the only member of the MMPs or TIMP family selected in the current microarray analysis. Consistent with the results from our microarray analysis, gelatinase activities were significantly higher in conditioned media from HT-29-SDC2 cells versus controls (Fig. 4D, p Ͻ 0.01), and syndecan-2 overexpression in HT-29 cells increased MMP-7 expression at the mRNA and protein levels but had no such effect on MMP-2 or MMP-9 (Fig. 4E). These findings collectively suggest that syndecan-2 leads to increased expression of MMP-7, which may be an important regulator in the effects of syndecan-2 in colon cancer cells.

MMP-7 Regulates Syndecan-2-mediated Migration of HT-29
Cells-We next investigated whether MMP-7 is involved in syndecan-2-mediated cell migration. We generated two unique 19-bp siRNA sequences targeted against human MMP-7 mRNA and used these siRNA to knock down MMP-7 expression in HT-29 cells. Co-transfection of HT-29 cells with the syndecan-2 expression vector and MMP-7targeting siRNA showed decreased expression of MMP-7 protein (Fig. 5A), as well as decreased syndecan-2-induced cell migration activity, compared with cells transfected with the syndecan-2 expression vector alone (Fig. 5B, p Ͻ 0.01). In contrast, the migration of HT-29 cells was not significantly influenced by the treatment of FGF-19 whose mRNA expression was increased in HT-29-SDC2 cells (Fig. 5C, p Ͻ 0.01). These results show that MMP-7 is involved in syndecan-2induced cell migration in HT-29 cells, suggesting that syndecan-2-mediated tumorigenic activity is at least in part dependent on MMP-7.

Syndecan-2 Functions as a Cell
Surface Docking Receptor for Pro-MMP-7-Because syndecans are known to interact with various ECM components, we investigated whether syndecan-2 functions as a cell surface receptor in regulating MMP-7. Purified GST-fused proteins of syndecan-2 (GST⅐SDC2), or syndecan-4 (GST⅐SDC4) were incubated with the conditioned media from wild-type HT-29, HT1080 fibrosarcoma, or HT1080 fibrosarcoma cells pretreated with tumor necrosis factor-␣. Western blotting of the precipitated GST⅐SDC2 complexes showed that the syndecan-2 fusion protein bound to 28-kDa pro-MMP-7 in the condition of equal loading of GST-fused protein (Fig. 6A, bottom panel). Furthermore, purified pro-MMP-7 interacted with GST⅐S2E and -SDC2, but not -S2TC, which lacks the extracellular domain (Fig. 6B). These findings indicate the direct interaction between the extracellular domain of syndecan-2 and pro-MMP-7. Cell lysates from HT-29 cells were immunoprecipitated with an anti-syndecan-2 antibody and then incubated with conditioned media from either HT-29vector or HT-29-SDC2 cells. Consistent with the above results, syndecan-2 interacted with pro-MMP-7 in the HT-29conditioned media (Fig. 6C), suggesting that, like recombinant protein, endogenous syndecan-2 expressed in HT-29 cells, which contain glycosaminoglycan chains, can interact with pro-MMP-7. We also found that FLAG-tagged syndecan-2 expressed in HT-29 cells interacted with pro-MMP-7 protein (Fig. 6D). Pretreatment with NaClO 4 to inhibit the maturation of the heparan sulfate chain on the cell surface heparan sulfate proteoglycan did not affect the interaction of syndecan-2 with MMP-7 (Fig. 6E), suggesting that MMP-7 binds to syndecan-2 through the core protein rather than the glycosaminoglycan chains. To further examine the effect of syndecan-2 on cell surface localization of MMP-7, HT-29-Vector or HT-29-SDC2 were co-immunostained with both syndecan-2-and MMP-7specific antibodies following staining with fluorescein isothiocyanate-and Texas Red-conjugated secondary antibodies. As expected, the levels of syndecan-2 were significantly elevated in HT-29-SDC2 cells, and syndecan-2 was localized on the cell membrane (Fig. 6F). Taken together these data suggest that syndecan-2 directly interacts with MMP-7. To characterize the binding of MMP-7 to syndecan-2, several recombinant syndecan-2 and MMP-7 constructs were made (Fig. 7A), and purified GST-fused proteins of syndecan-2 were incubated with purified recombinant MMP-7 mutant proteins. Western blotting of the precipitated GST⅐SDC2 complexes showed that N-terminal extracellular domain of syndecan-2 (amino acid residues 16 -50) protein bound to pro-domain of MMP-7 (Fig. 7, B and  C). Consistently, when aminophenylmercuric acid was used to convert pro-MMP-7 to the active form, active MMP-7 no lon-

Syndecan-2, a Docking for MMP-7 in Colon Cancer
ger interacted with syndecan-2 (Fig. 7D), suggesting that syndecan-2 interacts specifically with pro-MMP-7. Taken together these data suggest that syndecan-2 functions as a docking receptor for pro-MMP-7 on the cell surface to regulate colon cancer activity.

Syndecan-2 Regulates Activation of Pro-MMP-7-If synde-
can-2 functions as a docking receptor for MMP-7, it might also regulate the activity of MMP-7. To mimic cell surface docking of MMP-7 with syndecan-2, we coated enzyme-linked immunosorbent assay plates with either S2E or S4E (negative con-  Table 1. B, molecular functional categories of the 47 genes are shown in a pie map. C, the relative expression ratios of MMP7 in HT-29 (VEC) and HT-29 (SDC2), respectively, compared with common reference RNAs are shown as bar graphs with standard deviations. Each bar represents a mean expression level Ϯ S.D. from three independent experiments. D, conditioned media were collected from the transfectants, and gelatinase activity was measured using DQ-gelatin as a substrate. The relative activity was normalized by subtracting the background fluorescence from the media. Collagenase was used as a positive control (con). Data are shown as the mean fluorescence Ϯ S.E. from three independent experiments. E, total RNA was extracted from control HT-29 cells or HT-29 cells stably overexpressing syndecan-2 (SDC2), and the expression levels of various MMPs were evaluated with RT-PCR and the indicated primers. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified as a control. The given results are representative from four independent experiments (left panel). Conditioned media from control HT-29 cells or HT-29 cells stably overexpressing syndecan-2 (SDC2) was subjected to immunoblotting with the indicated MMP-specific antibodies (right panel). DECEMBER 18, 2009 • VOLUME 284 • NUMBER 51 trol), and then incubated conditioned media from HT-29 cells in the coated plates in the presence of DQ-gelatin as an exogenous substrate. As shown in Fig. 8A, GST⅐S2E, but not GST⅐S4E enhanced the gelatinase activity of MMP-7 in the HT-29-conditioned media. Consistently, both S2E and SDC2 enhanced the enzymatic activity of pure pro-MMP-7 by Ͼ10-fold in vitro, whereas pro-MMP-2 or pro-MMP-9 had no such effect (Fig.  8B). Interestingly, syndecan-2 had no effect on the catalytic activity of active MMP-7 (Fig. 8C) but did enhance the processing of pro-MMP-7 into the active MMP-7 (Fig. 8D). Consistently, in response to tumor necrosis factor-␣, syndecan-2 expression enhanced activation of MMP-7 and extracellular cleavage of E-cadherin, a known MMP-7 substrate (29) in HT-29 cells (Fig. 8E). These data strongly suggest that syndecan-2 might directly interact with pro-MMP-7 at the plasma membrane to regulate the processing of pro-MMP-7.

MMP-7 Causes Extracellular Shedding of Syndecan-2-Be-
cause MMP-7 is known to cleave the extracellular domain of syndecan-1 (31), we investigated whether MMP-7 regulates the shedding of the syndecan-2 extracellular domain. The soluble form of syndecan-2 was detected in the conditioned media from HT-29 cells (Fig. 9A), suggesting constitutive shedding of syndecan-2 in colon cancer cells. Immunoprecipitation with anti-syndecan-2 antibody confirmed that shed syndecan-2 was in the conditioned media (Fig. 9B). We next investigated whether MMP-7 was involved in the shedding of syndecan-2. After being washed with SFM to remove any accumulated shed syndecan-2 and proteases, HT-29 cells were incubated with SFM-containing pro-MMP-7. As expected, treatment with pro-MMP-7 increased shed syndecan-2 in the conditioned media (Fig. 9C). Taken together, these data suggest that MMP-7 causes extracellular domain shedding of syndecan-2.

TABLE 1 Genes differentially expressed in HT-29 colon cells overexpressing SDC-2 (p < 0.001)
A set of 47 genes shown in the tree view (Fig. 4A) is listed here in the order of relative expression. "GB acc" refers to the GenBank TM accession number. "VEC" and "SDC2" columns refer to average expression ratios of genes from HT-29 transfected with vector only and those from overexpressing syndecan-2, respectively. S.D. VEC and S.D. SDC2 are standard deviations for VEC and SDC2, respectively. "Ratio" in the last column is the ratios of SDC2 over VEC and represents -fold changes in expression levels of genes in HT-29 overexpressing syndecan-2 compared with HT-29 transfected with vector only.

DISCUSSION
Although our previous study showed that increased syndecan-2 expression led to enhanced cell proliferation and migration/inva-sion of cancer cell lines (16), the mechanism by which syndecan-2 regulates colon carcinogenesis is not yet known. Consistent with its effects in colon carcinoma cell lines, increased syndecan-2 expression was also observed in the tested human colon adenocarcinoma tissue samples (Fig. 1). We also observed that syndecan-2 overexpression caused the epithelial morphology of HT-29 cells to change to a more aggregated morphology characterized by a lack of clear discernible cell-cell boundaries (Fig.  2B). This is consistent with our previous report that reduction of syndecan-2 expression in the KM12SM colon carcinoma cell line using trichostatin A (a histone deacetylase inhibitor) or antisense cDNA triggered recovery of clear cell boundaries, flattened morphology, and reappearance of actin filaments in the normally aggregated cell morphology characterized by a lack of clear boundaries between cells and cell layering even at low cell density (30). Thus, syndecan-2 expression appears to promote tumorigenic characteristics in colon cancer cell lines. Correlated with these morphological changes, we herein report that overexpression of syndecan-2 increased the tumorigenic activity of HT-29 cells, as shown by enhancement of cell proliferation, migration/invasion, and anchorageindependent growth (Fig. 2). In addition, transplantation of stably syndecan-2-overexpressing HT-29 cells into nude mice led to enhanced primary tumor formation versus that by control HT-29 cells (Fig. 3). These data collectively support the hypothesis that syndecan-2 plays a critical role in colon tumorigenesis by altering the cell morphology and tumorigenic activity of colon cancer cells.
Genome-scale microarray analysis showed that syndecan-2 overexpression in HT-29 cells significantly increased the expression level of MMP-7, but not that of MMP-2 or MMP-9, which are centrally involved in tumor invasion and metastasis (Fig. 4). Notably, decreased expression of MMP-7 led to reductions in the syndecan-2-mediated enhancement of cell migration (Fig. 5). These  Purified GST, GST⅐SDC2, or GST-syndecan-4 (GST⅐SDC4) was incubated with the conditioned media, and the proteins were bound to glutathione-agarose bead and immunoblotted with the specific MMP antibodies (bottom panel). The membranes were then stripped and reprobed with anti-GST antibody (␣-GST). Arrowheads point to dimer (d) and monomer (m) of syndecan-2 and -4. B, equivalent amount of purified GST, GST⅐SDC2, GST-syndecan-2 extracellular domain (GST⅐S2E), or GST-syndecan-2 mutant lacking the extracellular domain (GST⅐S2TC) was either separated by 12% SDS-PAGE and stained with Coomassie Blue (Coomassie) or incubated with purified MMPs as indicated, bound to glutathione-agarose beads, and immunoblotted with the specific MMP antibody. C, cell lysates form HT-29 cells were immunoprecipitated with anti-syndecan-2 antibody, and the precipitates were incubated with the conditioned media from either HT-29 or HT-29-SDC2 cells, and then used for slot blotting with anti-syndecan-2 and anti-MMP-7 antibodies. The presented data are representative of three independent experiments. D, HT-29 cells were transfected with 4 g of either control (VEC) or FLAG-tagged syndecan-2 (SDC2) expression vectors. Cell lysates were subjected to slot blotting with anti-FLAG antibody (top panel). Anti-FLAG immunoprecipitates from each cell lysates were incubated with proMMP-7, and the proteins bound were immunoblotted with the specific MMP-7 antibodies (bottom panel). E, HT-29 cells were preincubated in the absence (Ϫ) or presence (ϩ) of 25 mM NaClO 4 for 24 h in SFM. The interaction of syndecan-2 and MMP-7 was assessed as described in D. F, cells were co-incubated with anti-syndecan-2 and anti-MMP-7-specific antibodies, stained with a Texas Red-and fluorescein isothiocyanate-conjugated secondary, and then photographed under confocal microscopy. The inset shows a high power view of the box. DECEMBER 18, 2009 • VOLUME 284 • NUMBER 51 findings are consistent with the notion that syndecan-2 plays a crucial role in colon cancer, because MMP-7 expression is a common feature of colorectal cancer development via multiple pathways (32). Furthermore, MMP-7 is activated at an early stage of colon carcinogenesis, and its expression has been correlated with the depth of invasion, lymph node metastasis, lymphatic invasion, and poor outcome. Therefore, it is likely that syndecan-2 enhances the tumorigenic potential of colon adenocarcinoma cells by inducing key molecules such as MMP-7. Indeed, MMP-7 is involved in syndecan-2-induced cell migration in HT-29 cells (Fig. 5) and the shedding of the syndecan-2 extracellular domain (Fig. 9). However, we cannot exclude the possibility that syndecan-2 regulates cancer cell migration in an MMP-7-independent manner, because syndecan-2-mediated cell migration was observed in MMP-7 knockdown HT29 cells (Fig. 5B, compare lanes 3 and 4). In contrast to our finding, a previous study has shown that syndecan-2 suppresses metastasis of Lewis lung carcinoma cells via inhibition of MMP-2 activation (33). The discrepancy may result from using different cancer cell types.

Syndecan-2, a Docking for MMP-7 in Colon Cancer
Previous studies have shown that cell surface heparan sulfate proteoglycans function to dock MMPs at the cell surface (25), and the cell surface localization of MMPs is important for their ability to regulate carcinogenesis (34). Because syndecan-2 is located in the plasma membrane and MMP-7 is known to localize at invasive front in colon cancer cells, it was considered likely that syndecan-2 acts as a cell surface docking receptor for MMP-7. It seemed as though synde-can-2 acts as a pro-MMP-7 docking receptor in at least four possible ways: (i) syndecan-2 could immobilize pro-MMP-7 to restrict its range of action; (ii) syndecan-2 could regulate the enzymatic activity of MMP-7; (iii) syndecan-2 could protect pro-MMP-7 from proteolytic activation; or (iv) syndecan-2 could regulate protease-directed activation of MMP-7. In the present study, we showed that syndecan-2 directly interacts with pro-MMP-7 at the cell surface (Fig. 6), where it regulates the processing of pro-MMP-7 (Fig. 8). These data clearly support the notion that syndecan-2 function as a cell surface docking receptor for pro-MMP-7 in regulating colon adenocarcinoma cells.  ). B, purified GST or GSTsyndecan-2 mutants were incubated with PD or full-length of pro-MMP-7, respectively. Bound materials to glutathione-agarose beads were immunoblotted with the MMP-7-specific antibody (top panel). The membranes were then stripped and reprobed with an anti-GST antibody (bottom panel). C, purified GST, GST⅐SDC2, or GST⅐SDC4 were incubated with PD, CD, or FULL. Bound materials to glutathione-agarose beads were immunoblotted with either pro-domain (pro) or catalytic domain-specific (cat) MMP-7 antibody. D, purified GST⅐SDC2 was incubated with purified pro-MMP-7 (-APMA, P) or active MMP-7 (ϩAPMA, A), bound to glutathione-agarose beads and then immunoblotted with the MMP-7-specific antibody (left panel). The membranes were then stripped and reprobed with an anti-GST antibody (right panel). Arrowheads point to syndecan-2 dimer (d) and monomer (m). FIGURE 8. Syndecan-2 regulates the processing of pro-MMP-7. A, a 96-well plate was coated with the indicated amount of extracellular domain of either syndecan-2 (S2E) or syndecan-4 (S4E) overnight at 4°C. Conditioned media from HT-29 cells was collected, and gelatinase activity was measured using DQ-gelatin as a substrate. The relative activity was normalized versus the background fluorescence from the media. Data are shown as the mean fluorescence Ϯ S.E. from three independent experiments. B, 96-well plate was coated with GST, S2E, or SDC2 (0.5 mg/ml) overnight at 4°C. Purified pro-MMP-7, -2, or -9 was added, and gelatinase activity was measured using DQgelatin. The relative activity was normalized versus the fluorescence of a GSTcoated well, and aminophenylmercuric acid (AMPA) was used to activate pro-MMP-2 and pro-MMP-9 as a positive activity control. C, purified recombinant active-MMP-7 was incubated with purified GST⅐S2E, and the catalytic activity was measured as described in A. The relative activity was normalized versus the background activity from GST. D, purified pro-MMP-7 was incubated with purified S2E or SDC2 as described in B. MMP-7 levels were analyzed by immunoblotting with the MMP-7-specific antibody. E, conditioned media from cells pretreated with 20 ng/ml tumor necrosis factor-␣ were collected and subjected to immunoblotting with the specific MMP-7 (top panel) or E-cadherin antibody (bottom panel). FIGURE 9. MMP-7 causes extracellular shedding of syndecan-2. A, conditioned media from HT-29 cells were collected and subjected to slot blotting with the specific syndecan-2 antibodies. The presented data are representative of five independent experiments. B, conditioned media from HT-29 cells overexpressing syndecan-2 (HT29-SDC-2) were immunoprecipitated with anti-syndecan-2 antibody, and the precipitates were used for slot blotting with anti-syndecan-2 antibody. IgG was used as a control. C, HT-29 cells were treated with 0.2 unit/ml pro-MMP-7 for the indicated time periods. Conditioned medium was collected and subjected to slot blotting with anti-syndecan-2 antibody. The presented data are representative of four independent experiments.

Syndecan-2, a Docking for MMP-7 in Colon Cancer
In summary, we herein show for the first time that syndecan-2 acts as a docking receptor for pro-MMP-7. At the cell surface, syndecan-2 interacts directly with pro-MMP-7 at the plasma membrane, enhancing its processing into active MMP-7, which in turn regulates tumorigenic activities of colon cancer cells. Although future studies will be required to fully elucidate the mechanism underlying syndecan-2-induced MMP-7-mediated signaling in colon cancer cells, our present findings provide important new insights into colon cancer.