Preprogastrin mRNA is abundantly expressed in G cells of the gastric antrum, pancreatic β cells, normal colon and colonic tumor cells.1, 2 The posttranslational processing produces a mature 17-amino acid C-terminally amidated form of gastrin, as well as the immature form including glycine-extended gastrin (G-Gly) and another progastrin-(6–80).3, 4 Amidated gastrin (gastrin) is secreted from the gastric antrum as the main product of this processing3 and stimulates gastric acid secretion, cellular proliferation and differentiation in the stomach, via the gastrin/CCK-B receptor.5 Progastrins are by-products in the antrum.3, 4 On the contrary, in colonic cancers G-Gly and other progastrins are main products because the processing by which progastrin is converted to gastrin is deficient in colon cancers.3, 4 However, it was reported that the amount of G-Gly secreted from colon cancers was less than that from the antrum of a patient suffering from nonulcer dyspepsia.6 G-Gly plays an important role in stimulating the proliferation of colon cancer cells.7, 8, 9 G-Gly also reportedly promotes colon carcinogenesis in rats10 and increases cell/cell dissociation leading to migration of mouse gastric epithelial cells via the G-Gly binding site (putative G-Gly receptor), which is distinct from the gastrin/CCK-B receptor (gastrin receptor).11 Recently, it was reported that G-Gly increased the level of MMP-2 in the human colon cancer cell line LoVo and enhanced invasion through an artificial basement matrix barrier, Matrigel,12 although the precise underlying mechanism was not clarified.
Colorectal cancer remains among the most common malignant tumors of the alimentary tract and often presents with local invasion and metastasis at the time of diagnosis. Furthermore, successful treatment of metastatic colonic tumors has been reported to be extremely difficult.13 Therefore, to develop new target therapies for metastatic colon cancer, elucidation of the precise mechanism underlying invasion and metastasis is critical.
Matrix metalloproteinases (MMPs) are a family of zinc-dependent enzymes, and their production is regulated by several growth factors (HGF, TGF-β), cytokines (IL-1β, TNF-α) and chemical agents (phorbol esters, actin stress fiber-disrupting drugs).14 MMPs play pivotal roles in tumor invasion and metastasis through the proteolysis of several extracellular matrix (ECM) proteins,15 the release and activation of ECM-sequestered growth factor16 and angiogenic factors17 and the digestion of molecules besides ECM components such as adhesion factors.18 Among them, basement membrane-degrading enzymes (MMP-2 and -9) have been given considerable attention for their roles in invasion and metastasis. Stromelysin-1 (MMP-3) is known to activate proMMP-1, -9 and -13, in addition to the degradation of a wide spectrum of substrates including proteoglycans, laminin and fibronectin.19, 20 Conversely, interstitial collagenases MMP-1, -13 and membrane-bound MT1-MMP degrade stromal ECM, consisting of native fibrillar collagen types I, II and III.21 Although the enhancement of collagenases has been indicated in multiple processes necessary for invasion through the stromal ECM, the precise mechanism remains unclear. MMP-1 is strongly expressed in the cytoplasm of colon cancers with liver and lymph node metastasis.22 However, the role of collagenase produced by cancer cells in invasion is unclear.
In our study, we examined the stimulating effect of G-Gly on the invasiveness of 2 colon cancer cell lines, LoVo and HT-29, and focused on the role of collagenases and other matrix metalloproteinases in invasion by LoVo cells. To examine G-Gly-mediated signaling, we used 2 nonselective gastrin receptor antagonists, proglumide and benzotript,23, 24 and the gastrin/CCK-B receptor antagonist YM02225 in our study.
LoVo and WiDr cells, derived from human colon adenocarcinoma, were obtained from the Health Service Research Resources Bank (Osaka, Japan), and HT-29 cells, derived from human colon adenocarcinoma, were from American Type Culture Collection (Manassas, VA). LoVo cells from passages 33–37 were grown in Ham's F-12 medium (Invitrogen, Carlsbad, CA) supplemented with 20% fetal bovine serum (Sigma, St. Louis, MO) and HT-29 cells from passages 129–133 in McCoy's 5A medium (Invitrogen) supplemented with 10% fetal bovine serum (Sigma) and WiDr cells from passages 3–6 in Dulbecco's modified Eagle medium (Invitrogen) supplemented with 10% fetal bovine serun (Sigma) in a humidified atmosphere containing 5% CO2 at 37°C. Media were changed 3 times weekly.
The Matrigel invasion assay was performed using a Matrigel Invasion Chamber consisting of a 24-well plate and cell culture inserts with an 8 μm pore size polyethylene terephthalate membrane coated with a reconstituted basement membrane matrix (Becton Dickinson, Bedford, MA). The type I collagen gel invasion assay was carried out using a 24-well Transwell chamber (Coster, Cambridge, MA) containing a polyvinylpyrrolidone-free polycarbonate filter (pore size 8.0 μm) coated with 165 μg of rat type I collagen (Becton Dickinson). A total of 1 × 105 cells suspended in serum-free medium with 0.1% BSA were seeded in the upper compartment, and conditioned medium (cultured in serum-free medium with 0.1% BSA for 24 hr) of LoVo or HT-29 cells was added to the lower chamber. G-Gly (Neosystem, Strasbourg, France), gastrin (Peptide Institute, Osaka, Japan), proglumide (Sigma), benzotript (kindly provided by Nihon Kaken, Tokyo, Japan), YM022 (kindly provided by Yamanouchi Pharmaceutical, Tokyo, Japan) or CGS27023A (kindly provided by Novartis Pharma, Basel, Switzerland) were added to both the upper compartment and the lower chamber at final concentrations of 10−7 M, 10−7 M, 3 × 10−3 M, 10−4 M, 10−9 M and 2 × 10−5 M, respectively. After incubation for 6 or 24 hr at 37°C, nonmigrating cells on the upper surfaces of membranes were removed with a cotton swab, and the invading cells on the lower surface were fixed with 4% paraformaldehyde followed by staining with Diff-Quick (International Reagent, Kobe, Japan). The number of cells were counted at high-power magnification (×200) in 13 different areas on each filter membrane, and the mean value for each assay was obtained from 3 to 4 wells. Two independent sets of experiments were undertaken. Also, to examine the effective dose of G-Gly, a type I collagen gel invasion assay was carried out in the presence of 10−10–10−6 M G-Gly. The purity of G-Gly was determined to be 93% by HPLC, and the absence of known gastrin-related compounds was confirmed. The purity of gastrin was determined to be 99% by HPLC.
Gelatin and casein zymography
Gelatin and casein zymography were carried out following the procedure reported by Koshikawa et al.26 Briefly, LoVo cells were cultured in Ham's F-12 medium supplemented with 20% FBS until they reached 70–80% confluence. After the cells have been starved by incubation in serum-free medium (containing 0.1% BSA) for 24 hr, they were stimulated by G-Gly with or without 30 min pre-incubation using 3 × 10−3 M proglumide or 10−4 M benzotript. Amounts of 2 × 10−5 M CGS27023A, 20 μg/ml anti-MMP-3 IgG (Daiich Fine Chemical, Tokyo, Japan) and 2 × 10−8 M TIMP-1 (Daiichi Fine Chemical) were added to the medium with G-Gly at the same time. After a 24 hr incubation, supernatants were collected, centrifuged and concentrated 7-fold for gelatinase and 15-fold for collagenase and stromelysin, using a membrane dialysis concentrator (Centricon YM-10, Millipore, Bedford, MA). Concentrated supernatants (30 μl) were mixed with 10 μl of sample buffer (250 mM Tris-HCl, pH 6.8, 8% SDS, 40% glycerol, 0.0096% BPB). Then the protein was separated on 9% SDS-PAGE copolymerized with 0.1% gelatin or casein. The gels were washed twice in 2.5% Triton X-100 for 30 min and 10 mM Tris-HCl (pH 8.0) for 30 min at room temperature after electrophoresis. The gels were then incubated for 24 hr at 37°C in the reaction buffer containing 50 mM Tris-HCl (pH 8.0), 0.5 mM CaCl2 and 1 μM ZnCl2. Gels were stained with PAGE Blue 83 (Daiich Pure Chemicals, Tokyo, Japan) for 30 min and destained with 5% acetic acid in 10% methanol. Clear zones of protein lysis (MMP-1, -2, -3 and -9) were scanned using a flat scanner GT-8200UF (Epson, Tokyo, Japan) and quantitatively analyzed with NIH Image software (version 1.61).
To determine whether proglumide and benzotript were MMP inhibitors, LoVo cells were starved by incubation in serum-free medium (containing 0.1% BSA) for 24 hr. The concentrated conditioned medium was applied to 8 lanes on 9% SDS-PAGE copolymerized with 0.1% gelatin or casein. After electrophoresis, the gel was washed in 2.5% Triton X-100 and 10 mM Tris-HCl (pH 8.0). Next, the gel was cut into 4 slips containing 2 lanes each and incubated for 24 hr at 37°C in reaction buffer containing 2 × 10−5 M CGS27023A, 3 × 10−3 M proglumide or 10−4 M benzotript and in nontreated reaction buffer. Then each slip of gel was stained with PAGE Blue 83.
Immunoprecipitation and western blotting
Cells were grown until they reached 70–80% confluence. After the cells had been starved by incubation in serum-free medium with 0.1% BSA for 24 hr, they were incubated in the presence of 10−7 M G-Gly. After a 24 hr incubation, the cells were washed with PBS and lysed in cold lysis buffer containing 10 mM Tris-HCl (pH 7.5), 1% NP-40, 0.1% SDS, 0.1% sodium deoxycholate, 150 mM NaCl, 1 mM PMSF and 10 μg/ml aprotinin. For preclarification, the obtained cell lysates containing 1–2 mg soluble protein and the conditioned medium containing 7 mg soluble protein were incubated with 30 μl of protein G-agarose (Immobilized Protein G, Pierce, Rockford, IL) for 2 hr. After centrifugation (10,000g, 2 min), the supernatants were incubated with 2 μg of anti-MMP-1 mouse monoclonal antibody (IgG1) (Oncogene, Boston, MA) or anti-MMP-3 mouse monoclonal antibody (IgG1) (Daiichi Fine Chemical) overnight at 4°C. Then 30 μl of protein G-agarose were added, followed by incubation with a rotator for 2 hr at 4°C. The immobilized protein G-bound anti-MMP antibody-MMP complexes were washed 5 times with 0.5 ml of the immunoprecipitation buffer. The bound anti-MMP antibody-MMP complexes were eluted from the gel by incubation with 25 μl of SDS-PAGE sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 2% β-mercaptoethanol) for 5 min at 95°C. The samples were separated on SDS-9% polyacrylamide gel, and proteins were transferred onto a polyvinylidene fluoride (PVDF) microporous membrane (Immobilon-P, Millipore) using a semidry blotting system (Nihon Eido, Tokyo, Japan). The membranes were then incubated with the anti-MMP-1 antibody (1:1,000) or anti-MMP-3 antibody (1:2,000), followed by incubation with HRP-labelled anti-mouse IgG. MMPs were visualized using an ECL Western Blotting Analysis System (Amersham Pharmacia Biotech, Little Chalfont, UK) and measured using a flat scanner and NIH Image software.
Transfection of MMP-3 siRNA
MMP-3 siRNA and inverted MMP-3 siRNA were purchased from Proligo Japan (Kyoto, Japan), including MMP-3 siRNA sense (5′-AUGAAGAGUCUUCCAAUCCUU-3′); MMP-3 siRNA antisense (5′-GGAUUGGAAGACUCUUCAUUU-3′); inverted MMP-3 siRNA sense (5′-CCUAACCUUCUGAGAAGUAUU-3′); and inverted MMP-3 siRNA antisense (5′-UACUUCUCAGAAGGUUAGGUU-3′).
LoVo cells were cultured in 0.5 ml of F-12 medium supplemented with 20% FBS without antibiotics in 24-well plates until they reached 50% confluence. Thirty pmoles of MMP-3 siRNA and inverted MMP-3 siRNA duplexes were diluted in 50 μl of Opti-MEM I reduced serum medium (Invitrogen), respectively. A 1.5 μl quantity of Lipofectamine 2000 (Invitrogen) was diluted in 50 μl of Opti-MEM I reduced serum medium and incubated for 5 min at room temperature. After diluted siRNA and lipofectamine 2000 had been combined and incubated for 20 min at room temperature, the siRNA and Lipofectamine 2000 complexes were added, and the cells were incubated at 37°C in a CO2 incubator for 48 hr. Then the cells were starved by incubation in serum-free F-12 medium (containing 0.1% BSA) for 24 hr with or without G-Gly. MMP-3 protein was estimated by immunoprecipitation and Western blotting using cell lysates, and the activity of MMP-9 was determined by gelatin zymography using concentrated conditioned medium.
MMP-1-EGFP expression vector construction and transfection
Full-length MMP-1 cDNA with the Hind III site and Kozak sequence added at the 5′ end and the Sma I site added at the 3′ end was generated from total RNA extracted from human melanoma cell line A2058 by RT-PCR using a forward primer, 5′-AAGCTTGCCACCATGCACAGCTTTCCTCCACT-3′ and a reverse primer, 5′-CCCGGGCATTTTTCCTGCAGTTGAACC-3′. The PCR product was subcloned into a pUC19 vector, and the sequence was verified using an ABI PRISM 310 DNA Analyzer (Applied Biosystems, Foster City, CA). To construct fusions of the enhanced green fluorescent protein (EGFP) and MMP-1, the cDNA fragment was digested with Hind III and Sma I from a pUC19-MMP-1 vector and inserted into a pEGFP-N1 vector (BD Biosciences Clontech, Palo Alto, CA). Transfection was carried out in LoVo cells using Lipofectamine (Invitrogen) and Plus Reagent (Invitrogen) according to the manufacturer' s protocol.
Cells were seeded on 12 mm round glass coverslips coated with poly-L-lysine (Becton Dickinson). To examine colocalization of MMP-1 and CD147, transient-MMP-1-EGFP-transfected cells were fixed with 3% paraformaldehyde for 10 min. After permeabilization with 0.1 M Tris-HCl buffer (pH 7.4) and blocking with 2% BSA containing 3% normal goat serum for 1 hr at room temperature, the cells were incubated with anti-CD147 mouse monoclonal antibody (1:100) (Ancell, Bayport, MN) for 3 hr and Texas-Red-conjugated goat anti-mouse IgG (Molecular Probes, Eugene, OR) for 30 min at room temperature. To examine the coat proteins of vesicles containing MMP-1, after fixation, permeabilization and blocking, transient-MMP-1-EGFP transfected cells were incubated with anti-GGA-2 (one each of Golgi-localized, γ-adaptin ear-containing, ARF-binding proteins) goat polyclonal antibody (1:100) (Santa Cruz, Santa Cruz, CA) for 2 hr, followed by incubation with Alexa Fluor 594-conjugated chicken anti-goat IgG (Molecular Probes) for 30 min at room temperature. Wild-type cells were fixed with 3% paraformaldehyde for 10 min, permeabilized with 0.1 M Tris-HCl buffer (pH 7.4) for 15 min for MMP-1 and with 0.01% Triton X-100 for 5 min for MMP-3 and incubated with blocking solution for 1 hr at room temperature. The cells were then incubated with anti-MMP-1 rabbit polyclonal antibody (1:100) (Chemicon, Temecula, CA) for 24 hr at 4°C or with anti-MMP-3 rabbit polyclonal antibody (1:100) (Sigma) for 3 hr at room temperature, respectively, followed by incubation with Alexa Fluor 488-conjugated goat anti-rabbit IgG (Molecular Probes) for 30 min. The localizations of MMP-1-EGFP fusion protein, CD147, GGA-2 and MMP-1 and -3 were observed using a confocal laser microscope, LaserSharp2000 Version 4.0 (Bio-Rad, Richmond, CA).
To examine the effect of G-Gly on proliferation of LoVo cells, cells (1.5 × 104/well) were seeded in a 96-well microplate. After 24 hr incubation, the cells were incubated for another 24 hr in F-12 medium (0.1% BSA) containing 10−12–10−6 M G-Gly. To confirm the effect of YM022 on the increased proliferation of WiDr cells expressing the gastrin/CCK-B receptor25 in the presence of gastrin, cells (1 × 104/well) were seeded in a 96-well microplate. After 24 hr incubation, the cells were incubated for another 24 hr in serum-free Dulbecco's modified Eagle medium (0.1% BSA) containing 10−8 or 10−7 M gastrin in the presence of 10−10 or 10−9 M YM022. Then MTT(3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) (Molecular Probes) was added to the medium at a final concentration of 0.5 mg/ml, followed by incubation in 5% CO2 for 4 hr at 37°C. After dissolution of MTT formazan with 0.04 N HCl in isopropanol, absorbance was measured at 570 nm using a microplate reader.
Student's t-test was used to evaluate the differences between 2 groups. p < 0.05 was considered to be statistically significant.
Invasion of colon cancer cells through type I collagen gel and Matrigel
We first examined the effect of G-Gly on the invasion of LoVo and HT-29 cells through type I collagen gel and Matrigel. As shown Figure 1, 10−7 M G-Gly significantly increased the number of LoVo and HT-29 cells invading through both gels. The effect of G-Gly was reversed by treating the cells with a broad-spectrum MMP inhibitor (CGS27023A),27 which competitively inhibited the binding of Zn2+ to the active sites of MMP-1, -2, -3, -9 and -13 (top and bottom panels, column 3, Fig. 1). Nonselective gastrin/CCK receptor antagonists (proglumide and benzotript) also attenuated this stimulated invasiveness (top and bottom panels, columns 4 and 5, Fig. 1). However, the effect of G-Gly was not inhibited by treating the cells with a gastrin/CCK-B receptor antagonist (10−9 M YM022) (top and bottom panels, column 6, Fig. 1). An amount of 10−10 or 10−9 M of YM022 inhibited the increased proliferation of WiDr cells expressing the gastrin/CCK-B receptor in the presence of 10−8 or 10−7 M gastrin. CGS27023A and benzotript slightly inhibited the basal invasiveness of LoVo cells through type I collagen gel in the absence of G-Gly (bottom panel, columns 7 and 9, Fig. 1).
In gelatin and casein zymography, MMPs (pro and active forms) that separated on SDS-polyacrylamide gel containing gelatin or casein were activated by washing out SDS with Triton X-100 and incubating gel with reaction buffer containing CaCl2 and ZnCl2. The MMP inhibitor CGS27023A inhibited the activation of proMMP-2 and -3 from LoVo lysates on gel, but 3 × 10−3 M proglumide and 10−4 M benzotript did not (Fig. 2). These results show that proglumide and benzotript are not MMP inhibitors.
Taken together, these results suggest that G-Gly promotes invasion of LoVo and HT-29 cells through type I collagen gel and Matrigel via a putative G-Gly receptor-mediated pathway.
Dose-dependence effect of G-Gly on invasiveness and proliferative activity of LoVo cells
An amount of 10−9–10−6 M G-Gly significantly increased the number of LoVo cells invading type I collagen gel, and the maximum stimulating effect, 255% compared to controls, was found at 10−8 M G-Gly for a 6 hr incubation (Fig. 3a). By contrast, 10−10–10−8 M G-Gly significantly but modestly increased the proliferation of LoVo cells, and the maximum effect, 118% compared to control cells, was obtained at 10−9 M (Fig. 3b). These results indicate that the optimal concentration of G-Gly for invasion of colon cancer cells through stromal extracellular matrix differs from that for cellular proliferation and suggest that G-Gly binds to G-Gly binding sites with 2 different affinity states or that fractional occupancy only is necessary to activate “proliferation” as a biologic effect.
Change of MMP-1 in LoVo cells by G-Gly
Next, we examined the effect of G-Gly on the level of MMP-1 expression. To detect MMP-1, immunoprecipitation followed by Western blotting using the cell lysate from 5 × 106 cells was performed. We found that 10−7 M G-Gly increased (2-fold) the level of proMMP-1 (53 kDa) in cell lysates compared to control cells (bottom, Fig. 4a). Casein zymography using 15-fold concentrated conditioned medium from LoVo cells demonstrated that G-Gly also significantly (1.8-fold) increased MMP-1 enzyme activity (Fig. 4b). We further examined the localization of MMP-1 using immunofluorescence microscopy. MMP-1 localized mainly in cytoplasmic vesicles, and G-Gly significantly increased the number of vesicles containing MMP-1 (Fig. 4c). These results indicate that G-Gly enhances the production of MMP-1.
Invasion of MMP-1-EGFP-transfected cells through type I collagen
To confirm the role of MMP-1 in LoVo cell invasion, an invasion assay with type I collagen gel was carried out employing MMP-1-EGFP-transfected cells. Transient transfection efficiency was approximately 20% in both MMP-1-EGFP and the mock transfectant. The number of cells expressing EGFP on the lower surfaces of Transwell insert filters was higher (2-fold) than that in the mock transfectant (Fig. 5a,b), suggesting that overexpression of MMP-1 enhances the invasion of LoVo cells through type I collagen gel.
Colocalization of MMP-1 and CD147 in MMP-1-EGFP-transfected cells
To further examine the role of MMP-1 in LoVo cells, we observed the localization of MMP-1 and CD147 in MMP-1-EGFP-transfected LoVo cells. CD147 was expressed in both the cell membrane and vesicles containing MMP-1 in transfected cells (Fig. 6c), and MMP-1 colocalized with CD147 in the vesicles (Fig. 6d). Since CD147 is reportedly an inducer of MMP-1, -2, and -3 in stromal cells adjacent to cancer cells,28 it suggests that MMP-1 interacts with CD147 to increase invasiveness. EGFP-fused MMP-1 was mainly localized in cytoplasmic vesicles, and the vesicles containing MMP-1 were found in perinuclear, submembranous regions and also outside of cells (Fig. 6b and 7b). To confirm whether the vesicles outside of cells were artifacts, they were stained with anti-GGA-2. GGA-2 is reportedly a ubiquitous coat protein.29 GGA-2 was found to be localized in MMP-1 vesicles on surface and outside of the cell, indicating that the extracellular vesicles had been released from cells (Fig. 7d). These results suggest that extracellular vesicles containing MMP-1 and CD147 contribute to stromal invasiveness by LoVo cells.
Changes of MMP-2 and -9 by G-Gly
Gelatin zymography revealed the proMMP-2 level to be significantly increased by 10−7 M G-Gly in the conditioned medium of LoVo cells but no intermediate or active forms of MMP-2 in either the presence or the absence of G-Gly (top panel, Fig. 8a). Because LoVo cells have been reported to be deficient in furin,30 no MMP-2 activation was observed even after treatment of the cells with 10 μg/ml concanavalin A, which promotes the production of MT1-MMP following cleavage of proMMP-2 into the active form in fibrosarcoma HT1080 cells (bottom panel, Fig. 8a). By contrast, 10−7 M G-Gly significantly increased MMP-9 activity (top panel, Fig. 8b). CGS27023A, proglumide, benzotript, anti-MMP-3 IgG and TIMP-1 inhibited this increase in the active form of MMP-9 (76 kDa) induced by G-Gly (bottom panel, lanes 3–7, Fig. 8b). Also, to confirm that MMP-3 enhanced activation of proMMP-9 in LoVo cells, the effect of MMP-3 siRNA on the MMP-9 increase by G-Gly was examined. MMP-3 siRNA decreased MMP-3 levels by 50% compared to inverted MMP-3 siRNA (left panel, Fig. 8c). At the same time, the MMP-9 increase by G-Gly was inhibited in the conditioned medium (right panel, Fig. 8c).
Taken together, these results indicate that MMP-3 is associated with the MMP-9 activity increase by G-Gly, which results in the enhancement of MMP-9-mediated invasion through Matrigel.
Collagenase, MT1-MMP, activates proMMP-2. However, Western blotting showed the MT1-MMP expression level to be low and unaffected by G-Gly in LoVo cells (data not shown). In addition, furin, an activator of proMT1-MMP, was deficient in LoVo cells.30 Thus, MT1-MMP appeared to be minimally involved in LoVo cell invasion through type I collagen gel and Matrigel.
Change of MMP-3 in LoVo cells by G-Gly
Because MMP-9 activity was significantly increased in the presence of G-Gly and reduced in the presence of anti-MMP-3 IgG and MMP-3 siRNA (Fig. 8b,c), we next examined the effects of G-Gly on the expression of MMP-3. Immunoprecipitation followed by Western blotting using cell lysates from 3 × 106 cells and the conditioned medium from 8 × 105 cells revealed that 10−7 M G-Gly significantly increased the proMMP-3 level (57 kDa) (Fig. 9a). Casein zymography showed MMP-3 in conditioned medium to be significantly increased in the presence of 10−7 M G-Gly and that this increase was inhibited by treating the cells with 3 × 10−3 M proglumide or 10−4 M benzotript (bottom, lanes 3 and 4, Fig. 9b). Immunofluorescence revealed MMP-3 to apparently be localized in cytoplasmic vesicles corresponding to the endoplasmic reticulum and Golgi apparatus. G-Gly increased the number of MMP-3-positive vesicles (Fig. 9c), suggesting that G-Gly stimulates the production of MMP-3 protein in these cells.
Effect of gastrin on invasion and MMP expression
Finally, we checked the effect of gastrin on the invasiveness and MMP activity of LoVo cells. An amount of 10−7 M gastrin did not significantly increase invasion of these cells through either type I collagen gel or Matrigel (Fig. 10a). The expressions of MMP-1, -2, -3 and -9 in the cells were unaffected by the presence of 10−7 M gastrin (Fig. 10b).
In our study, we demonstrated that G-Gly enhances the invasiveness of 2 colon cancer cell lines, LoVo and HT-29, through type I collagen gel and Matrigel via the putative G-Gly receptor. We also found that G-Gly increased the levels of collagenase-1 (MMP-1) and stromelysin-1 (MMP-3) expression and the activity of gelatinase (MMP-9) in LoVo cells. A broad-spectrum MMP inhibitor (CGS27023A) reversed the G-Gly-mediated invasiveness of LoVo and HT-29 cells, suggesting that G-Gly promotes MMP-mediated invasion. Because the expression of MMPs was tightly regulated at the gene transcription level,31 G-Gly appeared to mediate the production of both MMP-1 and -3.
G-Gly markedly enhanced the invasiveness of colon cancer cells through type I collagen gel. Collagenase MMP-1 was expressed in LoVo cells, and G-Gly increased the MMP-1 level. The expression of other collagenases, MMP-13 and MT1-MMP, was marginal, and their levels were unaffected by G-Gly (data not shown). Thus, MMP-1 was mainly involved in the enhancement of invasion through type I collagen gel by LoVo cells. Consequently, the invasion of MMP-1-EGFP transiently transfected LoVo cells through type I collagen gel was significantly increased compared to that of controls. To further clarify the role of MMP-1 in cancer cell invasion, the transfectants were examined with immunofluorescence microscopy. We found MMP-1-EGFP to be localized in vesicles that were perinuclear, submembranous or extracellular. Immunofluorescence study also showed that GGA-2 localized in the vesicles containing MMP-1 and suggested vesicle-associated MMP-1 to be released from cells because GGA-2 is among the ubiquitous ARF-dependent coat proteins29, 32 and does not associate with clathrin-coated vesicles.29 It appeared that extracellular MMP-1 degraded the type I collagen gel, followed by the enhancement of invasions in MMP-1-EGFP transfectant. Further, MMP-1 was also colocalized with CD147 (EMMPRIN). CD147 enrichment on the tumor cell surface has been reported to induce the production of MMP-1, -2, and -3 in stromal cells adjacent to cancer cells and to enhance the invasiveness of cancer cells.28 It was suggested that in the presence of stromal cells, CD147 containing vesicles outside of cancer cells are involved in stromal invasion. However, Guo et al.33 found that MMP-1 formed a complex with CD147 on the lung cancer cell surface and suggested that this complex modified the tumor cell pericellular matrix to promote invasion. MMP-2, -9, -13 and TGF-β were demonstrated in matrix vesicles produced in growth-plate chondrocytes, and these MMPs were suggested to activate growth factor during chondrocyte hypertrophy.34 These findings suggested that MMP-1 and CD147 interact to increase invasion, although the precise biologic crosstalk between MMP-1 and CD147 was unclear and awaits future examination. Recently, it was reported that progastrin-(6–80) regulates adherens and tight junctions via separate pathways in the colon cancer cell line DLD-1.35 The relationship between MMP-1 and cell-cell adhesion should also be examined.
G-Gly increased the number of vesicles containing MMP-3 in the LoVo cell cytoplasm. These MMP-3 containing vesicles appeared to be located on the endoplasmic reticulum and Golgi apparatus, suggesting that G-Gly increases the production of MMP-3. MMP-3 is involved in mediating the activations of MMP-1 and -9,19, 20 and the increase in the active form of MMP-9 via G-Gly was inhibited by anti-MMP-3 IgG and MMP-3 siRNA in LoVo cells. Thus, MMP-3 apparently activates proMMP-9 in LoVo cells. On the other hand, high levels of MMP-1 and -3 expression have been shown to correlate with human metastatic melanoma.36 The domain structures of MMP-1 and -3 are closely related,15 and their promoter regions contain TATA box, AP-1 and Ets-1 binding sites.37, 38 Furthermore, the expressions of MMP-1 and -3 are enhanced via activation of p38 mitogen-activated protein kinase in the presence of TNF-α.39 Taken together, these observations suggest the expressions of MMP-1 and -3 to be increased by G-Gly via similar pathways, resulting in the enhancement of LoVo cell invasion through the stromal extracellular matrix.
In terms of gelatinase activity, G-Gly increased the level of proMMP-2, but no active form was found on gelatin zymograms of the conditioned medium of cells treated with or without G-Gly. ProMMP-2 was mainly processed to MMP-2 by MT1-MMP, which was activated via a serine proteinase (furin).40, 41 Since LoVo cells are furin-deficient,30 we speculated that no activation of MMP-2 was found in LoVo cells stimulated with G-Gly. Therefore, it was mainly MMP-9 that contributed to the invasion of Matrigel by the cells. In gastrin/CCK-B-transfected gastric epithelial cells, gastrin was shown to be involved in MMP-9-mediated invasion through Matrigel.42
The effective dose of G-Gly for enhancing invasion through type I collagen gel was more than 10−9 M. Thus, the local G-Gly concentration or tumoral G-Gly content should be more than 10−9 M and 10−12 mole/g, respectively. Plasma G-Gly concentrations have been reported to be less than 2 × 10−11 M.43 However, the mean G-Gly content in colon cancers has been reported to be approximately 5 × 10−12 mole/g.44 Therefore, we speculate that G-Gly is a possible metastatic promoter in patients suffering from colon cancer. In addition, because the expression of preprogastrin mRNA in colon cancers was lower than gastric antrum2 and the amount of G-Gly secreted from the gastric antrum was high,6 it appears that antral G-Gly contributes to increasing invasion of colon cancer via the circulation.
The G-Gly receptor has not yet been cloned. However, binding assay raised the possibility of a G-Gly-specific binding site on the membranes of colon cancer cells.8, 9 It was reported that G-Gly binding sites were independent from gastrin/CCK-A and -B receptors8, 9, 45 but similar to gastrin/CCK-C receptor.45, 46 The binding of G-Gly in G-Gly binding sites was inhibited in the presence of nonselective gastrin/CCK receptor antagonists but not the gastrin/CCK-A and -B-receptor antagonists in colorectal cancer cells and normal colonic mucosa.45, 46 Also, the increased proliferation by G-Gly was not inhibited in the presence of the gastrin/CCK-B-receptor antagonist.8, 9 Therefore, G-Gly appeared to act via a G-Gly-specific binding site (putative G-Gly receptor). Our results also show 10−10–10−8 M G-Gly to modestly increase the proliferation of LoVo cells, but this effect was no longer observed at concentrations higher than 10−7 M G-Gly. This reversal of proliferation in the presence of G-Gly in LoVo cells has been confirmed by Stepan et al.9 Because proliferation did not fall below 100% in LoVo cells and invasion at a G-Gly concentration of 10−6 M was 245% compared to the controls, it is reasonable to assume the proliferative effect to be independent of G-Gly cytotoxicity.
In summary, we have demonstrated that G-Gly promotes MMP-mediated invasion of colon cancer cells. This raises the possibility of tumor metastasis developing via increasing levels of MMP-1 and -3. Measurement of G-Gly may be worthwhile in making a molecular diagnosis of invasion and metastasis of colon cancer, and G-Gly is a potential new molecular target for the clinical treatment of colon cancer.
We thank Dr. N. Koshikawa (Department of Cancer Cell Research, Institute of Medical Science, University of Tokyo) for his helpful advice and technical support in the gelatin zymography and Dr. M. Inoue (Department of Tumor Biochemistry, Research Institute, Osaka Medical Center for Cancer and Cardiovascular Diseases) for his advice in generating confocal microscopic images.