Invasiveness and MMP Expression in Pancreatic Carcinoma

Journal of Surgical Research 98, 33–39 (2001) doi:10.1006/jsre.2001.6150, available online at http://www.idealibrary.com on

Invasiveness and MMP Expression in Pancreatic Carcinoma Xinqun Yang, MSBS,* Edgar D. Staren, M.D., Ph.D.,* ,1 John M. Howard, M.D.,* Takeshi Iwamura, M.D.,† John E. Bartsch, MSBS,* and Hubert E. Appert, Ph.D.* *Department of Surgery, Medical College of Ohio, Toledo, Ohio 43614-5807; and †First Department of Surgery, Miyazaki Medical College, Miyazaki, Japan Presented at the Annual Meeting of the Association for Academic Surgery, Tampa, Florida, November 2– 4, 2000

Key Words: matrix metalloproteinase; pancreatic carcinoma; invasion.

Background. Previous investigations have suggested that expression of matrix metalloproteinases (MMPs) may be related to increased invasiveness of various tumors. This study evaluated a possible relation between pancreatic tumor cell invasiveness and MMPs. Methods. A Matrigel invasion assay was performed with pancreatic tumor cell line SUIT-2 and its sublines S2-007, S2-013, S2-020, and S2-028. The degree of invasiveness of stimulated and unstimulated cell lines was correlated with MMP gene expression measured by RT-PCR and MMP protein product measured by gelatin zymography. Cell lines were stimulated by 12-Otetradecanoylphorbol-13-acetate (TPA), concanavalin (Con-A), and polymerized collagen type I gel (Vitrogen). Results. For SUIT-2, S2-007, S2-013, S2-020, and S2028, 3.2, 1.0, 4.1, 6.4, and 0.4%, respectively, of the cells invaded the Matrigel membrane. TPA, Con-A, and Vitrogen resulted in the up-regulation of MMP-2 in S2-020. TPA and Vitrogen resulted in up-regulation of MMP-9 in each of the cell lines, while Con-A could up-regulate MMP-9 expression only in SUIT-2. There was no constitutive expression of either MMP-2 or MMP-9 in SUIT-2 or its sublines. There was a positive relationship between Matrigel invasiveness and upregulation of MMP-2 and MMP-9 expression. Conclusions. These data suggest that, while MMP-2 and MMP-9 are not constitutively expressed in pancreatic carcinoma cell lines, they may be up-regulated by TPA, Con-A, and Vitrogen. Since MMP-2 and MMP-9 expression correlated with degree of tumor cell invasiveness, the ability to up-regulate MMP-2 and MMP-9 expression may play a role in facilitating pancreatic tumor cell invasion. © 2001 Academic Press

INTRODUCTION

Pancreatic cancer is the fourth and fifth most common cause of cancer-related death in men and women, respectively, in developed countries and the incidence seems to be increasing [1]. Local invasion of adjacent structures, perineural invasion, early metastasis to lymph nodes and liver, and intense desmoplastic stromal reaction are characteristics of pancreatic cancer. Pancreatic cancer patients have one of the worst clinical prognoses of any form of cancer, with a 5-year survival rate of less than 0.4% and a median survival time of 4 – 6 months following diagnosis. This has been explained by difficulty of early detection of the neoplastic process, lack of effective treatment, and limited knowledge of the peculiarity in the biological nature of the cancer [2]. Invasion and metastasis are hallmarks of malignant tumors, and both involve multistep processes. Before they can invade tissues and metastasize, tumor cells must cross the extracellular matrix (ECM) that consists of combinations of proteinaceous fibers and carbohydrate macromolecules. Several classes of proteases may be involved in the proteolytic events which occur during tumor invasion [3]; these include serine proteinases such as urokinase-type plasminogen activator (u-PA), cysteine proteinases such as Cathepsin D, and matrix metalloproteinases (MMPs). While all of these enzymes may participate in the process of matrix degradation, the degradation of basement membrane type IV collagen is a critical early event in tumor invasion, suggesting that MMPs, especially MMP-2 and MMP-9, may be particularly important [4]. In addition to collagen, MMP-2 and MMP-9 are also able to de-

1

To whom correspondence should be addressed at Department of Surgery, Medical College of Ohio, 3065 Arlington Avenue, Toledo, OH 43614-5807. Fax: (419) 383-6636. E-mail: [email protected].

33

0022-4804/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

34

JOURNAL OF SURGICAL RESEARCH: VOL. 98, NO. 1, JUNE 1, 2001

grade fibronectin, laminin, elastin, and proteoglycan components of the ECM [5]. Some genes encoding MMPs are not constitutively expressed, but a variety of agents such as tumor promoters (TPA, Con-A), growth factors (EGF), cytokines (IL-1), and oncogene products (ras) could induce the expression of these genes [6]. Furthermore, MMPs may be secreted from cancer cells and/or adjacent stromal cells as latent forms which then may be converted into active forms under regulation of membrane-type matrix metalloproteinases (MT-MMPs) and tissue inhibitors of metalloproteinases (TIMPs). SUIT-2 is a pancreatic tumor cell line from a liver metastasis of a human pancreatic cancer [7]. Four of its clonal sublines, S2-007, S2-013, S2-020, and S2-028, have distinct morphological and immunohistochemical characteristics as well as different metastatic potentials. The SUIT-2 cell line and these four clonal variants were selected for this study because the clonal variants would be expected to possess fewer genotypic differences unrelated to their metastatic properties than non-clonally related cell lines. In this study, the relationship between the invasive ability of these cells into the extracellular matrix and the expression of matrix metalloproteinases was investigated. MATERIALS AND METHODS Cells and cell culture. The SUIT-2 cell line and its four sublines, S2-007, S2-013, S2-020, and S2-028, were kindly provided by Dr. Takeshi Iwamura (Miyazaki, Japan). The cells were maintained in McCoy’s medium supplemented with 10% fetal bovine serum, penicillin (62500 U/L), streptomycin (62.5 mg/L), and gentamicin (65 mg/L) at 37°C in humidified air containing 5% CO 2. McCoy’s 5A, fetal bovine serum (FBS), penicillin-streptomycin, gentamicin, sodium pyruvate, MEM amino acid (50X), MEM amino acid (100X), MEM vitamins, L-glutamine, serine, asparagine, and sodium bicarbonate were purchased from Sigma. Trypsin-EDTA (1X) was purchased from Cellgro. Tissue culture flasks and pipettes were purchased from Corning. Matrigel invasion assay. This assay is similar to that described previously [8, 9] with the exception that Costar Transwell polycarbonate membrane with 12 ␮m pore size was used instead of Boyden chambers in the assay system. DMEM, 200 ␮l, with 50 ␮g Matrigel was dried onto a Transwell polycarbonate membrane for 24 h. The Matrigel membrane was reconstituted with 0.5 ml McCoy’s medium in the incubator for 2 h and the medium was removed. NIH 3T3conditioned medium was added to the bottom of the well as the chemoattractant. Cells (10,000 cells/well) suspended in McCoy’s medium were then added to the top of the well. Wells were incubated at 37°C in humidified air containing 5% CO 2 for 24 h. The cells on the upper side of the polycarbonate membrane were wiped off and the remaining cells that traversed the Matrigel and spread on the lower surface of the filter were stained with hematoxylin 2 and counted using a light microscope. The relative invasive rate was calculated as the percentage of total cells applied to the top of the membrane that migrated to the bottom side of the membrane. The Matrigel assay was repeated four times for each of the cell lines. Matrigel basement membrane matrix was purchased from Becton Dickinson. Transwell polycarbonate membrane (12 mm diameter, 12 ␮m pore size) and cell culture cluster dishes were purchased from Costar. DMEM medium was purchased from Sigma. Hematoxylin 2 was purchased from Richard-Allan Scientific.

Isolation of total ribonucleic acid (RNA). Total cellular RNA was isolated by a single-step method using TRI reagent according to the manufacturer’s instructions. One milliliter of TRI reagent was used per 10-cm 2 culture flask. TRI reagent kits were purchased from Molecular Research Center, Inc. (Cincinnati, OH). Chloroform, isopropanol, and ethanol were purchased from Fisher Scientific; diethylpyrocarbonate (DEPC) was purchased from Sigma. Treatment of cells with 12-O-tetradecanoylphorbol-13-acetate (TPA), concanavalin (Con-A), and polymerized collagen type I gel (Vitrogen). Cells were plated on culture flasks in serum-containing medium until 90% confluence was reached and then washed three times with serum-free medium. The cells were incubated for 24 h in the serum-free medium which was then replaced with serum-free medium plus either TPA (100 nM) or Con-A (100 ␮g/ml). Medium was then collected for gelatin zymography, and total RNA was isolated for reverse transcriptase-polymerase chain reaction (RT-PCR) after 24 h of incubation. Vitrogen was diluted 8:1:1 with 10X PBS and 0.1 M NaOH, respectively. Five milliliters of diluted Vitrogen was allowed to gel in a 75-cm 2 tissue culture flask. Cells were plated on Vitrogen in serum-containing medium until 90% confluence was reached and then washed three times with serum-free medium. The cells were incubated in serum-free medium for an additional 24 h before the conditioned media were collected and total RNA was isolated. 12-O-Tetradecanoylphorbol-13-acetate was purchased from Sigma; concanavalin was purchased from EY Laboratories, Inc. (San Mateo, CA); Vitrogen was purchased from Collagen Biomaterials (Palo Alto, CA). Reverse transcriptase-polymerase chain reaction. Using previously reported primers [10], constitutive mRNA levels of matrix metalloproteinases and tissue inhibitors of metalloproteinases were determined by reverse transcriptase-polymerase chain reaction using a SuperScript one-step RT-PCR system kit. MMP2, MMP-9, MT1-MMP, and TIMP1 mRNA levels were measured after stimulation of cells with TPA, Con-A, and Vitrogen. Reactions were assembled by mixing 1.0 ␮g of total RNA with every 25 ␮l of 2X reaction mix, 6 ␮l of mixed primers (a final concentration of 0.2 ␮M), 1 ␮l of RT/Taq mix, 6 ␮l of 5 mM MgSO 4, and autoclaved distilled water to a final volume of 50 ␮l. RT-PCR was performed in a TEMP•TRONIC DNA thermal cycler with a 30-min incubation at 50°C, 2 min of predenaturation at 94°C, and 40 cycles of amplification consisting of 94°C for 30 s (denaturation), 55°C for 30 s (annealing), and 72°C for 1 min (elongation). ␤-Actin was used as external control. The PCR product (10 ␮l) for each sample was analyzed by electrophoresis in a 1% agarose gel containing ethidium bromide. Molecular Dynamics Imagine Analysis program was used to obtain quantitative RT-PCR results of mRNA expression. SuperScript one-step RT-PCR system kits and primers were purchased from Life Technologies Inc.; 100-bp DNA ladder was purchased from GIBCO BRL; agarose was purchased from Sigma. The primer sequences were as follows: MMP-2 (605 bp), 5⬘-GTGCTGAAGGACACACTAAAGAAGA-3⬘ (sense) and 5⬘-TTGCCATCCTTCTCAAAGTTGTAGG-3⬘ (antisense); MMP-9 (243 bp), 5⬘-CACTGTCCACCCCTCAGAGC-3⬘ (sense) and 5⬘GCCACTTGTCGGCGATAAGG-3⬘ (antisense); MT1-MMP (497 bp), 5⬘-CGCTACGCCATCCAGGGTCTCAAA-3⬘ (sense) and 5⬘-CGGTCATCATCGGGCAGCACAAAA-3⬘ (antisense); TIMP1 (667 bp), 5⬘ATCCTGTTGTTGCTGTGGCTGATAG-3⬘ (sense) and 5⬘-TGCTGGGTGGTAACTCTTTATTTCA-3⬘ (antisense). Secretion of proteolytic enzyme determined by gelatin zymography. Gelatin zymography was performed using precast polyacrylamide gels containing 0.1% gelatin. The collected media were centrifuged at 1500 rpm for 3 min. Equal amounts of supernatant media were mixed with 2X Tris-glycine SDS sample buffer. A 20-␮l mixed sample was loaded. Electrophoresis was performed at 125 V for 90 min. After electrophoresis, the gel was rinsed with renaturing buffer for 30 min and then incubated overnight at 37°C in developing buffer. The gel was stained with Coomassie blue and destained with 5% acetic acid in 20% methanol. The unstained bands corresponded to the areas of collagen digestion. XCell II Mini-Cell (Model EI9001), precast poly-

35

YANG ET AL.: INVASIVENESS AND MMP EXPRESSION

TABLE 1 Results of Matrigel Invasion Assay in Pancreatic Cell Line SUIT-2 and Its Four Sublines Cell lines

1 (%)

2 (%)

3 (%)

4 (%)

Average (%)

SUIT-2 S2-007 S2-013 S2-020 S2-028

2.27 0.41 5.53 8.88 0.15

3.08 0.61 3.11 4.28 0.32

2.78 1.00 3.29 7.19 0.45

4.58 2.01 4.30 5.03 0.74

3.18 ⫾ 0.99 1.01 ⫾ 0.71 4.06 ⫾ 1.11 6.35 ⫾ 2.09 0.42 ⫾ 0.24

Note. The difference in relative invasiveness between the cell lines was analyzed using the Mann-Whitney U statistical test. SUIT-2 relative to S2-028, S2-020, S2-007: P ⬍ 0.05. S2-020 relative to S2-028, S2-007: P ⬍ 0.05. S2-013 relative to S2-028, S2-007: P ⬍ 0.05. All others: not significant. acrylamide gels (containing 0.1% gelatin), Tris-glycine SDS sample buffer (2X), Tris-glycine SDS running buffer (10X), zymogram renaturing buffer (10X), and zymogram developing buffer (10X) were purchased from Novex (San Diego, CA). Coomassie blue was purchased from Sigma; acetic acid and methanol were purchased from Fisher Scientific.

RESULTS

Matrigel invasion assay. Table 1 shows the relative invasiveness from four assays with each pancreatic cell line SUIT-2 and its four sublines, S2-007, S2-013, S2020, and S2-028. The results of each experiment, as well as the mean, were reported. Each reading was the result of counting the total number of cells on the bottom of three membranes. Among the sublines of SUIT-2, the most invasive, S2-020, was 15 times more invasive than the least invasive, S2-028. The invasiveness of S2-013 and S2-007 was close to that of S2-020 and S2-028, respectively. SUIT-2 had an intermediate invasive ability compared to its sublines. Constitutive mRNA expression of MMPs and TIMPs in pancreatic cell line SUIT-2 and its four sublines by RT-PCR. The constitutive mRNA expression of MMPs and TIMPs of SUIT-2 and its four sublines is shown in Table 2. Using the Molecular Dynamics Imagine Analysis program, the relative extent of mRNA expression was estimated by comparing gel patterns of RT-PCR with those of ␤-actin (Fig. 1A). It was found that MMP-1, MMP-14, and MMP-15, as well as TIMP-1 and TIMP-2, were strongly expressed at the mRNA level in SUIT-2 and all of its sublines (Fig. 1B). No constitutive mRNA expression of MMP-2, MMP-3, MMP-8, MMP-9, or MMP-12 was observed in any of the above cell lines (Fig. 1C). For MMP-2 and MMP-9, no gene expression was detected even after the number of cycles for PCR amplification was increased. mRNA expression varied for other MMPs and TIMP-3 in these cells (Fig. 1D). However, no correlation was found between their invasive abilities and the gene expression.

Regulation of MMP-2, MMP-9, MT1-MMP, and TIMP-1 mRNA gene expression by TPA, Con-A, and Vitrogen. The effect of TPA, Con-A, and Vitrogen on the mRNA gene expression of MMP-2, MMP-9, MT1MMP, and TIMP-1 is summarized in Fig. 2. Cells were incubated in serum-free medium with TPA, Con-A, or Vitrogen for 24 h before the conditioned medium was collected and total mRNA was isolated. RT-PCR was employed and the Molecular Dynamics Imagine Analysis program was used to investigate the change of mRNA expression (Table 2). Gelatin zymography was used to measure the activity of MMP-2 and MMP-9 products (Fig. 3). MMP-2 mRNA gene expression was induced by TPA, Con-A, and Vitrogen only in the most invasive subline, S2-020. MMP-9 mRNA expression was up-regulated by TPA in all studied cells. However, the effect on the most invasive subline, S2-020, was significantly stronger than on the least invasive subline, S2-028. Con-A only up-regulated MMP-9 expression in SUIT-2. Vitrogen seemed to have a strong universal effect on the up-regulation of MMP-9 in all of these cell. Both MT1-MMP and TIMP-1 had strong constitutive mRNA expression and no change was observed with simulation by TPA or Con-A. Gelatin zymography results were consistent with those from RT-PCR in MMP-2 and MMP-9. TABLE 2 Constitutive mRNA Expression of MMPs and TIMPs in Pancreatic Cell Line SUIT-2 and Its Four Sublines by RT-PCR Genes

SUIT-2

S2-020

S2-013

S2-007

S2-028

␤-actin MMP-1 MMP-2 MMP-3 MMP-7 MMP-8 MMP-9 MMP-10 MMP-11 MMP-12 MMP-13 MMP-14 MMP-15 MMP-16 TIMP-1 TIMP-2 TIMP-3

⫹⫹⫹ ⫹⫹⫹⫹ ⫺ ⫺ ⫹⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹⫹⫹⫹ ⫹⫹⫹⫹ ⫺ ⫹⫹ ⫹⫹⫹ ⫹⫹⫹⫹

⫹⫹⫹ ⫹⫹⫹ ⫺ ⫺ ⫹⫹⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹⫹ ⫹⫹⫹⫹ ⫹⫹⫹ ⫺ ⫹⫹ ⫹⫹⫹ ⫹

⫹⫹⫹ ⫹⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹⫹⫹⫹ ⫹⫹⫹⫹ ⫺ ⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹

⫹⫹⫹ ⫹⫹⫹⫹ ⫺ ⫺ ⫹⫹⫹⫹ ⫺ ⫺ ⫹ ⫹ ⫺ ⫹⫹⫹ ⫹⫹⫹⫹ ⫹⫹⫹ ⫹ ⫹⫹ ⫹⫹⫹ ⫹

⫹⫹⫹ ⫹⫹⫹ ⫺ ⫺ ⫹⫹⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹⫹⫹ ⫹⫹⫹⫹ ⫺ ⫹⫹⫹ ⫹⫹⫹ ⫹⫹

Note. Molecular Dynamics Imagine Analysis program was used to obtain quantitative RT-PCR results of mRNA expression. ⫹⫹⫹⫹, volume report above 60,000; ⫹⫹⫹, volume report between 40,000 and 60,000; ⫹⫹, volume report between 20,000 and 40,000; ⫹, volume report between 1000 and 20,000; ⫺, volume report below 1000.

36

JOURNAL OF SURGICAL RESEARCH: VOL. 98, NO. 1, JUNE 1, 2001

FIG. 1. Constitutive mRNA expression of MMPs and TIMPs in SUIT-2 pancreatic tumor cell line and its sublines. SUIT-2, S2-028, S2-020, S2-013, and S2-007 cells were grown in McCoy’s medium, with serum, to 90% confluence and total RNA was isolated as described under Materials and Methods. Using previously reported primers, one-step RT-PCR was performed to measure mRNA expression of MMPs and TIMPs as described under Materials and Methods. (A) Gel pattern of mRNA gene expression for ␤-actin by RT-PCR. (B) Gel pattern of mRNA expression by RT-PCR for genes which showed expression in all studied cells. (C) Gel pattern of mRNA expression by RT-PCR for genes which lacked expression in all studied cells. (D) Gel pattern of mRNA expression by RT-PCR for genes whose expression varied in the studied cells. Line 1, DNA marker; line 2, MDA-MB-435; line 3, SUIT-2; line 4, S2-020; line 5, S2-013; line 6, S2-007; line 7, S2-028.

DISCUSSION

Although most malignant tumors are monoclonal in origin, their constituents are extremely heterogeneous. This is because multiple mutations occur independently and accumulate in different cells, thus producing subclones with different characteristics [11]. Studies of clonal variants of a single cell line have revealed differences in gene expression associated with their metastatic potentials [12, 13]. The four subclones investigated in this study were from a human pancreatic tumor cell line, SUIT-2; they were genetically similar yet distinct from each other not only in their morphology but also in their tumor marker expression [7]. In this study, the relative in vitro invasiveness was compared in SUIT-2 and its sublines. It was found that the most invasive subline, S2-020, was at least 15 times more invasive than the least invasive subline, S2-028. The invasiveness of the other two sublines, S2-013 and S2-007, was close to that of S2-020 and S2-028, respectively. SUIT-2 has an average invasive ability compared to its sublines. These results are consistent with the chemotactic activities of these cells toward Matrigel [14]. However, the in vitro invasive abilities of these cells did not correlate with their in

vivo metastatic potentials. In a tumorigenicity experiment in athymic nude mice [7], among the four sublines, S2-007 had the highest metastatic potential, S2028 had the lowest metastatic potential, and S2-020 and S2-013 were intermediate. Differences in gene expression that might explain these differences in metastatic potential were among the main questions addressed in this investigation. This study focused primarily on the role of matrix metalloproteinases (MMPs) in tumor cell invasiveness. Constitutive MMP and tissue inhibitor of metalloproteinase gene expression was found to occur and was investigated in this study. It was observed that collagenase (MMP-1), membrane-type matrix metalloproteinases (MT-MMPs) (MMP-14 and MMP-15), and TIMPs (TIMP-1 and TIMP-2) were strongly expressed in all cell lines. MMP-7, MMP-10, MMP-11, MMP-13, MMP-16, and TIMP-3 were expressed to varying extents although there was no evidence of correlation between the gene expression and invasiveness. Other known MMPs, including MMP-2 and MMP-9, were not detected at the messenger ribonucleic acid (mRNA) level. The lack of constitutive expression of MMP-2 and MMP-9 was not consistent with previous studies showing that MMP-2 and MMP-9 overexpression correlated

YANG ET AL.: INVASIVENESS AND MMP EXPRESSION

37

FIG. 2. Regulation of MMP-2, MMP-9, MT1-MMP, and TIMP1 expression by TPA, Con-A, and Vitrogen. SUIT-2, S2-028, S2-020, S2-013, and S2-007 cells were grown in McCoy’s medium, with serum, to 90% confluence. The seeding medium was removed, cells were washed three times with PBS, and serum-free medium was added with TPA (100 nM) or Con-A (100 ␮g/ml). For Vitrogen regulation, cells were cultured on a Vitrogen gel (3 mg/ml) in McCoy’s medium till 90% confluence was reached. The seeding medium was removed, cells were washed three times with PBS, and serum-free medium was added. For all TPA, Con-A, and Vitrogen regulation, medium was collected after 24 h and total RNA was isolated. RT-PCR gel pattern: MMP-2 mRNA gene expression was induced by TPA, Con-A, and Vitrogen only in S2-020; MMP-9 expression was up-regulated by TPA in all studied cells but the effect on SUIT-2, S2-020, and S2-013 was significantly stronger than on S2-028 and S2-007; Con-A only up-regulated MMP-9 expression in SUIT-2; Vitrogen had a strong effect on the up-regulation in all studied cells. Line 1, DNA marker; line 2, MDA-MB-435; line 3, SUIT-2; line 4, S2-020; line 5, S2-013; line 6, S2-007; line 7, S2-028.

38

JOURNAL OF SURGICAL RESEARCH: VOL. 98, NO. 1, JUNE 1, 2001

FIG. 3. Gelatin zymography pattern. Gelatin zymography results were consistent with those from RT-PCR. Lines 1 and 2, type IV collagenase, used as gelatinase markers; line 3, S2-028; line 4, S2-007; line 5, S2-013; line 6, S2-020; line 7, SUIT-2. The position of the 92-kDa band was determined in relationship to the collagenase standard based on Johnson et al. [23] which showed that a band of latent MMP-9 (92 kDa) appeared in a zymography gel as a result of TPA stimulation.

with the degree of basement membrane disruption in pancreatic cancer [15]. Since the degradation of type IV collagen is thought to be a crucial step for tumor cell invasion and metastasis [16], it was thought that upregulation of MMP-2 and MMP-9 might generate a more malignant tumor. Type I collagen is a potent modulator of cellular function which plays an important role in cell morphogenesis, proliferation, adhesion, contraction, migration, and enzymatic degradation [17]. Mediated by integrin receptors, type I collagen can up-regulate expression of 72- and 92-kDa type IV collagenases (MMP-2 and MMP-9) [18]. Integrins are a large family of cell surface receptors that bind different components of extracellular matrix and in some cases mediate cellcell interaction [19]. Most cells express several integrins and one integrin molecule can have several different ligands. Integrins are heterodimers of ␣ and ␤ subunits and the ligand-binding site is composed of parts of both chains. At least 20 integrin heterodimers, composed of 14 types of ␣ and 8 types of ␤, have been found in humans. A single ␤ subunit can bind different ␣ subunits, forming integrins that bind different ligands. Requirements for interaction between ligands and integrin receptors are the subject of intense study. The tripeptide sequence Arg-Gly-Asp (RGD) has been shown to be the minimal structure required for integrin recognition. Upon integrin binding by a ligand, the ligand signal is transduced into an activating signal that mediates the phosphorylation of transcription factors, hence the regulation of gene expression. Different signal transduction pathways have been found to be involved in integrin-mediated interaction with collagen, including protein kinase C, G-protein, and calcium [20, 21]. In the present study, MMP-2 expression was induced upon 12-O-tetradecanoylphorbol-13-acetate, concanavalin, and Vitrogen treatment only in the most invasive subline, S2-020, suggesting the importance of MMP-2

up-regulation in conferring invasive potential. Even though TPA up-regulated MMP-9 expression in SUIT-2 and all of its four sublines, a good correlation was found between the extent of gene expression and cell invasiveness: the more invasive the cells, the stronger the MMP-9 mRNA expression. However, upon Con-A stimulation, MMP-9 expression was found only in SUIT-2. A possible explanation for this is that MMP-9 regulation by Con-A is not as important as by TPA or MMP-2 regulation in these pancreatic tumor cells. Twenty-eight subclones of SUIT-2 have been found and other clones may be sensitive to the stimulation of Con-A on MMP-9. Vitrogen stimulated strong MMP-9 expression in all cell lines and no correlation was found between gene expression and cell invasiveness. Vitrogen may not be specific for certain cells upon stimulation of MMP-9. Taken together, MMP-2 upregulation by all of the above stimuli and selective upregulation of MMP-9 by TPA may be required for different sublines of SUIT-2 to obtain their distinct invasive potentials. MT-MMPs and TIMPs have been found to play an important role in the regulation of MMP expression [22]. The assembly of MT1-MMP and TIMP2 with MMP-2 and integrin on the cell surface is thought critical for MMP-2 activation. Only activated MMPs have enzymatic function. It is interesting that both MT1-MMP and TIMP1 mRNA are constitutively expressed at a high level in all these cells, and neither TPA, Con-A, nor Vitrogen can alter the gene expression. Because protein synthesis appears to be required for MT-MMP and TIMP function, the translation of constitutive mRNA may be mediated to regulate the gene expression. CONCLUSION

This study has shown that exposure of SUIT-2 and its four subclones to TPA, Con-A, and Vitrogen induces mRNA and protein expression of MMP-2 and MMP-9

YANG ET AL.: INVASIVENESS AND MMP EXPRESSION

at different levels. The extent of MMP-2 and MMP-9 expression has a good correlation with their in vitro invasive potentials. TPA, Con-A, and type I collagen may modulate invasion and metastasis in human pancreatic tumor cells by increasing the expression of MMP-2 and MMP-9, which are under differential transcriptional regulation. REFERENCES 1. 2.

3. 4.

5.

6. 7.

8.

9.

10.

11.

Murr, M. M., Sarr, M. G., Oishi, A. J., and Van Heerden, J. A. Pancreatic cancer. Ca. Cancer J. Clin. 44: 304, 1994. Iwamura, T., Katsuki, T., and Ide, K. Establishment and characterization of a human pancreatic cancer cell line (SUIT-2) producing carcinoembryonic antigen and carbohydrate antigen 19-9. Jpn. J. Cancer Res. 78: 54, 1987. Liotta, L. A. Tumor invasion and metastasis: Role of the extracellular matrix. Cancer Res. 46: 1, 1986. Evans, J. D., Ghaneh, P., Kawesha, A., and Neoptolemos, J. P. Role of matrix metalloproteinases and their inhibitors in pancreatic cancer. Digestion 58: 520, 1997. Senior, R. M., Griffin, G. L., Fliszar, C. L., Shapiro, S. D., Goldberg, G. I., and Welgus, H. G. Human 92- and 72-kilodalton type IV collagenases are elastases. J. Biol. Chem. 266: 7870, 1991. Matrisian, L. M. Metalloproteinases and their inhibitors in matrix remodeling. Trends Genet. 6: 121, 1990. Iwamura, T., Taniguchi, S., Kitamura, N., Yamanari, H., Kojima, A., Hidaka, K., Setoguchi, T., and Katsuki, T. Correlation between CA19-9 production in vitro and histological grades of differentiation in clones isolated from a human pancreatic cancer cell line (SUIT-2). J. Gastroenterol. Hepatol. 7: 512, 1992. Albini, A., Iwamoto, Y., Kleinman, H. K., Martin, G. R., Aaronson, S. A., Kozlowski, J. M., and McEwan, R. N. A rapid in vitro assay for quantitating the invasive potential of tumor cells. Cancer Res. 47: 3239, 1987. Thompson, E. W., Shima, T. B., Reich, R., Graf, J., Albini, A., Martin, G. R., Dickson, R. B., and Lippman, M. E. Differential regulation of growth and invasiveness of MCF-7 breast cancer cells by antiestrogen. Cancer Res. 48: 6764, 1988. Giambernardi, T. A., Grant, G. M., Taylor, G. P., Hay, R. J., Maher, V. M., Mccormich, J. J., and Klebe, R. J. Overview of matrix metalloproteinase expression in cultured human cells. Matrix Biol. 16: 483, 1998. Cortran, R. S., Kumar, V., and Collins, T. Bobbins Pathologic Basis of Disease, 6 th ed. Philadelphia: Saunders, 1999. Pp. 302– 305.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

39

Albini, A., Aukerman, S. L., Ogle, R. C., Noonan, D. M., Fridman, R., Martin, G. R., and Fidler, I. J. The in vitro invasiveness and interactions with laminin of K-1735 melanoma cells. Evidence for different laminin-binding affinities in high and low metastatic variants. Clin. Exp. Metastasis 7: 437, 1989. Maslow, D. E. Tabulation of results on the heterogeneity of cellular characteristics among cells from B16 mouse melanoma cell lines with different colonization potentials. Clin. Exp. Metastasis 9: 182, 1989. Taniguchi, S., Iwamura, T., Kitamura, N., Yamanari, H., and Setoguchi, T. Heterogeneities of attachment, chemotaxis, and protease production among clones with different metastatic potentials from a human pancreatic cancer cell line. Clin. Exp. Metastasis 12: 238, 1994. Gress, T. M., Muller-Pillasch, F., Lerch, M. M., Freiss, H., Buchler, M., and Alder, G. Expression and in-situ localization of genes coding for extracellular matrix degrading proteases in pancreatic cancer. Int. J. Cancer 62: 407, 1995. Lee, C. S., Montebello, J., Georgiou, T., and Rode, J. Distribution of type IV collagen in pancreatic adenocarcinoma and chronic pancreatitis. Int. J. Exp. Pathol. 75: 79, 1994. Thompson, E. W., Yu, M., Bueno, J., Jin, L., Maiti, S. N., Palao-Marco, F. L., Pulyaeva, H., Tamborlane, J. W., Tirgari, R., Wapnir, I., and Azzam, H. Collagen induced MMP-2 activation in human breast cancer. Breast Cancer Res. Treat. 31: 357, 1994. Sarret, Y., Woodley, D. T., Goldberg, G. S., Kronberger, A., and Wynn, K. C. Constitutive synthesis of a 92-kDa keratinocytederived type IV collagenase is enhanced by type I collagen and decreased by type IV collagen matrices. J Invest. Dermatol. 99: 836, 1992. Lodish, H., Baltimore, D., Berk, A., Zipursky, S. L., Matsudaira, P., and Darnell, J. Molecular Cell Biology, 3 rd ed. New York: Sci. Am., 1995. Pp. 1123–1200. Korczak, B., Whale, C., and Kerbel, R. S. Possible involvement of Ca 2⫹ mobilization and protein kinase C activation in the induction of spontaneous metastasis by mouse adenocarcinoma cells. Cancer Res. 49: 2579, 1989. Lester, B. R., and McCathy, J. B. Tumor cell adhesion to the extracellular matrix and signal transduction mechanisms implicated in tumor cell motility, invasion and metastasis. Cancer Metastasis Rev. 11: 31, 1992. Deryugina, E. I., Bourdon, M. A., Reisfeld, R. A., and Strongin, A. Remodeling of collagen matrix by human tumor cells requires activation and cell surface association of matrix metalloproteinse-2. Cancer Res. 58: 3743, 1998. Johnson, M. D., Torri, J. A., Lippman, M. E., and Dickson, R. B. Regulation of motility and protease expression in PKCmediated induction of MCF-7 breast cancer invasiveness. Exp. Cell Res. 247: 105, 1999.