Pancreatic ductal adenocarcinoma (PDAC) is associated with an intense fibrotic reaction around the tumor known as desmoplastic reaction. This tissue is composed of interstitial matrix, predominantly type I collagen, together with proliferating fibroblastic cells. Despite the recognized importance of tumor-stromal interactions, very little is known about the interactions among pancreatic cells, myofibroblasts, and the interstitial matrix. The current study was undertaken to test the hypothesis that the desmoplastic reaction alters PDAC gene expression and cellular behavior. Evaluation of human pancreatic specimens showed increased fibrosis and enhanced membrane type 1-matrix metalloproteinase (MT1-MMP) expression in tumor specimens compared with normal pancreas. Using an in vitro model of tumor cell-stromal interactions, type I collagen and the extracellular matrix deposited by pancreatic fibroblasts and PDAC cells regulated motility of human papillomavirus–immortalized human pancreatic ductal epithelial (HPDE) cells. These “stromal” matrices also regulated MT1-MMP expression by HPDE cells, without affecting the expression of tissue inhibitor of metalloproteinase 2. Treatment with transforming growth factor-β1 (TGF-β1) type I receptor kinase inhibitors and function-blocking anti-TGF-β1 antibody abrogated matrix-mediated MT1-MMP induction. TGF-β1 also promoted MT1-MMP–dependent migration by HPDE cells. Moreover, compared with normal tissue, there was increased TGF-β1 signaling in grade 3 tumor specimens as shown by increased phospho-Smad2 staining. These data show that the crosstalk between cancer cells and stromal elements mediated by TGF-β1 influences cell surface– and pericellular matrix–degrading potential in vitro and may contribute to pancreatic cancer progression in vivo. (Cancer Res 2006; 66(14): 7032-40)

Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer related death in the United States, with a median survival time after diagnosis of <6 months (1). The American Cancer Society had estimated that there would be 32,180 new cases of PDAC and 31,800 deaths projected for 2005 (1). The poor outcome is attributed to the advanced stage of disease at the time of diagnosis, with 80% of the patients presenting with locally invasive or metastatic disease (2). Thus, strategies aimed at prevention of invasion and metastasis would generate immediate clinical effect. PDAC is frequently associated with an intense fibrotic reaction around the tumor tissue known as desmoplastic reaction (3). PDAC tumors exhibit a 3-fold increase in interstitial fibrillar collagen (types I and III) compared with the normal pancreas, and with loss of basement membrane, malignant cells are exposed to interstitial collagen (4, 5). The desmoplastic reaction is also associated with proliferation of fibroblastic cells, which in some cases outnumber local tumor cells (6). Current evidence suggests that these are mesenchymal cells, known as stellate cells, which have differentiated into an activated myofibroblastic phenotype. These activated myofibroblasts have been identified as the principal source of type I collagen in the desmoplastic reaction (7). Despite the recognized importance of tumor-stromal interactions (8, 9), very little is known about the interactions among pancreatic cells, myofibroblasts, and the interstitial matrix.

Transforming growth factor-β1 (TGF-β1), which is frequently overexpressed in PDAC, is associated with advanced tumor stage and a significant decrease in survival (10, 11). Moreover, in human surgical specimens, there is a strong correlation between expression of TGF-β1 mRNA and type I collagen mRNA, suggesting that TGF-β1 may directly be responsible for the fibrotic reaction in PDAC tumors (12). TGF-β1 signals through cell surface serine-threonine kinases to activate cellular responses (13). Binding of TGF-β1 to its type II receptor (TβRII) promotes TβRII association with and phosphorylation of type I receptor (TβRI), which then phosphorylates receptor-associated Smads (R-Smads) Smad2 and Smad3. The R-Smads then bind to Smad4 and translocate to the nucleus, wherein the complex can associate with transcription factors and regulate target genes.

Interstitial collagens, such as type I, are highly resistant to proteolysis due to the triple helical structure and the fibrillar organization (14, 15). Genetic studies support membrane type 1-matrix metalloproteinase (MT1-MMP) as a primary regulator of interstitial collagenolysis, as mice genetically deficient in MT1-MMP have severe growth defects due to the inability to process interstitial collagens during bone and soft tissue formation (16). Moreover, gene expression studies of stromal and neoplastic cells at the site of primary PDAC invasion have shown that MT1-MMP is a primary interstitial collagenase overexpressed by pancreatic cancer cells (17). Immunohistochemistry and in situ hybridization studies have shown that MT1-MMP is overexpressed in pancreatic tumors relative to normal pancreas, and that expression of MT1-MMP is enhanced in metastatic PDAC lesions compared with the primary tumors (18, 19).

The current study was undertaken to test the hypothesis that the desmoplastic reaction alters PDAC gene expression and cellular behavior. Evaluation of human pancreatic specimens showed increased fibrosis and enhanced MT1-MMP expression in tumor specimens compared with normal pancreas. Using an in vitro model of tumor cell-stromal interactions to better understand the role of fibrosis in tumor progression, type I collagen and the extracellular matrix deposited by pancreatic fibroblasts and PDAC cells were found to promote cell motility of human papillomavirus (HPV)–immortalized human pancreatic ductal epithelial (HPDE) cells. These “stromal” matrices also regulated MT1-MMP expression by HPDE cells, without affecting the expression of tissue inhibitor of metalloproteinase 2 (TIMP2). Treatment with TGF-β1 type I receptor kinase inhibitors and function-blocking anti-TGF-β1 antibody abrogated matrix-mediated MT1-MMP induction. TGF-β1 also promoted MT1-MMP–dependent migration by HPDE cells. Moreover, compared with normal tissue, there was increased TGF-β1 signaling in grade 3 tumor specimens as shown by increased phospho-Smad2 staining. These data show that the crosstalk between cancer cells and stromal elements mediated by TGF-β1 influences cell surface– and pericellular matrix–degrading potential in vitro and may contribute to pancreatic cancer progression in vivo.

Materials. Type I collagen, cell culture reagents, TGF-β1, peroxidase-conjugated secondary antibodies, and the MT1-MMP antibody directed against the hinge region were purchased from Sigma (St. Louis, MO). Anti-extracellular signal-regulated kinase 2 (anti-ERK2) antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The TGF-β1 type I receptor kinase inhibitor (designated TbRi; ref. 20) was obtained from Calbiochem (La Jolla, CA); an additional TGF-β1 type I receptor kinase inhibitor SB431542 (21) was obtained from Tocris (Ellisville, MO), whereas function-blocking anti-TGF-β1 antibody and Quantikine human TGF-β1 ELISA kit were purchased from R&D Systems (Minneapolis, MN). DMEM and keratinocyte-SFM were purchased from Life Technologies (Grand Island, NY). The broad-spectrum MMP inhibitor GM6001, TIMP1, TIMP2, anti-laminin 1 polyclonal (AB19012), anti-laminin 5 (MAB19562), and function-blocking anti-MT1-MMP (AB8102) antibodies were purchased from Chemicon (Temecula, CA). Anti-type I collagen antibody was obtained from Southern Biotech (Birmingham, AL). Supersignal enhanced chemiluminescence reagent was obtained from Pierce (Rockford, IL). Polyvinylidene difluoride membranes were purchased from Millipore (Bedford, MA). Protease inhibitor cocktail was obtained from Roche Diagnostics (Indianapolis, IN).

Immunohistochemistry. Pancreatic tissue microarrays were obtained from U.S. Biomax (Rockville, MD) and consisted of 70 tissue cores measuring 1.5 mm in diameter and 5 μm in thickness. The slides were H&E and trichrome stained or stained for MT1-MMP and phospho-Smad2 (pSmad2) by the Pathology Core Facility of the Robert H. Lurie Comprehensive Cancer Center at Northwestern University. The microarray tissue specimen grades were verified by a pathologist (V.A.) and included nine normal pancreas, 10 grade 1 adenocarcinomas, 27 grade 2 adenocarcinomas, and 24 grade 3 adenocarcinomas. The extent of fibrosis was determined by digitizing the trichrome-stained slide using an automated cellular imaging system from Clarient (San Juan Capistrano, CA) and quantifying the blue signal relative to the total digitized signal from each of the core specimens. Immunohistochemical staining with antibodies for MT1-MMP (Neomarkers, Fremont, CA; 1:100 dilution) and pSmad2 (Cell Signaling, Beverly, MA; 1:200 dilution) was done according to standard procedures. Analysis of tissue sections was done by light microscopy by a pathologist (V.A.), and a total of 200 tumor cells in each core (six fields) were examined at ×40 magnification. Because many of the tumors showed nonhomogenous staining, the percentages of tumor cells (excluding stromal cells) that were positive at each of the intensity scores 0 (absent staining), 1+ (weak staining), 2+ (moderate staining), and 3+ (strong staining) were recorded, and the sum of these scores gave a final score for each of the cores ranging from 0 to 0.99, 1 to 1.99, or 2 to 2.99. Ps were calculated using unpaired t test comparing the scores of each of the tumor grades with the normal pancreatic tissue samples. The association between fibrosis and MT1-MMP expression in the tumor samples was assessed using Pearson correlation coefficient test. Statistical analyses were done using GraphPad Instat 3 (San Diego, CA).

Cell culture. Premalignant HPDE cells (HPV16-immortalized normal pancreatic ductal epithelium) were generously provided by Dr. M. Tsao (Ontario Cancer Institute; ref. 22). Malignant PDAC Panc1 cells were obtained from the American Type Culture Collection (Manassas, VA). A pancreatic fibroblast cell line (F5), which was derived from a patient with pancreatic cancer, was obtained from Dr. M. Goggins (Johns Hopkins University; ref. 23). HPDE cells were maintained in keratinocyte-SFM supplemented with 100 units/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL BPE (supplied with the medium), and 5 ng/mL epidermal growth factor. Panc1 cells and F5 fibrobalsts were routinely maintained in DMEM containing 10% FCS and supplemented with 100 units/mL penicillin and 100 μg/mL streptomycin. The HPDE cells were plated in supplement containing media and supplement starved overnight before treatment with TGF-β1. In additional experiments, inhibitors or other chemical reagents were added 30 minutes before TGF-β1 treatment.

Generation of extracellular matrix and type I fibrillar collagen. The extracellular matrices deposited by HPDE cells, Panc1 cells, and F5 fibroblasts were generated as described previously (24, 25). Briefly, the cells were grown in six-well plates to 48 to 72 hours after confluence before treatment for 5 minutes with 20 mmol/L ammonium hydroxide to remove cells. After three rapid washes each in sterile distilled water and PBS, the deposited extracellular matrix was then used. To determine the composition of the deposited matrices, the ammonium hydroxide–treated six-well plates were scraped in Laemmli sample dilution buffer (25), and equal volumes of the lysates were analyzed by SDS-PAGE (7%) and immunoblotted for type I collagen (1:400), laminin 1 (1:5,000), and laminin 5 (1:1,000). In additional experiments, cells were plated on top of type I fibrillar collagen surfaces generated as previously described (25, 26). Briefly, acid-solubilized rat tail type I collagen was neutralized with NaOH according to the manufacturer's specification to a final concentration of 1 mg/mL, and 1.2 mL was added to a six-well tissue culture plate and allowed to form a gel.

Cell scattering and colloidal gold assays. HPDE cells (104) were supplement starved overnight and then plated onto the various matrices (HPDE matrix, Panc1 matrix, F5 matrix, and type I collagen) in supplement-free media for 24 hours, and the effect on cell shape was photographed using Nikon camera. Haptotactic motility was assessed as described previously by plating 103 cells on matrices overlaid with colloidal gold (27). Cells were allowed to migrate for 18 hours, and phagokinetic tracks (including circular clearings) were monitored by visual examination using a Zeiss microscope with dark-field illumination and photographed using Nikon camera. The relative motility was determined by quantifying the area generated by the tracks and the circular clearings using Adobe Photoshop (San Jose, CA).

Western blot analysis. Cells plated on HPDE, Panc1, and F5 matrices for 24 hours were detached by scraping, lysed in modified radioimmunoprecipitation assay buffer [50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 5 mmol/L EDTA, 1% Triton X-100, and 0.1% SDS] with protease inhibitor cocktail, and clarified by centrifugation. Equal amounts of the lysates were analyzed by SDS-PAGE (9% gels) and immunoblotted for MT1-MMP (1:1,000) and for ERK2 (1:5,000; refs. 28, 29).

Reverse transcription-PCR. Total RNA was isolated using RNeasy Mini kit (Qiagen, Chatsworth, CA) according to the manufacturer's specifications. Following digestion with RQ1 DNase (Ambion, Austin, TX) for 30 minutes at 37°C, the total RNA concentration was determined by spectrophotometric measurement. Primer pairs for human MT1-MMP, human TGF-β1, and human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were as follows: forward primer, 5′-GCCCATTGGCCAGTTCTGGCGGG-3′ and reverse primer, 5′-CCTCGTCCACCTCAATGATGATC-3′ (MT1-MMP); forward primer, 5′-TGAACCGGCCTTTCCTGCTTCTCATG-3′ and reverse primer, 5′-GCGGAAGTCAATGTACAGCTGCCGC-3′ (TGF-β1); forward primer, 5′-CGGAGTCAACGGATTTGGTCGTAT-3′ and reverse primer, 5′-AGCCTTCTCCATGGTGGTGAAGAC-3′ (GAPDH; ref. 29). The length of the MT1-MMP, TGF-β1, and GAPDH amplicons were 530, 152, and 307 bp, respectively. The reverse transcription-PCR (RT-PCR) reactions were done using the One-step RT-PCR kit (Invitrogen, San Diego, CA), and the PCR products were visualized by UV transillumination of 1.5% agarose gels stained with ethidium bromide.

Real-time PCR. Reverse transcription of RNA to cDNA was done using GeneAmp RNA PCR kit (Applied Biosystems, Foster City, CA). Quantitative gene expression was done using MT1-MMP and GAPDH gene-specific probes (Applied Biosystems) using Taqman Universal PCR Master Mix and the 7500 Fast Real-time PCR System (Applied Biosystems). The data were then quantified with the comparative Ct method for relative gene expression.

TGF-β1 ELISA. HPDE cells were supplement starved overnight, and then equal numbers of HPDE cells (5 × 105) were plated onto the various matrices (HPDE matrix, Panc1 matrix, F5 matrix, and type I collagen) in 1.5 mL of supplement-free media for 24 hours. Active and total TGF-β1 present in the conditioned media were analyzed by ELISA according to the manufacturer's specifications. Active TGF-β1 was determined by excluding the activation step with 1 N HCl in the protocol, whereas total TGF-β1 (active and latent) was determined by adding 1 N HCl as directed in the protocol. Because the levels of active TGF-β1 in the conditioned media were below the sensitivity of the ELISA kit, only results of total TGF-β1 are presented.

Migration assay. Migratory activity was quantified using Transwell inserts (8-μm pore size) coated with type I collagen (5 μg) on the upper surface of the membrane (29). HPDE cells (2 × 105) were added to the upper chamber in 500 μL of supplement-free medium. In selected experiments, TGF-β1 (10 ng/mL) was added to the lower well containing 500 μL of media to promote migration. The proteinase dependence of migration was determined by quantifying migration in the presence of the MMP inhibitor GM6001 (10 μmol/L), TIMP1 (15 nmol/L), TIMP2 (15 nmol/L), rabbit IgG (10 μg/mL), or function-blocking anti-MT1-MMP antibody (AB8102, 10 μg/mL). Moreover, the effect of TGF-β1 type I receptor kinase inhibitor TbRi (5μmol/L) on basal and TGF-β1-mediated migration was also determined. Nonmigratory cells were removed 24 hours later from the upper chamber with a cotton swab; filters were fixed and stained with Diff-Quik Stain; and migratory cells adherent to the underside of the filter were enumerated using an ocular micrometer and counting a minimum of 5 high-powered fields.

Analysis of fibrosis and MT1-MMP in human pancreatic tissue. Using microarrayed human tissue, we examined the relationship between fibrosis and MT1-MMP expression in 70 human pancreatic cancer samples. The extent of fibrosis was determined by trichrome staining. A representative example of H&E, trichrome, and MT1-MMP staining in normal pancreas and a grade 2 adenocarcinoma of the pancreas is shown in Fig. 1A. There is enhanced blue fibrotic staining in the tumor specimen compared with normal pancreas (Fig. 1A,, middle). Quantification of the extent of fibrosis using automated cellular imaging system showed a statistically significant increase in fibrosis in the tumor specimens compared with the normal pancreas with no differentiation by grade (Table 1). We also quantified MT1-MMP expression in these pancreatic cancer specimens. In agreement with the previously published data in pancreatic cancer (17), MT1-MMP expression was detected mainly in the epithelial cells instead of the stromal cells. As observed with fibrosis, a statistically significant overall increase in MT1-MMP expression was also observed in the tumor specimens compared with the normal pancreas (Fig. 1A; Table 1). Interestingly, tumor areas with enhanced fibrosis also showed increased MT1-MMP expression (Pearson correlation coefficient r = 0.31, P = 0.0095).

Figure 1.

Analysis of fibrosis, MT1-MMP, and pSmad2 staining in human pancreatic tissue. A, H&E, trichrome, and MT1-MMP staining in a normal pancreas and a grade 2 pancreatic tumor specimen. Magnification, ×400. F, fibrosis. B, H&E and pSmad2 staining in a normal pancreas and a grade 3 pancreatic tumor specimen. Magnification, ×400.

Figure 1.

Analysis of fibrosis, MT1-MMP, and pSmad2 staining in human pancreatic tissue. A, H&E, trichrome, and MT1-MMP staining in a normal pancreas and a grade 2 pancreatic tumor specimen. Magnification, ×400. F, fibrosis. B, H&E and pSmad2 staining in a normal pancreas and a grade 3 pancreatic tumor specimen. Magnification, ×400.

Close modal
Table 1.

Analysis of fibrosis and MT1-MMP expression in human pancreatic samples

HistologynTrichrome
MT1-MMP
Mean intensityP0-0.991-1.992-2.99Mean intensityP
Normal 0.039 ± 0.017  1.23 ± 0.17  
Grade 1 10 0.124 ± 0.045 <0.01 2.40 ± 0.14 <0.0001 
Grade 2 27 0.123 ± 0.053 <0.001 20 2.04 ± 0.13 <0.01 
Grade 3 24 0.097 ± 0.058 <0.05 10 14 1.90 ± 0.15 <0.05 
HistologynTrichrome
MT1-MMP
Mean intensityP0-0.991-1.992-2.99Mean intensityP
Normal 0.039 ± 0.017  1.23 ± 0.17  
Grade 1 10 0.124 ± 0.045 <0.01 2.40 ± 0.14 <0.0001 
Grade 2 27 0.123 ± 0.053 <0.001 20 2.04 ± 0.13 <0.01 
Grade 3 24 0.097 ± 0.058 <0.05 10 14 1.90 ± 0.15 <0.05 

NOTE: Seventy human pancreatic samples were stained with trichrome and analyzed for blue fribrotic staining using an automated imaging system. The relative blue “signal” in each of the cores was normalized to the total area of the core. Immunohistochemical expression of MT1-MMP was quantified by determining the percentages of tumor cells (excluding stromal cells) that were positive at each of the intensity scores 0 (absent staining), 1+ (weak staining), 2+ (moderate staining), and 3+ (strong staining) recorded. The sum of these scores gave a final score for each of the cores ranging from 0 to 0.99, 1 to 1.99, or 2 to 2.99. Ps were calculated using unpaired t test comparing each of the tumor grades with the normal pancreatic tissue samples.

Extracellular matrix and type I collagen promote a scattered phenotype and induce cell motility. To further understand the contribution of the prominent fibrotic reaction to tumor progression, we established an in vitro model of cell-stromal interactions by examining the effect of matrix deposited by cells in the microenvironment of premalignant pancreatic epithelium. Because both pancreatic cancer cells and fibroblasts contribute to extracellular matrix generation in vivo (5, 30), we first examined the composition of extracellular matrix deposited by Panc1 cells and pancreatic F5 fibroblasts and compared it with the extracellular matrix deposited by premalignant HPDE cells. As shown in Fig. 2A, the HPDE cells deposit predominantly a laminin 5–rich matrix in contrast to Panc1 cells and F5 fibroblasts, which assemble a type I collagen matrix. Matrix from all three cell lines contained laminin 1. These data highlight the distinct matrix deposition patterns of premalignant HPDE cells relative to malignant Panc1 cells or F5 fibroblasts.

Figure 2.

Extracellular matrix and type I collagen promote a scattered phenotype and induce cell motility. Extracellular matrices deposited by HPDE cells, Panc1 cells, and pancreatic F5 fibroblasts were generated as described in Materials and Methods. A, matrices were then extracted using Laemmli sample dilution buffer and analyzed for type I collagen, laminin 1, and laminin 5 by Western blot analysis. B, HPDE cells were plated for 24 hours onto its own matrix, Panc1 matrix, pancreatic F5 fibroblast matrix, or type I collagen, and phase-contrast pictures were taken using Nikon camera. C, colloidal gold phagokinetic tracks generated by HPDE cells (103) plated for 18 hours onto its own matrix, Panc1 matrix, pancreatic F5 fibroblast matrix, or type I collagen were visualized using dark-field illumination. D, the area generated by the motile cells was quantified, and the average area of individual tracks (including circular clearings) was compared with the area of the tracks and clearings generated by HPDE cells. *, P < 0.05, significantly different from control HPDE matrix. Representative of three independent experiments.

Figure 2.

Extracellular matrix and type I collagen promote a scattered phenotype and induce cell motility. Extracellular matrices deposited by HPDE cells, Panc1 cells, and pancreatic F5 fibroblasts were generated as described in Materials and Methods. A, matrices were then extracted using Laemmli sample dilution buffer and analyzed for type I collagen, laminin 1, and laminin 5 by Western blot analysis. B, HPDE cells were plated for 24 hours onto its own matrix, Panc1 matrix, pancreatic F5 fibroblast matrix, or type I collagen, and phase-contrast pictures were taken using Nikon camera. C, colloidal gold phagokinetic tracks generated by HPDE cells (103) plated for 18 hours onto its own matrix, Panc1 matrix, pancreatic F5 fibroblast matrix, or type I collagen were visualized using dark-field illumination. D, the area generated by the motile cells was quantified, and the average area of individual tracks (including circular clearings) was compared with the area of the tracks and clearings generated by HPDE cells. *, P < 0.05, significantly different from control HPDE matrix. Representative of three independent experiments.

Close modal

We then examined the effect of Panc1 and F5 matrices on the behavior of HPDE cells. HPDE cells in contact with Panc1 or F5 fibroblast matrix adopt a more elongated, scattered phenotype compared with cells adherent to HPDE matrix. Similar changes were observed when HPDE cells were cultured on type I collagen, the predominant matrix component in PDAC tumors in vivo (Fig. 1; refs. 4, 5). To further examine the effect of the different matrices on cell scattering, a colloidal gold assay was done to evaluate matrix–driven motility (Fig. 2C). Purified type I collagen, HPDE matrix, Panc1 matrix, and F5 matrix were overlaid with colloidal gold particles (27). HPDE (1 × 103) cells were plated onto the coated surface, and generation of phagokinetic tracts (including circular clearings) was examined after 18 hours (Fig. 2C). Although little motility was observed on HPDE matrix, F5 fibroblast matrix, Panc1 matrix, and type I collagen significantly enhanced the motility of HPDE cells by 5- to 7-fold (Fig. 2D). Note, in this phagokinetic colloidal gold assay, the gold taken in by the cells becomes toxic, and thus, cells round up (27). However, the intent of this assay is to quantify the area of phagokinetic tracks rather than cell shape to evaluate relative motility on different matrices (27).

TGF-β1 signaling modulates extracellular matrix–mediated MT1-MMP expression. Because MT1-MMP regulates cell migration and invasion in three-dimensional collagen-rich matrices (31, 32), the effect of extracellular matrix on MT1-MMP expression by HPDE cells was evaluated. HPDE cells were plated onto its own matrix or matrix deposited by Panc1 cells or F5 fibroblasts for 24 hours, and MT1-MMP expression was determined by Western blotting and RT-PCR. Relative to HPDE cells plated onto its own matrix, contact with Panc1 or F5 matrix induced MT1-MMP protein and mRNA expression (Fig. 3A and B, respectively), indicating that contact between premalignant cells and matrix deposited by fibroblast and/or malignant cells regulates expression of a predominant matrix-degrading proteinase.

Figure 3.

Extracellular matrix promotes MT1-MMP expression. Extracellular matrices deposited by HPDE cells, Panc1 cells, and pancreatic F5 fibroblasts were generated as described in Materials and Methods. HPDE cells plated were then plated for 24 hours onto its own matrix, Panc1 matrix, or F5 matrix and analyzed for 55-kDa active form of MT1-MMP (top) and ERK2 (bottom) protein expression by Western blotting (A) or for MT1-MMP (top) and GAPDH (bottom) mRNA expression by RT-PCR (B). Representative of four independent experiments.

Figure 3.

Extracellular matrix promotes MT1-MMP expression. Extracellular matrices deposited by HPDE cells, Panc1 cells, and pancreatic F5 fibroblasts were generated as described in Materials and Methods. HPDE cells plated were then plated for 24 hours onto its own matrix, Panc1 matrix, or F5 matrix and analyzed for 55-kDa active form of MT1-MMP (top) and ERK2 (bottom) protein expression by Western blotting (A) or for MT1-MMP (top) and GAPDH (bottom) mRNA expression by RT-PCR (B). Representative of four independent experiments.

Close modal

As TGF-β1 is linked to both matrix deposition and protease expression, we examined the role of TGF-β1 as a potential regulator of MT1-MMP expression by HPDE cells. In control experiments, TGF-β1 induced MT1-MMP expression in HPDE cells (Fig. 4A, compare lanes 1 and 3), and this effect was blocked with the highly specific inhibitors TbRi (Fig. 4A, compare lanes 3 and 4) and the specific inhibitor of TGF-β1 type I receptor kinase activity SB431542 (data not shown). These data, in agreement with our previously published data in oral cancer (29), indicate that TGF-β1 regulates MT1-MMP expression in HPDE cells.

Figure 4.

TGF-β1 modulates extracellular matrix–mediated MT1-MMP expression. A, HPDE cells plated onto plastic culture plates were treated with TGF-β1 (10 ng/mL) in the presence or absence of TβRI kinase inhibitor (TbRi, 5 μmol/L) for 24 hours. Cell lysates were analyzed for the 55-kDa active form of MT1-MMP (top) and ERK2 (bottom) by Western blotting. B, extracellular matrices deposited by HPDE cells, Panc1 cells, and pancreatic F5 fibroblasts were generated as described in Materials and Methods. HPDE cells were plated in the presence or absence of TβRI kinase inhibitor (TbRi, 5 μmol/L) for 24 hours onto its own matrix, Panc1 matrix, or F5 fibroblast matrix. Cell lysates were analyzed for the 55-kDa active form of MT1-MMP (top) and ERK2 (bottom) protein expression by Western blotting. C, HPDE cells were plated onto its own matrix or onto Panc1 or F5 fibroblast matrix and allowed to condition the media for 24 hours. The cell lysates were analyzed for TGF-β1 (top) and GAPDH (bottom) mRNA expression by RT-PCR and the conditioned media for total TGF-β1 by ELISA. *, P < 0.05, significantly different from control HPDE matrix. Representative of three different experiments.

Figure 4.

TGF-β1 modulates extracellular matrix–mediated MT1-MMP expression. A, HPDE cells plated onto plastic culture plates were treated with TGF-β1 (10 ng/mL) in the presence or absence of TβRI kinase inhibitor (TbRi, 5 μmol/L) for 24 hours. Cell lysates were analyzed for the 55-kDa active form of MT1-MMP (top) and ERK2 (bottom) by Western blotting. B, extracellular matrices deposited by HPDE cells, Panc1 cells, and pancreatic F5 fibroblasts were generated as described in Materials and Methods. HPDE cells were plated in the presence or absence of TβRI kinase inhibitor (TbRi, 5 μmol/L) for 24 hours onto its own matrix, Panc1 matrix, or F5 fibroblast matrix. Cell lysates were analyzed for the 55-kDa active form of MT1-MMP (top) and ERK2 (bottom) protein expression by Western blotting. C, HPDE cells were plated onto its own matrix or onto Panc1 or F5 fibroblast matrix and allowed to condition the media for 24 hours. The cell lysates were analyzed for TGF-β1 (top) and GAPDH (bottom) mRNA expression by RT-PCR and the conditioned media for total TGF-β1 by ELISA. *, P < 0.05, significantly different from control HPDE matrix. Representative of three different experiments.

Close modal

These highly specific inhibitors of TGF-β1 type I receptor kinase activity were then used to examine the extent to which extracellular matrix–induced MT1-MMP expression was mediated by TGF-β1 signaling. HPDE cells were plated onto its own matrix or matrix deposited by Panc1 or F5 fibroblasts in the absence or presence of the TbRi kinase inhibitor. As shown previously (Fig. 3), both Panc1 and F5 fibroblast matrices induced MT1-MMP expression (Fig. 4B,, lanes 2 and 3). However, in the presence of TbRi kinase inhibitor, the matrix-induced MT1-MMP expression was completely abrogated (Fig. 4B , lanes 5 and 6), implicating TGF-β1 signaling in extracellular matrix–induced MT1-MMP regulation.

To evaluate the effect of matrix composition on TGF-β1 expression, HPDE cells were plated onto HPDE, Panc1, or F5 fibroblast matrix in serum-free media and allowed to condition the media for additional 24 hours, and the conditioned media were then analyzed for TGF-β1 by ELISA, and the cell lysates were examined for TGF-β1 message by RT-PCR. The HPDE cells plated onto Panc1 and F5 fibroblast matrices have increased TGF-β1 message compared with the cells plated onto its own matrix (Fig. 4C). Also significantly higher levels of TGF-β1 were detected in the conditioned media of HPDE cells in contact with Panc1 or F5 matrices relative to HPDE matrix (Fig. 4C).

Type I collagen-mediated MT1-MMP expression involves TGF-β1 signaling. Because our data indicate that the presence of type I collagen distinguishes Panc1 and F5 matrices from HPDE matrix and promotes cell scattering and motility as effectively as cell-deposited matrices, we examined the extent to which type I collagen can induce MT1-MMP and TGF-β1 expression in HPDE cells. HPDE cells were plated onto three-dimensional type I collagen or HPDE matrix in the presence or absence of TGF-β1 type I receptor kinase inhibitor or function-blocking anti-TGF-β1 antibody for 24 hours. Because it is technically difficult to normalize protein concentration of cell lysates extracted from cells plated onto three-dimensional type I collagen, MT1-MMP mRNAs were quantified by real-time PCR and normalized to GAPDH message levels. As shown in Fig. 5A, HPDE cells expressed a 2-fold increase in MT1-MMP mRNA when plated onto type I collagen. Moreover, quantification of TGF-β1 in the conditioned media by ELISA showed a concomitant increase in TGF-β1 production by HPDE cells plated onto type I collagen (Fig. 5A,, bottom inset). Consistent with our data in Fig. 4B, this enhanced MT1-MMP was blocked in the presence of TbRi kinase inhibitor and was also reduced to HPDE baseline levels in the presence of the function-blocking anti-TGF-β1 antibody (Fig. 5B). Additionally, Panc1 cells also increased MT1-MMP in response to type I collagen (Supplementary Figure). Moreover, the increase in MT1-MMP expression was blocked with the TbRi kinase inhibitor (Supplementary Figure). Because MT1-MMP activity is modulated post-translationally by TIMP2, we also examined and quantified the effect of type I collagen on TIMP2 expression using real-time PCR. In contrast to collagen-induced MT1-MMP expression, type I collagen did not affect TIMP2 expression in HPDE cells, and TIMP2 was not modulated by the TbRi kinase inhibitor or the function-blocking anti-TGF-β1 antibody (Fig. 5C).

Figure 5.

TGF-β1 modulates type I collagen-mediated MT1-MMP expression. HPDE cells were plated onto its own matrix or onto type I collagen as detailed in Materials and Methods in the presence of vehicle control DMSO, 5 μmol/L TβRI kinase inhibitor TbRi, or 20 μg/mL function-blocking anti-TGF-β1 antibody for 24 hours. The samples were analyzed for MT1-MMP, TIMP2, and GAPDH by real-time PCR. A, representative real-time data for MT1-MMP and GAPDH (top inset) gene expression. The conditioned media were also analyzed for total TGF-β1 levels by ELISA (bottom inset). B and C, the relative expression of MT1-MMP/GAPDH (B) and TIMP2/GAPDH (C) was quantified with the comparative Ct method for relative gene expression as detailed in Materials and Methods. Results are expressed relative to control cells plated on HPDE matrix, which was set as 1.0. Columns, mean of three different experiments; bars, SE. *, P < 0.05, significantly different from control cells plated on HPDE matrix.

Figure 5.

TGF-β1 modulates type I collagen-mediated MT1-MMP expression. HPDE cells were plated onto its own matrix or onto type I collagen as detailed in Materials and Methods in the presence of vehicle control DMSO, 5 μmol/L TβRI kinase inhibitor TbRi, or 20 μg/mL function-blocking anti-TGF-β1 antibody for 24 hours. The samples were analyzed for MT1-MMP, TIMP2, and GAPDH by real-time PCR. A, representative real-time data for MT1-MMP and GAPDH (top inset) gene expression. The conditioned media were also analyzed for total TGF-β1 levels by ELISA (bottom inset). B and C, the relative expression of MT1-MMP/GAPDH (B) and TIMP2/GAPDH (C) was quantified with the comparative Ct method for relative gene expression as detailed in Materials and Methods. Results are expressed relative to control cells plated on HPDE matrix, which was set as 1.0. Columns, mean of three different experiments; bars, SE. *, P < 0.05, significantly different from control cells plated on HPDE matrix.

Close modal

These data support a model of collagen-induced MT1-MMP expression in pancreatic tumors mediated by activation of TGF-β1 signaling. To provide evidence for enhanced TGF-β1 signaling in pancreatic tumors, histologic sections were examined for expression of pSmad2. Analysis of the pSmad2 staining showed no significant difference in the staining pattern between normal pancreas and grade 1 and 2 tumors (Table 2); however, an increased nuclear staining of pSmad2 in grade 3 tumors was observed (Fig. 1B), resulting in overall statistically enhanced pSmad2 staining (Table 2). These data, together with our results showing that the grade 3 tumors also have increased trichrome staining and MT1-MMP expression compared with the normal pancreas (Table 1), provide strong correlative evidence in support of TGF-β1-regulated MT1-MMP expression in high-grade pancreatic adenocarcinomas in vivo.

Table 2.

Immunohistochemical expression of pSmad2 in human pancreatic samples

HistologynpSmad2
0-0.991-1.992-2.99Mean intensityP
Normal 1.56 ± 0.22  
Grade 1 10 1.27 ± 0.17 NS 
Grade 2 27 11 10 1.61 ± 0.15 NS 
Grade 3 24 18 2.45 ± 0.11 <0.001 
HistologynpSmad2
0-0.991-1.992-2.99Mean intensityP
Normal 1.56 ± 0.22  
Grade 1 10 1.27 ± 0.17 NS 
Grade 2 27 11 10 1.61 ± 0.15 NS 
Grade 3 24 18 2.45 ± 0.11 <0.001 

NOTE: Immunohistochemical expression of nuclear pSmad2 was quantified in normal pancreatic ductal cells and in pancreatic cancer cells by determining the percentages of tumor cells (excluding stromal cells) that were positive at each of the intensity scores 0 (absent staining), 1+ (weak staining), 2+ (moderate staining), and 3+ (strong staining) recorded. The sum of these scores gave a final score for each of the cores ranging from 0 to 0.99, 1 to 1.99, or 2 to 2.99. Ps were calculated using unpaired t test comparing each of the tumor grades with the normal pancreatic tissue samples.

Abbreviation: NS, not significant.

TGF-β1 promotes type I collagen-mediated migration. To evaluate the functional consequences of increased MT1-MMP expression, the effect of TGF-β1 on collagen-mediated migration was examined. Migration was examined in the presence of TGF-β1, TbRi kinase inhibitor, or the broad-spectrum MMP inhibitor GM6001. As shown in Fig. 6A, GM6001 blocked migration, showing that basal collagen-mediated migration by HPDE cells requires MMP activity. Although data in Fig. 5A indicate that collagen can induce TGF-β1 production, basal levels of TGF-β1 or MT1-MMP were not sufficient to promote migration in this in vitro assay as shown by failure of TbRi kinase inhibitor to affect basal migration (Fig. 6A). In contrast, addition of exogenous TGF-β1 significantly increased type I collagen-mediated migration, and this effect was blocked by both TbRi kinase inhibitor and GM6001, supporting a role for TGF-β1 signaling and MMP activity in collagen-mediated migration.

Figure 6.

TGF-β1 promotes type I collagen-mediated migration. A, HPDE cells were added to porous polycarbonate filters coated with type I collagen (5 μg) and treated with TGF-β1 (10 ng/mL) in the presence or absence of broad-spectrum MMP inhibitor GM6001 (10 μmol/L) or TβRI kinase inhibitor (TbRi, 5 μmol/L) and allowed to migrate for 24 hours. Nonmigratory cells were removed from upper chamber, and migrating cells were enumerated using ocular micrometer. Columns, mean of three different experiments; bars, SE. *, significantly different from untreated samples. B, the TGF-β1-mediated type I collagen-dependent migration experiment was done in the presence of TIMP1 (15 nmol/L), TIMP2 (15 nmol/L), purified rabbit IgG (10 μg/mL), or MT1-MMP function-blocking antibody AB8102 (anti-MT, 10 μg/mL). Columns, mean of two different experiments; bars, SE. *, P < 0.05, significantly different from TIMP1-treated samples; **, P < 0.05, significantly different from IgG-treated samples.

Figure 6.

TGF-β1 promotes type I collagen-mediated migration. A, HPDE cells were added to porous polycarbonate filters coated with type I collagen (5 μg) and treated with TGF-β1 (10 ng/mL) in the presence or absence of broad-spectrum MMP inhibitor GM6001 (10 μmol/L) or TβRI kinase inhibitor (TbRi, 5 μmol/L) and allowed to migrate for 24 hours. Nonmigratory cells were removed from upper chamber, and migrating cells were enumerated using ocular micrometer. Columns, mean of three different experiments; bars, SE. *, significantly different from untreated samples. B, the TGF-β1-mediated type I collagen-dependent migration experiment was done in the presence of TIMP1 (15 nmol/L), TIMP2 (15 nmol/L), purified rabbit IgG (10 μg/mL), or MT1-MMP function-blocking antibody AB8102 (anti-MT, 10 μg/mL). Columns, mean of two different experiments; bars, SE. *, P < 0.05, significantly different from TIMP1-treated samples; **, P < 0.05, significantly different from IgG-treated samples.

Close modal

To further show that the enhanced migration was mediated by MT1-MMP, the ability of TIMP1 and TIMP2 to block migration was determined (Fig. 6B). TIMP1 is a poor inhibitor of MT1-MMP activity, whereas TIMP2 is a well-described endogenous inhibitor of MT1-MMP activity (32). As shown in Fig. 6B, TGF-β1 promoted type I collagen-mediated migration by 5-fold, which was not inhibited by the addition of TIMP1. In contrast, TIMP2 significantly blocked TGF-β1-mediated migration, indicating that the enhanced type I collagen-mediated migration was due to increased MT1-MMP activity. This is supported by data showing that an anti-catalytic function-blocking MT1-MMP antibody also inhibited migration (Fig. 6B), thus identifying MT1-MMP as a key protease in TGF-β1-mediated migration.

Recent reports have shown the crosstalk between pancreatic cancer cells and stromal fibroblasts contributes to cancer progression (33, 34). Gene expression profiling of tumor-stromal interactions between pancreatic cancer cells and stromal fibroblasts showed that multiple genes are differentially expressed in pancreatic cancer cells and in fibroblasts as a consequence of their mutual interactions (23). Desmoplastic reaction, a very prominent feature of PDACs in vivo (3), is composed of mainly type I collagen deposited by fibroblasts (12), and the extracellular matrix in the desmoplastic reaction has also been shown to modulate the behavior of pancreatic cancer cells (33, 35). In this report, we show that the extracellular matrix deposited by pancreatic fibroblasts and cancer cells modulates the behavior of pancreatic ductal cells by inducing a more scattered phenotype and enhancing cell motility. Interestingly, these changes in cell scattering and motility were similarly shown when the ductal cells were plated onto purified type I collagen matrix, suggesting that the desmoplastic reaction in vivo may also contribute to tumor progression by increasing cellular motility.

Our data show that the HPDE ductal cells express the membrane-anchored metalloproteinase MT1-MMP when these cells encounter type I collagen-enriched extracellular matrix deposited by fibroblasts and/or malignant cells. Interestingly, malignant Panc1 cells also increase MT1-MMP expression when these cells are exposed to type I collagen. Acquisition of this metalloproteinase will significantly enhance the ability of these cells to invade through a collagen-rich barrier, as recent reports have shown that MT1-MMP is the dominant regulator of tissue invasive activities by normal and neoplastic cells in the in vivo setting (36). Given the crucial role that MT1-MMP plays in both physiologic and pathologic conditions, MT1-MMP overexpression is seen in a number of different tumors, including PDACs as shown herein and in earlier reports (18, 19). Interestingly, MT1-MMP overexpression was particularly prominent in areas of the tumor with intense fibrotic reaction, suggesting that type I collagen-enriched desmoplastic reaction may also contribute to MT1-MMP expression in the in vivo setting.

Our results suggest a mechanism for type I collagen matrix regulation of MT1-MMP via induction of TGF-β1 signaling. Treatment with highly specific inhibitors of TGF-β1 type I receptor kinase activity blocked the expression of MT1-MMP at both the message and protein levels in HPDE cells. Moreover, type I collagen-mediated MT1-MMP expression by Panc1 cells was also blocked by TGF-β1 type I receptor kinase inhibitors. Although recent reports have shown that integrins can modulate TGF-β1 receptor tyrosine kinase activity in a ligand-independent manner (37, 38), our data show extracellular matrix and collagen-induced up-regulation of TGF-β1 as the function-blocking anti-TGF-β1 antibody abrogated the effect of extracellular matrix on MT1-MMP production. Moreover, increased TGF-β1 was detected by ELISA in the conditioned media of ductal cells plated onto extracellular matrix deposited by pancreatic fibroblasts and by cancer cells. In addition, there was increased Smad2 phosphorylation when the ductal cells were plated onto type I collagen compared with its own matrix (data not shown). Consistent with our findings, recent reports have shown that extracellular matrices can promote expression of growth factors and cytokines (39, 40). For example, human mesangial cells cultured on type I collagen showed increased TGF-β1 message and increased TGF-β1 production in the conditioned media (40). Moreover, this effect was blocked by the function-blocking anti-β1 integrin antibody and by dominant-negative integrin-linked kinase, suggesting that type I collagen-mediated TGF-β1 synthesis is regulated by β1 integrins. In addition to regulating TGF-β1 expression, integrins also have been shown to promote activation of latent complexes of TGF-β1 (41). Both αvβ8 and αvβ6 can activate latent complexes of TGF-β1, albeit by different mechanisms. The αvβ6-mediated activation is protease independent and without significant release of free TGF-β1, whereas αvβ8-mediated activation is mediated by proteolytic degradation of latency associated peptide (LAP) and release of free TGF-β1. Interestingly, MT1-MMP can degrade LAP and has been shown to play a role in αvβ8-mediated TGF-β1 activation (41). Although our experiments show that the extracellular matrix–mediated MT1-MMP involves TGF-β1 signaling as shown by inhibition of MT1-MMP expression using TGF-β1 type I receptor kinase inhibitors (Fig. 4B), we were not able to detect active TGF-β1 presumably due to the much shorter half-life of active relative to latent TGF-β1 (42).

Previous studies support a key role for TGF-β1 in pancreatic cancer progression. TGF-β1 treatment of pancreatic cancer cell lines increased MMP-2 expression and enhanced MMP-dependent Matrigel invasion (43, 44), whereas treatment with the TGF-β1 type I receptor kinase inhibitor SD-093 blocked Matrigel invasion (45). Our data show that TGF-β1 promotes type I collagen-mediated migration via an MT1-MMP–dependent mechanism, and this migratory activity is blocked using small molecule inhibitors of TGF-β1 type I receptor kinase activity. Because type I collagen is the predominant matrix in the desmoplastic reaction (4, 5), and because genetic studies have shown that MT1-MMP is a primary regulator of interstitial collagenolysis (16), our data suggest that inhibition of TGF-β1 signaling would also help to block tumor progression in vivo. Importantly, TGF-β1 has also been shown to promote a desmoplastic reaction in vivo following orthotopic transplantation of TGF-β1-transfected pancreatic cancer cells in an experimental model of human pancreatic carcinoma (46), whereas conditional loss of TGF-β1 signaling within the pancreas significantly ameliorated chemical-induced pancreatic fibrosis (47, 48). Based on our data and on recent reports showing that the desmoplastic reaction is detrimental to the host (7, 35), modulation of TGF-β1 signaling using small-molecule inhibitors, a number of which are currently under development (49), could be a potential approach for the treatment of this highly lethal cancer.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Grant support: National Cancer Institute grant K08CA94877, Zell Family Foundation, and H Foundation (H.G. Munshi).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. M. Sharon Stack for her support and encouragement throughout this project.

1
Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005.
CA Cancer J Clin
2005
;
55
:
10
–30.
2
Keleg S, Buchler P, Ludwig R, Buchler MW, Friess H. Invasion and metastasis in pancreatic cancer.
Mol Cancer
2003
;
2
:
14
.
3
Longnecker DS. Pathology and pathogenesis of diseases of the pancreas.
Am J Pathol
1982
;
107
:
103
–21.
4
Mollenhauer J, Roether I, Kern HF. Distribution of extracellular matrix proteins in pancreatic ductal adenocarcinoma and its influence on tumor cell proliferation in vitro.
Pancreas
1987
;
2
:
14
–24.
5
Imamura T, Iguchi H, Manabe T, et al. Quantitative analysis of collagen and collagen subtypes I, III, and V in human pancreatic cancer, tumor-associated chronic pancreatitis, and alcoholic chronic pancreatitis.
Pancreas
1995
;
11
:
357
–64.
6
Bachem MG, Schneider E, Gross H, et al. Identification, culture, and characterization of pancreatic stellate cells in rats and humans.
Gastroenterology
1998
;
115
:
421
–32.
7
Apte MV, Park S, Phillips PA, et al. Desmoplastic reaction in pancreatic cancer: role of pancreatic stellate cells.
Pancreas
2004
;
29
:
179
–87.
8
Hendrix MJ, Seftor EA, Kirschmann DA, Quaranta V, Seftor RE. Remodeling of the microenvironment by aggressive melanoma tumor cells.
Ann N Y Acad Sci
2003
;
995
:
151
–61.
9
Mueller MM, Fusenig NE. Friends or foes: bipolar effects of the tumour stroma in cancer.
Nat Rev Cancer
2004
;
4
:
839
–49.
10
Friess H, Yamanaka Y, Buchler M, et al. Enhanced expression of transforming growth factor β isoforms in pancreatic cancer correlates with decreased survival.
Gastroenterology
1993
;
105
:
1846
–56.
11
Yue H, Yang B, Zhang H, et al. Clinical significance of TGF- β1 and β-glucuronidase synchronous detection in human pancreatic cancer.
Hepatobiliary Pancreat Dis Int
2002
;
1
:
309
–11.
12
Aoyagi Y, Oda T, Kinoshita T, et al. Overexpression of TGF-β by infiltrated granulocytes correlates with the expression of collagen mRNA in pancreatic cancer.
Br J Cancer
2004
;
91
:
1316
–26.
13
Shi Y, Massague J. Mechanisms of TGF-β signaling from cell membrane to the nucleus.
Cell
2003
;
113
:
685
–700.
14
Sternlicht MD, Werb Z. How matrix metalloproteinases regulate cell behavior.
Annu Rev Cell Dev Biol
2001
;
17
:
463
–516.
15
Kadler K. Extracellular matrix 1: fibril-forming collagens.
Protein Profile
1995
;
2
:
491
–619.
16
Holmbeck K, Bianco P, Caterina J, et al. MT1-MMP-deficient mice develop dwarfism, osteopenia, arthritis, and connective tissue disease due to inadequate collagen turnover.
Cell
1999
;
99
:
81
–92.
17
Iacobuzio-Donahue CA, Ryu B, Hruban RH, Kern SE. Exploring the host desmoplastic response to pancreatic carcinoma: gene expression of stromal and neoplastic cells at the site of primary invasion.
Am J Pathol
2002
;
160
:
91
–9.
18
Bramhall SR, Neoptolemos JP, Stamp GW, Lemoine NR. Imbalance of expression of matrix metalloproteinases (MMPs) and tissue inhibitors of the matrix metalloproteinases (TIMPs) in human pancreatic carcinoma.
J Pathol
1997
;
182
:
347
–55.
19
Imamura T, Ohshio G, Mise M, et al. Expression of membrane-type matrix metalloproteinase-1 in human pancreatic adenocarcinomas.
J Cancer Res Clin Oncol
1998
;
124
:
65
–72.
20
Sawyer JS, Anderson BD, Beight DW, et al. Synthesis and activity of new aryl- and heteroaryl-substituted pyrazole inhibitors of the transforming growth factor-β type I receptor kinase domain.
J Med Chem
2003
;
46
:
3953
–6.
21
Laping NJ, Grygielko E, Mathur A, et al. Inhibition of transforming growth factor (TGF)-β1-induced extracellular matrix with a novel inhibitor of the TGF-β type I receptor kinase activity: SB-431542.
Mol Pharmacol
2002
;
62
:
58
–64.
22
Furukawa T, Duguid WP, Rosenberg L, Viallet J, Galloway DA, Tsao MS. Long-term culture and immortalization of epithelial cells from normal adult human pancreatic ducts transfected by the E6E7 gene of human papilloma virus 16.
Am J Pathol
1996
;
148
:
1763
–70.
23
Sato N, Maehara N, Goggins M. Gene expression profiling of tumor-stromal interactions between pancreatic cancer cells and stromal fibroblasts.
Cancer Res
2004
;
64
:
6950
–6.
24
Gospodarowicz D, Delgado D, Vlodavsky I. Permissive effect of the extracellular matrix on cell proliferation in vitro.
Proc Natl Acad Sci U S A
1980
;
77
:
4094
–8.
25
Munshi HG, Stack MS. Analysis of matrix degradation.
Methods Cell Biol
2002
;
69
:
195
–205.
26
Henriet P, Zhong ZD, Brooks PC, Weinberg KI, DeClerck YA. Contact with fibrillar collagen inhibits melanoma cell proliferation by up-regulating p27KIP1.
Proc Natl Acad Sci U S A
2000
;
97
:
10026
–31.
27
Albrecht-Buehler G. The phagokinetic tracks of 3T3 cells.
Cell
1977
;
11
:
395
–404.
28
Munshi HG, Wu YI, Ariztia EV, Stack MS. Calcium regulation of matrix metalloproteinase-mediated migration in oral squamous cell carcinoma cells.
J Biol Chem
2002
;
277
:
41480
–8.
29
Munshi HG, Wu YI, Mukhopadhyay S, et al. Differential regulation of membrane type 1-matrix metalloproteinase activity by ERK 1/2- and p38 MAPK-modulated tissue inhibitor of metalloproteinases 2 expression controls transforming growth factor-β1-induced pericellular collagenolysis.
J Biol Chem
2004
;
279
:
39042
–50.
30
Lohr M, Trautmann B, Gottler M, et al. Human ductal adenocarcinomas of the pancreas express extracellular matrix proteins.
Br J Cancer
1994
;
69
:
144
–51.
31
Itoh Y, Seiki M. MT1-MMP: a potent modifier of pericellular microenvironment.
J Cell Physiol
2006
;
206
:
1
–8.
32
Hotary K, Allen E, Punturieri A, Yana I, Weiss SJ. Regulation of cell invasion and morphogenesis in a three-dimensional type I collagen matrix by membrane-type matrix metalloproteinases 1, 2, and 3.
J Cell Biol
2000
;
149
:
1309
–23.
33
Miyamoto H, Murakami T, Tsuchida K, Sugino H, Miyake H, Tashiro S. Tumor-stroma interaction of human pancreatic cancer: acquired resistance to anticancer drugs and proliferation regulation is dependent on extracellular matrix proteins.
Pancreas
2004
;
28
:
38
–44.
34
Maehara N, Matsumoto K, Kuba K, Mizumoto K, Tanaka M, Nakamura T. NK4, a four-kringle antagonist of HGF, inhibits spreading and invasion of human pancreatic cancer cells.
Br J Cancer
2001
;
84
:
864
–73.
35
Armstrong T, Packham G, Murphy LB, et al. Type I collagen promotes the malignant phenotype of pancreatic ductal adenocarcinoma.
Clin Cancer Res
2004
;
10
:
7427
–37.
36
Sabeh F, Ota I, Holmbeck K, et al. Tumor cell traffic through the extracellular matrix is controlled by the membrane-anchored collagenase MT1-MMP.
J Cell Biol
2004
;
167
:
769
–81.
37
Owens DM, Romero MR, Gardner C, Watt FM. Suprabasal α6β4 integrin expression in epidermis results in enhanced tumourigenesis and disruption of TGFβ signalling.
J Cell Sci
2003
;
116
:
3783
–91.
38
Reynolds LE, Conti FJ, Lucas M, et al. Accelerated re-epithelialization in β3-integrin-deficient- mice is associated with enhanced TGF-β1 signaling.
Nat Med
2005
;
11
:
167
–74.
39
Ortega-Velazquez R, Diez-Marques ML, Ruiz-Torres MP, Gonzalez-Rubio M, Rodriguez-Puyol M, Rodriguez Puyol D. Arg-Gly-Asp-Ser (RGDS) peptide stimulates transforming growth factor β1 transcription and secretion through integrin activation.
FASEB J
2003
;
17
:
1529
–31.
40
Ortega-Velazquez R, Gonzalez-Rubio M, Ruiz-Torres MP, et al. Collagen I upregulates extracellular matrix gene expression and secretion of TGF-β 1 by cultured human mesangial cells.
Am J Physiol Cell Physiol
2004
;
286
:
C1335
–43.
41
Sheppard D. Integrin-mediated activation of latent transforming growth factor β.
Cancer Metastasis Rev
2005
;
24
:
395
–402.
42
Annes JP, Munger JS, Rifkin DB. Making sense of latent TGFβ activation.
J Cell Sci
2003
;
116
:
217
–24.
43
Ellenrieder V, Hendler SF, Ruhland C, Boeck W, Adler G, Gress TM. TGF-β-induced invasiveness of pancreatic cancer cells is mediated by matrix metalloproteinase-2 and the urokinase plasminogen activator system.
Int J Cancer
2001
;
93
:
204
–11.
44
Teraoka H, Sawada T, Yamashita Y, et al. TGF-β1 promotes liver metastasis of pancreatic cancer by modulating the capacity of cellular invasion.
Int J Oncol
2001
;
19
:
709
–15.
45
Subramanian G, Schwarz RE, Higgins L, et al. Targeting endogenous transforming growth factor β receptor signaling in SMAD4-deficient human pancreatic carcinoma cells inhibits their invasive phenotype1.
Cancer Res
2004
;
64
:
5200
–11.
46
Lohr M, Schmidt C, Ringel J, et al. Transforming growth factor-β1 induces desmoplasia in an experimental model of human pancreatic carcinoma.
Cancer Res
2001
;
61
:
550
–5.
47
Yoo BM, Yeo M, Oh TY, et al. Amelioration of pancreatic fibrosis in mice with defective TGF-β signaling.
Pancreas
2005
;
30
:
e71
–9.
48
Nagashio Y, Ueno H, Imamura M, et al. Inhibition of transforming growth factor β decreases pancreatic fibrosis and protects the pancreas against chronic injury in mice.
Lab Invest
2004
;
84
:
1610
–8.
49
Yingling JM, Blanchard KL, Sawyer JS. Development of TGF-β signalling inhibitors for cancer therapy.
Nat Rev Drug Discov
2004
;
3
:
1011
–22.

Supplementary data