The aim of this study was to examine the effects of transforming growth factor (TGF) β1 on the phenotype and the biological behavior of pancreatic cancer cell lines with and without mutations in the TGF-β signaling pathway and to elucidate whether the Ras signaling cascade participates in mediating these effects of TGF-β1. TGF-β-responsive (PANC-1, COLO-357, and IMIM-PC1) and nonresponsive (CAPAN1 and IMIM-PC2) pancreatic cancer cell lines with activating mutations of the Ki-Ras oncogene were treated with 10 ng/ml TGF-β1 over time. Phenotypic alterations were studied by electron and phase contrast microscopy and by immunohistochemistry and expression analyses of differentiation markers. The influence of TGF-β on tumor cell scattering, migration, and invasion was determined. The role of the Ras-mitogen-activated protein kinase kinase (MEK)-extracellular signal-regulated kinase (ERK) cascade in mediating TGF-β-induced morphological and functional effects were studied by pretreatment with the MEK1 inhibitor PD 98059 and by measuring ERK2 activation using immune complex kinase assays. TGF-β1 led to a reversible and time-dependent epithelial-mesenchymal transdifferentiation (EMT) in TGF-β-responsive pancreatic cancer cell lines, characterized by a fibroblastoid morphology and an up-regulation of mesenchymal markers and a down-regulation of epithelial markers. EMT was associated with an increase in tumor cell migration, invasion, and scattering. In the responsive cell lines, TGF-β1 induced a moderate but sustained activation of ERK2. EMT, the concomitant changes in gene expression, and the invasive and migratory potential were reduced or abolished by pretreatment with the selective MEK1 inhibitor. Thus, in TGF-β-responsive pancreatic cancer cells with activating Ki-Ras mutations, TGF-β1 treatment caused an EMT associated with a more invasive phenotype. Cross-talk with the Ras-MEK-ERK-signaling cascade appears to be essential for mediating these effects of TGF-β1.

TGF3-β and the receptor TGFβRII have been described to be overexpressed in pancreatic cancer tissues (1, 2), although very little is known about the functional implications of this overexpression. Because not all pancreatic tumors have mutations known to inactivate the TGF-β pathway, such as Smad4 (40–50%) or TGFβRII (3–10%), the overexpression of this growth factor and its receptor may well have an important role in pancreatic carcinogenesis, rather than being a simple epiphenomenon (3, 4). TGF-β is known to play a dual role in tumorigenesis. In an early phase of tumorigenesis, TGF-β is a potent mediator of antiproliferative effects. Upon binding of TGF-β to the respective type II receptor (TGFβRII), the type I receptor becomes phosphorylated and propagates the signal downstream through phosphorylation and thereby activation of the Smad2 and Smad3 proteins (receptor Smads). The activated receptor Smads form a complex with Smad4 and translocate into the nucleus, where they interact with other transcription factors to regulate the expression of genes involved in cell growth control (5, 6, 7). In later stages of tumorigenesis, however, TGF-β may as well contribute to tumor progression by inducing cell spreading, migration, angiogenesis, and tumor cell invasion (8, 9, 10). Furthermore, TGF-β has been associated with changes of cell morphology during embryonic development and carcinogenesis (11). In a small number of tumor models, it has for example been shown that TGF-β leads to an EMT of tumor cells, which appears to be associated with an increase in cell motility and invasiveness (12). It is not clear yet which signaling pathways are responsible for mediating these pleiotropic effects of TGF-β. Oft et al.(12) provided data for a mouse mammary tumor model that implicate the Ras pathway in mediating these effects of TGF-β on tumor cell morphology. Because both overexpression of TGF-βs and their receptors and activating mutations of the Ki-Ras oncogene occur at a high frequency in pancreatic tumors, these interactions of the Ras cascade and the TGF-β pathway may play an important role in pancreatic carcinogenesis (13, 14). To date, no data have been reported concerning the influence of TGF-β on the Ras cascade in pancreatic cancer cells, and the role of this interaction on crucial characteristics of tumor cells, such as differentiation, invasion, and migration. In the present study, we investigated the effect of TGF-β1 on phenotypic and functional characteristics of TGF-β-responsive pancreatic cancer cells and whether cross-talk between the TGF-β and the Ras signaling cascades is required to mediate these effects.

Materials.

The following human pancreatic cancer cell lines were used in the study and cultured in DMEM (Life Technologies, Inc., Gaithersburg, MD) supplemented with 10% FCS (Life Technologies, Inc.), 100 units/ml penicillin and 100 units/ml streptomycin: IMIM-PC1 and IMIM-PC2 (15) from F. X. Real (Institute Municipal de Investigacion Medica, Barcelona, Spain), COLO-357 (16) from H. Kalthoff (University of Kiel, Kiel, Germany), PANC-1 from the European Collection of Animal Cell Cultures (Salisbury, United Kingdom), and CAPAN-1 from the German Cancer Research Center (Heidelberg, Germany). The cDNA probes for cytokeratins 8 and 19 were cloned in our laboratory (17). The cDNA probes for human collagen type α1/I and TGF-β1 were obtained from American Type Culture Collection (Rockville, MD). A 750-bp human vimentin probe was PCR-amplified from a human chronic pancreatitis tissue sample (primer pair: 5′-TGGCACGTCTTGACCTTGAA-3′ and 5′-GGTCATCGTGATGCTGAGAA-3′). TGF-β was obtained from R&D (Wiesbaden, Germany), the MEK1 inhibitor PD 98059 from New England Biolabs (Schwalbach, Germany), and Matrigel from Becton 38 Dickinson (Bedford, MA).

Monoclonal guinea pig anti-cytokeratin 8/18 antibodies and monoclonal mouse anti-vimentin antibodies (Linaris, Wertheim, Germany) were used for immunostaining. For immunoblotting, we used monoclonal antibodies against cytokeratin 8, cytokeratin 19 (Boehringer, Mannheim, Germany), and vimentin (Serotec, Oxford, United Kingdom).

Morphological Studies.

Morphological alterations after TGF-β1 treatment were studied by phase contrast (Zeiss S 100 TV) and electron microscopy (Zeiss EM 100). To analyze the effect of TGF-β1 on cell morphology, pancreatic cancer cell lines were grown to 70% confluence in DMEM containing 10% FCS. Afterward, cells were washed twice in serum-free medium, starved for 24 h in serum-free medium, and finally treated for up to 72 h with TGF-β1 (10 ng/ml) or medium alone. For inhibition studies, the MEK1 inhibitor PD98059 was added at 20 μm concentrations 90 min before the application of TGF-β1. For electron microscopy, treated and untreated pancreatic cancer cells were fixed in 2% glutaraldehyde and 2% paraformaldehyde in 0.1 m cacodylate buffer (pH 7.4) for at least 90 min. In addition, cells were fixed in 1% osmium tetroxide for 1 h before dehydration in a graded ethanol series and embedded in Epon resin. Ultrathin sections were contrasted with uranyl acetate and lead citrate.

Northern Blot Analysis.

RNA was extracted using the RNAclean kit (AGS, Heidelberg, Germany) as indicated by the manufacturer. Northern blots with 30 μg of total RNA per lane were prepared and hybridized with 32P-labeled cDNA probes for TGF-β1, cytokeratins 8 and 19, vimentin, and collagen type α1/I, as described previously (18).

Western Blot Hybridization.

Protein expression of cytokeratins 8 and 19 and vimentin was measured in pancreatic cancer cell lines (PANC-1, COLO-357, and IMIM-PC1) before, 24, and 48 h after TGF-β1 treatment. Protein extraction was performed using an extraction buffer containing 10 mm Tris-HCl (pH 7.5), 2 m urea, 2 mm EDTA, 2 mm EGTA, and protease inhibitors. Equal amounts of protein were loaded onto SDS-polyacrylamide gels, size fractionated on SDS-PAGE gels, and transferred to a nitrocellulose membrane using the semidry technique. The membranes were blocked with 3% milk powder in TBS for 1 h and incubated with monoclonal antibodies against cytokeratin 8 (1:500), cytokeratin 19 (1:500), and vimentin (1:500). Afterward, membranes were washed and incubated with a secondary peroxidase-coupled antiserum. Detection of antibodies was performed by enhanced chemiluminescence.

Confocal Fluorescence Microscopy.

Pancreatic cancer cells (1 × 105 cells) were cultured on glass coverslips, serum starved, and treated with TGF-β1 for 48 h. After incubation, cells were fixed in −20°C methanol for 30 min, washed, and preincubated in blocking solution (5% goat serum, 5% glycerol, and 0.04% sodium azide). Cells were then costained with 1:100 diluted primary monoclonal antibodies recognizing cytokeratin 8/18 (guinea pig) and vimentin (mouse). After washing twice in PBS, cells were incubated with secondary FITC-labeled anti-pig IgG (green fluorescence for cytokeratin 8/18) and Cy3-labeled antimouse IgG (red fluorescence for vimentin). Cells were examined by confocal laser scanning microscopy (Zeiss LSM 510), and cells were also visualized by phase contrast microscopy.

Scattering Assay.

Pancreatic cancer cells were seeded at low density (2 × 104/60-mm dish) and allowed to grow as discrete colonies. After the formation of small colonies (5–10 cells), the medium was replaced by fresh medium containing 5% serum and 10 ng/ml TGF-β1. The effect of TGF-β1 on cell scattering was evaluated by phase contrast microscopy. Photographs were taken of the same field before, 4, 12, 18, and 24 h after TGF-β1 treatment.

In Vitro Invasion and Migration Assays.

Migration and invasion of pancreatic cancer cell lines were examined in a modified two-chamber assay as described previously by Albini et al.(19). For this purpose, 2–3 × 105 tumor cells were seeded on the upper side of 6-well Transwell plates (Costar; Cambridge, MA) uncoated (for migrations assays) or coated (for invasion assays) with Matrigel (Becton Dickinson, Heidelberg, Germany). Cells were treated either with TGF-β1 (10 ng/ml), MEK1 inhibitor PD 98059 (20 μm), or with a combination of both substances for 12, 24, and 48 h. After the incubation period, cells on the upper side of the membrane were wiped off, the membranes were fixed, and invading cells on the lower side of the membrane were stained and counted as described previously (20).

Cell Growth and Apoptosis Assay.

Cell growth assays were performed in the cell lines COLO-357, PANC-1, and IMIM-PC1. Cells were seeded at a low density (1 × 104/well) in 24-well plates and cultured in medium containing 10% FCS for 24 h. Cells were washed twice with PBS and then treated with medium containing 5% FCS, with medium alone, or with medium containing 5% FCS plus either 10 ng/ml TGF-β1, 20 μm PD 98059, or a combination of both. After 24 h, cells were counted using a hemocytometer and calculated as a percentage of control (100%). Growth assays were done in triplicate. Apoptosis was determined using the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling technique (Boehringer Mannheim, Indianapolis IN).

Immune Complex Kinase Assay for ERK2.

Pancreatic cancer cell lines were cultured in serum-free DMEM for 24 h, washed twice in DMEM, and treated with TGF-β1 for 5 min, 180 min, and 24 h. Subsequently, cells were lysed at 4°C in 1 ml of a solution containing 10 mm Tris/HCl (pH 7.6), 5 mm EDTA, 50 mm NaCl, 30 mm sodium PPi, 50 mm NaF, 2 mm Na3VO4, 1% Triton X-100, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1 mm phenylmethylsulfonyl fluoride (lysis buffer). Immunoprecipitation was performed as described previously (21) by using polyclonal ERK2 antibodies. After addition of protein A-agarose beads (40 μl; 1:1 slurry), immune complexes were collected by centrifugation and washed twice in lysis buffer and twice in kinase buffer [15 mm Tris-HCl (pH 7.4), 15 mm MgCl2]. The kinase reaction was performed by resuspending the pellet in 25 μl of kinase assay mixture containing kinase buffer, 100 μm ATP, 100 μCi/ml [γ-32P]ATP, 200 μm microcystin LR, and 1 mg/ml myelin basic protein. Incubations were performed for 10 min (linear assay conditions) at 30°C and terminated by spotting 20 μl of the supernatant onto P81 chromatography paper (Whatman). Filters were washed four times for 5 min in 0.5% ortho-phosphoric acid, immersed in acetone, and dried before Cerenkov counting. The average radioactivity of two blank samples containing no immune complex was subtracted from the result of each sample. The specific activity of [γ-32P]ATP used was 900-1200 cpm/pmol.

TGF-β1 Induces a Reversible EMT in Pancreatic Cancer Cell Lines with Functional TGF-β Pathway.

Cell morphology was analyzed before and after up to 72 h of TGF-β1 treatment in three cell lines with a functional TGF-β pathway (PANC-1, IMIM-PC1, and COLO-357), and in two cell lines lacking Smad4 expression (IMIM-PC2 and CAPAN-1; Refs. 22 and 23). All cell lines have activating Ki-Ras mutations. TGF-β1 treatment did not affect cell morphology of IMIM-PC2 and CAPAN-1 cells but induced an EMT in all three TGF-β-responsive cell lines (Fig. 1). Fig. 1 shows that upon treatment with TGF-β (10 ng/ml), responsive cells acquire a spindle type cell morphology and that the number of cell-cell contacts is reduced. These alterations started after 12 h and were maximal after 48 h of treatment. Interestingly, the changes in cellular differentiation were present, even 72 h after a single-dose application of TGF-β1, and were fully reversible within 24 h after removal of the growth factor from the culture medium. To further characterize the alterations of cell morphology, PANC-1 cells were studied by electron microscopy (Fig. 2 A). Treated cells were characterized by the presence of multiple electron-dense cytoplasmic fibers characteristic for cells of mesenchymal origin. In line with the light microscopic observations, electron microscopy showed a higher degree of dispersion of TGF-β1-treated as compared with untreated cells accompanied by a reduction of the number of cell-cell contacts.

In line with these morphological observations, EMT was characterized by repression of epithelial and induction of mesenchymal differentiation marker proteins. Fig. 2,B shows decreased immunostaining of cytokeratin as a marker protein for epithelial differentiation in TGF-β1-treated PANC-1 cells after 48 h. In contrast, the mesenchymal marker protein vimentin, a cytoskeletal intermediate filament protein, was weakly expressed before TGF-β1 treatment but highly expressed in transdifferentiated PANC-1 cells. Similar effects of TGF-β1 on cytokeratin and vimentin immunostaining were also observed in the responsive cell lines COLO-357 and IMIM-PC1 (data not shown). We next examined the effects of TGF-β1 on gene and protein expression of differentiation markers in all cell lines used in our study. As expected, TGF-β1 treatment resulted in reduced cytokeratins 8 and 19 expression but increased the expression of the mesenchymal markers vimentin and collagen type α1/I in COLO-357, PANC-1, and IMIM-PC1 cells (Fig. 3, A and B). In contrast, TGF-β1 treatment failed to induce any significant change in the expression pattern of differentiation markers in cell lines with a nonfunctional TGF-β1 signaling pathway (IMIM-PC2 and CAPAN-1). The TGF-β1-mediated changes of gene expression were also reproducibly observed on the protein level in all responsive cell lines in multiple experiments; however, down-regulation of epithelial marker proteins was less prominent on the RNA level (Fig. 3 C). As expected from the morphological studies, removal of TGF-β1 from the medium reversed these effects (data not shown).

TGF-β1-induced EMT Is Associated with Cell Scattering, Migration, and Tumor Cell Invasion.

To elucidate whether the phenotypic alterations were associated with functional changes, we measured cell scattering, migration, and invasion of pancreatic cancer cell lines before and after TGF-β1-induced EMT. To analyze the effect of TGF-β1 on cell scattering, we seeded pancreatic cancer cells at a very low density and allowed them to form colonies. After cells formed small complexes containing 5–10 cells, we treated the cells with 10 ng/ml TGF-β1 or medium alone. Fig. 4 shows that TGF-β1-induced cell scattering occurs exclusively in the responsive cell lines COLO-357, PANC-1, and IMIM-PC1. The effect is not observed in IMIM-PC2 and CAPAN-1 cells, in which TGF-β1 also failed to induce EMT. In conjunction with our electron microscopic data, these findings suggest that TGF-β1-induced EMT is of functional significance in pancreatic cancer and therefore prompted us to study the effects of TGF-β on critical oncogenic functions of tumor cells such as migration and invasion. Indeed, as shown in Fig. 5, TGF-β1 treatment resulted in a significant increase of tumor cell migration (Fig. 5,A) and invasion (Fig. 5 B) in the responsive cell lines COLO-357, PANC-1, and IMIM-PC1. The increase in tumor cell invasion and migration was time dependent and occurred in parallel with the morphological alterations. Tumor cell invasion and migration begun 12–16 h after treatment, reached statistical significance after 24 h, and continued increasing even 48 h after treatment. In contrast, TGF-β affected neither migration nor invasion in nonresponsive pancreatic cancer cell lines.

TGF-β1-mediated Morphological and Functional Changes Require ERK Signaling and Are Prevented by Inhibition of MEK1.

To examine whether the Ras-MEK-ERK signaling cascade is involved in the mediation of TGF-β-induced EMT, we studied the effect of the selective MEK1 inhibitor PD 98059, which inhibits ERK1 and ERK2 activity, on TGF-β-mediated changes of cell morphology and function. Pretreatment of responsive pancreatic cancer cell lines with PD 98059 significantly lowered the impact of TGF-β on cell morphology in all responsive cell lines. As shown in Fig. 6,A, application of PD 98059 dramatically reduced the potential of TGF-β1 to induce mesenchymal transdifferentiation of PANC-1 cells (Fig. 6,A). Interestingly, pretreatment of the cells with PD 98059 also appeared to prevent cell scattering (Fig. 6,A). In addition, pretreatment with the MEK1 inhibitor significantly attenuated the effects of TGF-β on changes of epithelial and mesenchymal differentiation marker expression (Fig. 6,B). Similar effects of the MEK1 inhibitor on cell morphology and gene expression were seen in COLO-357 and IMIM-PC1 cells (data not shown), suggesting that Ras-MEK-ERK signaling is required for TGF-β1 to exert maximal effects on the phenotype of pancreatic cancer cells. To examine whether the Ras-MEK-ERK signaling pathway also modulated TGF-β1-induced tumor cell migration and invasion, we measured both parameters in two-chamber assays after pretreatment with PD 98059. Because the effects of PD 98059 pretreatment on TGF-β-induced migration and invasion were almost identical, only the migration data are shown in Fig. 7. Application of PD 98059 alone did not significantly change the basal migratory and invasive potential of the different cell lines. However, it strongly attenuated TGF-β1-mediated tumor cell migration and invasion in PANC-1 and IMIM-PC1 cells (P < 0.001; Fig. 7,A). To exclude that the observed effects on migration and invasion were attributable to growth-inhibitory effects of the inhibitor, we measured the growth inhibition by PD 98059 alone and in combination with TGF-β1 in the responsive cell lines PANC-1, COLO-357, and IMIM-PC1. As demonstrated in Fig. 7 B, TGF-β1 (10 ng/ml) treatment for 24 h inhibited cell growth in all three responsive cell lines. Maximal values of ∼25% growth inhibition by TGF-β were observed in PANC-1 cells. The MEK inhibitor PD 98059 (20 μm) alone inhibited cell growth to a similar degree in all examined cell lines. However, combined treatment with 20 μm PD 98059 and 10 ng/ml TGF-β1 did not have an additive effect on cell growth inhibition. This suggests that inhibition of TGF-β1-mediated tumor cell migration by PD 98059 is not simply attributable to the growth-inhibitory effects of this MEK1 inhibitor. Because it has been shown previously that PD 98059 can induce apoptosis in a dose-dependent manner, we measured the number of cells in apoptosis at various doses of PD 98059. At 20 μm PD 98059, the concentration used to inhibit TGF-β1-induced EMT, the number of cells in apoptosis was not increased significantly (data not shown).

These data suggest that the Ras-MEK-ERK signaling pathway is required for TGF-β1 to induce morphological and functional changes in responsive pancreatic cancer cells. We therefore examined whether TGF-β1 can induce ERK activation in the responsive pancreatic cancer cells using immune complex kinase assays for ERK2. As shown in Fig. 8, after short treatment periods TGF-β1 had no effect on ERK2 activation in PANC-1 and IMIM-PC1 cells. However, after extended treatment periods both cell lines displayed a moderate (1.6–1.9-fold) and sustained increase in ERK2 activity. In contrast, TGF-β1 induced only a minor and transient activation of ERK2 in the nonresponsive cell line IMIM-PC-2 between 5 min and 3 h after the start of treatment (Fig. 8).

Progression of epithelial tumors to the metastatic stage is characterized by altered cellular plasticity, increased motility, down-regulation of cell-cell contacts and elevated expression and activation of matrix-degrading proteinases (8). It has been suggested that TGF-β is an important regulator of these processes associated with tumor progression (5, 24, 25). In view of the data implicating TGF-β in pancreatic tumorigenesis (1, 2, 4), the present study had two major aims: (a) to examine the effects of TGF-β1 treatment on the phenotype and biological behavior of pancreatic cancer cell lines with functional and nonfunctional TGF-β pathways; and (b) to elucidate whether the Ras signaling cascade participates in mediating the phenotypic and functional effects of TGF-β1 on pancreatic cancer cells.

Prolonged treatment with TGF-β1 induced an EMT of three pancreatic cancer cell lines with a functional TGF-β signaling pathway (PANC-1, IMIM-PC1, and COLO-357), leading to profound morphological changes. Because this effect could not be observed in cell lines lacking Smad4 (IMIM-PC2 and CAPAN-1), a functional TGF-β pathway appears to be an essential prerequisite. The epithelial to fibroblastoid conversion was time dependent and started 12 h after a single application of TGF-β1, became maximal after 72 h, and was reversible after removal of TGF-β1 from the medium. It was characterized by a strong down-regulation of epithelial differentiation markers and an up-regulation of mesenchymal markers, such as the intermediate filament protein vimentin (26). Similar effects of TGF-β on epithelial cell morphology have been observed after long-term treatment of keratinocytes, as well as in highly metastatic mouse mammary epithelial cells (24, 27, 28). Interestingly, in parallel to the acquisition of a spindle-shape cell morphology, we observed an enhanced scattering of pancreatic cancer cells, which was accompanied by a strong increase of tumor cell migration and invasion. Our data strongly support recent studies showing that TGF-β is a strong inducer of tumor cell invasion and metastasis in late-stage tumors. Overexpression of TGF-β is often seen in advanced stages of mouse and human carcinomas, such as pancreatic cancer (2, 29, 30). Transfection of a dominant-negative type II receptor into mouse keratinocytes blocked the formation of invasive spindle tumors in vivo and decreased tumorigenicity (24). In a previous study, we showed that in responsive pancreatic cancer cells, TGF-β1 increased the expression as well as the activation of matrix metalloproteinase-2 and the urokinase plasminogen activator system (31), which could contribute to enhanced invasion and metastasis in pancreatic tumors. Thus, data from the present and previous studies support the hypothesis that TGF-β1 contributes to tumor progression by inducing profound changes of cellular differentiation and function. In pancreatic cancer and other tumor models, this transdifferentiation appears to be associated with a switch from a sessile epithelial phenotype to a migratory mesenchymal phenotype (24, 27, 28, 32).

The signaling pathway by which TGF-β exerts its phenotypic effects is not well characterized. In a recent study, Oft et al.(12) demonstrated that induction of EMT of mouse mammary tumors cells by TGF-β is dependent on the presence of a mutated H-Ras gene. In fact, there is increasing evidence for cross-talk between TGF-β and the Ras signaling cascade (25). Ras has been shown to be activated by TGF-β and participates in TGF-β-mediated effects including negative growth control (33, 34). Furthermore, the downstream mediators of the TGF-β superfamily Smad2 and Smad3 can be phosphorylated via the Ras-MEK-ERK cascade (35, 36, 37). This cross-talk may be of importance for pancreatic carcinogenesis, in view of the fact that most pancreatic tumor cells display activating mutations of the Ki-Ras oncogene (14). It thus appears possible that an activation of the Ras-MEK-ERK cascade interferes with the induction of the specific gene program associated with TGF-β1-induced EMT. To obtain further evidence for this hypothesis, we studied ERK2 activation by TGF-β1 and the effect of MEK1 inhibition on the TGF-β1-induced phenotypic and functional changes of pancreatic cancer cells. MEK1 inhibition by PD 98059 prevented the TGF-β1-induced transdifferentiation of pancreatic cancer cells with activating mutations of the Ki-Ras oncogene. In the same way as the MEK1 inhibitor prevented EMT, it reduced the TGF-β1-induced increase of tumor cell scattering, migration and invasion. Moreover, in pancreatic cancer cell lines with an intact TGF-β signaling pathway, TGF-β1 induced a sustained increase in ERK2 activity. The TGF-β1-mediated increase in ERK activity was observed 24 h after TGF-β1 treatment and might either reflect a direct effect of TGF-β1 or an indirect effect through activation of growth factors acting in an autocrine manner. Nevertheless, in conjunction with the data obtained from the MEK inhibitor PD 98059, these observations strongly suggest that TGF-β exerts its influence on both tumor cell morphology and oncogenic functions as migration and invasion via an interaction with the Ras-MEK-ERK signaling cascade. Mutations resulting in constitutive activation of Ras are found in the majority of pancreatic cancer cell lines and may thus be a prerequisite for the remarkable morphological and functional alterations induced by TGF-β1 in our study.

In summary, our data provide evidence that in pancreatic cancer cells with a functional TGF-β pathway and with activating Ki-Ras mutations, TGF-β1 induces a specific gene program that is associated with an EMT. This EMT leads to increased tumor cell scattering, migration, and invasion and may thus be one of the mechanisms by which TGF-β contributes to tumor progression. Cross-talk between the TGF-β pathway and the Ras-MEK-ERK cascade appears to be essential to mediate these effects of TGF-β on the phenotype of pancreatic tumor cells.

Fig. 1.

Phase-contrast photomicrographs documenting the effect of a 48-h TGF-β1 treatment on the phenotype of TGF-β-responsive (COLO-357, IMIM-PC1, and PANC-1) and nonresponsive (IMIM-PC2 and CAPAN-1) pancreatic cancer cell lines. TGF-β1 treatment induced an EMT in COLO-357, IMIM-PC1, and PANC-1 but did not affect cell morphology of IMIM-PC2 and CAPAN-1 cells. Bars, 30 μm.

Fig. 1.

Phase-contrast photomicrographs documenting the effect of a 48-h TGF-β1 treatment on the phenotype of TGF-β-responsive (COLO-357, IMIM-PC1, and PANC-1) and nonresponsive (IMIM-PC2 and CAPAN-1) pancreatic cancer cell lines. TGF-β1 treatment induced an EMT in COLO-357, IMIM-PC1, and PANC-1 but did not affect cell morphology of IMIM-PC2 and CAPAN-1 cells. Bars, 30 μm.

Close modal
Fig. 2.

A, electron microscopic examination of TGF-β1-treated and untreated PANC-1 cells. TGF-β1-treated, transdifferentiated cells expressed multiple electron-dense cytoplasmic fibers (arrows) characteristic for cells of mesenchymal origin. ×8000. B, confocal images showing double staining of TGF-β1-treated (+TGFβ) or untreated (control) PANC-1 cells with antibodies against cytokeratin 8/18 and vimentin. Cells were also visualized by phase contrast microscopy. Bars, 10 μm.

Fig. 2.

A, electron microscopic examination of TGF-β1-treated and untreated PANC-1 cells. TGF-β1-treated, transdifferentiated cells expressed multiple electron-dense cytoplasmic fibers (arrows) characteristic for cells of mesenchymal origin. ×8000. B, confocal images showing double staining of TGF-β1-treated (+TGFβ) or untreated (control) PANC-1 cells with antibodies against cytokeratin 8/18 and vimentin. Cells were also visualized by phase contrast microscopy. Bars, 10 μm.

Close modal
Fig. 3.

Expression pattern of epithelial and mesenchymal differentiation markers by Northern blot (A and B) and Western blot analyses (C). A, TGF-β1 stimulation leads to a down-regulation of cytokeratins 8 and 19 mRNA levels in PANC-1 and IMIM-PC1 cells and to a minor extent in COLO-357 cells. The mesenchymal differentiation marker vimentin is strongly up-regulated in all responsive cell lines. B, the TGF-β1-mediated regulation of differentiation markers occurs in a time-dependent manner, as demonstrated by Northern blot analysis for IMIM-PC1 cells. The 28S ethidium bromide band is shown as loading control. C, TGF-β1-induced effects on differentiation markers were reproducible on Western blots, although down-regulation of cytokeratin 8 and 19 protein was less prominent than on the mRNA level. Representative results obtained in triple experiments for PANC-1 cells are shown. Similar results were obtained with COLO-357 and IMIM-PC1 cells. Actin is shown as a loading control.

Fig. 3.

Expression pattern of epithelial and mesenchymal differentiation markers by Northern blot (A and B) and Western blot analyses (C). A, TGF-β1 stimulation leads to a down-regulation of cytokeratins 8 and 19 mRNA levels in PANC-1 and IMIM-PC1 cells and to a minor extent in COLO-357 cells. The mesenchymal differentiation marker vimentin is strongly up-regulated in all responsive cell lines. B, the TGF-β1-mediated regulation of differentiation markers occurs in a time-dependent manner, as demonstrated by Northern blot analysis for IMIM-PC1 cells. The 28S ethidium bromide band is shown as loading control. C, TGF-β1-induced effects on differentiation markers were reproducible on Western blots, although down-regulation of cytokeratin 8 and 19 protein was less prominent than on the mRNA level. Representative results obtained in triple experiments for PANC-1 cells are shown. Similar results were obtained with COLO-357 and IMIM-PC1 cells. Actin is shown as a loading control.

Close modal
Fig. 4.

TGF-β1-induced cell scattering. Pancreatic cancer cells were allowed to grow as discrete colonies by seeding 2 × 104 cells per 60-mm dish. After 36 h, the medium was replaced by fresh medium containing 5% serum and 10 ng/ml TGF-β1. The effect of TGF-β1 on cell scattering was evaluated by comparing phase contrast microscopy of the same cell colony before treatment 0 h (0h) and 12 h (12h) after treatment with TGF-β1. Bars, 15 μm.

Fig. 4.

TGF-β1-induced cell scattering. Pancreatic cancer cells were allowed to grow as discrete colonies by seeding 2 × 104 cells per 60-mm dish. After 36 h, the medium was replaced by fresh medium containing 5% serum and 10 ng/ml TGF-β1. The effect of TGF-β1 on cell scattering was evaluated by comparing phase contrast microscopy of the same cell colony before treatment 0 h (0h) and 12 h (12h) after treatment with TGF-β1. Bars, 15 μm.

Close modal
Fig. 5.

TGF-β1-mediated pancreatic cancer cell migration (A) and invasion (B) were measured in modified two-chamber assays. Tumor cells (2–3 × 105) were seeded on the upper side of uncoated (for migrations assays) or coated (for invasion assays) six-well Transwell plates for 24 h, serum starved, and finally treated either with or without TGF-β1 (10 ng/ml). After treatment, the membrane was fixed, and penetrated cells were stained and counted. TGF-β1-induced tumor cell migration (A) and invasion (B) were measured after 48 and 24 h, respectively, and shown as the number of cells/field (cm2). Data are expressed as the mean number of penetrated cells (significantly different at *P < 0.0005; **P < 0.001 and +P < 0.01); bars, SE. TGF-β1 significantly increased tumor cell migration and invasion in COLO-357, PANC-1, and IMIM-PC1 cells but did neither affect cell migration nor the invasive potential of IMIM-PC2 or CAPAN-1 cells. Both migration and invasion assays were done at least in triplicate. Statistical analyses were done using Student’s t test.

Fig. 5.

TGF-β1-mediated pancreatic cancer cell migration (A) and invasion (B) were measured in modified two-chamber assays. Tumor cells (2–3 × 105) were seeded on the upper side of uncoated (for migrations assays) or coated (for invasion assays) six-well Transwell plates for 24 h, serum starved, and finally treated either with or without TGF-β1 (10 ng/ml). After treatment, the membrane was fixed, and penetrated cells were stained and counted. TGF-β1-induced tumor cell migration (A) and invasion (B) were measured after 48 and 24 h, respectively, and shown as the number of cells/field (cm2). Data are expressed as the mean number of penetrated cells (significantly different at *P < 0.0005; **P < 0.001 and +P < 0.01); bars, SE. TGF-β1 significantly increased tumor cell migration and invasion in COLO-357, PANC-1, and IMIM-PC1 cells but did neither affect cell migration nor the invasive potential of IMIM-PC2 or CAPAN-1 cells. Both migration and invasion assays were done at least in triplicate. Statistical analyses were done using Student’s t test.

Close modal
Fig. 6.

Prevention of TGF-β1-induced EMT by the MEK1 inhibitor PD 98059. A, phase-contrast photomicrographs show the preventive effect of MEK1 inhibition on TGF-β1-induced EMT in PANC-1 cells. Cells (5 × 105) were plated in 60-mm dishes and incubated in DMEM containing 10% FCS. After serum starvation, cells were either treated with medium alone (a), 10 ng/ml TGF-β1 (b), 20 μm PD 98059 for 90 min before application of 10 ng/ml TGF-β1 (c), or PD 98059 alone for 24 h (d). Bars, 15 μm. B, effect of the selective MEK1 inhibitor PD 98059 on the mRNA expression of differentiation markers in the cell line PANC-1 after 48 h. TGF-β1-induced down-regulation of cytokeratins 8 and 19 in PANC-1 was attenuated or blocked by pretreatment with PD 98059. In a similar manner, PD 98059 reduced the up-regulation of vimentin by TGF-β1, even 48 h after the application of TGF-β1 and the MEK1 inhibitor. Similar effects of PD 98059 on differentiation markers were observed in the cell lines COLO-357 and IMIM-PC1 (not shown). The 28S rRNA ethidium bromide band is included as loading control.

Fig. 6.

Prevention of TGF-β1-induced EMT by the MEK1 inhibitor PD 98059. A, phase-contrast photomicrographs show the preventive effect of MEK1 inhibition on TGF-β1-induced EMT in PANC-1 cells. Cells (5 × 105) were plated in 60-mm dishes and incubated in DMEM containing 10% FCS. After serum starvation, cells were either treated with medium alone (a), 10 ng/ml TGF-β1 (b), 20 μm PD 98059 for 90 min before application of 10 ng/ml TGF-β1 (c), or PD 98059 alone for 24 h (d). Bars, 15 μm. B, effect of the selective MEK1 inhibitor PD 98059 on the mRNA expression of differentiation markers in the cell line PANC-1 after 48 h. TGF-β1-induced down-regulation of cytokeratins 8 and 19 in PANC-1 was attenuated or blocked by pretreatment with PD 98059. In a similar manner, PD 98059 reduced the up-regulation of vimentin by TGF-β1, even 48 h after the application of TGF-β1 and the MEK1 inhibitor. Similar effects of PD 98059 on differentiation markers were observed in the cell lines COLO-357 and IMIM-PC1 (not shown). The 28S rRNA ethidium bromide band is included as loading control.

Close modal
Fig. 7.

A, TGF-β1-induced EMT and tumor cell migration is prevented by the MEK1 inhibitor PD 98059. TGF-β1-responsive pancreatic cancer cells (COLO-357, IMIM-PC1, and PANC-1) were seeded on the upper side of uncoated six-well Transwell plates for 24 h, serum starved, and treated with serum-free medium, 10 ng/ml TGF-β1, 20 μm PD 98059, or pretreated with PD 98059 for 90 min before treatment with 10 ng/ml TGF-β1. TGF-β1-induced tumor cell migration is shown as the number of cells/field (cm2). The application of the MEK inhibitor PD 98059 did not significantly alter baseline invasive potential but strongly reduced TGF-β1-mediated effects on tumor cell migration in PANC-1 and IMIM-PC1 cells (∗, P < 0.001) and to a lower extent in COLO-357. The experiment was done in triplicate. Means are shown; bars, SE. Statistical analyses were done using Student’s t test. Invasion assays are not shown because they yielded similar results. B, effect of TGF-β1 and PD 98059 on tumor cell growth. Cells were plated in 24-well plates at an initial density of 1 × 104 cells/well and either treated with medium containing 5% FCS, medium alone (control), or with medium containing 10 ng/ml TGF-β1, 20 μm PD 98059, or a combination of both. Cell numbers were counted after 24 h using a hemocytometer and shown in percentage of control. Data shown represent mean numbers; bars, SE. Growth assays were done in triplicate.

Fig. 7.

A, TGF-β1-induced EMT and tumor cell migration is prevented by the MEK1 inhibitor PD 98059. TGF-β1-responsive pancreatic cancer cells (COLO-357, IMIM-PC1, and PANC-1) were seeded on the upper side of uncoated six-well Transwell plates for 24 h, serum starved, and treated with serum-free medium, 10 ng/ml TGF-β1, 20 μm PD 98059, or pretreated with PD 98059 for 90 min before treatment with 10 ng/ml TGF-β1. TGF-β1-induced tumor cell migration is shown as the number of cells/field (cm2). The application of the MEK inhibitor PD 98059 did not significantly alter baseline invasive potential but strongly reduced TGF-β1-mediated effects on tumor cell migration in PANC-1 and IMIM-PC1 cells (∗, P < 0.001) and to a lower extent in COLO-357. The experiment was done in triplicate. Means are shown; bars, SE. Statistical analyses were done using Student’s t test. Invasion assays are not shown because they yielded similar results. B, effect of TGF-β1 and PD 98059 on tumor cell growth. Cells were plated in 24-well plates at an initial density of 1 × 104 cells/well and either treated with medium containing 5% FCS, medium alone (control), or with medium containing 10 ng/ml TGF-β1, 20 μm PD 98059, or a combination of both. Cell numbers were counted after 24 h using a hemocytometer and shown in percentage of control. Data shown represent mean numbers; bars, SE. Growth assays were done in triplicate.

Close modal
Fig. 8.

Immune complex kinase assays for ERK2 in pancreatic cancer cell lines after treatment with TGF-β1. Basal activity ranged between 1500 and 3000 cpm in all examined cell lines with no significant variations between the individual cell lines. Because treatment with PD 98059 reduces basal activity in all cell lines, some degree of constitutive activation of the ERKs has to be assumed. Cultured cells were treated with 10 ng/ml TGF-β1 for 5 min, 3 h, and 24 h. TGF-β1-induced ERK2 activation is shown as an increase over unstimulated controls in percentages. After extended treatment periods with TGF-β1, both responsive cell lines (PANC-1 and IMIM-PC1) displayed a moderate (1.6–1.9-fold) and sustained increase of ERK2 activity. In contrast, TGF-β1 induced only a minor and transient activation of ERK2 in IMIM-PC-2, a cell line with a homozygous Smad4 deletion, between 5 min and 3 h after the start of treatment. Experiments were done in triplicate, and the results shown in this figure are representative for three independent experiments; bars, SE.

Fig. 8.

Immune complex kinase assays for ERK2 in pancreatic cancer cell lines after treatment with TGF-β1. Basal activity ranged between 1500 and 3000 cpm in all examined cell lines with no significant variations between the individual cell lines. Because treatment with PD 98059 reduces basal activity in all cell lines, some degree of constitutive activation of the ERKs has to be assumed. Cultured cells were treated with 10 ng/ml TGF-β1 for 5 min, 3 h, and 24 h. TGF-β1-induced ERK2 activation is shown as an increase over unstimulated controls in percentages. After extended treatment periods with TGF-β1, both responsive cell lines (PANC-1 and IMIM-PC1) displayed a moderate (1.6–1.9-fold) and sustained increase of ERK2 activity. In contrast, TGF-β1 induced only a minor and transient activation of ERK2 in IMIM-PC-2, a cell line with a homozygous Smad4 deletion, between 5 min and 3 h after the start of treatment. Experiments were done in triplicate, and the results shown in this figure are representative for three independent experiments; bars, SE.

Close modal

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.

1

This work was supported by Grant SFB 518/Project B1 from the Deutsches Forschungsgemeinschaft and Grant BMH4-CT98-3085 from the European Community.

3

The abbreviations used are: TGF, transforming growth factor; TGFβRII, TGF-β receptor type II; EMT, epithelial-mesenchymal transdifferentiation; ERK, extracellular signal-regulated kinase.

1
Friess H., Yamanaka Y., Buchler M., Berger H. G., Kobrin M. S., Baldwin R. L., Korc M. Enhanced expression of the type II transforming growth factor β receptor in human pancreatic cancer cells without alteration of type III receptor expression.
Cancer Res.
,
53
:
2704
-2707,  
1993
.
2
Friess H., Yamanaka Y., Buchler M., Ebert M., Beger H. G., Gold L. I., Korc M. Enhanced expression of transforming growth factor β isoforms in pancreatic cancer correlates with decreased survival.
Gastroenterology
,
105
:
1846
-1856,  
1993
.
3
Hahn S. A., Schutte M., Hoque A. T., Moskaluk C. A., da Costa C. L., Rozenblum E., Weinstein C. L., Fischer A., Yeo C. J., Hruban R. H., Kern S. E. DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1.
Science (Wash. DC)
,
271
:
350
-353,  
1996
.
4
Goggins M., Shekher M., Turnacioglu K., Yeo C. J., Hruban R. H., Kern S. E. Genetic alterations of the transforming growth factor β receptor genes in pancreatic and biliary adenocarcinomas.
Cancer Res.
,
58
:
5329
-5332,  
1998
.
5
Markowitz S. D., Roberts A. B. Tumor suppressor activity of the TGF-β pathway in human cancers.
Cytokine Growth Factor Rev.
,
7
:
93
-102,  
1996
.
6
Massagué J., Wotton D. Transcriptional control by the TGF-β/Smad signaling system.
EMBO J.
,
19
:
1745
-1754,  
2000
.
7
Massagué J., Chen Y. G. Controlling TGF-β signaling.
Genes Dev.
,
14
:
627
-644,  
2000
.
8
Ellenrieder V., Adler G., Gress T. M. Invasion and metastasis in pancreatic cancer.
Ann. Oncol.
,
10 (Suppl. 4)
:
S46
-S50,  
1999
.
9
Welch D. R., Fabra A., Nakajima M. Transforming growth factor β stimulates mammary adenocarcinoma cell invasion and metastatic potential.
Proc. Natl. Acad. Sci. USA
,
87
:
7678
-7682,  
1990
.
10
Wikstrom P., Stattin P., Franck L. I., Damber J. E., Bergh A. Transforming growth factor β1 is associated with angiogenesis, metastasis, and poor clinical outcome in prostate cancer.
Prostate
,
37
:
19
-29,  
1998
.
11
Cui W., Fowlis D. J., Bryson S., Duffie E., Ireland H., Balmain A., Akhurst R. J. TGFβ1 inhibits the formation of benign skin tumors, but enhances progression to invasive spindle carcinomas in transgenic mice.
Cell
,
86
:
531
-542,  
1996
.
12
Oft M., Peli J., Rudaz C., Schwarz H., Beug H., Reichmann E. TGF-β1 and Ha-Ras collaborate in modulating the phenotypic plasticity and invasiveness of epithelial tumor cells.
Genes Dev.
,
10
:
2462
-2477,  
1996
.
13
Malats N., Porta M., Corominas J. M., Pinol J. L., Rifa J., Real F. X. Ki-ras mutations in exocrine pancreatic cancer: association with clinico-pathological characteristics and with tobacco and alcohol consumption. PANK-ras I Project Investigators.
Int. J. Cancer
,
70
:
661
-667,  
1997
.
14
Lemoine N. R., Jain S., Hughes C. M., Staddon S. L., Maillet B., Hall P. A., Kloppel G. Ki-ras oncogene activation in preinvasive pancreatic cancer.
Gastroenterology
,
102
:
230
-236,  
1992
.
15
Vila M. R., Lloreta J., Schussler M. H., Berrozpe G., Welt S., Real F. X. New pancreas cancers cell lines that represent distinct stages of ductal differentiation.
Lab. Investig.
,
72
:
395
-404,  
1995
.
16
Morgan R. T., Woods L. K., Moore G. E., Quinn L. A., McGavran L., Gordon S. G. Human cell line (COLO-357) of metastatic pancreatic adenocarcinoma.
Int. J. Cancer
,
25
:
591
-598,  
1980
.
17
Geng M., Ellenrieder V., Wallrapp C., Müller-Pillasch F., Sommer G., Adler G. Identification of TGFβ1 target genes in pancreatic cancer cells by cDNA representational difference analysis.
Genes Chromosomes Cancer
,
26
:
70
-79,  
1999
.
18
Gress T. M., Muller P. F., Lerch M. M., Friess H., Buchler M., Adler G. Expression and in-situ localization of genes coding for extracellular matrix proteins and extracellular matrix degrading proteases in pancreatic cancer.
Int. J. Cancer
,
62
:
407
-413,  
1995
.
19
Albini A., Iwamoto Y., Kleinman H. K., Martin G. R., Aaronson S. A., Kozlowski J. M., McEwan R. N. A rapid in vitro assay for quantitating the invasive potential of tumor cells.
Cancer Res.
,
47
:
3239
-3245,  
1987
.
20
Ellenrieder V., Alber B., Lacher U., Hendler S., Menke A., Wenger C., Boeck W., Wagner M., Wilda M., Friess H., Buchler M., Adler G., Gress T. M. The role of MT-MMPs and MMP-2 in the progression of pancreatic cancer.
Int. J. Cancer
,
85
:
14
-20,  
2000
.
21
Seufferlein T., van Lint J., Liptay S., Adler G., Schmid R. M. Transforming growth factor α activates Ha-Ras in human pancreatic cancer cells: their role for signal transduction and mitogenesis.
Gastroenterology
,
116
:
1
-14,  
1999
.
22
Wenger C., Ellenrieder V., Alber B., Lacher U., Menke A., Hameister H., Wilda M., Iwamura T., Beger H. G., Adler G., Gress T. M. Expression and differential regulation of connective tissue growth factor in pancreatic cancer cells.
Oncogene
,
18
:
1073
-1080,  
1999
.
23
Baldwin R. L., Friess H., Yokoyama M., Lopez M. E., Kobrin M. S., Buchler M. W., Korc M. Attenuated ALK5 receptor expression in human pancreatic cancer: correlation with resistance to growth inhibition.
Int. J. Cancer
,
67
:
283
-288,  
1996
.
24
Portella G., Cumming S. A., Liddell J., Cui W., Ireland H., Akhurst R. J., Balmain A. Transforming growth factor β is essential for spindle cell conversion of mouse skin carcinoma in vivo: implications for tumor invasion.
Cell Growth Differ.
,
9
:
393
-404,  
1998
.
25
Akhurst R. J., Balmain A. Genetic events and the role of TGF β in epithelial tumour progression.
J. Pathol.
,
187
:
82
-90,  
1999
.
26
Dumortier J., Daemi N., Pourreyron C., Anderson W., Bellaton C., Jacquier M. F., Bertrand S., Chayvialle J. A., Remy L. Loss of epithelial differentiation markers and acquisition of vimentin expression after xenograft with laminin-1 enhance migratory and invasive abilities of human colon cancer cells LoVo C5.
Differentiation
,
63
:
141
-150,  
1998
.
27
Caulin C., Scholl F. G., Frontelo P., Gamallo C., Quintanilla M. Chronic exposure of cultured transformed mouse epidermal cells to transforming growth factor-β1 induces an epithelial-mesenchymal transdifferentiation and a spindle tumoral phenotype.
Cell Growth Differ.
,
6
:
1027
-1035,  
1995
.
28
Miettinen P. J., Ebner R., Lopez A. R., Derynck R. TGF-β induced transdifferentiation of mammary epithelial cells to mesenchymal cells: involvement of type I receptors.
J. Cell Biol.
,
127
:
2021
-2036,  
1994
.
29
Arrick B. A., Lopez A. R., Elfman F., Ebner R., Damsky C. H., Derynck R. Altered metabolic and adhesive properties and increased tumorigenesis associated with increased expression of transforming growth factor β1.
J. Cell Biol.
,
118
:
715
-726,  
1992
.
30
Derynck R., Jarrett J. A., Chen E. Y., Eaton D. H., Bell J. R., Assoian R. K., Roberts A. B., Sporn M. B., Goeddel D. V. Human transforming growth factor-β complementary DNA sequence and expression in normal and transformed cells.
Nature (Lond.)
,
316
:
701
-705,  
1985
.
31
Hendler S., Ellenrieder V., Alber B., Lacher U., Adler G., Gress T. M. TGFβ induced invasiveness of pancreatic cancer cell lines is mediated by upregulation and activation of both the MMP-2 and the u-PA enzymatic system.
Gastroenterology
,
116
:
A398
-A399,  
2000
.
32
Oft M., Heider K. H., Beug H. TGFβ signaling is necessary for carcinoma cell invasiveness and metastasis.
Curr. Biol.
,
8
:
1243
-1252,  
1998
.
33
Hartsough M. T., Frey R. S., Zipfel P. A., Buard A., Cook S. J., McCormick F., Mulder K. M. Altered transforming growth factor signaling in epithelial cells when ras activation is blocked.
J. Biol. Chem.
,
271
:
22368
-22375,  
1996
.
34
Frey R. S., Mulder K. M. TGFβ regulation of mitogen-activated protein kinases in human breast cancer cells.
Cancer Lett.
,
117
:
41
-50,  
1997
.
35
Kretzschmar M., Doody J., Timokhina I., Massagué J. A mechanism of repression of TGFβ/Smad signaling by oncogenic Ras.
Genes Dev.
,
13
:
804
-816,  
1999
.
36
Yue J., Hartsough M. T., Frey R. S., Frielle T., Mulder K. M. Cloning and expression of a rat Smad1: regulation by TGFβ and modulation by the Ras/MEK pathway.
J. Cell Physiol.
,
178
:
387
-396,  
1999
.
37
de Caestecker M. P., Parks W. T., Frank C. J., Castagnino P., Bottaro D. P., Roberts A. B., Lechleider R. J. Smad2 transduces common signals from receptor serine-threonine and tyrosine kinases.
Genes Dev.
,
12
:
1587
-1592,  
1998
.