Thrombin differentiates normal lung fibroblasts to a myofibroblast phenotype via the proteolytically activated receptor-1 and a protein kinase C-dependent pathway.

Myofibroblasts are ultrastructurally and metabolically distinctive fibroblasts that express smooth muscle (SM)-alpha actin and are associated with various fibrotic lesions. The present study was undertaken to investigate the myofibroblast phenotype that appears after activation of normal lung fibroblasts by thrombin. We demonstrate that thrombin induces smooth muscle-alpha actin expression and rapid collagen gel contraction by normal lung fibroblasts via the proteolytically activated receptor-1 and independent of transforming growth factor-beta pathway. Using antisense oligonucleotides we demonstrate that a decreased level of PKCepsilon abolishes SM-alpha actin expression and collagen gel contraction induced by thrombin in normal lung fibroblasts. Inhibition of PKCepsilon translocation also abolishes thrombin-induced collagen gel contraction, SM-alpha actin increase, and its organization by normal lung fibroblasts, suggesting that activation of PKCepsilon is required for these effects. In normal lung fibroblasts PKCepsilon binds to SM-alpha actin after thrombin treatment, but in activated fibroblasts derived from scleroderma lung they associate even in untreated cells. This suggests that SM-alpha actin may serve as a substrate for PKCepsilon in lung fibroblasts when activated by thrombin. We propose that thrombin differentiates normal lung fibroblasts to a myofibroblast phenotype via a PKC-dependent pathway. Thrombin-induced differentiation of normal lung fibroblasts to a myofibroblast phenotype resembles the phenotype observed in scleroderma lung fibroblasts. Therefore, we conclude that chronic exposure to thrombin after microvascular injury leads to activation of normal lung fibroblasts and to the appearance of a myofibroblast phenotype in vivo. Our study provides novel, compelling evidence that thrombin is an important mediator of the interstitial lung fibrosis associated with scleroderma.

Myofibroblasts are ultrastructurally and metabolically distinctive fibroblasts that express smooth muscle (SM)-␣ actin and are associated with various fibrotic lesions. The present study was undertaken to investigate the myofibroblast phenotype that appears after activation of normal lung fibroblasts by thrombin. We demonstrate that thrombin induces smooth muscle-␣ actin expression and rapid collagen gel contraction by normal lung fibroblasts via the proteolytically activated receptor-1 and independent of transforming growth factor-␤ pathway. Using antisense oligonucleotides we demonstrate that a decreased level of PKC⑀ abolishes SM-␣ actin expression and collagen gel contraction induced by thrombin in normal lung fibroblasts. Inhibition of PKC⑀ translocation also abolishes thrombininduced collagen gel contraction, SM-␣ actin increase, and its organization by normal lung fibroblasts, suggesting that activation of PKC⑀ is required for these effects. In normal lung fibroblasts PKC⑀ binds to SM-␣ actin after thrombin treatment, but in activated fibroblasts derived from scleroderma lung they associate even in untreated cells. This suggests that SM-␣ actin may serve as a substrate for PKC⑀ in lung fibroblasts when activated by thrombin. We propose that thrombin differentiates normal lung fibroblasts to a myofibroblast phenotype via a PKC-dependent pathway. Thrombin-induced differentiation of normal lung fibroblasts to a myofibroblast phenotype resembles the phenotype observed in scleroderma lung fibroblasts. Therefore, we conclude that chronic exposure to thrombin after microvascular injury leads to activation of normal lung fibroblasts and to the appearance of a myofibroblast phenotype in vivo. Our study provides novel, compelling evidence that thrombin is an important mediator of the interstitial lung fibrosis associated with scleroderma.
The presence of fibroblasts expressing smooth muscle-␣ actin (SM-␣ actin), 1 called myofibroblasts has been extensively documented in active fibrotic lesions in many diseases, including pulmonary fibrosis (1)(2)(3)(4)(5). Myofibroblasts are ultrastructurally and metabolically distinctive connective tissue cells identified as a key participant in tissue remodeling, wound healing, and various fibrotic disorders (6,7). They contribute to the increase of extracellular matrix deposition and contractility of lung parenchyma associated with pulmonary fibrosis (1,8). Studies on bleomycin-induced pulmonary fibrosis identified myofibroblasts as a primary source of increased collagen expression and a major source of cytokines and chemokines (9,10). We have demonstrated that myofibroblasts are present in bronchoalveolar lavage fluid of scleroderma (SSc) patients with active lung disease, and that SSc lung myofibroblasts express more collagen I, III, and fibronectin than normal lung fibroblasts (11). They show a greater proliferative response upon exposure to transforming growth factor-␤ (TGF-␤) and platelet-derived growth factor than do normal lung fibroblasts (11,12). Recently, we have shown myofibroblasts to be present in the interstitium of lung tissue from scleroderma patients with active pulmonary fibrosis, where a large amount of extracellular matrix is deposited (13). The factors responsible for the differentiation of normal lung fibroblasts to a myofibroblast phenotype are not well known, although TGF-␤ has been postulated to participate in such transition (3,6,14).
Another mediator of lung fibroblast activation is thrombin, a multifunctional serine protease and G protein-coupled receptor ligand, which is generated immediately at sites of vascular injury (15,16). Thrombin is well known for its role in hemostasis and thrombosis, and it also induces a wide range of cellular responses associated with both normal and disease processes. Evidence suggests that thrombin is an important mediator of the interstitial lung fibrosis accompanied by microvascular injury associated with SSc (17)(18)(19). It has been demonstrated that thrombin activity is significantly greater in bronchoalveolar lavage fluid from SSc patients compared with healthy controls (17,20). Thrombin is mitogenic for lung fibroblasts (17,21), and enhances the proliferative effect of fibrinogen on fibroblasts (22). Thrombin is a potent inducer of fibrogenic cytokines, such as TGF-␤ (23), platelet-derived growth factor-AA (17), chemokines (18), and extracellular matrix proteins such as collagen, fibronectin, and tenascin in mesenchymal cells, including lung fibroblasts (19,(23)(24)(25)(26).
Protein kinase C (PKC) signal transduction has been impli-cated in many cellular responses to thrombin (15,18,19,(27)(28)(29). Recently we have shown that PKC␥ is involved in the increased expression of interleukin-8 by lung fibroblasts, while PKC⑀ regulates thrombin-induced tenascin-C expression in lung fibroblasts (18,19). We have demonstrated that thrombin promotes PKC⑀ expression in normal lung fibroblasts, yet inhibits PKC⑀ expression in SSc cells. Subcellular localization of PKC⑀ is also markedly different in both cell types (19). Therefore, thrombin may trigger distinct PKC signaling in normal and SSc lung fibroblasts, and the differences in signaling may be responsible for some of the different behaviors of these cell types.
The present study was undertaken to investigate the role of thrombin in lung fibroblast activation and the signaling pathway(s) by which thrombin affects lung fibroblast behaviors. Upon determining that thrombin induces smooth muscle-␣ actin and collagen gel contraction, two characteristics of myofibroblasts, we examined the cellular mechanisms mediating differentiation of normal lung fibroblasts to a myofibroblast phenotype. We demonstrate that PKC⑀ regulates the transition of normal lung fibroblasts to myofibroblasts by binding to SM-␣ actin upon exposure to thrombin. Interestingly, in SSc lung fibroblasts PKC⑀ and SM-␣ actin are associated even in the untreated state. We also show that another feature of myofibroblasts, their proliferative capacity, is mediated by classical PKC isoforms.

EXPERIMENTAL PROCEDURES
Reagents-Antisense (AS) and sense (S) oligonucleotides for PKC⑀ were synthesized in the Oligonucleotide Synthesis Facility at the Medical University of South Carolina. The sequences were as follows: PKC⑀ antisense 5Ј-ATTGAACACTACCAT-3Ј and sense 5Ј-ATGGTAGTGT-TCCAAT-3Ј. Synthetic peptides TRAP-6 (SFLLRN) and F-14 (LLYPP-WNKNFTEND) were synthesized employing the tBOC method as described previously (18). Thrombin from human plasma and protein kinase C⑀ translocation inhibitor peptide were obtained from Calbiochem, La Jolla, CA. [ 3 H]Thymidine (specific activity 82.1 Ci/mmol) was purchased from PerkinElmer Life Sciences, Boston, MA.
Cell Culture-Lung fibroblasts were derived from lung tissues obtained at autopsy from scleroderma patients and from age-, race-, and sex-matched normal subjects. Lung tissue was diced (0.5 ϫ 0.5 mm pieces) and cultured in Dulbecco's modified Eagle's medium (DMEM; Life Technologies, Inc., Grand Island, NY) supplemented with 10% fetal calf serum, 2 mM L-glutamine, 50 g/ml gentamicin sulfate, and 5 g/ml amphotericin B at 37°C in 10% CO 2 . Medium was changed in every 3 days to remove dead and nonattached cells until fibroblasts reached confluency. Monolayer cultures were maintained in the same medium. Lung fibroblasts were used between the second and fourth passages in all experiments. Purity of isolated lung fibroblasts was determined by crystal violet staining (19) and by immunofluorescent staining (monoclonal antibody against human fibroblasts as described previously (11) followed by fluorescein isothiocyanate-conjugated goat anti-mouse IgG staining) (Santa Cruz Biotechnology Inc., Santa Cruz, CA).

Western Blot Analysis of Smooth Muscle-␣ Actin in Normal and SSc
Lung Fibroblasts-Normal and SSc lung fibroblasts were stimulated with various concentration of thrombin (0.001-0.5 units/ml) for 24 h, or with 0.5 unit/ml for different times, solubilized using sample buffer, and boiled for 5 min. For each sample, 40 g of protein determined by Bio-Rad protein assay was analyzed by immunoblotting as described (30) using monoclonal anti-smooth muscle-␣ actin antibody (Oncogene Research Products, Boston, MA). The enhanced chemiluminescence system was used for detection of bound antibodies (ECL System; Amersham Pharmacia Biotech, Arlington Heights, IL).
Preparation of Collagen Lattices and Measurement of Collagen Gel Contraction by Lung Fibroblasts-Collagen lattices were prepared using type I collagen from rat tail tendon (BD Bioscience, Bedford, MA) adjusted to a final value of 2.5 mg/ml with 0.01% acetic acid, 10 ϫ DMEM, and 0.1 N NaOH. Normal and SSc lung fibroblasts (2.5 ϫ 10 5 cells/ml final concentration) were suspended in collagen (1.25 mg/ml of collagen final concentration) and aliquoted into 24-well plates (300 l/well). Collagen lattices were polymerized for 45 min in a humidified 10% CO 2 atmosphere at 37°C followed by incubation with DMEM containing 10% fetal calf serum for 4 h, followed by overnight incuba-tion in serum-free medium. To initiate collagen gel contraction, polymerized gels were gently released from the underlying culture dish and cells were immediately stimulated with various concentration of thrombin (0.001-0.5 unit/ml) in serum-free DMEM, or other factors as stated in the figure legends. The degree of collagen gel contraction was determined after 0.5, 2, 6, and 24 h. The diameter of the gels were measured in mm and recorded as the average values of the major and minor axes.
DNA Synthesis-Lung fibroblasts were plated onto 12-well plates, allowed to grow to 80% confluency and synchronized with serum-free DMEM for 24 h. [ 3 H]Thymidine incorporation was measured as previously described (11) with small modifications. Cells were stimulated with various concentrations of thrombin (0.001 to 0.5 units/ml) for 24 h after preincubation in the absence or presence of PKC inhibitors for 30 min. Following the 24-h incubation, 1 Ci/ml [ 3 H]thymidine was added to the cells for an additional 6 h incubation. Cells were then washed 3 times with PBS followed by 3 washings with ice-cold trichloroacetic acid (5%, w/v) and solubilized with NaOH/SDS (both 0.1% w/v). [ 3 H]Thymidine incorporation was determined by liquid scintillation counting.
PKC⑀ Oligonucleotide Treatment of Normal and Scleroderma Lung Fibroblasts-Oligonucleotides were introduced into the cells as described previously (18). Antisense oligonucleotide for PKC⑀ and appropriate sense oligonucleotide (control) were dissolved in culture medium and sterilized by filtration through 0.2-m cellulose acetate filters. Oligonucleotide (2 M) and Lipofectin (10 g/ml) were mixed and preincubated for 30 min at room temperature, the mixture was added to the cells, and the samples were incubated for 6 h at 37°C. After washing cells with DMEM to remove Lipofectin, fresh oligonucleotides (2 M) in DMEM with 10% fetal calf serum were added and incubation continued for 24 h at 37°C. For collagen gel contraction assay normal lung fibroblasts were suspended in 1.5 mg/ml collagen and seeded in 24-well plates as described above, and after polymerization the medium was replaced with serum-free DMEM containing fresh oligonucleotides (2 M) for an additional 24 h. For determination of SM-␣ actin organization after PKC⑀ depletion, antisense and sense oligonucleotides were introduced to normal and SSc lung fibroblasts as described above. The cells were seeded on glass slides, incubated with fresh oligonucleotides in serum-free DMEM overnight, and then stimulated with thrombin (0.5 units/ml) for an additional 24 h.
Distribution and Translocation of PKC⑀ in Lung Fibroblasts-PKC⑀ translocation assay was performed as described previously (19). Normal and SSc lung fibroblasts were grown to confluency, kept in serum-free DMEM overnight, and then treated with thrombin (0.5 unit/ml) for 15 min, washed twice with cold PBS, harvested, then collected by centrifugation at 500 ϫ g for 1 min, and suspended in 600 l of digitonin buffer (0.55% digitonin, 25 mM Tris-HCl, pH 7.6, 5 mM EGTA, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 25 g/ml leupeptin). Cells were incubated on ice for 30 min, centrifuged at 10,000 ϫ g for 10 min at 4°C, and the resulting supernatant was called the cytosolic fraction.
The pellet was subsequently dissolved in 500 l of digitonin extraction buffer containing 1% Triton X-100, passed through a 26-gauge needle, and centrifuged again for 10 min at 10,000 ϫ g at 4°C. The resulting supernatant was called the Triton X-100-soluble fraction (membrane fraction). Forty micrograms of protein from the cytosolic and membrane fractions was analyzed for PKC⑀ expression by Western blotting as described above. Monoclonal antibodies against PKC⑀ (Transduction Laboratories, Lexington, KY) were used.
Introduction of PKC⑀ Translocation Inhibitor Peptide into Lung Fibroblasts-PKC⑀ translocation inhibitor peptide (TIP) and scrambled peptide for TIP (Calbiochem, La Jolla, CA) were introduced into the permeabilized cells with saponin (Sigma) as described by Johnson et al. (35). Cells were grown to confluency, then incubated with PBS at room temperature twice for 2 min, and in fresh, cold PBS for 2 min on ice, followed by 10 min incubation on ice with permeabilization buffer (20 mM HEPES, pH 7.4, 10 mM EGTA, 140 mM KCl, 50 g/ml saponin) in the presence or absence of TIP (150 g/ml). Cells were then washed five times with cold PBS, followed by 20 min incubation on ice, 2 min incubation at room temperature, and 2 min at 37°C with PBS. Finally, the cells were incubated 30 min with serum-free DMEM at 37°C, and then stimulated with thrombin (0.5 unit/ml) for 15 min.

Co-immunoprecipitation of PKC⑀ and Smooth Muscle-␣ Actin by Lung Fibroblasts Stimulated with
Thrombin-Co-immunoprecipitation assay was performed as described by Budd et al. (31) with some modifications (32). Normal and SSc lung fibroblasts were grown to confluency, kept in serum-free DMEM overnight and then treated with thrombin (0.5 unit/ml) for 15 min, and/or pretreated with cytochalasin D (10 M) for 30 min, and then stimulated with or without thrombin (0.5 unit/ml) for 15 min, washed with ice-cold PBS and collected with 1 ml of ice-cold solubilization buffer (10 mM Tris, 10 mM EDTA, 500 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, pH 7.4). Samples were rotated for 3 h and then cleared by microcentrifugation at 4°C. Next, anti-smooth muscle-␣ actin monoclonal antibody (1 g) was added, and the samples were rotated for additional 90 min at 4°C. Immune complexes were isolated on protein G-Sepharose beads (Amersham Pharmacia Biotech, Piscataway, NJ) washed three times with buffer containing 10 mM Tris, 10 mM EDTA, pH 7.4. Isolated immune complexes were then resolved on 8% SDS-polyacrylamide electrophoresis gel and immunoblotted with anti-⑀-PKC monoclonal antibody overnight at 4°C.

Smooth Muscle-␣ Actin Expression and Organization in Lung Fibroblasts Using Confocal Microscopy-Normal and SSc lung fibroblasts
were cultured to subconfluence on glass slides in DMEM containing 10% fetal calf serum. PKC⑀ antisense and sense oligonucleotides or PKC⑀ translocation inhibitor peptide and scrambled peptide (negative control) were introduced into the cells as described above. Cells were stimulated with or without thrombin (0.5 unit/ml) for 24 h, washed with cold PBS, fixed in methanol, and immunostained with SM-␣ actin antibody. CY-2 anti-mouse goat IgG (Jackson ImmunoResearch) was used as secondary antibody at 1 g/ml for 1 h at room temperature. Confocal microscopy was performed using Olympus Merlin Imaging System (Life Science Resource, Melville, NY).

Thrombin Increases Smooth Muscle-␣ Actin Protein Expression in Normal Lung
Fibroblasts-Lung fibroblasts stimulated with thrombin exhibited a dose-and time-dependent increase in SM-␣ expression with maximum expression at 0.5 unit/ml thrombin in normal fibroblasts (Fig. 1). Thrombin, at concentrations as low as 0.005 unit/ml, increased SM-␣ actin in normal lung fibroblasts (Fig. 1A). Thrombin only slightly induced SM-␣ actin in SSc lung fibroblasts (Fig. 1B). Among fibroblast lines derived from various stages of SSc lung, differences in the level of SM-␣ expression were observed. We observed a 2.5-8fold higher basal level in SSc than in normal cells. In SSc cells with lower basal SM-␣ levels, a further increase of actin was observed upon exposure to thrombin (data not shown). In the present study we investigated SSc cell lines with high levels of SM-␣ actin, where thrombin only slightly up-regulated its expression.
Thrombin Induces Rapid Contraction of Lung Fibroblastpopulated Collagen Gels-Contractile phenotype is another characteristic feature of myofibroblasts. We observed that thrombin induces collagen gel contraction by normal lung fibroblasts in a dose-dependent manner with the maximum effect at 0.5 unit/ml ( Fig. 2A). Collagen gels rapidly contracted from ϳ15 mm in diameter (serum-free DMEM) to less than 4 mm in diameter within 30 min, reaching maximum contraction 1-2 h after thrombin stimulation. There were no further changes in collagen gel contraction after 24 h treatment with thrombin. Specific thrombin inhibitors, hirudin and PPACK, abolished thrombin-induced collagen gel contraction (Fig. 2B). We also found that untreated SSc lung fibroblasts contract collagen gels in serum-free DMEM, while normal lung fibroblasts do not (Fig. 2C). Additionally, SSc lung fibroblasts containing higher levels of SM-␣ actin contracted collagen gels to a higher extent than did those with lower levels of SM-␣ actin (data not shown). We conclude that thrombin induces a contractile phenotype that is similar to that observed in SSc lung fibroblasts at basal conditions.
Thrombin Receptor Agonist Peptide Mimics Thrombininduced Collagen Gel Contraction-Next we sought to establish whether thrombin's effect on inducing collagen gel contraction by normal lung fibroblasts is mediated via cleavage of the proteolytically activated receptor-1. We and others showed that the 6-amino acid synthetic thrombin receptor agonist peptide, TRAP-6, mimics thrombin's effect on a variety of cellular responses (18,19). Using TRAP-6 (proteolytic cleavage pathway) and the 14-amino acid peptide, F-14 (nonproteolytic pathway, binding to the receptor), we have shown that collagen gel contraction (Fig. 3A, left panel) and smooth muscle-␣ actin (Fig.  3B) are induced by TRAP-6 in a dose-dependent manner, while F-14 had no effect (Fig. 3A, right panel). Collagen gel contraction was induced by TRAP-6 to the same extent as native thrombin, suggesting that proteolytic cleavage of the thrombin receptor mediates collagen gel contraction when normal lung fibroblasts are stimulated by thrombin.
Thrombin-induced Transition of Normal Lung Fibroblasts to a Myofibroblast Phenotype Is Mediated via a TGF-␤-independent Pathway-One of the factors known to induce SM-␣ actin and a contractile phenotype in isolated fibroblasts, i.e. differentiation to a myofibroblast phenotype, is TGF-␤1 (4,14). Therefore, we investigated the possibility that thrombin-induced myofibroblast phenotype is mediated via a TGF-␤dependent mechanism. We observed that thrombin induces SM-␣ actin in normal lung fibroblasts within 24 h, while TGF-␤1 stimulation requires 48 h for such induction (Fig. 4).

FIG. 1. Thrombin increases smooth muscle-␣ actin expression in normal lung fibroblasts.
A, human fibroblasts from normal (Nml) lungs were serum-deprived overnight, then treated with various concentrations of thrombin for 24 h. B, normal and SSc lung fibroblasts treated with thrombin (0.5 unit/ml) for 0 -24 h. Cell lysate proteins (40 g) were separated by SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-SM-␣ actin antibody. The results presented are representative of an experiment that was performed four times.

FIG. 2. Thrombin induces collagen gel contraction by lung fibroblasts.
Normal (Nml) and SSc lung fibroblasts were cultured in the presence of 1.5 mg/ml collagen in 24-well plates in DMEM with 2% fetal calf serum overnight, followed by 24 h incubation in serum-free DMEM. A, 0.5 unit/ml thrombin was added and collagen gel contraction was measured at different time points (between 15 min and 24 h). B, normal lung fibroblasts were stimulated with various concentrations of thrombin (0.01-0.5 unit/ml), and with thrombin (0.5 unit/ml) in the presence of hirudin or PPACK. Collagen gel contraction was measured after 24 h incubation. C, normal (Nml) and SSc cell lines were incubated in serum-free medium, and collagen gel contraction was determined after 24 h. Data represent mean values Ϯ S.D. of three experiments, each in duplicate. The asterisk represents statistically significant differences (p Ͻ 0.05) between cells stimulated with 0.5 unit/ml thrombin for 15 and 30 min, 2 and 24 h versus nonstimulated cells and between Nml versus SSc lung fibroblast cell lines (C). Double asterisk represents statistically significant differences (p Ͻ 0.001) between cells stimulated with 0.5 unit/ml thrombin and stimulated with 0.5 unit/ml thrombin with hirudin or PPACK (B).
Pretreatment of the cells with antibodies against TGF-␤1 did not inhibit thrombin-induced SM-␣ actin, whereas TGF-␤1induced SM-␣ actin was inhibited (Fig. 4). Similarly, collagen gel contraction induced by TGF-␤1 required up to 24 h in contrast to 2 h when cells were treated with thrombin (data not shown). Therefore, we conclude that thrombin-induced differentiation of normal lung fibroblasts to a myofibroblast phenotype is not dependent on a TGF-␤-mediated pathway, although the possibility exists that this mechanism may be involved at later time points.
Other factors such as platelet-derived growth factor-AA, platelet-derived growth factor-BB, and tenascin C, which are stimulated by thrombin in human lung fibroblasts or other mesenchymal cells, might be responsible for thrombin's effect on collagen gel contraction. We have observed that all these factors induce collagen gel contraction. However, the effect was not as rapid as with thrombin, occurring within 8 -24 h (data not shown).
Protein Kinase C Is Involved in Thrombin-induced Fibroblast Proliferation and Collagen Gel Contraction-Cells with a myofibroblast phenotype are also characterized by an increase in proliferative capacity (11,12). Thrombin is a well known mitogen (33,34) and has been shown to induce human lung fibroblast proliferation (17). Basal levels of [ 3 H]thymidine incorporation were elevated 2.2-fold in SSc cells compared with normal lung fibroblasts (Fig. 5). Each cell line demonstrated a dosedependent proliferative response to the thrombin, which was much more profound in normal lung fibroblasts. At concentrations as small as 0.1 unit/ml, thrombin induced a 5.8-fold increase in [ 3 H]thymidine incorporation in normal lung fibroblasts, and only a 2-fold increase in scleroderma fibroblasts. Many cellular responses to thrombin are regulated by PKC signaling (15,18,19,27). To determine whether PKC is essential for thrombin-stimulated growth in lung fibroblasts, we employed two different PKC inhibitors: calphostin C, which inhibits almost all PKC isoforms, and Go 6976, which inhibits mainly ␣ and ␤ PKC isoforms. We found that both of them blocked thrombin-induced DNA synthesis (Fig. 6A). Interest-ingly, calphostin C abolished thrombin-induced collagen gel contraction as well, while Go 6976 did not, even at a concentration of 50 nM. To determine the PKC isoform that mediates thrombin-induced collagen gel contraction, we used the specific PKC inhibitor, Ro 320432, whose inhibitory effect is concentration-dependent. Ro 320432, which in low concentrations inhibits PKC␣ and -␤, while in high concentrations inhibits PKC␣, -␤, and -⑀, significantly reduced thrombin-induced collagen gel contraction only at high concentrations (Fig. 6B). These results suggest that classical PKC isoforms might be involved in the mitogenic response to thrombin, whereas PKC⑀ might be involved in phenotypic modifications of the myofibroblast, including smooth muscle-␣ actin expression, organization and collagen gel contraction.
PKC⑀ Mediates Collagen Gel Contraction, SM-␣ Actin Expression, and Its Organization by Lung Fibroblasts Stimulated with Thrombin-To establish whether PKC⑀ is indeed involved in thrombin-induced collagen contraction, we employed an antisense oligonucleotide technique. Antisense oligonucleotides were used to decrease PKC⑀ synthesis in normal lung fibroblasts. This treatment decreased the level of PKC⑀ protein by 60 -70%, compared with untreated cells or cells treated with sense oligonucleotide for PKC⑀ (Fig. 7C). As we previously reported (19), antisense or sense oligonucleotide treatment did not induce cell injury, nor did it have any effect on overall protein synthesis, suggesting that inhibition of PKC⑀ protein synthesis was due directly to oligonucleotide treatment. We demonstrate that PKC⑀ depletion in normal lung fibroblasts inhibits thrombin-induced collagen gel contraction, SM-␣ actin expression and organization, whereas pretreatment with PKC⑀ sense oligonucleotide has no effect (Fig. 7, A and C, and Fig. 8B) PKC isoforms have been shown to have specific subcellular localization before activation. Inactive PKC isoforms are localized to the cytosol, and after activation they translocate to the plasma membrane and/or cytoskeleton (32). Recently we reported that thrombin induces and activates PKC⑀ in lung fibroblasts (19). Thrombin treatment results in PKC⑀ translocation from the cytosolic fraction to both the membrane and cytoskeletal fraction (19). To establish whether inhibition of PKC⑀ activation affects thrombin-induced collagen gel contraction and SM-␣ expression, we employed a PKC⑀ translocation inhibitor peptide to inhibit translocation. We have shown that inhibition of PKC⑀ translocation completely abolished thrombinmediated collagen gel contraction and significantly decreased SM-␣ expression and organization in normal lung fibroblasts (Fig. 7, B and C, and Fig. 8C). Antisense oligonucleotides for PKC⑀, or translocation inhibitor for PKC⑀, did not affect DNA synthesis either in normal or in SSc lung fibroblasts. Moreover, it did not affect expression and organization of SM-␣ actin in scleroderma lung fibroblasts (data not shown).
Normal lung fibroblasts express small amounts of SM-␣ actin, which is not fully organized. Thrombin affects SM-␣ actin organization within 2 h in normal lung fibroblasts, and after 24 h of thrombin treatment cells express large amounts of highly organized SM-␣ actin (Fig. 8A). In contrast, SSc lung fibroblasts innately express highly organized SM-␣ actin, and thrombin only slightly increases its expression (Fig. 8D). Cytochalasin D, and inhibitor of actin organization, inhibited thrombin-induced SM-␣ actin in normal and scleroderma lung fibroblasts (Fig. 8E). We have also shown that cytochalasin D completely abolishes thrombin-induced collagen gel contraction in lung fibroblasts (Fig. 9). These results suggest that PKC⑀ depletion and inhibition of PKC⑀ activation prevents thrombin-induced SM-␣ actin organization in normal lung fibroblasts but does not disrupt existing and highly organized SM-␣ actin in scleroderma lung fibroblasts. Disruption of SM-␣ actin organization by cytochalasin D inhibits collagen gel contraction by normal and scleroderma lung fibroblasts.

PKC-⑀ Associates with Smooth Muscle-␣ Actin in Thrombinstimulated Normal Lung Fibroblasts-Recently
we have shown that thrombin treatment promotes PKC⑀ expression in normal lung fibroblasts, yet inhibits PKC⑀ expression in SSc lung fibroblasts (19), suggesting that this isoform may be differentially activated in normal and SSc lung fibroblasts. Both active and inactive PKC isoforms are believed to be localized to specific intracellular sites due to binding by specific receptor(s) (35). Thus, it is possible that normal and SSc lung fibroblasts exhibit different receptors for active and inactive forms of PKC⑀, or PKC⑀ is differentially activated in each cell type.
The common marker for both thrombin-treated normal and SSc lung fibroblasts is SM-␣ actin. To test the idea that SM-␣ actin may serve as a receptor or specific anchoring protein for PKC⑀, we used an immunoprecipitation technique. Interestingly, we found that in normal lung fibroblasts PKC⑀ binds to SM-␣ actin after thrombin stimulation while in SSc lung fibroblasts PKC⑀ and SM-␣ actin are associated even in the untreated cells (Fig. 10). Based upon our results showing that PKC⑀ depletion or inhibition affects SM-␣ organization and that disruption of actin organization affects collagen gel contraction, we asked whether the organization of SM-␣ might also be responsible for binding PKC⑀ in lung fibroblasts. Therefore, we examined whether inhibition of actin organization would affect PKC⑀⅐SM-␣ actin complex formation in normal lung fibroblasts stimulated with thrombin, and also whether cytochalasin D treatment would prevent PKC⑀⅐SM-␣ actin complex formation in SSc lung fibroblasts. Thrombin, by enhancing the amount of SM-␣ actin and by activating PKC⑀, binds this isoform to SM-␣ actin. In contrast, high amounts of SM-␣ actin in SSc lung fibroblasts cause binding to PKC⑀ even prior to any treatment. It is also possible that PKC⑀ is already activated in SSc lung fibroblasts and, therefore, binds to SM-␣ actin forming a complex. This would explain our observation that thrombin does not activate this PKC isoform in SSc lung fibroblasts (19). As expected, cytochalasin D did not affect co-immunoprecipitation of PKC⑀/SM-␣ actin in either cell line, confirming that this complex does not require polymerized actin and can be formed with G actin, particularly with SM-␣ actin (Fig. 10). DISCUSSION The present studies were carried out to characterize the myofibroblast phenotype that occurs after activation of normal lung fibroblasts by thrombin. Phenotypic modifications induced by thrombin such as SM-␣ actin expression and collagen gel contraction were observed. We demonstrate that PKC⑀ regulates transition of normal lung fibroblasts to myofibroblasts by binding to SM-␣ actin. Interestingly, in SSc lung fibroblasts PKC⑀ and SM-␣ actin are associated even in the untreated state. This suggests that SM-␣ actin may serve as a substrate or anchoring protein for PKC⑀ in lung fibroblasts when acti-

FIG. 4. Thrombins induction of SM-␣ actin in lung fibroblasts is not dependent on a TGF-␤ pathway.
Normal lung fibroblasts were stimulated with thrombin (0.5 unit/ml) (left panel) or TGF-␤1 (1 ng/ml) (right panel) for 24 or 48 h in the presence or absence of antibodies against TGF-␤1 (10 g/ml). SM-␣ actin protein levels in cell extracts (40 g of protein) were determined by Western blot analysis. Experiment was performed three times and the representative blots are presented. The asterisk represents statistically significant differences between cells stimulated with thrombin versus cells pretreated with PKC inhibitors and then stimulated with thrombin (p Ͻ 0.001). B, lung fibroblasts (10 5 cells/well) were cultured within 1.5 mg/ml collagen in 24 well plates in DMEM with 2% fetal calf serum overnight, followed by 24 h incubation in serum-free DMEM. PKC inhibitors (calphostin C, Go 6976, and Ro 320432) were added to serumfree medium 30 min before thrombin. The level of collagen gel contraction was measured at different time points between 15 min and 24 h. The results after 2 h stimulation are presented. The experiment was performed three times and representative results are presented. vated by thrombin. Using antisense oligonucleotides we demonstrate that decreasing the level of PKC⑀ in normal lung fibroblasts abolishes SM-␣ actin expression and collagen gel contraction induced by thrombin. We also observed that inhibition of PKC⑀ translocation in normal lung fibroblasts abolishes thrombin-induced collagen gel contraction and SM-␣ actin induction, suggesting that activation of PKC⑀ is required for both effects. We propose that thrombin differentiates normal lung fibroblasts to a myofibroblast phenotype via PKC⑀ signaling. Moreover, thrombin-induced transition of normal lung fibroblasts to a myofibroblast phenotype resembles the phenotype of SSc lung fibroblasts.
The myofibroblast phenotype is present in a variety of fibrotic diseases, including scleroderma. Scleroderma is an autoimmune connective tissue disease characterized by microvascular injury and fibrosis of skin and visceral organs, including the lung (38,39). Pulmonary fibrosis in scleroderma is a frequent complication and major cause of death; however, the pathogenesis is unclear and no effective therapy exists, The pathology of pulmonary fibrosis in SSc demonstrates features of dysregulated and abnormal repair, fibroproliferation and deposition of various extracellular matrix proteins (38).
It is postulated that activated fibroblasts (myofibroblasts) are involved in the pathogenesis of lung fibrosis. Recently, we

FIG. 7. Depletion of PKC⑀ or inhibition of PKC⑀ translocation abolishes thrombin-induced collagen gel contraction and smooth muscle-␣ actin expression in normal lung fibroblasts.
Antisense and sense oligonucleotides for PKC⑀ (2 M) were introduced into the cells using Lipofectin (10 g/ml), and collagen gel contraction was measured as described under "Experimental Procedures." A, right panel, PKC⑀ expression in cells treated with antisense oligonucleotides (AS), sense oligonucleotide (S) for PKC⑀ or in serum-free medium analyzed by Western blot of cell extracts followed by densitometric analysis. The difference between cells treated with antisense oligonucleotides and control cells treated with sense oligonucleotides is statistically significant (p Ͻ 0.05). Shown is a representative blot from three experiments. A, left panel, collagen gel contraction by lung fibroblasts treated with thrombin (0.5 unit/ml), or cells transfected with oligonucleotides for PKC⑀ AS and S in the presence or absence of thrombin. B, left panel, PKC⑀ TIP and its negative control (scrambled PKC⑀ translocation inhibitor peptide, TIPN) were introduced into cells permeabilized with saponin (50 g/ml) as described under "Experimental Procedures." Cells were then incubated in the presence or absence of thrombin (0.5 unit/ml) for 24 h. B, right panel, the inhibition of PKC⑀ translocation was confirmed by Western blot analysis after separation of cytosolic and membrane fraction, as described under "Experimental Procedures." The difference between cells treated with TIP plus thrombin versus samples treated with thrombin and/or TIP only is statistically significant (p Ͻ 0.05). The results presented are representative of three experiments performed in duplicate. C, smooth muscle-␣ actin expression in normal lung fibroblasts after treatment with thrombin, PKC⑀(S), PKC⑀(AS), thrombin plus PKC⑀(S), or PKC⑀(AS), TIPN, TIP, and thrombin plus TIPN or TIP in concentrations as described above. Representative results from three experiments are presented.  reported increased numbers of myofibroblasts in lung tissue from SSc in association with high extracellular matrix accumulation (13). Several studies have demonstrated a correlation between fibrosis and SM-␣ actin-expressing myofibroblasts (4,5,40). The isolated myofibroblasts from various fibrotic tissues, including lungs, are thought to be a primary source of collagen and other extracellular matrix proteins (10,11,19). Although myofibroblasts have been postulated to participate in the development of lung fibrosis, the mechanism of myofibroblast phenotype induction is not well understood.
In the present study we investigated thrombin-induced differentiation of normal lung fibroblasts to a myofibroblast phenotype and the signaling pathway that modulates such transition. Previously, we reported a significant increase in thrombin activity in bronchoalveolar lavage fluid from SSc patients (17). It has also been observed that procoagulant activity is increased in the lungs of patients with idiopathic pulmonary fibrosis, especially patients with progressive disease (41), and in bleomycin-induced pulmonary fibrosis in mice (42). Thrombin, besides its importance in thrombosis and hemostasis, also has several important functions at a cellular level in the stimulation of platelets, leukocytes, endothelial cells, smooth muscle cells, and lung fibroblasts (15,18,21,42,43). The multiple effects of thrombin on cells include regulation of proliferation, stimulation and expression of various cytokines, and regulation of several extracellular matrix proteins (19,(23)(24)(25)(26). Proliferation, contraction, and increased extracellular matrix protein production by myofibroblasts thereby generate forces within the parenchyma, which may be major contributors to the severely impaired lung function observed in SSc and other types of pulmonary fibrosis.
We observed that SSc lung fibroblasts express significantly higher amounts of SM-␣ actin compared with normal lung fibroblasts. We found that thrombin induces SM-␣ actin in a dose-and time-dependent manner in normal lung fibroblasts, as well as in SSc lung fibroblasts, despite the already high levels of SM-␣ actin in SSc cells. Treatment of normal lung fibroblasts with thrombin induced SM-␣ actin to the levels observed in SSc lung fibroblasts. Because thrombin and SM-␣ actin co-exist in the inflammatory and early fibrotic stages in many pathologic situations, including pulmonary fibrosis (5,12,44), and because thrombin induces SM-␣ actin in cultured lung fibroblasts, we sought to determine the mechanism which mediates this interaction(s).
We observed that lung fibroblasts stimulated with thrombin have the ability to contract collagen gels. When cultured within collagen gel, fibroblasts are able to recognize collagen fibers and contract the gel. This is believed to reflect the in vivo phenomenon of wound contraction and extracellular remodeling in connective tissue. In lung fibrosis it might reflect the pathologic stiffness observed in SSc and other restrictive lung diseases. We demonstrate that thrombin is a potent inducer of rapid fibroblast-populated collagen gel contraction. Contraction is inhibited by two thrombin inhibitors, hirudin and PPACK, suggesting a specific effect of thrombin on collagen gel contraction. We also observed that normal lung fibroblasts when incubated in serum-free medium do not contract collagen gels even after 24 h, whereas SSc lung fibroblast-populated collagen gels contract within 8 -24 h. The level of contraction among SSc cells is dependent on the level of SM-␣-actin, with more SM-␣ actin being associated with a higher degree of collagen gel contraction.
Proteolytically activated receptor-1, cell types including fibroblasts mediates many of the pathophysiological responses to thrombin (43,44). Proteolytically activated receptor-1 can couple to the G 12/12 , G q , and G I proteins. The ␣-subunit of G 12 and G 13 binds Rho guanidine-nucleotide exchange factors, which activate small G-protein RhoA and mediate cytoskeletal reorganization (46,47). G␣ q activates phospholipase C␤, triggering phosphoinositide hydrolysis, which results in calcium mobilization and activation of protein kinase C (48). We observed that differentiation of lung fibroblasts to myofibroblasts is mediated via proteolytically activated receptor-1.
Many cellular responses to thrombin have been found to be regulated by a PKC-dependent signal transduction pathway including thrombin's effects on lung fibroblasts (18,19,42,49). Recently, we demonstrated that PKC␥ is involved in interleukin-8 up-regulation in normal lung fibroblasts, while PKC⑀ regulates thrombin-induced tenascin-C expression in normal lung fibroblasts (18,19). We show that proliferative capacity of SSc lung fibroblasts was increased more then two times when compared with DNA synthesis in normal lung fibroblasts. We also demonstrate that thrombin-induced DNA synthesis in lung fibroblasts is mediated by classical PKC isoforms (␣ or ␤). The mitogenic effect of thrombin was more profound in normal lung fibroblasts then in SSc cells, possibly due to desensitization of the thrombin receptor in SSc cells (50,51), which have been already exposed to thrombin in vivo (17). There are several PKC-dependent pathways associated with growth in a variety of cell types (33,34). Among them, the classical mitogen-activated protein kinase cascade (52) and/or protein kinase B pathway (53) may be involved in lung fibroblast proliferation as a dominant pathway, and this will be a subject of future studies.
PKC isoforms respond differently to various stimuli, and the patterns of activation of these isoenzymes vary in extent, duration and extracellular localization (54 -56). Although individual PKC isoforms demonstrate only subtle differences in enzymatic properties, ligand binding, and substrate specificity in vitro, the isoforms exhibit different tissue and cell-type specific FIG. 10. PKC⑀ co-immunoprecipitates with smooth muscle-␣ actin in thrombin-stimulated normal lung fibroblasts. Cytochalasin D does not affect PKC⑀⅐SM-␣ actin complex formation. Co-immunoprecipitation was performed as described under "Experimental Procedures." Normal and scleroderma lung fibroblasts grown to 90% confluency were stimulated with thrombin (0.5 unit/ml) for 15 min with or without pretreatment with cytochalasin D (10 M) for 20 min followed by immunoprecipitation with SM-␣-actin antibody and then immunobloted with anti-PKC⑀ monoclonal antibody. Mouse IgG (bottom panel) served as a loading control. expression patterns in vivo, suggesting unique, specific functions for each isoform (37). PKC is localized in the cytosol in an inactive form and, after cell stimulation, translocates to the plasma membrane where it becomes activated. Individual PKC isoforms can translocate to subcellular locations other then the plasma membrane, including other membrane vesicles, nuclear structures, and cytoskeleton components. Several studies have demonstrated that after activation individual PKC isoforms translocate to different subcellular sites in various cell types (37,56,57,64). The subcellular location of a specific isoform may directly control the potential of that isoform to perform distinct functions, since the targeting of PKCs to discrete subcellular compartments would restrict their access to potential substrates.
We demonstrate that PKC⑀ regulates transition of normal lung fibroblasts to myofibroblasts by binding to SM-␣ actin after thrombin treatment. Interestingly, in SSc lung fibroblasts PKC⑀ and SM-␣ actin are associated even in untreated cells. This suggests that SM-␣ actin may serve as a substrate or anchoring protein for PKC⑀ in lung fibroblasts when activated by thrombin. Using antisense oligonucleotides and inhibition of PKC⑀ translocation, we demonstrate that decreased levels of PKC⑀ or inhibition of its activation abolishes collagen gel contraction, SM-␣ actin expression and organization induced by thrombin in normal lung fibroblasts, thus confirming that PKC⑀ is responsible for this phenomenon. Antisense oligonucleotides for PKC⑀, as well as translocation inhibitor for PKC⑀, did not affect either expression or organization of SM-␣ actin in scleroderma lung fibroblasts (data not shown).
Activation of PKC in a variety of different cell types leads to changes in the cytoskeleton and actin (55,56,(62)(63)(64). It has been shown that specific PKC isoforms can associate with several cytoskeletal proteins including intermediate filament proteins (vimentin, cytokeratins), membrane-cytoskeletal cross-linking proteins (MARCKS, ankyrin), and components of the actin filaments (F-actin) and microtubules (tubulin) (58 -60). Some of the cytoskeletal proteins, including F-actin, have also been shown to be substrates for PKC (56). Several recent studies have shown that individual PKC isoforms colocalize with the polymerized form of actin, F-actin, as a response to external stimuli (36,56,61,64). It has been demonstrated that various PKC isoforms, including PKC⑀, bind to F-actin with different affinities, and that these interactions result in an elevated level of isozyme activity (55).
A specific F-actin-binding motif has been identified as being unique to PKC⑀ (63). Later, it was demonstrated that PKC⑀ contains a putative actin-binding motif that is unique to this individual member of the PKC gene family. An actin-binding motif in the PKC⑀ sequence is exposed upon activation of this isoform and functions as a dominant localization signal in NIH 3T3 fibroblasts (37). These studies also indicate that this protein-protein interaction of F-actin/PKC⑀ is sufficient to maintain PKC⑀ in a catalytically active conformation within cytoskeletal structures.
We have found that an inhibitor of actin organization, cytochalasin D, completely abolishes collagen gel contraction by normal and SSc lung fibroblasts, suggesting that actin disorganization is sufficient to affect contraction in both cell types. Because cytochalasin D affects all actins, we performed confocal microscopy to determine whether SM-␣ actin organization was affected in normal and SSc lung fibroblasts. Normal lung fibroblasts express small amounts of SM-␣ actin that is not fully organized. Thrombin organizes SM-␣ actin within 2 h in normal lung fibroblasts. After 24 h of thrombin treatment normal cells express large amounts of highly organized SM-␣ actin. In contrast, SSc lung fibroblasts express highly orga-nized SM-␣ actin, and thrombin only slightly increases its expression. Cytochalasin D, an inhibitor of actin organization, inhibits thrombin-induced SM-␣ actin in both normal and scleroderma lung fibroblasts. These data suggest that PKC⑀ depletion or inhibition of PKC⑀ activation prevents thrombininduced SM-␣ actin organization in normal lung fibroblasts but does not disrupt existing and highly organized SM-␣ actin in scleroderma lung fibroblasts like cytochalasin D. We have also demonstrated that cytochalasin D does not interfere with complex formation of PKC⑀ and SM-␣ actin in lung fibroblasts, suggesting that inhibition of actin organization/polymerization is not required for this protein-protein interaction.
Our study demonstrates for the first time that thrombin differentiates normal lung fibroblasts to a myofibroblast phenotype via PKC signaling. In normal lung fibroblasts stimulated with thrombin SM-␣ actin interacts with PKC⑀ and serves as a substrate for this PKC isoform, while in SSc lung fibroblasts PKC⑀⅐SM-␣ actin complex exists even in untreated cells. This differential protein-protein interaction between PKC⑀ and SM-␣ actin in normal and SSc lung fibroblasts may thus explain some of the differences observed in the behavior of these two cell types. We conclude that the chronic presence of thrombin generated by ongoing microvascular injury in SSc leads to activation of lung fibroblasts and to the appearance of a myofibroblast phenotype. Moreover, our data present a compelling link between thrombin and SM-␣ actin, each of which are elevated in active stages of pulmonary fibrosis, and provide additional support for the notion that thrombin is an important mediator of interstitial lung fibrosis associated with scleroderma.