SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Objective

Members of the transforming growth factor β (TGFβ) cytokine superfamily play critical roles in both homeostasis and disease. In light of their profibrotic effects, these molecules are implicated in the pathogenesis of fibrosis. In fibroblasts, TGFβ signals through the activin receptor–like kinase 5 (ALK-5) type I TGFβ and triggers Smad and MAP kinase signaling pathways. Because targeting of TGFβ signaling represents a potential approach to the treatment of systemic sclerosis (SSc) and other fibrotic disorders, we investigated the modulation of intracellular TGFβ signal transduction by SB431542, the first small-molecule inhibitor of ALK-5 to be described.

Methods

Ligand-induced activation of the Smad signaling pathway in human dermal fibroblasts was examined by Western blot analysis and confocal immunocytochemistry. Modulation of profibrotic gene expression was investigated using Northern blot analysis, transient transfection assays, and confocal microscopy. Induction of TGFβ production was evaluated by enzyme-linked immunosorbent assay.

Results

SB431542 abrogated TGFβ-induced phosphorylation and nuclear importation of endogenous Smad2/3 and Smad4, and inhibited Smad3- and Smad2-dependent gene transcription. Treatment with SB431542 prevented TGFβ-induced stimulation of collagen, fibronectin, plasminogen activator inhibitor 1, and connective tissue growth factor gene expression, TGFβ autoinduction, and myofibroblast transdifferentiation, and it could reverse stimulation even when added to the cultures after TGFβ. In contrast, STAT-6–mediated stimulation of collagen gene expression induced by interleukin-13 was not prevented by SB431542, indicating the specificity of blockade for ALK-5–dependent signaling. Furthermore, in contrast to its effects on receptor-activated Smad activation, SB431542 failed to prevent TGFβ-induced activation of MAP kinases.

Conclusion

The results indicate that SB431542 is a potent inhibitor of intracellular TGFβ signaling in normal fibroblasts through selective interference with ALK-5–mediated Smad activation and Smad-dependent transcriptional responses. Therefore, SB431542 is useful as a novel experimental tool for gaining a detailed understanding of normal and aberrant TGFβ signaling in SSc. Furthermore, as an anti-TGFβ agent, SB431542 may represent a potential new approach to the treatment of fibrosis.

Systemic sclerosis (SSc) is a chronic disease of unknown etiology characterized by autoimmunity, vascular damage, and progressive fibrosis of the skin and internal organs. The pathogenesis of fibrosis is not well understood, and there are no effective treatments. Fibroblasts from lesional tissues show evidence of activation, with increased synthesis of collagen, fibronectin, tissue inhibitor of metalloproteinases 1 (TIMP-1), and plasminogen activator inhibitor 1 (PAI-1); secretion of profibrotic cytokines such as transforming growth factor β (TGFβ), interleukin-4 (IL-4), IL-13, and connective tissue growth factor (CTGF); myofibroblast differentiation, with elevated levels of α-smooth muscle actin (α-SMA) and stress fiber formation; enhanced expression of cell surface receptors for TGFβ; and resistance to apoptosis (1). Although the nature of the extracellular signals that initially trigger and subsequently sustain and amplify fibroblast activation and the corresponding intracellular signal transduction pathways are still a subject of controversy, TGFβ is considered to play a fundamental role (2).

TGFβ is the prototype of a superfamily of multifunctional cytokines that includes activins and bone morphogenetic proteins (3). These cytokines regulate cell differentiation, proliferation, function, and survival, and in mesenchymal cells, they stimulate extracellular matrix synthesis and organization. In a variety of in vitro and in vivo experimental models, TGFβ induces the phenotypical features of activation in fibroblasts. Accordingly, TGFβ represents an important potential therapeutic target for preventing or arresting the fibrotic process in SSc.

The mechanisms underlying cellular responses to TGFβ are now understood in some detail. Members of the TGFβ superfamily signal through the sequential activation of 2 distinct serine/threonine kinase cell surface receptors that are expressed on virtually all cell types. Upon ligand binding, a type II receptor (TβRII) recruits and phosphorylates a type I receptor (TβRI), which is also known as activin receptor–like kinase (ALK). Of the 7 known type I (ALK) receptors, ALK-5 is most specific for TGFβ, whereas the closely related ALK-4 and ALK-7 interact with other members of the TGFβ superfamily (4). In addition, ALK-1 was recently shown to function as a type I TGFβ receptor, but it is expressed primarily on endothelial cells or at sites of epithelial–mesenchymal interactions (5).

The Smads have been identified as major signaling molecules downstream of activated TβRI. These highly conserved modular proteins function as signal transducers/transcriptional activators that shuttle between the cytoplasm and the nucleus. In response to TGFβ, ALK-5 phosphorylates Smad2 and Smad3 on serine residues, whereas ALK-1 activates Smad1 and Smad5. Activin signals are also transduced by Smad2 and Smad3, but via the ALK-4 and ALK-7 receptors, whereas Smads 1, 5, and 8 are substrates for the bone morphogenetic protein–activated ALKs. In contrast to these receptor-activated Smads (R-Smads) that are phosphorylated by type I TGFβ receptors, Smad4 serves as a Smad cofactor, and Smad7 functions as an inhibitor of TGFβ-Smad signaling. Upon activation, R-Smads interact with Smad4, and the heteromeric complex is then imported into the nucleus. Within the nucleus, the DNA-bound Smad complex regulates the transcription of target genes directly or in association with transcription factors such as Sp-1 and forkhead activin signal transducer 1 (FAST-1), coactivators such as p300/CREB binding protein, or corepressors such as Ski and SnoN. Inhibitory Smad7 interacts with activated TβRI in competition with R-Smads and enhances receptor ubiquitination and proteasomal degradation in caveolae.

In light of the diversity of cellular responses elicited by TGFβ, it is not surprising that in addition to the canonical Smad pathway, TGFβ also activates alternate signal transduction pathways in a cell type– and context-specific manner. Non-Smad signaling pathways induced by TGFβ in cell lines include protein kinase A, protein kinase C, calmodulin-dependent protein kinase II, the MAP kinases ERK, JNK, and p38, and phosphatidylinositol 3 kinase (PI 3-kinase) (for review, see ref.6). By regulating Smad activation or through intracellular cross-talk with the Smad pathway, these kinases can influence the amplitude and duration of Smad-dependent signaling and may directly induce Smad-independent TGFβ responses. The mechanisms linking non-Smad signaling pathways with activated TGFβ receptors are unknown. Furthermore, because most studies examining the cross-talk between ligand-induced Smad and non-Smad signaling pathways have been performed in transformed or immortalized cell lines, the biologic consequences for non-Smad signaling in normal fibroblasts and their relevance in the physiologic context remain incompletely understood.

Small-molecule inhibitors of individual protein kinases are useful not only for dissecting the mechanisms of signal transduction for specific ligands and for delineating their distinct roles in biologic responses, but they have potential as therapeutic agents as well. A group of competitive ATP binding site inhibitors of ALK-5 has recently been described (7). In immortalized cell lines, the ALK-5 inhibitor SB431542 prevented Smad-dependent transcriptional responses and TGFβ-induced R-Smad phosphorylation in vitro; inhibition of ALK-4 and ALK-7 was also noted (8). To date, however, SB431542 has been studied primarily in the context of in vitro phosphorylation assays with immobilized targets or in transformed epithelial cell lines transfected with constitutively active TβRI kinases and substrates. These artificial experimental models fail to reflect either signal strength and duration or intracellular modulation. The results, therefore, cannot accurately predict the effects on nontransformed primary fibroblasts within the physiologic context of TGFβ signaling.

In the present studies, we examined the effects of SB431542 on newborn foreskin fibroblasts and adult dermal fibroblasts activated by TGFβ in vitro, focusing on cellular responses implicated in the development of fibrosis. The results showed that SB431542 could prevent, as well as reverse, TGFβ-induced stimulation of extracellular matrix synthesis and messenger RNA (mRNA) expression, TGFβ autoinduction, and Smad3- and Smad2-dependent transcriptional responses in vitro, and could prevent fibroblast transdifferentiation into myofibroblasts. Inhibition of profibrotic responses was associated with marked selective suppression of TGFβ-induced phosphorylation and nuclear translocation of endogenous R-Smads. Together, these results indicate that signaling via ALK-5 and downstream Smads plays a fundamental role in mediating TGFβ profibrotic responses in normal fibroblasts. Furthermore, these observations suggest that pharmacologic inhibition of the ALK-5 receptor may represent a novel targeted approach to the control of fibrosis in SSc.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Culture of dermal fibroblasts

Primary dermal fibroblasts were established from anonymous newborn foreskin specimens by explantation, as previously described (9). Normal adult dermal fibroblasts from a 36-year-old woman were obtained from Cascade Biologics (Portland, OR). Hep-G2 cells were obtained from American Type Culture Collection (Rockville, MD). Culture media were from BioWhittaker (Walkersville, MD); all other tissue culture reagents were from Gibco BRL (Grand Island, NY).

For these experiments, fibroblasts were studied between passages 4 and 8. Cells were grown at 37°C in an atmosphere of 5% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (FCS), 1% vitamin solution, 100 units/ml of penicillin/streptomycin, and 2 mML-glutamine. When the fibroblasts reached confluence, fresh medium containing various concentrations of TGFβ1 (Amgen, Thousand Oaks, CA) or IL-13 (R&D Systems, Minneapolis, MN) was added. Cultures were incubated with SB431542 (GlaxoSmithKline, King of Prussia, PA), U0126 (Cell Signaling Technology, Beverly, MA), or vehicle for 1 hour prior to the addition of TGFβ1 or IL-13. In selected experiments, SB431542 was added to the cultures simultaneously with, or 60 minutes after, the addition of TGFβ1. Both SB431542 and U0126 were dissolved in DMSO. Toxicity of SB431542 was evaluated by determining cell viability with 1% trypan blue dye, and by the MTT viability assay using the TOX-1 kit (an MTT-based in vitro toxicology assay kit; Sigma, St. Louis, MO), according to the manufacturer's instructions.

Extraction and analysis of RNA

For determination of mRNA levels, total RNA was isolated from fibroblasts using TRIzol Reagent (Gibco BRL) and examined by Northern blot analysis, as previously described (9). Filters were sequentially hybridized with 32P-labeled human complementary DNA (cDNA) probes for Smad3, Smad4, Smad7, COL1A1, COL1A2, TIMP-1, PAI-1, α-SMA, fibronectin, TGFβ1, CTGF, GAPDH, and 18S ribosomal RNA. Signal intensities were quantitated by densitometry, and results were normalized against the signal intensities of GAPDH mRNA or 18S ribosomal RNA in each sample.

Immunoblot analysis

Whole cell lysates or nuclear and cytoplasmic extracts were prepared from fibroblasts and subjected to immunoblot analysis as previously described (10). Equal aliquots (20 μg/lane) were separated by reducing electrophoresis in 4–20% gradient gels and transferred onto Immobilon-P (polyvinylidene difluoride) membranes (Millipore, Bedford, MA). Following blocking with 5% nonfat dry milk, membranes were incubated with antibodies against Smad2/3, Smad4, or actin (all from Santa Cruz Biotechnology, Santa Cruz, CA) or against ERK, p38, JNK, phosphoserine Smad2, phosphothreonine/tyrosine ERK, phosphothreonine/tyrosine p38, or phosphothreonine/tyrosine JNK (all from Cell Signaling Technology), followed by horseradish peroxidase–conjugated secondary antibodies. After washing, immunoblots were developed with chemiluminescence reagents according to the manufacturer's protocol (Pierce, Rockford, IL).

Total Smad and MAP kinases were examined in the same immunoblots as their phosphorylated forms, following stripping of the membranes. Signal intensities of the phosphorylated Smad and MAP kinase bands were quantitated by densitometry, and the results, which were normalized against the intensities of the corresponding total Smad or MAP kinase bands in each sample, were expressed as the magnitude of increase compared with controls. To exclude cross-contamination of the cytosolic and nuclear fractions, membranes were immunoblotted with antibodies specific for histone H3 or Hsp70 (Santa Cruz Biotechnology).

Cellular immunofluorescence imaging and in situ hybridization

The expression and intracellular localization of endogenous Smads was studied by immunocytochemistry and fluorescence confocal microscopy, as described elsewhere (10). Briefly, fibroblasts incubated in media with 0.1% FCS were pretreated with various concentrations of SB431542, followed 1 hour later by TGFβ1. At the end of the incubation period, cultures were fixed with 100% methanol and stained with primary antibodies against Smad2/3 or Smad4 (Santa Cruz Biotechnology) or α-SMA (Sigma). Slides were incubated with secondary antibodies and stained with fluorescein isothiocyanate or rhodamine, and nuclei were identified by 4′,6-diamidino-2-phenylindole (DAPI). Nonimmunized IgG was used as a negative control.

Following stringent washing of the slides, the pattern and subcellular distribution of fluorescence were evaluated by confocal laser scanning microscopy. Each experiment was performed at least 3 times and the results were consistent. Quantitative analysis was performed by scoring 100 individual fibroblasts from different microscopic fields as showing a predominantly nuclear or a predominantly cytoplasmic distribution of immunofluorescence. The observer was blinded to the identity of each section.

In some experiments, the modulation of fibroblast mRNA expression was further examined by in situ hybridization. For this purpose, confluent fibroblasts on glass chamber slides were incubated with SB431542 or DMSO, and 60 minutes later, TGFβ1 was added. At the end of the incubation periods indicated below, fibroblasts were fixed with paraformaldehyde/Triton X-100 and prehybridized in buffer containing 50 μg/ml of single-stranded DNA. For hybridization, cDNA probes specific for human COL1A1, Smad7, or GAPDH were labeled using a random primer fluorescein labeling kit (Perkin Elmer, Boston, MA) according to the manufacturer's protocol and as described previously (11). Signals were amplified using the tyramide signal amplification fluorescence system (Perkin Elmer). For quantitative analysis of the results, the proportion of positive fibroblasts was determined by scoring 100 fibroblasts in multiple high power fields. The observer was blinded to the identity of each slide.

Transient transfections

Subconfluent cultures of fibroblasts were transfected using Superfect reagent (Qiagen, Valencia, CA), as described previously (10). The reporter plasmids SBE4-Luc (containing 4 tandem copies of an 8-bp palindromic consensus Smad binding element that is specifically recognized by Smad3/4) (12), AR3-Lux (containing 3 copies of the Xenopus Mix.2 gene activin response element that is specifically recognized by FAST-1) (13), and 772COL1A2-CAT (containing the −772/+58-bp fragment of the human α2[I] collagen gene) were used. To study the regulation of AR3-Lux, reporter gene transfections were carried out in Hep-G2 cells cotransfected with expression vectors for Smad2 plus Xenopus FAST-1. The amount of plasmid DNA was equalized for transfection by adding the empty vector pcDNA3. A fixed amount (0.5 μg) of internal control reporter Renilla luciferase under the thymidine kinase promoter (pRL-TK; Promega, Madison, WI) was also cotransfected in every experiment to correct for variations in transfection efficiencies between the samples.

Following transient transfection, fibroblasts were pretreated with DMSO or with various concentrations of SB431542, followed by TGFβ1 for 48 hours. Luciferase and chloramphenicol acetyltransferase (CAT) activities in equal aliquots were determined as described previously (14). All experiments were performed in triplicate and performed at least 3 times.

In vitro kinase activity assays

The effects of SB431542 on cellular ERK activity were determined by in vitro kinase assays, as described previously (15). Briefly, confluent fibroblasts incubated with TGFβ1, with or without pretreatment with SB431542, were harvested after 15 minutes, and whole cell lysates were immunoprecipitated with antibodies against ERK-1/2 (Cell Signaling Technology). Immunoprecipitated proteins were washed in phosphorylation lysis buffer followed by kinase buffer (25 mM HEPES, pH 7.4, 25 mM MgCl2, 25 mM glycerophosphate, 100 μM sodium orthovanadate, 2 mM dithiothreitol, and 20 μM ATP), and resuspended in 30 μl of kinase buffer containing 3 μg of Elk-1 (Ser383) peptide substrate (Santa Cruz Biotechnology) and 10 μCi of γ32P-ATP. Following incubation for 30 minutes at room temperature, 30 μl of the reaction supernatant was spotted onto P81 paper, washed with 0.7% phosphoric acid and acetone, and counted in a scintillation counter.

Quantitation of TGFβ1

In selected experiments, the effect of SB431542 on autoinduction by TGFβ1 was determined. For this purpose, confluent fibroblasts in serum-free media were pretreated with SB431542 (10 μM) for 30 minutes, followed by incubation with 12.5 ng/ml TGFβ2 (Genzyme, Framingham, MA) for up to 48 hours. The cultures were harvested at various time points, and the conditioned media were activated with HCl (to measure both latent and active TGFβ1). Concentrations of TGFβ1 were quantitated by colorimetric enzyme-linked immunosorbent assay (ELISA) using an ELISA kit (R&D Systems) according to the manufacturer's instructions. Each experiment was performed in triplicate and performed 3 times and the results were consistent.

Statistical analysis

For comparisons of the means in control and treated fibroblasts, the Mann-Whitney U test was used. P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Regulation of TGFβ-induced intracellular signaling by SB431542

Cellular signaling by TGFβ involves the receptor-mediated phosphorylation of cytoplasmic Smads. Recent studies indicate that in addition to Smads, TGFβ also induces the activation of multiple MAP kinases in a cell type–dependent manner. In order to investigate the induction of downstream signaling pathways in primary dermal fibroblasts, we first examined the effects of TGFβ by Western blot analysis. Confluent fibroblasts were incubated with TGFβ1 for various periods in media containing 0.1% FCS. At the end of the incubation periods, whole cell lysates were prepared and subjected to immunoblot analysis.

The results, shown in a representative experiment, demonstrated that in addition to Smad2 phosphorylation, TGFβ also induced rapid and transient phosphorylation of the MAP kinase ERK-1/2 (Figure 1A). Maximal 6-fold induction compared with untreated fibroblasts was observed within 15 minutes. The MAP kinase JNK showed a more modest (∼3-fold) increase in phosphorylation induced by TGFβ (Figure 1A), and p38 was inconsistently phosphorylated by TGFβ, with a slow and time-dependent increase reaching maximal 2–3-fold induction at 24 hours (data not shown).

thumbnail image

Figure 1. Modulation of intracellular transforming growth factor β (TGFβ) signaling. Confluent dermal fibroblasts were A, incubated with 12.5 ng/ml of TGFβ1 or B, pretreated with 10 μM SB431542 or DMSO for 30 minutes prior to the addition of TGFβ1. Following incubation for the indicated periods, cultures were harvested, and whole cell lysates were subjected to Western blot analysis using antibodies against phosphothreonine/tyrosine ERK-1/2 (p-ERK) and total ERK-1/2, phospho-JNK or total JNK, or phosphoserine Smad2 (p-Smad2) and total Smad2/3. Immunoblots representative of several independent experiments are shown. C, ERK-1 activity following 15 minutes' incubation of the fibroblasts with TGFβ1 in the presence or absence of SB431542 was determined by in vitro kinase assays. Results are expressed as the fold increase relative to untreated fibroblasts; values are the mean of 2 independent experiments.

Download figure to PowerPoint

In order to examine the effects of ALK-5 inhibition on signaling pathways activated by TGFβ, fibroblasts were pretreated with various concentrations of SB431542 for 60 minutes prior to addition of TGFβ1. The results showed that while SB431542 had no effect on fibroblasts, it reversed the TGFβ-induced rapid phosphorylation of endogenous Smad2 (Figure 1B). Maximal inhibition was noted at 10 μM SB431542, which is consistent with previous results with NIH3T3 cells (7). In contrast, SB431542 had no effect on TGFβ-induced ERK phosphorylation, whereas the MEK inhibitor U0126 fully prevented this response (Figure 2B). The modest and inconsistent early stimulation of JNK phosphorylation induced by TGFβ, and delayed stimulation of p38 phosphorylation, were not altered by preincubation of the fibroblasts with SB431542 (data not shown). There was no evidence of cellular toxicity when fibroblasts were incubated with SB431542 at concentrations up to 10 μM for up to 72 hours, as measured by trypan blue dye exclusion and by MTT assays.

thumbnail image

Figure 2. Effects of SB431542 on intracellular Smad trafficking. Fibroblasts were pretreated with the indicated concentrations of SB431542, followed by further incubation with transforming growth factor β1 (TGFβ1) for 60 minutes. A, At the end of the incubation, fibroblasts were fixed, stained with anti-Smad2/3 antibodies, examined by confocal microscopy, and visualized with fluorescein isothiocyanate (green) or 4′,6-diamidino-2-phenylindole to identify nuclei (blue). Representative photomicrographs are shown (original magnification × 100). The proportion of fibroblasts unambiguously showing predominantly nuclear Smad2/3 localization was determined as described in Materials and Methods. The results (bottom) represent the mean and SD of at least 3 individual experiments. B, Nuclear or cytosolic fractions were isolated and examined by immunoblot analysis using antibodies against phospho-Smad2, total Smad2/3, total Smad4, phospho–ERK-1/2, histone H3, or Hsp70. The localization of histone H3 in the nuclear fraction and its absence in the cytosolic fraction, which contains Hsp70, indicate the lack of cross-contamination by these 2 fractions.

Download figure to PowerPoint

Next, we examined the modulation of TGFβ-induced MAP kinase activity by SB431542. In these experiments, we focused our attention on ERK-1/2, the MAP kinase that was most consistently and markedly phosphorylated by TGFβ in normal fibroblasts. To examine the effect of SB431542 on ERK-1 activity directly, in vitro assays were performed. The results from 2 independent experiments indicated that SB431542 (10 μM) did not prevent the activation of ERK kinase induced by TGFβ at 15 minutes (Figure 1C). Together, these results indicated that in normal fibroblasts, SB431542 acted as a potent and selective inhibitor of TGFβ-induced R-Smad phosphorylation, presumably mediated through blockade of ALK-5, the only known type I TGFβ receptor on fibroblasts.

Effects of SB431542 on intracellular R-Smad trafficking

We have previously demonstrated that brief exposure of normal dermal fibroblasts to TGFβ induced the translocation of cellular Smad2/3 and Smad4 from the cytoplasm into the nucleus (9). Because the nuclear importation of R-Smads is directly linked to their ALK-5–mediated phosphorylation, we examined whether inhibition of phosphorylation with SB431542 influenced intracellular Smad trafficking. For this purpose, confluent fibroblasts were incubated with SB431542 or vehicle, followed by TGFβ1 for 60 minutes, and examined using immunofluorescence confocal microscopy. The results showed that while TGFβ induced the rapid nuclear accumulation of Smad2/3 (as expected), pretreatment of the cultures with SB431542 resulted in dose-dependent inhibition of this response, with 10 μM consistently causing a >90% reduction (Figure 2A). In contrast, pretreatment with U0126 (10 μM) did not prevent TGFβ-induced Smad nuclear accumulation.

To further evaluate the effect of SB431542 on the regulation of Smad distribution, nuclear and cytosolic fractions were prepared from stimulated fibroblasts and examined in parallel by immunoblot analysis with antibodies against the phosphorylated form of Smad2 or against total Smad2/3 and Smad4. The results showed that pretreatment with SB431542 markedly reduced the levels of phosphorylated Smad2 and total Smad2/3 and Smad4 that were accumulated within the nucleus (Figure 2B). In contrast, the MEK/ERK inhibitor U0126 (10 μM) failed to prevent TGFβ-induced phosphorylation or cytoplasmic–nuclear translocalization of Smad2/3 and Smad4. These results indicated that SB431542 selectively prevented the ligand-induced nuclear accumulation of R-Smad/Smad4 complex in normal fibroblasts. No Hsp70 or histone H3 could be detected in the nuclear or cytosolic fractions, respectively, even after prolonged exposure of the immunoblots, indicating the absence of cross-contamination.

Effects of SB431542 on TGFβ target gene expression

Activation of the Smad signaling pathway is essential for maximal TGFβ responsiveness in fibroblasts. To examine the functional consequences of Smad inhibition on relevant TGFβ responses in primary fibroblasts, the effects of SB431542 on extracellular matrix synthesis were evaluated. Confluent cultures incubated in media with 0.1% FCS were pretreated with various concentrations of SB431542, followed by addition of TGFβ1. Whole cell lysates were then prepared and examined by immunoblot analysis. As expected, the levels of type I collagen, fibronectin, and PAI-1 were increased by TGFβ1 (Figure 3A). SB431542 alone consistently induced a modest decrease in the cellular levels of these proteins; furthermore, in the presence of TGFβ1, SB431542 reversed their stimulation in a dose-dependent manner.

thumbnail image

Figure 3. Effects of SB431542 on transforming growth factor β (TGFβ)–induced extracellular matrix synthesis. Fibroblasts were pretreated with the indicated concentrations of SB431542, followed by A and B, TGFβ1 or C, interleukin-13 (IL-13) for 24 hours. A, Fibroblasts were harvested and whole cell lysates were examined by immunoblot analysis using antibodies against type I collagen, fibronectin, plasminogen activator inhibitor 1 (PAI-1), and actin. Molecular size (kd) markers are shown at the left. B and C, Total RNA was isolated from fibroblasts incubated with DMSO or 10 μM SB431542 followed by TGFβ1 (B) or 10 ng/ml of IL-13 (C) for the indicated periods and examined by Northern blot analysis. Bar graphs show the relative intensities of the COL1A2 mRNA levels, normalized to those of GAPDH. TIMP-1 = tissue inhibitor of metalloproteinases 1; CTGF = connective tissue growth factor; 18S = 18S ribosomal RNA. D, Fibroblasts were harvested after 120 minutes of incubation with TGFβ1 and examined by in situ hybridization using rhodamine-labeled Smad7 or COL1A1 cDNA probes. Nuclei were identified by 4′,6-diamidino-2-phenylindole staining. Representative photomicrographs are shown (original magnification × 400). E, Following incubation of serum-starved cultures with 12.5 ng/ml of TGFβ2 for the indicated periods, the concentrations of TGFβ1 in conditioned supernatants were determined by enzyme-linked immunosorbent assay. Cultures were incubated with DMSO (open bars) or SB431542 (solid bars). Results of a representative experiment are shown. Values are the mean and SD pg/ml from triplicate determinations. = P < 0.005.

Download figure to PowerPoint

To examine changes in mRNA expression, confluent fibroblasts pretreated with SB431542 were examined by Northern blot analysis. The results showed that in both neonatal foreskin fibroblasts (Figure 3B, left) and adult dermal fibroblasts (data not shown), the TGFβ-stimulated expression of mRNA for both early-response genes, such as PAI-1, and late-response genes, such as COL1A1, COL1A2, fibronectin, TIMP-1, CTGF, and TGFβ, was completely abrogated by SB431542. Addition of SB431542 to the cultures simultaneously with, or even 60 minutes following (rather than preceding), TGFβ1 yielded identical results (Figure 4 and data not shown). In contrast to SB431542, U0126 had no effect on TGFβ-induced stimulation of any of these responses. In addition, SB431542 abrogated the rapid (90 minutes) induction of Smad7 mRNA and prevented the delayed (48 hours) suppression of Smad3 mRNA in TGFβ-treated fibroblasts, whereas levels of Smad4 mRNA remained unaltered (Figure 3B, right).

thumbnail image

Figure 4. Prevention and reversal of transforming growth factor β (TGFβ)–induced gene expression by SB431542. Fibroblasts were pretreated for 30 minutes with SB431542 followed by TGFβ1 (pre) or SB431542 was added 60 minutes after TGFβ1 (post). Following a further 24 hours of incubation, fibroblasts were harvested, and total RNA was isolated and examined by Northern blot analysis. Representative autoradiograms are shown. CTGF = connective tissue growth factor; 28S = 28S ribosomal RNA.

Download figure to PowerPoint

To further characterize the selectivity of SB431542 for TGFβ/ALK-5–mediated responses, IL-13 was used to activate fibroblasts. Like TGFβ, IL-13 is a potent stimulus for the synthesis of collagen and myofibroblast transdifferentiation (16, 17). However, in contrast to TGFβ, IL-13 exerts it stimulatory activities in fibroblasts via Smad-independent and STAT-6–dependent signaling. Therefore, confluent fibroblasts were incubated with IL-13 (10 ng/ml) following 30 minutes of pretreatment with SB431542. As shown in Figure 3C, IL-13 induced a >2-fold increase in COL1A2 mRNA expression, and this response was unaffected by pretreatment with SB431542, thus confirming the selectivity of the inhibitor for ALK-5–mediated TGFβ responses.

During in vitro culture, fibroblasts display considerable cell–cell heterogeneity in terms of collagen gene expression levels and TGFβ responses. To further examine the regulation of COL1A1 and Smad7 mRNA expression by TGFβ and its modulation by the ALK-5 inhibitor in individual fibroblasts, in situ hybridizations were performed. The results showed that the levels of fibroblast COL1A1 mRNA were uniformly up-regulated in a time-dependent manner by treatment with TGFβ; quantitation of the results indicated an 8-fold increase in COL1A1 mRNA expression at 48 hours (data not shown). The levels of Smad7 mRNA were increased 10-fold at 120 minutes of TGFβ exposure (Figure 3D), whereas GAPDH mRNA levels showed no detectable change. Consistent with the results from the Northern blot analyses, in situ hybridization revealed that pretreatment with SB431542 almost completely abrogated both the rapid induction of Smad7 mRNA expression and the delayed induction of COL1A1 mRNA expression in these fibroblasts.

One of the most important profibrotic effects of TGFβ is the induction of its own transcription in a variety of mesenchymal cell types. Such autoinduction may be responsible for sustaining or amplifying TGFβ responses in an autocrine or paracrine manner. We examined the effect of ALK-5 inhibition or TGFβ autoinduction in normal fibroblasts. For this purpose, fibroblasts pretreated with SB431542 were incubated with TGFβ2 for up to 48 hours, and TGFβ1 concentrations in the culture supernatants were determined at various time points by a colorimetric ELISA. The results from multiple experiments showed that TGFβ2 induced a time-dependent increase in TGFβ1 expression, with a nearly 2-fold increase at 48 hours compared with unstimulated controls (mean ± SD 385 ± 42 versus 208 ± 12 pg/ml; P < 0.005); SB431542 completely inhibited these effects of TGFβ2 on TGFβ autoinduction (Figure 3E). These results paralleled the effects of SB431542 on the stimulation of TGFβ mRNA expression as shown in Figure 3B.

Modulation of TGFβ-induced myofibroblast transdifferentiation

Myofibroblasts are fibroblast-derived cells characterized by a high level of α-SMA expression and incorporation into cytoskeletal stress fibers. Because they are a rich source of extracellular matrix proteins, secrete profibrotic cytokines, and are relatively resistant to apoptosis, myofibroblasts play important roles in the development of scar tissue and fibrosis (18). Previous studies have indicated that TGFβ is a potent inducer of myofibroblast transdifferentiation of normal fibroblasts (19, 20). Therefore, we examined the effect of ALK-5 inhibition on this process.

Confluent fibroblasts were pretreated with SB431542 for 30 minutes, followed by TGFβ1 for 48 hours, and mRNA expression was examined by Northern blot analysis. The results showed that while TGFβ induced a marked increase in α-SMA mRNA expression, SB431542 almost completely abrogated this response (Figure 5A). Pretreatment with the MEK inhibitor U0126 resulted in a 30–40% decrease in α-SMA mRNA levels. Because α-SMA is one of the stress fiber proteins, we examined by immunocytochemistry alterations in cell shape and cytoskeletal organization in monolayer cultures of fibroblasts. The results showed that TGFβ induced a time-dependent increase in fiber formation, reaching maximal levels at 5 days (Figure 5B). This was associated with a progressive increase in cell size. When fibroblasts were pretreated with SB431542 followed by incubation with TGFβ1 for 3 days, the dramatic increase in α-SMA fibril formation was completely blocked (Figure 5C). Staining of the nuclei with DAPI indicated that the cell numbers were unchanged.

thumbnail image

Figure 5. Modulation of myofibroblast transdifferentiation by SB431542. A. Confluent fibroblasts were pretreated with DMSO, SB431542, or U0126 (10 μM), followed by transforming growth factor β1 (TGFβ1). Total RNA was isolated after another 48 hours of incubation and examined by Northern blot analysis. Representative autoradiograms are shown. α-SMA = α-smooth muscle actin. B, Confluent fibroblasts were incubated with TGFβ1 (12.5 ng/ml) for up to 5 days, then fixed and stained with antibodies against α-SMA, examined by confocal microscopy, and visualized using rhodamine (red) or 4′,6-diamidino-2-phenylindole to identify nuclei (blue). Representative photomicrographs are shown (original magnification × 250). C, Cultures were pretreated with SB431542 (10 μM) or DMSO, followed by TGFβ1, for 72 hours prior to fixation, and α-SMA was visualized using fluorescein isothiocyanate (green) (original magnification × 250).

Download figure to PowerPoint

Effects of SB431542 on TGFβ/Smad-dependent transcriptional responses

To further investigate whether the inhibitory effects of SB431542 on fibrotic responses described above involved antagonism of Smad-mediated signaling, Smad-dependent transcriptional responses were examined in transient transfection assays. In addition to a well-characterized CAT reporter driven by a 772-bp fragment of human COL1A2, which was previously shown to be regulated by TGFβ through Smad3/Smad4, we used 2 different Smad-responsive minimal promoters widely used in the study of TGFβ signaling. The plasmid SBE4-Luc is specifically regulated by TGFβ through direct Smad3/4 binding to the CAGA sequence (12), whereas AR3-Lux contains an activin response element that is recognized by the DNA-binding transcription factor FAST-1 in response to TGFβ and is therefore a target for Smad2-dependent TGFβ signaling (12). Cells were transiently transfected with either 772COL1A2-CAT or SBE4-Luc (fibroblasts), or AR3-Lux plus expression vector for FAST-1 (Hep-G2 cells), and following pretreatment with SB431542, cells were incubated with TGFβ1 for 48 hours.

The results showed that, as expected, TGFβ induced strong (3–10-fold) activation of each of these constructs (Figure 6). In transfected Hep-G2 cells, marked activation of the AR3 reporter was induced by TGFβ when the FAST-1 expression vector was cotransfected with Smad2, which is consistent with the ability of DNA-bound FAST-1 to recruit Smad2/4 complex to the activin response element to activate transcription. Whereas by itself, SB431542 had no significant effect on the basal activities of these reporter constructs, pretreatment with SB431542 in the presence of TGFβ reduced the stimulation of the promoters in a dose-dependent manner (Figure 6). In contrast, SB431542 did not alter the activity of the internal control plasmid pRL-TK-Luc, indicating its specificity of inhibition for ALK-5/R-Smad–mediated transcriptional responses.

thumbnail image

Figure 6. Effects of SB431542 on transforming growth factor β (TGFβ)/Smad–dependent transcriptional responses. A and B, Confluent fibroblasts or C, Hep-G2 cells were transiently transfected with 772COL1A2-CAT or SBE4-Luc reporter constructs or with AR3-Lux plus a plasmid expressing Smad3 or Smad2 plus forkhead activin signal transducer 1 (FAST-1). The Renilla luciferase vector pRL-TK-Luc was cotransfected as an internal control. Cultures were then pretreated with SB431542 (10 μM or the indicated concentrations) or DMSO for 30 minutes, followed by TGFβ1 (12.5 ng/ml or the indicated concentrations). Following a further 48-hour incubation, fibroblasts were harvested and chloramphenicol acetyltransferase (CAT) or luciferase activities were determined. Values are the mean and SD arbitrary units (normalized for transfection efficiencies in each sample) of triplicate determinations from 2–3 representative experiments.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

By inducing fibroblast activation through paracrine mechanisms and by sustaining and amplifying this process via autocrine stimulation, TGFβ plays a fundamental role in normal wound repair and in pathologic fibrosis. Therefore, the expression, activity, and intracellular signaling of TGFβ must be held under tight regulation. The precise molecular mechanisms controlling the responses of the target cell to TGFβ in physiologic tissue remodeling and their deregulation in the pathogenesis of fibrosis, however, remain incompletely understood. In most cells, TGFβ signals are transduced to their transcriptional targets through a dual mechanism involving Smad family proteins and non-Smad pathways. A variety of approaches to the delineation of Smad-dependent versus Smad-independent cellular and transcriptional responses to TGFβ have yielded confusing, and sometimes even contradictory, results. For example, TGFβ stimulation of fibronectin and α-SMA expression has been shown to be Smad-dependent or Smad-independent in different studies and different cell types (21–24). However, global analysis of gene expression has revealed that >95% of transcriptional responses induced by TGFβ in mouse embryonic fibroblasts were Smad3-dependent (25). Furthermore, by using a mutant ALK-5 receptor that retains its kinase activity but is unable to activate downstream Smad signaling, a great majority of TGFβ-induced cellular responses have been shown to be dependent on Smad activation in a variety of cell lines (26).

The present studies revealed that the novel small-molecule ALK-5 inhibitor SB431542 caused a dramatic suppression of TGFβ-induced R-Smad phosphorylation and nuclear accumulation of the R-Smad/Smad4 complex in normal dermal fibroblasts. Blockade of ligand-induced Smad activation was accompanied by inhibition of Smad2- and Smad3-dependent transcriptional responses in transiently transfected dermal fibroblasts. Inhibition by SB431542 appeared to be selective for the Smad pathway, because under the same experimental conditions, TGFβ induction of ERK-1/2 phosphorylation and ERK kinase activity were unaffected by SB431542, as was the more modest stimulation of JNK and p38 phosphorylation. Previous studies have also indicated that ERK-1/2 is activated by TGFβ in mesenchymal cell types (27), but the precise mechanisms underlying this response remain unknown. The results of the present study indicate that TGFβ-induced rapid activation of ERK-1/2 in dermal fibroblasts was Smad-independent. Pretreatment with SB431542 abrogated TGFβ-induced stimulation of extracellular matrix gene products that play important roles in the development of fibrosis. These include type I collagen (the principal structural component of connective tissue in the skin), fibronectin, and TIMP-1 and PAI-1 (which prevent the activities of proteolytic enzymes and thereby contribute to connective tissue accumulation in fibrosis). Significantly, SB431542 not only prevented TGFβ-induced stimulation, but it was also effective in reversing these responses when added to cultures following TGFβ.

Although it has previously been reported that stimulation of fibronectin synthesis by TGFβ is independent of Smads (21), the present results indicate that in dermal fibroblasts, the fibronectin response was abrogated by SB431542, suggesting a Smad-dependent mechanism, which is also consistent with previous reports (28). Northern blot analysis and in situ hybridizations indicated that SB431542 itself had no effect on basal Smad7 mRNA expression levels, but prevented their rapid induction by TGFβ, consistent with the Smad3/4 dependence of this response (29). Therefore, the inhibitory activities of SB431542 on basal and TGFβ-induced extracellular matrix gene expression could not be attributed to the induction of an endogenous inhibitor of TGFβ signaling. Of interest was the finding that by itself, SB431542 reduced the cellular levels of extracellular matrix proteins, suggesting that their basal expression levels in fibroblasts were regulated, at least in part, by endogenous TGFβ through autocrine stimulatory loops involving ALK-5 and Smad activation. Furthermore, the results also demonstrated that ALK-5 blockade reduced the secretion of endogenous TGFβ1 in TGFβ2-stimulated fibroblasts and prevented the increase in TGFβ and CTGF mRNA expression. Autoinduction of TGFβ, a phenomenon important in amplifying the magnitude and duration of cellular responses to TGFβ, has been shown in previous studies with Smad3-null fibroblasts to be Smad3-dependent (24, 30). By inhibiting TGFβ autoinduction, SB431542 could potentially exert a powerful effect, arresting the progression of fibrotic responses.

Importantly, SB431542 prevented TGFβ-induced mRNA expression of α-SMA, its intracellular accumulation, and its incorporation into cytoskeletal stress fibers, the hallmarks of myofibroblast transdifferentiation. Myofibroblasts are terminally differentiated cells with morphologic and functional characteristics that are between those of fibroblasts and those of smooth muscle cells (18). In wound healing, myofibroblasts are thought to arise locally from quiescent fibroblasts under the influence of TGFβ released during tissue injury. In vitro, TGFβ is a potent stimulus of α-SMA expression and organization of filamentous actin into stress fibers, resulting in the transdifferentiation of various types of mesenchymal cells into myofibroblasts. The present findings, together with a recent report demonstrating the requirement for an intact L45 loop of the TβRI (24), suggest that TGFβ-induced fibroblast transdifferentiation into myofibroblasts is an ALK-5/Smad-dependent process. Furthermore, the fact that U0126 partially prevented myofibroblast differentiation induced by TGFβ suggests a contribution of MAP kinase pathways to this response as well. Because myofibroblasts play critical roles in the pathogenesis of tissue fibrosis, these inhibitory effects of SB431542 may be particularly significant in ameliorating the fibrotic process.

Previous studies have indicated that while TGFβ was pivotal in the regulation of myofibroblast differentiation, overexpression of Smad2, but not Smad3, induced myofibroblast features in the absence of TGFβ in lung fibroblasts (19). Although Smad2 and Smad3 are highly homologous, these 2 R-Smads show distinct biologic activities. In particular, Smad2 is considered to play pivotal roles during embryogenesis, whereas Smad3 is critical for cellular responses after birth. In our own previous studies, we have shown that dermal fibroblasts derived from Smad3-null mice retained their ability to induce α-SMA expression and to organize filamentous actin into stress fibers in response to TGFβ, consistent with involvement of Smad2 (31). In contrast to these results, Smad3 was shown to be both necessary and sufficient for myofibroblast transdifferentiation in lung fibroblasts (23). The reasons for these apparently discrepant findings relating to the roles of Smad2 and Smad3 in TGFβ regulation of myofibroblast differentiation in different cell types remain unknown.

Taken together, the present results indicate that the novel small-molecule ALK-5 inhibitor SB431542 is a potent and selective antagonist of TGFβ-induced R-Smad activation in normal dermal fibroblasts. Furthermore, inhibition of ALK-5–mediated Smad signaling was associated with a striking reduction of multiple TGFβ-induced cellular responses involved in fibrogenesis in the absence of detectable cytotoxicity. These results establish a compelling direct link between ligand-induced R-Smad phosphorylation and nuclear accumulation and the transcriptional activation of important target genes. Furthermore, the results indicate that in primary dermal fibroblasts, TGFβ stimulation of extracellular matrix gene expression and myofibroblast transdifferentiation are largely ALK-5/Smad-dependent processes, which is consistent with our own previous findings (32) as well as those of other investigators utilizing mutant ALK-5 receptors or Smad-null embryonic fibroblasts (26). Pharmacologic inhibition of ALK-5 and downstream signaling events may provide a novel approach to the precise delineation of the mechanisms of individual cellular responses elicited by TGFβ.

Selective inhibition of ALK-5 signaling using SB431542 may have important implications for the therapeutic control of pathologic fibrotic responses. In light of the widely recognized pivotal role of Smads and TGFβ in the pathogenesis of fibrosis (33), therapeutic strategies targeting TGFβ are currently under active study. Traditional approaches focus on prereceptor targeting of TGFβ by the use of neutralizing antibodies, naturally occurring inhibitors, or soluble TGFβ receptors in order to reduce the local expression or biologic activity of TGFβ. However, in fibrotic conditions such as SSc, lesional fibroblasts display intrinsic abnormalities in Smad signaling, such as loss of endogenous repressors, that could potentially result in their abnormal responsiveness to TGFβ (for review, see ref.34). Some of the autonomous alterations in the TGFβ pathway described to date in SSc fibroblasts include reduced expression of the inhibitory Smad7 (35), defective activation-induced degradation of cell surface receptors for TGFβ (36), and constitutive nuclear accumulation of activated R-Smads (10). Additional abnormalities of SSc fibroblasts, such as defective regulation of the Smad-associated transcriptional corepressors Ski and SnoN, are currently under investigation.

Together, these intrinsic abnormalities in the regulation of TGFβ/Smad signaling would have the net effect of reducing the target cell's threshold for exogenous TGFβ, markedly enhancing its sensitivity to stimulation. In such a situation, even very small amounts of active TGFβ would be sufficient to elicit full-scale biologic responses. Current therapeutic strategies targeting TGFβ at the prereceptor level are unlikely to be able to reduce TGFβ signaling below the subthreshold levels required for activating hypersensitive SSc fibroblasts. Therefore, strategies for inhibiting TGFβ responses by blocking intracellular signaling via the TGFβ/ALK-5/Smad axis may be more effective in preventing the activation of fibroblasts that are hypersensitive to TGFβ. Whether ALK-5 inhibitors can prevent the constitutive activation of Smad signaling observed in SSc fibroblasts (10) is currently under investigation.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We are grateful to Drs. N. Laping (GlaxoSmithKline, King of Prussia, PA), R. Derynck (University of California, San Francisco), P. ten Dijke (Ludwig Institute for Cancer Research, Uppsala, Sweden), J. Massague (Howard Hughes Medical Institute, New York, NY), B. Vogelstein (Johns Hopkins University, Baltimore, MD), and K. Miyazono (Cancer Institute, Tokyo, Japan) for providing us with plasmids and reagents. Members of the laboratory staff provided helpful discussions.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES