ARTHRITIS & RHEUMATISM Vol. 46, No. 11, November 2002, pp 3000–3009 DOI 10.1002/art.10621 © 2002, American College of Rheumatology Transforming Growth Factor ␤ Induces Fibroblast Fibrillin-1 Matrix Formation Eugene Y. Kissin, Raphael Lemaire, Joseph H. Korn, and Robert Lafyatis normal or scleroderma fibroblasts. However, TGF␤ did not alter the expression of either soluble fibrillin protein or fibrillin mRNA. Conclusion. Our data show that TGF␤ induces fibrillin protein incorporation into the extracellular matrix without affecting fibrillin gene expression or protein synthesis, suggesting that fibrillin matrix assembly is regulated extracellularly. TGF␤ might increase fibrillin matrix by activating myofibroblasts. Such TGF␤-mediated effects could account for the increased fibrillin matrix observed in SSc skin. Objective. Fibrillin, an extracellular matrix protein implicated in dermal fibrosis, is increased in the reticular dermis of systemic sclerosis (SSc) skin. We undertook this study to investigate the hypothesis that transforming growth factor ␤ (TGF␤) or other cytokines regulate fibrillin matrix formation by normal and SSc fibroblasts. We further investigated the mechanism of TGF␤-induced fibrillin fibrillogenesis and its relationship to myofibroblasts. Methods. Fibrillin and fibronectin matrix deposition and ␣-smooth muscle actin expression by fibroblast cultures from normal and SSc skin treated with TGF␤ or other cytokines were analyzed by immunofluorescence. Supernatant and extracellular matrix from normal and SSc fibroblasts treated with or without TGF␤ were evaluated by Western blot and Northern blot for fibrillin protein and messenger RNA (mRNA) expression, respectively. Results. Immunofluorescence demonstrated increased fibrillin matrix formation by normal and scleroderma fibroblasts after TGF␤ treatment. Other cytokines, including tumor necrosis factor ␣, interleukin-1␤ (IL-1 ␤ ), IL-4, granulocyte–macrophage colonystimulating factor, and platelet-derived growth factor, did not affect fibrillin fibrillogenesis. Fibrillin matrix formed in proximity to myofibroblasts and independently of up-regulation of fibronectin matrix or cell number. Western blot analysis of extracellular matrix confirmed increased fibrillin after TGF␤ stimulation of Transforming growth factor ␤ (TGF␤) is an important mediator of fibrosis. TGF␤ stimulates the synthesis and assembly of matrix proteins such as fibronectin and collagen (1,2). TGF␤ also promotes the formation of myofibroblasts, activated fibroblasts responsible for the generation of contractile force (3). Many studies have suggested a role for TGF␤ in the pathogenesis of scleroderma, or systemic sclerosis (SSc), a systemic disease that leads to fibrosis of the skin and internal organs (4). The effects of TGF␤ in vitro and in vivo mimic many of the observed changes in SSc skin pathology. These include increased collagen and fibronectin matrix formation (1,4) and increased connective tissue growth factor expression (5). Elevated levels of TGF␤ in several organs in SSc patients have been reported. TGF␤2 has been shown in some studies, but not in all, to be overexpressed in SSc skin (6–9). Some reports, but not all, also indicate increased levels of TGF␤ (particularly TGF␤1) in bronchoalveolar lavage fluids from SSc patients with pulmonary fibrosis (10,11). Other studies have indicated that fibroblasts from SSc patients are more sensitive to TGF␤ than are normal fibroblasts (12,13). A potential mechanism for increased sensitivity was suggested through the observation that TGF␤ receptor type I (TGF␤RI) and TGF␤RII are overexpressed in scleroderma fibroblasts compared with control fibroblasts (14). Genetic studies have also sup- Supported by USPHS grants AR-32343 and AR-07598, by the Scleroderma Research Fund, and by an Arthritis Foundation Basic Science Research Grant. Eugene Y. Kissin, MD, Raphael Lemaire, PhD, Joseph H. Korn, MD, Robert Lafyatis, MD: Boston University School of Medicine, Boston, Massachusetts. Address correspondence and reprint requests to Robert Lafyatis, MD, Boston University Medical Center, Arthritis Center, K-5, 80 East Concord Street, Boston, MA 02118. E-mail: rlafyatis@ medicine.bu.edu. Submitted for publication February 15, 2002; accepted in revised form August 5, 2002. 3000 INDUCTION OF FIBRILLIN FIBRILLOGENESIS BY TGF␤ ported a role for TGF␤ in SSc. Certain alleles for TGF␤2 and TGF␤3 are encountered more frequently in patients with SSc (15). Fibrillin is a connective tissue protein produced by fibroblasts and smooth muscle cells that helps form extracellular microfibrils and has also been implicated in the pathogenesis of fibrosis (16). Fibrillin-containing microfibrils interact both with basement membranes and with nearby elastic fibers and may help stabilize the extracellular matrix by forming a scaffold (17). Several studies have implicated fibrillin in the pathogenesis of scleroderma (18). In biopsy samples from affected skin, fibrillin lacking associated elastin is markedly increased in the lower reticular dermis in a randomly arranged pattern (19). Further supporting the role of fibrillin in scleroderma is the observation that the murine tight skin 1 (Tsk1) phenotype is due to an in-frame duplication of exons 17–40 of the fibrillin-1 gene (20). Dermal pathology in these mice has similarities to human SSc (21). Studies in humans have also supported a genetic link between fibrillin mutations and SSc. A study of Choctaw American Indians, a population with a high prevalence of SSc, has shown an association of a 2-cM haplotype that contains 2 markers for the fibrillin-1 gene with SSc (22). Further studies have indicated that a single-nucleotide polymorphism in the 5⬘-untranslated region of fibrillin is strongly associated with SSc (23). Together, these observations suggest that abnormal fibrillin deposition might contribute to fibrosis in SSc. We have investigated the potential role on fibrillin regulation of a variety of cytokines implicated in fibrosis or SSc. We show here that TGF␤, but not other cytokines tested, stimulates the incorporation of fibrillin into a fibrillar matrix by both normal and SSc fibroblasts. Strikingly, we show that TGF␤ stimulates fibrillin fibrillogenesis without affecting messenger RNA (mRNA) expression or protein secretion. Thus, TGF␤ stimulates fibrillin fibrillogenesis at the step of fibrillin incorporation into the extracellular matrix. MATERIALS AND METHODS Cell culture. Fibroblasts were grown from skin biopsy samples obtained from the forearms of normal volunteers or patients with SSc. The epidermis of the biopsy specimen was removed and the dermis was sectioned into submillimeter pieces. These dermal fragments were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% heatinactivated fetal calf serum and penicillin G (100 units/ml)/ streptomycin (100 g/ml) at 37°C in an incubator in 8% CO2. 3001 Cells reached confluence in 3–4 weeks and were used at passages 3–7. Immunofluorescence. Cultured human dermal fibroblasts (0.2 ml of 3 ⫻ 105/ml ⫽ 6 ⫻ 104/well) were plated onto 8-well chamber slides in 10% fetal bovine serum (FBS) containing DMEM. Chamber slides were examined 24 hours after plating to ensure confluent cell cultures prior to replacement of the media with serum-free DMEM. Human TGF␤1 (R&D Systems, Minneapolis, MN), platelet-derived growth factor type BB (PDGF-BB; Gibco BRL, Rockville, MD), interleukin-1␤ (IL-1␤; Endogen, Cambridge, MA), IL-4 (R&D Systems), tumor necrosis factor ␣ (TNF␣; Gibco BRL), or granulocyte–macrophage colony-stimulating factor (GM-CSF; McKesson BioServices, Rockville, MD) was added to some chamber wells at the indicated concentrations. After 4 days, supernatants were aspirated and the cells were fixed with 100% methanol for 5 minutes. Cells were washed with Tris buffered saline (TBS; 50 mM Tris [pH 8.0], 150 mM NaCl) and then incubated with rabbit antifibrillin antibody (pAb9543; kindly provided by Dr. Lynn Sakai [17,24]) at 1:250 dilution in TBS for 2 hours at 37°C. This antibody was raised against the amino-terminal half of human fibrillin-1 and cross-reacts with mouse fibrillin-1, but not with fibrillin-2 (17,24). Some cells were also incubated with mouse monoclonal anti–␣-smooth muscle actin (anti–␣-SMA) antibodies (clone 1A4; Sigma, St. Louis, MO) at a 1:250 dilution in TBS during the same time period. Cells were washed 3 times and incubated with rhodamine-conjugated donkey anti-rabbit antibody as well as with either fluorescein isothiocyanate (FITC)– conjugated goat anti-mouse antibody (Jackson ImmunoResearch, West Grove, PA) or FITC-conjugated mouse antifibronectin antibody (Jackson ImmunoResearch) for 1 hour. Cells were washed again and nuclei were stained for 1 minute using Hoechst reagent (100 ng/ml in phosphate buffered saline [PBS]). Chambers were removed from the slides, and cells were visualized using a fluorescence microscope (Olympus, Lake Success, NY) and photographed using a DC 120 Zoom digital camera (Eastman Kodak, Rochester, NY). Immunofluorescence results were quantified using image analysis software (version 1.62; NIH Image, National Institutes of Health, Bethesda, MD; online at http://rsb.info.nih.gov/nih-image/). Images were analyzed by setting the density slice option so that positively stained fibers were measured. Western blot. Dermal fibroblasts were passaged in 10% serum containing DMEM for 24 hours, and the media were replaced with serum-free DMEM. Cells were then treated with 5 ng/ml of TGF␤ or were left untreated. After 48 hours, supernatants, cell lysates, and extracellular matrix were prepared for sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) analysis similar to that described by Kitahama et al (25). Supernatants (1 ml) were mixed with 273 l of 10% trichloroacetic acid for 30 minutes on ice and centrifuged at 16,000g for 20 minutes at 4°C, and the resulting pellet was resuspended in 100 l of 2⫻ SDS-PAGE sample buffer (0.125M Tris [pH 6.8], 20% [weight/volume] glycerol, 4.6% SDS). Cells and extracellular matrix were scraped in 400 l of 1% Nonidet P40 in 50 mM Tris, pH 8.0, supplemented with 1 mM phenylmethylsulfonyl fluoride. Fifty microliters of sample was saved for bicinchoninic acid assay for total protein con- 3002 KISSIN ET AL Figure 1. Induction of fibrillin fibrillogenesis in dermal fibroblasts by transforming growth factor ␤ (TGF␤). A, Left panels show nuclei stained with Hoechst reagent (blue); right panels show overlays with nuclei stained with Hoechst reagent (blue) and fibrillin fibers stained with rhodamine (red). Fibrillin can be seen most prominently in areas of greatest cell density. Cultures were left untreated (control) or treated with 5 ng/ml TGF␤, as indicated. B, Rhodamine immunofluorescence in A is shown graphically after image analysis. C, In contrast to cells treated with 5 ng/ml TGF␤, treatment with tumor necrosis factor ␣ (TNF␣; 10 ng/ml), interleukin-1␤ (IL-1; 1 ng/ml), IL-4 (5 ng/ml), or granulocyte–macrophage colony-stimulating factor (GM-CSF; 5 ng/ml) did not increase matrix fibrillin (red fibers) over that seen in control cells. (Original magnification ⫻ 20 in A and C.) centration according to the supplied protocol (Micro BCA Protein Assay; Pierce, Rockford, IL), while the remainder was vortexed and centrifuged at 16,000g for 10 minutes at 4°C. The supernatant, cell lysate fraction, was mixed with 400 l of 2⫻ SDS-PAGE sample buffer. The pellet, matrix fraction, was resuspended in 100 l of 1.2⫻ SDS-PAGE sample buffer containing 10% ␤-mercaptoethanol. Samples were heated to 95°C for 3 minutes. Fifteen micrograms of total protein from normal fibroblast culture supernatants and matrix fractions and 40 g of total protein from scleroderma fibroblast culture supernatants and matrix fractions were analyzed on 5% Tris/glycine SDS– polyacrylamide gels. Proteins were transferred to nitrocellulose in 10 mM borax buffer (50 mM Tris, 380 mM glycine, 0.1% SDS, 20% methanol, and 10 mM borax). Membranes were blocked with 5% nonfat milk in PBS and incubated with rabbit antifibrillin primary antibody (pAb9543, described above) at a 1:1,000 dilution in PBS for matrix protein and a 1:4,000 dilution in PBS for supernatant protein for 2 hours. After washing in 0.05% PBS–Tween (PBST), blots were incubated with peroxidase-conjugated anti-goat antibody (Jackson ImmunoResearch) for 1 hour at 1:1,000 dilution in PBS for matrix protein and 1:4,000 dilution in PBS for supernatant protein and again washed in 0.05% PBST. Blots were developed by enhanced chemiluminescence detection (SuperSignal, West PICO chemiluminescent substrate; Pierce). Northern blot. Cultured human dermal fibroblasts were passaged in DMEM containing 10% FBS for 24 hours and the media replaced with serum-free DMEM supplemented with 0.1 mg/ml bovine serum albumin. For Northern blot analyses, cells were harvested for RNA using the RNeasy mini protocol for isolation of total RNA according to the supplied protocol (Qiagen, Valencia, CA). RNAs were analyzed by formaldehyde/1% agarose gel electrophoresis and INDUCTION OF FIBRILLIN FIBRILLOGENESIS BY TGF␤ 3003 Figure 2. Induction of fibrillin fibrillogenesis in systemic sclerosis (SSc) fibroblasts by transforming growth factor ␤ (TGF␤). Shown is a representative SSc fibroblast culture stained for fibrillin with rhodamine (red) and counterstained with Hoechst reagent (blue). SSc fibroblasts show little baseline deposition of fibrillin matrix (control), but considerable fibrillin matrix formed after treatment with TGF␤. (Original magnification ⫻ 40.) transferred to Nytran as described previously (26). Fibrillin mRNA was detected by hybridization to a 32P-labeled complementary DNA fragment of a 593-nucleotide human fibrillin-1 complementary DNA. By BLAST comparison, this region shows no homology with fibrillin-2. This fragment was obtained by reverse transcription of poly(dT)-primed human dermal fibroblast RNA followed by polymerase chain reaction using primers (hFBN-1 5⬘, bp 1970–1991, 5⬘-TGCGGAGCACATGCTATGGTGG-3⬘ and hFBN-1 3⬘, bp 2562–2539, 5⬘-CCAGCAAGTGCCCTTGATGGTTTC-3⬘) spanning bp 1970–2562 of the fibrillin-1 mRNA. Signals were detected using a Cyclone phosphorimager (Hewlett-Packard, McMinnville, OR) and quantified using the OptiQuant software provided. RESULTS Fibrillin fibrillogenesis stimulated by TGF␤, but not by other cytokines. To investigate the effect of cytokines implicated in fibrosis or SSc, cultured dermal fibroblasts from normal volunteers were stimulated with TGF␤, IL-4, TNF␣, IL-1␤, GM-CSF, or PDGF-BB. Only stimulation with TGF␤ resulted in increased fibrillin fiber formation (Figures 1A and B). After 4 days, TGF␤-stimulated fibroblasts appeared more dense and aggregated to discrete areas on the slide surface. These cultures revealed heavy fibrillin deposition over cell clusters, with more sparse deposition in other areas. While this pattern was observed in cultures both with and without TGF␤ treatment, cultures treated with TGF␤ showed increased cell number, increased cell aggregation, and increased fibrillin deposition (Figures 1A and B). Low-level fibrillin staining in control cultures may have been partly due to lower plating density of cells used in the current study compared with that used in other studies (27). Dermal fibroblasts from skin affected by scleroderma also showed increased fibrillin matrix with TGF␤ (Figure 2). Next, to determine whether TGF␤ increases fibrillin production simply by increasing the cell number, we counterstained cells for nuclei. We were thus able to compare cell clusters with roughly equivalent cell densities. We found that TGF␤ increased fibrillin deposition even when cell density was controlled for (Figure 3A, panels a and b). To further control for the effects of cell number, normal fibroblasts were stimulated with PDGF, another known fibroblast mitogen. While PDGF also increased the cell number, its effect on fibrillin matrix deposition was minimal compared with that of TGF␤ 3004 Figure 3. No dependence on increased cell number of transforming growth factor ␤ (TGF␤) induction of fibrillin incorporation into matrix. A, Identical regions of normal dermal fibroblasts costained for fibrillin (with rhodamine; red) (a–c) and fibronectin (with fluorescein isothiocyanate; green) (d–f). Fibroblast cultures were treated with 5 ng/ml TGF␤ (b and e) or 5 ng/ml platelet-derived growth factor (PDGF) (c and f). Hoechst staining (blue) shows nuclei in all panels. To control for the effect of TGF␤ on cell proliferation, areas of the culture slide containing approximately equal numbers of cells with and without TGF␤ treatment are shown (92 control cells in a and d; 75 TGF␤-treated cells in b and e). PDGF induces fibronectin fibrillogenesis (f), but not fibrillin fibrillogenesis (c). Arrowheads in a and d indicate colocalization of some fibrillin and fibronectin matrix strands (original magnification ⫻ 20). B, Fibrillin and fibronectin immunofluorescence in A shown graphically after image analysis. (Figure 3A, panels b and c). Therefore, the effect of TGF␤ on fibrillin matrix formation is not mediated only through changes in cell number. KISSIN ET AL Since other groups of investigators have shown that fibronectin colocalizes with fibrillin and have suggested that fibronectin may act as a template for fibrillin fibrillogenesis (28), we tested whether TGF␤ increases fibrillin matrix as a byproduct of increasing fibronectin formation. We first costained for fibrillin and fibronectin matrix in cultured dermal fibroblasts and found partial colocalization of the two proteins (Figure 3A, panels a and d). Investigators in our group have previously shown that PDGF and TGF␤ stimulated fibronectin fibrillogenesis by synovial fibroblasts (29). We therefore compared the effects of TGF␤ with those of PDGF on fibrillin and fibronectin formation. PDGF induced substantially greater fibronectin fibrillogenesis with relatively little effect on fibrillin fibrillogenesis, while TGF␤ up-regulated both processes (Figures 3A, panels b, c, e, and f, and 3B). Association of fibrillin fibrillogenesis with myofibroblasts. Fibronectin and other matrix proteins are produced at increased levels by myofibroblasts. Since TGF␤ is known to induce myofibroblast formation, we costained cell cultures for ␣-SMA, a marker of myofibroblasts, in order to determine whether fibrillin matrix is associated with myofibroblast formation. Untreated cultures showed only occasional ␣-SMA–positive cells. Cultures treated with TGF␤ showed a dramatic increase in ␣-SMA–positive cells (Figure 4), as previously reported (30). TGF␤-induced fibrillin fibers were generally associated with ␣-SMA–positive cells (Figure 4, white arrows). Myofibroblasts were associated with cell aggregates, and such foci also showed increased fibrillin. Although immunofluorescent staining for fibrillin generally localized to culture regions containing myofibroblasts, the overlap was not complete (Figure 4, blue arrow). Some ␣-SMA–positive cells showed no fibrillin, and some fibrillin fibers could be seen in association with ␣-SMA–negative cells. Induction of fibrillin fibrillogenesis by TGF␤ without change in fibrillin mRNA or protein expression. Many effects of TGF␤ are mediated through changes in gene expression. To further elucidate the mechanism by which TGF␤ exerts its effect on fibrillin matrix formation, we analyzed the effect of TGF␤ on fibrillin mRNA expression. Surprisingly, TGF␤ had no effect on fibrillin mRNA expression by either normal or SSc dermal fibroblasts (Figure 5). We next determined whether TGF␤ affects fibrillin fibrillogenesis through increased fibrillin protein synthesis or secretion. We analyzed supernatant and matrix fractions from dermal fibroblasts with and without TGF␤ treatment. Matrix was solubilized using SDS-PAGE sample buffer containing 10% INDUCTION OF FIBRILLIN FIBRILLOGENESIS BY TGF␤ 3005 Figure 4. Fibrillin matrix production by myofibroblasts. Dermal fibroblasts that were untreated (top panels) or were treated with 5 ng/ml of transforming growth factor ␤ (TGF␤) (bottom panels) were costained for fibrillin fibers with rhodamine (red) (left panels) and for myofibroblasts with fluorescein isothiocyanate (green) (right panels). Cell cultures were costained with ␣-smooth muscle actin (␣-SMA), a marker of myofibroblasts, to determine whether fibrillin matrix is associated with myofibroblast formation. White arrows indicate colocalization of myofibroblasts with fibrillin matrix (␣-SMA–positive cells). Blue arrow (upper right panel) indicates myofibroblasts without surrounding fibrillin fibers. Stimulation with TGF␤ produced increased numbers of myofibroblasts and surrounding fibrillin fiber formation (bottom panels). (Original magnification ⫻ 20 .) ␤-mercaptoethanol. Reduction of disulfide bonds by ␤-mercaptoethanol extracts a proportion of fibrillin bound to matrix, although perhaps not all (31). In normal fibroblast cell lines, TFG␤ did not significantly affect the amount of fibrillin in the soluble/supernatant fraction of cell cultures (Figure 6A). Despite this, TGF␤ increased the amount of fibrillin in the matrix fraction of cultured fibroblasts, as expected from our immunofluorescence results (compare Figure 6 with Figures 1–3). Antibody specificity for fibrillin was confirmed in control experiments using fibroblast supernatants from control (pa/pa) and Tsk1 (tsk/pa) mice. Control mouse fibroblast supernatants showed a single high molecular weight band of the same size as that seen in the human immunoblots (Figure 6B). An additional higher molecular weight band was seen in the supernatant from Tsk1 fibroblasts corresponding to mutated fibrillin, containing a large in-frame insertion (Figure 6B). Similar results were obtained from SSc fibroblast cell lines (Figure 6C). As in normal fibroblasts, matrix fibrillin was increased in all cell lines. Although TGF␤ led to a modest increase in soluble fibrillin in one SSc cell line, the change in matrix-associated protein was much more dramatic. In all cell lines tested, the amount of soluble fibrillin greatly exceeded that found in the matrix. This finding suggests that the availability of soluble fibrillin is not the limiting step in fibrillin fibrillogenesis. There was no consistent difference in basal or TGF␤-induced soluble or matrix-associated fibrillin between normal and scleroderma cell lines (data not shown). Differences in matrix formation at baseline and after TGF␤ treatment correlated with the passage number. Both normal and SSc fibroblast cultures of low passage formed greater amounts of fibrillin matrix than did those of higher passage (data not shown). DISCUSSION The generation of extracellular matrix begins with synthesis of matrix proteins and is followed by 3006 Figure 5. Fibrillin expression in fibroblasts treated with TGF␤. Normal and SSc dermal fibroblast cultures were left untreated or were treated with 5 ng/ml of TGF␤ for 48 hours, and then mRNA was purified, blotted, and hybridized with a fibrillin-1 cDNA probe. Signal was detected using a phosphorimager. The blot was then rehybridized with an 18S rRNA probe, and the signal was detected again on a phosphorimager. Signal intensities were analyzed using OptiQuant software, fibrillin expression was normalized to the 18S rRNA signal, and fibrillin expression in TGF␤-treated cells was normalized to fibrillin-1 expression in untreated cells. See Figure 2 for definitions. formation of an insoluble matrix. Incorporation of collagen and fibronectin into the extracellular matrix provides examples of matrix fibrillogenesis. Type I collagen assembles passively into matrix by self-polymerization (32), although certain proteins, including other collagen types, fibromodulin, lumican, decorin, and tenascin, can modify this process (33–35). In contrast, fibronectin matrix assembly is an active process and requires several conditions: an intact intracellular cytoskeleton, an activated fibronectin-binding integrin receptor, and tension across the developing fibrillar structure (36,37). Investigators in our group showed previously that TGF␤ and PDGF stimulate fibronectin matrix assembly, possibly through changes in integrin receptor activity (29). Our results demonstrate that TFG␤ also stimulates fibrillin fibrillogenesis. The action of TGF␤ on fibrillin was not mediated through increased cell number alone, since stimulation of fibroblasts with PDGF increased cell number and density, but had no effect on fibrillin fibril formation. Further, TGF␤ did not change fibrillin mRNA expression or secretion, despite dramatically increasing fibrillin fibrillogenesis. Thus, our data indicate that TGF␤ induces fibrillar fibrillin by affecting the incorporation of soluble fibrillin into cross-linked fibrillin matrix. Our understanding of this process is KISSIN ET AL incomplete, but the process starts with release of soluble fibrillin, which is then incorporated into matrix fibrils. Like fibronectin, fibrillin incorporation into matrix fibrils involves several steps, including covalent cross-linking through intermolecular cysteine bonds (31). Heparin inhibits fibrillin fibrillogenesis, suggesting that as-yet-undefined active mechanisms might regulate fibrillin matrix assembly (38). Our data, which show that the formation of fibrillin matrix is independent of soluble fibrillin concentration in the media, suggest that, like fibronectin matrix formation, fibrillin matrix formation is an active process. Our data showing that TGF␤ stimulates this process further support active, perhaps cell-directed, fibrillin matrix assembly. The mechanisms that underlie this process and the effect of TGF␤ remain unclear. Some investigators have suggested that fibrillin matrix depends on fibronectin matrix assembly (28). While a minimal amount of fibronectin may be necessary for fibrillin scaffold formation, up-regulation of fibronectin with PDGF did not produce similar increased fibrillin matrix. It is noteworthy, however, that fibrillin, like fibronectin, contains an RGD integrin binding site (39), with highest affinity for ␣v␤3 receptors, suggesting that like fibronectin, fibrillin binding to integrins might stimulate fibrillin matrix formation. Fibrillin is found in increased amounts in the lower reticular dermis of scleroderma skin. Previous investigation has shown that TGF␤ regulates collagen and fibronectin matrix formation and suggests that it may induce fibrosis in SSc skin. The present study did not show a difference in basal or TGF␤-induced fibrillin matrix formation between normal and SSc fibroblasts, consistent with the observations by Wallis et al (40). Wallis et al also showed that fibrillin matrix produced by SSc fibroblasts is less stable than that produced by normal fibroblasts (40). Although these intriguing observations are similar to an observed decrease in stability of fibrillin produced by Tsk1 mouse fibroblasts (41), they do not readily explain the presence of increased fibrillin in the dermis of SSc skin. Our results suggest that the excess fibrillin found in SSc skin may be the result of increased TGF␤ secreted by endogenous or infiltrating cells in SSc skin, or of increased sensitivity to TGF␤. Potentially, TGF␤ may trigger both fibroblast production of excess collagen and excess fibrillin fibrillogenesis. Our observation that TGF␤, but not other cytokines that induce collagen and/or are implicated in SSc (including PDGF, TNF␣, IL-1␤, IL-4, or GM-CSF), causes increased fibrillin fibrillogenesis supports the notion that TGF␤ plays a particularly prominent role in this disease. Fibrillin incorporation into the matrix was asso- INDUCTION OF FIBRILLIN FIBRILLOGENESIS BY TGF␤ 3007 Figure 6. Increased extracellular matrix accumulation, but not secreted fibrillin, induced in dermal fibroblasts by TGF␤. A, Normal or C, SSc fibroblasts were left untreated or treated with 5 ng/ml TGF␤ for 24 hours, and proteins were precipitated from supernatants (soluble) or extracted from extracellular matrix (matrix). Total protein was assayed and equal amounts were loaded in each lane. Nonreduced (soluble) and reduced (matrix) proteins were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and blotted to nitrocellulose, and fibrillin was detected as described in Materials and Methods. B, Proteins from wild-type (WT) control (pa/pa) mouse or tight skin 1 (tsk/pa) mouse fibroblast supernatants were precipitated and analyzed as in A and C. See Figure 2 for other definitions. ciated with myofibroblasts, and TGF␤ is known to induce myofibroblast formation. The ability of TGF␤ to up-regulate ␣-SMA expression depends on the cellular growth phase. In proliferating fibroblast cultures at low density, TGF␤ stimulates less production of ␣-SMA than it does under nonproliferating, high-density conditions (42). This may explain the association of fibrillin matrix with regions of increased cell density, if myofibroblasts are responsible for integrating soluble fibrillin into matrix. Myofibroblasts share many properties with smooth muscle cells, and others have shown that fibrillin matrix is assembled faster by smooth muscle cells than by fibroblasts (43). Furthermore, ultrastructural analyses revealed fibrillin fibers assembled by smooth muscle cells to be of greater length than those assembled by skin fibroblasts. Myofibroblasts are also increased in SSc and could therefore account for increased fibrillin matrix in SSc skin. Alternatively, fibrillin matrix might protect myofibroblasts from apoptosis. Jelaska and Korn have shown previously that prolonged treatment with TGF␤ induces myofibroblasts and protects dermal fibroblasts from apoptosis (30). Perhaps fibrillin, alone or through interactions with other matrix proteins, protects myofibroblasts from apoptosis (44–46). Observations in mice with mutations in fibrillin suggest that fibrillin may play a more primary role in SSc by regulating connective tissue homeostasis and fibrosis. Most significantly, Tsk1 mice show dermal fibrosis histologically similar to that seen in SSc skin (20). It is not known how mutant fibrillin leads to fibrosis, but this observation suggests that fibrillin may normally regulate connective tissue structure. Pereira et al have suggested 3008 KISSIN ET AL a similar role for fibrillin based on the vascular pathology of mice harboring a targeted homozygous mutation in fibrillin that leads to an internally deleted protein (47). Homozygous mutant mice develop Marfan’s syndrome–like features and die perinatally due to vascular complications. Surprisingly, despite markedly disorganized structure of the vascular wall, these animals have relatively preserved elastic fibers. These and other results have challenged the previously held belief that fibrillin functions primarily in elastic fibers, and suggest that fibrillin helps to organize connective tissue structure (48). Therefore, not only does increased fibrillin deposition suggest increased TGF␤ activity, it also suggests that TGF␤ may mediate increased fibrosis in SSc skin. In summary, we show that TGF␤ induces markedly increased fibrillin matrix by stimulating soluble fibrillin incorporation into the extracellular matrix. These observations indicate a possible role for TGF␤ in both human and murine fibrotic diseases through increased accumulation of matrix fibrillin. REFERENCES 1. Ignotz RA, Massague J. Transforming growth factor-b stimulates expression of fibronectin and collagen and their incorporation into extracellular matrix. J Biol Chem 1986;261:4337–45. 2. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS, Wakefield LM, et al. Transforming growth factor type beta: rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Natl Acad Sci U S A 1986;83: 4167–71. 3. Vaughan MB, Howard EW, Tomasek JJ. Transforming growth factor-beta1 promotes the morphological and functional differentiation of the myofibroblast. Exp Cell Res 2000;257:180–9. 4. Varga J, Jimenez SA. Pathogenesis of scleroderma: cellular aspects. In: Clements PJ, Furst DE, editors. Systemic sclerosis. Philadelphia: Williams & Wilkins; 1996. p. 123–152. 5. Igarashi A, Nashiro K, Kikuchi K, Sato S, Ihn H, Grotendorst GR, et al. Significant correlation between connective tissue growth factor gene expression and skin sclerosis in tissue sections from patients with systemic sclerosis. J Invest Dermatol 1995;105:280–4. 6. Sfikakis PP, McCune BK, Tsokos M, Aroni K, Vayiopoulos G, Tsokos GC. Immunohistological demonstration of transforming growth factor-beta isoforms in the skin of patients with systemic sclerosis. Clin Immunol Immunopathol 1993;69:199–204. 7. Gabrielli A, Di Loreto C, Taborro R, Candela M, Sambo P, Nitti C, et al. Immunohistochemical localization of intracellular and extracellular associated TGF beta in the skin of patients with systemic sclerosis (scleroderma) and primary Raynaud’s phenomenon. Clin Immunol Immunopathol 1993;68:340–9. 8. Falanga V, Gerhardt CO, Dasch JR, Takehara K, Ksander GA. Skin distribution and differential expression of transforming growth factor beta 1 and beta 2. J Dermatol Sci 1992;3:131–6. 9. Gruschwitz M, Muller PU, Sepp N, Hofer E, Fontana A, Wick G. Transcription and expression of transforming growth factor type beta in the skin of progressive systemic sclerosis: a mediator of fibrosis? J Invest Dermatol 1990;94:197–203. 10. Ludwicka A, Ohba T, Trojanowska M, Yamakage A, Strange C, Smith EA, et al. Elevated levels of platelet derived growth factor and transforming growth factor-beta 1 in bronchoalveolar lavage fluid from patients with scleroderma. J Rheumatol 1995;22: 1876–83. 11. Moreland LW, Goldsmith KT, Russell WJ, Young KR Jr, Garver RI Jr. Transforming growth factor beta within fibrotic scleroderma lungs. Am J Med 1992;93:628–36. 12. Falanga V, Tiegs SL, Alstadt SP, Roberts AB, Sporn MB. Transforming growth factor-beta: selective increase in glycosaminoglycan synthesis by cultures of fibroblasts from patients with progressive systemic sclerosis. J Invest Dermatol 1987;89:100–4. 13. Kikuchi K, Kadono T, Ihn H, Sato S, Igarashi A, Nakagawa H, et al. Growth regulation in scleroderma fibroblasts: increased response to transforming growth factor-beta 1. J Invest Dermatol 1995;105:128–32. 14. Kawakami T, Ihn H, Xu W, Smith E, LeRoy C, Trojanowska M. Increased expression of TGF-beta receptors by scleroderma fibroblasts: evidence for contribution of autocrine TGF-beta signaling to scleroderma phenotype. J Invest Dermatol 1998;110:47–51. 15. Susol E, Rands AL, Herrick A, McHugh N, Barrett JH, Ollier WE, et al. Association of markers for TGFbeta3, TGFbeta2 and TIMP1 with systemic sclerosis. Rheumatology (Oxford) 2000;39:1332–6. 16. Sakai LY, Keene DR, Engvall E. Fibrillin, a new 350-kD glycoprotein, is a component of extracellular microfibrils. J Cell Biol 1986;103:2499–509. 17. Reinhardt DP, Sasaki T, Dzamba BJ, Keene DR, Chu ML, Gohring W, et al. Fibrillin-1 and fibulin-2 interact and are colocalized in some tissues. J Biol Chem 1996;271:19489–96. 18. Bona C, Saito S. Fibrillin-1 protein in tight skin mice and scleroderma. Clin Rev Allergy Immunol 2000;18:119–26. 19. Fleischmajer R, Jacobs L, Schwartz E, Sakai LY. Extracellular microfibrils are increased in localized and systemic scleroderma skin. Lab Invest 1991;64:791–8. 20. Siracusa LD, McGrath R, Ma Q, Moskow JJ, Manne J, Christner PJ, et al. A tandem duplication within the fibrillin 1 gene is associated with the mouse tight skin mutation. Genome Res 1996;6:300–13. 21. Jimenez SA, Gershwin ME. Animal models of systemic sclerosis. In: Clements PJ, Furst DE, editors. Systemic sclerosis. Philadelphia: Williams & Wilkins; 1996. p. 251–73. 22. Tan FK, Stivers DN, Foster MW, Chakraborty R, Howard RF, Milewicz DM, et al. Association of microsatellite markers near the fibrillin 1 gene on human chromosome 15q with scleroderma in a Native American population. Arthritis Rheum 1998;41:1729–37. 23. Tan FK, Wang N, Kuwana M, Chakraborty R, Bona CA, Milewicz DM, et al. Association of fibrillin 1 single-nucleotide polymorphism haplotypes with systemic sclerosis in Choctaw and Japanese populations. Arthritis Rheum 2001;44:893–901. 24. Pereira L, Lee SY, Gayraud B, Andrikopoulos K, Shapiro SD, Bunton T, et al. Pathogenetic sequence for aneurysm revealed in mice underexpressing fibrillin-1. Proc Natl Acad Sci U S A 1999;96:3819–23. 25. Kitahama S, Gibson MA, Hatzinikolas G, Hay S, Kuliwaba JL, Evdokiou A, et al. Expression of fibrillins and other microfibrilassociated proteins in human bone and osteoblast-like cells. Bone 2000;27:61–7. 26. Sarkissian M, Lafyatis R. Integrin engagement regulates proliferation and collagenase expression of rheumatoid synovial fibroblasts. J Immunol 1999;162:1772–9. 27. Hollister DW, Godfrey M, Sakai LY, Pyeritz RE. Immunohistologic abnormalities of the microfibrillar-fiber system in the Marfan syndrome. N Engl J Med 1990;323:152–9. 28. Dallas SL, Keene DR, Bruder SP, Saharinen J, Sakai LY, Mundy GR, et al. Role of the latent transforming growth factor beta binding protein 1 in fibrillin-containing microfibrils in bone cells in vitro and in vivo. J Bone Miner Res 2000;15:68–81. INDUCTION OF FIBRILLIN FIBRILLOGENESIS BY TGF␤ 29. Sarkissian M, Lafyatis R. Transforming growth factor-beta and platelet derived growth factor regulation of fibrillar fibronectin matrix formation by synovial fibroblasts. J Rheumatol 1998;25: 613–22. 30. Jelaska A, Korn JH. Role of apoptosis and transforming growth factor ␤1 in fibroblast selection and activation in systemic sclerosis. Arthritis Rheum 2000;43:2230–9. 31. Reinhardt DP, Gambee JE, Ono RN, Bachinger HP, Sakai LY. Initial steps in assembly of microfibrils: formation of disulfidecross-linked multimers containing fibrillin-1. J Biol Chem 2000; 275:2205–10. 32. Birk DE, Silver FH, Trelstad RL. Matrix assembly. In: Hay ED, editor. Cell biology of extracellular matrix. New York: Plenum; 1991. p. 221–54. 33. Ezura Y, Chakravarti S, Oldberg A, Chervoneva I, Birk DE. Differential expression of lumican and fibromodulin regulate collagen fibrillogenesis in developing mouse tendons. J Cell Biol 2000;151:779–88. 34. Neame PJ, Kay CJ, McQuillan DJ, Beales MP, Hassell JR. Independent modulation of collagen fibrillogenesis by decorin and lumican. Cell Mol Life Sci 2000;57:859–63. 35. Schalkwijk J, Zweers MC, Steijlen PM, Dean WB, Taylor G, van Vlijmen IM, et al. A recessive form of the Ehlers-Danlos syndrome caused by tenascin-X deficiency. N Engl J Med 2001;345:1167–75. 36. Wu C, Keivens VM, O’Toole TE, McDonald JA, Ginsberg MH. Integrin activation and cytoskeletal interaction are essential for the assembly of a fibronectin matrix. Cell 1995;83:715–24. 37. Zhong C, Chrzanowska-Wodnicka M, Brown J, Shaub A, Belkin AM, Burridge K. Rho-mediated contractility exposes a cryptic site in fibronectin and induces fibronectin matrix assembly. J Cell Biol 1998;141:539–51. 38. Tiedemann K, Batge B, Muller PK, Reinhardt DP. Interactions of 3009 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. fibrillin-1 with heparin/heparan sulfate, implications for microfibrillar assembly. J Biol Chem 2001;276:36035–42. Sakamoto H, Broekelmann T, Cheresh DA, Ramirez F, Rosenbloom J, Mecham RP. Cell-type specific recognition of RGD- and non-RGD-containing cell binding domains in fibrillin-1. J Biol Chem 1996;271:4916–22. Wallis DD, Tan FK, Arnett FC, Milewicz DM. Fibrillin-1 instability in fibroblast cultures from systemic sclerosis patients. Arthritis Rheum 2000;43 Suppl 9:S91. Gayraud B, Keene DR, Sakai LY, Ramirez F. New insights into the assembly of extracellular microfibrils from the analysis of the fibrillin 1 mutation in the tight skin mouse. J Cell Biol 2000;150: 667–80. Arora PD, McCulloch CA. The deletion of transforming growth factor-beta-induced myofibroblasts depends on growth conditions and actin organization. Am J Pathol 1999;155:2087–99. Kielty CM, Shuttleworth CA. Synthesis and assembly of fibrillin by fibroblasts and smooth muscle cells. J Cell Sci 1993;106:167–73. Iwamoto H, Sakai H, Tada S, Nakamuta M, Nawata H. Induction of apoptosis in rat hepatic stellate cells by disruption of integrinmediated cell adhesion. J Lab Clin Med 1999;134:83–9. Mooney A, Jackson K, Bacon R, Streuli C, Edwards G, Bassuk J, et al. Type IV collagen and laminin regulate glomerular mesangial cell susceptibility to apoptosis via beta(1) integrin-mediated survival signals. Am J Pathol 1999;155:599–606. McGill G, Shimamura A, Bates RC, Savage RE, Fisher DE. Loss of matrix adhesion triggers rapid transformation-selective apoptosis in fibroblasts. J Cell Biol 1997;138:901–11. Pereira L, Andrikopoulos K, Tian J, Lee SY, Keene DR, Ono R, et al. Targetting of the gene encoding fibrillin-1 recapitulates the vascular aspect of Marfan syndrome. Nat Genet 1997;17:218–22. Ramirez F, Gayraud B, Pereira L. Marfan syndrome: new clues to genotype-phenotype correlations. Ann Med 1999;31:202–7.