Induction of the Myofibroblastic Phenotype in Human Gingival

Home , Aponeurosis

Patricio C. Smith1, Mnica Induction of the Cceres2, Jorge Martinez2 1Faculty of Odontology and 2Institute of Nutrition and Food Technology (INTA), University of myoﬁbroblastic phenotype Chile, Santiago, Chile. in human gingival ﬁbroblasts by transforming growth factor-b1: role of RhoA-ROCK and c-Jun N-terminal kinase signaling pathways

Smith PC, Caćeres M, Martinez J. Induction of the myofibroblastic phenotype in human gingival fibroblasts by transforming growth factor-b1: role of RhoA-ROCK and c-Jun N-terminal kinase signaling pathways. J Periodont Res 2006; 41: 418–425. Blackwell Munksgaard 2006

Background and Objectives: Myofibroblastic differentiation is an important event in gingival wound healing and chronic inflammation. Transforming frowth factor- b 1 (TGF-b1) is a potent growth factor that has been implicated in this process. Gingival myofibroblasts have an increased ability to remodel the extracellular matrix and this feature has been associated with changes in the distribution of F-actin and the expression of the myofibroblast marker a-smooth muscle actin. In the present study we have analyzed the role of TGF-b1 and the signaling routes activated by this factor in the cytoskeletal changes that characterize the myofibroblastic differentiation process in human gingival fibroblasts. Materials and methods: The signalling pathways involved in myofibroblastic differentiation were studied in primary cultures of human gingival fibroblasts using several signal transduction inhibitors. RhoA activation was analyzed through a pull-down assay. Distribution of focal adhesions and actin cytoskeleton was assessed by means of immunofluorescence and western blot. A cell adhesion assay was performed in TGF-b1-stimulated cells. Smooth muscle actin expression was studied through western blot and imunofluorescence. c-Jun N-terminal kinase phosphorylation was assessed through western blot. Patricio C. Smith, Faculty of Odontology, Results: Our observations show that TGF-b1 activated the GTPase RhoA, a key University of Chile, Olivos 943, Casilla 1903, Santiago, Chile regulator of the actin cytoskeleton. As a consequence of this event, this growth Tel: 56 2 978 1433 factor stimulated the generation of actin stress fibers and the reinforcement of Fax: 56 2 231 4030 vinculin-enriched focal adhesions. These responses were blocked after inhibiting e-mail: [email protected], ROCK, the main target of RhoA activation. TGF-b1 also stimulated the adhesion Keywords: transforming growth factor-b1; gingival fibroblast; RhoA; c-Jun N-terminal kinase of fibroblasts over fibronectin, an extracellular matrix molecule involved in myofibroblastic differentiation. Finally, induction of the myofibroblast marker Accepted for publication January 5, 2006 TGF-b1 and gingival myofibroblasts. 419 a-smooth muscle actin by TGF-b1 was abolished by the c-Jun N-terminal protein kinase inhibitor SP600125, suggesting a role for this signaling pathway during the induction of this phenotype. Conclusions: We propose that TGF-b1 may promote the differentiation of myofibroblasts through the stimulation of cell spreading and adhesion, the reinforcement of focal adhesions, the maturation of the actin cytoskeleton, and the induction of a-smooth muscle actin. Activity of RhoA-ROCK and c-Jun N-terminal protein kinase signaling pathways are probably involved in these cellular events.

Gingival fibroblasts play an active role RhoA, a small GTP-binding protein anti-p-JNK was purchased from Up- in tissue remodeling events during that regulates the dynamics of the actin state Biotechnology (Lake Placid, NY, wound healing and inflammation. cytoskeleton through the activation of USA) and mouse monoclonal anti- Despite their somewhat xuniform the kinase ROCK (12), may be modu- JNK-2 was obtained from Santa Cruz appearance, fibroblasts exhibit a high lated by TGF-b1 in epithelial cells Biotechnology (Santa Cruz, CA, degree of heterogeneity in tissues sub- (13,14). In spite of the fact that TGF-b1 USA). Human plasma fibronectin was jected to remodeling (1). A subpopula- is highly expressed in gingival tissue (5), obtained from Calbiochem (San Die- tion of cells termed myofibroblasts the role of this growth factor on the go, CA, USA). express high levels of a-smooth muscle activation of the above-mentioned pathways has not been studied in gin- actin (a-sma) and are responsible for Cell cultures cell-mediated matrix contraction dur- gival fibroblasts. ing wound healing and chronic inflam- In the present study we show that Primary cultures of human gingival mation (2,3). In vivo, wound TGF-b1 may promote several cellular fibroblasts were established by the myofibroblasts are thought to arise changes associated with the myofib- explant method (15). Tissue explants from quiescent a-sma negative fibro- roblastic phenotype including the for- were obtained from the retromolar blasts (4). Transforming growth factor- mation of actin stress fibers and focal tissue of one female patient undergoing b1 (TGF-b1) is a multifunctional adhesions, responses that were regula- extraction of a third molar at a private growth factor expressed at high levels in ted by the activity of the RhoA-ROCK dental practice in Santiago, Chile. The gingival tissues subjected to inflamma- signaling cascade. Moreover, induction tissue sample was harvested with the tion and wound healing (5,6) and has of the myofibroblast marker a-sma by informed consent of the patient and been implicated in the induction of the TGF-b1 was associated with the acti- the Ethical Commitee of the Faculty of myofibroblastic phenotype (7). vation of the JNK signaling pathway. Dentistry of the University of Chile Although the role of TGF-b1 on myo- We propose that the activation of these approved the protocol for tissue fibroblast differentiation and a-sma signaling routes by TGF-b1 represents obtention. No previous history of expression has been studied previously significant molecular events during inflammation of the retromolar tissue (7,8), the cellular mechanisms activated gingival wound healing and chronic was reported. No relevant pre-existing by this growth factor during this event inflammation. medical or drug histories were cited are still not completely understood. A during the last 6 months. Cells were cultured in Dulbecco’s modified striking feature of myofibroblasts is the Materials and methods development of a strong actin network Eagle’s medium (DMEM) (Gibco BRL, Grand Island, NY, USA) con- reinforced by a-sma and the presence of Antibodies and reagents enhanced cell adhesion mechanisms (3). taining 10% fetal bovine serum (Gibco Since TGF-b1 is able to regulate the Monoclonal antibody against RhoA BRL), 100 lg/ml penicillin (Sigma), actin cytoskeleton in several cell types was obtained from Cytosqueleton 100 lg/ml streptomycin (Sigma), and (9,10), this growth factor has been (Denver, CO, USA). Anti-vinculin, 50 lg/ml gentamycin at 37Cina5% proposed as a modulator of the changes anti-a-sma and b-actin were supplied CO2 atmosphere. All experiments were observed during the differentiation of by Sigma (St Louis, MO, USA). performed using cells between the myofibroblasts (7). Recombinant human TGF-b1 was fourth and tenth passages. TGF-b1 may modify cell behavior purchased from US Biological (Swampscott, MA, USA). The Rho- through activation of several signal Affinity precipitation of GTP-bound kinase inhibitor Y27632 and PD98059 transduction pathways including the RhoA Smads and the mitogen-activated pro- were obtained from Calbiochem (San tein kinases, extracellular signal-regu- Diego, CA, USA). SP600125 was pur- Dr Keith Burridge (University of lated kinase (ERK) and c-Jun N- chased from Biolmol (Plymouth North Carolina, USA) kindly provided terminal protein kinase (JNK) (11). Meeting, PA, USA). Rabbit polyclonal a construct that encodes the Rho 420 Smith et al. binding domain of Rhotekin (RBD) as levels were detected as loading con- Cell spreading and adhesion assays a glutathione S-transferase fusion trols. Fibroblasts were treated with protein (GST-RBD). Cell cultures were Gingival fibroblasts were stimulated, 10 ng/ml TGF-b1 in DMEM without stimulated with 10 ng/ml TGF-b1 (US or not, with 10 ng/ml TGF-b1 for serum at various time-points and lysed Biological, Swampscott, MA, USA) 12 h. Cells detached from the substra- in PBS containing 150 mM NaCl, 1% for 7 min. Lysophosphatidic acid tum by a brief exposure to trypsin/ Triton X-100 pH 7.4 and 10 lg/ml (Sigma) was used as a positive control EDTA were plated onto fibronectin- leupeptin, 1 mM PMSF, 2 mM N-eth- 7 of RhoA activation (16). Cells (10 ) coated plates (10 lg/ml) for 10 min. ylmaleimide, 1 mM orthovanadate, and were washed once with ice-cold phos- Cell spreading was stopped by pouring 1mM NaF. Proteins were resolved by phate-buffered saline (PBS) and lysed off the medium, washing adhered cells 10% sodium dodecyl sulfate-poly- with 50 mM Tris–HCl, pH 7.6, 0.5 mM once with PBS, fixing them with acrylamide gel electrophoresis and

MgCl2, 500 mM NaCl, 1% Triton X- methanol for 2 min, and then incuba- transferred to polyvinylidene difluoride 100, 0.1% sodium dodecyl sulfate, ting with 0.2% crystal violet for 5 min transfer membranes (PerkinElmer Life 0.5% deoxycholate, 10 lg/ml each of (17). After several washes, adherent Sciences, Boston, MA, USA). Mem- aprotinin and leupeptin and 1 mM cells were photographed using a Sony branes were exposed to primary anti- phenylmethylsulfonyl fluoride DSC camera through a Zeiss Axiovert bodies, secondary antibodies coupled (PMSF). Lysates were clarified by 25 inverted microscope. To quantify to horseradish peroxidase, and finally centrifugation at 19,000 g for 10 min, cell adhesion, crystal violet-stained developed with an enhanced chemilu- and the supernatant was incubated on cells were dissolved with 0.1% sodium minescence kit (Amersham Corp., a roller for 1 h with 30 lg GST-RBD citrate in 50% methanol. The absorb- Arlington Heights, FL, USA). [the GST fusion protein containing the ance at 570 nm was read on an ELX RhoA-binding domain (amino acids 7– 800 Microplate reader as previously TGF-b1-stimulated a-sma expres- 89) of Rhotekin] bound to glutathione- described (18). sion — Serum-starved cells were sti- agarose beads (Sigma). Samples were mulated with 10 ng/ml TGF-b1 for washed three times with lysis buffer 3 days in the presence or absence of Preparation of Triton-soluble and and bound proteins were eluted by matogen-activated protein kinase Triton-insoluble fractions boiling in sample buffer and then were kinase (MEK) (PD98059), JNK immunoblotted with RhoA monoclo- Cells in 60-mm culture dishes were lysed (SP600125), or ROCK (Y27632) inhi- nal antibodies. Whole cell lysates were for 15 min with a buffer containing bitors. a-sma and b-actin expression also immunoblotted for RhoA as 50 mM HEPES pH 7.4, 150 mM NaCl, levels were detected using Western loading controls. 2mM MgCl2,2mM EGTA, 1% Triton blots. Cells were lysed and processed for X-100, 10% glycerol, 2 mM PMSF, Western blot as described in the JNK 2 lg/ml pepstatin, 2 lg/ml leupeptin, activation assay. Immunofluorescence and 1 mM sodium orthovanadate. The Cells plated on coverslips were washed solubilized material (hereafter the Results once with PBS and fixed with 4% Triton-soluble fraction) was obtained paraformaldehyde for 10 min, perme- after centrifugation for 10 min at RhoA activation, actin cytoskeleton abilized with 0.25% Triton X-100 for 19,000 g The material that remained re-organization and focal adhesion 5 min and incubated with PBS con- adhered to the plastic dish after Triton formation by TGF-b1 taining 4% bovine serum albumin solubilization was scraped with a rubber (Rockland, Gilbertsville, PA, USA) for policeman in the presence of 1% NP-40, Since myofibroblasts display a well- 30 min at room temperature. Primary 0.5% sodium deoxycholate, 0.2% developed actin cytoskeleton network, antibodies diluted in PBS containing sodium dodecyl sulphate, 150 mM we investigated whether RhoA, a well- 1% bovine serum albumin were used in NaCl, 50 mM Tris HCl, pH 7.4 (RIPA). known regulator of cytoskeletal chan- a dilution of 1/100 for both anti-a-sma This suspension was boiled in sodium ges (12), is activated by TGF-b1in and anti-vinculin and were incubated dodecyl sulfate–polyacrylamide gel human gingival fibroblasts. To test this, for 30 min at room temperature. electrophoresis sample buffer to disso- semiconfluent cultures of gingival Afterwards, antigen–antibody com- ciate proteins and then centrifuged at fibroblasts were stimulated with 10 ng/ plexes were washed and incubated with 19,000 g. This supernatant correspon- ml TGF-b1 for 7 min and the activation fluorescein isothiocyanate-conjugated ded to the Triton-insoluble fraction. of RhoA was assayed as described in the anti-mouse immunoglobulin G (Rock- Materials and methods. As a positive land). F-Actin was stained with Alexa control, cells were stimulated with Western blot analysis fluor 594 Phalloidin-rhodamine 10 lM lysophosphatidic acid. As shown (Molecular Probes, Eugene, OR, USA). Assay of mitogen-activated protein in Fig. 1(A), TGF-b1 stimulation Fluorescence images were collected kinase activation - JNK activation induced a potent activation of RhoA as on a Zeiss Axioplan microscope and was determined by Western blotting demonstrated by a pull-down assay, photographed using a 63 · immersion using antibodies specific for phos- which allows the identification of the objective and an Axiocam camera. phorylated forms of JNK. Total JNK GTP-bound form of this protein. TGF-b1 and gingival myofibroblasts. 421

A insoluble fraction was increased in TGF-b1-stimulated cells, suggesting that the growth factor may stimulate the establishment of strong focal adhesions enriched in vinculin. This difference in the distribution of surface-associated vinculin was not a result of an enhanced expression of the B protein as shown by immunoblot analysis (Fig. 2A). To determine whether the TGF-b1- stimulated phenotypical changes were able to modulate the adhesive or spreading behavior of these cells, serum-starved human gingival fibroblasts were stimulated with 10 ng/ml of the growth factor for 12 h. Cells were washed twice with PBS and detached by a brief exposure to a trypsin/EDTA solution. Trypsin was inactivated by adding DMEM supplemented with 10% fetal bovine serum and cells were allowed to recover for 15 min in an Eppendorf tube for 10 min. TGF-b1- stimulated and control cells were see- Fig. 1. RhoA activation, actin cytoskeleton re-organization, and focal adhesion formation ded onto fibronectin-coated plates and by TGF-b1 in human gingival fibroblasts. (A) Cultures of gingival fibroblasts were stimu- allowed to attach for 10 min. As lated with 10 ng/ml TGF-b1 for 7 min. As a positive control, cells were stimulated with observed in Fig. 2(B), TGF-b1 10 ng/ml LPA. RhoA activation was assessed by a pull-down assay, which allows the strongly stimulated the adhesion of identification of the GTP-bound form of this protein. As a loading control, total levels of human gingival fibroblasts when com- RhoA were detected. (B) Gingival fibroblasts were stimulated with 10 ng/ml TGF-b1 for pared to vehicle-treated cells. As 12 h and processed for immunofluorescence. Y27632 was added 30 min before TGF-b1 Fig. 2(C) shows, TGF-b1-stimulated stimulation. The actin cytoskeleton was identified after staining with Alexa fluor Phalloidin- cells displayed obvious signs of cell rhodamine (red) and distribution of vinculin was detected through immunofluorescence spreading. In contrast, vehicle-stimu- (green). Results are representative of at least two independent experiments. Bar ¼ 10 lm. lated cells remained with a rounded morphology, lacking detectable cellular extensions. To analyze the morphological con- abolished the basal ROCK activity in sequences of TGF-b1-induced RhoA these cells. Signaling pathways involved in activation, serum-starved gingival fi- From these results we may conclude a-sma induction in human gingival broblasts were stimulated with 10 ng/ that TGF-b1 is able to stimulate the fibroblasts ml TGF-b1 for 12 h and processed for generation of focal adhesions and immunofluorescence. As observed in stress fiber formation in human gingi- Previous studies have demonstrated Fig. 1(B), TGF-b1-stimulated cells val fibroblasts through activation of that TGF-b1 may stimulate a-sma displayed strong actin stress fibers and the RhoA–ROCK signaling pathway. expression in human gingival fibro- high levels of focal adhesions, as blasts (19). In the present study we revealed by staining of the actin investigated the signaling pathways Modulation of gingival fibroblast cytoskeleton or immunostaining for presumptively involved in TGF-b1-sti- adhesion and spreading by TGF-b1 vinculin. When the ROCK inhibitor mulated a-sma production. For this Y27632 was added before TGF-b1 To confirm the stimulatory effect of purpose, serum-starved gingival fibro- stimulation, the cytoskeletal changes TGF-b1 on the assembly of vinculin- blasts were exposed to the MEK1/2 promoted by the factor were abro- enriched focal adhesions suggested in inhibitor PD98059 (40 lM), the JNK gated, suggesting that RhoA and its Fig. 1(B), serum-starved cells were sti- inhibitor SP600125 (10 lM), or the kinase ROCK are implicated in these mulated with 10 ng/ml TGF-b1 for ROCK inhibitor Y27632 (10 lM) for cellular responses. Y27632-treated 12 h. Cells were lysed and processed as 30 min before the stimulation with cells had almost no focal adhesions described in the Materials and methods 10 ng/ml TGF-b1 for 3 days. This sti- and a poorly developed actin network, section. As shown in Fig. 2(A), the mulation time was derived from pre- suggesting that this inhibitor also presence of vinculin in the Triton- vious studies that have demonstrated 422 Smith et al.

A the induction of the myofibroblastic phenotype (7). In the present study, we have analyzed the cellular mechanisms induced by TGF-b1 during differentiation of gingival myofibroblasts. A striking feature of the myofibroblastic phenotype is the formation of increased cell adhesions onto insoluble B substrates present in the extracellular matrix and the generation of strong stress fibers along the axis of the cells (21). A morphological continuity called fibronexus, a specialized form of focal adhesion enriched with vinculin, has been described between actin stress fibers that contain a-sma and the extracellular matrix, which contains C fibronectin (21,22). The Rho family of GTPases has been involved in the regulation of cytoskeletal organization and therefore in cell morphology (12). Rho family members such as Cdc42, Rac, and RhoA are part of the Ras superfamily of proteins that cycle between an active, GTP-bound state and an inactive, GDP-bound state (12). Fig. 2. Modulation of gingival fibroblast adhesion and spreading by TGF-b1. (A) Human Rac and Cdc42 control membrane gingival fibroblasts seeded over fibronectin were stimulated with 10 ng/ml TGF-b1 for 12 h ruffling, lamellipodia formation, and and lysed as described in the Material and methods. Presence of vinculin in the Triton- filopodia extensions (23–25). Consid- soluble and insoluble fractions was detected through Western blot. b-actin levels were used as ering that RhoA, and its main target a loading control. The Triton-soluble/insoluble ratio was determined after the digital analysis Rho kinase (ROCK) have been shown of vinculin expression normalized against b-actin expression levels. (B) Control and TGF-b1- to be essential for the formation of stimulated cells were seeded over a fibronectin matrix for 10 min. Cell adhesion was quan- focal complexes and actin bundling tified after staining cells with crystal violet. Cells were lysed and absorbance was read at (26–28), we investigated whether the 570 nm (C) Under the same experimental conditions as described in B, cell spreading was signaling pathway initiated by RhoA evaluated through direct microscopy of crystal violet-stained cells. Bar ¼ 10 lm. activation was controlled by TGF-b1 in gingival fibroblasts. The present that a-sma is induced after 72 h of fibroblasts. As shown in Fig. 3(C), study shows that stimulation of human TGF-b1 treatment in gingival fibro- TGF-b1 activated JNK with a peak of gingival fibroblasts by TGF-b1 acti- blasts (19). As revealed by Western phosphorylation detected between 15 vates RhoA and stimulates the forma- blotting of the cell lysates, TGF-b1 and 30 min of stimulation. tion of strong actin stress fibers and the stimulated a-sma expression in these distribution of vinculin into focal cells and only the JNK inhibitor was Discussion adhesion contacts, changes that were able to attenuate this response sensitive to the ROCK inhibitor Y- (Fig. 3A). Neither the MEK1/2 During gingival wound healing and 27632. Moreover, Y-27632 also abol- (PD98059) nor the ROCK (Y-27632) chronic inflammation, a specific sub- ished the basal assembly of focal inhibitors were able to modify the population of cells known as myo- adhesions and the organization of the induction of a-sma in these cells. This fibroblasts is differentiated (1). actin cytoskeleton, suggesting that result was confirmed by an immuno- Myofibroblasts probably originate ROCK is also involved in the main- fluorescence experiment in which we from local resting fibroblasts, which are tenance of these phenotypical features observed that the JNK inhibitor com- stimulated by soluble growth factors under non-stimulated conditions. pletely abrogated the expression of and/or molecular components of the These data strongly suggest that the a-sma in TGF-b1-stimulated cells extracellular matrix (7,19,20). Among RhoA–ROCK signaling pathway, (Fig. 3B). To test whether the JNK the growth factors expressed in the gin- activated by TGF-b1, is the main signaling route was activated by TGF- giva, TGF-b1 is abundantly produced component of the Rho GTPase family b1, we analyzed the time–course when this tissue is subjected to inflam- involved in the cytoskeletal changes phosphorylation of JNK through mation or wound healing (5,6). This observed during myofibroblastic dif- Western blotting in human gingival growth factor has also been involved in ferentiation. TGF-b1 and gingival myofibroblasts. 423

A pathways that regulate the induction of a-sma during the stimulation of human gingival fibroblasts by TGF-b1. Induction of a-sma in gingival fibroblasts depends on the interaction of the cell with the ECM and the level of tension of the actin cytoskeleton (28). Moreover, expression of a-sma depends on the adhesive state of the cell and signaling via focal adhesion kinase (32). These data suggest that B the molecular events regulating the dynamic state of the actin cytoskeleton are probably involved in the induction of a-sma. However, contradictory studies have shown that induction of a-sma in pancreatic stellate cells can be attenuated after treatment with the ROCK inhibitor Y27632 (33). On the other hand, in an epithelial–mesenchymal transition model of kidney cells, TGF-b1-stimulated a-sma induction was dependent on the activity of RhoA but was not interrupted after inhibiting ROCK (14). Our results show that TGF-b1-induced a-sma expression is not affected by the ROCK inhibitor Y27632, suggesting that this kinase might not be directly involved in this event. The variability observed in the above-presented studies may be explained by the fact that regulation of the a-sma promoter is a Fig. 3. Signaling pathways involved in a-sma induction in human gingival fibroblasts. complex process (34) and that it shows Gingival fibroblasts were stimulated with 10 ng/ml TGF-b1 for 3 days. Cell cultures were substantial tissue specificity (35). Fur- exposed to the MEK1/2 inhibitor PD98059 (40 lM), the JNK inhibitor SP600125 (10 lM)or ther studies should be performed to the ROCK inhibitor Y27632 (10 lM) for 30 min before TGF-b1 stimulation. (A) a-sma determine whether RhoA is indeed expression was detected through Western blotting of the cell lysates and b-actin was used as a involved in a-sma induction in human loading control. (B) a-sma and total actin were detected through immunofluorescence (green) gingival fibroblasts. and Alexa fluor Phalloidin-rhodamine staining (red), respectively. Bar ¼ 10 lm. (C) JNK TGF-b1 is able to activate several activation was evaluated through Western blot of TGF-b1-stimulated cells using specific signaling pathways which may lead to antibodies against the phosphorylated forms of JNK. Total JNK levels were detected as a expression or repression of genes loading control. These results are representative of three independent experiments. involved in wound healing and inflammation (36). This growth factor Since cell adhesion is involved in matrix. These results are in agreement exerts its effects through heterodimeric events such as wound healing and with previous studies, which have receptor complexes composed of type I myofibroblast differentiation, we also shown that TGF-b1 may increase the and type II serine/threonine kinase studied whether TGF-b1 was able to expression levels of several integrin receptors (11). Besides the role played modulate the adhesive phenotype of subunits, vinculin, and F-actin in by the Smads pathway (the prototypic human gingival fibroblasts. Our human peritoneal fibroblasts (9) and signaling route activated by TGF-b1), observations show that cell adhesion may reinforce the formation of vincu- activation of the mitogen-activated and spreading over fibronectin are lin-containing fibronexus adhesion protein kinase signaling pathways strongly promoted after stimulation complexes in palmar aponeurosis ERK, p38, and JNK by TGF-b1 may with this growth factor. This effect may myofibroblasts (10). also mediate several responses to this be explained by the re-distribution of It is known that a-sma is one of the growth factor in distinct cell types (11). vinculin towards the focal adhesions, best markers of the myofibroblastic In human lung fibroblasts, TGF-b1- which allows the generation of strong phenotype (29–31). We have analyzed induced a-sma expression is sensitive adhesive contacts with the extracellular the intracellular signal transduction to the JNK inhibitor CEP-1347, but is 424 Smith et al. not affected after blocking the ERK of Chile) for critical reading of this 14. Masszi A, Di Ciano C, Sirokmańy G et al. pathway (37). Similarly, our results manuscript. Central role of Rho in TGF-b1-induced show that inhibition of JNK by a-smooth muscle actin expression during epithelial-mesenchymal transformation. SP600125, but not the inhibition of References Am J Physiol Renal Physiol 2003; 284: ERK, is able to attenuate the induction F911–F924. of a-sma in human gingival fibroblasts 1. Ha¨kkinen L, Westermarck J, Ka¨ha¨ri V-M, 15. Larjava H, Heino J, Ka¨ha¨ri V-M, Krusius after TGF-b1 treatment. These find- Larjava H. Human granulation-tissue T, Vuorio E. Characterization of one ings suggest that a-sma expression by fibroblasts show enhanced proteoglycan phenotype of human periodontal granu- gene expression and altered response to TGF-b1 may be regulated by the same lation-tissue fibroblasts. J Dent Res 1989; TGF-b1. J Dent Res 1996; 75: 1767–1778. signaling pathway in cells derived from 68: 20–25. 2. Arora PD, McCulloch CAG. Dependence 16. Arthur WT, Burridge K. RhoA inactiva- distinct tissue origins such as lung or of collagen remodeling on a-smooth tion by p190RhoGAP regulates cell gingiva. muscle actin expression by fibroblasts. spreading and migration by promoting It is interesting to note that the JNK J Cell Physiol 1994; 159: 161–175. membrane protrusion and polarity. Mol signaling pathway has been involved in 3. Tomasek JJ, Gabbiani G, Hinz B, Cha- Biol Cell 2001; 12: 2711–2720. the production of fibronectin after ponnier C, Brown RA. Myofibroblasts 17. Wakatsuki T, Wysolmerski RB, Elson El. and mechano-regulation of connective TGF-b1 stimulation (38). Expression Mechanism of cell spreading: role of tissue remodelling. Nat Rev Mol Cell Biol myosin II. J Cell Sci 2003; 116: 1617–1625. of a spliced form of fibronectin (ED-A 2002; 3: 349–363. 18. Liu L, Schwarts BR, Lin N, Winn RK, fibronectin) is an early and crucial 4. Scmitt-Graff A, Desmouliere A, Gabbiani Harlan JM. Requirement for RhoA acti- event detected during myofibroblastic G. Differentiation repertoire of fibroblas- vation in leukocyte de-adhesion. J Immu- differentiation that involves a-sma tic phenotypic features: an example of nol 2002; 169: 2332–2336. expression (20). It remains to be fibroblastic cell plasticity. Virch Arch 19. Arora PD, Narani N, McCulloch CAG. 1994; 425: 3–24. determined whether activation of JNK The compliance of collagen gels regulates 5. Steinsvoll S, Alstensen TS, Schenck K. by TGF-b1 may modulate a-sma transforming growth factor-b induction of Extensive expression of TGF-b1in smooth muscle actin in fibroblasts. Am J induction through expression of ED-A chronically-inflamed periodontal tissue. Pathol 1999; 154: 871–882. fibronectin or by other unknown J Clin Periodontol 1999; 26: 366–373. 20. Serini G, Bochaton-Piallat M-L, Ropraz P molecular mechanisms. 6. Kuru L, Griffiths GS, Petri A, Olsen I. et al. The fibronectin domain ED-A is Finally, we may conclude that TGF- Changes in transforming growth factor- crucial for myofibroblastic phenotype b1 promotes the induction of the beta1 in gingival crevicular fluid following induction by transforming growth factor- periodontal surgery. J Clin Periodontol myofibroblastic phenotype through the b1. J Cell Biol 1998; 142: 873–881. 2004; 31: 527–533. activation of the RhoA–ROCK and 21. Singer II, Kawka DW, Kazazis DM, 7. Desmouliere A, Geiniz A, Gabbiani F, Clark RAF. The in vivo codistribution of JNK signaling pathways in human Gabbiani G. Transforming growth factor- fibronectin and actin fibers in granulation gingival fibroblasts. These phenotypi- b1 induces a-smooth muscle actin expres- tissue: immunofluorescence and electron cal changes include the ability of cells sion in granulation tissue myofibroblasts microscopic studies of the fibronexus at to adhere and spread over fibronectin, and in quiescent and growing cultured the myofibroblast surface. J Cell Biol the reinforcement of focal adhesion fibroblasts. J Cell Biol 1993; 122: 103–111. 1984; 98: 2091–2106. 8. Heino J, Ignotz RA, Hemler ME, Crouse complexes, the organization of actin 22. Dugina V, Fontao L, Chaponnier C, C, Massague J. Regulation of cell adhe- Vasiliev J, Gabbiani G. Focal adhesion stress fibers, and the further induction sion receptors by transforming growth features during myofibroblastic differen- of the myofibroblast marker a-sma. factor-b: concomitant regulation of inte- tiation are controlled by intracellular and These cellular and molecular events grins that share a common b1 subunit. extracellular factors. J Cell Sci 2001; 114: may represent significant steps J Biol Chem 1989; 264: 380–388. 3285–3296. involved in the acquisition of an ECM- 9. Rout UK, Saed GM, Diamond MP. 23. Ridley AJ, Paterson HF, Johnston CL, Transforming growth factor-b1 modulates remodeling phenotype during physio- Diekmann D, Hall A. The small GTP- expression of adhesion and cytoskeletal logical and pathological tissue binding protein rac regulates growth fac- proteins in human peritoneal fibroblasts. tor-induced membrane ruffling. Cell 1992; remodeling in gingival fibroblasts. Fertil Steril 2002; 78: 154–161. 70: 401–410. 10. Vaughan MB, Howard EW, Tomasek JJ. 24. Kozma R, Sarner S, Ahmed S, Lim L. Transforming growth factor-b1 promotes Acknowledgements Rho family GTPases and neuronal growth the morphological and functional differ- cone remodelling. Relationship between This work received financial support entiation of the myofibroblast. Exp Cell increased complexity induced by Res 2000; 257: 180–189. from the Departamento de Investiga- Cdc42Hs, Rac1, and acetylcholine and 11. Massague J. How cells read TGF-beta cioń y Desarrollo, Universidad de collapse induced by RhoA and lysophos- signals. Nat Rev Mol Cell Biol 2000; 1: phatidic acid. Mol Cell Biol 1997; 17: Chile, grant number 02/10–2 (to J.M.), 169–178. 1201–1211. FONDECYT, grant number(1040734) 12. Mackay DJG, Hall A. Rho GTPases. 25. Nobes CD, Hall A. Rho, rac, and cdc42 (to J.M.). We would like to thank Dr J Biol Chem 1998; 273: 20685–20688. GTPases regulate the assembly of multi- Keith Burridge (University of North 13. Bhowmick NA, Ghiassi M, Bakin A et al. molecular focal complexes associated with Caroline) for providing us with the Transforming growth factor-beta1 medi- actin stress fibers, lamellipodia, and filo- ates epithelial to mesenchymal transdif- construct that encodes the Rho podia. Cell 1995; 81: 53–62. ferentiation through a RhoA-dependent 26. Hotchin NA, Hall A. The assembly of binding domain of Rhotekin and Dr mechanism. Mol Biol Cell 2001; 12: 27–36. integrin adhesion complexes requires both Mariana Cifuentes (INTA, University TGF-b1 and gingival myofibroblasts. 425

extracellular matrix and intracellular rho/ cells: expression of cytoskeletal proteins as cells. Arterioscler Thromb Vasc Biol 1999; rac GTPases. J Cell Biol 1995; 131: 1857– markers of phenotypic modulations. Lab 19: 2049–2058. 1865. Invest 1994; 63: 144–161. 36. Prime SS, Pring M, Davies M, Paterson 27. Ridley AJ, Hall A. The small GTP-bind- 32. Tannickal VJ, Lee DY, White ES et al. IC. TGF-beta signal transduction in oro- ing protein rho regulates the assembly of Myofibroblast differentiation by trans- facial health and non-malignant disease focal adhesions and actin stress fibers in forming growth factor-b1 is dependent on (Part I). Crit Rev Oral Biol Med 2004; 15: response to growth factors. Cell 1992; 70: cell adhesion and integrin signaling via 324–336. 389–399. focal adhesion kinase. J Biol Chem 2003; 37. Hashimoto S, Yasuhiro G, Takeshita I, 28. Riento K, Ridley AJ. Rocks: multifunc- 278: 12384–12389. Matsumoto K, Maruoka S, Horie T. tional kinases in cell behaviour. Nat Rev 33. Masamune A, Kikuta K, Satoh M, Satoh Transforming growth factor-b1 induces Mol Cell Biol 2003; 4: 446–456. K, Shimosegawa T. Rho kinase inhibitors phenotypic modulation of human lung 29. Skalli O, Schu¨rch W, Seemayer T et al. block activation of pancreatic stellate fibroblasts to myofibroblast through a c-

Myofibroblasts from diverse pathologic cells. Br J Pharmacol 2003; 140: 1292– Jun NH2-terminal kinase-dependent settings are heterogeneous in their content 1302. pathway. Am J Respir Crit Care Med of actin isoforms and intermediate fil- 34. Mack CP, Thompson MM, Lawrenz- 2001; 163: 152–157. ament proteins. Lab Invest 1989; 60: 275– Smith S, Owens GK. Smooth muscle actin 38. Hocevar BA, Brown TI, Howe PH. TGF-b 285. differentiation marker gene expression is induces fibronectin synthesis through a c- 30. Darby I, Skalli O, Gabbiani G. a-smooth regulated by RhoA-mediated actin poly- Jun N-terminal kinase-dependent, Smad4- muscle actin is transiently expressed by merization. J Biol Chem 2001; 276: 341– independent pathway. EMBO J 1999; 18: myofibroblasts during experimental 347. 1345–1356. wound healing. Lab Invest 1990; 63: 21– 35. Hautmann MB, Adam PJ, Owens GK. 29. Similarities and differences in smooth 31. Sappino AP, Schurch W, Gabbiani G. muscle alpha-actin induction by TGF-b in Differentiation repertoire of fibroblastic smooth muscle versus non-smooth muscle