Advertisement

Caveolin-1 Controls Hyperresponsiveness to Mechanical Stimuli and Fibrogenesis-Associated RUNX2 Activation in Keloid Fibroblasts

  • Author Footnotes
    11 These authors contributed equally to this work.
    Chao-Kai Hsu
    Footnotes
    11 These authors contributed equally to this work.
    Affiliations
    Institute of Clinical Medicine, College of Medicine, National Cheng Kung University, Tainan, Taiwan

    Department of Dermatology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan

    International Research Center of Wound Repair and Regeneration, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Author Footnotes
    11 These authors contributed equally to this work.
    Hsi-Hui Lin
    Footnotes
    11 These authors contributed equally to this work.
    Affiliations
    Department of Physiology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Hans I Harn
    Affiliations
    International Research Center of Wound Repair and Regeneration, National Cheng Kung University, Tainan, Taiwan

    Department of Physiology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Rei Ogawa
    Affiliations
    Department of Plastic, Reconstructive and Aesthetic Surgery, Nippon Medical School, Tokyo, Japan
    Search for articles by this author
  • Yang-Kao Wang
    Affiliations
    Department of Cell Biology and Anatomy, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Yen-Ting Ho
    Affiliations
    Department of Physiology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Wan-Rung Chen
    Affiliations
    Department of Dermatology, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Yi-Chao Lee
    Affiliations
    PhD Program for Neural Regenerative Medicine, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan
    Search for articles by this author
  • Julia Yu-Yun Lee
    Affiliations
    Institute of Clinical Medicine, College of Medicine, National Cheng Kung University, Tainan, Taiwan

    International Research Center of Wound Repair and Regeneration, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Shyh-Jou Shieh
    Affiliations
    Division of Plastic and Reconstructive Surgery, Department of Surgery, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Chao-Min Cheng
    Affiliations
    Institute of Nanoengineering and Microsystems, National Tsing Hua University, Hsinchu, Taiwan
    Search for articles by this author
  • John A. McGrath
    Affiliations
    St. John’s Institute of Dermatology, King’s College London (Guy’s Campus), London, UK
    Search for articles by this author
  • Ming-Jer Tang
    Correspondence
    Correspondence: Ming-Jer Tang, Department of Physiology, College of Medicine, National Cheng-Kung University, 1 Da-Hsueh Road, Tainan 70101, Taiwan.
    Affiliations
    International Research Center of Wound Repair and Regeneration, National Cheng Kung University, Tainan, Taiwan

    Department of Physiology, College of Medicine, National Cheng Kung University, Tainan, Taiwan
    Search for articles by this author
  • Author Footnotes
    11 These authors contributed equally to this work.
Open ArchivePublished:September 09, 2017DOI:https://doi.org/10.1016/j.jid.2017.05.041
      Keloids are pathological scars characterized by excessive extracellular matrix production that are prone to form in body sites with increased skin tension. CAV1, the principal coat protein of caveolae, has been associated with the regulation of cell mechanics, including cell softening and loss of stiffness sensing ability in NIH3T3 fibroblasts. Although CAV1 is present in low amounts in keloid fibroblasts (KFs), the causal association between CAV1 down-regulation and its aberrant responses to mechanical stimuli remain unclear. In this study, atomic force microscopy showed that KFs were softer than normal fibroblasts with a loss of stiffness sensing. The decrease of CAV1 contributed to the hyperactivation of fibrogenesis-associated RUNX2, a transcription factor germane to osteogenesis/chondrogenesis, and increased migratory ability in KFs. Treatment of KFs with trichostatin A, which increased the acetylation level of histone H3, increased CAV1 and decreased RUNX2 and fibronectin. Trichostatin A treatment also resulted in cell stiffening and decreased migratory ability in KFs. Collectively, these results suggest a role for CAV1 down-regulation in linking the aberrant responsiveness to mechanical stimulation and extracellular matrix accumulation with the progression of keloids, findings that may lead to new developments in the prevention and treatment of keloid scarring.

      Abbreviations:

      AFM (atomic force microscopy), COL (collagen), ECM (extracellular matrix), FN (fibronectin), HDAC (histone deacetylase), KFs (keloid fibroblasts), NFs (normal fibroblasts), PA (polyacrylamide), siNC (nontargeting control small interfering RNA), siRNA (small interfering RNA), TCP (tissue culture plastic), TSA (trichostatin A)

      Introduction

      Keloids occur as a result of abnormal wound healing and are characterized by excessive deposition of collagen in the dermis with extension of scar tissue beyond the original borders of wounds (
      • Bran G.M.
      • Goessler U.R.
      • Hormann K.
      • Riedel F.
      • Sadick H.
      Keloids: current concepts of pathogenesis (review).
      ,
      • Niessen F.B.
      • Spauwen P.H.
      • Schalkwijk J.
      • Kon M.
      On the nature of hypertrophic scars and keloids: a review.
      ). Pathologically, keloids are characterized by thick hyalinized eosinophilic collagen fibers (keloidal collagen) and a horizontal, tongue-like advancing edge in the upper dermis (
      • Jumper N.
      • Paus R.
      • Bayat A.
      Functional histopathology of keloid disease.
      ,
      • Lee J.Y.
      • Yang C.C.
      • Chao S.C.
      • Wong T.W.
      Histopathological differential diagnosis of keloid and hypertrophic scar.
      ). The etiology and molecular mechanism of keloid formation have not been fully elucidated, and effective prevention or treatment for keloids is still lacking (
      • Tziotzios C.
      • Profyris C.
      • Sterling J.
      Cutaneous scarring: Pathophysiology, molecular mechanisms, and scar reduction therapeutics Part II. Strategies to reduce scar formation after dermatologic procedures.
      ).
      Human skin is a highly specialized mechanoresponsive organ that continually senses and adapts to various mechanical stimuli (
      • Ogawa R.
      • Hsu C.K.
      Mechanobiological dysregulation of the epidermis and dermis in skin disorders and in degeneration.
      ,
      • Wong V.W.
      • Akaishi S.
      • Longaker M.T.
      • Gurtner G.C.
      Pushing back: wound mechanotransduction in repair and regeneration.
      ). Keloids tend to form in areas of the body subjected to increased skin tension or stiffness, such as the presternum and shoulders (
      • Ogawa R.
      • Akaishi S.
      • Huang C.
      • Dohi T.
      • Aoki M.
      • Omori Y.
      • et al.
      Clinical applications of basic research that shows reducing skin tension could prevent and treat abnormal scarring: the importance of fascial/subcutaneous tensile reduction sutures and flap surgery for keloid and hypertrophic scar reconstruction.
      ). A simulated finite element analysis showed that mechanical stimulation (i.e., skin-stretching tension) strongly influences the cellular behavior that leads to keloid growth (
      • Akaishi S.
      • Akimoto M.
      • Ogawa R.
      • Hyakusoku H.
      The relationship between keloid growth pattern and stretching tension: visual analysis using the finite element method.
      ). Thus, decreasing skin tension has been suggested as potentially beneficial for treating keloids (
      • Ogawa R.
      • Akaishi S.
      • Huang C.
      • Dohi T.
      • Aoki M.
      • Omori Y.
      • et al.
      Clinical applications of basic research that shows reducing skin tension could prevent and treat abnormal scarring: the importance of fascial/subcutaneous tensile reduction sutures and flap surgery for keloid and hypertrophic scar reconstruction.
      ).
      As observed for chemical stimuli, matrix stiffness also has a large impact on the regulation of cell behavior, such as survival, proliferation, differentiation, and migration (
      • Stroka K.M.
      • Konstantopoulos K.
      Physical biology in cancer. 4. Physical cues guide tumor cell adhesion and migration.
      ,
      • Wang Y.H.
      • Chiu W.T.
      • Wang Y.K.
      • Wu C.C.
      • Chen T.L.
      • Teng C.F.
      • et al.
      Deregulation of AP-1 proteins in collagen gel-induced epithelial cell apoptosis mediated by low substratum rigidity.
      ). Moreover, several proteins, such as FAK and integrins, are found to play key roles in the mechanosensing and mechanotransduction of cells (
      • Jansen K.A.
      • Donato D.M.
      • Balcioglu H.E.
      • Schmidt T.
      • Danen E.H.
      • Koenderink G.H.
      A guide to mechanobiology: Where biology and physics meet.
      ). Recently, we found that decreased CAV1 contributes to cell softening and loss of rigidity sensing ability in Ha-RasV12-transformed fibroblasts (
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ). CAV1, a major component of caveolae, plays an important role in signal transduction, membrane traffic, and cholesterol transport (
      • Boscher C.
      • Nabi I.R.
      Caveolin-1: role in cell signaling.
      ). In addition to its primary role as a tumor suppressor, CAV1 has been linked to the regulation of focal adhesions and integrin-mediated actin remodeling, which have been widely studied for their roles in mechanotransduction (
      • Nethe M.
      • Hordijk P.L.
      A model for phospho-caveolin-1-driven turnover of focal adhesions.
      ,
      • Radel C.
      • Carlile-Klusacek M.
      • Rizzo V.
      Participation of caveolae in beta1 integrin-mediated mechanotransduction.
      ,
      • Yang B.
      • Radel C.
      • Hughes D.
      • Kelemen S.
      • Rizzo V.
      p190 RhoGTPase-activating protein links the beta1 integrin/caveolin-1 mechanosignaling complex to RhoA and actin remodeling.
      ). In various fibrotic disorders, including idiopathic pulmonary fibrosis, systemic sclerosis, and keloids, Cav1 was down-regulated (
      • Del Galdo F.
      • Sotgia F.
      • de Almeida C.J.
      • Jasmin J.F.
      • Musick M.
      • Lisanti M.P.
      • et al.
      Decreased expression of caveolin 1 in patients with systemic sclerosis: crucial role in the pathogenesis of tissue fibrosis.
      ,
      • Wang X.M.
      • Zhang Y.
      • Kim H.P.
      • Zhou Z.
      • Feghali-Bostwick C.A.
      • Liu F.
      • et al.
      Caveolin-1: a critical regulator of lung fibrosis in idiopathic pulmonary fibrosis.
      ,
      • Zhang G.Y.
      • Yu Q.
      • Cheng T.
      • Liao T.
      • Nie C.L.
      • Wang A.Y.
      • et al.
      Role of caveolin-1 in the pathogenesis of tissue fibrosis by keloid-derived fibroblasts in vitro.
      ). Collectively, we speculate that, in keloid fibroblasts (KFs), a decrease in CAV1 may result in an aberrant response to the mechanical stimuli from the local environment.
      Epigenetic mechanisms are important in regulating the wound healing processes, including keratinocyte proliferation, differentiation, and migration, together with dermal regeneration and neoangiogenesis (
      • Lewis C.J.
      • Mardaryev A.N.
      • Sharov A.A.
      • Fessing M.Y.
      • Botchkarev V.A.
      The epigenetic regulation of wound healing.
      ). The histone deacetylase (HDAC) inhibitor trichostatin A (TSA) has potent antifibrogenic effects in a mouse model of bleomycin-induced skin fibrosis (
      • Huber L.C.
      • Distler J.H.
      • Moritz F.
      • Hemmatazad H.
      • Hauser T.
      • Michel B.A.
      • et al.
      Trichostatin A prevents the accumulation of extracellular matrix in a mouse model of bleomycin-induced skin fibrosis.
      ). Treatment of fibroblasts with TSA causes abrogated transforming growth factor-β1–induced collagen gene expression (
      • Ghosh A.K.
      • Mori Y.
      • Dowling E.
      • Varga J.
      Trichostatin A blocks TGF-beta-induced collagen gene expression in skin fibroblasts: involvement of Sp1.
      ,
      • Rombouts K.
      • Niki T.
      • Greenwel P.
      • Vandermonde A.
      • Wielant A.
      • Hellemans K.
      • et al.
      Trichostatin A, a histone deacetylase inhibitor, suppresses collagen synthesis and prevents TGF-beta(1)-induced fibrogenesis in skin fibroblasts.
      ).
      • Russell S.B.
      • Russell J.D.
      • Trupin K.M.
      • Gayden A.E.
      • Opalenik S.R.
      • Nanney L.B.
      • et al.
      Epigenetically altered wound healing in keloid fibroblasts.
      found epigenetic modifications in KFs that include an altered pattern of DNA methylation and histone acetylation. Although HDAC inhibition by TSA has been shown to inhibit collagen synthesis of KFs (
      • Diao J.S.
      • Xia W.S.
      • Yi C.G.
      • Wang Y.M.
      • Li B.
      • Xia W.
      • et al.
      Trichostatin A inhibits collagen synthesis and induces apoptosis in keloid fibroblasts.
      ), the molecular mechanism underpinning this response remains unknown.
      Our study aimed to understand the mechanical properties of KFs and the mechanisms underlying the hyperresponsiveness of KFs to mechanical stimuli.

      Results

      Keloid tissues are stiffer than normal dermal tissues, but KFs are softer than normal dermal fibroblasts

      Comparison of extracellular matrix (ECM) levels in dermal tissues from keloid lesions and normal dermis showed a relative increase in fibronectin (FN) and collagen (COL) 1A1, COL3A1, and COL11A1 in the keloid tissues (Figure 1a and b). Atomic force microscopy (AFM) indentation measurements showed that keloid tissues (14,213 ± 1,047 Pa) were approximately 10-fold stiffer than normal skin tissue (2,406 ± 1,035 Pa) (P < 0.01) (Figure 1c, and see Supplementary Figure S1 online). This ECM accumulation in keloid lesions could be due to a corresponding increase in the amounts of ECM produced by the KFs (Figure 1d and e). However, at the cellular level, KFs (1,133 ± 77 Pa) were softer than normal fibroblasts (NFs) (1,539 ± 113 Pa) from healthy subjects (P < 0.01) (Figure 1f and g).
      Figure 1
      Figure 1Keloid tissues are stiffer than normal tissue, whereas keloid fibroblasts (KFs) are softer than normal fibroblasts (NFs). (a) Western blot results of skin tissue from control subjects (n = 6) and keloid patients (n = 6). The protein levels of fibronectin (FN), collagen (COL) 1A1, COL3A1, COL11A1, and GAPDH (internal control) were analyzed. (b) Quantification results of FN, COL1A1, COL3A1, and COL11A1 from a. (c) Atomic force microscopy indentation results of skin tissue dissected from control subjects (n = 6) and keloid patients (n = 6). (d) Western blot results of NFs and KFs derived from control subjects (n = 6) and keloid patients (n = 6), respectively. The protein levels of FN, COL1A1, COL3A1, COL11A1, and GAPDH were analyzed. (e) Quantification results of FN, COL1A1, COL3A1, and COL11A1 from d. (f) Stiffness distribution histograms of NFs (n = 8) and KFs (n = 8). (g) Atomic force microscopy indentation results of NFs (n = 8) and KFs (n = 8). P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

      RUNX2 is a potential key regulator for ECM overproduction in keloids

      We analyzed the microarray results from seven Japanese patients with keloids. A total of 345 up-regulated genes (>4-fold), 86 down-regulated genes (<4-fold), and 15 transcription regulators were identified in the keloid tissue compared with adjacent normal tissue (discarded skin as “dog ear” for aesthetic reasons). Using Ingenuity Pathway Analysis software (http://www.ingenuity.com), canonical pathway analysis showed that the pathway of the role of osteoblasts, osteoclasts, and chondrocytes in rheumatoid arthritis was the most important signaling pathway, highlighting 17 significantly up-regulated and one down-regulated gene from the comparative dataset (see Supplementary Figure S2 online). The upstream regulatory analysis showed three key transcription regulators in the keloid group, including RUNX2, NKX3-2, MLXIPL (see Supplementary Table S1 online). Among these, RUNX2 plays a vital role in regulating the downstream ECM proteins, including COL11A1, which is predominantly found in cartilage (Figure 2a). The expression of RUNX2 in dermal tissues was examined by immunohistochemistry on various tissue sections of normal skin, normal scar, hypertrophic scar, and keloid tissue (Figure 2b). Among these tissues, the highest percentage of fibroblasts with nuclear RUNX2 was found in keloid tissue (Figure 2c). Immunoblotting of lysates from normal and keloid tissues showed that pRUNX2 (the active form of RUNX2) and RUNX2 were up-regulated, although not statistically significantly, in keloid tissue (Figure 2d and e). However, NFs expressed similar levels of pRUNX2 and RUNX2 compared with KFs after culture on tissue culture plastic (TCP) dish (Figure 2f and g). The role of RUNX2 in the overproduction of ECMs in KFs was validated using small interfering RNA (siRNA) to silence RUNX2. Suppression of RUNX2 down-regulated the mRNA level of FN and COL11A1 in KFs (see Supplementary Figure S3 online). Altogether, these data suggest that the highly active RUNX2 in KFs could be involved in excessive ECM production in keloid tissue.
      Figure 2
      Figure 2RUNX2 is highly up-regulated in keloid tissues/fibroblasts. (a) Ingenuity pathway analysis results showing the network of RUNX2 and their close interactions with several extracellular matrix proteins and matrix metalloproteinases, which were related to wound healing and scar formation. (b) Immunohistochemical staining results for RUNX2 expression in normal skin tissue, normal scar tissue, hypertrophic scar, and keloid tissue (n = 4 in each). Scale bars = 100 μm. (c) Percentage of fibroblasts with nuclear RUNX2 in tissues from b. (d) Representative immunoblots of skin tissue from control subjects and keloid patients. The protein levels of phosphorylated RUNX2, RUNX2, and GAPDH were analyzed. (e) Quantification results of phosphorylated RUNX2 and RUNX2 from skin tissue of control subjects (n = 6) and keloid patients (n = 6). (f) Representative immunoblots of NFs and KFs. The protein levels of p-RUNX2, RUNX2, and GAPDH were analyzed. (g) Quantification results of p-RUNX2 and RUNX2 from NFs (n = 6) and KFs (n = 6). P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. KF, keloid fibroblast; NF, normal fibroblast; p-, phosphorylated.

      Hyperresponsiveness of KFs to dermal tissue-equivalent matrix stiffness causes increased expression of RUNX2 and COL11A1

      It is known that primary cells cultured on TCP dish with a physical stiffness greater than 109 Pa are spontaneously activated toward myofibroblast differentiation (
      • Achterberg V.F.
      • Buscemi L.
      • Diekmann H.
      • Smith-Clerc J.
      • Schwengler H.
      • Meister J.J.
      • et al.
      The nano-scale mechanical properties of the extracellular matrix regulate dermal fibroblast function.
      ,
      • Chen W.C.
      • Lin H.H.
      • Tang M.J.
      Regulation of proximal tubular cell differentiation and proliferation in primary culture by matrix stiffness and ECM components.
      ). We postulate that the increased RUNX2 expression in NFs is associated with the extreme change in matrix stiffness, from soft dermal tissue to stiff TCP dish, during primary culture. The effects of matrix stiffness on the expression of RUNX2 were investigated by culturing the primary dermal fibroblasts on a stiff TCP dish (>109 Pa), 2 kPa polyacrylamide (PA) gel (normal dermis-equivalent stiffness), and 20 kPa PA gel (keloid-equivalent stiffness). As shown in Figure 3a and b, the expression of nuclear RUNX2 was increased with increasing matrix stiffness in both NFs and KFs. However, KFs expressed higher nuclear RUNX2 than NFs in all conditions. The expression of FN in KFs was prominently elevated with increasing matrix stiffness (Figure 3c). Nevertheless, the expression of FN in NFs was only slightly enhanced, despite culturing on the TCP dish. We further performed explant culture of fibroblasts on type I collagen-coated PA gels of 2 and 20 kPa to diminish the stiff matrix-induced RUNX2 activation during primary culture. The mRNA level of RUNX2 and several ECM genes were evaluated on day 3. As shown in Figure 3d and 3e, NFs derived from explant culture on 2- and 20-kPa PA gels were devoid of RUNX2 and COL11A1. The mRNA levels of FN and COL3A1 of NFs were low on the 2-kPa PA gel and increased slightly on the 20-kPa PA gel. KFs derived from explants cultured on 2-kPa PA gels expressed higher RUNX2, COL11A1, FN, and COL3A1 mRNA levels than NFs, which were intensified on 20-kPa PA gels. Collectively, these data show that physiologically equivalent mechanical culture conditions suppress the expression of fibrosis-associated RUNX2 and COL11A1 of NFs compared with conventional stiff TCP. Moreover, it seems that KFs show hyperresponsiveness to dermal tissue-equivalent matrix stiffness, which causes increased expression of RUNX2 and COL11A1.
      Figure 3
      Figure 3Keloid dermal fibroblasts show hyperresponsiveness to dermal tissue-equivalent matrix stiffness, which causes increased expression of RUNX2 and COL11A1. (a) Immunofluorescence study of NFs and KFs cultured on matrices of varying stiffness for 24 hours. Cells were stained for RUNX2 (green), the nucleus (blue), and F-actin (red). Scale bars = 20 μm. (b) Percentage of cells with nuclear RUNX2 in NFs (n = 5) and KFs (n = 5) from a. (c) Western blot results of NFs (n = 4) and KFs (n = 4) cultured on matrices of varying stiffness for 24 hours. (d) Representative reverse transcription-PCR results of NFs (n = 3) and KFs (n = 3) derived from explants on matrices of varying stiffness at day 3. The mRNA expressions of RUNX2, COL11A1, fibronectin (FN), and COL3A1 were analyzed. (e) Quantification results of RUNX2, COL11A1, FN, and COL3A1 mRNA from d. GAPDH-normalized data were compared with those of NFs cultured on 2 kPa PA gel. P < 0.05, ∗∗∗P < 0.001. COL, collagen; FN, fibronectin; KF, keloid fibroblast; NF, normal fibroblast; PA, polyacrylamide.

      Decreased CAV1 is associated with cell softening and the up-regulation of fibrogenesis-associated RUNX2 and migratory ability in KFs

      Although keloids are categorized as benign fibroproliferative lesions selectively occurring in the dermis, their aggressive and recurrent behavior resembles that of malignant tumors. Our previous study showed that decreased CAV1 contributes to changes in the cell mechanics of cancer cells (
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ). The cDNA microarray data showed that CAV1 level was down-regulated in keloid tissues (–1.548) compared with adjacent normal tissue. Thus, it is important to understand whether CAV1 is also involved in the pathogenesis of keloids, particularly from a mechanobiological perspective. Immunofluorescence showed that CAV1 was expressed in the cytoplasm and cell membrane of dermal fibroblasts of normal skin, normal scar, and perilesional normal skin of keloid tissue (see Supplementary Figure S4 online). In contrast, CAV1 expression was markedly decreased in keloid lesions (see Supplementary Figure S4) and KFs (Figure 4a). The potential association of CAV1 with cell mechanics of human dermal fibroblasts was studied using siRNA to knock down CAV1 in NFs (Figure 4a and b). CAV1 knockdown caused softening of NFs (Figure 4c). When cultured on matrices of varying stiffness, non-targeting control siRNA (siNC)-treated NFs altered their stiffness, but siCAV1-treated NFs and KFs did not (Figure 4d). The results indicated the loss of stiffness-sensing ability in siCAV1-treated NFs and KFs. Furthermore, CAV1 knockdown increased the expression and nuclear translocation of RUNX2 (Figure 4a, b, e, and f) and increased FN expression (Figure 4a and b) in NFs. Knockdown of CAV1 did not change the expression pattern of nuclear RUNX2 in NFs or KFs grown on matrices of 20 kPa and 2 kPa (Figure 4g). However, siCAV1-treated NFs showed a higher percentage of cells with nuclear RUNX2 when cultured on 2-kPa PA gel, compared with siNC-treated NFs. These data suggest that a decrease in CAV1 might contribute to keloid pathogenesis via elevating fibrogenic RUNX2 expression. Immunostaining results showed that dermal fibroblasts in normal and peripheral normal skin near the keloid lesion expressed high CAV1 and extremely low RUNX2, whereas dermal fibroblasts in keloid lesions expressed low CAV1 and high RUNX2 (Figure 4h). Pearson’s correlation analysis showed that CAV1 was negatively correlated with RUNX2 in keloid lesions (Figure 4i).
      Figure 4
      Figure 4Suppression of CAV1 in NFs causes cell softening, loss of stiffness-sensing ability, and increased expression of RUNX2 and FN. (a) Western blot results of NFs transfected with nontargeting control siRNA (siNC) or siCAV1 for 48 hours and KFs. (b) Quantification results of CAV1, RUNX2, and FN from a. GAPDH-normalized data were compared with those of NFs (n = 3) transfected with siNC. CAV1 knockdown in NFs causes (c) cell softening and (d) inability to change cell stiffness on matrices of varying stiffness. (e) Confocal immunofluorescence images obtained with anti-RUNX2 (green) in NFs and KFs cultured on dishes and treated with siNC or siCAV1 for 24 hours. Scale bars = 50 μm. (f) Percentage of cells with nuclear RUNX2 in dermal fibroblasts from e. (g) Percentage of cells with nuclear RUNX2 in siNC- or siCAV1-transfected NFs and KFs grown on PA gels of 20 kPa and 2 kPa for 24 hours. (h) Confocal immunofluorescence images obtained with anti-CAV1 (red), anti-RUNX2 (green), and nucleus (blue) in skin tissues from control subjects (n = 2) and keloid patients (n = 3). Scale bar = 50 μm. (i) Pearson correlation comparing CAV1 and RUNX2 intensity in skin tissues from h. P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. FN, fibronectin; KF, keloid fibroblast; NC, nontargeting control; NF, normal fibroblast; PA, polyacrylamide; siRNA, small interfering RNA.
      One characteristic of keloids is their ability to continually migrate beyond the original border of a wound (
      • Witt E.
      • Maliri A.
      • McGrouther D.A.
      • Bayat A.
      RAC activity in keloid disease: comparative analysis of fibroblasts from margin of keloid to its surrounding normal skin.
      ). In vitro studies showed that KFs migrated faster and displayed a higher total force and force per post than NFs (
      • Harn H.I.
      • Wang Y.K.
      • Hsu C.K.
      • Ho Y.T.
      • Huang Y.W.
      • Chiu W.T.
      • et al.
      Mechanical coupling of cytoskeletal elasticity and force generation is crucial for understanding the migrating nature of keloid fibroblasts.
      ,
      • Witt E.
      • Maliri A.
      • McGrouther D.A.
      • Bayat A.
      RAC activity in keloid disease: comparative analysis of fibroblasts from margin of keloid to its surrounding normal skin.
      ). Also, Cav1-depleted NIH3T3 fibroblasts displayed increased cell migration (
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ). Here, we found that CAV1 knockdown by siRNA increased migratory ability (Figure 5a, and see Supplementary Figure S5 online) and force per post of NFs (siCav/NFs vs. siNC/NFs = 4.18 ± 0.14 nN vs. 3.53 ± 0.08 nN) (Figure 5b and c). Furthermore, both MLC2 protein and MLC2 phosphorylation, the relevant mediators of transcellular contractility (
      • Iwabu A.
      • Smith K.
      • Allen F.D.
      • Lauffenburger D.A.
      • Wells A.
      Epidermal growth factor induces fibroblast contractility and motility via a protein kinase C delta-dependent pathway.
      ), were markedly increased in KFs and in CAV1-depleted NFs (Figure 5d and e). Collectively, these data show that KFs are softer and migrate faster than NFs, which may correlate with the decrease in CAV1.
      Figure 5
      Figure 5CAV1 knockdown promotes migratory ability of NFs. NFs (n = 3) and KFs (n = 3) were transfected with siNC or siCAV1 for 24 hours. After reaching a confluent monolayer, wound healing experiments were performed. (a) Quantification results of wound scratch (0 hours) and recovery (8 hours and 24 hours) from . Wound recovery (%) = [(wound area at 0 hours – wound area at 24 hours)/wound area at 0 hours] × 100%. (b) The representative force maps of fibroblasts under indicated conditions plated on polydimethylsiloxane micropost arrays. The color scale indicates the magnitude of traction force (nN). (c) Quantification results of total force and force per post generated in cells from b. (d) Western blot results of NFs or KFs transfected with siNC and siCAV1 for 48 hours. The protein levels of CAV1, pMLC2, MLC2, and GAPDH were analyzed. (e) Quantification results of CAV1, pMLC2, and MLC2 from d. P < 0.05, ∗∗P < 0.01. KF, keloid fibroblast; NC, nontargeting control; NF, normal fibroblast; p, phosphorylated; siRNA, small interfering RNA.

      TSA, an HDAC inhibitor, inhibited histone deacetylase, increased CAV1, and decreased RUNX2 in KFs

      Considering that CAV1 down-regulation occurs only in actual keloid lesions, CAV1 might be under epigenetic regulation. HDAC2, an epigenetically regulated enzyme, is strikingly up-regulated in keloid scars (
      • Fitzgerald O’Connor E.J.
      • Badshah I.I.
      • Addae L.Y.
      • Kundasamy P.
      • Thanabalasingam S.
      • Abioye D.
      • et al.
      Histone deacetylase 2 is upregulated in normal and keloid scars.
      ). Moreover, inhibition of HDAC by TSA diminished collagen synthesis in KFs (
      • Diao J.S.
      • Xia W.S.
      • Yi C.G.
      • Wang Y.M.
      • Li B.
      • Xia W.
      • et al.
      Trichostatin A inhibits collagen synthesis and induces apoptosis in keloid fibroblasts.
      ). Here, treatment of KFs with TSA increased acetylation of lysine residue at position 9 of histone H3 (histone H3AK9) and CAV1 expression (Figure 6a–d). Consequently, TSA-treated KFs showed the phenotype toward normal, including decreased expression of RUNX2 and FN (Figure 6e and f), migratory ability (Figure 6g), and increased cell stiffness (Figure 6h). Finally, TSA-treated KFs showed a decreased percentage of cells with nuclear RUNX2 when cultured on 2-kPa PA gel (Figure 6i). Collectively, these data suggest the important role of epigenetics-modulated CAV1 in the pathogenesis of keloid scars.
      Figure 6
      Figure 6Epigenetic control of CAV1 affects cell mechanics and RUNX2 expression in KFs. (a) Confocal immunofluorescence images of KFs plated onto culture dishes and treated with various doses of TSA for 24 hours. Scale bars = 20 μm. (b) Western blot results of KFs (n = 3) under indicated conditions. The protein levels of CAV1, RUNX2, FN, histone H3 acetyl K9 (histone H3AK9), and GAPDH were analyzed. Quantification results of (c) histone H3AK9, (d) CAV1, (e) RUNX2, and (f) FN from b. GAPDH-normalized data were compared with those of KFs treated with DMSO (TSA = 0 μmol/L). Treatment of KFs with 200 nmol/L TSA (g) hindered wound recovery, (h) increased cell stiffness, and (i) decreased the percentage of cells with nuclear RUNX2 on 2 kPa. P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001. C, control; FN, fibronectin; hr, hour; KF, keloid fibroblast; M, mol/L; TSA, trichostatin A.

      Discussion

      In this study, we showed that CAV1 down-regulation plays a critical role in the pathogenesis of keloids. The finding is consistent with that of
      • Zhang G.Y.
      • Yu Q.
      • Cheng T.
      • Liao T.
      • Nie C.L.
      • Wang A.Y.
      • et al.
      Role of caveolin-1 in the pathogenesis of tissue fibrosis by keloid-derived fibroblasts in vitro.
      , who showed that CAV1, antagonizing profibrotic processes by decreasing transforming growth factor-β receptor type I and Smad2/3 phosphorylation, was markedly decreased in keloid-derived fibroblasts. Given that CAV1 levels in keloid lesions were substantially lower compared with perilesional skin of the keloid (see Supplementary Figure S4) and that HDAC inhibitor TSA increased CAV1 in KFs (Figure 6a), epigenetic modification involvement in the decrease of CAV1 is suggested. In vivo studies of mouse and human skin wounds showed that HDAC2 is markedly increased in both normal and keloid scar tissues (
      • Fitzgerald O’Connor E.J.
      • Badshah I.I.
      • Addae L.Y.
      • Kundasamy P.
      • Thanabalasingam S.
      • Abioye D.
      • et al.
      Histone deacetylase 2 is upregulated in normal and keloid scars.
      ). Whether the heightened HDAC2 is relevant to epigenetically down-regulated CAV1 in keloid lesions remains unknown. Given the normal scar tissues normally expressed CAV1 (see Supplementary Figure S4), it is possible that other epigenetic-modifying enzymes dominate CAV1 down-regulation or that other factors are required to potentiate the action of HDAC2 in keloid lesions.
      CAV1 has been implicated in pathological conditions of abnormal collagen expression in the skin. When Cav1 was knocked out, mice exhibited scleroderma-like pathological skin features, characterized by up-regulated expression of dermal collagens (
      • Castello-Cros R.
      • Whitaker-Menezes D.
      • Molchansky A.
      • Purkins G.
      • Soslowsky L.J.
      • Beason D.P.
      • et al.
      Scleroderma-like properties of skin from caveolin-1-deficient mice: implications for new treatment strategies in patients with fibrosis and systemic sclerosis.
      ), or had greater bone size and stiffness, along with an increase in alkaline phosphatase protein and expression of osterix and Runx2 (
      • Rubin J.
      • Schwartz Z.
      • Boyan B.D.
      • Fan X.
      • Case N.
      • Sen B.
      • et al.
      Caveolin-1 knockout mice have increased bone size and stiffness.
      ). In human dermal fibroblasts, CAV1 inhibition by methyl-β-cyclodextrin led to up-regulation of collagen expression (
      • Lee J.A.
      • Choi D.I.
      • Choi J.Y.
      • Kim S.O.
      • Cho K.A.
      • Lee J.B.
      • et al.
      Methyl-beta-cyclodextrin up-regulates collagen I expression in chronologically-aged skin via its anti-caveolin-1 activity.
      ). In this study, the Ingenuity Pathway Analysis data show that “the role of osteoblasts, osteoclast, and chondrocytes in rheumatoid arthritis” is the most important signaling pathway in keloids (see Supplementary Figure S2), which is consistent with results from
      • Naitoh M.
      • Kubota H.
      • Ikeda M.
      • Tanaka T.
      • Shirane H.
      • Suzuki S.
      • et al.
      Gene expression in human keloids is altered from dermal to chondrocytic and osteogenic lineage.
      , who showed that gene expression in human keloids is altered from dermal to chondrocytic and osteogenic lineage. RUNX2, the osteogenesis-specific transcription factor, is ectopically expressed in KFs in vitro and in vivo (Figure 4) (
      • Naitoh M.
      • Kubota H.
      • Ikeda M.
      • Tanaka T.
      • Shirane H.
      • Suzuki S.
      • et al.
      Gene expression in human keloids is altered from dermal to chondrocytic and osteogenic lineage.
      ). In osteoblasts, Runx2 was found to be a target of mechanical signals mediated by Ras–extracellular signal-regulated kinase 1/2–mitogen-activated protein kinase signaling (
      • Kanno T.
      • Takahashi T.
      • Tsujisawa T.
      • Ariyoshi W.
      • Nishihara T.
      Mechanical stress-mediated Runx2 activation is dependent on Ras/ERK1/2 MAPK signaling in osteoblasts.
      ,
      • Ziros P.G.
      • Gil A.P.
      • Georgakopoulos T.
      • Habeos I.
      • Kletsas D.
      • Basdra E.K.
      • et al.
      The bone-specific transcriptional regulator Cbfa1 is a target of mechanical signals in osteoblastic cells.
      ). Runx2 was also shown to be an inducer of aortic fibrosis and stiffness (
      • Raaz U.
      • Schellinger I.N.
      • Chernogubova E.
      • Warnecke C.
      • Kayama Y.
      • Penov K.
      • et al.
      Transcription factor Runx2 promotes aortic fibrosis and stiffness in type 2 diabetes mellitus.
      ). Silencing CAV1 with siRNA in NFs increased RUNX2 expression (Figure 4a), whereas enhancing CAV1 with TSA treatment in KFs suppressed RUNX2 expression (Figure 6a and b), suggesting the repressive role of CAV1 in RUNX2 expression. Cav1, the main component of caveolae, have been shown to serve as scaffolds for a variety of signaling molecules and negatively regulate signal penetrance (
      • Boscher C.
      • Nabi I.R.
      Caveolin-1: role in cell signaling.
      ). Accordingly, silencing CAV1 might relieve the signals that lead to the expression of RUNX2 in keloids. However, the expression of RUNX2 was also reported to be modulated by HDAC inhibitors in different cell types.
      • Saito T.
      • Nishida K.
      • Furumatsu T.
      • Yoshida A.
      • Ozawa M.
      • Ozaki T.
      Histone deacetylase inhibitors suppress mechanical stress-induced expression of RUNX-2 and ADAMTS-5 through the inhibition of the MAPK signaling pathway in cultured human chondrocytes.
      showed that cyclic tensile strain-induced expression of RUNX2 was suppressed by TSA in human chondrocytes. Conversely,
      • Cho H.H.
      • Park H.T.
      • Kim Y.J.
      • Bae Y.C.
      • Suh K.T.
      • Jung J.S.
      Induction of osteogenic differentiation of human mesenchymal stem cells by histone deacetylase inhibitors.
      showed that valproic acid treatment increased RUNX2 expression and osteogenic differentiation of human bone marrow stromal cells and human adipose tissue-derived stromal cells. This discrepancy may be explained by cell type differences, which initiate various signaling pathways. Although we suggest that TSA–up-regulated CAV1 plays a critical role in the suppression of RUNX2, we cannot rule out the possibility that TSA treatment suppresses RUNX2 expression, bypassing CAV1 restoration.
      Expression of COL11A1 is very low or absent in normal skin. In several types of cancer, high expression of COL11A1 occurred in the invasive carcinoma and activated stromal cells of the desmoplastic reaction, and this expression is correlated with carcinoma aggressiveness and progression and lymph node metastasis (
      • Raglow Z.
      • Thomas S.M.
      Tumor matrix protein collagen XIalpha1 in cancer.
      ,
      • Vazquez-Villa F.
      • Garcia-Ocana M.
      • Galvan J.A.
      • Garcia-Martinez J.
      • Garcia-Pravia C.
      • Menendez-Rodriguez P.
      • et al.
      COL11A1/(pro)collagen 11A1 expression is a remarkable biomarker of human invasive carcinoma-associated stromal cells and carcinoma progression.
      ). Here, the cDNA microarray data indicate that COL11A1 is the most highly up-regulated collagen gene in KFs related to NFs, which coincides with previous findings (
      • Chen W.
      • Fu X.
      • Sun X.
      • Sun T.
      • Zhao Z.
      • Sheng Z.
      Analysis of differentially expressed genes in keloids and normal skin with cDNA microarray.
      ,
      • Naitoh M.
      • Kubota H.
      • Ikeda M.
      • Tanaka T.
      • Shirane H.
      • Suzuki S.
      • et al.
      Gene expression in human keloids is altered from dermal to chondrocytic and osteogenic lineage.
      ,
      • Seifert O.
      • Bayat A.
      • Geffers R.
      • Dienus K.
      • Buer J.
      • Lofgren S.
      • et al.
      Identification of unique gene expression patterns within different lesional sites of keloids.
      ). RUNX2 knockdown by siRNA significantly decreased COL11A1 (see Supplementary Figure S3), suggesting that RUNX2 is vital for the expression of COL11A1 in KFs. Although the function of COL11A1 in keloid remains unknown, it is plausible that COL11A1 might confer keloid fibroblasts the invasive ability to spread outside the original injury site. However, more in-depth studies are needed to determine the mechanism through which COL11A1 influences KF behavior. In summary, RUNX2/COL11A1 expression might be a remarkable biomarker of KFs and an ideal target for future therapies in keloid.
      Matrices of physiologic stiffness potently inhibit normal cell proliferation (
      • Klein E.A.
      • Yin L.
      • Kothapalli D.
      • Castagnino P.
      • Byfield F.J.
      • Xu T.
      • et al.
      Cell-cycle control by physiological matrix elasticity and in vivo tissue stiffening.
      ,
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ) and transforming growth factor-β–induced fibrosis (
      • Chen W.C.
      • Lin H.H.
      • Tang M.J.
      Regulation of proximal tubular cell differentiation and proliferation in primary culture by matrix stiffness and ECM components.
      ,
      • Leight J.L.
      • Wozniak M.A.
      • Chen S.
      • Lynch M.L.
      • Chen C.S.
      Matrix rigidity regulates a switch between TGF-beta1-induced apoptosis and epithelial-mesenchymal transition.
      ); thus, regulation of tissue stiffness level is fundamental for the physiological function of organs. Discordance of cell and matrix stiffness has been hypothesized to contribute to the pathogenesis of cancer and scarring (
      • Janmey P.A.
      • Miller R.T.
      Mechanisms of mechanical signaling in development and disease.
      ). Here, we found that KFs shared similar mechanical phenotypes with cancer cells, including cell softening and loss of stiffness sensing ability due to decreased CAV1 (
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ). Cav1 is linked to the formation of actin cap, which dominates cell stiffness and mechanosensing in NIH3T3 fibroblasts (
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ). However, immunostaining results showed that microtubule polymerization, but not actin cap organization, was weakened in KFs compared with NFs (data not shown). Microtubules function as rigid struts and contribute to cytoskeletal stiffness and cellular contractions.
      • Kawabe J.
      • Okumura S.
      • Nathanson M.A.
      • Hasebe N.
      • Ishikawa Y.
      Caveolin regulates microtubule polymerization in the vascular smooth muscle cells.
      showed that down-regulation of caveolins attenuated microtubule polymerization, which indirectly regulates cellular signaling. Microtubule disruption has been reported to increase the isometric cell force, an effect attributed to either change in intracellular Ca2+ (
      • Paul R.J.
      • Bowman P.S.
      • Kolodney M.S.
      Effects of microtubule disruption on force, velocity, stiffness and [Ca(2+)](i) in porcine coronary arteries.
      ), which regulates actomyosin contractility, or increases in myosin light chain (MLC) phosphorylation, resulting from the release of tubulin monomers (
      • Birukova A.A.
      • Smurova K.
      • Birukov K.G.
      • Usatyuk P.
      • Liu F.
      • Kaibuchi K.
      • et al.
      Microtubule disassembly induces cytoskeletal remodeling and lung vascular barrier dysfunction: role of Rho-dependent mechanisms.
      ,
      • Kolodney M.S.
      • Elson E.L.
      Contraction due to microtubule disruption is associated with increased phosphorylation of myosin regulatory light chain.
      ). Altogether, it seems reasonable to propose that decreased CAV1 may inhibit microtubule polymerization, which decreases cell stiffness (Figure 4c) but increases contractile forces, resulting from MLC2 activation (Figure 5b–e) in KFs. Cells sense stiffness through mechanosensing of strains on their cytoskeletons by outside-in signaling mechanisms (
      • Discher D.E.
      • Janmey P.
      • Wang Y.L.
      Tissue cells feel and respond to the stiffness of their substrate.
      ). These forces, exerted on the cells, trigger further changes in morphology and phenotype (
      • Engler A.J.
      • Sen S.
      • Sweeney H.L.
      • Discher D.E.
      Matrix elasticity directs stem cell lineage specification.
      ,
      • Goldmann W.H.
      Mechanosensation: a basic cellular process.
      ). Although CAV1 is linked to the stiffness sensing ability of KFs, the underlying mechanism remains unclear. Previous studies showed that β1-integrin–mediated mechanotransduction is mediated by caveolae domains (
      • Radel C.
      • Carlile-Klusacek M.
      • Rizzo V.
      Participation of caveolae in beta1 integrin-mediated mechanotransduction.
      ,
      • Wei W.C.
      • Lin H.H.
      • Shen M.R.
      • Tang M.J.
      Mechanosensing machinery for cells under low substratum rigidity.
      ). CAV1 has been shown to inhibit metastatic potential in melanomas through suppression of the integrin/Src/FAK signaling pathway (
      • Trimmer C.
      • Whitaker-Menezes D.
      • Bonuccelli G.
      • Milliman J.N.
      • Daumer K.M.
      • Aplin A.E.
      • et al.
      CAV1 inhibits metastatic potential in melanomas through suppression of the integrin/Src/FAK signaling pathway.
      ).
      • Suarez E.
      • Syed F.
      • Alonso-Rasgado T.
      • Mandal P.
      • Bayat A.
      Up-regulation of tension-related proteins in keloids: knockdown of Hsp27, alpha2beta1-integrin, and PAI-2 shows convincing reduction of extracellular matrix production.
      found that three specific tension-related genes and proteins (Hsp27, α2β1-integrin, and PAI-2), are overexpressed in keloids and regulate ECM production. Altogether, it is plausible that CAV1 interacts with the integrin/FAK network to regulate the mechanotransduction of the skin. There are some animal models developed to mimic hypertrophic scar (
      • Ibrahim M.M.
      • Bond J.
      • Bergeron A.
      • Miller K.J.
      • Ehanire T.
      • Quiles C.
      • et al.
      A novel immune competent murine hypertrophic scar contracture model: a tool to elucidate disease mechanism and develop new therapies.
      ,
      • Wong V.W.
      • Rustad K.C.
      • Akaishi S.
      • Sorkin M.
      • Glotzbach J.P.
      • Januszyk M.
      • et al.
      Focal adhesion kinase links mechanical force to skin fibrosis via inflammatory signaling.
      ) or keloid (
      • Lee Y.S.
      • Hsu T.
      • Chiu W.C.
      • Sarkozy H.
      • Kulber D.A.
      • Choi A.
      • et al.
      Keloid-derived, plasma/fibrin-based skin equivalents generate de novo dermal and epidermal pathology of keloid fibrosis in a mouse model.
      ). It would be interesting to further test our hypothesis through these in vivo models.
      In conclusion, we highlight a role of CAV1 in the progression of keloids (see also schematic model in Supplementary Figure S5 online). Epigenetically decreased CAV1 increases fibrogenesis-associated RUNX2 and alters cell mechanics, including cell softening, loss of stiffness sensing ability, and increase in contractile forces. Consequently, keloid-derived fibroblasts exhibit hyperresponsiveness to both normal dermis-equivalent stiffness (2 kPa) and pathological keloid-equivalent stiffness (20 kPa) or have established an activated state of fibrogenesis-associated RUNX2 that is independent of mechanical cues. These results could explain why KFs produce excessive ECM and prominently migrate into the surrounding normal dermal tissue, thus generating the main clinicopathologic hallmarks of keloids.

      Materials and Methods

      Human samples

      The study was approved by the Institutional Review Board at the National Cheng Kung University Hospital (NCKUH-10105017/BR-100-102). Participants gave their written informed consent as detailed in the Supplementary Materials and Methods online.

      Primary culture of fibroblasts

      Fibroblasts were isolated from dermal tissues, as described by
      • Zhang G.Y.
      • Yi C.G.
      • Li X.
      • Ma B.
      • Li Z.J.
      • Chen X.L.
      • et al.
      Troglitazone suppresses transforming growth factor-beta1-induced collagen type I expression in keloid fibroblasts.
      , and as detailed in the Supplementary Materials and Methods.

      Measurements of cell/tissue mechanical properties by AFM

      For measurements of cell/tissue stiffness, a JPK NanoWizard II AFM with BioCell (JPK Instruments, Berlin, Germany) was equipped and manipulated as previously described (
      • Chiou Y.W.
      • Lin H.K.
      • Tang M.J.
      • Lin H.H.
      • Yeh M.L.
      The influence of physical and physiological cues on atomic force microscopy-based cell stiffness assessment.
      ,
      • Harn H.I.
      • Wang Y.K.
      • Hsu C.K.
      • Ho Y.T.
      • Huang Y.W.
      • Chiu W.T.
      • et al.
      Mechanical coupling of cytoskeletal elasticity and force generation is crucial for understanding the migrating nature of keloid fibroblasts.
      ,
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ) and as detailed in the Supplementary Materials and Methods.

      Preparation and fabrication of PA gels

      PA gels with uniform stiffness were prepared as previously described (
      • Chen W.C.
      • Lin H.H.
      • Tang M.J.
      Regulation of proximal tubular cell differentiation and proliferation in primary culture by matrix stiffness and ECM components.
      ,
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ) and as detailed in the Supplementary Materials and Methods. The Young’s moduli of the PA gels used were as follows: 2- and 20-kPa PA gels represented the elastic modulus of normal and keloid dermis, respectively. PA gels from each polymerization batch were checked to verify consistent matrix mechanical properties by AFM.

      cDNA microarray analysis and ingenuity pathway analysis

      The microarray study was approved by the ethical committee at Nippon Medical School, Tokyo, Japan (no. 22-11-143) and is detailed in the Supplementary Materials and Methods.

      Western blot analyses

      Cell lysates were harvested in modified RIPA buffer. After sonication, lysates were resolved on SDS-PAGE and analyzed by Western blot as detailed in the Supplementary Materials and Methods.

      Reverse transcription-PCR

      Total RNA was extracted from cells using Trizol reagent (Invitrogen-Molecular Probes, Carlsbad, CA) according to the manufacturer’s instructions and as detailed in the Supplementary Materials and Methods.

      Immunofluorescence staining and confocal microscopy

      Immunofluorescence staining was performed as previously described (
      • Chen W.C.
      • Lin H.H.
      • Tang M.J.
      Matrix-stiffness-regulated inverse expression of Kruppel-like factor 5 and Kruppel-like factor 4 in the pathogenesis of renal fibrosis.
      ) and as detailed in the Supplementary Materials and Methods.

      Immunohistochemistry

      Immunohistochemistry was performed as previously described (
      • Lee P.T.
      • Lin H.H.
      • Jiang S.T.
      • Lu P.J.
      • Chou K.J.
      • Fang H.C.
      • et al.
      Mouse kidney progenitor cells accelerate renal regeneration and prolong survival after ischemic injury.
      ) and as detailed in the Supplementary Materials and Methods.

      Cell transfection with siRNA

      The expressions of RUNX2 and CAV1 were knocked down using commercially available RUNX2 and CAV1 siRNA kits purchased from Dharmacom (Lake Placid, NY). Fibroblasts at a density of 60% confluence were serum-deprived for 24 hours and then transfected with human RUNX2- and caveolin-1–specific siRNA or nontargeting control siRNA using the Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s protocol.

      Wound migration assay

      In vitro wound assays were performed using IBIDI (ibidi GmbH, Am Klopferspitz, Germany) culture inserts according to the manufacturer’s instructions, and as detailed in the Supplementary Materials and Methods.

      Fabrication of micropost arrays and quantification of traction force

      Polydimethylsiloxane micropost arrays were fabricated using standard microfabrication techniques as previously described (
      • Fu J.
      • Wang Y.K.
      • Yang M.T.
      • Desai R.A.
      • Yu X.
      • Liu Z.
      • et al.
      Mechanical regulation of cell function with geometrically modulated elastomeric substrates.
      ,
      • Lin H.H.
      • Lin H.K.
      • Lin I.H.
      • Chiou Y.W.
      • Chen H.W.
      • Liu C.Y.
      • et al.
      Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
      ,
      • Yang M.T.
      • Reich D.H.
      • Chen C.S.
      Measurement and analysis of traction force dynamics in response to vasoactive agonists.
      ) and as detailed in the Supplementary Materials and Methods.

      Statistical analysis

      Most data are shown as the mean ± standard error of the mean of independent experiments. Some data were normalized as described in the figure legends and are expressed as mean relative value ± standard error of the mean. Analyses of the results used analysis of variance and Student t tests by GraphPad Prism 3.0 (GraphPad Software, San Diego, CA). Each experiment was repeated at least three times to ensure the validity of the data. Values of P less than 0.05 were considered significant.

      Conflict of Interest

      The authors state no conflict of interest.

      Acknowledgments

      We thank Yuan-Yu Hsueh for assistance with tissue sampling. This work was supported by grants of the National Science Council/the Ministry of Science and Technology (NSC 99-2314-B-006-008-MY3 and MOST105-2320-B-006-044 to C-KH and MOST103-2320-B-006-044-MY3 to M-JT) and by the grants of National Cheng Kung University Hospital (NCKUH-10105017 and NCKUH-10204011 to C-KH).

      Supplementary Material

      References

        • Achterberg V.F.
        • Buscemi L.
        • Diekmann H.
        • Smith-Clerc J.
        • Schwengler H.
        • Meister J.J.
        • et al.
        The nano-scale mechanical properties of the extracellular matrix regulate dermal fibroblast function.
        J Invest Dermatol. 2014; 134: 1862-1872
        • Akaishi S.
        • Akimoto M.
        • Ogawa R.
        • Hyakusoku H.
        The relationship between keloid growth pattern and stretching tension: visual analysis using the finite element method.
        Ann Plast Surg. 2008; 60: 445-451
        • Birukova A.A.
        • Smurova K.
        • Birukov K.G.
        • Usatyuk P.
        • Liu F.
        • Kaibuchi K.
        • et al.
        Microtubule disassembly induces cytoskeletal remodeling and lung vascular barrier dysfunction: role of Rho-dependent mechanisms.
        J Cell Physiol. 2004; 201: 55-70
        • Boscher C.
        • Nabi I.R.
        Caveolin-1: role in cell signaling.
        Adv Exp Med Biol. 2012; 729: 29-50
        • Bran G.M.
        • Goessler U.R.
        • Hormann K.
        • Riedel F.
        • Sadick H.
        Keloids: current concepts of pathogenesis (review).
        Int J Mol Med. 2009; 24: 283-293
        • Castello-Cros R.
        • Whitaker-Menezes D.
        • Molchansky A.
        • Purkins G.
        • Soslowsky L.J.
        • Beason D.P.
        • et al.
        Scleroderma-like properties of skin from caveolin-1-deficient mice: implications for new treatment strategies in patients with fibrosis and systemic sclerosis.
        Cell Cycle. 2011; 10: 2140-2150
        • Chen W.
        • Fu X.
        • Sun X.
        • Sun T.
        • Zhao Z.
        • Sheng Z.
        Analysis of differentially expressed genes in keloids and normal skin with cDNA microarray.
        J Surg Res. 2003; 113: 208-216
        • Chen W.C.
        • Lin H.H.
        • Tang M.J.
        Regulation of proximal tubular cell differentiation and proliferation in primary culture by matrix stiffness and ECM components.
        Am J Physiol Renal Physiol. 2014; 307: F695-F707
        • Chen W.C.
        • Lin H.H.
        • Tang M.J.
        Matrix-stiffness-regulated inverse expression of Kruppel-like factor 5 and Kruppel-like factor 4 in the pathogenesis of renal fibrosis.
        Am J Pathol. 2015; 185: 2468-2481
        • Chiou Y.W.
        • Lin H.K.
        • Tang M.J.
        • Lin H.H.
        • Yeh M.L.
        The influence of physical and physiological cues on atomic force microscopy-based cell stiffness assessment.
        PLoS One. 2013; 8e77384
        • Cho H.H.
        • Park H.T.
        • Kim Y.J.
        • Bae Y.C.
        • Suh K.T.
        • Jung J.S.
        Induction of osteogenic differentiation of human mesenchymal stem cells by histone deacetylase inhibitors.
        J Cell Biochem. 2005; 96: 533-542
        • Del Galdo F.
        • Sotgia F.
        • de Almeida C.J.
        • Jasmin J.F.
        • Musick M.
        • Lisanti M.P.
        • et al.
        Decreased expression of caveolin 1 in patients with systemic sclerosis: crucial role in the pathogenesis of tissue fibrosis.
        Arthritis Rheum. 2008; 58: 2854-2865
        • Diao J.S.
        • Xia W.S.
        • Yi C.G.
        • Wang Y.M.
        • Li B.
        • Xia W.
        • et al.
        Trichostatin A inhibits collagen synthesis and induces apoptosis in keloid fibroblasts.
        Arch Dermatol Res. 2011; 303: 573-580
        • Discher D.E.
        • Janmey P.
        • Wang Y.L.
        Tissue cells feel and respond to the stiffness of their substrate.
        Science. 2005; 310: 1139-1143
        • Engler A.J.
        • Sen S.
        • Sweeney H.L.
        • Discher D.E.
        Matrix elasticity directs stem cell lineage specification.
        Cell. 2006; 126: 677-689
        • Fitzgerald O’Connor E.J.
        • Badshah I.I.
        • Addae L.Y.
        • Kundasamy P.
        • Thanabalasingam S.
        • Abioye D.
        • et al.
        Histone deacetylase 2 is upregulated in normal and keloid scars.
        J Invest Dermatol. 2012; 132: 1293-1296
        • Fu J.
        • Wang Y.K.
        • Yang M.T.
        • Desai R.A.
        • Yu X.
        • Liu Z.
        • et al.
        Mechanical regulation of cell function with geometrically modulated elastomeric substrates.
        Nat Methods. 2010; 7: 733-736
        • Ghosh A.K.
        • Mori Y.
        • Dowling E.
        • Varga J.
        Trichostatin A blocks TGF-beta-induced collagen gene expression in skin fibroblasts: involvement of Sp1.
        Biochem Biophys Res Commun. 2007; 354: 420-426
        • Goldmann W.H.
        Mechanosensation: a basic cellular process.
        Prog Mol Biol Transl Sci. 2014; 126: 75-102
        • Harn H.I.
        • Wang Y.K.
        • Hsu C.K.
        • Ho Y.T.
        • Huang Y.W.
        • Chiu W.T.
        • et al.
        Mechanical coupling of cytoskeletal elasticity and force generation is crucial for understanding the migrating nature of keloid fibroblasts.
        Exp Dermatol. 2015; 24: 579-584
        • Huber L.C.
        • Distler J.H.
        • Moritz F.
        • Hemmatazad H.
        • Hauser T.
        • Michel B.A.
        • et al.
        Trichostatin A prevents the accumulation of extracellular matrix in a mouse model of bleomycin-induced skin fibrosis.
        Arthritis Rheum. 2007; 56: 2755-2764
        • Ibrahim M.M.
        • Bond J.
        • Bergeron A.
        • Miller K.J.
        • Ehanire T.
        • Quiles C.
        • et al.
        A novel immune competent murine hypertrophic scar contracture model: a tool to elucidate disease mechanism and develop new therapies.
        Wound Repair Regen. 2014; 22: 755-764
        • Iwabu A.
        • Smith K.
        • Allen F.D.
        • Lauffenburger D.A.
        • Wells A.
        Epidermal growth factor induces fibroblast contractility and motility via a protein kinase C delta-dependent pathway.
        J Biol Chem. 2004; 279: 14551-14560
        • Janmey P.A.
        • Miller R.T.
        Mechanisms of mechanical signaling in development and disease.
        J Cell Sci. 2011; 124: 9-18
        • Jansen K.A.
        • Donato D.M.
        • Balcioglu H.E.
        • Schmidt T.
        • Danen E.H.
        • Koenderink G.H.
        A guide to mechanobiology: Where biology and physics meet.
        Biochim Biophys Acta. 2015; 1853: 3043-3052
        • Jumper N.
        • Paus R.
        • Bayat A.
        Functional histopathology of keloid disease.
        Histol Histopathol. 2015; 30: 1033-1057
        • Kanno T.
        • Takahashi T.
        • Tsujisawa T.
        • Ariyoshi W.
        • Nishihara T.
        Mechanical stress-mediated Runx2 activation is dependent on Ras/ERK1/2 MAPK signaling in osteoblasts.
        J Cell Biochem. 2007; 101: 1266-1277
        • Kawabe J.
        • Okumura S.
        • Nathanson M.A.
        • Hasebe N.
        • Ishikawa Y.
        Caveolin regulates microtubule polymerization in the vascular smooth muscle cells.
        Biochem Biophys Res Commun. 2006; 342: 164-169
        • Klein E.A.
        • Yin L.
        • Kothapalli D.
        • Castagnino P.
        • Byfield F.J.
        • Xu T.
        • et al.
        Cell-cycle control by physiological matrix elasticity and in vivo tissue stiffening.
        Curr Biol. 2009; 19: 1511-1518
        • Kolodney M.S.
        • Elson E.L.
        Contraction due to microtubule disruption is associated with increased phosphorylation of myosin regulatory light chain.
        Proc Natl Acad Sci USA. 1995; 92: 10252-10256
        • Lee J.A.
        • Choi D.I.
        • Choi J.Y.
        • Kim S.O.
        • Cho K.A.
        • Lee J.B.
        • et al.
        Methyl-beta-cyclodextrin up-regulates collagen I expression in chronologically-aged skin via its anti-caveolin-1 activity.
        Oncotarget. 2015; 6: 1942-1953
        • Lee J.Y.
        • Yang C.C.
        • Chao S.C.
        • Wong T.W.
        Histopathological differential diagnosis of keloid and hypertrophic scar.
        Am J Dermatopathol. 2004; 26: 379-384
        • Lee P.T.
        • Lin H.H.
        • Jiang S.T.
        • Lu P.J.
        • Chou K.J.
        • Fang H.C.
        • et al.
        Mouse kidney progenitor cells accelerate renal regeneration and prolong survival after ischemic injury.
        Stem Cells. 2010; 28: 573-584
        • Lee Y.S.
        • Hsu T.
        • Chiu W.C.
        • Sarkozy H.
        • Kulber D.A.
        • Choi A.
        • et al.
        Keloid-derived, plasma/fibrin-based skin equivalents generate de novo dermal and epidermal pathology of keloid fibrosis in a mouse model.
        Wound Repair Regen. 2016; 24: 302-316
        • Leight J.L.
        • Wozniak M.A.
        • Chen S.
        • Lynch M.L.
        • Chen C.S.
        Matrix rigidity regulates a switch between TGF-beta1-induced apoptosis and epithelial-mesenchymal transition.
        Mol Biol Cell. 2012; 23: 781-791
        • Lewis C.J.
        • Mardaryev A.N.
        • Sharov A.A.
        • Fessing M.Y.
        • Botchkarev V.A.
        The epigenetic regulation of wound healing.
        Adv Wound Care. 2014; 3: 468-475
        • Lin H.H.
        • Lin H.K.
        • Lin I.H.
        • Chiou Y.W.
        • Chen H.W.
        • Liu C.Y.
        • et al.
        Mechanical phenotype of cancer cells: cell softening and loss of stiffness sensing.
        Oncotarget. 2015; 6: 20946-20958
        • Naitoh M.
        • Kubota H.
        • Ikeda M.
        • Tanaka T.
        • Shirane H.
        • Suzuki S.
        • et al.
        Gene expression in human keloids is altered from dermal to chondrocytic and osteogenic lineage.
        Genes Cells. 2005; 10: 1081-1091
        • Nethe M.
        • Hordijk P.L.
        A model for phospho-caveolin-1-driven turnover of focal adhesions.
        Cell Adh Migr. 2011; 5: 59-64
        • Niessen F.B.
        • Spauwen P.H.
        • Schalkwijk J.
        • Kon M.
        On the nature of hypertrophic scars and keloids: a review.
        Plast Reconstr Surg. 1999; 104 (1435–5)
        • Ogawa R.
        • Akaishi S.
        • Huang C.
        • Dohi T.
        • Aoki M.
        • Omori Y.
        • et al.
        Clinical applications of basic research that shows reducing skin tension could prevent and treat abnormal scarring: the importance of fascial/subcutaneous tensile reduction sutures and flap surgery for keloid and hypertrophic scar reconstruction.
        J Nippon Med Sch. 2011; 78: 68-76
        • Ogawa R.
        • Hsu C.K.
        Mechanobiological dysregulation of the epidermis and dermis in skin disorders and in degeneration.
        J Cell Mol Med. 2013; 17: 817-822
        • Paul R.J.
        • Bowman P.S.
        • Kolodney M.S.
        Effects of microtubule disruption on force, velocity, stiffness and [Ca(2+)](i) in porcine coronary arteries.
        Am J Physiol Heart Circ Physiol. 2000; 279: H2493-H2501
        • Raaz U.
        • Schellinger I.N.
        • Chernogubova E.
        • Warnecke C.
        • Kayama Y.
        • Penov K.
        • et al.
        Transcription factor Runx2 promotes aortic fibrosis and stiffness in type 2 diabetes mellitus.
        Circ Res. 2015; 117: 513-524
        • Radel C.
        • Carlile-Klusacek M.
        • Rizzo V.
        Participation of caveolae in beta1 integrin-mediated mechanotransduction.
        Biochem Biophys Res Commun. 2007; 358: 626-631
        • Raglow Z.
        • Thomas S.M.
        Tumor matrix protein collagen XIalpha1 in cancer.
        Cancer Lett. 2015; 357: 448-453
        • Rombouts K.
        • Niki T.
        • Greenwel P.
        • Vandermonde A.
        • Wielant A.
        • Hellemans K.
        • et al.
        Trichostatin A, a histone deacetylase inhibitor, suppresses collagen synthesis and prevents TGF-beta(1)-induced fibrogenesis in skin fibroblasts.
        Exp Cell Res. 2002; 278: 184-197
        • Rubin J.
        • Schwartz Z.
        • Boyan B.D.
        • Fan X.
        • Case N.
        • Sen B.
        • et al.
        Caveolin-1 knockout mice have increased bone size and stiffness.
        J Bone Miner Res. 2007; 22: 1408-1418
        • Russell S.B.
        • Russell J.D.
        • Trupin K.M.
        • Gayden A.E.
        • Opalenik S.R.
        • Nanney L.B.
        • et al.
        Epigenetically altered wound healing in keloid fibroblasts.
        J Invest Dermatol. 2010; 130: 2489-2496
        • Saito T.
        • Nishida K.
        • Furumatsu T.
        • Yoshida A.
        • Ozawa M.
        • Ozaki T.
        Histone deacetylase inhibitors suppress mechanical stress-induced expression of RUNX-2 and ADAMTS-5 through the inhibition of the MAPK signaling pathway in cultured human chondrocytes.
        Osteoarthritis Cartilage. 2013; 21: 165-174
        • Seifert O.
        • Bayat A.
        • Geffers R.
        • Dienus K.
        • Buer J.
        • Lofgren S.
        • et al.
        Identification of unique gene expression patterns within different lesional sites of keloids.
        Wound Repair Regen. 2008; 16: 254-265
        • Stroka K.M.
        • Konstantopoulos K.
        Physical biology in cancer. 4. Physical cues guide tumor cell adhesion and migration.
        Am J Physiol Cell Physiol. 2014; 306: C98-C109
        • Suarez E.
        • Syed F.
        • Alonso-Rasgado T.
        • Mandal P.
        • Bayat A.
        Up-regulation of tension-related proteins in keloids: knockdown of Hsp27, alpha2beta1-integrin, and PAI-2 shows convincing reduction of extracellular matrix production.
        Plast Reconstr Surg. 2013; 131 (158e–73e)
        • Trimmer C.
        • Whitaker-Menezes D.
        • Bonuccelli G.
        • Milliman J.N.
        • Daumer K.M.
        • Aplin A.E.
        • et al.
        CAV1 inhibits metastatic potential in melanomas through suppression of the integrin/Src/FAK signaling pathway.
        Cancer Res. 2010; 70: 7489-7499
        • Tziotzios C.
        • Profyris C.
        • Sterling J.
        Cutaneous scarring: Pathophysiology, molecular mechanisms, and scar reduction therapeutics Part II. Strategies to reduce scar formation after dermatologic procedures.
        J Am Acad Dermatol. 2012; 66: 13-24
        • Vazquez-Villa F.
        • Garcia-Ocana M.
        • Galvan J.A.
        • Garcia-Martinez J.
        • Garcia-Pravia C.
        • Menendez-Rodriguez P.
        • et al.
        COL11A1/(pro)collagen 11A1 expression is a remarkable biomarker of human invasive carcinoma-associated stromal cells and carcinoma progression.
        Tumour Biol. 2015; 36: 2213-2222
        • Wang X.M.
        • Zhang Y.
        • Kim H.P.
        • Zhou Z.
        • Feghali-Bostwick C.A.
        • Liu F.
        • et al.
        Caveolin-1: a critical regulator of lung fibrosis in idiopathic pulmonary fibrosis.
        J Exp Med. 2006; 203: 2895-2906
        • Wang Y.H.
        • Chiu W.T.
        • Wang Y.K.
        • Wu C.C.
        • Chen T.L.
        • Teng C.F.
        • et al.
        Deregulation of AP-1 proteins in collagen gel-induced epithelial cell apoptosis mediated by low substratum rigidity.
        J Biol Chem. 2007; 282: 752-763
        • Wei W.C.
        • Lin H.H.
        • Shen M.R.
        • Tang M.J.
        Mechanosensing machinery for cells under low substratum rigidity.
        Am J Physiol Cell Physiol. 2008; 295: C1579-C1589
        • Witt E.
        • Maliri A.
        • McGrouther D.A.
        • Bayat A.
        RAC activity in keloid disease: comparative analysis of fibroblasts from margin of keloid to its surrounding normal skin.
        Eplasty. 2008; 8: e19
        • Wong V.W.
        • Akaishi S.
        • Longaker M.T.
        • Gurtner G.C.
        Pushing back: wound mechanotransduction in repair and regeneration.
        J Invest Dermatol. 2011; 131: 2186-2196
        • Wong V.W.
        • Rustad K.C.
        • Akaishi S.
        • Sorkin M.
        • Glotzbach J.P.
        • Januszyk M.
        • et al.
        Focal adhesion kinase links mechanical force to skin fibrosis via inflammatory signaling.
        Nat Med. 2011; 18: 148-152
        • Yang B.
        • Radel C.
        • Hughes D.
        • Kelemen S.
        • Rizzo V.
        p190 RhoGTPase-activating protein links the beta1 integrin/caveolin-1 mechanosignaling complex to RhoA and actin remodeling.
        Arterioscler Thromb Vasc Biol. 2011; 31: 376-383
        • Yang M.T.
        • Reich D.H.
        • Chen C.S.
        Measurement and analysis of traction force dynamics in response to vasoactive agonists.
        Integr Biol. 2011; 3: 663-674
        • Zhang G.Y.
        • Yi C.G.
        • Li X.
        • Ma B.
        • Li Z.J.
        • Chen X.L.
        • et al.
        Troglitazone suppresses transforming growth factor-beta1-induced collagen type I expression in keloid fibroblasts.
        Br J Dermatol. 2009; 160: 762-770
        • Zhang G.Y.
        • Yu Q.
        • Cheng T.
        • Liao T.
        • Nie C.L.
        • Wang A.Y.
        • et al.
        Role of caveolin-1 in the pathogenesis of tissue fibrosis by keloid-derived fibroblasts in vitro.
        Br J Dermatol. 2011; 164: 623-627
        • Ziros P.G.
        • Gil A.P.
        • Georgakopoulos T.
        • Habeos I.
        • Kletsas D.
        • Basdra E.K.
        • et al.
        The bone-specific transcriptional regulator Cbfa1 is a target of mechanical signals in osteoblastic cells.
        J Biol Chem. 2002; 277: 23934-23941