Recombinant Human Decorin Inhibits TGF-b1 Induced Contraction of Collagen Lattice by Keloid Fibroblasts

While wound contraction plays an important role in healing, it may lead to excessive scar formation and pathological wound contracture in extreme conditions. To date, the key regulator of wound contraction and keloid formation is transforming growth factor-beta (TGF-b1). Decorin has been reported to bind TGF-b1 and neutralize some of its activities. The present study investigated whether decorin affected TGF-b1-induced fibroblast contractile activity by using fibroblast-populated collagen lattice (FPCL), which has been generally used as an in-vitro model thought to mimic wound contraction in vivo, modified by the incorporation of recombinant human decorin into collagen gel. As expected, TGF-b1 significantly enhanced the contraction of collagen gel at hour 12, 24, 48, 72, and 96 (P < 0.05). Recombinant human decorin inhibited both the basal and TGF-b1-enhanced contraction of collagen gel by keloid fibroblasts (P < 0.05). These inhibitory effects of recombinant human decorin were associated with suppression of TGF-b1-induced filamentous actin (F-actin) expression in keloid fibroblasts. Furthermore, recombinant human decorin inhibited TGF-b1 induced a-smooth muscle actin (a-SMA), PAI-1 (plasminogen activator inhibitor-1) protein, and mRNA expressions in keloid fibroblasts (P < 0.05). These data indicate that recombinant human decorin can suppress TGF-b1-induced contraction of collagen gel by keloid fibroblasts. Moreover, decorin can inhibit basal contraction of collagen gel by keloid fibroblasts. These results suggest that decorin may have therapeutic potential for excessive skin contraction as observed in a keloid.

The wound healing process consists of inflammation followed by granulation tissue formation, and tissue remodeling.¹ Dermal fibroblasts actively and dynamically contribute to wound healing by migrating to the wound site, synthesizing extracellular matrix components such as collagen, forming granulation tissue,² and generating mechanical forces within the wound to initiate wound contraction.³ Wound contraction can be beneficial to overall wound healing by decreasing the wound area and forming a mechanically strong reparative scar. Undesirable wound contraction can occur and is particularly a consequence of burn and trauma wounds and can result in cosmetic and functional problems.^4–6 Human skin is firmly attached to the underlying tissues, therefore, the consequences of contraction can be detrimental ranging from minimal cosmetic scarring to major body deformation and loss of joint mobility.7 Additionally, the accumulation of an abnormally large number of fibroblasts within a healing wound can result in excessive wound contraction or contractures.⁸

Early skin grafting is an effective means of reducing wound contraction, but it gives variable results depending on the thickness and site from which it is taken.⁹ Thin split-thickness skin grafts give almost as much contraction as healing by secondary intention alone, whereas full-thickness skin grafts can markedly reduce it. Physical factors alone do not account for this since full-thickness skin grafts from thinner areas of skin (eg, the upper eyelid or post-auricular region) do not lead to contraction, whereas the much thicker split-thickness grafts from abdominal skin do result in contraction.¹⁰ Furthermore, it is not the graft itself that contracts, but rather the bed of new granulation tissue forming underneath the graft. Fibroblasts play a major role in wound contraction.^2,9,11 In order to quantify the contractile ability of fibroblasts during wound contraction, a floating collagen gel with fibroblasts incorporated has been frequently used as an in-vitro model.^12–15

Numerous studies suggest that transforming growth factor (TGF)-βb1 affects the process of wound healing at various phases including wound contraction.^13,16,17 Transforming growth factor-β1 augments fibroblast contraction of collagen gels in vitro^13,16 and enhances the wound contraction process in vivo.¹⁷ Increased synthesis of actin (including a-SMA), fibronectin, and matrix receptors may represent the potential mechanism by which TGF-b1 enhances the efficiency of collagen gel contraction by fibroblasts.

Decorin is a secreted 45 kDa proteoglycan with a core protein comprised primarily of leucine-rich repeats,18 and is found in the extracellular matrix (ECM) of several tissues such as skin,¹⁹ cartilage,²⁰ and bone.²¹ Due to the leucine-rich repeats, decorin is thought to interact with several other proteins and lipid molecules. Decorin has been reported to bind TGF-b1 and neutralize some of its activities.^22,23 Decorin, either injected or synthesized in vivo from an expression vector, has been shown to have a beneficial effect of antifibrosis in various tissues such as kidney,²⁴ muscle,²⁵ and lung.²⁶ Previous work from the authors’ group showed that recombinant human decorin inhibited cell proliferation and downregulated TGF-b1 production in keloid fibroblasts.²⁷ The question addressed in the present report is whether decorin influences not only proliferative activity, but also the ability of fibroblasts to cause contraction in the collagen gel assay.

The present study investigated whether decorin affected TGF-b1-induced fibroblast contractile activity by using a floating collagen gel with incorporated keloid-derived fibroblasts.

Cell culture
The keloid fibroblast was established²⁷ as a primary cell line from keloid tissue obtained from 6 patients whose scars extended above and beyond the sites of original injury. Scarring resulted in functional impairment and aesthetic deformity, and an operation was needed. Exemption to use operative specimens that would otherwise be discarded was obtained from the Human Subjects Committee of Jinan University, China.

Using a sterile technique under a laminar flow hood, the dermal specimen was minced into approximately 1-mm³ fragments with a sterile scalpel blade on a Petri dish. The specimens were washed in phosphate-buffered saline (PBS) solution with a combination of 1% penicillin, streptomycin sulfate, and amphotericin B (Gibco®, Invitrogen, Beijing, China). The specimens were then placed in 75-mm² tissue culture flasks (T75, Falcon, Becton, Dickinson and Company, Franklin Lakes, NJ) with 10-mL of culture medium (10% fetal calf serum in Dulbecco-modified Eagle medium with 1% penicillin-streptomycin sulfate-amphotericin B). The specimens were then maintained in a humidified incubator at 37˚C with a 5% carbon dioxide atmosphere. After 24 hours, the medium was changed to 7-mL of culture medium. The medium was then changed every 2 days until fibroblasts were seen growing outward from the explanted tissue under light microscopy. At that time, the tissue was removed. With sufficient outgrowth of fibroblasts, cells were subcultured into 75-mm² culture flasks. Culture medium was changed every 3 to 4 days. Successive cultures were passed at confluence. Experiments were performed with early passage cells (passage 3 through 6). Cells at the same passage were used to examine the effect of decorin on collagen lattice contraction induced by TGF-b1.

Collagen lattice contraction assay
Fibroblast populated collagen lattice was prepared as previously described.^14,15 Briefly, the collagen solution was prepared by mixing acid-soluble type I collagen (2 mg/mL), a 5-fold concentration of DMEM, and a buffer solution (0.05 M NaOH, 2.2% NaHCO3, 200 mM HEPES) in the ratio 7:2:1 (all purchased from Sigma-Aldrich, St. Louis, MO). A mixture of cell suspension in serum-free DMEM with or without recombinant human decorin (R&D Systems, Minneapolis, MN) at final concentrations of 10, 100, or 200 nM and collagen solution was divided into the wells (1.0 mL) of a 6-well cell culture cluster (Costar®, Corning, Lowell, MA) and then gelled at 37˚C for 30 minutes. The final concentration of collagen was 1.0 mg/mL with a cell density of 2.0 × 10⁵ cells/mL in the presence or absence of 10-, 100-, or 200-nM recombinant human decorin. Serum-free DMEM (2 mL) was then poured onto the gel to prevent the surface from dehydrating. After 12 hours of incubation, the gel was separated from each well, floated and treated with 5 ng/mL recombinant human TGF-b1 or a control vehicle (4 mM HCl containing 1 mg/mL of bovine serum albumin).

At 0, 12, 24, 48, 72, and 96 hours after the application of TGF-b1 or the control vehicle, the major and minor axis of each gel sample was measured, and the surface area was calculated. The contraction of the gel was expressed as a percentage of initial lattice area, with the surface area of the noncontracted state serving as 100%: percentage of initial area = A2 ⁄ A1 × 100, where A1 = initial gel area and A2 = area at the observed interval. Six culture plates were used for each experimental group (n = 6).

Immunofluorescence microscopy for F-actin
Cultured keloid fibroblasts in collagen gels treated with 5 ng/mL TGF-b1 in the presence or absence of 200 nM recombinant human decorin for 24 hours, were rinsed 3 times with DPBS and fixed with 4% paraformaldehyde containing 5% sucrose at room temperature for 30 minutes. The cells were then washed with Dulbecco’s Phosphate-buffered Saline (DPBS) 3 times every 5 minutes, and treated with 0.2% Triton X-100 (Sigma-Aldrich) at room temperature for 5 minutes. After washing with DPBS 3 times every 5 minutes, they were stained for filamentous actin (F-actin) with fluorescent isothiocyanate (FITC)-conjugated phalloidin (2 µg/mL, Sigma-Aldrich) for 40 minutes, washed with DPBS 3 times every 5 minutes, and embedded in 80% glycerol. All stained samples were examined with a laser scanning confocal microscopy (LSM 410, Zeiss, Germany).

Western blot analysis
Keloid fibroblasts were stimulated with or without 5 ng/mL TGF-b1 for 24 hours in the presence or absence of 200-nM decorin. Cells were washed with sterile PBS twice and then 100-µL cell lysis buffer (35-mM Tris-HCl, pH 7.4, 0.4-mM ethylene glycol tetraacetic acid [EGTA], 10-mM MgCl2, 100-µg/mL aprotinin, 1-µM phenylmethylsulfonyl fluoride, 1-µg/mL leupeptin, and 0.1% Triton X-100) was added. Lysates were centrifuged at 10,000 g for 10 minutes. The protein concentration in the cell lysates was measured using a protein assay kit (Bio-Rad, Hercules, CA). Cell lysate samples heated at 95˚C for 5 minutes before loading (20-µg/lane), and 10% sodium dodecyl sulfate (SDS)—polyacrylamide gel electrophoresis (PAGE) was performed. After SDS-PAGE, proteins were transferred onto a polyvinylidene fluoride (PVDF) membrane (Bio-Rad). The membrane was blocked for 1 hour at room temperature with 5% skim milk in PBS-Tween and incubated overnight at 4˚C with mouse anti-human plasminogen activator inhibitor 1 (PAI-1) monoclonal antibody (Novocastra, Bucks, UK), mouse anti-human a-SMA (a-smooth muscle actin) monoclonal antibody (Sigma-Aldrich), and rabbit anti-human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) monoclonal antibody (Cell Signaling Technology, Danvers, MA). After incubation with horseradish peroxidase (HRP)-conjugated anti-mouse IgG, or anti-rabbit IgG, an ECL Western blots detection system (Amersham Biosciences, Piscataway, NJ) was used according to the manufacturer’s instructions. The densities of the bands were quantified with the Fluor-s MultiImager™ and Quantity One® 4.1 photo analysis software (Bio-Rad). The expression levels of PAI-1 and a-SMA were normalized to the corresponding GAPDH determined as an internal control. Data presented are from means of triplicate from one experiment that was replicated on 4 separate occasions (n = 4).

RNA isolation and reverse transcriptase PCR (RT-PCR) analysis
Total ribonucleic acid (RNA) was prepared using the TRIzol® reagent (Life Technologies, Gaithersburg, MD), according to the manufacturer’s instructions. The RT-PCR was performed with a SuperScript One-Step™ RT-PCR System (Life Technologies) according to the manufacturer’s instructions. Briefly, 1-µg of RNA was reverse-transcribed. The primer sets used were as follows: for PAI-1, forward 5'-TGCTGGTGAATGCCCTCTACT-3' and reverse 5'-CGGTCATTCCCAGGTTCTCTA-3'; for a-SMA, forward 5'-AGGAAGGACCTCTAT GCTAACAAT-3' and reverse 5'-AACACATAGGTAACGAGTCAGAGC-3'; for βb-actin, forward 5'-ACTTAGTTGCGTTACACCCTTTCT-3' and reverse 5'-TTCATACATCTCA AGTTGGGGGAC-3'. The following was done for each RT-PCR that was included: a cDNA synthesis and predenaturation cycle at 95˚C for 5 minutes; the cDNA was amplified for 30 cycles involving a denaturation step at 95˚C for 1 minute; a primer annealing step at 55˚C for 60 seconds; and an extension step at 72˚C for 1 minute. The PCR products were analyzed by electrophoresis on a 2% agarose gel containing ethidium bromide and visualized and photographed under UV light. In all experiments, two control reactions, one containing no mRNA and another containing mRNA but no reverse transcriptase or Taq polymerase, were included. Densitometric scanning was performed with the Fluor-s MultiImager. The data were analyzed with Quantity One 4.1 photo analysis software and normalized to those of β-actin used as an internal control. The mRNA values are expressed as relative units calculated according to the following formula: density of the PAI-1 or a-SMA / density of the βb-actin. Data are presented as means of results from 4 experiments each performed in duplicate (n = 4).

Statistical analysis
Data were expressed as mean ± standard error of mean (SEM). Statistical evaluation of the continuous data was performed by one-way analysis of variance, followed by Dunnett’s t-test for between group comparisons. The level of significance was P < 0.05.

Decorin inhibits basal and TGF-b1 enhanced contraction of collagen gel by keloid fibroblasts 
To determine whether decorin inhibited TGF-b1 mediated fibroblast contraction in vitro, an established in-vitro wound contraction model was used with a type I collagen gel with embedded keloid fibroblasts and some modifications.^14,15 Briefly, a mixture of human keloid fibroblasts and type I collagen in a serum-free medium containing various concentrations of recombinant human decorin was gelled, and then separated from the well after a 12-hour incubation when the fibroblasts had fully stretched. TGF-b1 or a control vehicle was then applied to the medium, and contractile activity was analyzed by measuring the amount of surface area contraction.

Keloid-derived fibroblasts showed basal level of contraction of collagen gel to 68.5%±6.5%, 39.8%±5.3%, 32.4%±3.4%, 26.2%±3.4%, and 22.6%±2.7% of the initial area at hour 12, 24, 48, 72, 96, respectively (Figure 1). Addition of TGF-b1 further enhanced the contraction of collagen gel at 12, 24, 48, 72, 96 hours after TGF-b1 stimulation (P < 0.05 as compared with control). Two hundred-nM decorin significantly inhibited the basal level of collagen gel contraction (P < 0.05, compared to control) and the TGF-b1 mediated enhancement of contraction (P < 0.05 as compared with TGF-b1 group). There were no significant differences in 10-nM, 100-nM, and 200-nM of decorin in inhibiting TGF-b1 mediated contraction of collagen gel. These results indicated that recombinant human decorin inhibited the basal and TGF-b1 mediated contraction of collagen gel by keloid dermal fibroblasts.

Decorin inhibits TGF-b1 induced F-actin expression in keloid fibroblasts 
Fibroblasts, rich in F-actin bundles, generate the force of wound contraction.^28,29 Fluorescent isothiocyanate-conjugated phalloidin (FITC) staining was performed to visualize actin filaments. A bright green fluorescence marked the positive expression of F-actin (Figure 2). Many of the F-actin bundles were formed in keloid fibroblasts (Figure 2A), which were significantly enhanced by addition of TGF-b1 (extensive F-actin bundle formation, Figure 2C). Recombinant human decorin inhibited both the basal level of F-actin expression (Figure 2B) and the TGF-b1 mediated enhancement of F-actin expression (Figure 2D). Three pairs of keloid-derived dermal fibroblasts were examined and a representative result is shown. Similar results were obtained in two other independent experiments.

Decorin inhibits TGF-b1 induced αa-SMA expression in keloid fibroblasts 
Myofibroblasts that differentiate from fibroblasts by TGF-b1 have been suggested to be responsible for wound contraction.^2,9 a-SMA, an actin isoform of the vascular smooth muscle cell type, is thought to be one of the most useful markers for myofibroblast phenotype. Importantly, a-SMA is not only a marker for myofibroblasts, but the expression of αa-SMA may be involved in fibroblast contractile activity.^30,31 Thus, the effect of recombinant human decorin on αa-SMA expression was examined. Western blot results showed that the expressions of a-SMA protein were significantly increased in keloid dermal fibroblasts after 5-ng/mL TGF-b1 stimulation (P < 0.05), and recombinant human decorin inhibited TGF-b1 induced a-SMA expression in keloid fibroblasts (P < 0.05; Figure 3).

Decorin inhibits TGF-b1 induced PAI-1 expression in keloid fibroblasts
Fibroblast locomotion is thought to generate tractional forces which lead to contraction of collagen lattices.³² Plasminogen activator inhibitor-1 (PAI-1) is an important physiological regulator of cell motility. PAI-1 synthesis and its matrix deposition, in response to TGF-b1, correlated with a significant increase in cell motility.³³ PAI-1 expression was an important aspect in cellular movement, as PAI-1-deficient cells had significantly impaired basal locomotion and were unresponsive to TGF-b1.³⁴ Thus, we examined the effect of recombinant human decorin on PAI-1 expression. Western blot results showed that the expressions of PAI-1 protein were significantly increased in keloid dermal fibroblasts after 5-ng/mL TGF-b1 stimulation (P < 0.05), and recombinant human decorin inhibited TGF-b1 induced PAI-1 expression in keloid fibroblasts (P < 0.05; Figure 4).

Decorin inhibits TGF-b1 induced αa-SMA mRNA expression in keloid fibroblasts
Reverse transcriptase PCR results showed that the expressions of a-SMA mRNA were significantly increased in keloid dermal fibroblasts after 5-ng/mL TGF-b1 stimulation (P < 0.05), and recombinant human decorin inhibited TGF-b1 induced αa-SMA mRNA expression in keloid fibroblasts (P < 0.05; Figure 5).

Decorin inhibits TGF-b1-induced PAI-1 mRNA expression in keloid fibroblasts
Reverse transcriptase PCR results showed that the expressions of PAI-1 mRNA were significantly increased in keloid dermal fibroblasts after 5-ng/mL TGF-b1 stimulation (P < 0.05), and recombinant human decorin inhibited TGF-b1 induced PAI-1 mRNA expression in keloid fibroblasts (P < 0.05; Figure 6).

Fibroblasts cultured in a 3-dimensional collagen matrix attach to the collagen fibers and generate mechanical tension.² If maintained in floating culture, fibroblasts will cause the surrounding matrix to contract. ³⁵ The fibroblast collagen matrix contraction model provides a unique way to study wound contraction. Analogous to the in-vivo situation, cell adhesion occurs 3-dimensionally to attachment sites comprised of protein fibrils, rather than 2-dimensionally along a protein-coated interface. Moreover, fibroblasts develop reciprocal geometric and mechanical relationships with the collagen matrix that are difficult to appreciate or to examine in routine cell culture because the culture surface is, for all practical purposes, fixed in place. The most conspicuous activity of the fibroblasts is to contract the collagen network or lattice, as one would expect to occur within wounds in vivo.³⁶ Measuring the degree of reduction of collagen contraction is a widely accepted method of assessing drugs that can affect this process.^14,15,37

TGF-b1 plays an important role in the pathogenesis of the keloid. Increased mRNA expression of TGF-b1 and collagen has been observed in the keloid.³⁸ Previous studies demonstrated that keloid fibroblasts produced increased TGF-b1.³⁹ TGF-b1 has also been demonstrated to enhance the contraction of collagen gels in vitro^13,16 and wound contraction in vivo,¹⁷ in addition to supporting its role in wound contraction. Moreover, contraction of collagen gel, populated with keloid fibroblasts, was attenuated with the addition of anti- TGF-b1 neutralizing antibody,⁴⁰ indicating the importance of TGF-b1 in keloid scar contracture.

In the present study, the fibroblast-populated collagen lattice (FPCL), which is an established 3-dimensional, in-vitro wound contraction and scar contracture model, modified by the incorporation of recombinant human decorin into collagen gel, by creating a hybrid lattice where keloid fibroblasts interact simultaneously with decorin and collagen mimicking the biological environment found in vivo. The presented results show that keloid fibroblasts had the basal contraction of collagen gel, which was enhanced after TGF-b1 stimulation as previously described, ^13,16,17 and recombinant human decorin inhibited both the basal and TGF-b1 enhanced contraction of collagen gel by keloid fibroblasts.

Although decorin may inhibit the proliferation of keloid fibroblasts,²⁷ it is not likely the basis of the contraction-inhibiting activity since fibroblasts usually do not grow during the collagen gel contraction assay.⁴¹ Since fibroblasts, which are rich in F-actin bundles, generate the force of contraction,^28,29 it was observed that the expression and distribution of F-actin in keloid fibroblasts was stimulated by TGF-b1 in the presence or absence of recombinant human decorin. These results suggest that the gel contraction-inhibitory effect of decorin was associated with suppression of TGF-b1 induced F-actin expression in keloid fibroblasts.

Although it was demonstrated that decorin inhibited both the basal and TGF-b1 enhanced contraction of collagen gel, only when it was incorporated into the gel did the decorin present in the media not exhibit this inhibitory activity (unpublished observations). Further studies are needed to clarify the exact mechanisms.

Myofibroblasts, a group of actin-rich fibroblasts, are known to play an important role in wound contraction.² Wound contraction is an important step in wound repair involving the conversion of protomyofibroblasts to differentiated myofibroblasts by the production of a-SMA in order to generate more force for contracture. In normal wound repair these myofibroblasts disappear by apoptosis. In pathological conditions the myofibroblasts persist and continue to remodel the extracellular matrix, resulting in connective tissue contracture,⁴² as increased expression of TGF-b1 directly induces αa-SMA expression.⁴³ The results of the present study show that the expressions of αa-SMA protein and mRNA increased after TGF-b1 stimulation and that recombinant human decorin prevented the increases in a-SMA expression that were observed with TGF-b1 stimulation in keloid fibroblasts.

In some animal wounds, daily removal of the newly formed granulation tissue does not prevent wound contraction. In the early phases of wound repair before mechanical loading occurs, wound contraction might occur by cell-migratory activity. Harris et al⁴⁴ demonstrated that the contraction of the collagen lattice occurred as a consequence of fibroblast migration through the matrix. This process was termed “tractional remodeling.” Migration of fibroblasts into and through the extracellular matrix during the initial phase of wound healing, prior to the expression of alpha-SMA, appears to be a fundamental component of wound contraction.⁹ Fibroblast locomotion is thought to generate tractional forces, which lead to contraction of collagen lattices.³² PAI-1 is an important physiological regulator of cell motility. PAI-1 synthesis and its matrix deposition in response to TGF-b1 correlated with a significant increase in cell motility.³³ PAI-1 expression was an important aspect in cellular movement, as PAI-1-deficient cells had significantly impaired basal locomotion and were unresponsive to TGF-b1.³⁴ Thus, the effect of decorin on PAI-1 expression was examined, and the presented results show that recombinant human decorin inhibited the increases in PAI-1 protein and mRNA expression that were observed with TGF-b1 stimulation in keloid fibroblasts.

Decorin efficiently inhibited TGF-b1 induced contraction of collagen lattice by keloid fibroblasts, which was associated with suppression of TGF-b1 induced F-actin, αa-SMA, and PAI-1 expression. Moreover, decorin reduced the basal level of contraction and F-actin expression in keloid fibroblasts. Therefore, decorin may have therapeutic potential for excessive skin contraction as observed in a keloid.

Address correspondence to:
Zhi Zhang, MD
Department of Burn and Plastic Surgery
Guangzhou Red Cross Hospital,
Jinan University No. 396 Tong Fu Zhong Road
Guangzhou 510220 China
Phone: 86 20 3437 1669
E-mail: zhangzhicc48@163.com