Extracellular Matrix Elasticity Modulates TGF-β–Induced p38 Activation and Myofibroblast Transdifferentiation in Human Tenon Fibroblasts

Abstract

Purpose.: Extracellular matrix and the cytokine TGF-β influence scar formation in an interdependent fashion. In this study, the impact of extracellular matrix elasticity on TGF-β–induced signal transduction and myofibroblast transdifferentiation was examined.

Methods.: Primary human tenon fibroblasts were seeded on collagen-coated glass coverslips (rigid environment) or collagen or polyacrylamide gels (elastic environment) of different compliance and stimulated with TGF-β. Myofibroblast transdifferentiation was assessed by reverse transcription–quantitative polymerase chain reaction (RT-qPCR) and Western blot analysis for the marker gene α-smooth muscle actin (SMA), and SMA incorporation into stress fibers was determined by confocal immunofluorescence microscopy. CTGF transcription was assessed by RT-qPCR. Signaling pathways were examined by Western blot using phosphospecific antibodies and by immunofluorescence microscopy.

Results.: TGF-β–dependent myofibroblast transdifferentiation was enhanced in a stiff environment. Increasing matrix elasticity attenuated TGF-β–induced myofibroblast transdifferentiation and the associated CTGF expression. TGF-β–induced p38 activation was reduced on elastic substrates.

Conclusions.: The results suggest that matrix elasticity influences TGF-β–dependent activation of p38 signaling and subsequent myofibroblast transdifferentiation. Biomechanical cues represent an important determinant of scarring processes. Therefore, cellular signals elicited by mechanotransduction deserve consideration in the design of novel antifibrotic strategies.

Conjunctival scarring is considered the main cause of long-term failure after filtering glaucoma surgery. On the cellular level, transdifferentiation of tenon fibroblasts to myofibroblasts is the hallmark of scarring. ¹ The cytokine TGF-β is regarded as the most important factor driving this process. TGF-β activates several cellular signaling cascades, among them the canonical SMAD pathway and mitogen-activated protein kinases such as p38. 

Besides growth factor stimulation, biomechanical factors such as extracellular matrix elasticity are known to influence cell functions. Seminal studies by Pelham and Wang ² revealed that a stiffer environment promotes cell spreading, assembly of focal adhesions, and actin stress fiber formation. ² More recently, an essential influence of ECM elasticity on mesenchymal stem cell differentiation has been demonstrated. ³  

The influence of tissue strain or mechanical load on healing and scar formation is clinically well established, but its molecular underpinnings are just beginning to be understood. In a rat wound-healing model, stretching of wound edges caused a significant increase in myofibroblast transdifferentiation. ⁴ TGF-β-driven transdifferentiation of fibroblasts seeded in collagen gels in vitro has been positively associated with gel stiffness, and release of mechanical tension causes myofibroblast dedifferentiation and apoptosis. ^{4 –6}  

Earlier studies have shown that TGF-β–induced p38 activation is essential for the transdifferentiation process in human tenon fibroblasts. ⁷ Hampering the contractile apparatus of the cell by ROCK inhibitors also interferes with TGF-β–driven myofibroblast transdifferentiation. Moreover, inhibition of cell contractility prevents TGF-β–induced activation of p38 with little effect on the SMAD pathway. ^8,9 This suggests a link of biomechanical cues, TGF-β–induced p38 signaling, and subsequent myofibroblast transdifferentiation. 

We report that extracellular matrix elasticity modulates TGF-β–induced p38 activation and subsequent myofibroblast transdifferentiation, whereas SMAD signaling appears to be less dependent on matrix elasticity. TGF-β–induced CTGF expression, an important cofactor in scarring, is also enhanced on rigid substrates. Our results suggest that matrix elasticity should be regarded as a relevant factor in conjunctival scarring. This appears particularly relevant, as long-standing, low-grade conjunctival inflammation induced by topical medication often precedes filtering glaucoma surgery and may result in significant alterations of biomechanical tissue characteristics. 

Materials and Methods

Reagents

A total RNA purification kit (RNeasy) and RNase-free DNase were purchased from Qiagen (Hilden, Germany), Taq polymerase and M-MLV reverse transcriptase (Impro II; Promega, Mannheim, Germany). Antibodies raised against the following proteins were used: α-SMA (SMA; Sigma-Aldrich, St. Louis, MO), active p38 (Promega), active SMAD2/3 (Zymed/Zytomed, Berlin, Germany), total SMAD2/3 (Chemicon, Temecula, CA), and Alexa-488-conjugated goat anti-mouse (Molecular Probes, Eugene, OR) and HRP-conjugated secondary antibodies (Jackson/Dianova, Hamburg, Germany). Recombinant TGF-β1 was obtained from Tebu-Bio (Offenbach, Germany) and used at 2 ng/mL (final concentration) in all experiments. 

Cell Culture

Small tenons biopsies were obtained during standard intraocular surgery after comprehensive information and written consent of selected patients. The protocol complied with the Declaration of Helsinki, and an institutional ethics committee approval was granted. Primary human tenon fibroblasts (HTFs) were obtained from an expansion culture of human tenon explants and propagated in Dulbecco's modified Eagle's medium (DMEM; PAA Laboratories GmbH, Pasching, Austria) supplemented with 10% heat-inactivated fetal calf serum (Biochrom, Berlin, Germany), 100 U/mL penicillin, and 100 μg/mL streptomycin (both PAA). The cells were maintained in the logarithmic growth phase. For all experiments, cells from passages 3 to 6 were used. The cells were starved overnight in 0.2% FCS before stimulation with vehicle or TGF-β1, as indicated. All experiments were performed at least three times with similar results. 

Flexible Substrates

Polyacrylamide substrates were prepared as described, with modifications. ^2,10,11 The gels were composed (final concentrations) of 7% acrylamide (Sigma-Aldrich), 3% acrylamidopropyl-trimethylammonium-chloride (Sigma-Aldrich), 0.1% to 0.8% bis-acrylamide to control elasticity (Sigma-Aldrich), 0.1% ammoniumpersulfate, and 0.2% N,N,N′,N′-tetramethylmethylendiamine (TEMED, Sigma-Aldrich) in PBS. The charged trimethylammonium-chloride compound allows for ECM binding. Gels containing 0.1% and 0.2% bis-acrylamide will be termed soft gels, as opposed to stiff gels containing 0.4% or 0.8% bis-acrylamide. For immunofluorescent stains, 24-mm round coverslips were coated with amino-silane (Sigma-Aldrich) for 2 minutes, washed in water, activated with 0.5% glutaraldehyde (Carl Roth, Karlsruhe, Germany) for 30 minutes, washed again, and air-dried. Polyacrylamide solution, (14 μL; as above), was pipetted onto the larger coverslip and covered with an 18-mm round coverslip. After polymerization for 30 minutes, the smaller coverslip was removed, and the gel was washed in PBS and coated with collagen (2.5 μg/mL PBS) for 1 hour in a cell culture incubator. To cultivate cells for Western blots, gels of 1 mm thickness were cast in a regular minigel casting stand (Bio-Rad, Hercules, CA), removed after a 60-minute polymerization, and washed in PBS. A 34-mm customized trephin was used to punch out smaller round gels, which were transferred to 35-mm dishes of respective size and coated with collagen as above. 

Collagen Gels

A 0.4-mg/mL collagen solution was prepared using 130 μL bovine collagen 3 mg/mL (PureCol; AdvancedBioMatrix, San Diego, CA) in 100 μL 10× DMEM (Sigma-Aldrich), 10 μL NaHCO₃ (Sigma-Aldrich), and 760 μL DMEM. The collagen solution (200 μL) was pipetted onto glass coverslips to form a collagen gel. After polymerization, HTFs were seeded on top of the collagen gels. 

RNA Isolation and Reverse Transcription–Quantitative Polymerase Chain Reaction (RT-qPCR)

HTFs were plated in 60-mm dishes or 60-mm polyacrylamide gels at 2 × 10⁵ cells/dish, serum-starved for 12 hours, and stimulated with TGF-β. Total RNA was harvested by scraping the cells off in PBS, spinning down the cell pellet, and using a kit for cell lysis and mRNA extraction, according to the manufacturer's recommendations (RNeasy; Qiagen, Hilden, Germany). To eliminate contamination with genomic DNA, DNase digestion was performed for 15 minutes. First-strand cDNA was synthesized for reverse transcription (Superscript II Reverse Transcriptase; Invitrogen, Karlsruhe, Germany) at 42°C using 500 ng total RNA extract. 

Quantitative PCR was conducted in micropipettes on a thermal cycler (LightCycler; Roche, Mannheim, Germany) with a total reaction volume of 20 μL containing 2 μL of a 1:4 dilution of first-strand reaction product, 0.4 μM specific upstream and downstream primers, 200 μM dNTP-mix, 1× reaction buffer, Taq polymerase (HotStarTaq; Qiagen), and SYBR-green (Sigma-Aldrich). Cycling conditions were initial denaturation at 95°C for 3 minutes, followed by 40 cycles consisting of a 15-second denaturation interval and a 1-minute interval for annealing and primer extension at 60°C. Amplification of the reference gene hypoxanthin-guanin-phosphoribosyl-transferase 1 (HPRT1) mRNA, which served as a normalization standard, was performed with HPRT1-forward (GACCAGTCAACAGGGGACAT) and HPRT1-reverse (ACACTTCGTGGGGTCCTTTT) primers. Side-strand–specific primers for α-SMA were a-SMA-forward (CTGTTCCAGCCATCCTTCAT), a-SMA-reverse (CCGTGATCTCCTTCTGCATT), and for CTGF were CTGF-forward (CCTGGTCCAGACCACAGAGT), and CTGF-reverse (TGG AGATTTTGGGAGTACGG). a-SMA and CTGF mRNA levels were measured as CT threshold levels and normalized with the individual HPRT1 control CT values. 

Western Blot Analysis

Cells were rinsed with ice-cold PBS, and total cell protein extracts were prepared using a RIPA lysis buffer (20 mM Tris, 150 mM NaCl, 0.1 mM EDTA, 1% Triton X-100, 1% deoxycholate, and 0.1% SDS) containing phosphatase inhibitors (1 mM sodium vanadate and 50 mM NaF) and protease inhibitors (0.1% PMSF; Complete Protease Inhibitor; Roche). Protein extracts were boiled in Laemmli sample buffer and subjected to SDS polyacrylamide gel electrophoresis. Proteins were transferred onto a PVDF membrane (Amersham, Braunschweig, Germany) using a gel blotting apparatus (Bio-Rad). Membranes were blocked in 3% BSA in TBST (10 mM Tris HCl [pH 6.8], 150 mM NaCl and 0.1% Tween 20) for 1 hour. Membranes were incubated with primary antibody overnight at 4°C and with a peroxidase-conjugated secondary antibody for 45 minutes at room temperature. After each incubation step, the membranes were washed in TBST for 30 minutes. Peroxidase was visualized by enhanced chemoluminescence (ECL) and exposure to ECL films (Hyperfilm; both Amersham) for the appropriate times. 

Immunofluorescence Confocal Microscopy

Glass coverslips or polyacrylamide gels on coverslips were coated with collagen (2.5 μg/mL) for 1 hour at 37°C in the cell incubator and washed with PBS. HTF cells were then plated in FCS-supplemented DMEM on collagen-coated glass coverslips or polyacrylamide gels and incubated for 24 hours to allow attachment. The cells were serum deprived overnight and subsequently stimulated with TGF-β [2 ng/mL], as indicated. The cells were fixed in 2% paraformaldehyde and permeabilized in 0.2% Triton X-100, blocked in 2% normal goat serum (Jackson ImmunoResearch, Hamburg, Germany), and labeled with a primary antibody against SMA (Sigma-Aldrich) at a dilution of 1:500 in blocking buffer. Alexa-488-conjugated secondary antibody against mouse IgG was used at a 1:500 dilution in blocking buffer. Rhodamine-phalloidin was used to counterstain the F-actin cytoskeleton. The cells were viewed with a laser scanning confocal microscope (TCS SP-2; Leica Microsystems, Bensheim, Germany). 

Results

Substrate Elasticity Modulates TGF-β–Induced p38 Signaling

While a heavy p38 isoform is often phosphorylated independently of TGF-β stimulation, a second p38 isoform is activated 12 to 24 hours after TGF-β stimulation. In previous experiments we had observed that activation of this smaller p38 isoform is associated with and probably mandatory for TGF-β–driven myofibroblast transdifferentiation. ⁷  

On regular rigid cell culture substrates, TGF-β–induced p38 phosphorylation has been reported (Fig. 1). In contrast, p38 activation was significantly attenuated on elastic substrates (Fig. 1) with stronger reduction on softer matrices (0.1% polyacrylamide) compared with stiffer matrices (0.4% polyacrylamide). Baseline p38 activation also inversely correlated with matrix elasticity. TGF-β–induced SMAD phosphorylation was also attenuated on softer polyacrylamide gels to some extent, but SMAD nuclear translocation was preserved (Fig. 2). 

Figure 1.

 

Matrix elasticity modulates TGF-β–induced signaling. HTFs were plated on collagen-coated tissue culture plastic and collagen-coated polyacrylamide gels of different elasticity and starved overnight. The cells were stimulated for the indicated times with TGF-β and harvested for Western blot analysis. TGF-β–induced phosphorylation of p38 or SMAD is shown. PA, polyacrylamide.

Figure 2.

 

SMAD translocation is independent of matrix elasticity. HTFs were plated on collagen-coated tissue culture plastic and collagen-coated polyacrylamide gels of different elasticity and starved overnight. The cells were stimulated for 1 hour with TGF-β and stained for SMAD2/3. TGF-β–induced nuclear translocation of SMAD2 is independent of matrix elasticity. PA, polyacrylamide.

Substrate Elasticity Modulates TGF-β–Induced SMA Expression

Previous observations had suggested an indispensible role of p38 signaling in TGF-β–induced SMA expression and myofibroblast transdifferentiation. ⁷ We therefore investigated the effect of substrate elasticity on TGF-β–induced SMA expression. Human tenon fibroblasts were seeded on 0.1% bis-acrylamide polyacrylamide gels (soft substrate), 0.4% bis-acrylamide polyacrylamide gels (intermediate substrate), or tissue culture plastic (rigid substrate). On the transcriptional level, TGF-β induced SMA only on rigid substrates (Fig. 3A). On the protein level, TGF-β stimulation elicited some SMA expression on gel substrates but significantly more when the HTFs were plated on rigid plastic dishes (Fig. 3B). 

Figure 3.

 

Matrix elasticity modulates TGF-β–induced SMA expression. HTF were plated on collagen-coated tissue culture plastic and collagen-coated polyacrylamide gels of different elasticity, stimulated with TGF-β for indicated times and harvested for real time rtPCR and Western blot analysis. (A) TGF-β caused SMA mRNA increase only on glass coverslips. No increase was observed in HTFs on polyacrylamide gels. (B) TGF-β–induced SMA expression was significantly stronger on glass coverslips compared with polyacrylamide gels. Moreover, this increase was stronger on stiff gels (0.4% bis-acrylamide gel) compared with soft gels (0.1% bis-acrylamide gel). PA, polyacrylamide.

TGF-β–Induced SMA Incorporation into Stress Fibers Is Reduced on Elastic Substrates

SMA is incorporated into actin stress fibers to enhance cellular force generation. ¹² We examined the influence of environmental elasticity on the TGF-β–driven recruitment of SMA to stress fibers. HTFs were seeded on collagen-coated polyacrylamide gels or glass coverslips and stimulated with TGF-β or vehicle control. Increasing substrate elasticity prevented incorporation of SMA into stress fibers. On glass coverslips, 43% of TGF-β–treated HTFs showed SMA-positive stress fibers, whereas on 0.4% or 0.1% bis-acrylamide polyacrylamide gels, only 15% or 5%, respectively (Fig. 4). On elastic substrates, diffuse cytosolic SMA staining was observed. 

Figure 4.

 

Recruitment of SMA to actin stress fibers is reduced on softer matrices. HTF were plated on collagen-coated coverslips and collagen-coated polyacrylamide gels of different elasticity, stimulated with TGF-β for 3 days. (A) Gray scale view of SMA staining. (B) Approximately 45% of HTFs on rigid glass coverslips showed SMA incorporation into actin stress fibers, whereas only 15% on gels of intermediate elasticity (0.8% bis-acrylamide gel) and only 5% on soft gels (0.2% bis-acrylamide gel). PA, polyacrylamide.

HTFs on Collagen Gels Show Reduced SMA Recruitment

We also studied the effects of thin collagen gels as an alternative to polyacrylamide substrates. Cells plated on elastic thin collagen gels are able to partially invade the substrate. Therefore, this model represents an intermediate combining features of two- and three- dimensional cell culture techniques. In particular, more elaborate cell–matrix interactions are possible than in two dimensional culture, while assessment of stress fibers is easier than in three-dimensional culture. In this system, TGF-β–treated HTFs showed significantly less SMA incorporation into stress fibers than did the cells on rigid glass substrates (Fig. 5). 

Figure 5.

 

Recruitment of SMA to actin stress fibers is reduced on collagen gels. HTF were plated on collagen-coated coverslips and 0.4 m/mL collagen gels and stimulated with TGF-β for 3 days. Images show F-actin (red), SMA (green), and the DAPI-stained nuclei (blue). Whereas the main part of HTFs grown on rigid coverslips showed SMA incorporation into stress fibers, only a few HTFs grown on collagen gel did so. PA, polyacrylamide.

TGF-β–Induced CTGF Expression Depends on Substrate Elasticity

CTGF is believed to be the major cofactor of TGF-β–driven fibrosis. ¹³ TGF-β induces CTGF mRNA, ¹⁴ and CTGF promotes TGF-β–driven fibrosis in vitro and in vivo. ¹⁵ Since TGF-β–mediated fibrosis requires a stiff environment, we sought to explore the matrix-stiffness dependence of CTGF expression. Whereas TGF-β stimulation led to a dramatic increase of CTGF mRNA on stiff plastic culture dishes, this increase was significantly reduced on intermediate 0.8% polyacrylamide gels and even further decreased on soft 0.2% polyacrylamide gels (Fig. 6). 

Figure 6.

 

TGF-β–induced expression of CTGF is reduced on softer matrices. HTF were plated on tissue culture dishes and collagen-coated polyacrylamide gels of different elasticities, stimulated with TGF-β for 24 hours, and harvested for real-time RT-PCR. Real-time PCR was performed for CTGF mRNA and normalized for respective HPRT mRNA values. TGF-β caused a marked increase in CTGF mRNA in HTFs on glass coverslips. The increase was attenuated in HTF on flexible substrate in a dose-dependent manner. PA, polyacrylamide.

Discussion

The link between mechanical strain and scar formation is clinically well established. To avoid keloid formation after surgery, surgeons follow skin tension lines for incisions and suture wounds with the least possible traction. These clinical paradigms have been confirmed in animal models where increasing mechanical tension by splinting of wound edges enhanced myofibroblast transdifferentiation and formation of granulation tissue. ⁴ Tissue elasticity is an associated basic biomechanical feature with a decisive role in wound healing. Scar formation is enhanced in rigid environments. ^16,17 To further elucidate a possible role of extracellular matrix elasticity in conjunctival scarring, we used polyacrylamide gels cross-linked with different quantities of bis-acrylamide as elastic cell culture substrates and rigid regular tissue culture glass or plastic for comparison in vitro. 

In this study, TGF-β–induced myofibroblast transdifferentiation of human tenon fibroblasts, the hallmark of conjunctival scarring, was attenuated on elastic extracellular matrices in vitro. On the molecular level, substrate elasticity modulates TGF-β–induced activation of the p38 signaling pathway and subsequent cytoskeletal rearrangements. SMAD phosphorylation is partially affected, whereas SMAD nuclear translocation is preserved (Fig. 1, 2). Previous studies have shown that p38 activation is necessary for myofibroblast transdifferentiation, and inhibition of p38 signaling was sufficient to block the transdifferentiation process. ⁷  

The molecular details of elasticity-dependent TGF-β signaling await further characterization. It has been shown that myofibroblast transdifferentiation is adhesion-dependent, while SMAD-signaling is largely independent of cell adhesion. ¹⁸ Along these lines, our own previous experiments had revealed that inhibition of cell contractility, which reduces actin stress fibers and diminishes integrin-mediated adhesion signaling, blocks TGF-β–induced p38 activation with little effect on SMAD signaling. ⁸ Matrix elasticity and cell matrix interactions influence activation of various signaling molecules: FAK phosphorylation is reduced on soft surfaces ¹⁹ and several other proteins, such as talin, paxillin, p130Cas, and Src have been characterized as mechanosensitive. ²⁰ Growth factor signaling by the MAPK pathway also requires cell–matrix interaction and is attenuated in a soft environment. ^11,21,22  

Adhesion-dependent receptor localization may have an important role. Integrins govern the localization and internalization of caveolar lipid rafts, ²³ and distinct endocytic pathways have been shown to regulate TGF-β receptor signaling, ²⁴ but data on the subcellular localization of non-SMAD signaling initiation sites are still lacking. 

To elucidate whether the effects of substrate elasticity on TGF-β signaling elicit subsequent functional changes, we studied the expression of SMA, a myofibroblast marker. HTFs plated on rigid tissue culture plastic showed SMA expression after 24 or 48 hours of TGF-β stimulation on the transcriptional (Fig. 3A) and translational levels (Fig. 3B). In contrast, TGF-β failed to enhance SMA mRNA expression when the cells were plated on elastic polyacrylamide gels, and SMA protein expression was significantly reduced in an elasticity-dependent manner. These findings are supported by earlier observations in three-dimensional cell culture models. Arora et al. ⁵ showed that gingival fibroblasts embedded in free-floating and thus highly compliant collagen gels failed to induce SMA protein expression on TGF-β stimulation. When the gel edges were tethered to the wall of the culture dish to decrease gel compliance, TGF-β did induce SMA expression, but to a lesser extent than in regular two-dimensional cell cultures on plastic dishes. They also found this effect to be adhesion-dependent. ⁵ Since we used only two-dimensional substrates, possible spatial confounding effects that may hamper comparisons of two- and three-dimensional systems are ruled out. Our data suggest that substrate elasticity is sufficient to modulate cell functions in HTF. 

Localization of SMA is also critical for the development of a myofibroblast phenotype. We therefore studied the subcellular SMA distribution by confocal laser scanning immunofluorescence microscopy. Although up to 50% of TGF-β–treated fibroblasts on rigid glass coverslips showed incorporation of SMA into actin stress fibers, their ratio decreased to 15% on intermediate and 5% on soft polyacrylamide gels (Fig. 4). These results are in line with a previous study on the influence of substrate elasticity on focal adhesion size in primary rat myofibroblasts. ²⁵ The authors showed that an elastic modulus above 15 kPa is necessary to allow for SMA incorporation into stress fibers in this cell type. Similar to our results, portal fibroblasts needed an elastic module of greater than 3 kPa to incorporate SMA in stress fibers on TGF-β stimulation. ²⁶ Also human trabecular meshwork cells need a certain rigidity to establish SMA-positive stress fibers. ¹¹  

In vivo, ECM elastic modules are reported to range from below 1 kPa in neuronal, 20 to 25 kPa in muscle, to more than 40 kPa in cartilage tissue. ²⁷ Currently, no data are available on the elastic modulus of normal and cicatrized conjunctival tissue. 

To validate our observations, we expanded our experiments using thin collagen gel films as another elastic cell culture substrate. This setup may be more physiological than polyacrylamide gels. It shares some features with three-dimensional systems, since cells can protrude into the collagen films allowing for more intimate cell–matrix interactions. However, the cells are not completely immersed in the collagen gel. HTFs on collagen gel films showed reduced incorporation of SMA in stress fibers compared with HTF on glass coverslips (Fig. 3) similar to the cells plated on polyacrylamide gels. 

Our results, combined with those in other reports, suggest that a more rigid environment facilitates TGF-β–driven myofibroblast transdifferentiation. Interestingly, other aspects of cell–matrix interaction seem also to be important in this process. Integrins, heterodimeric receptors which connect ECM molecules with the intracellular actin cytoskeleton, take a central role. Inhibiting the interaction of integrin receptors with the ECM can hinder myofibroblast transdifferentiation. Competitive blockade of integrin receptors by RGD peptides, which represent the major binding motive for integrin collagen interaction, block TGF-β–induced SMA expression in dermal fibroblasts, ²⁸ and form SMA-positive stress fibers in hepatic stellate cells. ²⁹ Moreover, an intact cellular actin cytoskeleton is important: ROCK inhibitors blocked TGF-β–induced SMA expression and recruitment to stress fibers in HTFs ⁸ and inhibited SMA expression in hyalocytes. ³⁰ Statins, which also interfere with Rho signaling, hampered TGF-β–induced transdifferentiation of HTFs. ⁹ Thus, TGF-β–induced SMA expression and incorporation into stress fibers requires a sufficiently rigid environment and an intact intracellular contractile apparatus. 

Furthermore, a more rigid environment renders the fibroblasts more resistant to apoptosis both due to serum deprivation and loss of adhesion, termed anoikis. Especially on soft polyacrylamide gels, the media had to be supplemented with a certain threshold of FCS to keep the fibroblasts alive (own data not shown). Horowitz et al. described a TGF-β–driven p38-PI3K-Akt-dependent prosurvival pathway in lung fibroblasts. ^31,32 PI3K-Akt signaling may also be elasticity-dependent and may contribute to myofibroblast transdifferentiation in our system. 

TGF-β–induced myofibroblast transdifferentiation is also characterized by a robust increase in the expression of CTGF, an essential cofactor of TGF-β–driven fibrosis. ¹³ Earlier studies indicated that inhibition of cell contractility blocks TGF-β–dependent CTGF expression in HTF. ⁹ We were therefore compelled to examine the effect of substrate elasticity on CTGF expression. In HTF plated on regular plastic cell culture dishes, TGF-β induced a significant increase in CTGF transcription. In cells plated on polyacrylamide gels, TGF-β–induced CTGF transcription was attenuated with increasing substrate compliance (Fig. 6). These findings are in line with our observations on SMA expression. To our knowledge, data on elasticity-dependent CTGF expression are currently lacking. 

Taken together, our data suggest that ECM elasticity may have a decisive role in conjunctival scarring since it modulates TGF-β signaling and subsequent myofibroblast transdifferentiation in human tenon fibroblasts. The p38 signaling pathway appears as a mechanosensitive element in TGF-β–induced myofibroblast transdifferentiation. Biomechanical cues should be considered in our quest for improved strategies to modulate conjunctival wound healing. It appears that alterations in (sub)conjunctival tissue elasticity due to ageing, previous surgery, or precedent long-term inflammation, as observed with topical treatment, could predispose to enhanced scarring. At the same time, sudden changes in tissue prestress (due to edema) may reduce tissue compliance in young patients and contribute to scarring. Both hypotheses would call for a modulation of mechanotransduction (e.g., by ROCK inhibitors) to prevent scar formation. 

Footnotes

 Supported by the Interdisziplinäres Zentrum für Klinische Forschung (IZKF) Würzburg.

Footnotes

 Disclosure: T. Meyer-ter-Vehn, None; H. Han, None; F. Grehn, None; G. Schlunck, None

The authors thank Michael Sendtner and his group at the Institute for Clinical Neurobiology, Würzburg, for generously sharing equipment.