BASIC INVESTIGATION

Inhibitory Effect of Tranilast on Transforming Growth Factor-Beta-Induced in Granular Corneal Dystrophy Type 2 Corneal Fibroblasts

Tae-im Kim, MD, PhD,* Hun Lee, MD,* Hye Kyoung Hong, PhD,* Kyu Seo Kim, BS,† Seung-Il Choi, PhD,* Yong-Sun Maeng, PhD,* and Eung Kweon Kim, MD, PhD*‡

recurrence of corneal opacity in TGFBI-linked corneal dystrophies if Purpose: To investigate the effects of tranilast, an inhibitor of clinical studies confirm these findings. chemical mediators and fibroblast proliferation, on the expression of transforming growth factor-beta (TGF-b)-induced protein (TGFBIp) Key Words: tranilast, TGFBIp, TGF-b, pSmad, a-SMA, corneal in wild-type (WT) and homozygous (HO) granular corneal dystro- dystrophy fi phy type 2 corneal broblasts. (Cornea 2015;34:950–958) Methods: Cell proliferation and cytotoxicity were measured by Cell Counting Kit-8 and lactate dehydrogenase assay. Western blotting and real-time polymerase chain reaction were used to bnormal accumulation of (ECM) determine changes in the expression of TGFBIp and TGFBI mRNA. Aproteins produced by activated or transformed corneal We determined the effects of tranilast on phosphorylated Smad2 fibroblasts is associated with loss of corneal transparency in (pSmad2) and pSmad3, wound-healing, and expression of alpha- corneal dystrophies. The extracellular transforming growth smooth muscle actin (a-SMA), type I , and . factor beta (TGF-b)-induced (TGFBI, formerly known as BIGH3), located on 5q31, is mutated in corneal Results: High concentrations of tranilast decreased proliferation of dystrophies.1–3 The TGFBI protein (TGFBIp) contains 4 tandem corneal fibroblasts but did not cause elevation of lactate dehydroge- repeats of the fasciclin I domain and an N-terminal cysteine-rich nase, except at 1.0 mM tranilast. TGF-b increased the expression of domain (EMI domain) of EMILIN-1, which facilitate protein– TGFBIp and TGFBI mRNA in WT and HO corneal fibroblasts. protein interactions. These domains are important for a variety Cotreatment of corneal fibroblasts with tranilast and TGF-b reduced of cellular responses, including adhesion, migration, pro- the levels of TGFBIp and TGFBI mRNA. In addition, application of liferation, angiogenesis, and wound healing.4,5 tranilast reduced pSmad2 in WT and HO corneal fibroblasts and Granular corneal dystrophy type 2 (GCD2) is an pSmad3 in HO corneal fibroblasts, both of which were increased autosomal dominant disorder caused by an arginine-to- initially by TGF-b. Tranilast delayed wound healing and reduced the histidine substitution at codon 124 (R124H) in TGFBI on expression of a-SMA, type I collagen, and some of integrins in WT chromosome 5q31.6 Age-dependent progressive accumula- and HO corneal fibroblasts. tion of hyaline and amyloid containing TGFBIp in the corneal 3,7 Conclusions: Application of tranilast in WT and HO corneal stroma is a hallmark of GCD2. fibroblasts inhibited the expression of TGFBIp by blocking TGF-b Although phototherapeutic keratectomy (PTK) is the signaling. Thus, tranilast may be useful in delaying or preventing the recommended treatment for GCD2, a major limitation is the recurrence of corneal opacity after PTK.8 Moreover, cotreat- ment of GCD2 with mitomycin C and PTK does not significantly preserve long-term corneal clarity.9 Penetrating Received for publication October 24, 2014; revision received March 3, 2015; keratoplasty can be performed to replace the opaque cornea, accepted March 22, 2015. Published online ahead of print May 28, 2015. but deposits may occur along the suture tract, graft/host From the *The Institute of Vision Research, Corneal Dystrophy Research junction, or in the grafted donor cornea within several years.10 Institute, Yonsei University College of Medicine, Seoul, South Korea; 9 9 †Emory University School of Medicine, Atlanta, GA; and ‡Severance Tranilast (N-[3 ,4 -dimethoxycinnamoyl]-anthranilic Biomedical Science Institute, Brain Korea 21 Plus Project for Medical acid) has been applied for bronchial asthma, allergic rhinitis, Science, Yonsei University College of Medicine, Seoul, South Korea. and atopic dermatitis.11 Tranilast inhibits the release of Supported in part by the National Research Foundation of Korea (NRF) grant chemical mediators from mast cells, synthesis of chemokines funded by the Korea government (MEST No. 2013R1A1A2058907) and Basic Science Research Program through the National Research Foun- in various cell types, and production of cytokines such as 12–14 dation of Korea (NRF) funded by the Ministry of Education (NRF-2009- interleukins and TGF-b1. In addition, tranilast suppresses 351-E00061). the abnormal proliferation of fibroblasts and collagen syn- The authors have no conflicts of interest to disclose. thesis, thereby preventing and reducing the formation of Reprints: Eung Kweon Kim, MD, PhD, Department of Ophthalmology, 15 Yonsei University College of Medicine, 50 Yonsei-ro, Seodaemun-gu, keloids or hypertrophic scars. Seoul 120-752, Korea (e-mail: [email protected]). Molecular mechanisms and cellular roles of tranilast in Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. TGFBIp expression during GCD2 pathogenesis have not

950 | www.corneajrnl.com Cornea  Volume 34, Number 8, August 2015

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea  Volume 34, Number 8, August 2015 Inhibitory Effect of Tranilast on TGFBIp in GCD2

been fully elucidated. In this study, we hypothesize that (5.0 ng/mL, recombinant human TGF-b, R&D Systems, tranilast is a potential therapy for TGFBI-linked corneal Emeryville, CA) for 24 hours. TGFBI mRNA and TGFBIp dystrophy due to the reduced TGF-b-dependent production levels were analyzed by real-time polymerase chain reaction of TGFBIp. Therefore, we investigated the effects of tranilast (PCR) and Western blot analysis. To investigate the inhibition on TGFBIp expression in wild-type (WT) and homozygous of the TGF-b signaling pathway by tranilast, the levels of (HO) cultured human corneal fibroblasts. phosphorylated Smad2 (pSmad2) and pSmad3 were analyzed by Western blot analysis. MATERIALS AND METHODS Isolation and Culture of Corneal Fibroblasts RNA Isolation and Real-Time PCR Primary corneal fibroblasts were prepared from healthy TGFBI and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels were examined by isolating total RNA corneas obtained from the eye bank of Yonsei University from corneal fibroblasts with TRIzol reagent (Invitrogen Life Severance Hospital and from HO patients after keratoplasty. Donor confidentiality was maintained in accordance with the Technologies, Carlsbad, CA). Using the Power SYBR Green RNA-to-CT 1-Step kit, StepOnePlus (Applied Biosystems), Declaration of Helsinki and was approved by the Severance we measured the mRNA expression of GAPDH and Hospital IRB Committee (4-2010-0013). The medical records TGFBI according to the manufacturer’s instructions. The PCR of the control donors from the eye bank of Yonsei University conditions for all were as follows: 48°C for 30 minutes, Severance Hospital did not show any histories of genetic or 95°C for 10 minutes, 40 cycles of 95°C for 15 seconds, and systemic metabolic diseases. GCD2 was diagnosed by DNA 60°C for 1 minute. Primers for TGFBI amplification were as sequencing for TGFBI mutations. Cultures and isolation of 9 9 primary corneal fibroblasts were prepared as previously follows: TGFBI-F, 5 -CACAGTCTTTGCTCCCACAA-3 and TGFBI-R, 59-CTCCGCTAACCAGGATTTCA-39.Primers described.16 Immortalization of cells was performed based for GAPDH amplification were as follows: GAPDH-F, on a protocol previously published for primary HO corneal 9 9 fibroblasts.17 All experiments were performed in triplicate. 5 -ATGGGGAAGGTGAAGGTCG-3 and GAPDH-R, 59-GGGGTCATTGATGGCAACAATA-39. TGFBI mRNA levels were normalized to the levels of GAPDH. Cell Proliferation Assay Cells were incubated with various concentrations of tranilast (Sigma-Aldrich, St Louis, MO) for 24 hours, after Western Blot Analysis which the cell viability was measured with Cell Counting Kit-8 Cells were washed with ice-cold PBS and then scraped (CCK-8; Dojindo Laboratories, Kumamoto, Japan). Cells were with a cell scraper and collected by centrifugation. Cells were cultured in a 96-well plate (BD Falcon, Bradford, MA) at 1 · lysed in a radioimmunoprecipitation assay buffer (RIPA 104 cells per well overnight and treated with various concen- buffer; Biosesang, Inc, Seoul, Korea) containing 5 mM fl trations of tranilast. For the CCK-8 assay, 10 mL of CCK-8 EDTA, 1 mM phenylmethylsulfonyl uoride, 1 mM sodium m m solution was added to each well containing the samples in orthovanadate, 1 g/mL pepstatin, and 10 g/mL leupeptin 100 mL of culture media. The cultures were incubated at 37°C for 20 minutes at 4°C. The crude lysates were centrifuged at 15,000g for 10 minutes at 4°C. Supernatants were collected in 5% CO2 for 1 to 4 hours. The optical density was measured on a plate reader at 450 nm. by centrifugation at 15,000g for 10 minutes at 4°C and were boiled in Laemmli sample buffer (Bio-Rad, Hercules, CA) for 5 minutes. were separated by sodium dodecyl Cytotoxicity Assay sulfate–polyacrylamide gel electrophoresis (SDS-PAGE; After incubating the cells with tranilast for 24 hours, the Bio-Rad) on 10% gels and transferred to polyvinylidene cell cytotoxicity was determined with the CytoTox 96R Non- difluoride membranes (Millipore, Billerica, MA). The mem- Radioactive Cytotoxicity Assay (Promega, Madison, WI) as branes were blocked overnight at 4°C in 5% nonfat dry milk described by the manufacturer. Briefly, 100 mM medium and in a buffer containing 20 mM Tris-HCl, pH 7.5 (Sigma- the supernatant of freeze-thawed cells were mixed in assay Aldrich), 150 mM NaCl (Sigma-Aldrich), and 0.1% Tween- buffer with substrate mix in a 96-well plate (BD Falcon) and 20 (Sigma-Aldrich). Blots were incubated overnight at 4°C incubated at 37°C for 30 minutes. The reaction was with primary antibodies for TGFBIp (1:1000 dilution; R&D terminated by adding 50 mL of stop solution, and the Systems), pSmad2 (1:1000 dilution; Cell Signaling Technol- absorbance at 490 nm was measured on a plate reader. ogy, Beverly, MA), pSmad3 (1:1000 dilution; Cell Signaling Technology), type I collagen (1:1000 dilution; Abcam, Cambridge, MA), a-SMA (1:500 dilution; Sigma-Aldrich), Cotreatment of Corneal Fibroblasts With or b-actin (1:1000 dilution; Santa Cruz Biotechnology, Santa Tranilast and TGF-b Cruz, CA). Subsequent incubation with secondary antibodies Tranilast was freshly dissolved in phosphate-buffered conjugated to horseradish peroxidase was performed at room saline (PBS; Gibco-BRL, Grand Island, NY) at room temperature for 1 hour. Horseradish peroxidase-linked anti- temperature before use. To analyze the effects of tranilast mouse IgG (1:5000 dilution; Amersham Pharmacia Biotech, on TGFBI mRNA and TGFBIp expression, cells were Piscataway, NJ) or anti-rabbit IgG (1:5000 dilution; Amersham cotreated with tranilast (0, 0.2, 0.5, or 1.0 mM) and TGF-b Pharmacia Biotech) was used as a secondary antibody. b-actin

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. www.corneajrnl.com | 951

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Kim et al Cornea  Volume 34, Number 8, August 2015

served as a loading control. Immunoreactive bands were (0.5 or 1.0 mM) and TGF-b (5.0 ng/mL) for 24 hours. Cells visualized with an enhanced chemiluminescence Western were rinsed with ice-cold PBS and fixed with 50% acetone/ blotting kit (Amersham Pharmacia Biotech) according to the methanol in PBS for 10 minutes. After washing with PBS, manufacturer’sinstructions. cells were incubated at room temperature with 10 mg/mL mouse monoclonal anti-a-SMA antibody (Sigma-Aldrich) for 1 hour. After washing twice with PBS, cells were Wound-Healing Assay incubated with anti-mouse IgG tagged with Alexa Fluor Corneal fibroblasts were cultured in supplemented 594 (10 mg/mL; Molecular Probes, Eugene, OR). Additional media for 2 days, and monolayers in culture plates were counterstaining was performed with DAPI (10 mg/mL; wounded with a scraper. Wounded monolayers were washed Sigma-Aldrich) to identify cell nuclei. Finally, the cover several times with media to remove cell debris. Injured cells glasses were mounted on glass slides and imaged with were incubated with tranilast (0, 0.5, or 1.0 mM) and TGF-b confocal microscopy (Leica Microsystems, Mannheim, (5.0 ng/mL) in cell culture media at 37°C in 5% CO2 for 24 Germany). hours and allowed to migrate. Cells were photographed with a microscope 24 hours after wounding. To determine the distance between the leading edges of growing cells, the Measurement of Expression Levels pixels were counted on digitized photographs with ImageJ To identify cell surface integrin, we used the Alpha software, version 1.37 (National Institutes of Health, Bethes- Integrin-Mediated Cell Adhesion Array Kit and Beta Integrin- da, MD). Mediated Cell Adhesion Array Kit (Chemicon, Temecula, CA), in which 96-well plates were coated with mouse monoclonal antibodies against human adhesion molecules. Immunofluorescent Confocal Microscopy The plates were used to capture cells that express specific To observe a-SMA expression, WT or HO cells were surface adhesion molecules. Corneal fibroblasts were cotreated seeded on 22- · 22-mm uncoated cover glasses placed in the with tranilast (0, 0.5, or 1.0 mM) and TGF-b (5.0 ng/mL) for wells of a 6-well culture plate for 24 hours to allow efficient 24 hours and prepared as single-cell suspensions in non- attachment. Cells were treated with TGF-b alone or tranilast enzymatic dissociation buffer consisting of PBS with 2 to 5

FIGURE 1. Proliferation and cytotoxicity of WT and HO GCD2 corneal fibroblasts after application of tranilast. A, CCK-8 assay. B, Lactate dehydrogenase assay. Error bars represent SEM (***P , 0.001, *P , 0.05).

952 | www.corneajrnl.com Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea  Volume 34, Number 8, August 2015 Inhibitory Effect of Tranilast on TGFBIp in GCD2

mM EDTA. After adjusting the cell density to 1 · 105 to 1 · San Diego, CA). Differences were considered statistically 107 cells per milliliter with assay buffer, the mixture was significant for P , 0.05. incubated at 37°C in 5% CO2 for 45 minutes. After several washes with PBS, cell-staining solution (100 mL/well) was added and incubated for 5 minutes. After several washes with PBS, extraction buffer (100 mL/well) was added. The suspen- RESULTS sion was rotated on an orbital shaker for 5 to 10 minutes. Reaction intensities were measured at 570 nm. Effects of Tranilast on Cell Proliferation and Cytotoxicity The CCK-8 assay showed a dose-dependent decrease in Statistical Analysis the number of viable WT and HO corneal fibroblasts at all Data are presented as mean 6 SEM. The data from all concentrations except 0.01 mM (Fig. 1A). Tranilast (1.0 mM) experiments were analyzed by 1-way analysis of variance, increased lactate dehydrogenase levels in WT corneal fibro- and the statistical significance was determined with a Bonfer- blasts compared with those in untreated fibroblasts, whereas roni post hoc comparison. Statistical analyses were performed no significant cellular damage in HO corneal fibroblasts was with GraphPad PRISM (version 4; GraphPad Software, Inc, reported (Fig. 1B).

FIGURE 2. Changes in the expression of TGFBIp and TGFBI mRNA after application of tranilast in WT and HO GCD2 corneal fibroblasts cotreated with TGF-b. A, Western blot analysis. B, Real-time PCR. b-actin was used as an internal control. Error bars represent SEM (***P , 0.001, **P , 0.01, *P , 0.05). TGFBI, transforming growth factor beta-induced gene; TGFBIp, trans- forming growth factor beta-induced protein.

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. www.corneajrnl.com | 953

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Kim et al Cornea  Volume 34, Number 8, August 2015

Effects of Tranilast on the Expression of corneal fibroblasts, which were significantly decreased in TGFBIp and TGFBI mRNA response to .0.5 and 1.0 mM tranilast, respectively (Fig. 3). The TGF-b-treated cells showed significantly increased expression levels of TGFBIp and TGFBI mRNA in WT and HO corneal fibroblasts (Figs. 2A, B). Cotreat- Effects of Tranilast on Wound Healing of ment of WT and HO corneal fibroblasts with tranilast and Corneal Fibroblasts TGF-b reduced the levels of TGFBIp in a dose-dependent WT and HO corneal fibroblasts that had been treated manner compared with the cells treated with TGF-b alone with TGF-b alone showed increased proliferation and (Fig. 2A). Tranilast also decreased TGFBI mRNA expres- migration, resulting in complete closure of the wounded area sion in a dose-dependent manner in WT and HO corneal compared with that in untreated fibroblasts (Fig. 4). Cotreat- fibroblasts (Fig. 2B). ment of WT and HO corneal fibroblasts with tranilast and TGF-b demonstrated that closure of the wounded area was increasingly delayed as the concentration of tranilast increased in WT and HO corneal fibroblasts (Fig. 4). Inhibitory Effects of Tranilast on TGF-b-Induced Smad2/3 Phosphorylations In WT corneal fibroblasts, pSmad2 significantly Effects of Tranilast on the Expression increased in response to TGF-b compared with that in of a-SMA untreated fibroblasts. Cotreatment of WT corneal fibroblasts In WT and HO corneal fibroblasts, TGF-b treatment with tranilast (0.5 and 1.0 mM) and TGF-b significantly increased the expression of a-SMA, which were significantly reduced the levels of pSmad2 compared with that in cells decreased in response to .0.5 mM tranilast, respectively treated with TGF-b alone (Fig. 3). TGF-b treatment signif- (Fig. 5A). WT and HO corneal fibroblasts treated with TGF-b icantly increased the levels of pSmad2 and pSmad3 in HO alone showed increased numbers of a-SMA-positive cells in

FIGURE 3. Changes in the expression of pSmad2 and pSmad3 after application of tranilast in WT and HO GCD2 corneal fibroblasts cotreated with TGF-b. b-actin was used as an internal control. Error bars represent SEM (***P , 0.001, **P , 0.01, *P , 0.05). pSmad, phosphorylated Smad.

954 | www.corneajrnl.com Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea  Volume 34, Number 8, August 2015 Inhibitory Effect of Tranilast on TGFBIp in GCD2

FIGURE 4. Wound-healing assay of WT and HO GCD2 corneal fibroblasts. Injured cells were incubated with tranilast and TGF-b in cell culture media for 24 hours. Error bars represent SEM (***P , 0.001, **P , 0.01). All micrographs are of the same magni- fication.

immunofluorescence confocal microscopy. However, the tranilast (#0.5 mM) without cytotoxic effects. The tranilast- numbers of TGF-b-induced a-SMA-positive cells were mediated reduction in the proliferation of WT and HO corneal reduced by cotreatment of WT and HO corneal fibroblasts fibroblasts was consistent with the reduced expression of with tranilast (Fig. 5B). TGFBIp and a-SMA, as well as delayed migration in wound- healing assay. Notably, we also demonstrated that decreased levels of TGFBIp may be related to the inhibition of TGF-b Effects of Tranilast on the Expression of Type I signaling, which was confirmed by decreases in pSmad2 and Collagen and Integrins pSmad3 levels upon treatment with tranilast. In WT and HO corneal fibroblasts, TGF-b induced the In this study, we demonstrated that TGFBIp expres- expression of type I collagen, and increased expression was sion can be induced by TGF-b and tranilast inhibits the reduced by tranilast (Fig. 6). TGF-b significantly induced the TGF-b-induced expression of TGFBIp in WT and HO expression of a1, a2, a3, a4, a5, av, avb3, b1, avb5, and corneal fibroblasts. It was reported that the loss of TGFBI a5b1 in WT corneal fibroblasts. Similarly, TGF-b induced expression is a frequent event in human cancer and can be the expression of a2, a3, a4, a5, av, avb3, b1, b3, b6, the cause of acquisition of the tumorigenic phenotype in avb5, and a5b1 in HO corneal fibroblasts (Fig. 7). Cotreat- asbestos-treated human papillomavirus-immortalized human ment of WT corneal fibroblasts with tranilast and TGF-b bronchial epithelial cells.18 Although the inhibitory effect of significantly reduced the expression of a1, a2, b1, avb5, and tranilast on the TGF-b-induced cellular change in WT and a5b1 compared with that in cells treated with TGF-b alone. HO corneal fibroblasts was slightly different, tranilast might Cotreatment of HO corneal fibroblasts with tranilast and be of use as a potential cotreatment with surgical ap- TGF-b reduced the expression of a2, a3, a4, a5, av, avb3, proaches, such as PTK, to delay or prevent the recurrence b1, b3, b6, avb5, and a5b1 (Fig. 7). of corneal opacity in TGFBI-linked corneal dystrophies if there is no tumorigenic effect of the topical application of tranilast in vivo. DISCUSSION Injection of tranilast inhibited the progression of pro- In this study, we discovered a decrease in the expres- liferative vitreoretinopathy in rabbits by reducing the vitreous sion of TGFBIp and TGFBI mRNA after treatment with levels of TGF-b.19 In one study that evaluated the effects of

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. www.corneajrnl.com | 955

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Kim et al Cornea  Volume 34, Number 8, August 2015

FIGURE 5. Changes in the expression of a-SMA after application of tranilast in WT and HO GCD2 corneal fibroblasts cotreated with TGF-b. A, Western blot analysis. B, Immunofluorescence confocal microscopy analysis. a-SMA: alpha-smooth muscle actin.

tranilast on corneal haze after photorefractive keratectomy, Smad2 and Smad3, which then form a heterotrimeric complex tranilast reduced corneal haze by suppressing TGF-b expres- with Smad4. The Smad complex translocates to the nucleus sion, as demonstrated by immunohistochemistry and Western and regulates the expression of target genes.21,22 Therefore, blot analysis.20 These results suggest that tranilast could affect the decrease in TGFBIp expression induced by tranilast is in TGF-b signaling pathways, which have been implicated as line with the observed reductions in pSmad2 and pSmad3. a regulator of a variety of cellular and physiological Although we did not identify the specific mechanisms processes.21 involved in the tranilast-induced decreases in the phosphor- TGF-b signaling is activated by the binding of TGF-b ylation of Smad2 and Smad3, our data suggest that TGFBIp to type I and type II TGF-b receptors. The activated receptor reduction is mediated by the inhibition of TGF-b signaling complex phosphorylates the downstream transcription factors through Smad2/3.

FIGURE 6. Changes in the expression of type I collagen after application of tranilast in WT and HO GCD2 corneal fibroblasts cotreated with TGF-b. Error bars represent SEM (***P , 0.001, *P , 0.05).

956 | www.corneajrnl.com Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea  Volume 34, Number 8, August 2015 Inhibitory Effect of Tranilast on TGFBIp in GCD2

FIGURE 7. Changes in the expression of integrins after application of tranilast in WT and HO GCD2 corneal fibroblasts cotreated with TGF-b. Error bars represent SEM (***P , 0.001, **P , 0.01, *P , 0.05).

In this study, cotreatment of WT and HO corneal reactions of the corneal stroma by inhibiting TGF-b-induced fibroblasts with tranilast and TGF-b delayed closure of the corneal myofibroblast transition. wounded area. A previous study showed that rapidly pro- The increased expression of type I collagen in response liferating cells, such as pterygial fibroblasts, were more to TGF-b stimulation was reduced after treatment with susceptible to the inhibitory effects of tranilast compared tranilast. The ECM of the corneal stroma primarily consists with cells in the G0 phases of the cell cycle.23 Tranilast may of type I collagen, and interactions with type I and type V be beneficial immediately after PTK for GCD2 because cell regulate the diameter of the collagen fibril, which is proliferation occurs more actively at the injured lesion. important for corneal transparency.28,29 Tranilast does not We demonstrated that TGF-b increased the a-SMA inhibit collagen synthesis in human pterygial fibroblasts and expression and tranilast inhibited the TGF-b-induced normal human skin fibroblasts.23,30 Rather, tranilast exerts a-SMA expression in WT and HO corneal fibroblasts. inhibitory effects on TGF-b-induced collagen synthesis in Among the TGF-b/Smad-signaling molecules, a-SMA is keloid fibroblasts, indicating that tranilast inhibits enhanced a recognized biochemical marker for the transition of stimulation of collagen synthesis.31 Likewise, in our study, fibroblasts to myofibroblasts.24 TGF-b-induced a-SMA tranilast exerted inhibitory effects on TGF-b-induced type I expression and myofibroblast transformation have been collagen synthesis in WT and HO corneal fibroblasts. shown to initiate ECM remodeling and wound healing in Most of the a and b integrins were induced by TGF-b, the corneal stroma.24–26 Thus, antibodies against TGF-b can and some of the integrins were reduced by tranilast. TGFBIp reduce corneal fibrosis and fibronectin deposition in vivo.27 mediates adhesion and migration of several cell types through Accordingly, we expect, based on our results, that tranilast interactions with integrins, which transduce intracellular may be of use in controlling excessive wound-healing signals that modulate adhesion, migration, proliferation, and

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. www.corneajrnl.com | 957

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Kim et al Cornea  Volume 34, Number 8, August 2015

survival.32 We previously suggested that alterations in the 12. Hida RY, Takano Y, Okada N, et al. Suppressive effects of tranilast on b eotaxin-1 production from cultured conjunctival fibroblasts. Curr Eye TGF- receptor- and integrin-mediated signaling pathway are – involved in GCD2 pathophysiology by using cDNA micro- Res. 2008;33:19 22. 33 13. Chikaraishi A, Hirahashi J, Takase O, et al. Tranilast inhibits interleukin- arrays and cell adhesion assay. Considering the inhibitory 1beta-induced monocyte chemoattractant protein-1 expression in rat effects of tranilast on integrins and TGFBIp, we conclude that mesangial cells. Eur J Pharmacol. 2001;427:151–158. tranilast can inhibit adhesion of corneal fibroblasts and other 14. Komatsu H, Kojima M, Tsutsumi N, et al. Study of the mechanism of integrin-related events. inhibitory action of tranilast on chemical mediator release. Jpn J Pharmacol. 1988;46:43–51. Recently, tranilast ophthalmic solution containing ben- 15. Yamada H, Tajima S, Nishikawa T, et al. Tranilast, a selective inhibitor zalkonium chloride (KRIX; Choongwae Pharmaceutical Co, of collagen synthesis in human skin fibroblasts. J Biochem. 1994;116: South Korea) is on the market for the treatment of allergic 892–897. conjunctivitis. Kato et al34 showed that benzalkonium chlo- 16. Choi SI, Kim TI, Kim KS, et al. Decreased catalase expression and fi increased susceptibility to oxidative stress in primary cultured corneal ride solution accelerates the formation of the amyloid brils fibroblasts from patients with granular corneal dystrophy type II. Am J of corneal dystrophy-associated peptides in vitro. Pathol. 2009;175:248–261. In conclusion, we provided evidence that treatment with 17. Choi SI, Kim KS, Oh JY, et al. Melatonin induces autophagy via an tranilast decreased the expression of TGFBIp and TGFBI mTOR-dependent pathway and enhances clearance of mutant-TGFBIp. – mRNA. Most notably, we found that a tranilast-induced J Pineal Res. 2013;54:361 372. a 18. Zhao YL, Piao CQ, Hei TK. Downregulation of Betaig-h3 gene is reduction in TGFBIp, accompanying a decrease in -SMA causally linked to tumorigenic phenotype in asbestos treated immortal- expression and integrins, could be due to the inhibition of ized human bronchial epithelial cells. Oncogene. 2002;21:7471–7477. TGF-b signaling through the inhibition of Smad2/3 phos- 19. Ito S, Sakamoto T, Tahara Y, et al. The effect of tranilast on experimental phorylation. These results support the need for further studies proliferative vitreoretinopathy. Graefes Arch Clin Exp Ophthalmol. 1999;237:691–696. of tranilast as a novel treatment for preventing recurrence of 20. Song JS, Jung HR, Kim HM. Effects of topical tranilast on corneal haze GCD2 after surgical procedures. after photorefractive keratectomy. J Cataract Refract Surg. 2005;31: 1065–1073. 21. Shi Y, Massague J. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell. 2003;113:685–700. 22. Derynck R, Zhang YE. Smad-dependent and Smad-independent path- ACKNOWLEDGMENTS ways in TGF-beta family signalling. Nature. 2003;425:577–584. The authors thank Hye Jeong Hwang for her technical 23. Isaji M, Kikuchi S, Miyata H, et al. Inhibitory effects of tranilast on the support in performing the experiments. proliferation and functions of human pterygium-derived fibroblasts. Cornea. 2000;19:364–368. 24. Huet E, Vallee B, Szul D, et al. Extracellular matrix metalloproteinase REFERENCES inducer/CD147 promotes myofibroblast differentiation by inducing 1. Munier FL, Korvatska E, Djemai A, et al. Kerato-epithelin mutations in alpha-smooth muscle actin expression and collagen gel contraction: four 5q31-linked corneal dystrophies. Nat Genet. 1997;15:247–251. implications in tissue remodeling. FASEB J. 2008;22:1144–1154. 2. Kannabiran C, Klintworth GK. TGFBI gene mutations in corneal 25. Serini G, Gabbiani G. Mechanisms of myofibroblast activity and dystrophies. Hum Mutat. 2006;27:615–625. phenotypic modulation. Exp Cell Res. 1999;250:273–283. 3. Klintworth GK. Advances in the molecular genetics of corneal dystro- 26. Jester JV, Barry-Lane PA, Cavanagh HD, et al. Induction of alpha- phies. Am J Ophthalmol. 1999;128:747–754. smooth muscle actin expression and myofibroblast transformation in 4. Nam JO, Jeong HW, Lee BH, et al. Regulation of tumor angiogenesis by cultured corneal keratocytes. Cornea. 1996;15:505–516. fastatin, the fourth FAS1 domain of betaig-h3, via alphavbeta3 integrin. 27. Jester JV, Barry-Lane PA, Petroll WM, et al. Inhibition of corneal Cancer Res. 2005;65:4153–4161. fibrosis by topical application of blocking antibodies to TGF beta in the 5. Park SW, Bae JS, Kim KS, et al. Beta ig-h3 promotes renal proximal rabbit. Cornea. 1997;16:177–187. tubular epithelial cell adhesion, migration and proliferation through the 28. Marshall GE, Konstas AG, Lee WR. Immunogold fine structural interaction with alpha3beta1 integrin. Exp Mol Med. 2004;36:211–219. localization of extracellular matrix components in aged human cornea. 6. Kocak-Altintas AG, Kocak-Midillioglu I, Akarsu AN, et al. BIGH3 gene II. Collagen types V and VI. Graefes Arch Clin Exp Ophthalmol. 1991; analysis in the differential diagnosis of corneal dystrophies. Cornea. 229:164–171. 2001;20:64–68. 29. Birk DE, Fitch JM, Babiarz JP, et al. Collagen fibrillogenesis in vitro: 7. Korvatska E, Henry H, Mashima Y, et al. Amyloid and non-amyloid interaction of types I and V collagen regulates fibril diameter. J Cell Sci. forms of 5q31-linked corneal dystrophy resulting from kerato-epithelin 1990;95:649–657. mutations at Arg-124 are associated with abnormal turnover of the 30. Isaji M, Nakajoh M, Naito J. Selective inhibition of collagen accumu- protein. J Biol Chem. 2000;275:11465–11469. lation by N-(3,4-dimethoxycinnamoyl)anthranilic acid (N-5’) in granu- 8. Dogru M, Katakami C, Nishida T, et al. Alteration of the ocular surface lation tissue. Biochem Pharmacol. 1987;36:469–474. with recurrence of granular/avellino corneal dystrophy after photo- 31. Suzawa H, Kikuchi S, Arai N, et al. The mechanism involved in the therapeutic keratectomy: report of five cases and literature review. inhibitory action of tranilast on collagen biosynthesis of keloid Ophthalmology. 2001;108:810–817. fibroblasts. Jpn J Pharmacol. 1992;60:91–96. 9. Ha BJ, Kim TI, Choi SI, et al. Mitomycin C does not inhibit exacerbation 32. Kim JE, Kim SJ, Lee BH, et al. Identification of motifs for cell adhesion of granular corneal dystrophy type II induced by refractive surface within the repeated domains of transforming growth factor-beta-induced ablation. Cornea. 2010;29:490–496. gene, betaig-h3. J Biol Chem. 2000;275:30907–30915. 10. Holland EJ, Daya SM, Stone EM, et al. Avellino corneal dystrophy. 33. Choi SI, Yoo YM, Kim BY, et al. Involvement of TGF-{beta} receptor- Clinical manifestations and natural history. Ophthalmology. 1992;99: and integrin-mediated signaling pathways in the pathogenesis of granular 1564–1568. corneal dystrophy II. Invest Ophthalmol Vis Sci. 2010;51:1832–1847. 11. Azuma H, Banno K, Yoshimura T. Pharmacological properties of 34. Kato Y, Yagi H, Kaji Y, et al. Benzalkonium chloride accelerates the N-(3’,4’-dimethoxycinnamoyl) anthranilic acid (N-5’), a new anti- formation of the amyloid fibrils of corneal dystrophy-associated peptides. atopic agent. Br J Pharmacol. 1976;58:483–488. J Biol Chem. 2013;288:25109–25118.

958 | www.corneajrnl.com Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.