Molecular Vision 2015; 21:1368-1377
Received 31 August 2015 | Accepted 30 December 2015 | Published 31 December 2015
1Department of Ophthalmology, Yamaguchi University Graduate School of Medicine, Ube City, Yamaguchi, Japan; 2Department of Ophthalmology, First Hospital of Jilin University, Jilin, PR China; 3Yamaguchi University School of Medicine, Ube City, Yamaguchi, Japan
Correspondence to: Kazuhiro Kimura, Department of Ophthalmology, Yamaguchi University Graduate School of Medicine, 1-1-1 Minami-Kogushi, Ube City, Yamaguchi 755-8505, Japan; Phone: +81-836-22-2278; FAX: +81-836-22-2334. email: firstname.lastname@example.org
Purpose: Scar formation is most frequently responsible for the failure of glaucoma filtration surgery. Retinoic acids are vitamin A derivatives that play diverse roles in development, immunity, and tissue repair. The effects of the retinoic acid receptor (RAR) γ agonist R667 on the contractility of human Tenon fibroblasts (HTFs) cultured in a three-dimensional collagen gel as well as on intraocular pressure (IOP) in a rat model of glaucoma filtration surgery were investigated.
Methods: HTFs were cultured in a type I collagen gel, the contraction of which was evaluated by measurement of the gel diameter. The release of matrix metalloproteinases (MMPs) into culture supernatants was assessed with immunoblot analysis and gelatin zymography. Phosphorylation of focal adhesion kinase (FAK) was examined with immunoblot analysis, and production of fibronectin and type I collagen was measured with immunoassays.
Results: R667 inhibited transforming growth factor-β1 (TGF-β1)-induced collagen gel contraction mediated by HTFs in a concentration- and time-dependent manner, whereas an RARα agonist inhibited this process to a lesser extent and an RARβ agonist had no effect. TGF-β1-induced MMP-1 and MMP-3 release, FAK phosphorylation, and fibronectin and type I collagen production in HTFs were also attenuated by R667. Furthermore, R667 lowered IOP in rats after glaucoma filtration surgery.
Conclusions: R667 inhibited TGF-β1-induced contraction and extracellular matrix synthesis in HTFs. Such effects might have contributed to the lowering of IOP by R667 in a rat model of glaucoma filtration surgery. RARγ agonists might thus prove effective for inhibition of scar formation after such surgery.
Glaucoma is the major cause of irreversible blindness worldwide. Filtration surgery remains the gold standard for the treatment of glaucoma in individuals whose intraocular pressure (IOP) is not well controlled by medication or laser surgery. The long-term success of filtration surgery depends on the postoperative wound healing response at the subconjunctival filtering bleb site [1,2]. Scar formation due to excessive synthesis of new extracellular matrix (ECM) and contraction of subconjunctival tissue can obstruct the aqueous flow and thus result in an increase in IOP and surgery failure . Antimitotic agents such as mitomycin C and 5-fluorouracil are administered to reduce the extent of postoperative scarring and improve the outcome of filtration surgery, but the severe side effects of these agents limit their application . Alternative strategies that are safe and modulate the wound healing response more effectively to prevent scar formation are thus needed.
The wound healing response at the subconjunctival filtering bleb site is mediated in part by matrix metalloproteinases (MMPs) and is regulated by various molecules, including growth factors and inflammatory mediators. The cytokine transforming growth factor (TGF)-β is a pivotal regulator of wound healing and fibrosis and a major driving force of conjunctival scarring [5,6]. TGF-β1 and TGF-β2 have been detected at subconjunctival wounds after filtering surgery ; these factors, thus, have become major targets of evolving antifibrotic strategies. Antisense oligonucleotides that target TGF-β for downregulation were found to prolong bleb survival in vivo and to improve surgical outcome . Neutralizing antibodies to TGF-β2 were also shown to reduce scarring in animal models of glaucoma  as well as to maintain good bleb morphology in a pilot clinical study , but the antibodies did not have a significant effect on bleb survival in a randomized clinical trial .
Retinoic acids are derivatives of vitamin A that play complex roles in eye development and physiology . They also possess anti-inflammatory properties and antifibrotic potential as a result of their attenuation of TGF-β actions . Such effects of retinoic acids are mediated by nuclear receptors—α, β, and γ isoforms of retinoic acid receptors (RARs)—that function as ligand-inducible transcriptional regulators and are expressed in most parts of the eye . We recently showed that the RARγ agonist R667 attenuated the epithelial-mesenchymal transition (EMT) in RPE cells as well as subretinal fibrosis . In the present study, we examine the effects of R667 on collagen gel contraction mediated by human Tenon’s capsule fibroblasts (HTFs) as well as on fibronectin and collagen production by these cells exposed to TGF-β1. We also assess the effect of R667 on IOP in a rat model of glaucoma filtration surgery.
Fetal bovine serum, Eagle’s minimum essential medium (MEM), and 10X MEM were obtained from Invitrogen-Gibco (Rockville, MD). Reconstitution buffer and acid-solubilized native porcine type I collagen were from Nitta Gelatin (Osaka, Japan). Antibodies to phospho-FAK (phospho-Tyr576/577) and to FAK were obtained from Cell Signaling (Beverly, MA). The RARα agonist Am580, the RARβ agonist BmS453, collagenase, and antibodies to β-actin and to alpha-smooth muscle actin (α-SMA) were from Sigma-Aldrich (St. Louis, MO), and the RARγ agonist R667 (palovarotene) was from NARD Institute (Hyogo, Japan). Recombinant human TGF-β1 and antibodies to MMPs and to tissue inhibitors of metalloproteinases (TIMPs) were obtained from R&D Systems (Minneapolis, MN). Enzyme immunoassay (EIA) kits for fibronectin and the C peptide of procollagen type I were from Takara Shuzo (Shiga, Japan). Viscoelastic material (Healon) was obtained from Abbott Medical Optics (Santa Ana, CA), 30G needles were from NIPRO (Osaka, Japan), and silicon microtubes with a diameter of 0.3 mm were from ARAM (Osaka, Japan).
This study adhered to the tenets of the ARVO statement on human subjects. Human tissue was used in strict accordance with the basic principles of the Declaration of Helsinki and with approval of the human ethics committee of the Yamaguchi University Graduate School of Medicine. Tenon’s capsule fibroblasts were isolated with informed consent from subconjunctival tissue of three individuals undergoing eyelid or strabismus surgery. The subconjunctival tissue was digested with collagenase (2 mg/ml) for 1 h at 37 °C to provide a suspension of Tenon fibroblasts. The cells derived from each patient were maintained separately at 37 °C under a humidified atmosphere containing 5% CO2 and in MEM supplemented with 10% fetal bovine serum. Cells were used for experiments before the sixth passage.
Free-floating collagen gels containing relaxed HTFs were prepared as previously described . Briefly, type I collagen, reconstitution buffer, 10X MEM, and HTF suspension were mixed on ice. Portions (0.5 ml) of the mixture containing HTFs at a density of 2 × 105/ml and collagen at 1.9 mg/ml were transferred to the wells of a 24-well culture plate that had been coated with 1% bovine serum albumin (BSA). After gelation of the mixture, the collagen gels were detached from the wells of the culture plate, and serum-free MEM (0.5 ml) containing vehicle (4 mM HCl containing 0.1% BSA), TGF-β1, or RAR agonists was then added on top of each gel. The diameter of the gels was determined daily with a ruler for calculation of the extent of gel contraction.
Immunoblot analysis was performed as described . For immunoblot analysis of FAK and α-SMA, HTFs incubated in collagen gels were lysed in an ice-cold lysis buffer, and portions (10 µg of protein) of the lysates were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) on a 10% gel. Immunoblot analysis of MMP-1, MMP-3, TIMP-1, and TIMP-2 was performed as described , with culture supernatants subjected to SDS–PAGE on a 10% gel. After electrophoresis, the separated proteins were transferred electrophoretically to a nitrocellulose membrane. The membrane was exposed to 5% dried skim milk in Tris-buffered saline containing 0.05% Tween-20 (TBST) for 1 h at room temperature before incubation overnight at 4 °C with primary antibodies (each at a 1:1,000 dilution). The membrane was then washed extensively with TBST before incubation for 1 h at room temperature with horseradish peroxidase–conjugated secondary antibodies and detection of immune complexes with enhanced chemiluminescence reagents.
Gelatin zymography of culture supernatants was performed as described previously . Briefly, 8 µl of culture supernatants mixed with 4 µl of non-reducing SDS sample buffer were subjected to SDS–PAGE at 4 °C on a 10% gel containing 0.1% gelatin. The gel was then incubated first for 1 h at room temperature in 2.5% Triton X-100 and then for 18 h at 37 °C in a reaction buffer containing 50 mM Tris-HCl (pH 7.5), 1% Triton X-100, and 5 mM CaCl2. The gel was then stained with 0.5% Coomassie brilliant blue.
Fibronectin and the C peptide of procollagen type I, which reflects the amount of newly synthesized type I collagen, were measured with EIAs . Conditioned medium from collagen gel incubations was collected and stored at –80 °C until analysis. Standards were included with each assay to ensure accuracy.
Animal experiments were approved by the animal ethics committee of the Yamaguchi University Graduate School of Medicine. Surgery was performed on the left eye of 6-week-old male Wistar rats (five animals per treatment group) by a clinical glaucoma specialist according to a modified version of a previously described procedure . Briefly, a fornix-based conjunctival flap was created at a position 3 to 4 mm behind the limbus with a conjunctival incision and dissection of the underlying Tenon’s capsule from the sclera. No antimetabolite was applied. A full-thickness scleral tunnel into the intraocular space (either anterior or posterior chamber) was generated by insertion of a 30G needle with care so as to avoid iridal hemorrhage. Viscoelastic material was injected through the needle to maintain the anterior chamber, and the needle was then withdrawn. A silicon microtube with a bevel at the distal end and a slit at the proximal end was inserted through the scleral tunnel. R667 (50 μg/ml) or PBS (1X; 155 mM NaCl, 1 mM KH2PO4, 3 mM Na2HPO4-7H2O, pH 7.4; 0.2 μl) was injected into Tenon’s capsule, and the conjunctival flap was closed with a 10–0 nylon suture. IOP was measured with a portable applanation tonometer (Tono-Pen AVIA Vet; Reichert Technologies, Depew, NY) before and for 2 weeks after surgery with the animals under inhalation anesthesia with diethyl ether.
Data are presented as means ± standard deviation (SD). All in vitro experiments were performed in triplicate and repeated three times with three different cell isolates, and the data were analyzed with the Tukey-Kramer test. Nonparametric statistical analysis of IOP was performed with the Holm pairwise t test and the Mann–Whitney U test. A p value of <0.05 was considered statistically significant.
RARα, RARβ, and RARγ have been found to be expressed at the human ocular surface . We first examined whether RAR agonists affect TGF-β1-induced collagen gel contraction mediated by HTFs. The cells were incubated with 1 μM Am580 (RARα agonist), BmS453 (RARβ agonist), or R667 (RARγ agonist) for 12 h before exposure to TGF-β1 (1 ng/ml) for 3 days. The stimulatory effect of TGF-β1 on collagen gel contraction was inhibited by 36%, 12%, or 73% in the presence of Am580, BmS453, or R667, respectively (Figure 1). None of the RAR agonists affected HTF-mediated collagen gel contraction in the absence of TGF-β1.
We further characterized the inhibitory effect of R667 on TGF-β1-induced collagen gel contraction mediated by HTFs. Incubation of the cells for 3 days with various concentrations of R667 (0.01 to 1 µM) revealed that the drug attenuated TGF-β1-induced gel contraction in a concentration-dependent manner (Figure 2A). The stimulatory effect of TGF-β1 on collagen gel contraction was thus inhibited by 41%, 57%, or 66% by R667 at concentrations of 0.01, 0.1, or 1 µM, respectively. The inhibitory effect of R667 (1 µM) on TGF-β1-induced gel contraction was also time dependent (Figure 2B), becoming statistically significant at day 2. Vehicle (4 mM HCl containing 0.1% BSA) for TGF-β1 had no effect on collagen gel contraction mediated by HTFs (Figure 2B).
MMPs play an important role in cell-mediated collagen contraction . Therefore, we next examined the effects of R667 on the release of MMPs and TIMPs by HTFs with the use of immunoblot analysis and gelatin zymography. HTFs were cultured in collagen gels for 3 days in the presence of TGF-β1 (1 ng/ml) and various concentrations of R667. Analysis with antibodies to MMP-1 revealed that the culture supernatants of cells maintained in the absence of TGF-β1 and R667 contained small amounts of 49- and 45-kDa immunoreactive proteins corresponding to active forms of MMP-1 (Figure 3A). Culture of the cells in the presence of TGF-β1 resulted in an increase in the intensity of these bands. The additional presence of R667 in the cultures induced a concentration-dependent decrease in the abundance of active MMP-1 in the culture supernatants. Immunoblot analysis with antibodies to MMP-3 did not detect either pro or active forms of MMP-3 in the culture supernatants of cells incubated in the absence of TGF-β1 and R667 (Figure 3A). In the presence of TGF-β1, such analysis revealed 49- and 45-kDa immunoreactive proteins corresponding to active forms of MMP-3. The further addition of R667 induced a concentration-dependent decrease in the abundance of active MMP-3 in the culture supernatants.
Gelatin zymography of culture supernatants revealed that culture of the cells with TGF-β1 resulted in an increase in the intensity of the bands corresponding to pro-MMP-2 and active MMP-2, but that this effect of TGF-β1 was not influenced by R667 (Figure 3A). Immunoblot analysis with antibodies to TIMP-1 or to TIMP-2 revealed that neither TGF-β1 nor R667 affected the abundance of these proteins in culture supernatants (Figure 3B).
We next examined whether R667 might affect the phosphorylation of FAK in HTFs cultured in collagen gels. Immunoblot analysis revealed that TGF-β1 (1 ng/ml) induced the phosphorylation of FAK in HTFs and that this effect was inhibited by R667 in a concentration-dependent manner (Figure 4). Immunoblot analysis also revealed that culture of the cells for 3 days in the presence of TGF-β1 (1 ng/ml) and various concentrations of R667 had no effect on the abundance of α-SMA (Figure 5).
With the use of specific EIAs, we examined whether R667 might affect the release of fibronectin and collagen type I induced by TGF-β1 from HTFs. The release of fibronectin (Figure 6A) and the C peptide of procollagen type I (Figure 6B) induced by TGF-β1 (1 ng/ml) was inhibited by R667 in a concentration-dependent manner.
Finally, we examined the effect of R667 on the change in IOP in rats after glaucoma filtration surgery (Figure 7A). IOP was significantly reduced at 7 days after the surgery in rats treated with R667 during the procedure compared with those that received PBS vehicle (Figure 7B).
We have shown that the RARγ agonist R667 inhibited TGF-β1-induced collagen gel contraction mediated by HTFs as well as MMP-1 and MMP-3 release, FAK phosphorylation, and the production of fibronectin and type I collagen induced in these cells by TGF-β1. Furthermore, we found that R667 lowered IOP in a rat model of glaucoma filtration surgery.
Retinoic acids have diverse effects on development, immunity, and tissue repair and regulate the differentiation and function of many cell types . RARs, including RARα, RARβ, and RARγ, belong to the steroid and thyroid hormone receptor superfamily of ligand-dependent transcription factors  and bind all-trans retinoic acid (ATRA) and its isomer 9-cis retinoic acid . We have previously shown that ATRA attenuates TGF-β signaling in HTFs . Failure of filtration surgery due to excessive postoperative scarring remains a major problem, and recent studies into alternative methods for the prevention of such tissue fibrosis have focused on the regulation of HTFs by various growth factors [9,25]. TGF-β is a key mediator of scarring after glaucoma surgery and has become a major target for modulation of the wound healing response. We have now shown that the RARγ agonist R667 inhibited TGF-β1-induced collagen gel contraction mediated by HTFs in a concentration- and time-dependent manner, whereas the RARα agonist Am580 inhibited this process to a lesser extent and the RARβ agonist BmS453 had no such effect. Our results thus suggest that R667 may prove effective as a drug to reduce scar formation after glaucoma filtration surgery.
Matrix contraction and remodeling during wound healing are mediated in part by the release of MMPs and the formation of focal adhesions by fibroblasts. The regulation of MMP activity is a complex and finely tuned process in which TIMPs as well as MMP synthesis and activation play an important role and to which TGF-β is thought to contribute . Increased expression of MMPs and TIMPs is associated with the formation of filtering blebs . An MMP inhibitor was found to have a pronounced antiscarring effect in an experimental model of glaucoma filtration surgery . We have previously shown that MMPs contribute to TGF-β1-induced collagen gel contraction mediated by HTFs . In the present study, R667 inhibited the TGF-β1-induced release of MMP-1 and MMP-3 by HTFs in a concentration-dependent manner, without affecting the release of MMP-2, TIMP-1, or TIMP-2, suggesting that this inhibition of MMP-1 and MMP-3 production contributes to the attenuation of TGF-β1-induced collagen gel contraction by R667. The formation of actin stress fibers as well as the assembly of focal adhesions that connect cells to the ECM contribute to TGF-β-induced cell contractility and tissue contraction . FAK is a protein tyrosine kinase that plays an important role in the assembly of focal adhesions and cell contractility . We have now shown that the inhibition of TGF-β1-induced HTF contraction by R667 was associated with attenuation of the TGF-β1-induced phosphorylation of FAK at Tyr576/577 in these cells. TGF-β also induces FAK phosphorylation at other tyrosine residues, including Tyr397 and Tyr925, in some cell types [29,30]. The effect of R667 on FAK phosphorylation at other tyrosine sites remains to be determined.
Transdifferentiation of fibroblasts into myofibroblasts, which is characterized by the onset of expression of the contractile protein α-SMA, also plays a key role during wound contraction . Consistent with our previous results , we found that α-SMA expression was not affected by TGF-β1 or R667 in HTFs cultured in a floating collagen gel, likely because tension is required for the induction of α-SMA expression and is not generated to a sufficient extent in the floating matrix . We found that α-SMA was expressed by HTFs without TGF-β1 stimulation (Figure 5), suggesting that prior exposure of the cells to serum induced myofibroblast differentiation to some extent, consistent with previous observations showing a low level of α-SMA expression in cultured HTFs . Further examination of the effects of R667 on native Tenon fibroblasts is warranted.
Tissue scarring is associated with changes in the composition of the ECM, which acts as a primary scaffold for cell attachment and migration. Excessive accumulation of ECM proteins such as collagen and fibronectin in the subconjunctival space contributes to bleb failure from wound obstruction in glaucoma filtration surgery . The expression of fibronectin is strongly associated with hypertrophic scar formation and enhanced TGF-β-driven scarring responses after glaucoma revision surgery . We have previously shown that fibronectin contributes to the contractility of fibroblasts cultured in a three-dimensional collagen gel . In the present study, the TGF-β1-induced production of type I collagen and fibronectin by HTFs was attenuated by R667 in association with the inhibitory effect of this drug on TGF-β1-induced collagen gel contraction mediated by these cells. We found that R667 also effectively reduced IOP after experimental glaucoma filtration surgery in rats. Inhibition of the contractility of Tenon fibroblasts and ECM synthesis by R667 may contribute to this in vivo action. Comparison of surgical outcomes for animals treated with R667 and mitomycin C is now warranted for further evaluation of the potential of RARγ agonists as alternative, more physiologic agents to protect against postoperative scarring.
Several studies have shown that ATRA possesses antifibrotic potential as a result of its attenuation of TGF-β actions . TGF-β signaling via Smad and other pathways and retinoid signaling via nuclear receptors intersect in complex ways that are frequently cell dependent . A derivative of ATRA was previously shown to inhibit TGF-β-induced collagen expression by blocking Smad2/3 phosphorylation in hepatic stellate cells . We recently showed that the RARγ agonist R667 attenuated the EMT in RPE cells by suppressing mitogen-activated protein kinase, Smad2, and Akt signaling . Modulation of such signaling pathways may also contribute to the inhibitory effects of R667 on ECM remodeling mediated by HTFs.
Humans express three isoforms of TGF-β (TGF-β1, TGF-β2, and TGF-β3), all of which were detected in association with the conjunctival scarring response . All three isoforms of TGF-β also stimulate the proliferation, migration, and contractility of fibroblasts in vitro, and they show similar stimulatory effects on conjunctival scarring in vivo . Inflammatory processes as well as other growth factors such as vascular endothelial growth factor play a role in conjunctival scarring . Neutralizing antibodies to TGF-β2 did not have a significant effect on bleb survival after glaucoma filtration surgery in a randomized clinical trial , suggesting that an agent that targets all three isoforms of TGF-β or a combination of treatments may be necessary for such an effect. In the present study, the RARγ agonist R667 lowered IOP in a rat model of glaucoma filtration surgery in vivo. We also previously showed that R667 attenuated TGF-β2-induced EMT in RPE cells . Attenuation of TGF-β2 activity might thus also contribute to the lowering of IOP by R667.
Retinoic acids play complex roles in eye development and physiology . The availability of retinoic acids in the anterior chamber has been implicated as a determinant of the pathological changes associated with primary congenital glaucoma . Whether RARγ agonists affect other aspects of ocular physiology, such as aqueous flow or the metabolism of aqueous humor, remains to be determined. Whereas the pharmacokinetics and toxicity of R667 in the eye have not been examined to date, those after oral administration of the drug have been studied in human subjects [41,42]. The half-life of R667 in plasma was found to be about 10 h, and a single dose of 1 mg was well tolerated, with no unexpected safety concerns . Moreover, we have previously shown that injection of R667 into the eye ameliorated subretinal fibrosis in mice . Further studies are warranted to assess the pharmacokinetics and possible side effects of ocular administration of this drug.
In conclusion, our results show that the RARγ agonist R667 inhibited the contractility of HTFs as well as ECM production by the cells induced by TGF-β1 in vitro. Moreover, this agent lowered IOP in a rat model of glaucoma filtration surgery in vivo. RARγ agonists might thus prove effective for inhibition of scar formation during subconjunctival wound healing after such surgery.
This study was supported by Grants in Aid from the Japan Society for the Promotion of Science (JSPS, to K.S. and Y.L.), by a JSPS postdoctoral fellowship for foreign researchers (Y.L.), and by a grant (no. 81100635) from the National Science Foundation of China (Y.L.).