YAP/TAZ Are Essential for TGF-β2–Mediated Conjunctival Fibrosis

Abstract

Purpose: To investigate the roles of Yes-associated protein (YAP)/transcriptional co-activator with PDZ-binding motif (TAZ), the major effector molecules of the Hippo pathway, in TGF-β2–mediated conjunctival fibrosis.

Methods: Primary human conjunctival fibroblasts were treated with TGF-β2. The expression of YAP/TAZ was examined by Western blot analyses and immunocytochemistry. The expression of fibrotic proteins and genes were evaluated by Western blot analyses and quantitative real-time PCR, respectively. The effects of YAP/TAZ on fibrotic changes were examined by knockdown experiments and the YAP/TAZ inhibitor, verteporfin.

Results: TGF-β2 stabilized YAP/TAZ and subsequently activated Smad2/3, which led to the transcription of fibrotic genes in human primary conjunctival fibroblasts. These fibrotic genes were differently regulated by YAP/TAZ. Notably, α-smooth muscle actin, fibronectin, collagen I, and collagen IV were primarily regulated by YAP. In contrast, CCN family proteins (CTGF and CYR61) depended on both YAP and TAZ. Mechanistically, YAP/TAZ were located in close proximity to Smad2/3, and in particular, YAP was required for TGF-β2-mediated phosphorylation and the nuclear translocation of Smad2/3. Furthermore, a YAP/TAZ inhibitor markedly suppressed TGF-β2-mediated fibrotic changes in conjunctival fibroblasts.

Conclusions: YAP/TAZ acted as a molecular hub of TGF-β2 signaling in a cellular model of conjunctival fibrosis. Moreover, verteporfin, a YAP/TAZ inhibitor exerted potent antifibrosis effects by suppressing TGF-β2-YAP/TAZ-Smad signaling. Our study highlights YAP/TAZ as essential regulators of conjunctival fibrosis and shows that inhibition of YAP/TAZ might potentially improve the outcomes of glaucoma filtration surgery.

Glaucoma is one of the leading causes of blindness and affects more than 70 million people worldwide.^1,2 Glaucoma filtration surgery is the most commonly performed surgery to reduce IOP. In this category of surgical modalities, a filtering bleb is created to drain the aqueous humor into the lymphatic and/or venous systems. Glaucoma filtration surgery is highly effective at reducing the IOP, but many clinical studies have reported a reelevation of IOP due to fibrosis-mediated disappearance of the filtering bleb.³ 

Wound-healing process and the resulting fibrosis at the conjunctiva involves four successive but overlapping phases: hemostasis phase, inflammatory phase, proliferative phase, and remodeling phase.^4,5 These processes involve a variety of cells, signals, and soluble extracellular factors,⁴ including TGF-β, platelet-derived growth factor (PDGF), TNF-α, epidermal growth factor, insulin-like growth factor (IGF)-1, interleukin (IL)-1, IL-6, IL-4, interferons, BFGF, and macrophage-colony stimulating factor (m-CSF).^4,6–8 

TGF-β is one of the most important profibrogenic cytokines in the proliferative phase of wound healing. TGF-β induces the differentiation of fibroblasts into myofibroblasts, which are characterized by the expression of α-smooth muscle actin (α-SMA), a high contractile capacity, and the synthesis of extracellular matrix (ECM).^5,9 Although myofibroblasts usually disappear in the later phase of wound healing, these cells sometimes persist at the wound site, resulting in conjunctival fibrosis and ultimately leading to the reelevation of the IOP. 

Cellular mechanical signal transduction has been linked to pathological wound healing and fibrosis in liver and lung tissues.^10–21 One of the most famous mechanical stress-related signaling pathways is the Hippo pathway.^22,23 The Hippo pathway constitutes a kinase cascade that controls the stability and localization of YAP and TAZ. In the absence of mechanical stress, the kinase cascade is activated, and YAP/TAZ are quickly destabilized by proteasome-mediated degradation.^24,25 Conversely, in the presence of mechanical stress, the kinase cascade is inactivated, and YAP/TAZ are stabilized and translocated to the nucleus. In the nucleus, YAP/TAZ subsequently interacts with TEAD transcription factors and drives the transcription of genes related to cell proliferation, cell survival, and fibrosis.^26–31 

The filtering bleb, which is created after glaucoma filtration surgery, is constantly subjected to mechanical stress induced by the draining of the aqueous humor. Although a high level of TGF-β2 has been detected in the aqueous humor of patients with glaucoma,³² it was unknown how mechanical stress and TGF-β2 are integrated and lead to conjunctival fibrosis. In this study, we sought to elucidate the role of YAP/TAZ in human conjunctival fibrosis. YAP/TAZ acted as a molecular hub of TGF-β2 signaling in a cellular model of conjunctival fibrosis. Moreover, verteporfin, a YAP/TAZ inhibitor, exerted potent antifibrosis effects by suppressing TGF-β2-YAP/TAZ-Smad signaling. 

Materials and Methods

Cell Culture

Two lots of human primary conjunctival fibroblasts (lot 3072 and 5965) were obtained from ScienCell Research Laboratories (Carlsbad, CA, USA) and cultured according to the manufacturer's recommendations. Cells at passages 4 to 6 were used in all experiments. We checked each cell strain for the morphology and the marker of fibroblasts (fibroblast-specific protein 1 [FSP1]; cat. no. 27957; Abcam, Cambridge, UK). We confirmed that the two cell strains did not exhibit the morphological changes during the passages, and that FSP1 was clearly detected in these cells (data not shown). Each experiment using these cells was repeated at least four times. 

Cell Stimulation

Human primary conjunctival fibroblasts were treated with 5 ng/mL recombinant human TGF-β2 (R&D Systems, Minneapolis, MN, USA) for 48 hours. The concentration of 5 ng/mL could be comparable to the physiological level of human aqueous TGF-β2, because the amount of TGF-β2 in the aqueous humor was reported to range from 2.3 to 8.1 ng/mL, with 60% being in the active form.³³ In experiments using verteporfin, human primary conjunctival fibroblasts were protected from light and were treated with different concentrations of verteporfin (0–2 μM; Sigma-Aldrich Corp., St. Louis, MO, USA) 1 hour before the addition of TGF-β2. Verteporfin has been clinically used through intravenous injection for targeting intraocular blood vessels of retina and choroid. To date, there are no data regarding the concentration of this drug on the ocular surface. Therefore, in the present study, we tested several doses of verteporfin, which have also been used in previously published articles.^34,35 Before treatment, cells were starved overnight in serum-free Dulbecco's Modified Eagle's Medium (DMEM; Wako Pure Chemical Industries, Osaka, Japan). 

Transfections and Small Interfering RNA (siRNA)

RNAiMAX was used to transfect siRNAs (Life Technologies, Carlsbad, CA, USA). All siRNAs were Silencer Select predesigned siRNAs (ID# s20366, ID# s24789) from Thermo Fischer Scientific (Waltham, MA, USA). 

Western Blot Analysis

Western blot analyses were performed using previously described methods,^36,37 with primary antibodies recognizing the following proteins: YAP (cat. no. 14074; Cell Signaling Technology, Danvers, MA, USA), TAZ (cat. no. 8418; Cell Signaling Technology), TEAD1 (cat. no. 12292; Cell Signaling Technology), α-SMA (cat. no. A2547; Sigma-Aldrich Corp.), fibronectin (cat. no. ab6328; Abcam), type I collagen (cat. no. ab138492; Abcam), type IV collagen (cat. no. ab6586; Abcam), CCN2 (cat. no. sc-14939; Santa Cruz Biotechnology, Dallas, TX, USA), CCN1 (cat. no. sc-13100; Santa Cruz Biotechnology), pSmad2 (cat. no. 3108; Cell Signaling Technology), pSmad3 (cat. no. ab52903; Abcam), Smad2 (cat. no. 5339; Cell Signaling Technology), and Smad3 (cat. no. ab40854; Abcam). The band densities of each sample were normalized to β-actin (cat. no. A1978; Sigma-Aldrich Corp.). 

Immunocytochemistry

Immunocytochemistry was performed as previously described.³⁷ Briefly, cells were fixed with 4% paraformaldehyde (PFA), permeabilized with 0.5% Triton X-100, blocked with 10% fetal bovine serum, and then incubated overnight with monoclonal antibodies against YAP/TAZ (cat. no. sc-101199; Santa Cruz Biotechnology), TEAD1 (cat. no. 610922; BD Biosciences, San Jose, CA, USA), and Smad2/3 (cat. no. 8685; Cell Signaling Technology). Cells were then probed with an Alexa Fluor 488-conjugated goat anti-mouse or anti-rabbit IgG secondary antibody (Thermo Fisher Scientific) and then mounted with antifade reagent containing DAPI (4′,6-diamidino-2-phenylindole). Images were obtained using a fluorescence microscope (model BZ 700; Keyence, Osaka, Japan). 

Quantitative RT-PCR

RNA was isolated from human primary conjunctival fibroblasts using TRIzol reagent (Thermo Fischer Scientific), according to the manufacturer's instructions. RNA was reverse transcribed using the PrimeScript RT-PCR kit (TaKaRa Bio, Shiga, Japan) and subjected to quantitative PCR (SYBR Premix Ex Taq II; TaKaRa Bio) using a 7300 Real-time PCR System (Thermo Fischer Scientific). All expression data were normalized to levels of the 18S RNA. The primer sequences are listed in Supplementary Table S1. 

Proximity Ligation Assay (PLA)

The PLA was performed according to the manufacturer's instructions (Duolink; Sigma-Aldrich Corp.), with minor modifications. Briefly, human primary conjunctival fibroblasts seeded on glass coverslips were treated with or without 1.5 μM verteporfin for 1 hour before stimulation with 5 ng/mL TGF-β2. After 12 hours of TGF-β2 treatment, cells were fixed with 4% PFA, permeabilized with 0.5% Triton X-100, blocked with Blocking solution, and incubated overnight with a mixture of primary antibodies against mouse YAP/TAZ and rabbit Smad2/3. After three washes with Wash buffer, cells were incubated with PLA probes (anti-mouse PLUS and anti-rabbit MINUS) for 1 hour. Cells were washed twice, followed by a 30-minute incubation with the ligation mixture and a subsequent incubation with the amplification mixture for 100 minutes. After a final wash, the cells were mounted, and PLA signals were detected using a Keyence microscope. 

Collagen Gel Contraction Assay

The collagen gel contraction assay was performed as previously described.³⁸ Briefly, type I collagen (Nitta Gelatin; Yao, Osaka, Japan), a 10-fold concentration of DMEM (Sigma-Aldrich Corp.), reconstitution buffer (Nitta Gelatin; Yao), and suspended human primary conjunctival fibroblasts were mixed at an 8:1:1:1 ratio. The mixture was divided into each well of a BSA-coated 24-well plate. After gelation, DMEM aliquots containing increasing concentrations of verteporfin (0–2 μM) were poured onto the gels for 1 hour, followed by the addition of 5 ng/mL TGF-β2. Gels were separated from the side of the wells with 26-gauge needles and incubated for 72 hours after initiation. 

Cell Proliferation Assay

Human primary conjunctival fibroblasts were seeded in 96-well plates at a density of 1 × 10⁴ cells per well. Cells were treated with increasing concentrations of verteporfin (0–2 μM) for 1 hour, followed by 5 ng/mL TGF-β2 for 48 hours. Proliferation was assessed using the WST-8 assay (Dojindo Laboratories, Kumamoto, Japan). 

Cytotoxicity Analysis

The cytotoxicity of verteporfin was evaluated by dual staining with Hoechst 33342/propidium iodide (PI) (Dojindo Laboratories). PI is permeant to only dead/damaged cells, whereas Hoechst stains live/damaged/dead cells.³⁹ The cells were observed using a Keyence microscope. 

Quantification and Statistical Analysis

The results are presented as the means ± SEM. All data were analyzed using GraphPad Prism 7 software (GraphPad, La Jolla, CA, USA). Comparisons among groups were analyzed using Dunnett's test and the Tukey-Kramer multiple comparisons test. In all analyses, P < 0.05 was considered statistically significant. 

Results

TGF-β2 Stabilizes YAP/TAZ and Promotes Their Nuclear Localization

We treated human primary conjunctival fibroblasts with TGF-β2 and examined protein levels to investigate the effects of TGF-β2 on effector molecules of the Hippo pathway. The TGF-β2 treatment gradually increased levels of the YAP/TAZ proteins (Fig. 1A). The level of TEAD1, a transcription factor that interacts with YAP/TAZ, was not changed on TGF-β2 stimulation. YAP/TAZ shuttles between the cytosol and nucleus in response to external stimulation.^24,25 We performed immunofluorescence staining to examine their localization following TGF-β2 stimulation. The TGF-β2 treatment induced YAP/TAZ translocation to the nucleus, whereas TEAD1 was already located in the nucleus in control cells, and its localization remained unchanged after TGF-β2 stimulation (Fig. 1B). Based on these results, YAP/TAZ are regulated by TGF-β2. 

Figure 1

 

Effects of TGF-β2 stimulation on the Hippo pathway. (A) Human primary conjunctival fibroblasts were stimulated with 5 ng/mL TGF-β2, and then protein lysates were collected at different time points and subjected to Western blot analysis. Representative bands from eight independent samples are shown. (B) Human primary conjunctival fibroblasts were stimulated with or without TGF-β2 for 12 hours. Cells were then stained with an anti-YAP, TAZ, or TEAD1 antibody. Arrows indicate the nucleus. Scale bar: 500 μm.

YAP/TAZ Differentially Regulate the Expression of TGF-β2–Induced Fibrotic Proteins and Genes

Next, we silenced YAP/TAZ and examined the impact of YAP/TAZ inactivation on TGF-β2–induced fibrotic changes in conjunctival fibroblasts. We confirmed the efficacy and specificity of siRNAs by Western blotting (Supplementary Fig. S1). Knockdown of YAP, but not TAZ, completely suppressed TGF-β2–induced increases in the levels of the fibrotic protein α-SMA and ECM proteins, such as fibronectin, and collagen type I and type IV (Fig. 2A). In contrast, levels of CCN family proteins (CTGF and CYR61) were decreased only by the simultaneous knockdown of YAP and TAZ (Fig. 2B). We also performed quantitative PCR to examine changes in the transcription of fibrotic genes. Consistent with the changes in protein levels, silencing of YAP alone effectively suppressed TGF-β2–induced expression of ACTA2 (encodes α-SMA), FN1 (encodes fibronectin), COL1A1 (encodes part of collagen type I), and COL4A1 (encodes part of collagen type IV) to basal levels (Fig. 2C). Silencing of TAZ did not suppress TGF-β2–induced FN1 and COL1A1 expression (Fig. 2C). Interestingly, silencing of TAZ moderately suppressed TGF-β2–induced ACTA2 and COL4A1 expression, but the expression levels were still significantly higher than the basal levels (Fig. 2C). In addition, silencing of YAP or TAZ significantly suppressed the TGF-β2–mediated induction of CTGF and CYR61 expression, but the complete suppression of the expression of these genes was achieved only by the simultaneous knockdown of YAP and TAZ. Moreover, we found that TGF-β2 itself upregulated the TGFB2 mRNA. YAP silencing was sufficient to reduce the expression of the TGFB2 mRNA to the basal level. Interestingly, TAZ silencing also significantly reduced the TGFB2 mRNA level but with less efficacy than YAP knockdown. Thus, TGF-β2–mediated fibrotic changes are transcriptionally regulated by YAP and TAZ with either overlapping or distinctive mechanisms. 

Figure 2

 

Effects of YAP/TAZ on TGF-β2–induced changes in the levels of fibrotic proteins and genes. Human primary conjunctival fibroblasts were transfected with siRNAs against YAP or TAZ alone or in combination and then stimulated with TGF-β2 for 48 hours. (A, B) Western blot analysis of TGF-β2–induced changes in the levels of fibrotic proteins. (A) Knockdown of YAP, but not TAZ, inhibited TGF-β2–induced expression of the α-SMA and ECM proteins, including FN, Col I, and Col IV. Data are presented as the means ± SEM from four independent samples per group. *P < 0.05 compared with the TGF-β2 group; **P < 0.01 compared with the TGF-β2 group; ***P < 0.001 compared with the TGF-β2 group. (B) TGF-β2–induced expression of the CCN family of matricellular proteins (CTGF and CYR61) was significantly suppressed only when YAP and TAZ were simultaneously silenced. Data are presented as the means ± SEM from four independent samples per group. *P < 0.05 compared with the TGF-β2 group; **P < 0.01 compared with the TGF-β2 group; ***P < 0.001 compared with the TGF-β2 group. (C–E) Real-time RT-PCR analysis of genes related to fibrosis. (C) TGF-β2–mediated increases in the levels of the ACTA2 (α-SMA), FN1 (fibronectin), COL1A1 (collagen type I), and COL4A1 (collagen type IV) mRNAs were all significantly suppressed by YAP knockdown, whereas levels of the ACTA2 mRNA were also reduced by TAZ knockdown. Data are presented as the means ± SEM from three independent samples per group. ***P < 0.001 compared with the TGF-β2 group. (D) RT-PCR analysis of CCN genes. TGF-β2–induced increases in the levels of the CTGF and Cyr61 mRNAs were significantly suppressed by YAP/TAZ single knockdown, but simultaneous knockdown of YAP and TAZ further reduced their expression. (E) RT-PCR analysis of the TGF-β2 gene. TGF-β2 upregulated the expression of the TGFB2 mRNA, which was also inhibited by knockdown of YAP/TAZ.

YAP Knockdown Suppresses TGF-β2–Smad Signaling

Given the marked suppression of TGF-β2–induced fibrotic changes by YAP/TAZ inactivation, we investigated whether YAP/TAZ play substantial roles in TGF-β2 signaling. Human primary conjunctival fibroblasts were stimulated with TGF-β2 in the presence or absence of YAP/TAZ silencing, and the levels of phosphorylated Smad2/3 were examined. TGF-β2–induced Smad2/3 phosphorylation in human primary conjunctival fibroblasts is bimodal (an early and transient peak at 1 hour, with a nadir at 3 hours, and then a second long-lasting peak between 6 and 12 hours),⁴⁰ and this bimodal phosphorylation was suppressed in YAP-silenced cells (Fig. 3A). Interestingly, TAZ silencing did not affect TGF-β2–induced Smad2/3 phosphorylation. Because phosphorylation of Smad2/3 is required for the nuclear translocation of Smad2/3,⁴¹ we examined the cellular localization of Smad2/3 using immunofluorescence staining. Consistent with the Western blotting results, YAP silencing effectively suppressed TGF-β2–induced Smad2/3 nuclear translocation, whereas TAZ silencing had no effect (Fig. 3B). Taken together, these results suggest that YAP, but not TAZ, is required for TGF-β2–induced Smad2/3 signaling in human primary conjunctival fibroblasts. 

Figure 3

 

Effects of YAP/TAZ on canonical TGF-β-Smad signaling. (A) Time-course analysis of the levels of phosphorylated Smad2/3. An early and transient phosphorylation signal was observed 1 hour after stimulation with TGF-β2, with a second long-lasting phosphorylation observed after 6 hours. The levels of the phosphorylated proteins were markedly reduced by YAP knockdown. TAZ knockdown did not affect Smad2/3 phosphorylation. (B) Immunofluorescence staining for Smad2/3. YAP knockdown, but not TAZ knockdown, inhibited nuclear translocation of Smad2/3 after 12 hours of TGF-β2 treatment. Scale bar: 500 μm.

A YAP/TAZ Inhibitor, Verteporfin, Suppresses TGF-β2–Smad Signaling

We next examined the effects of a potent YAP/TAZ inhibitor, verteporfin,²⁰ on the TGF-β2–induced fibrotic changes in conjunctival fibroblasts. Verteporfin reduced YAP/TAZ levels in a dose-dependent manner, and moreover, verteporfin reduced the level of TEAD1 (Fig. 4). Accordingly, the levels of fibrotic proteins, including α-SMA, fibronectin, type I and type IV collagens, as well as CCN family proteins (CTGF and CYR61), were all markedly decreased in verteporfin-treated cells. Notably, verteporfin exerted a strong inhibitory effect on the production of ECM at a very low concentration. 

Figure 4

 

Effects of a YAP/TAZ inhibitor, verteporfin, on TGF-β2–induced changes in the levels of fibrotic proteins. Verteporfin was administered 1 hour before the TGF-β2 treatment, and proteins were collected after 48 hours. Western blot analysis showed the dose-dependent decreases in the levels of the YAP, TAZ, and TEAD1 proteins induced by verteporfin. TGF-β2 induced α-SMA and ECM production (fibronectin, type I and type IV collagens), whereas the levels of CCN family proteins (CTGF and CYR61) were dose-dependently decreased by verteporfin. VP, verteporfin.

We next examined whether verteporfin was also capable of regulating TGF-β2–Smad signaling. As expected, verteporfin suppressed TGF-β2–induced Smad2/3 phosphorylation and nuclear translocation (Figs. 5A, 5B). Notably, levels of the total Smad2/3 proteins were also reduced by verteporfin, suggesting that verteporfin affects both the phosphorylation-mediated signaling and protein stability of Smad2/3. YAP/TAZ are located in close proximity to Smad2/3.⁴² We performed an in situ PLA, which specifically visualizes the interactions of proteins located in close proximity, to examine the spatial distribution of YAP/TAZ and Smad. In unstimulated cells, a weak but clear fluorescence signal was detected in the cytosol. On TGF-β2 stimulation, a very strong fluorescence signal was detected in the nucleus. These signals had almost completely disappeared following treatment with verteporfin (Fig. 5C). Thus, YAP/TAZ regulates TGF-β2 signaling through a physical interaction with Smad2/3. 

Figure 5

 

Effects of verteporfin on canonical TGF-β–Smad signaling. (A) TGF-β2–induced Smad2/3 phosphorylation was evaluated by analyzing the time course of changes on Western blots. Phosphorylation of Smads was inhibited by verteporfin (1.5 μM). Moreover, total cellular Smad2/3 protein levels were also reduced by verteporfin. (B) Images of immunocytochemistry showing the verteporfin-mediated inhibition of the nuclear translocation of Smad2/3 after 12 hours of TGF-β2 treatment. Scale bar: 500 μm. (C) The PLA showed that TGF-β2 stimulation induced a robust increase in the signals for Smad2/3-YAP/TAZ in nucleus after 12 hours of TGF-β2 treatment, which was completely abolished by verteporfin. Scale bar: 500 μm.

Verteporfin Inhibits TGF-β2–Induced Fibrosis in a Cellular Model

During pathological wound healing after glaucoma surgery, excess ECM production and its contraction lead to surgical failure.⁴³ Thus, we examined whether YAP/TAZ inactivation was capable of suppressing contractility using an in vitro model. Conjunctival fibroblasts were cultured in a collagen gel and stimulated with TGF-β2. The diameter of the gel was measured in the presence or absence of verteporfin as an indicator of contractility. Verteporfin effectively suppressed the contraction of the collagen gel in a dose-dependent manner (Figs. 6A, 6B). 

Figure 6

 

Effects of verteporfin on collagen gel contraction. (A) Representative photos of gels 72 hours after stimulation with TGF-β2. Dotted lines indicate the edges of the collagen gels. (B) The extent of gel contraction was assessed by measuring the gel diameter. Data are presented as the means ± SEM from 10 independent samples per group. ***P < 0.001 compared with the control or TGF-β2 group.

In addition to the excess ECM production and contractility, TGF-β signaling also induces aberrant cell proliferation.⁴ We then investigated the effects of YAP/TAZ inactivation on fibroblast proliferation. Conjunctival fibroblasts were treated with TGF-β2 in the presence or absence of verteporfin. Verteporfin alone suppressed cell growth at high concentrations (>1.5 μM). Importantly, verteporfin inhibited the TGF-β2–induced increase in the number of viable cells, even at a low concentration (0.5 μM); the effect was dose-dependent (Fig. 7A). Because verteporfin is capable of producing reactive oxygen species,⁴⁴ the decrease in the cell number might be attributed to the cytotoxicity of verteporfin. Double staining with Hoechst and PI was conducted to examine the cytotoxic effects. Few PI-positive (dead or damaged) cells were detected following treatment with verteporfin, suggesting that the decrease in cell proliferation induced by the verteporfin treatment occurred through a cell death–independent mechanism (Fig. 7B). 

Figure 7

 

Effects of verteporfin on cell proliferation and cytotoxicity. (A) Results from the WST-8 assay showed that treatment with verteporfin alone significantly reduced the number of viable cells in a dose-dependent manner. Moreover, the TGF-β2–induced increase in cell viability was also significantly reduced by verteporfin. Data are presented as the means ± SEM from four independent samples per group. **P < 0.01 compared with the control or TGF-β2 group; ***P < 0.001 compared with the control or TGF-β2 group. (B) Double staining with Hoechst and PI showed few PI-positive (dead or damaged) cells following treatment with 2 μM verteporfin. Scale bar: 500 μm.

Discussion

In the present study, we demonstrated that YAP/TAZ were essential for TGF-β2–mediated signaling in a cellular model of conjunctival fibrosis. TGF-β2 stabilized YAP/TAZ and induced their nuclear translocation, which led to the transcription of fibrotic genes in human primary conjunctival fibroblasts. These fibrotic genes were differently regulated by YAP/TAZ. Mechanistically, YAP/TAZ were located in close proximity to Smad2/3, and in particular, YAP was required for TGF-β2–mediated phosphorylation and the nuclear translocation of Smad2/3. Furthermore, a YAP/TAZ inhibitor markedly suppressed TGF-β2–mediated fibrotic changes in conjunctival fibroblasts. 

The TGF-β family proteins are considered master regulators of pathogenic fibrosis.⁴⁵ TGF-β1 induces Smad2/3 phosphorylation, which subsequently induces the transcription of profibrotic genes in various tissues, including the lung, liver, and kidney. Increased expressions of TGF-β1 and CTGF were detected in the conjunctival fibroblasts from ocular cicatricial pemphigoid patients, which is one of the representatives of ocular surface fibrosis, and these cytokines could contribute to the pathogenesis and the process of conjunctival fibrosis.^46,47 In contrast to the pathogenic role of TGF-β1 in diverse tissues, TGF-β2 has predominantly been detected in the aqueous humor and is associated with eye diseases such as glaucoma.^32,48 Considering the constant mechanical stress to which the filtering bleb is subjected after glaucoma filtration surgery, the interplay between canonical TGF-β2 signaling and mechanical signaling potentially contributes to abnormal fibrosis and leads to surgical failure. Our study demonstrated that YAP/TAZ, the major components of mechanical response signaling, directly regulated TGF-β2–mediated fibrosis by modulating Smad2/3 pathway. Based on these results, the YAP/TAZ-mediated mechanical stress response is intimately coupled to TGF-β2 signaling and contributes to the pathogenesis of conjunctival fibrosis. 

Myofibroblasts at the injury site continuously deposit ECM, which increases the stiffness of the microenvironment and leads to fibrosis.⁴⁹ Given the central role of YAP/TAZ in the cellular response to the stiffness of the microenvironment, YAP/TAZ have been implicated in pathological fibrosis. Indeed, overexpression of YAP/TAZ has been observed in liver, lung, and renal fibrosis.^15,17,19,20 Notably, YAP/TAZ are associated with TGF-β1–Smad signaling in renal fibrosis.²⁰ Consistent with these results, YAP/TAZ were also essential for TGF-β2–mediated fibrosis in primary conjunctival fibroblasts in the present study. TGF-β2 stimulation rapidly increased levels of the YAP/TAZ proteins. Genetic and pharmacological suppression of YAP/TAZ abolished TGF-β2–induced production of ECM and contractility. Importantly, inhibition of YAP/TAZ drastically suppressed the transcription of the TGFB2 gene in fibroblasts stimulated with TGF-β2. Thus, YAP/TAZ activation leads to the production of ECM and TGF-β2, which in turn sustains YAP/TAZ activity and creates a feedforward loop in conjunctival fibroblasts. 

An important finding of our study is that YAP/TAZ differently regulated the transcription of fibrotic genes in conjunctival fibroblasts. In general, YAP/TAZ share many common target genes. For example, CTGF (CCN2) and CYR61 (CCN1) are matricellular proteins that act as signaling molecules and have been implicated in pathogenic fibrosis.^50,51 TGF-β2–mediated induction of CTGF/CYR61 proteins were not suppressed unless both YAP and TAZ were silenced, but on the transcriptional level, silencing of YAP or TAZ alone significantly suppressed the TGF-β2–mediated induction of CTGF/CYR61 mRNA expression (the complete suppression of the expression of these genes was achieved only by the simultaneous knockdown of YAP and TAZ). This mRNA-protein discrepancy may be due to posttranscriptional or posttranslational regulation. The precise mechanism still requires further research. Based on these results, YAP and TAZ have overlapping functions in the transcription of these genes. In contrast to CTGF/CYR61, silencing of YAP alone completely suppressed TGF-β2–induced transcription of genes such as FN1, COL1A1, and COL4A1. Importantly, TAZ silencing had no effects on the expression of these genes. Thus, a subset of fibrotic genes is specifically regulated by the TGF-β2–YAP axis. 

It is generally accepted that most of the molecular functions of YAP and TAZ are overlapped. YAP and TAZ share the same transcription factors, such as TEADs, and thus activate the same target genes. Nevertheless, a number of articles also indicate that YAP and TAZ have distinct functions. For example, TAZ, but not YAP, modulates mesenchymal stem cell differentiation.⁵² In hepatocellular carcinoma, TAZ is predominantly expressed under normal conditions, but a switch into YAP-dominant expression is thought to be a key step to acquire a cancer stem-like property.⁵³ Moreover, YAP- and TAZ-deficient mice show different phenotypes: YAP-deficient mice are embryonic lethal, whereas TAZ knockout mice are viable.^54–56 The differential functions of YAP and TAZ can be explained by the differential binding partners. For example, YAP specifically binds to ErbB4 and p73, whereas TAZ specifically binds to PPARγ, Pax3, TBX5, and TTF-1.⁵⁷ In line with these results, we observed that FN1, COL1A1, and COL4A1 were specifically suppressed by silencing of YAP. Thus, it is likely that YAP, but not TAZ, can interact with a unique subset of transcription factors to induce the expression of FN1, COL1A1, and COL4A1 on TGF-β2 stimulation. Further study is needed to elucidate the precise molecular mechanism in the future. 

In support of these inhibitory effects of YAP, TGF-β2–induced Smad2/3 phosphorylation and nuclear translocation was blocked only by YAP, but not by TAZ, in human primary conjunctival fibroblasts. These data partly differ from previously published data in human embryonic stem cells and other immortalized cell lines wherein TAZ silencing does not block TGF-β1–induced phosphorylation but does inhibit Smad2/3 nuclear translocation.⁵⁸ As we used human primary cells in our experiments, the difference between cell types could account for the discrepancy, and further studies should be explored. 

Verteporfin is a potent YAP/TAZ inhibitor that reduces the abundance of YAP/TAZ.²⁰ Consistent with previous findings, verteporfin effectively reduced the levels of both Smad2 and 3 proteins in the present study, thereby leading to the suppression of TGF-β2–Smad2/3 signaling and subsequent fibrotic changes. Curiously, this effect of verteporfin on Smad2/3 differed from the effect of genetic knockdown of YAP/TAZ because siRNA treatments decreased levels of only phosphorylated Smad2/3 on TGF-β2 stimulation. Consistent with our results, the verteporfin treatment also decreases Smad2/3 protein levels in renal fibroblasts treated with TGF-β1.²⁰ The inhibitory effects of verteporfin on TGF-β2–treated human primary conjunctival fibroblasts is at least partly YAP/TAZ-dependent, but the reason the effect of verteporfin on Smad2/3 differed from that of YAP/TAZ siRNAs is unknown. Interestingly, the biological function of verteporfin was originally attributed to the generation of reactive oxygen species. Conceivably, excess production of reactive oxygen species induced by the verteporfin treatment might stimulate the proteasome-mediated protein degradation pathway to decrease Smad2/3 levels. 

Fibrosis is a major cause of surgical failure and other eye diseases. To date, an effective drug that specifically prevents fibrosis is not available, with the exception of antiproliferative drugs, such as mitomycin C and 5-fluorouracil, which have been shown to improve the surgical outcomes of glaucoma filtration surgery.^59,60 Given the critical role of TGF-β2 in the pathogenesis of fibrosis, researchers have devoted substantial efforts to target TGF-β2 signaling with different molecules, including antibodies, siRNA, and small molecule inhibitors.^61–63 Notably, monoclonal antibodies have been attracting increasing attention because of their specificity and successful clinical applications in other diseases. However, an anti-TGF-β2 antibody failed to show efficacy in a randomized clinical trial to prevent scarring after first-time trabeculectomy.⁶⁴ Therefore, the development of an effective drug that targets molecules downstream of TGF-β2 instead of TGF-β2 itself may be needed. In the present study, verteporfin dose-dependently inhibited TGF-β2–mediated fibrotic changes in conjunctival fibroblasts. Verteporfin is approved by the Food and Drug Administration, and has been clinically used as a treatment for AMD.⁶⁵ Its phototoxicity may impede the practical use of the drug, but our results suggest that strategies targeting YAP/TAZ represent promising candidates that prevent pathogenic fibrosis in patients after glaucoma surgery. 

As TGF-β2 is the predominant isoform of the TGF-β family proteins inside the eye, targeting YAP/TAZ may lead to breakthroughs of the other intraocular fibrotic pathology, such as proliferative vitreoretinopathy (PVR),⁶⁶ and corneal fibrosis.⁶⁷ As a previous report demonstrated the association between YAP/TAZ and TGF-β1–Smad signaling, YAP/TAZ may also regulate the pathophysiology of conjunctival fibrosis mediated by TGF-β1, such as ocular cicatricial pemphigoid.⁴³ 

In conclusion, the present study revealed indispensable roles for YAP/TAZ in TGF-β2–mediated fibrosis in human conjunctival fibroblasts. Moreover, verteporfin, a YAP/TAZ inhibitor, exhibited therapeutic potential as a treatment that prevents fibrosis after glaucoma surgery. Further characterization of the in vivo function of YAP/TAZ is needed in future studies. 

Acknowledgments

Supported by Japan Society for the Promotion of Science Grant 17K11458 (TI) and 17H04351 (HT). The authors alone are responsible for the content and writing of the paper. 

Disclosure: A. Futakuchi, None; T. Inoue, Novartis Pharmaceutical (F); F.-Y. Wei, None; M. Inoue-Mochita, None; T. Fujimoto, None; K. Tomizawa, None; H. Tanihara, Senju Pharmaceutical (F), Santen Pharmaceutical (F), Alcon Japan (F), Pfizer Japan (F), Kowa (F, C), Otsuka Pharmaceutical (F), Merck Sharp & Dohme Corp. (C)