Effect of Smad7 Gene Overexpression on Transforming Growth Factor β–Induced Retinal Pigment Fibrosis in a Proliferative Vitreoretinopathy Mouse Model

Objective  To determine the effects of Smad7 gene transfer in the prevention of fibrogenic responses by the retinal pigment epithelium, a major cause of proliferative vitreoretinopathy after retinal detachment, in mice.

Methods  Retinal detachment–induced proliferative vitreoretinopathy in a mouse model. Forty-eight eyes received either an adenoviral gene transfer of Smad7 or Cre recombinase gene only. The eyes were histologically analyzed. A retinal pigment epithelial cell line, ARPE-19, was used to determine whether Smad7 gene transfection suppresses the fibrogenic response to transforming growth factor (TGF) β2 exposure.

Results  The Smad7 gene transfer inhibited TGF-β2/Smad signaling in ARPE-19 cells and expression of collagen type I and TGF-β1 but had no effect on their basal levels. In vivo Smad7 overexpression resulted in suppression of Smad2/3 signals and of the fibrogenic response to epithelial-mesenchymal transition by the retinal pigment epithelium.

Conclusion  Smad7 gene transfer suppresses fibrogenic responses to TGF-β2 by retinal pigment epithelial cells in vitro and in vivo.

Clinical Relevance  Smad7 gene transfer might be a new strategy to prevent and treat proliferative vitreoretinopathy.

Proliferative vitreoretinopathy (PVR), characterized by the formation of scarlike fibrous tissue, is a major complication of rhegmatogenous retinal detachment surgery.¹ It is the leading cause of failure of retinal detachment surgery and usually results in visual loss.¹ The PVR tissue contains myofibroblasts derived from retinal pigment epithelial (RPE) cells and other cell types, such as glial cells.^1,2 Such fibrocellular tissue may then contract the retina by means of a cell-mediated process and ultimately reduce the flexibility of the detached retina.^2-5 The RPE cells disseminated into the vitreous cavity and onto the retinal surface transdifferentiate into mesenchymal-like α-smooth muscle actin (α-SMA)–positive cells that produce extracellular matrix and contribute to the accumulation of fibrous scar tissue.^2-5 Such transdifferentiation (or dedifferentiation) is considered to be a process of epithelial-mesenchymal transition (EMT), a program of differentiation whereby cells lose their epithelial morphologic features and expression of epithelial markers and acquire more mesenchymal-like morphologic features in association with expression of mesenchymal markers such as α-SMA.^2-5 Although various growth factors are reportedly involved in the pathogenesis of PVR,^6-12 transforming growth factor β2 (TGF-β2) is expressed at much higher levels than the other TGF-β isoforms in the vitreous humor and is also more up-regulated in eyes with PVR.^13-16 As such, its fibrogenic effect contributes to assist the surgical closure of idiopathic macular holes.^17,18 Smad3, a major signaling transmitter for TGF-β, is essential for EMT and fibrogenic responses by RPE cells induced by retinal detachment. Its importance is based on the finding that in Smad3-null mice such responses were suppressed.¹⁹ Similarly, in the lens epithelium, injury-induced EMT is mediated by Smad3 and blocked by adenoviral gene introduction of anti–TGF-β/Smad signal molecules; that is, Smad7, Id2, Id3, or bone morphogenetic protein 7.²⁰ Moreover, Smad7 gene transfer blocks scar formation in a cornea alkali burn mouse model, resulting in restoration of its transparency.²¹ These findings prompted us to test the hypothesis that Smad7 gene transfer suppresses EMT and fibrogenic responses by RPE after retinal detachment in mice. We also tested this hypothesis in ARPE-19 cells by determining whether Smad7 gene transfection disrupted TGF-β2 signaling.

ARPE-19 cells²² were cultured in a mixture of Dulbecco modified Eagle's minimum essential medium and Hanks balanced salt solution supplemented with 10% fetal calf serum and antibiotic agents as previously reported.¹⁹

We used the Adenovirus Cre/LoxP-Regulated Expression Vector Set (code 6151; Takara Bio Inc, Tokyo, Japan) to make recombinant adenovirus–expressing mouse Smad7 as previously reported.^20,21 A mixture of recombinant adenoviruses carrying CAG (cytomegalovirus enhancer, chicken β-actin promoter plus a part of the 3′ untranslated region of rabbit β-globin) promoter-driven Cre (Cre-Ad) and LoxP–neomycin resistance gene–LoxP (LNL)–mouse Smad7 complementary DNA was applied to the targets to express Smad7 protein. Herein, Smad7-Ad refers to coinfection of Cre-Ad and Smad7-Ad. Efficacy of gene transfer was confirmed as previously reported²³ by means of transfection of green fluorescence protein–carrying adenovirus. Approximately 30% to 40% of ARPE-19 cells exhibited marked fluorescence, and the others showed low to moderate fluorescence (data not shown).

To examine whether exogenous Smad7 gene transfection induces Smad7 protein overexpression, ARPE-19 cells were seeded and grown to subconfluence in 60-mm plastic culture dishes for Western blotting. Subconfluent cells were infected with either Cre-Ad or Smad7-Ad as previously reported.^20,21 Adenoviral vector was used at a concentration of 2.0 × 104 plaque-forming units (PFU)/μL in each culture. The viral vector was allowed to infect the cells for 2 hours in serum-free medium. The cells were incubated for 56 hours before experiments. The cells were then incubated for 24 hours in serum-free medium, followed by treatment with TGF-β2, 2.0 ng/mL (R&D Systems, Minneapolis, Minn), for 48 hours. The cells were harvested in cell lysis buffer and were processed for sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blotting for Smad7, β-actin, and α-SMA as previously reported.^20,21,24

Cells grown to subconfluence in wells of chamber slides (Nalge Nunc, Naperville, Ill) were gene transfected and exposed to TGF-β2 for 48 hours as described previously herein.^20,21,24 The cells were fixed and dual immunostained as described previously herein. Goat polyclonal anti–Smad7 antibody (1: 100) (Southern Biotechnology, Birmingham, Ala) and monoclonal anti–α-SMA antibody (1:100) (Sigma-Aldrich Corp, St Louis, Mo) were used.

Immunocytochemical analyses for phosphorylated Smad2/3 and Western blotting for phospho-Smad2/total Smad2 were conducted to evaluate the effect of Smad7 overexpression on TGF-β2/Smad signal. The cells were seeded in the wells of a chamber slide (Nalge Nunc) or in 60-mm plastic culture dishes for immunocytochemical analysis or Western blotting, respectively. Subconfluent cells received gene transfer and TGF-β2 treatment as described previously herein.^20,21,24 The cells on a chamber slide were fixed in 4.0% paraformaldehyde, and those in 60-mm dishes were harvested in cell lysis buffer. The specimens were then processed for immunofluorescence staining or sodium dodecyl sulfate–polyacrylamide gel electrophoresis and Western blotting using monoclonal anti–phosphorylated Smad2 (Ser465/467) antibody (Chemicon, Temecula, Calif) and polyclonal anti–phosphorylated Smad3 (Ser423/425) antibody (Biosource International, Camarillo, Calif) as previously reported.^20,21,24 Antibodies were diluted 1:100 in phosphate-buffered saline solution for immunocytochemical analysis and 1:1000 for Western blotting. In immunocytochemical analysis, after a wash in phosphate-buffered saline solution the specimens were treated with fluorescein- or rhodamine-conjugated secondary antibodies and were mounted in a medium containing 4′-6-diamidino-2-phenylindole nuclear dye (VectaShield H-1200; Vector Laboratories, Burlingame, Calif). Negative control immunocytochemical analysis was performed using nonimmune IgG.

Real-time reverse transcriptase–polymerase chain reaction (RT-PCR) for TGF-β1and collagen type Ia2 messenger RNAs and enzyme immunoassay for collagen type I were conducted to examine the effects of Smad7 gene transfection on the expression of these entities. The cells were grown to subconfluence in 60-mm plastic culture dishes (for RNA extraction) or in wells of 12-well plates (for medium harvest) and received gene transfer and TGF-β2 treatment for 24 hours (for real-time RT-PCR) or 48 hours (for immunoassay) as described previously herein.^18,19,22 RNA was extracted from the cells and was processed for real-time RT-PCR for TGF-β1 and collagen Iα2 messenger RNAs as previously reported.^20,21,24 Medium from 48-hour cultures was processed for collagen type I C-terminal peptide enzyme-linked immunosorbent assay (Takara Bio Inc) according to the protocol provided by the manufacturer.²⁴ Three dishes (for real-time RT-PCR) or wells (for enzyme-linked immunosorbent assay) were prepared to obtain each measure.

Cells in wells of chamber slides were transfected and exposed to TGF-β2. They were then immunostained as described previously herein. Goat polyclonal anti–type I collagen antibody (1:100) (Southern Biotechnology) and monoclonal anti–fibronectin antibody (1:100) (Santa Cruz Biotechnology, Santa Cruz, Calif) were used.

All the experimental procedures were approved by the DNA Recombination Experiment Committee and the Animal Care and Use Committee of Wakayama Medical University and were conducted in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

The PVR model was established in the right eye of each mouse (n = 48) as described elsewhere.^19,23 Briefly, with the mouse under general anesthesia (intraperitoneal pentobarbital sodium), a linear incision was made in the central cornea, and the crystalline lens was removed. After the vitreous cavity was filled with 1.0% hyaluronan, the retina was gently detached using forceps without damaging the pigment epithelium. The corneal incision was sutured using 10-0 nylon string. The left eye served as the uninjured control.

Because we had confirmed that adenoviral genes were transferred to RPE cells in vivo by observing specimens treated with green fluorescence protein–Ad as previously reported,²³ 5 μL of Cre-Ad (n = 24) or Smad7-Ad (n = 24) was administered into the vitreous cavity using a hypodermic needle introduced through the sutured corneal incision. Each eye received viral vectors at a concentration of 2.0 × 107 PFU/μL. Ofloxacin ointment was instilled into the eye, and the mice were allowed to heal. On day 5, 10, or 21 after treatment the mice were humanely killed by means of carbon dioxide asphyxia and cervical dislocation; each eye was enucleated, fixed in 4% paraformaldehyde in 0.1M phosphate buffer, and embedded in paraffin. The number of animals (eyes) examined in each experimental condition at each time point was 6.

Deparaffinized 5-μm sections of eyes were processed for hematoxylin-eosin staining and indirect immunohistochemical analysis with mouse monoclonal anti–α-SMA antibody (1:100) (Neomarker, Fremont, Calif), anti–collagen I antibody, or nonimmune IgG (control). Rhodamine- or fluorescein-conjugated secondary antibodies were used to visualize the specific immunoreaction as previously reported.²³

Western blotting detected marked protein expression of Smad7 in cells that had received Smad7-Ad (Figure 1A). Exogenous TGF-β2 up-regulated protein expression of α-SMA, the major marker of EMT (Figure 1A). Western blotting showed that in control cells transfected with the Ad-Cre gene, TGF-β2 induced an increase in α-SMA expression that was similar to that induced by this cytokine in cells transfected with Smad7-Ad (Figure 1A). Dual immunohistochemical analysis showed that approximately 30% to 40% of the cells were labeled with anti–Smad7 antibody in Smad7-Ad–transfected cells, whereas such immunoreactivity was not observed in control culture. In the presence of TGF-β2, many cells were positive for α-SMA in Cre-Ad culture, whereas faint α-SMA was detected in cells without Smad7 immunoreactivity in Smad7-Ad culture (Figure 1B).

To examine whether Smad7 blocks Smad2/3/4 signaling, we then examined the effect of Smad7 overexpression on phosphorylation of Smad2 after TGF-β2 addition. Western blotting showed suppression of protein expression of phospho-Smad2 in cells with Smad7-Ad compared with control (Cre-Ad–treated) cells (Figure 2A and B). Nuclear accumulation of phospho-Smad2/3 peaked 1 hour after TGF-β2 addition in control (Cre-Ad–treated) cells. In cells that had received Smad7-Ad, cytoplasmic expression and nuclear accumulation of phospho-Smad2 and phospho-Smad3 were suppressed, but not completely blocked, in some cells (Figure 2C and D).

Treatment of ARPE-19 cells for 48 hours with exogenous TGF-β2 caused a significant increase in TGF-β1 expression and induced collagen Iα2 messenger RNA expression (Figure 3A). Smad7 overexpression partially suppressed these increases. Enzyme-linked immunosorbent assay also showed that Smad7 gene transfer counteracted the up-regulation of collagen I protein (Figure 3B). Immunocytochemical analysis also showed that Smad7 overexpression suppressed deposition of collagen I (Figure 3Ca and c) and fibronectin (Figure 3Cb and d).

To address a possible role for Smad7 overexpression in the pathogenic response of RPE cells to retinal detachment we developed a new mouse model of retinal detachment based on a variation of a previously published model.^19,23 A previous study showed that 5 days is required for the development of EMT and the appearance of α-SMA–positive myofibroblasts in the RPE cell layer after retinal detachment. On days 5 (Figure 4Aa), 10 (Figure 4Ac), and 21 (Figure 4Ae) after retinal detachment, RPE cells in the posterior pole region became multilayered in control (Cre-Ad–treated) eyes, whereas cells in Smad7-Ad–treated eyes retained their monolayer organization (Figure 4Ab, d, and f). Fibrous tissues containing multilayered RPEs were formed only in Cre-Ad–treated eyes and never in Smad7-Ad–treated eyes throughout the healing period. Although it is likely that multilayered cells with pigmented cytoplasm seen in fibrous tissues represent RPEs that have undergone EMT, this cannot be determined from histologic examination of hematoxylin-eosin–stained sections. Also, there was no morphologic abnormality in corneal endothelium outside the wound incision (data not shown).

Because the histologic findings suggested perturbed EMT by Smad7 in response to retinal detachment, we examined whether the expression of α-SMA, a hallmark of EMT and of acquisition of a myofibroblast phenotype,^6-11,19,24 would also be reduced or absent in eyes treated with Smad7 overexpression. On days 5 (Figure 4Ba), 10 (Figure 4Bc) and 21 (Figure 4Be), elongated, pigmented, fibroblast-like multilayered cells under the detached retina, likely derived from RPE cells, were labeled by the anti–α-SMA antibody in control mice, whereas monolayer cells of Smad7 overexpression–treated mice were negative for α-SMA throughout the interval examined (Figure 4Bb, d, and f). Collagen type I, the major component of fibrous tissue of PVR, was detected in cell multilayer formed on the Bruch membrane in a control eye on day 21 (Figure 4Ca), whereas such immunoreactivity was not seen in an Smad7 gene transfer–treated eye (Figure 4Cb).

We examined whether overexpressed Smad7 affects Smad2/3/4 signaling in vivo in RPE cells after retinal detachment by performing immunohistochemical analysis for phospho-Smad2 and phospho-Smad3. Marked Smad7 immunoreactivity was detected in RPE cells of Smad7-Ad–treated eyes (Figure 5B and H) compared with RPE cells in control (Cre-Ad–treated) eyes (Figure 5A and G) at days 5 and 10. However, Smad7 protein expression levels in RPE cells were similar to each other between control (Figure 5M) and treated (Figure 5N) eyes. As for Smad signaling expression, nuclear accumulation of phospho-Smad2 was observed in RPE cells throughout the interval examined, with a peak on day 10 in control eyes (Figure 5C, I, and O). On the other hand, this immunoreactivity was barely detected in Smad7-Ad–treated eyes (Figure 5D, J, and P). The expression pattern of phospho-Smad3 was similar to that of phospho-Smad2 (Figure 5E, F, K, L, Q, and R).

The present study was undertaken to examine whether Smad7 gene transfection, which inhibits TGF-β/Smad2/3/4 signaling, could be used as a new strategy for the prevention and treatment of PVR, similar to previous results with Smad3-null mice.¹⁹ Herein we showed that in vivo EMT of RPE cells and induction of collagen type I expression, both major components of PVR development after retinal detachment, are inhibited by Smad7 overexpression.

Although efficiency of adenoviral gene transfer seemed less in ARPE-19 cells compared with in vivo RPE cells, Smad7 gene transfection also exhibited an antifibrogenic reaction in this cell line, which further supports the present in vivo findings. Immunocytochemical analysis showed that Smad7 overexpression attenuated the formation of α-SMA–labeled contractile cytoskeletal fibers in ARPE-19 cells in response to the addition of TGF-β2, although Western blotting did not reveal any change in the protein expression level of α-SMA. Smad2 signaling is reportedly involved in the expression of α-SMA in fibroblasts.²⁵ Although Western blotting showed that exogenous Smad7 reduced, but did not completely block, the expression of phospho-Smad2/3 in a total cell lysate, the protein expression level of α-SMA was not affected, and only the formation of α-SMA–labeled cytoskeletal fiber was inhibited. The efficacy of adenoviral gene transfer did not reach 100% (approximately 30%-40%) as determined by comparing green fluorescence protein gene expression and Smad7 immunostaining. In some cells, phospho-Smad2/3–positive cells were observed even with Smad7 overexpression. Immunocytochemical analysis revealed α-SMA–labeled cytoskeleton in ARPE-19 cells with less exogenous Smad7 protein expression. Nevertheless, again, such impaired formation of α-SMA–labeled fiber was not due to the reduction of total protein levels as detected by Western blotting. Smad7 gene transfection might not be sufficient to inhibit cofilin, LIM kinase, and slingshot, which are essential to α-SMA fiber formation. A similar finding was reported in cultured hepatic stellate cells, the main component of mesenchymal cells involved in liver fibrosis; Smad7 gene transfer blocks formation of α-SMA–labeled cytoskeleton, but its protein expression level was not altered.²⁶ On the other hand, expression of TGF-β1, collagen type I, and fibronectin in response to exogenous TGF-β2 was suppressed by Smad7 overexpression.

In the present study, Smad7 overexpression almost completely blocked development of α-SMA–positive cytoskeleton and collagen type I expression by RPE cells after retinal detachment in vivo, whereas in vitro Smad7 overexpression partially inhibited up-regulation of collagen type Iα2 in ARPE-19 cells. Most RPE cells in vivo after retinal detachment were Smad7–positive labeled on day 5, whereas ARPE-19 cells were not consistently showing such expression. In the in vivo experiment, on day 21 after retinal detachment, Smad7 expression was similar in the control and gene transfer groups. Nevertheless, the expression level of phospho-Smad2/3 was higher in control eyes compared with eyes injected with the Smad7-containing virus. This might be because suppression of the fibrogenic process in the earlier phase after retinal detachment on days 5 and 10 might result in deactivation of the intraocular cells on day 21.

Although many animal PVR models have been used to identify the pathogenic mechanisms of this disease,^27-33 we tried to mimic spontaneous EMT of RPE cells after retinal detachment. Injection of cultured fibroblasts into the vitreous induces PVR-like lesions in the animal eye.³¹ This model is based on the fact that RPE cells undergo EMT and behave like fibroblasts during the process of PVR. Other models of PVR are based on transgenic mice in which overexpression of various growth factors in the photoreceptor cells leads to traction detachment of the retina.^25,26 These approaches are not suitable because forced alteration of the cytokine expression profile likely affects the signal transduction system in RPE cells. Induction of PVR by means of an intravitreal injection of dispase is likely to stimulate retinal glial cells to migrate toward the subretinal space.³³ The PVR model used herein has been modified from one originally described elsewhere.³⁴ In the present model the fibrogenic response by RPE cells was observed mainly on the Bruch membrane but not on the surface of the detached retina. This is a disadvantage of the present model; the RPE cells are confined to the subretinal space and do not migrate into the vitreous as in PVR. Nevertheless, the exposure of RPE cells to subretinal fluid in a model of retinal detachment induced by de novo expression of α-SMA (occurrence of EMT), leading ultimately to fibrosis, simulates the key features of the human disease PVR.²

We previously reported that blocking p38 mitogen-activated protein kinase by a chemical inhibitor or gene transfer of a dominant negative form inhibits the fibrogenic response of RPE cells induced by TGF-β2, as well as in vivo EMT and fibrogenesis of RPE cells in response to retinal detachment in mice.²³ Although it has not been determined if blocking p38 signaling affects TGF-β/Smad 2/3/4 signaling, it is known that TGF-β–derived non-Smad mitogen-activated protein kinases affect Smad signaling via interference of phosphorylation in the Smad middle linker regions.^35-38 Thus, gene transfer of dominant negative p38 mitogen-activated protein kinase might augment the effects of Smad7 overexpression.

These results show a specific signaling pathway that might be an important new target for design of therapeutics against PVR. We showed that the PVR reaction is blocked by Smad7 gene transfer in mice. A clinical application could be developed using ocular viral or nonviral vectors that carry Smad7 complementary DNA.

Correspondence: Shizuya Saika, MD, PhD, Department of Ophthalmology, Wakayama Medical University, 811-1 Kimiidera, Wakayama 641-0012, Japan (shizuya@wakayama-med.ac.jp).

Submitted for Publication: July 18, 2006; final revision received August 25, 2006; accepted September 18, 2006.

Financial Disclosure: None reported.

Funding/Support: This study was supported by grant 15591871 from the Ministry of Education, Science, Sports, and Culture of Japan and by the Uehara Memorial Foundation (Dr Saika).

Acknowledgment: We thank the late Anita B. Roberts, PhD, for her continuous help, support, and encouragement in the authors' projects on TGF-β/Smad and tissue fibrosis for the past 4 years.