Transforming Growth Factor-β2 and Connective Tissue Growth Factor in Proliferative Vitreoretinal Diseases: Possible Involvement of Hyalocytes and Therapeutic Potential of Rho Kinase Inhibitor

Recombinant human TGF-β2 and CTGF were purchased from Sigma-Aldrich (Tokyo) and PeproTech (London), respectively. Y27632, SB203580, PD98059, and lysophosphatidic acid were purchased from Calbiochem. A mouse monoclonal antibody against Smad4 was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal antibodies against phospho-p38 MAPK, p38 MAPK, phospho-p44/p42 MAPK, and p44/p42 MAPK were obtained from Cell Signaling. Horseradish peroxidase–conjugated secondary antibody was obtained from Bio-Rad (Hercules, CA). Fasudil, a potent and selective Rho-kinase inhibitor, was kindly provided by Asahi Kasei Pharma (Tokyo).

This study was carried out with approval from an institutional review board and performed in accordance with the ethical standards of the 1989 Declaration of Helsinki. We obtained written informed consent from the patients. Vitreous samples were collected from patients who underwent pars plana vitrectomy because of nonproliferative vitreoretinal diseases (macular hole and rhegmatogenous retinal detachment [RRD]) and proliferative vitreoretinal diseases (PDR and PVR). Concentrations of TGF-β2 in the vitreous were measured by a human TGF-β2 immunoassay kit (R&D Systems) according to the manufacturer’s protocols. To detect CTGF in the vitreous, we performed sandwich enzyme-linked immunosorbent assays with two different anti-human CTGF antibodies, as previously reported (36). Enzyme-linked immunosorbent assay plates (96 wells; Falcon) were coated with 100 μl of 1 μg/ml goat polyclonal anti-human CTGF antibody, which had an epitope within the internal region of CTGF (Santa-Cruz), in 0.1 mol/l carbonate (pH 9.5) overnight at 4°C. After washing, unbound sites were blocked by incubation with 200 μl of PBS containing 1% BSA for 2 h at 37°C. After washing with 200 μl of PBS three times, 100 μl of vitreous samples and recombinant human CTGF as standard were added to the well and incubated for 2 h at room temperature. After washing three times, we added 100 μl of 1 μg/ml monoclonal anti-human CTGF COOH-terminal peptide antibody (R&D systems) in PBS containing 0.01% BSA. The plates were incubated for 2 h at room temperature and subsequently washed three times, and 100 μl of 1:50 dilution EnVision+ (DakoCytomation, Carpinteria, CA), anti-mouse horseradish peroxidase–labeled polymer, with PBS containing 0.01% BSA was added. After a 1-h incubation, plates were washed three times and incubated with 100 μl of OPD (1,2-phenylenediamine, dihydrochloride) solution for 30 min in the dark, and the intensity of the wells was determined at 490 nm (Immuno Mini NJ-2300; Nalge Nunc International).

Bovine eyes were obtained from a local abattoir. The posterior part of the vitreous body was extracted and washed once in Dulbecco’s modified Eagle’s medium (DMEM; Sigma). The vitreous was chopped into several pieces and incubated in type I collagen–coated dishes in DMEM containing 10% fetal bovine serum (Invitrogen-Gibco, San Diego, CA) for 1 week. The cells that proliferated on the dishes were then subcultured in type I collagen–coated dishes in DMEM supplemented with 10% fetal bovine serum. Cultured hyalocytes obtained between passages 4 and 7 that had shown no obvious morphological changes were used in the following experiment. Isolated cells were immnocytochemically identified as hyalocytes expressing S-100 protein, but neither expressed glial fibrillary acidic protein nor cytokeratin, as previously described (37). Bovine retinal pigment epithelial cells and retinal capillary endothelial cells were also isolated from bovine eyes and cultured as previously reported (38,39). MIO-M1, a cell line of the human Muller cell, was a kind gift from Dr. G.A. Limb, Division of Cell Biology, Institute of Ophthalmology, London.

Northern blot analysis was performed as previously described (38). Subconfluent hyalocytes, bovine retinal pigment epithelial cells, retinal capillary endothelial cells, and MIO-M1 under normal growth medium conditions were collected, and total RNA was extracted by the AGPC (acid guanidine thiocyanate-phenol-chloroform) extraction method. Radioactive CTGF cDNA probe was generated using a labeling kit (Multiprime; Amersham) and [³²P]dCTP (NEN, Boston, MA). Quantitation of Northern blot analysis was performed using a computing phosphorescence imager (PhosphorImager with ImageQuant software analysis; Molecular Dynamics, Sunnyvale, CA). Lane loading differences were normalized by rehybridization with rediolabeled 36Β4 cDNA probe as an internal control gene.

Nuclear extracts were isolated from hyalocytes with a CelLytic NuCLEAR extraction kit (Sigma) according to the manufacturer’s protocols. Protein samples were separated by SDS-PAGE, followed by electrophoretic transfer to nitrocellulose membranes (New England Biolabs, Beverly, MA). The blots were blocked with skim milk and incubated overnight at 4°C with antibody against Smad4 (1:1,000). After being washed with t-TBS (20 mmol/l Tris, pH 7.5, 500 mmol/l NaCl, and 0.1% Tween-20) three times for 10 min each, the membranes were incubated with horseradish peroxidase–labeled rabbit anti-goat IgG (1:4,000; Bio-Rad, Richmond, CA) for 30 min at room temperature. Visualization was performed with an enhanced chemiluminesence detection system (Amersham, Arlington Heights, IL) according to the manufacturer’s protocols.

Rho activity in hyalocytes under the presence or absence of TGF-β2 was analyzed using a Rho activation assay kit (Upstate Biotechnology). Briefly, subconfluent hyalocytes were starved and unstimulated or stimulated with TGF-β2 (0.3 nmol/l) for indicated periods of time (5, 15, 30, and 60 min), and cell lysates were prepared according to the manufacturer’s protocols. After centrifugation, supernatants were incubated with rhotekin Rho-binding peptide immobilized on agarose, and activated GTP-Rho bound to rhotekin-agarose was detected by Western blotting with anti-Rho antibody.

Hyalocytes were starved and untreated or treated with TGF-β2 (0.3 nmol/l) for the indicated time (2, 5, 15, 30, and 60 min), and the cells were lysed with 1× Laemmli buffer (50 mmol/l Tris, pH 6.8, 2% SDS, and 10% glycerol) containing protease inhibitor and phosphatase inhibitors (1 mmol/l phenylmethylsulfonyl fluoride, 1 mmol/l NaF, and 0.5 mmol/l Na₃VO₄). Western blot analysis was performed as described above with antibodies against phospho-p38 MAPK (1:1,000) and phospho-p44/p42 MAPK (1:1,000), respectively. Lane loading differences were normalized by reblotting the membranes with antibodies against p38 MAPK (1:1,000) and p44/p42 MAPK (1:1,000), respectively.

The experimental data are expressed as the means ± SD. P < 0.05 was considered statistically significant using Student’s t test in normally distributed populations. The correlation between the differences in CTGF and the differences in TGF-β2 was statistically analyzed using StatView version 5.0 (SAS Institute, Cary, NC), and statistical significance was assumed at P < 0.05.

Average concentrations of TGF-β2 protein in the vitreous were 2.63 ng/ml (macular hole, n = 15), 3.24 (RRD, n = 18), 5.09 (PDR, n = 41), and 8.54 (PVR, n = 10). Concentrations of TGF-β2 with PDR or PVR were significantly higher than those with macular hole and RRD, respectively (P < 0.01) (Fig. 1A). Average concentrations of CTGF protein in the vitreous were 38.90 ng/ml (macular hole, n = 15), 39.87 (RRD, n = 18), 56.63 (PDR, n = 41), and 69.47 (PVR, n = 10). Concentrations of CTGF in eyes with PDR or PVR were also significantly elevated compared with those with macular hole and RRD, respectively (P < 0.01) (Fig. 1B). In addition, there was a significant correlation between the concentrations of TGF-β2 and CTGF (r = 0.320, P = 0.0028, n = 84) (Fig. 1C), whereas there was no significant correlation between their concentrations in the vitreous with PDR (r = 0.238, P = 0.134, n = 41) (Fig. 1D).

As shown in Fig. 2, all of the bovine hyalocytes, bovine retinal pigment epithelial cells, bovine retinal capillary endothelial cells, and Muller cells (MIO-M1) expressed CTGF mRNA under normal culture conditions. These results indicated that various cell types could be the sources of CTGF in the eye.

We investigated whether the expression of CTGF mRNA by hyalocytes is promoted in the presence of TGF-β. The cells were starved and left unstimulated or stimulated with TGF-β2 (0.3 nmol/l) for 1, 4, 10, and 24 h, and then total RNA was extracted for Northern blot analysis. As shown in Fig. 3A, CTGF mRNA expression was promoted by TGF-β2 and reached its maximum expression level at 10 h after stimulation (4.05-fold compared with control, P < 0.01) and was maintained at least up to 24 h.

We also examined the levels of Smad4 protein in the nuclei in the presence of TGF-β2 to investigate the intracellular signaling pathway of hyalocytes. Subconfluent hyalocytes were starved and unstimulated or stimulated with TGF-β2 (0.3 nmol/l) for 15, 30, 60, 120, and 180 min. Although the amount of Smad4 protein in the nuclei was low in the quiescent cells, its concentration was significantly elevated by TGF-β2 treatment within 15 min and was maintained at least 180 min (P < 0.01) (Fig. 3B).

To examine whether TGF-β2 activates the Rho/Rho kinase, p38 MAPK, and p44/p42 MAPK pathways in hyalocytes, we investigated Rho activation, p38 MAPK phosphorylation, and p44/p42 MAPK phosphorylation in hyalocytes in the presence of TGF-β2. Levels of Rho-GTP were elevated by TGF-β2 initially after 15 min of stimulation and were sustained even after 60 min of stimulation (Fig. 4A). Phosphorylation of p38 MAPK in the presence of TGF-β2 was observed from 30 min of stimulation and maintained at least up to 60 min (Fig. 4B), whereas p44/p42 MAPK was not phosphorylated by TGF-β2 at any stimulation time from 0 to 60 min (Fig. 4C).

We confirmed the regulation of TGF-β2–dependent CTGF gene expression by the Rho/Rho kinase, p38 MAPK, or p44/p42 MAPK pathways. We examined the effect of Y27632, SB203580, and PD98059, potent and selective inhibitors of Rho kinase, p38 MAPK, and p44/p42 MAPK, respectively. As shown in Figs. 5A and B, Y27632 and SB203580 significantly suppressed nuclear accumulation of Smad4 protein and CTGF gene expression in both the absence and presence of TGF-β2 (P < 0.05). On the other hand, PD98059 had no significant effect on either nuclear accumulation of Smad4 protein or CTGF gene expression. Our results suggest that the suppressive effect of Y27632 and SB203580 on the expression of CTGF mRNA is mediated via the inhibition of nuclear translocation of Smad4 protein.

To confirm the involvement of the Rho signaling pathway in CTGF gene expression, the effects of lysophosphatidic acid, an agonist of the Rho signaling pathway, on nuclear translocation of Smad4 and the expression of CTGF mRNA were examined. Levels of Smad4 in the nuclei and expression of CTGF mRNA were significantly promoted in the presence of lysophosphatidic acid (1.71- and 1.79-fold compared with respective controls) (Figs. 6A and B), suggesting that the Rho kinase pathway lies, at least in part, upstream of Smad pathway.

We further examined the effect of fasudil, a Rho kinase inhibitor that has already been in clinical use for the inhibition of cerebrovascular spasm, on Smad pathway and CTGF mRNA expression. The concentration of Smad4 in the nuclei rose in the presence of TGF-β2, whereas the effect was significantly suppressed by fasudil in a dose-dependent manner (Fig. 7A). Similarly, CTGF mRNA expression was significantly prohibited by fasudil in a dose-dependent manner in both the absence and presence of TGF-β2 (Fig. 7B).

In the current study, we demonstrated that concentrations of TGF-β2 and CTGF are elevated in the vitreous of eyes with proliferative vitreoretinal diseases (PDR and PVR), compared with those with nonproliferative diseases (macular hole and RRD), and that their concentrations are positively correlated. We further demonstrated that hyalocytes are one of the possible sources of CTGF, with expression upregulated in the presence of TGF-β2 mediated through Rho kinase and at least partially by p38 MAPK. In addition, our results indicated that fasudil might have therapeutic potential for the treatment of vitreoretinal interface diseases via inhibition of CTGF synthesis, at least in part.

Our findings of elevated concentrations of TGF-β2 and CTGF in the vitreous of eyes with proliferative vitreoretinal diseases and their positive correlation suggest their combinational role, at least in part, in the pathogenesis of proliferative vitreoretinal diseases. Although it is known that CTGF is a downstream mediator of TGF-β in vitro, this is the first demonstration of their correlation in vivo. Even though the mean concentrations of CTGF in PDR and PVR were higher than those of macular hole and RRD, respectively, the data showed some spread. It is possible that the concentrations depend on the activity of the diseases (e.g., active or quiescent PDR). It is also possible that the concentration of CTGF in the vitreous may be elevated before or during the formation of proliferative membranes and that afterwards, at a stage of contraction of the membrane (which is often the time of removal by vitreoretinal surgery), it may be lower. CTGF is known to promote extracellular matrix production and cause fibrosis, but it is difficult to ascertain with precision the onset time of fibrosis. In addition, even in simple RRD without fibrosis, which in the current study we classified in the control group, a recent study has revealed that the concentration of CTGF in subretinal fluid is increased with time between the diagnosis of retinal detachment and surgery (40). This suggests that long-standing RRD might predispose to PVR through CTGF and that the level of CTGF can be elevated even without morphologic fibrous change yet evident. Indeed, in this study, some RRD samples also had higher concentrations of CTGF compared with lower values in PDR or PVR. Further study is necessary to fully characterize the dynamics of CTGF expression in PDR or PVR. Moreover, the correlation between the concentrations of TGF-β2 and CTGF was significant but not so strong. This might have occurred because of the three TGF-β isoforms (TGF-β1, -β2, and -β3), we investigated only TGF-β2, even though it is the predominant isoform in the posterior segment of the eye (41). In addition, when only PDR samples were analyzed, there was no significant correlation between the concentrations of TGF-β2 and CTGF, and there was a better correlation between them in the other vitreous samples with macular hole, RRD, and PVR (r = 0.400, P = 0.0074, n = 43) (data not shown), and, despite low TGF-β2 concentrations, concentrations of CTGF in several vitreous samples with PDR were high. According to some recent studies, the expression of CTGF can be induced not only by TGF-β, but also by vascular endothelial growth factor (42), hypoxia (43), or high glucose (44). So it is possible that CTGF induced by vascular endothelial growth factor, for example, released from retinal endothelial cells (42) might diffuse into the vitreous independent of TGF-β in PDR eyes.

Although it was reported that CTGF has been overexpressed in preretinal membranes of eyes with PVR or PDR, as expected (17,18), the origin of CTGF and its regulation still remain to be clarified. We demonstrated in this study that hyalocytes, bovine retinal pigment epithelial cells, retinal capillary endothelial cells, and Muller cells can be sources of CTGF. Hinton et al. (17) indicated that most CTGF-immunoreactive cells observed in PVR membrane were cytokeratin-positive cells associated with retinal pigment epithelial cells and also that most CTGF-immunoreactive cells observed in PDR membrane were smooth muscle actin–positive myofibroblast.

Recent studies have revealed that hyalocytes might be involved in the pathogenesis of vitreoretinal diseases (28,29). Our previous study revealed that hyalocytes could be transdifferentiated into myofibroblastic cells and come to express α-smooth muscle actin in the presence of TGF-β2 (9). We thus hypothesized that hyalocytes might be involved in the formation of fibrotic membrane via the expression of CTGF in proliferative vitreoretinal diseases. In the current study, we demonstrated that hyalocytes expressed CTGF and that its expression was significantly enhanced in the presence of TGF-β2, suggesting that the hyalocytes are one of the possible sources of CTGF synthesis in the vitreous cavity. Additionally, CTGF could stimulate the proliferation of cultured hyalocytes, suggesting that the hyalocytes themselves might be one of the elements of fibrous membrane (data not shown), consistent with our previous report demonstrating that transdifferentiated hyalocytes cause collagen gel contraction in the presence of TGF-β2 (9). The hyalocytes thus might participate in the initiation and progression of proliferative vitreoretinal diseases through the formation of proliferative membrane and its cicatricial contraction.

TGF-β signals through transmembrane receptor serine/threonine kinases to activate signaling intermediates called Smad proteins, which modulate the transcription of target genes (45). Eight kinds of Smad proteins were identified (Smad1–8) (46). TGF-β binding to its receptor causes Smad2/3 phosphorylation. Phosphorylated Smad2/3 then forms a complex with Smad4, and this complex translocates into the nuclei, where it regulates the transcription of target genes, such as CTGF, as a transcription factor. In the current study, we also demonstrated that TGF-β2 caused nuclear translocation of Smad4 in hyalocytes and that Smad2/3 was translocated into the nuclei after stimulation with TGF-β2 in the same manner as Smad4 (data not shown). These data suggest that enhanced expression of CTGF in the presence of TGF-β2 by hyalocytes might be caused by activation of the Smad pathway, similar to that observed in other types of cells such as fibroblasts (47) and mesangial cells (48).

Some recent studies revealed cross talk between Rho and MAPK pathways and Smad signaling downstream of TGF-β. As reviewed by Javelaud and Mauviel (49), TGF-β has been shown in numerous cell types to activate all Rho and p38 and p44/p42 MAPKs, but the activation by TGF-β is cell type–dependent, and, furthermore, the activated Rho and MAPKs might be involved in the phosphorylation of Smad proteins and regulate either Smad transcriptional activity or capacity to translocate into the cell nucleus. In this study, we demonstrated that TGF-β2 induced Rho activity and phosphorylation of p38 MAPK, but it did not phosphorylate p44/p42 MAPK in hyalocytes. Y27632 and SB203580 could partially but significantly prohibit TGF-β2–dependent nuclear accumulation of Smad4 and CTGF mRNA expression. However, PD98059 did not suppress either of them, different from many other types of cells reported, such as fibroblasts and smooth muscle cells (49). Furthermore, Nishida et al. (50) reported that Rho/Rho kinase participates in p38 MAPK activation. We also confirmed Y27632 inhibits TGF-β2–dependent phosphorylation of p38 MAPK by hyalocytes (data not shown), suggesting that the Rho/Rho kinase pathway might lie upstream of the p38 MAPK pathway in TGF-β2–dependent CTGF gene regulation.

With the prospect of a clinical application, we further demonstrated the inhibitory effect of fasudil, a Rho kinase inhibitor that has already been in clinical use for the inhibition of cerebrovascular spasm, on Smad pathway and CTGF mRNA expression. The results suggested that fasudil might suppress aberrant fibrosis of proliferative tissue through the inhibition of CTGF gene expression. In addition, our previous study revealed that fasudil inhibits collagen gel contraction by hyalocytes and thus might have an inhibitory effect on the contraction of the fibrotic membrane (9). Taken together, fasudil might have therapeutic potential for proliferative vitreoretinal diseases. However, further studies, not only in vitro but also in vivo, are necessary to more fully evaluate the therapeutic potential of fasudil for clinical use in the management of proliferative vitreoretinal diseases.

This study was supported in part by Grant-in-Aid for Scientific Research 17591839 from the Ministry of Education, Science, Sports and Culture of Japan.

We acknowledge Asahi Kasei Pharma (Tokyo) for its generous provision of fasudil.