A Role for Connective Tissue Growth Factor in the Pathogenesis of Choroidal Neovascularization

Objective  To evaluate the expression of connective tissue growth factor (CTGF) in choroidal neovascular membranes from patients with age-related macular degeneration and the effect of CTGF on choroidal endothelial cell (CEC) function.

Methods  Using immunohistochemical methods, we analyzed CTGF expression in 13 surgically excised choroidal neovascular membranes related to age-related macular degeneration. The expression of CTGF in retinal pigment epithelial and CEC cultures was determined by means of reverse transcriptase polymerase chain reaction and Western blot, and its regulation by vascular endothelial growth factor and transforming growth factor β was determined. The effects of CTGF on bovine CEC proliferation, attachment, migration, and tube formation were measured.

Results  Vascularized human choroidal neovascular membranes showed strong CTGF immunoreactivity. Double staining disclosed colocalization of CTGF with retinal pigment epithelial cells and CECs. The CTGF induced a significant increase in attachment and migration of CECs; however, it did not stimulate CEC proliferation. The CTGF protein was up-regulated in retinal pigment epithelial cells and CECs by stimulation with transforming growth factor β and vascular endothelial growth factor, respectively.

Conclusions  The expression of CTGF in choroidal neovascular membranes, its regulation by angiogenic growth factors, and its proangiogenic effects on CEC function suggest that CTGF may play a role in the pathogenesis of choroidal neovascularization.

Clinical Relevance  Multiple growth factors are involved in the pathogenesis of choroidal neovascularization in age-related macular degeneration.

AGE-RELATED MACULAR degeneration is the leading cause of lost reading vision in the elderly population.¹ The exudative form of the disease, characterized by choroidal neovascularization, is the major cause of severe vision loss in this disorder.^1,2 The pathogenesis of choroidal neovascularization is a complex process that involves growth factor–mediated alterations in choroidal endothelial cell (CEC) function and remodeling of the extracellular matrix, resulting in the development of a fibrovascular subretinal membrane.^3-6

Experimental evidence suggests that retinal pigment epithelial (RPE) cells play a key role in the development of choroidal neovascularization.^7-9 A number of angiogenic factors produced by RPE and localized to human choroidal neovascular membranes(CNVMs) have been identified, including fibroblast growth factor, platelet-derived growth factor, vascular endothelial growth factor (VEGF), transforming growth factor β (TGF-β), angiopoietin 1, and angiopoietin 2.^9-13

The secreted, cysteine-rich polypeptide connective tissue growth factor(CTGF) was originally identified from conditioned medium of human umbilical vein endothelial cells.¹⁴ It is expressed in a variety of cells, including fibroblasts, epithelial cells, and smooth muscle cells. It is highly expressed in active endothelial cells and vessels, as well as in scars, fibrotic lesions, and atherosclerotic lesions.^15,16 This angiogenic factor was recently found to induce endothelial cell migration, proliferation, and tube formation and to increase endothelial cell attachment.¹⁷ Application of recombinant CTGF in rat and chicken embryos induces angiogenic responses in vivo.^17,18 Antisense deoxynucleotides directed against CTGF inhibit endothelial cell tube formation, ¹⁹ and anti-CTGF antibody blocks in vivo angiogenesis induced by recombinant CTGF.¹⁹ This suggests that CTGF could play an important role in angiogenesis by regulating endothelial cell function.

The role of CTGF in ocular angiogenesis is currently under investigation. Recent publications report that VEGF up-regulates CTGF messenger RNA (m RNA)expression in retinal endothelial cells.^20,21 However, to the best of our knowledge, there have been no reports about the role of CTGF in the pathogenesis of choroidal neovascularization and no studies of CTGF effects on CECs. It is important to specifically study CECs, since endothelial cell heterogeneity is well described in vivo and in vitro, and there are marked functional differences between CECs and endothelial cells from other regions.^22-25

In the present study, we show that CTGF expression is present in CNVMs of patients with age-related macular degeneration. We then demonstrate that CTGF m RNA is expressed in RPE cells and CECs and that the production of CTGF protein is up-regulated in RPE cells and CECs by stimulation with TGF-β and VEGF, respectively. Furthermore, we show that CTGF induces proangiogenic responses in CECs in vitro.

Thirteen surgically excised subfoveal CNVMs from patients with age-related macular degeneration (aged 66-85 years) and 3 normal adult postmortem donor retinas (from the Lions Doheny Eye Bank, aged 55-97 years) were prepared for immunostaining. The study was approved by the institutional review board of the University of Southern California, Los Angeles. Five of the membranes were predominantly vascular, 5 were predominantly fibrotic, and 3 were mixed in appearance. Tissues were snap frozen and sectioned at 6 µm with the cryostat. Thawed tissue sections were air dried, rehydrated with phosphate-buffered saline (p H 7.4), and blocked with 5% normal goat serum for 15 minutes. Sections were incubated with anti-CTGF polyclonal antibody (Fibro Gen, Inc, South San Francisco, Calif) for 60 minutes and then treated with biotinylated secondary anti–rabbit antibody (1:400; Vector Laboratories, Burlingame, Calif) followed by streptavidin peroxidase. The red color was developed with an aminoethyl carbazole kit (Zymed Laboratories, Inc, South San Francisco). After each step, sections were given three 5-minute washes with phosphate-buffered saline. Slides were counterstained with hematoxylin and mounted with glycerin-gelatin medium. Controls for the immunoperoxidase studies included (1) staining of adjacent tissue sections by means of an identical protocol except that the primary antibody was omitted and (2) staining of adjacent tissue sections with an identical protocol with the use of antiserum preadsorbed with recombinant CTGF. In both controls, negligible background or nonspecific staining was found (Figure 1C and F).

Thawed tissue sections were fixed for 10 minutes with 10% neutral buffered formalin (Polysciences, Inc, Warrington, Pa), given three 5-minute washes with Tris hydrochloride buffer (p H 7.4), and blocked with 5% goat serum for 15 minutes. After cells were immunostained for CTGF as described above, the slides were washed with the Tris buffer 3 times. Primary antibodies were added to cover the tissue, and the slides were incubated for 1 hour at room temperature. The sections were washed 3 times with Tris buffer, and secondary alkaline phosphatase anti–mouse antibody (Vector Laboratories) was added for 30 minutes. After another triple wash with Tris buffer, an alkaline phosphatase chromogen substrate (Vector Blue; Vector Laboratories) was added to the slides for 15 minutes for color development.

Human RPE cells were obtained from human fetal eyes (18-26 weeks of gestation; Advanced Bioscience Resources, Inc, Alameda, Calif).²⁶ Second- through fourth-passage cells were used for all experiments. The RPE cells were cultured in Dulbecco modified Eagle medium (Fisher Scientific, Pittsburgh, Pa) supplemented with 10% fetal bovine serum (Gibco BRL, Gaithersburg, Md), 2m M glutamine, 100-µg/mL streptomycin, and 100-U/mL penicillin (Sigma-Aldrich Corp, St Louis, Mo). The purity of cultured RPE cells was confirmed by means of immunocytochemical staining; more than 95% of cells were positive for cytokeratin(DAKO, Carpinteria, Calif), while no cells were positive for the endothelial cell antigen von Willebrand factor (DAKO) or macrophage antigen CD11c (DAKO). Magnetic beads carrying the specific endothelial marker Lycopersicon esculentum (Sigma-Aldrich Corp) were used to isolate CECs from bovine eyes, as previously described.²⁷ Positive immunostaining for von Willebrand factor and binding of diacetylated low-density lipoprotein confirmed that the CECs were vascular endothelial cells. Although it is likely that the distinct functional characteristics of CECs may be lost after prolonged passage, previous work has shown that low-passage endothelial cells of diverse origin retain functional and structural differences.²⁵ In our experiments, all CECs were in passage 2 or 3.

Poly A (A)+RNA was isolated by means of a kit (Fast Track; Invitrogen Corp, Carlsbad, Calif). First-strand complementary DNA was synthesized at 42°C with Moloney murine leukemia virus reverse transcriptase (Gibco BRL). The polymerase chain reaction was initiated in a DNA thermal cycler (Perkin-Elmer/Cetus, Norwalk, Conn) with CTGF primer pairs (human CTGF: GCATCCGTACTCCCAAAATC [sense] and CTTCTTCATGACCTCGCCGT [antisense]; bovine CTGF: GATCATAGGACTGTATTAGTGC[sense] and CTGACTTAGAGACGAACTTG [antisense]). After 30 cycles of polymerase chain reaction, the products were resolved on a 1.2% agarose gel and stained with ethidium bromide, and the gel was photographed under ultraviolet illumination.

The RPE cells and CECs were grown in 6-well plates and starved for 24 hours in Dulbecco modified Eagle medium or essential growth media (Bio Whittaker, Walkersville, Md) with 0.1% bovine serum albumin. The medium was then removed and replaced with fresh medium. For stimulation studies, the CEC replacement medium included 1% fetal bovine serum and recombinant VEGF (10-50 ng/mL); for RPE cells, the medium was serum-free and included TGF-β₂ (1-30 ng/mL). Lysed cells were collected after 48 hours of additional incubation; proteins were resolved on Tris hydrochloride 10% polyacrylamide gels (Ready Gel; Bio-Rad Laboratories, Hercules, Calif) at 120 V, with 8 µg of protein added to each lane. The proteins were transferred to a polyvinylidene fluoride blotting membrane (Millipore, Billerica, Mass); these membranes were then probed with polyclonal anti-CTGF antibody (Fibro Gen, Inc), followed by horseradish peroxidase–conjugated goat anti–rabbit antibody (Vector Laboratories) for 30 minutes at room temperature. Images were developed by adding chemiluminescence detection solution (ECL; Amersham Pharmacia Biotech, Piscataway, NJ).

The RPE cells or CECs (10⁵/mL) were trypsinized and resuspended in cell culture medium with 0.4% fetal bovine serum. After a 48-hour treatment with CTGF, 100 mL of cell suspension (10⁴ cells) was added to each well of a 96-well fibronectin-coated plate (Becton, Dickinson Labware, Bedford, Mass), and cells were allowed to attach for 60 minutes.²⁶ The cells were gently washed twice with phosphate-buffered saline, and fresh medium(150 mL) was added to each well with 3-[4, 5-dimethylthiazol-2-yl]-2, 5-diphenyltetrazolium bromide, 5 mg/mL, 20 mL (Sigma-Aldrich Corp). After a 5-hour incubation, the supernatants were decanted; the formazan precipitates were solubilized by adding 150 mL of 100% dimethyl sulfoxide (Sigma-Aldrich Corp) and placed on a plate shaker for 10 minutes. Absorbance at 550 nm was determined on a microplate reader (Benchmark Microplate Reader; Bio-Rad Laboratories).

Migration of RPE cells and CECs was measured by means of a modified Boyden chamber assay in 24-well plates with fibronectin-coated inserts.²⁶ Recombinant CTGF, 10 to 50 ng/mL (Fibro Gen, Inc), was added to the lower chamber as the stimulant. After a 5-hour incubation, the inserts were washed 3 times with phosphate-buffered saline, fixed with cold (4°C) methanol for 10 minutes, and counterstained with hematoxylin for 20 minutes. The number of migrated cells was counted by means of phase-contrast microscopy (×320). Four randomly chosen fields were counted per insert.

Two methods of measuring cell proliferation were used. Subconfluent CECs grown in 6-well plates were treated with recombinant CTGF (0, 1, 10, 50, and 100 ng/mL) for 48 hours. Cell proliferation was measured by cell counting of representative triplicate samples with a hemocytometer, and by tritiated thymidine uptake assay as previously described.²⁸

The CECs were subcultured on fibronectin-coated, 6-well plates. The original medium was replaced with essential growth media containing 1% fetal bovine serum and 20-ng/mL CTGF, and cells were incubated for 9 days. Tube formation was monitored every day by phase-contrast microscopy.

All 13 CNVMs stained positively for CTGF, with the most prominent staining found in the vascularized regions of the membranes (Figure 1A). Portions of the intact RPE monolayer were found in 5 membranes; occasional RPE in these regions stained strongly positive for CTGF(Figure 1D). Fibrotic regions of the membranes showed focal CTGF staining in the fibroblastic stromal cells(Figure 1C). Double staining of the CTGF-positive cells in 4 membranes disclosed that many of the CTGF-positive cells were also cytokeratin positive, indicating that these cells were transdifferentiated RPE cells (Figure 1B). A smaller number of CTGF-positive cells were endothelial cells, as confirmed by double labeling with CD-31 and CTGF (Figure 1E). A series of 3 normal adult retinas were also stained for CTGF, and no expression was found in the RPE or CECs (results not shown). Specificity of the CTGF antiserum was demonstrated by staining serial sections with CTGF antiserum(Figure 1C) and antiserum preadsorbed with recombinant CTGF (Figure 1F); staining with preadsorbed antiserum resulted in negligible backgound staining.

The expression of CTGF m RNA was demonstrated in cultured human RPE cells and bovine CECs by means of reverse transcriptase polymerase chain reaction with species-specific primer sets (Figure 2). As predicted by the complementary DNA sequences, the human CTGF polymerase chain reaction product was 210 base pairs (bp), while the bovine product was 220 bp. When complementary DNA was omitted from the reaction mixture, no product was obtained (negative control; result not shown).

Cultures of both human RPE and bovine CECs expressed CTGF protein that was identified on Western blot as a doublet of 36 to 38 k Da (Figure 3). Stimulation of RPE cells with TGF-β₂ for 48 hours at concentrations of between 1 and 30 ng/mL caused a marked increase of CTGF protein expression (Figure 3A), with the greatest effect at 1 and 20 ng/mL. In contrast, VEGF stimulation had no effect on RPE production of CTGF (Figure 3B). In CECs, VEGF stimulation for 48 hours at concentrations between 10 and 50 ng/mL resulted in increased expression of CTGF protein with a maximal effect at 50 ng/mL (Figure 3D). In contrast, culture of CECs with TGF-β did not stimulate expression of CTGF in CECs (Figure 3C).

The migration of CECs in the modified Boyden chamber assay was stimulated by recombinant CTFG (Figure 4). The stimulation was concentration dependent, with the maximal response to CTGF at a concentration of 30 ng/mL in the medium (P<.05). Interestingly, the maximal CTGF response was as effective as VEGF (10 ng/mL) in stimulating CEC migration. The dose of VEGF shown is submaximal and the CTGF effect is significantly less than a maximal VEGF stimulation. Stimulation of CEC with CTGF (10-100 ng/mL) for 48 hours also increased attachment of the cells to fibronectin (P<.05) (Figure 5). The effect was dose dependent, with a maximal effect at 50 to 100 ng/mL. Stimulation of CECs with CTGF (1-100 ng/mL) for 48 hours produced no proliferative response as measured by thymidine uptake or by cell counting (results not shown). When CECs were plated on fibronectin for 9 days in the presence of 1% serum, there was minimal, if any, tube formation (Figure 6A). In the presence of CTGF, 20 ng/mL, the attached endothelial cells extended cell processes, formed cell-cell contact, and established a branched network of tubelike structures (Figure 6B).

Growth factors play an important role in the development, progression, and regression of pathological angiogenesis. In choroidal neovascularization, studies have demonstrated the presence of a complex mixture of angiogenic and antiangiogenic molecules, whose interactions likely result in the neovascular response.^9-13,29 In this study, we demonstrate that our series of CNVMs uniformly express CTGF and that this expression is most prominent in the vascular regions of the membrane. The expression of CTGF is localized to stromal RPE and CECs as well as RPE in the residual monolayer. The pattern of immunoreactivity is similar to that seen previously for VEGF and provides support for the central role of RPE and their secreted growth factors in the pathogenesis of the disorder.⁹ No CTGF immunoreactivity was found in the RPE or CECs of normal adult retinas, consistent with previous reports that CTGF is undetectable in normal vessels, ^30,31 but high expression levels are found in endothelial cells of atherosclerotic lesions and other angiogenic sites.³² The presence of CTGF in the residual RPE monolayer of CNVMs suggests the possibility that CTGF has the potential to activate normal underlying choroidal endothelium.

This article demonstrates for the first time, to our knowledge, that RPE and CECs each express CTGF m RNA and protein in vitro. There is considerable evidence that CTGF is a major downstream effector of TGF-β action in many cell types, ^15,16,33,34 and TGF-β clearly up-regulates CTGF protein expression in RPE. The effect of TGF-β on endothelial cell subtypes, however, is more controversial. The expression of CTGF is up-regulated by TGF-β in periaortic, omental, and retinal endothelial cells, but not pulmonary or aortic endothelial cells.³⁵ In this study, we saw no increase in CTGF protein when bovine CECs were stimulated with TGF-β. The mechanism of the differential response to TGF-β in various cell types has not been defined; however, induction of CTGF by TGF-β in fibroblasts is largely at the level of transcription initiation.³⁶

The regulation of CTGF by VEGF in endothelial cells is also controversial. Although VEGF up-regulates CTGF in bovine retinal endothelial cells, there was no such up-regulation in retinal vascular endothelial cells from Macaca mulatta.^20,22 Our results clearly show that VEGF stimulation of bovine CEC results in increased CTGF protein expression. The possibility that other factors play a role in stimulating CTGF expression should also be considered, since CTGF has also been shown to be regulated by advanced glycosylation end products, mechanical stress, or hypoxia.^37-39

In this report, we show that CTGF stimulates CEC migration, adhesion to fibronectin, and tube formation. One of the characteristic features of CTGF is its induction of a fibrotic response, typified by the increased production of extracellular matrix molecules by fibroblasts.³³ In support of this possibility, CTGF m RNA was recently demonstrated by in situ hybridization in transdifferentiated RPE cells in association with type 1 collagen in an epiretinal fibrovascular membrane.⁴⁰

Our results demonstrate that CTGF is expressed by RPE and CECs, and that CTGF has the potential to stimulate choroidal angiogenesis through regulation of CEC function. These results, together with the observation that CTGF is prominently expressed in CNVM, suggest that CTGF may participate in the pathogenesis of CNVM in age-related macular degeneration.

Corresponding author: David R. Hinton, MD, FRCPC, Department of Pathology and Ophthalmology, 2011 Zonal Ave, HMR 209, Los Angeles, CA 90033 (e-mail: dhinton@usc.edu).

Submitted for publication October 22, 2002; final revision received April 10, 2003; accepted May 14, 2003.

This study was supported in part by a grant from Prevent Blindness America, New York, NY; the Arnold and Mabel Beckman Foundation, Irvine, Calif; grant EY03040 from the National Institutes of Health, Bethesda, Md; and an unrestricted grant to the Doheny Eye Institute from Research to Prevent Blindness Inc, New York.

Fibro Gen, Inc provided CTGF reagents used in this study.

We thank Noelynn Oliver, PhD, for her critical review of the manuscript, Christine Spee for isolation and culture of the ocular cells, and Ernesto Barron for preparation of the figures. We appreciate the editorial assistance of Susan Clarke and Thomas E. Ogden, MD, PhD.