Expression of Vascular Endothelial Growth Factor in Eyes with Coats' Disease

Abstract

Purpose.: To examine the expression of vascular endothelial growth factor (VEGF) and VEGF receptor (VEGFR)-2 in enucleated eyes with Coats' disease.

Methods.: Formalin-fixed, paraffin-embedded tissue sections from nine globes with Coats' disease were submitted for hematoxylin and eosin staining and immunohistochemistry with anti-VEGF and VEGFR antibodies.

Results.: Histologically, the enucleated eyes demonstrated the presence of macrophage infiltration and cholesterol clefts in the subretinal space. There were marked retinal vascular abnormalities, including dilated vessels with hyalinized vessel walls in six globes. Exudative retinal detachment was noted in all globes. VEGF immunoreactivity was observed in macrophages infiltrating the subretinal space, and in the detached retina including several blood vessels. VEGF-positivity in macrophages was significantly higher in cases containing retinal vessel abnormalities than those without the abnormalities (P < 0.01). VEGFR-2 immunoreactivity was detected in endothelial cells lining the abnormal retinal vessels, where VEGFR-1 or VEGFR-3 was not expressed.

Conclusions.: Immunoreactivity for VEGF and VEGFR-2 was detected in macrophages and endothelia of abnormal vessels in eyes with Coats' disease. These results suggest that anti-VEGF approach is a promising therapy for patients with Coats' disease.

Introduction

Coats' disease typically induces idiopathic retinal telangiectasias in all components of the retinal vasculature. ¹ In addition, it leads to significant lipid deposition in the outer retina and subretinal exudation due to capillary nonperfusion. A staging classification of Coats' disease has been devised as follows: stage 1 is characterized by telangiectasias, stage 2 by telangiectasia and exudation, stage 3 by exudative retinal detachment (RD), stage 4 by total RD with glaucoma, and stage 5 by advanced end-stage changes. ² The abnormal retinal vasculature and occlusion of the dilated blood vessels are treated by cryotherapy and laser photocoagulation for stages 1–3A. Surgical retinal reattachment is performed for most cases of stage 3B showing total exudative RD, whereas enucleation is consequently indicated for stage 4 patients, ^1,2 which may be possibly complicated by painful neovascular glaucoma (NVG). Histologic and cytologic evaluations of Coats' disease are limited and rare, but they prove the presence of macrophages and lymphocytes in the eye. ^3,4 These results suggest that Coats' disease involves abnormal vasculature as well as inflammatory cell infiltration. 

Anti–vascular endothelial growth factor (VEGF) agent injection into the vitreous is now widely used to treat patients with various neovascular diseases of the eye, including age-related macular degeneration (AMD). Moreover, anti-VEGF treatments have been tried for patients with Coats' disease. In fact, injections of intravitreal anti-VEGF antibody contributed to the improvement of exudation, hemorrhage, and reattachment of the retina in selected patients. ^1,5 Subsequent studies have demonstrated that elevated VEGF levels were observed in intraocular fluid, such as subretinal fluid, vitreous, and aqueous humor of Coats' disease. ^1,5,6 Moreover, the elevated VEGF levels in the vitreous decreased markedly after the injection of an anti-VEGF agent. ^1,5 Thus, the pathogenesis and progression of the disease are possibly regulated by VEGF secretion into the ocular fluid. Although anti-VEGF treatments may be a significant option for patients with Coats' disease, it is an enigma that immunolocalization of VEGF has yet to be elucidated in eyes with Coats' disease. 

Therefore, we collected enucleated globes, histologically diagnosed with Coats' diseases, to examine VEGF expression using immunohistochemistry. In addition, the expression of VEGF receptor (VEGFR)-2, which plays a key role in the pathology of endothelial cell permeability, ⁷ was analyzed in this study. 

Materials and Methods

The institutional review boards of the University of Southern California and Hokkaido University approved our use of human specimens obtained from the Doheny Eye Institute, Pathology Laboratory, and Hokkaido University Hospital, respectively. All procedures conformed to the Declaration of Helsinki for research involving human subjects. Nine eyes were obtained from donors who underwent enucleation through the Doheny Eye and Tissue Transplant Bank, Los Angeles, California. A subretinal membrane was obtained during vitrectomy in a 28-year-old male who was clinically diagnosed with stage 3B Coats' disease. Two normal retinae obtained by orbital exenteration in a 50-year-old patient due to squamous cell carcinoma invading the orbit, and evisceration in a 69-year-old patient due to blindness, ocular pain, and corneal opacity, without diabetes mellitus, were also examined as controls. All eyeballs, retinal tissues, and a membrane were fixed in 4% paraformaldehyde in an operating room soon after the enucleation, evisceration, or membrane removal. After fixation, all human samples were preserved as paraffin-embedded blocks. All unstained slides were made from the blocks in April 2011, and immunohistochemical study was performed soon after the unstained sections were ready for use. The slides were also stained with hematoxylin and eosin staining, and periodic acid-Schiff (PAS) staining for detection of eosinophilic exudation. 

Immunohistochemistry

The slides were dewaxed, rehydrated, and rinsed in phosphate-buffered saline twice for 10 minutes. As a pretreatment, microwave-based antigen retrieval was performed in 10 mM citrate buffer (pH 6.0). These slides were incubated with 3% hydrogen peroxide for 10 minutes and then with normal goat serum for 30 minutes. Sections were incubated with anti-VEGF (dilution 1:50; Santa Cruz Biotechnology, Santa Cruz, CA), VEGFR-2 (1:50; Acris Antibodies, San Diego, CA), VEGFR-1 (1:50, ab2350; Abcam, Tokyo, Japan), and VEGFR-3 (1:50, AF349; R&D Systems, Abingdon, UK) antibodies at 4°C overnight. Positive signals were visualized using 3,3′-diaminobendizine (DAB) as a substrate. Sections with a DAB reaction, in which primary antibody was omitted, served as negative controls. Human tissues of the pterygium, microvessels in the epiretinal membrane of diabetic retinopathy, and kidney were used as positive controls for VEGF and VEGFR-3, VEGFR-2, and VEGFR-1, respectively. Slides were examined using a commercial digital microscope (Keyence BZ-9000; Keyence, Osaka, Japan). 

Double-Staining Immunohistochemistry

In double-staining immunohistochemistry, after incubation with the first antibody of anti-VEGF antibody, anti-CD68 monoclonal antibody (1:50; Abcam), a marker for macrophage, ⁸ or anti-VEGFR-2 phosphorylation on Tyr1214, Tyr951, and Tyr1054 polyclonal antibodies (1:50; Abcam), Alexa Fluor 546 goat anti-rabbit antibody for 30 minutes and FITC-conjugated anti-CD31 monoclonal antibody (1:50; Abcam), for 30 minutes at room temperature were then incubated. After being washed, sections were mounted using mounting media with 4′,6-diamino-2-phenylindole (DAPI; SlowFade Gold Antifade Reagent with DAPI; Life Technologies, Carlsbad, CA). All slides were examined using a multiphoton laser scanning microscope (Zeiss LSM 510; Carl Zeiss, Inc., Thornwood, NY). 

Evaluation of Tissue Specimens

First of all, the presence of macrophages was confirmed in one slide available of each case at low magnification. The number of VEGF-immunopositive macrophages was determined by direct counting at a high magnification (objective lens: ×40) in all fields of each specimen. The VEGF-positive rate (%) was calculated as the number of positive cells of the total number of cells present in the tissues. In this study, histologically, vascular abnormalities were defined as blood vessels with marked irregular dilated vessel lumen and/or with hyalinized vessel walls. When two or more such abnormal blood vessels were seen in the largest cut section of eyes, the marked vascular abnormalities were evaluated as the presence in each case. Student's t-test was applied for the evaluation of differences in VEGF-positive rates and the presence of marked vascular abnormalities. The level of significance for all tests was P < 0.05. 

Reverse Transcription (RT)-PCR Analyses

Total RNA was isolated from three formalin-fixed, paraffin-embedded (FFPE) tissue sections, consisting of two Coats' disease and one retinoblastoma (High Pure FFPE RNA Micro Kit; Roche Applied Science, Indianapolis, IN). Reverse transcription was performed with a first-strand synthesis system for RT-PCR (SuperScript III; Life Technologies) and random primers, essentially as described. ⁹ RT-PCR analyses were performed with the following primers for human genes used: VEGF-A (forward 5′-AGA TCG AGT ACA TCT TCA AGC CATC-3′; reverse 5′-CGT CAT TGC AGC AGC CC-3′), VEGFR-2 (forward 5′-GGC CCA ATA ATC AGA GTG GCA-3′; reverse 5′-TGT CAT TTC CGA TCA CTT TTG GA-3′), and ACTB (forward 5′-CTG GAA CGG TGA AGG TGA CA-3′; reverse 5′-AAG GGA CTT CCT GTA ACA ATG CA-3′). 

Results

Histopathologic Findings in Eyes with Coats' Disease

The Table summarizes the clinicopathologic profiles examined in this study. Ages ranged from 1 to 22 years (mean: 6.7 years). All patients were male. Histologic examination demonstrated NVG and cataract in three enucleated eyes with Coats' disease. There were cholesteric clefts and macrophage infiltration in the subretinal space in all cases examined (Fig. 1B). In some eyes, there were collections of foreign body-type multinuclear giant cells (Fig. 1D) and lipid-laden histiocytes. In six of nine globes, marked vascular abnormalities, including irregular dilated vessel lumen and/or with hyalinized vessel walls, were noted in the retina (Fig. 1F). Dense serous eosinophilic materials filled the retina (Fig. 1H, arrows) and subretinal space, which were positive for PAS staining. RD was noted in all patients. Morphologic retinal pigment epithelium (RPE) changes included metaplasia and cell proliferation (Fig. 1J) in four eyes. Choroidal inflammatory cell infiltration (Fig. 1L, arrow) was observed in five eyes. 

Figure 1. 

 

H&E staining in enucleated eyes with Coats' disease. Low magnifications reveal entire globes including squares (A, C, E, G, I, K). High magnifications are consistent with the squares, respectively (B, D, F, H, J, L). There are cholesteric clefts and macrophages in the subretinal space (B). Foreign body-type multinuclear giant cells are surrounding the cholesteric clefts (D). There are marked vascular abnormalities in the retina, including dilated vessels and hyalinized vessel walls (F). Dense serous eosinophilic materials fill in the retina (H, arrows). Morphologic retinal pigment epithelium changes include metaplasia and cell proliferation (J). Choroidal inflammatory cell infiltration is observed (L, arrow). SRS, subretinal space; NFL, nerve fiber layer; Vit, vitreous space.

Table. 

 

Clinicopathologic Profiles of Patients with Coats' Disease Examined in This Study

No.	Age, y	Sex	Side	Size, mm	NVG	Eosinophilic Fluid	Macrophages	PAS-Positive Exudate	Dilated Retinal Vessels	Giant Cells	Cholesteric Clefts	RD	Histiocytes	RPE Changes	Choroiditis	VEGF-Positive Rate in Macrophages
1	4	M	L	22 × 21 × 21	–	+	+	–	–	+	+	+	+	Metaplasia	+	16.7
2	4	M	L	26 × 22 × 22	–	+	+	–	+	+	+	+	+	–	–	79.8
3	10	M	R	25 × 24 × 24	+	+	+	–	+	–	+	+	+	–	+	22.5
4	8	M	R	23 × 24 × 23	–	+	+	–	–	–	+	+	+	–	+	5.1
5	2	M	L	21 × 22 × 20	–	+	+	+	+	–	+	+	–	–	+	22.2
6	22	M	R	19 × 19 × 18.5	+	+	+	+	+	+	+	+	–	Proliferation	+	93.5
7	4	M	L	23.5 × 23 × 23	+	+	+	+	+	–	+	+	–	Proliferation	–	60.3
8	5	M	R	20 × 21 × 20	–	+	+	+	+	–	+	+	–	Proliferation	–	61.6
9	1	M	R	20 × 20 × 22	–	+	+	–	–	–	+	+	+	–	–	4.4

 M, male; L, left; R, right; RPE, retinal pigment epithelium.

Immunohistochemical Results for VEGF and VEGFR

VEGF immunoreactivity was observed in cells infiltrating the subretinal space (Figs. 2A, 2B), whereas no immunoreactivity was noted on omitting the anti-VEGF antibody reaction, serving as a negative control (Fig. 2C). These cells expressing VEGF were also positive for CD68, a marker for macrophages, using double-staining immunohistochemistry (Figs. 2D–G). The Table represents VEGF-positive macrophages in each globe. All globes contained VEGF-positive macrophages, ranging from 4.4% to 93.5%. The VEGF-positive rates were 56.7 ± 29.3% and 8.7 ± 6.9% in globes with and without marked vascular abnormalities, respectively. The VEGF-positive rate in macrophages was significantly higher in cases containing the vascular abnormalities than in those without the abnormalities (P < 0.01). 

Figure 2. 

 

H&E staining (A), VEGF immunoreactivity (B), and double-staining immunohistochemistry with DAPI nuclear staining (D, G: blue), VEGF (E, G: red), CD68 (F, G: green) in an eye with Coats' disease. A variety of macrophages are present in the subretinal space (A), where cytoplasmic immunoreactivity for VEGF is noted (B). Immunoreactivity is not detected in macrophages in negative control (NC) omitting the first antibody reaction (C). VEGF immunoreactivity is colocalized with cells positive for CD68, a macrophage marker (D–G). ONL, outer nuclear layer.

Immunoreactivity for VEGF was observed in the detached retina and dilated vessels (Fig. 3A, asterisk), where the immunoreactivity was detected in CD31-positive retinal endothelial cells (Figs. 4A–D, arrows). There were no significant correlations between VEGF immunoreactivity in retinal endothelial cells and the presence of retinal vascular abnormalities. No immunoreactivity for VEGF was observed in multinuclear giant cells. VEGF immunoreactivity was detected in macrophages infiltrating the subretinal proliferative tissue obtained during vitrectomy (see Supplementary Material and Supplementary Fig. S1). 

Figure 3. 

 

H&E staining (A, C) and VEGF immunoreactivity (B, D) in an eye with Coats' disease (A, B) and a normal retina (C, D). The detached retina contains abnormal dilated vessels (A, asterisk). Immunoreactivity for VEGF is observed in the detached retina and abnormal vessels (B). In contrast, VEGF immunoreactivity is not observed in the blood vessel situated in the normal retina (C, D: arrow).

Figure 4. 

 

Double-staining immunohistochemistry with VEGF (A, D: red), CD31 (B, D: green), and DAPI nuclear staining (C, D: blue) in eyes with Coats' disease. VEGF immunoreactivity is detected in some of the CD31-positive blood vessels (A–D, arrows) in the retinal tissue, where VEGF-negative retinal endothelial cells are intermingled (A–D, arrowheads).

In contrast, normal retinae did not contain macrophage infiltration or vascular abnormalities. VEGF immunoreactivity was not observed in the blood vessel situated in the normal retina (Figs. 3C, 3D, arrow). Cytoplasmic immunoreacitivity for VEGF was weakly noted in the inner nuclear layer (INL). 

VEGFR-2 immunoreactivity was noted in endothelial cells located in abnormal retinal vessels with hyalinization of the vessel wall in six cases (Figs. 5A, 5B). VEGFR-2 immunoreactivity was also detected in the inner layer of the detached retina, whereas macrophages infiltrating the subretinal space showed no expression of VEGFR-2. Immunoreactivity for VEGFR-2 was not observed in a subretinal membrane obtained during vitrectomy. Moreover, VEGFR-2 was phosphorylated on Tyr1214 in blood vessels situating in eyes with Coats' disease (Figs. 5C, 5D), whereas the phosphorylation was not observed in normal retinal vessels (data not shown). Phosphorylation of VEGFR-2 on Tyr1054 was marginally detected in CD31-positive retinal vessels (Figs. 5E, 5F), whereas phosphorylation on Tyr951 was not observed (Figs. 5G, 5H) in Coats' disease. Immunoreactivity for other types of VEGFR, such as VEGFR-1 and VEGFR-3, was not detected in blood vessels in Coats' disease or normal retina using immunohistochemistry (data not shown). 

Figure 5. 

 

H&E staining (A) and double-staining immunohistochemistry with DAPI nuclear staining (B, D, F, H: blue), VEGF receptor (VEGFR)-2 (B, red), VEGFR-2 phosphorylation on Tyr1214 (C, D: red), Tyr1054 (E, F: red), Tyr951 (G, H: red), and CD31 (B, D, F, H: green) in an eye with Coats' disease (A–H). Retinal vessel abnormalities including irregular dilatation of vessel lumen are noted (A, arrowheads). VEGFR-2 immunoreactivity is detected in the cytoplasm of the endothelial cells constituting the abnormal vessels (B, arrowheads). Bar, 50 μm. VEGFR-2 is phosphorylated on Tyr1214 in CD31-positive blood vessels (C, D). Phosphorylation of VEGFR-2 on Tyr1054 is marginally detected in CD31-positive retinal vessels (E, F, arrow), whereas phosphorylation on Tyr951 is not observed (G, H).

Gene Expression of VEGF and VEGFR-2

Both VEGF-A and VEGFR-2 genes were clearly observed in FFPE tissue sections (Fig. 6) using RT-PCR in Coats' disease (nos. 2 and 3; refer to the Table), and one case of retinoblastoma. Expression levels of VEGF-A and VEGFR-2 genes in Coats' disease were similar to those in retinoblastoma. 

Figure 6. 

 

Gene expression analysis of VEGF-A and VEGFR-2 in human retinas. RT-PCR products of VEGF-A and VEGFR-2 observed in cDNA from Coats' disease retinas and retinoblastoma.

Discussion

In this study, we found that a variety of macrophages were present in the subretinal space in all nine eyes and in a surgically removed subretinal membrane of Coats' disease. VEGF immunoreactivity was clearly detected in the cytoplasm of infiltrated CD68-positive macrophages. Moreover, the VEGF-positive rate in macrophages was significantly higher in cases containing retinal vessel abnormalities than those without the abnormalities (P < 0.01). Together with a previous publication on VEGF, a major product of activated macrophages in the pathology of AMD, ¹⁰ these results suggest that macrophages play a critical role in the promotion of vascular permeability and angiogenesis by expressing VEGF in Coats' disease. Indeed, previous reports have demonstrated that the VEGF protein concentration was high in subretinal fluid in patients with Coats' disease, ^1,5,6 correlated with the location of infiltrating macrophages and VEGF immunolocalization. 

It has been reported that anti-VEGF agents together with the addition of triamcinolone acetonide, a corticosteroid, are effective for treatment in patients with Coats' disease. Cakir et al. ¹¹ proposed that both anti-VEGF drug and triamcinolone injections were responsible for reducing retinal vascular permeability. However, the mechanisms underlying the effectiveness have yet to be elucidated. This study demonstrated that VEGF was expressed in infiltrating macrophages, indicating the reason why anti-VEGF therapy and corticosteroid use are effective. Singhal et al. ¹² demonstrated that triamcinolone can attenuate macrophage accumulation when activated macrophages accumulate in the retina during neural degeneration. These results suggest that not only an anti-VEGF agent but also triamcinolone lead to the suppression of macrophage activity and VEGF expression, which eventually contribute to effective treatment in patients with Coats' disease. 

This study also showed that VEGF was expressed in the detached retina, where the expression in retinal vessels was also noted. In contrast, VEGF immunoreactivity was hardly detected in normal retinae. Together with the previous research on elevated VEGF concentrations in the vitreous of Coats' disease, ⁶ the present data suggest that VEGF was upregulated in the retina during the progression of Coats' disease, which might be subsequently secreted into the vitreous. On the other hand, it has been shown that VEGF levels were elevated in rhegmatogeneous RD, ¹³ despite the absence of neovascular changes. Therefore, the expression of VEGF in the detached retina of Coats' disease may be associated with RD, which might not correlate with abnormal vasculature. 

Jun et al. ¹ demonstrated that an elevated VEGF level could increase the vascular permeability in Coats' disease. This study also showed VEGFR-2 expression in the cytoplasm of endothelial cells constituting the retinal vessel abnormalities in enucleated eyes with Coats' disease. Since VEGFR-2 is known to play a key role in the pathology of endothelial cell permeability, ⁷ these results suggest that VEGFR-2 contributes to the pathology of vascular abnormalities, together with VEGF originating from macrophages. It is known that VEGFR-2 undergoes autophosphorylation and becomes activated. Major phosphorylation sites of VEGFR-2 are located in the kinase insert domain (Tyr951), in the tyrosine kinase catalytic domain (Tyr1054), and in the C-terminal portion of the receptor (Tyr1214). ¹⁴ In addition, phosphorylation of VEGFR-2 on Tyr1214 could lead to endothelial migration in response to VEGF, which plays an important role in angiogenesis. ¹⁴ Indeed, our immunohistochemical results suggested VEGFR-2 phosphorylation on Tyr1214 in blood vessels situated in eyes with Coats' disease. Further studies are needed to clarify a critical role of VEGF signaling in the pathology of Coats' disease. 

In this study, immunohistochemical results on VEGF and VEGFR-2 have been shown in formalin-fixed, paraffin-embedded tissue sections. Indeed, the positive rate of immunohistochemical study on human tissue is affected not only by the fixation, but also by the preserved conditions or duration. To reduce the influence, the fixation and immunohistochemical study were performed soon after obtaining the human tissues and preparation of unstained slides, respectively. Moreover, gene expression of VEGF and VEGFR-2 was clearly detected in mRNA isolated from formalin-fixed, paraffin-embedded sections of Coats' disease. These results suggest that the pathologic sections can be available, and the results on RT-PCR analysis may support the data in an immunohistochemical study. 

A limitation of this study was our analysis of the expression of VEGF and VEGFR-2 in enucleated eyes, which represent the advanced stage of Coats' disease. Therefore, it may be hard to evaluate whether macrophages (and the VEGF they express) are the cause or the effect in the advanced stage disease. Moreover, it remains unclear whether the results on VEGF immunoreactivity determined in this study can also be applied in the early stage of the disease. Actually, a previous cytologic analysis revealed that macrophage infiltration was observed in intraocular fluid, ³ where VEGF concentrations might be elevated at a relatively early stage. ⁶ We also confirmed that VEGF immunoreactivity was noted in infiltrated macrophages in enucleated eyes, as well as a proliferative tissue obtained during vitrectomy (see Supplementary Material and Supplementary Fig. S1). Therefore, these results suggest that macrophages play a key role in the production of VEGF at various stages of the disease. 

Supplementary Materials

jpg S1 

Acknowledgments

The authors thank Ikuyo Hirose and Shiho Namba for their technical assistance. 

Footnotes

 Supported by grants-in-aid for Scientific Research from The Ministry of Education, Culture, Sports, Science, and Technology, Japan.

Footnotes

 Disclosure: S. Kase, None; N.A. Rao, None; H. Yoshikawa, None; J. Fukuhara, None; K. Noda, None; A. Kanda, None; S. Ishida, None