Increased intercellular adhesion molecule-1 immunoreactivity in the sclera-choroid complex in hypercholesterolemia experimental model

1 Postdoctoral Programme, Pontifical Catholic University of Paraná, Curitiba/PR, Brazil; 2 Pontifical Catholic University of Paraná, Curitiba/PR, Brazil; 3,4 Medical School of the Pontifical Catholic University of Paraná, Curitiba/PR, Brazil; 5 Positivo University, Curitiba/PR, Brazil; 6 Curitiba Ophthalmic Centre, Curitiba/PR, Brazil; 7 Angelina Caron Hospital, Campina Grande do Sul/PR, Brazil; 8 Pontifical Catholic University of Paraná, Curitiba/PR, Brazil;

INTRODUCTION P athological intraocular neovascularisation is a complication that usually leads to blindness. Conditions associated with intraocular neovascularisation include retinopathy of prematurity, diabetic retinopathy, neovascular glaucoma, and age-related macular degeneration (AMD). AMD is a multifactorial disease that can be triggered by genetic, environmental, demographic, dietary, clinical, ocular, and lifestyle factors (1,2) .
For decades, AMD was attributed to excessive, pathological accumulation of lipids in Bruch's membrane due to a dysfunction of retinal pigment epithelium (RPE) cells (3) . More recently, the role of vascular endothelial growth factor (VEGF) in subretinal neovascularisation was clarified, which was critical to better understand the causes of AMD. As a result, major studies have been conducted to identify the mechanisms and the main cells responsible for releasing VEGF. Previous experimental and clinical studies aiming to identify the causes and cures for other organ diseases had demonstrated that hypoxia is the main stimulus for VEGF release (4) , thus helping understand the formation of the subretinal neovascular network seen in exudative AMD (1) . However, angiogenesis is a complex multi-stage process that involves not only growth factors, but also adhesion molecules and inflammatory cytokines, both of which have a key role in the process (5,6) . Intercellular adhesion molecule 1 (ICAM-1), or CD54, is a glycoprotein of the immunoglobulin superfamily. Similarly to other adhesion molecules, ICAM-1 is present in leukocytes and endothelial cells and participates in leukocyte recruitment to injured or inflamed tissue (7) . A study performed in human donor eyes that aimed to assess ICAM-1 distribution in the choriocapillaris found a greater concentration of this immunoglobulin in the macula, as opposed to the peripheral region (8) . This finding suggests a higher susceptibility of the macula to trafficking of immune cells such as macrophages, which helps explain the higher incidence of subretinal neovascular membrane in this region. In addition to producing VEGF (9) , macrophages are also a source of inflammatory and pro-angiogenic cytokines such as interleukin 6 and tumour necrosis factors (TNF) (10,11) , which mediate inflammatory response and contribute significantly to the formation of subretinal neovascular membranes (6,(12)(13) .
Treatment with anti-angiogenic drugs has been more effective than previous therapies, although the visual prognosis is still limited (14) . Therefore, an experimental model that induces the expression of adhesion molecules in the sclera-choroid complex can potentially simulate the changes observed in AMD. Improved understanding of the pathophysiology of AMD might support the development of more effective treatments for the condition.
The objective of this study was to assess ICAM-1 expression in the sclera-choroid complex of hypercholesterolemic rabbits.

METHODS
The study protocol was approved by the Ethics Committee for Animal Research of the Pontifical Catholic University of Paraná, in accordance with the principles of the Declaration of Helsinki (1964) and the Association for Research in Vision and Ophthalmology (ARVO).
Increased intercellular adhesion molecule-1 immunoreactivity in the sclera-choroid complex in hypercholesterolemia experimental model

Experimental setting
The procedures described in this study were carried out in the Laboratory of Surgical Technique at PUC-PR and the Study Centre of the Angelina Caron Hospital. Animals were kept in a vivarium (macroenvironment) on 12/12-hour light cycles with air circulation and controlled temperature between 19 and 23°C. During the experimental period, animals were fed water and standard Nuvital™ (Nuvital, Colombo, Brazil) rabbit feed ad libitum.

Animals and experimental design
Twenty-one male New Zealand albino rabbits (Oryctolagus cuniculus) were selected for this study from the Central Vivarium of PUC-PR. The animals had a mean age of 110 days and a mean weight of 2,770 grams. They were divided into two groups: a normal diet group (NG) with 8 rabbits and a hypercho-lesterolemic diet group (HG) with 13 rabbits. NG animals were fed Nuvital™ (Nuvital, Colombo, Brazil) standard laboratory rabbit feed and euthanised after 4 weeks. Rabbits from the HG group were fed Nuvital™ (Nuvital, Colombo, Brazil) standard laboratory rabbit feed with additional 1% cholesterol throughout the study (supplementary feed) and were euthanised after 8 weeks.
Serum total cholesterol, triglycerides, HDL cholesterol, and fasting glucose levels were measured for each rabbit at the baseline and at the time of euthanasia. Blood samples were taken by puncturing the marginal ear vein under general anaesthesia by intramuscular injection of ketamine 5 mg/kg and xylazine 35 mg/kg. Plasma levels of glucose, total cholesterol, HDL cholesterol, and triglycerides were determined using the automated enzymatic-colorimetric method (Architect™, Abbott Diagnostics). A total of 500 microlitres of serum from each animal were frozen for future analysis. The weight of rabbits was measured on a weekly basis.
The animals were euthanised by intravenous injection of 5 mL pentobarbital and their eyes were immediately fixed in 4% paraformaldehyde (Merck, Darmstadt, Germany) in a 0.1M phosphate buffer at pH 7.4 for 4 hours for histomorphometric and immunohistochemical analysis.

Preparation of supplementary feed
For the 1% hypercholesterolemic feed used throughout the 8 weeks of the experiment, 200 g of cholesterol 95% (Sigma-Aldrich™) were dissolved in 800 mL of chloroform (Biotec™) and the solution was uniformly distributed over 20 kg of Nuvital™ (Nuvital, Colombo, Brazil) rabbit feed. Before feeding it to the rabbits we waited for 24 hours to allow the chloroform to evaporate. Each animal was offered 600 g of feed per day (15) .

Histomorphometric Analysis (quantitative)
Both eyes of each animal were removed and fixed, but only one was chosen at random for analysis. After fixation, the samples were evaluated macroscopically. An axial section was performed at the level of the optic nerve, dividing the ocular globe into two equal halves (upper and lower). The lower half was stored for future studies. The upper half was dehydrated, diaphanised and embedded in paraffin using a Leica™ TP 1020 tissue processor (Leica, Wetzlar, Germany). Paraffin blocks were produced using a Leica™ EG1160 embedding station. A Leica™ RM2145 microtome (Leica, Wetzlar, Germany) was used to obtain 5-µ-thick histological sections. The sections were placed on glass slides smeared with albumin, stained with haematoxylineosin, and mounted with 24x90-mm coverslips using Entellan  For quantitative analysis the selected sections were stained with haematoxylin-eosin and evaluated microscopically with a 4x objective lens. Using a blue marker, the posterior segment of the half ocular globe was divided manually into 10 equal segments (from one pars plana to the contralateral pars plana). One image was obtained per segment using an Olympus BX50 microscope connected to a Sony camera and Image-Pro Plus™ software. In each captured image four linear morphometric measurements were taken using the Image-Pro Plus™ software in order to evaluate scleral and choroidal thickness in the 10 segments. Finally, the mean values of the four measurements in each segment were calculated for all samples. Thickness was expressed in micrometres.

Tissue preparation and immunohistochemical analysis
The histological cuts were deparaffinised and rehydrated, and then subjected to blocking of endogenous peroxidase. They were subsequently washed in deionised water and incubated in a moist chamber at 95°C for 20 minutes for antigen retrieval, after which they were again subjected to blocking of endogenous peroxidase. Sections were covered with primary mouse anti-ICAM-1 monoclonal antibody diluted at 1:100 (Novocastra™, Newcastle upon Tyne, England). They were then covered with a secondary antibody, HRP-labelled anti-mouse polymer (Advance™ System, DakoCytomation, Inc., Carpinteria, CA, USA), and incubated at room temperature for 30 minutes. The sections were dripped with mixed DAB substrate (DakoCytomation, Inc., CA, USA) and incubated for another 3 to 5 minutes. Finally, the slices were counterstained with Mayer's haematoxylin and mounted.
Positive and negative controls were used in all markings, and the slides were first analysed by an observer who did not know to which group the slides belonged (masked analysis). In this analysis, positive and negative results for the ICAM-1 marker were recorded. The positive areas acquired a brownish colour and were analysed using colour morphometry. For this procedure, images were captured in five consecutive fields, from the pars plana to the contralateral pars plana, using an Olympus BX50 microscope with a 40x objective lens, a Sony DXC-107A camera, and Image-Pro Plus software. This software allowed the observer to select and colour the positive areas. The immunoreactive surface area was then automatically calculated by the software and expressed in micrometres². Data were compiled in a Microsoft Excel (Redmond, WA, USA) spreadsheet for statistical analysis. The variable immunoreactive area represents the sum of all positive areas in each of the five fields.

Statistical analysis
To compare groups with regard to the quantitative variables we used either Student's t-test for independent samples or the Mann-Whitney nonparametric test, as appropriate. Normality was assessed using the Shapiro-Wilk test. Values with p<0.05 were considered statistically significant. Data were analysed using Statistica 8.0 software.

RESULTS
Comparing groups with regard to glucose, total cholesterol, HDL, and triglycerides.
Laboratory values in NG (normal diet group) animals remained stable from the baseline to the end of the experiment.
On the other hand, HG (hypercholesterolemic diet group) animals showed a significant increase in total cholesterol at the time of euthanasia. At the baseline, mean total cholesterol levels for both groups were approximately 41.3 mg/dL, and they reached 2146.8 mg/dL (p<0.001) in HG rabbits at the end of the experiment. Triglyceride levels also changed significantly in HG rabbits. Mean values at the baseline were approximately 46.5 mg/dL for both groups, and they reached 168.5 mg/dL (p=0.001) in HG animals at the time of euthanasia. No significant change was observed for glucose and HDL levels in this group at the end of the experiment.

Sclera and choroid staining with haematoxylin-eosin
Very few histiocytes were found in the sclera-choroid complex of NG animals (Figure 1), whereas HG samples revealed a large number of macrophages in these structures, causing their thickness to increase (Figure 2).
NG: Normal diet group; HG: Hypercholesterolemic diet group.  Comparing groups with regard to the total ICAM-1immunoreactive area HG animals showed a significant increase in ICAM-1 expression in the sclera-choroid complex relative to NG animals (p<0.001) ( Table 2), characterised by the brownish colour acquired by immunoreactive areas (Figure 3-B). Conversely, the sclera-choroid of NG animals was thinner and predominantly blue, indicating low immunoreactivity to anti-ICAM-1 antibodies (Figure 3-A). Table 2 Total surface area of the sclera-choroid immunoreactive to anti-ICAM-1 antibodies through colour morphometry.
Values are expressed in micrometres².

DISCUSSION
In a recent experimental study, our group demonstrated that a hypercholesterolemic diet increased MCP-1 expression in the sclera and choroid, suggesting that the changes in great vessels that induce atherosclerotic plaque formation might also occur in the sclera-choroid complex (16) . After observing an increase in chemotactic molecules such as MCP-1, adhesion molecules had to be assessed to verify the hypothesis that atherosclerosis and AMD have similar pathophysiological mechanisms (17)(18)(19) .
In a normal eye, immunoglobulin ICAM-1 expression is low in the vascular endothelium of the choroid and retina, as well as the RPE, Bruch's membrane, and the external limiting membrane (20)(21)(22) . However, under pathological conditions such as exudative AMD, there is a significant increase in ICAM-1 expression in choroidal vessels and RPE cells. This increase in immunoreactivity has been observed mainly around the subretinal neovascular membrane (SNM), where a large number of vessels is observed (23) . Increased ICAM-1 expression in the RPE and choroid during SNM formation was also observed in experimental studies, suggesting this immunoglobulin is involved in the pathogenesis of AMD (24)(25) . This hypothesis was supported by another experiment where mice genetically deficient for ICAM-1 or its counter-receptor CD18 were found to be less prone to SNM formation than normal mice (5) .
In the present study, a cholesterol-enriched diet induced a significant increase in ICAM-1 expression in the sclera-choroid complex. As it is involved in leukocyte recruitment to inflamed tissues (26) , the increased expression of this immunoglobulin may explain the significant rise in the number of macrophages and the consequent increase in scleral-choroidal thickness observed in this and other studies.
Inflammatory cells and the complement system play an important role in the pathogenesis of AMD (27) . Furthermore, it has been demonstrated that macrophages produce inflammatory cytokines, tissue factor, VEGF, and other angiogenic factors (9,(28)(29) . It has also been experimentally observed that the widespread depletion of macrophages decreases SNM activity (30) . Hence, there is a consensus regarding the importance of macrophages in the pathogenesis of AMD. In this context, it is worth noting that oxidised LDL is the major stimulus for an increase in chemotactic molecules, adhesion molecules that attract macrophages to the inflammatory site, both in atherosclerosis and in AMD (31) . However, increased ICAM-1 expression is induced not only by oxidised LDL, but also by native LDL (32) . Thus, LDL deposited in Bruch's membrane, originating from the degradation of photoreceptor ends by RPE cells (2,31) , can potentially increase the expression of these adhesion molecules. This may explain the increased number of macrophages and the onset of AMD. Therefore, experimental models that induce the accumulation of macrophages in the choroid by increasing the number of chemotactic or adhesion molecules may mimic macular degeneration, paving the way for future studies on therapies for the condition.

CONCLUSION
This study has demonstrated that increased serum cholesterol levels in an experimental model lead to increased ICAM-1 expression, which may explain the accumulation of macrophages in the sclera-choroid complex. Future research using drugs that interfere with this adhesion molecule may shed light on their possible use in the treatment of AMD.