Alterations of retinal pigment epithelium cause AMD-like retinopathy in senescence-accelerated OXYS rats

Ophthalmoscopy examination

Retinas from 20 day-aged OXYS rats did not revealed the presence of any pathological signs. The first AMD-like alterations in the fundus of the eye were found in 22% of the OXYS rats when they reached the age of 1.5 months (Figure 1). Alterations manifested itself as distorted reflectance of the fundus, swelling of the retina, the emergence of distinct foci of ischemia, and the signs of atrophy of choriocapillaris and RPE.

Up to the age of 3-4 months, the morbidity had already reached 100% when OXYS rats exhibited the prevalent 1st stage of the disease: drusen with signs of atrophy of RPE and partial loss of choriocapillaris, whereas the major blood vessels of choroid remained unchanged. In some animals, we observed detachment of pigment epithelium in the form of dome elevation in the posterior pole of the eye. Four to five percent of the animals at this age have features of the 2nd stage: large soft drusen, edema in the central zone, enlargement of the zone, exudative detachment of pigmented and neuroepithelium of the retina. The proportion of animals with this stage of the disease increased with age (Figure 1), and by the age of 12 months, it reached 31%.

Up to the age of 12 months, clinical symptoms of retinopathy worsened. The third stage was registered in 3% of animals at 9 months, but at 12 months, it was apparent in 44% of the animals and was characterized by severe irreversible changes. The distinctive signs of this stage are hemorrhagic detachment of retina, dotted bleeding in retina, and new blood vessel formation and cicatrization. Importantly, neovascularization did not develop in all animals, but rather, in isolated cases, as it normally happens in people. By the age of 24 months, the changes were irreversible in nearly all animals. In contrast, first alteration in Wistar retina started to develop by the age of 24 months and could not be strictly classified as signs of AMD.

Histological analysis and electron microscopy

Light- and electron microscopy did not relive significant differences in structure of retina of OXYS and Wistar rats at 20 days and 3 months. The RPE layer has one line of cells with normal localized nucleus, endoplasmic reticulum, ribosome and mitochondria (Figure 2). On the Figure 2 endothelial cells, which formed the wall of choriocapillaries, cavity of choriocapillaries and erythrocyte in it are good visible. But already at the age of 20 days, along with normally functioning choriocapillaries, we observed appearance choriocapillaries with stasis of blood cells. (Quantity analysis of choroid vessels and RPE cells see below).

Towards the 12 months of age, destructive changes in OXYS rat's retina progressed and affect RPE/ Bruch's/choriocapillaris complex in contrast to Wistar rats (Figure 3a) with small anomaly in Bruch's membrane ultrastructure and open choriocapillaries. Main ultra-structural alterations are revealed in the cell apical part. First of all, this is the absence of a continuous layer of electron-dense inclusions as well as practically completes absence of phagosomes – debris of the photoreceptor outer segments phagocytized by pigment epithelium, whirling extensions of the basement membrane into the cytoplasm (Figure 3b). Underlying these areas was pronounced thickening of Bruch's membrane with disturbance of collagen and elastin layers. The choriocapillaris was found to have severe endothelial degeneration and transformation to fibrous tissue in the most severely affected regions.

In the retina of 24 months-aged OXYS rats, there is almost complete fibrosis of choroid and destruction of endothelium and, as a result, obliteration of cavity of choriocapillaries. Ultrastructure analysis of OXYS eyes from 24-month-old animals shown, that area of degenerating retina is shown from the inner nuclear layer to retinal pigment epithelium/Bruch's/ choriocapillaris complex (Figure 4b). Outer nuclear layer and the photoreceptors are disorganized and fragmented. Changes of the RPE included significantly increased number of lipofuscin granules, atrophy of some RPE cells and vascularization (Figure 4a). Really, as the photoreceptor cells degenerated and the outer nuclear layer and photoreceptor cell layers disappeared, vessels migrated toward the RPE (arrow on Figure 4a). Thickening of Brunch membrane with disturbance of collagen and elastin layers were seen in all eyes (Figure 5a).Also there is almost complete fibrosis of choroid and destruction of endothelium and, as a result, obliteration of cavity of choriocapillaris.

Ultrastructure of Wistar retina had some changes that typical for aging, but it was not so dramatic pathology (Figure 5) as in OXYS rats. We can see all outer retina layers. It is also increased amount of lipofuscin granules. Cavity of most choriocapillaries decreased, but vessels are functionally active, as it visible.

Quantitative analyses of the morphometry data showed that the average area of RPE cell was affected both by age (F_4.414= 15.8, p<0.000) and by genotype (F_1.39=73.4, p<0.000) and there was also a significant interaction between these two factors (F_4.414=3.7, p<0.007). Overall, we observed well-pronounced growth with age the average area of RPE cells in both OXYS and Wistar rats from 20 days to 17 months of age (1.4- and 1.6-fold respectively, p<0.000). By the age of 24 months, the area extent of RPE cells decreased slightly compared to 17-month-old animals, but the decline was significant only in the Wistar rats (p<0.00004), while in the OXYS strain, we detected only a tendency for lowering (p<0.056) as shown on Figure 6a.

Pairwise comparisons within the strains revealed reduction of average area of RPE cell already in 20-day-old OXYS rats by 22% compared to Wistar rats (p<0.000). At ages 3, 6, 17 and 24 months, the differences remained and this parameter was lower in OXYS rats by 15% (p<0.023), 33% (p<0.000), 28% (p<0.001) and 16% (p<0.026), respectively.

ANOVA analysis of morphometry data revealed age- related increasing of the specific area of choriocapillaris with stasis and thrombosis (F_4.68= 22.4, p<0.000), which was observed in rats of both strains. In addition, its quantity was affected by genotype (F_1.68=56.7, p<0.000), and since the age of 20 days, specific area of vessels with signs of partial occlusion of retinal vessels was greater in OXYS rats compared to the age-matched Wistar controls (Figure 7b). There was also a significant interaction between these two factors (F_4.68=4.8, p<0.002). As shown in the Figure 6b, in the Wistar retina, the number of vessels with thrombosis and stasis remained unchanged between 20 days and 17 months of age. Only at age 24 months the percentage of choriocapillaris with signs of abnormality became 5.6-fold greater than that in the 20-day-old animals.

OXYS rats exhibited a relatively rapid expansion of specific area of choriocapillaris with stasis and thrombosis already by age 6 months. Up to the age of 24 months, these pathological changes progressed, but at a slower rate. As a result, at ages 3, 6, and 24 months, this parameter in the OXYS retina was 2.6-, 3.6-, and 4-fold higher compared to 20-day-old OXYS rats (p<0.006, p<0.0002, p<0.0001, respectively).

Gene Expression

Expression of VEGF gene was affected both by age (F_4.39= 18.3, p<0.000) and by genotype (F_1.39=10.0, p<0.003) and there was also a significant interaction between these two factors (F_4.39=3.3, p<0.020). The same was observed for expression of PEDF gene: age – F_4.39=46.9; p<0.000; genotype – F_1.39=4.5; p<0.040. Factors also interacted with each other (F_4.39=3.1; p<0.026), which is probably due to the differences in age-dependent changes of retinal VEGF and PEDF expression between Wistar and OXYS strains. Post hoc analysis revealed that the maximum level of genes mRNA was in the retina at the age of 20 days (Figure 6) and had not difference between strains. Expression of VEGF and PEDF genes decreased 10-fold (p<0.000) and 5-fold (p<0.001) respectively between 20 days and 24 months of age in both Wistar and OXYS rats. At the same time, in OXYS rats, the reduction progressed more rapidly than in Wistar (Figure 7).

At the age of 20 days, there was no differencebetween Wistar and OXYS rats in terms of expression of VEGF and PEDF. But already at the age of 3 months, VEGF expression (Figure 7a) in OXYS rats was almost 10-fold lower (F_1.8=21.9, p<0.002) compared to 20-day-old animals and to age-matched Wistar rats (F_1.8=21.7, p<0.002) and PEDF expression (Figure 7b) decreased three-fold (p<0.0001) from age 20 days to 3 months and was about two-fold lower (F_1.8=7.29; p<0.027) compared to the age-matched Wistar animals (F_1.8=7.29; p<0.027). Expression of both genes in OXYS rats remained without significant changes throughout the rest of the observation period.

VEGF expression at the age of 24 months was 4- (p<0.002) and 2.6-fold higher (p<0.008) than in 12 and 17-month-old Wistar rats respectively, and it reached half of the expression level at the age of 20 days (p<0.0003). By the age of 24 months, PEDF expression increased by 2.4-fold (p<0.013) compared to 17-month-old Wistar rats (Figure 7).

Animals

The OXYS rat strain was developed at the Institute of Cytology and Genetics from Wistar stock by selection on the basis of susceptibility to cataractogenic effects of a galactose-rich diet and by inbreeding of highly susceptible rats as described earlier [39,40] and registered in the Rat Genome Database [41]. Today we have the 94nd generation of OXYS rats with spontaneously developing cataract and accelerated-senescence syndrome inherited in a linked manner. Cataract still serves as the key parameter for maintaining the phenotype of OXYS strain whereas the other signs of accelerated senescence, including retinopathy, appeared as concomitant.

Male OXYS and male Wistar (as controls) rats were born and reared at the Center for Genetic Resources of Laboratory Animals of Institute of Cytology and Genetics SB RAS (Novosibirsk, Russia). At the age of 4 weeks, the pups were taken away from their mothers and housed in groups of five animals per cage (57×36×20 cm) and kept under standard laboratory conditions (at 22±2^oC, 60% relative humidity, and natural light), provided with a standard rodent feed, PK-120-1, Ltd. (Laboratorsnab, Russia), and given water ad libitum.

Ophthalmoscopic examination

Ophthalmoscopic examination of OXYS and Wistar rats was carried out using a Beta direct ophthalmoscope (Germany) equipped with a slit lamp, after dilatation with 1% tropicamide. Assessment of stages of retinopathy was carried out according to Age-Related Eye Disease Study (AREDS) grade protocol (eyephoto.ophth.wisc.edu). In total, we examined 790 animals between 20 days and 24 months of age.

Histological analysis and electron microscopy

A total of 42 OXYS and 37 age-matched Wistar rats were used for histological and morphometry studies. Animals were examined at the age of 20 days, 3, 6, 12 and 24 months (6 – 10 per group). For studies of the main parameters, we selected retinas on the basis of fundoscopy findings in OXYS rats. Morphological analysis was performed as described earlier [21]. The posterior wall of the eye was collected, fixed in 2.5% glutaraldehyde, postfixed in 2% OsO₄, dehydrated in ascending concentrations of alcohol and in acetone, and embedded in epon-812. Semi-thin sections were sliced on an LKB-4 ultratome and stained with 0.1% toluidine blue. Counting was carried out on photo-graphs covering section regions from 50 to 250 mkm. RPE cells' area and specific area of choroid vessels were measured using quantitative analyses of the images were performed using Axiovision software (Zeiss, Thornwood, NY). Estimation was performed by examination 50 fields of view for each animal.

Ultrastructural study was performed on OXYS and Wistar rats at the age of 3, 12 and 24 months. Per 5 animals were in each group and in each of eye 3 areas were analyzed. Material for electron microscopy was fixed for 2 h at 4°C in 3% glutaraldehyde solution in buffer, pH 7.4. Then the material was additionally fixed with 1 % osmium tetroxide solution for 1.5 h and dehydrated in solutions of increasing alcohol concentrations (70% ethanol saturated with uranyl acetate). The material was embedded in Epon-812 epoxide resin. Serial ultrathin sections were prepared on a Leica Ultracut ultramicrotome (Leica, Austria) and stained with lead according to Reynolds [42]. The preparations were examined and photographed in a HU-1lB electron microscope (Hitachi, Japan).

RNA extraction and real-time PCR

Analysis of steady-state mRNA expression of VEGF in the retina of OXYS rats was completed at 20 days, 3, 12, 17, and 24 months of age (Wistar strain as control, 5 animals in each age group). At the appropriate age, rats were decapitated and eyes were removed. Retina for gene expression analysis was separated from the other tissue, placed into tubes for RNA isolation and frozen in liquid nitrogen. All specimens were stored at -70°C prior to the analysis.

Total cell RNA was isolated from rat retina by TRI-Reagent (Ambion) according to the manufacturer's instructions. The amount of isolated RNA was evaluated by means of electrophoresis of 1 μl of each RNA sample in 1.5% agarose gel. RNA content was determined in each sample using spectrophotometry at 260 nm and also by absorbance ratios 260/280 nm and 260/320 nm. RNA was stored at -70°C. Contaminating genomic DNA was removed by treatment with DNase I (Promega, USA) according to the manufacturer's instructions and then by repeated RNA extraction with the phenol-chloroform mixture and pure chloroform followed by precipitation with propanol. Reverse transcription was performed using M-MLV Reverse Transcriptase (Promega, USA) following the manufacturer's protocol. For subsequent PCR, we used of 0.25–1.0 μl of the resulting cDNA mixture.

Aliquots (4 μl) from all cDNA samples were mixed and the “average” solution was used for preparation of calibration curves, which were used for determination of relative cDNA level for genes of interest and a reference gene in experimental samples.

Age-related changes of vegf and pedf gene expression were investigated using iCycler iQ4 real-time PCR detection system (Bio-Rad Laboratories, USA) and SYBR Green I dye (Molecular Probes, USA). The housekeeping gene Rpl30 (encoding large ribosomal subunit protein 30) was used as a reference gene. The following primers were used: Rpl305′_ATGGTGGCTGCAAAGAAGAC_3′ and 5′_CAAAGCTGGACAGTTGTTGG_3′; vegf 5′_CTGGCTTTACTGCTGTACCTCCACC_3′ and 5′_GGCACACAGGACGGCTTGAA_3′; pedf 5′_GATTGCCCAGCTGCCTTTGACA_3′ and 5′_GGGACAGTCAGCACAGCTTGGATAG_3′.

The reaction mixture (final volume of 25 μl) contained the standard PCR buffer (67 mM Tris-HCl, pH 8.9, 16 mM (NH₄)₂SO₄, 0.01% Tween 20, and 10 mM β-mercaptoethanol), MgCl₂ (3 mM for vegf, 1.5 mM for RPL30, 3 mM for pedf), 0.2 mM dNTPs, SYBRGreen I (1: 20,000 dilution), 150 nM primers, and 0.8 U of Taq polymerase (Institute of Cytology and Genetics, Russia). Reaction was carried out under the following conditions: heating at 95°C for 3 min, 40 cycles: denaturation at 95°C for 20 sec, annealing 20 sec, elongation at 72°C 30 sec; data collection was based on fluorescence for Rpl30 at 84°C for 30 sec, data collection by fluorescence for vegf and pedf genes at 87°C for 10 sec. After completion of PCR, the melting curves for specificity control were recorded. In each experiment, samples of investigated cDNA were mixed with primers for a gene of interest (four repeats per cDNA sample) in one microtube plate; similar samples were mixed with primers for the reference gene (also four repeats). “Standard” cDNA was diluted from 1:2 to 1:64 with the same primers (2-3 repeats). For each cDNA sample, PCR was repeated at least twice. To confirm amplicon size and reaction specificity, PAGE electrophoresis was performed with DNA molecular weight markers.

The initial level of expression of investigated cDNA was determined by standard calibration curves (versus “standard” cDNA) and this value was obtained for each gene of interest and normalized according to the amount of the reference gene cDNA [43].

Statistical analysis

The data were analyzed using repeated measures ANOVA and nonparametric tests with the statistical package Statistica 6.0. Two-way ANOVA was used to evaluate the differences between OXYS and Wistar rats across ages (age × genotype). A Newman-Keuls post hoc test was applied to significant main effects and interactions in order to estimate the differences between particular sets of means. One-way ANOVA was used for individual group comparisons. Data are represented as mean ± S.E.M. Results were considered statistically significant if p value was less than 0.05.