Introduction

Senescent changes in the retinal pigment epithelium (RPE) and Bruch’s membrane are suspected causal factors for age-related macular degeneration [6, 7, 16, 23–25, 35]. Our understanding of the relationship of these changes to the pathogenesis of human macular degeneration is limited by the dearth of animal models. There is evidence that similar changes occur in aging non-human primates [2, 10, 11, 14, 36, 38, 40], but the slow development of this process in primates handicaps experimentation. It would be valuable to have models for these changes in the RPE and Bruch’s membrane in small animals, such as rodents, which occur over a more rapid time course, providing a greater range of possible experimentation. There have been several reports of age-related changes in the RPE and/or Bruch’s membrane of rats [20] and mice [4, 5, 8, 15, 22, 27, 29–32, 42]. Some of these reports describe unusual changes in the basal infoldings and basement membrane of rodents that develop between 1 and 2 years of age [20, 30, 31]. We find a similar pattern of age-related changes in the Rpe65 −/− mouse, but the alterations occur earlier and more extensively than in wild-type mice. This is interesting because the Rpe65 mutant evinces a drastic reduction of lipofuscin fluorophores and excessive accumulation of all-trans retinyl palmitate in the RPE [19] that may influence this process.

Methods

Two strains of mice were studied, the Rpe65 mutant and a wild-type strain. The Rpe65 −/− mice were established in a C57Bl/6 strain and obtained from breeding pairs provided by Michael Redmond (National Institutes of Health, Bethesda, MD). Wild-type C57Bl/6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME) and served as normal controls. Both eyes of eight Rpe65 −/− and nine wild-type mice were studied. Two mice from each strain were examined at 3, 6, 12–13 and 16 months of age; one wild-type mouse was examined at 6 weeks of age. Mice were killed with an intraperitoneal injection of Euthasol (40 mg/kg). The eyes were rapidly removed and punctured with a 30-gauge needle at the temporal limbus for orientation and placed in 3% glutaraldehyde in phosphate-buffered saline, pH 7.4. They were kept at 4°C for 3–20 days before being washed and dissected. The anterior segment and the lens were removed and the remaining eye cup sectioned into two pieces with a razor blade along the vertical meridian through the optic nerve head. These segments were osmicated, dehydrated using ethanol and propylene oxide and embedded in epoxy resin. Blocks were sectioned semi-serially at 1–2 μm in thickness, stained with 2% toluidine blue and examined by light microscopy. At selected points, mainly at the posterior pole, the block was trimmed and ultra-thin sections were cut, stained with uranyl acetate and lead citrate and examined using an electron microscope (Zeiss 100B). In two eyes from Rpe65 −/− mice, aged 13 and 16 months of age, blocks were trimmed and sectioned in the inferior peripheral retina to compare this area with changes seen at the posterior pole. All sections were viewed and selected areas photographed at magnifications of 5,000–27,000×.

Results

No abnormalities were detected in wild-type mice up to 1 year of age. Figure 1 shows an example of the normal-appearing basal surface of the RPE of a 6-week-old and a 1-year-old wild-type mouse. There are numerous, slender basal infoldings of the plasma membrane of the RPE cell adjacent to Bruch’s membrane without significant thickening of the basement membrane. Above the basal infoldings are numerous mitochondria and pigment granules adjacent to the nucleus of the RPE cells.

Fig. 1
figure 1

Electron micrographs of a 6-week-old and b 1-year-old wild-type mice show typical infoldings of the basal plasma membrane of RPE cells adjacent to Bruch’s membrane. There is no significant thickening of the basement membrane (magnification a 4,000×; b 8,000×)

Full size image

After 13 months of age changes could be observed at the basal surface of the RPE layer of wild-type mice; these are illustrated in Fig. 2. The most frequent change was a thickening of the basement membrane with small protrusions into the infoldings of the plasma membrane of the RPE cell (Fig. 2a). In some cases there were areas where the basal infoldings were completely absent (Fig. 2a,b). In some locations there was a prolongation of these protrusions so that they extended deeply into the basal surface of the RPE cell (Fig. 2c,d); a similar protrusion is also visible on the right side of Fig. 2b. The mitochondria, melanin and melano-lysosomal bodies as well as the apical surface of the RPE appeared to be normal.

Fig. 2
figure 2

Electron micrographs show the basal surface of RPE cells of a 13-month-old wild-type mouse. a Protuberances of the basement membrane into plasma membrane of the cell (arrow) and local loss of basal infoldings. b Thickening of the basement membrane with irregularities and some loss of the basal infolding. c, d Abnormal prolongations (arrow in c) of the basement membrane deep into the cytoplasm of the cell. These prolongations of basal laminar-like material are marked with an x. In both c and d, a thin stalk connects this large protuberance with the rest of the basement membrane. A small asterisk indicates an area of basal laminar infoldings that is being isolated from the remainder of the cell by the prolongations of the basal lamina (magnification 12,500×)

Full size image

In Rpe65 −/− mice, similar changes were apparent at the basal surface of the RPE by 6 months of age (Fig. 3) but not at 3 months of age (data not shown). In Fig. 3a there is an area at the basal surface that appears normal with characteristic basal infoldings (short arrow on the left). But on the right side of this arrow, there are abnormal protrusions of the basal lamina into the plasma membrane (longer arrow) and an absence of basal infoldings. This change is even more striking in Fig. 3b–d. At higher magnification (Fig. 3d) one sees an amorphous, granular appearance to these protrusions or deposits extending from the basement membrane into the basal plasma membrane of the RPE cell. There are a number of vacuoles, presumably containing lipid, in the cytoplasm of these cells seen as gray homogeneous round structures, which are considered to contain retinyl esters.

Fig. 3
figure 3

Electron micrographs show the basal surface of an RPE cell of a 6-month-old Rpe65 −/− mouse. a An area of normal basal infoldings (short arrow) adjacent to an area where they are absent. At this point there is an abnormal protuberance of basement membrane into the plasma membrane of the cell (long arrow) (magnification 10,000×). b An area where basal infoldings are absent and the basement membrane is thickened by the accumulation of amorphous material (arrow) (magnification 8,000×). c A prolongation of such a basement membrane protuberance penetrating the cytoplasm and contacting a lipid droplet (arrow) (magnification 8,000×). d A more extreme example of the protuberances of the basement membrane contacting the plasma membrane of the cell where basal infoldings are absent (Magnification 25,000×)

Full size image

At 16 months of age, these changes at the basement membrane the RPE of Rpe65 −/− mice have become most extraordinary, with long fingers of amorphous material penetrating deeply into the basal surface of the RPE cell (Fig. 4a,b). In some places in the section, islands of this material extend deeply into the cytoplasm of the RPE cell. In other areas these processes appear to completely encircle and isolate parts of the cytoplasm from the rest of the RPE cell (Fig. 4c,d). A slight suggestion of this can also be seen in the aged normal retina (Fig. 2d). Basal infoldings of the basement membrane are completely absent. This appearance was characteristic of many areas of the basal surface of the RPE layer of aged Rpe65 −/− retinas but seemed to be more prominent in the central area than the periphery of the retina.

Fig. 4
figure 4

Electron micrographs show the basal surface of RPE cells of a 16-month-old Rpe65 −/− mouse. a, b Enormous prolongations of the basement membrane into the plasma membrane of the cell (arrows) with total absence of basal infoldings (magnification 10,000×). c, d These prolongations of the basement membrane tend to isolate whole segments of the cytoplasm (asterisks) from the remainder of the cell (magnification 12,500× and 10,000×, respectively)

Full size image

Discussion

Age produces characteristic changes in the basal surface of RPE of wild-type mice [30, 31] and rats [20]. Mishima and Kondo [31] considered that these changes were due to a fusion of the plasma membranes of the basal infoldings of the RPE cell. Katz and Robison [20] noted a thickening of the basement membrane of RPE as well as an enlargement and prolongation of the basal infoldings. They also noted areas of loss of basal infoldings. In both studies [20, 31] these changes appeared only in animals that were more than 1 year old. We have found similar changes in the Rpe65 −/− mutant which appear much earlier, being detectable at 6 months of age. In these mice the thickening appears to be due to an excessive formation of basement membrane material, which extends into the basal plasma membrane of the RPE and is associated with a loss of basal infoldings. In extreme cases these extensions surround areas of cytoplasm of the RPE cell, separating these areas from the rest of the cell, which could lead to local areas of degeneration.

Any correlation with the pathology of human age-related macula degeneration is difficult because the human changes are much more extensive, involving most of Bruch’s membrane [6, 7, 16, 23–25, 35]. They characteristically involve the formation of drusen which are located external rather than internal to the basement membrane of the RPE and have not been reported in rodent eyes. There are, however, changes in the aging human retina that occur between the basement membrane and the basal plasma membrane of the RPE cell which extend over large areas, often being more extensive than drusen. In their late stages these are characterized by a linear palisade-like appearance but in earlier stages as excrescences on the basement membrane of the RPE, just as we and others [20, 30, 31] have described in murine retina. In humans, these changes in the basement membrane lead to the formation of putative long-spacing collagen [6, 24, 41], which is also found in aging basement membrane in other parts of the body. We have not observed such long-spacing collagen in our mice, but Katz and Robison [20] have found an example in rats. In man the increase in the basement membrane deposits that accumulate under the plasma membrane of the RPE can lead to detachment of the RPE cell and have been associated with choroidal neovascularization [16, 34, 37], although choroidal neovascularization in humans seems to occur in the plane of basal linear deposits [6]. This excessive deposition of basement membrane material observed in these aging rodent eyes could be related to the pathogenesis of human age-related macular degeneration.

This tendency towards a thickening of the basement membrane of the RPE is consistent with most other rodent models of possible age-related macular degeneration [4, 5, 8, 15, 17, 20, 22, 27, 29–32]. Why this occurs with age is difficult to explain. It seems that this amorphous material is an overexpression of basement membrane. The basement membrane is far more complex than standard electron microscopy has traditionally shown. All basement membranes contain laminins, entactin-1/nidogen-1, type IV collagen, heparan sulfate proteoglycans including type XVIII collagen, agrin, perlecan and various growth factors and proteases [12, 13]. Many growth factors and cytokines are stored in the basement membrane matrix and are released and activated only on dissolution. Little is known about the specific constituents of RPE basement membrane and the turnover of these constituents. Therefore it is difficult to hypothesize why this material seems to increase with age. Further studies with immunohistochemical methods, especially at the electron-microscopic level, will be necessary to elucidate the evolution of these changes in murine retina and their possible relevance to human age-related macular degeneration.

Perhaps the excessive accumulation of this material detected in the Rpe65 mutant provides an insight into the factors responsible. Katz and Robison [20] noted that there was an increase in the ratio of retinyl palmitate to stearate with age, which could reflect RPE cell senescence. This is interesting with regard to the Rpe65 −/− mutant, where there is a considerable increase in the amount of retinyl palmitate in the RPE due to a failure of isomerization of all-trans to 11-cis retinol [33]. Perhaps this is a sign of premature senescence in the mutant RPE. Human subjects with defects in the Rpe65 gene have a retinal degeneration with early signs of macular degeneration [18], but there is no evidence that they have thickened basal laminar deposits.

It is also interesting that the Rpe65 −/− strain has smaller amounts of lipofuscin accumulation in the RPE than wild-type mice, presumably because of the reduced turnover of rhodopsin [19]. Excessive lipofuscin accumulation is characteristic of autosomal recessive Stargardt’s dystrophy [3, 9, 21, 26, 28, 39], which is usually manifested as a severe macular degeneration with possible genetic relevance to age-related macular degeneration [1]. In the Rpe65 mutant we have an example of excessive basement membrane deposition, a possible causal factor in age-related macular degeneration, with minimum lipofuscin accumulation [19].