Ultrastructural Changes Associated with Accumulation of Inclusion Bodies in Rat Retinal Pigment Epithelium

Abstract

purpose. To determine the structural changes in the retinal pigment epithelium (RPE) and neighboring structures induced by intravitreal injection of a lysosomal protease inhibitor.

methods. Eleven-week-old Sprague–Dawley rats were injected with 5 μl of a lysosomal protease inhibitor, E-64 (2.22 μM), intravitreally once and killed at 24 hours, 48 hours, or 7 days later. Others received two or three injections at 48-hour intervals or three daily injections, and killed at 1, 4, and 7 days after the last injection. Eyes were enucleated and retinal tissues were processed for light and electron microscopy.

results. A single injection of E-64 caused only a transient accumulation of phagosome-like and phagolysosome-like inclusion bodies in the RPE. By contrast, repeated injection caused progressive accumulation of these inclusions followed by altered RPE cell conformation, and changes in organelles such as loss of smooth endoplasmic reticulum (SER). This was accompanied by shortening and loss of photoreceptor outer segments without prior dysmorphic changes, alteration of choroidal capillaries, and invasion of Bruch’s membrane by fibroblasts and pericytes. Intravitreal injection of vehicle as control induced no structural changes.

conclusions. E-64 treatment induced structural changes in the outer retina. The causal relationship between accumulation of inclusions in RPE and changes in other subcellular organelles and neighboring cells systems is not clear. However, there are possible explanations: physical disturbance of organelles, particularly SER by inclusions; cellular damage by consequent upon accumulation of A2-E; or, shortage of recycled material due to reduced degradation of phagosomes.

The retinal pigment epithelium (RPE) plays an important role in mediating nutritional exchange between the photoreceptors and the choriocapillaris, and in processing the shed tips of outer segments (OS). The RPE also exerts an inductive effect on the choriocapillaris by liberating trophic factors. It is thought to play a central role in the pathogenesis of age-related macular disease and certain monogenic retinal disorders. 

A universal biomarker of cellular aging in eukaryotic postmitotic cells is the appearance over time of autofluorescent lysosomal residual bodies that have been variously termed age pigments or lipofuscin granules. The potential role of accumulation of lipofuscin granules in RPE cells in age-related macular disease, and in many macular dystrophies such as fundus flavimaculatus and Bulls eye dystrophy has been extensively studied. ^{1 2 3 4 5 6 7 8 9 10 11 12 13} Although there is a linear relationship between the quantity of residual bodies and autofluorescence, marked variation exists in the ratio between the two in human donors, ¹⁴ and it can be manipulated by varying dietary vitamin A. ¹⁵ Increase in lipofuscin-like granules in the RPE has been reproduced experimentally by dietary deficiency of antioxidants, ^{16 17} exposure of RPE to oxidized photoreceptor OS components, ¹⁸ and intravitreal injection of lysosomal enzyme inhibitors. ^{19 20 21 22} Lipofuscin granules or residual bodies are generated by incomplete degradation of both heterophagosomes produced by phagocytosis of shed OS and autophagosomes. Efficiency of degradation of phagosomes may depend upon the nature of the phagosomal contents, ^{17 18 23 24 25} and on the qualitative or quantitative attributes of degradative enzymes. ^{19 20} Despite increasing knowledge concerning the mechanisms of lipofuscin granule accumulation, little is known of the causal relationship between lipofuscin accumulation and RPE cell dysfunction, which may in turn influence photoreceptors and choroidal capillaries. It has been proposed that accumulation of lipofuscin granules may interfere with cell function by reducing the cytoplasmic space ²⁶ or by acting as a free-radical generator. ²⁷ Eldred ²⁸ hypothesized that one of lipofuscin components (A2-E: N-retinylidence-N-retinylethanolamine) may inhibit lysosomal enzyme activity by elevating lysosomal pH resulting in accumulation of substrates and formation of residual storage granules. A2-E, at a critical concentration, might cause leaky lysosomes. Leaked A2-E may induce changes of cellular plasma membranes, and leaked lysosomal enzymes may contribute to RPE cell death. 

The purpose of this study is to seek a causal relationship between lipofuscin accumulation and RPE cell dysfunction by recording ultrastructural changes following accumulation of lipofuscin-like granules in RPE cells induced experimentally by intravitreal injection of E-64. E-64 is known to inhibit lysosomal cysteine proteases such as cathepsins B, H, and L²⁹ and some nonlysosomal cysteine proteases such as calpains ³⁰ ; E-64 treatment has proved to induce an accumulation of lipofuscin-like granules in brain and RPE. ²⁰  

Materials and Methods

Animals

Thirty-six Sprague–Dawley rats, obtained at 11 weeks of age, were maintained in a 12-hour light/12-hour dark cycle for at least 7 days with a daytime illumination of 150 lux. 

Factors

E-64 (Boehringer, Mannheim, Germany) was dissolved in a minimal volume of a mixture of ethanol and saline (1:1), and then diluted with a saline solution to a concentration of 2.22 μM used in this study. 

Injection and Histologic Procedures

Animals were anesthetized with ketamine (84 mg/kg)-xylazine (6 mg/kg) mixture by intramuscular injection. Using a 32-gauge beveled needle, 5 μl of E-64 solution was injected at 10:00 AM into the vitreous cavity of one eye through the sclera, choroid, and retina at a point midway between the ora serrata and the equator of the eye. The other eye of each rat received injection of 5 μl of vehicle as control. Nine animals were injected once, and three animals were killed after 24 hours, 48 hours, and 7 days. Nine animals were given two injections and nine were given three injections on alternate days; and, nine were injected on 3 consecutive days. In each group three animals were killed 1, 4, and 7 days after the last injection. All animals were killed at 10:00 AM to control for photoreceptor disc shedding by an overdose of carbon dioxide and perfused intravascularly with a phosphate-buffered mixture of 2% paraformaldehyde and 2.5% glutaraldehyde. Eyes were enucleated, bisected along the vertical meridian, and immersed overnight in the fixative at 4°C after removal of the anterior segment and rinsing in phosphate-buffer solution containing 0.2 M sucrose. For electron microscopy the specimens were dissected into smaller pieces, fixed in 2% phosphate-buffered (pH 7.4) osmium tetroxide for 1.5 hours, dehydrated in a graded series of ethanol, and embedded in epoxy resin. Semithin sections stained with toluidine blue were examined by light microscopy. Ultrathin sections were obtained from the posterior retina, stained with uranyl acetate and lead citrate, and examined using a Hitachi 7100 electron microscope (Hitachi, Katsuta, Japan). Remaining portions of the eyes were embedded in paraffin and processed for light microscopy. 

We quantified the number of phagosomes and phagolysosomes in RPE of animals injected with E-64 or vehicle once. Then, 10 to 15 electron micrographs (X4,000) of RPE in each posterior retina (average 150 μm) were taken as contiguously as possible. The negative images of electron microphotographs were altered to the positive images and stored using a Film Scanner LS-4500 AF (Nikon, Tokyo, Japan) and Photoshop 4.0 (Adobe, San Jose, CA). A phagosome was identified as an inclusion body that contains lamellar materials and surrounded by a limiting membrane on the screen. A phagolysosome was identified as an inclusion body containing both lamellar and amorphous materials. The length of the sampled RPEs were measured on the calibrated screen using NIH image. The number of the inclusion bodies per millimeter of the RPE was then calculated in each eye. Data from three eyes were averaged to produce the number per group. 

We quantified the area of profiles of SER per cytoplasmic unit of RPE of animals injected with E-64 or vehicles on 3 consecutive days and killed 4 days after the last injection. To obtain these data, electron micrographs (X10000) of five cells of RPE in each posterior retina were taken. In the same way that inclusions were measured, we measured the area of each cell, the total area of all profiles of SER, and the total areas of other organelles such as nucleus, mitochondria, phagosomes, phagolysosomes, lysosomes, and peroxisomes. Then, in each eye, we calculated the ratio of the area occupied by SER to the total area of five cells from which the total areas of other organelles were subtracted. Data from three eyes in each group were averaged to produce the mean. Statistical differences between experimental animals and controls were evaluated using the nonparametric Mann–Whitney U test. 

We calculated the ratio of the cytoplasmic area occupied by inclusions to the cytoplasmic area of the RPE using published methods. ¹⁴  

All experimental procedures followed the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the guidelines for Animal Research of the Kagoshima University Faculty of Medicine. 

Results

Intravitreal injection of E-64 induced variable structural changes in the outer, but not the inner, retinal layers. The degree of change depended on the number of injections and on the treatment time, while intravitreal injection of saline/ethanol as control induced no structural changes. The changes were more marked in the posterior than the anterior region. 

A single injection of E-64 induced changes limited to the RPE cells. After 24 hours, the RPE cells showed an increased number of membrane-bound inclusion bodies that were 1 to 3 μm in diameter. Most of them contained lamellar material that appeared to be photoreceptor OS disks, and they resembled the phagosomes observed in normal and control RPE cells. Some inclusions contained both lamellar and amorphous material, representing phagolysosomes (secondary lysosomes). After 48 hours, the number of inclusions increased further in experimental eyes (Fig. 1B ), and was consistently much greater than in control eyes in which inclusions were mostly smaller secondary lysosomes (0.1–1.0 μm in diameter) and only occasionally phagosomes (Fig. 1A) . After 1 week, the injected and control eyes appeared to be more similar in number and appearance of the inclusion bodies (Table 1) . Apart from these inclusions, the RPE cells appeared normal and there were no morphologic differences between experimental and control eyes (Fig. 1) . Morphologic differences were observed in neither the photoreceptor outer and inner segments nor the inner retina between the experimental and control eyes that received a single injection. 

Repeated injections of E-64 caused more profound alterations in the RPE, shortening and absence of the photoreceptor OS and changes of the choroid. The RPE cells exhibited more numerous phagosomes and phagolysosomes than those cells treated with a single injection (Figs. 2B , 3 ). Noticeably, the inclusions in the injected eyes were distributed throughout the cytoplasm, while the secondary lysosomes in the control eyes were present only in the apical region as occurs in normal eyes. The ratio of the cytoplasmic area occupied by inclusions to the cytoplasmic area of the RPE was 39% just 4 days after three daily injections. These RPE cells were taller in height and showed changes in other organelles. Smooth-surfaced endoplasmic reticulum (SER), which forms close-meshed networks of branching tubules throughout the cytoplasm in the normal and control RPE (Fig. 2A) , was decreased, showing sparse networks particularly in the apical region of RPE cells (Fig. 2B) . Quantification demonstrated a significant change in the ratio of the area occupied by SER per cytoplasmic unit (Table 2) . A relatively large number of autophagic vacuoles was observed (Fig. 2B 3A 3B 3C) . Most of them contained SER that was degraded to variable degrees (Figs. 3A B ). There were also vacuoles enclosing possible microperoxisomes (Fig. 3C) . Small lipid droplets were frequently seen in clusters (Fig. 3D) . These changes were seen at days 1 and 4 in the animals that received multiple injections and were more intense and extensive after three injections. 

In the rats killed 4 to 7 days after three daily injections (Fig. 4) , the apical processes of some RPE cells were dilated, shortened, and decreased in number when compared with normal and control RPE cells. The basal infoldings of these cells were expanded or deepened and disappeared in places. Granular material was found between the plasma membrane and the basal lamina. The photoreceptor OS adjacent to the abnormal RPE cells were shortened or absent. Apart from rudimental OS, the remaining disk membranes of the shortened OS did not show an abnormal appearance such as transformation into vesicular or tubular structures that are usually seen in degenerative OS. The choriocapillaris showed changes such as decreased number of fenestrations. 

After three injections on alternate days, local thinning and degeneration of the RPE cells was observed in the posterior globe. This was associated with inward folding of the outer nuclear layer, causing detachment of the photoreceptors from the RPE. At these sites, the photoreceptor OS were shortened or lost, and macrophage-like cells were seen occasionally in the subretinal space. Although the inner nuclear layer was displaced, the inner retinal layers appeared otherwise normal, except that some processes of Müller cells appeared to be altered. In such areas, the endothelial cells of the choriocapillaris showed degenerative alterations, and the fenestrations were mostly absent. Probable fibroblasts and processes of pericytes had invaded Bruch’s membrane, which showed slight widening (Fig. 5) . 

Discussion

That intravitreal injection of E-64 causes accumulation of inclusions in rat RPE cells, which increased with repeated injection of E-64, is in accord with the observations of others. ^{19 20} Most of the accumulated inclusions observed resembled phagosomes and phagolysosomes derived from shed OS observed in normal RPE cells. Several possible mechanisms exist to account for this observation. First, E-64 injection might cause accelerated photoreceptor OS shedding. However, the lack of abnormality of photoreceptor OS, such as vesicular or tubular change of disc membranes as seen in light-damaged retinas, would argue against this explanation. Second, inhibition of lysosomal cysteine proteases by E-64 may decrease the catabolism of phagocytized OS disks, resulting in generation of inclusions, a view favored by others. ^{20 21 22} Also, our study showed that the phagosomes that accumulated after one injection of E-64 were mostly degraded into residual bodies by about 1 week, whereas the accumulated phagosomes after repeated injections remained mostly nondegraded or little degraded for a longer period. Our observations after a single E-64 injection resemble those reported by Katz ³¹ in that a single injection of leupeptin, a similar protease inhibitor, resulted in a rapid massive accumulation of lipofuscin-like inclusions that returned to normal levels. These results seem to imply that the rate of phagosomal degradation was inversely related to the presence of residual E-64 in the RPE and reversed by the production of newly synthesized proteases. 

In contrast to a single injection, repeated injections of E-64 caused additional cellular changes at the level of the RPE such as enlargement of the cells, decrease in the amount of SER, increase in the number of autophagic vacuoles (particularly of SER), appearance of lipid droplets, disorganization of apical processes, and basal infoldings. All indicate generalized cellular dysfunction. As these changes followed accumulation of inclusion, they seem not to be the cause of this accumulation. Although it is not possible to resolve whether these cellular changes are the consequence of the accumulated inclusion bodies, or the consequence of other factors such as A2-E, or the inhibition of cytosol cysteine proteases by E-64, there are several possible explanations for the observations. First, the inclusions may generate free radicals with consequent damage to cellular membranes and other components by peroxidation. ³¹ However, we think that the nondegraded inclusions observed in this study would not act as a free-radical generator as do lipofuscin granules in vivo. ²⁷ Second, A2-E may be formed in the inclusions with consequent damage to cellular membrane. ²⁸ Eldred ²⁸ stated “Only when the membrane is disrupted (as in the RCS retinal debris, or in the RPE lysosome, or possibly in threshold light damage) can the aldehyde group approach the ethanolamine moiety of phosphatidylethanolamine or free ethanolamine. The acidic and hydrophobic conditions provided by the RPE lysosome would be well-suited for the promotion of the A2-E synthetic reaction.” A2-E might be formed only at low level in those inclusions in which degradation of OS disc membranes would be inhibited by E-64. Nevertheless, in some phagolysosomes, A2-E could reach a critical concentration, at which point A2-E might cause leaky lysosomal membranes and leaked A2-E might affect cell membranes. Third, the increased number of inclusions may reduce the cytoplasmic space available to the organelles, and may cause a mechanical disruption of the cellular organization. In nerve cells of the human brain, lipofuscin granules increase with age and displace or destroy the protein synthesizing endoplasmic reticulum with its ribosomal RNA. ³² In this context, the ratio of 39% of the cytoplasmic area occupied by inclusions in the eye after 4 days following three daily injections is much larger than that of the maximum of the ratio reported for residual bodies in aged human RPE. ¹⁴ The enlargement of the RPE cell body full of inclusions seems to be due to overloading with the inclusions. SER in RPE cells, characteristically forming closed-meshworks throughout the cytoplasm, may be susceptible to damage. If such were the case, accumulated lipofuscin granules may damage RPE cells even by their physical effects. Fourth, retarded degradation of the inclusions may cause reduction of material, the recycling of which may be necessary for renewal of intracellular organelles such as SER and photoreceptor OS. 

Noteworthy is finding an increase in autophagic vacuoles, particularly of SER and possible microperoxisomes, and decrease in SER that was demonstrated quantitatively. These changes seemed to be the initial detectable events following the accumulation of inclusions. SER and microperoxisomes play an important role in lipid metabolism that is highly active in the RPE. The increase in autophagic vacuoles may be due either to retarded degradation of their contents by E-64 or to increase in autophagic activity triggered either by damage to organelles such as SER or by some mechanisms regulating autophagy. It is plausible that retarded degradation of OS-derived phagosomes may lead to decreased metabolism of lipid, which may cause excess SER and microperoxisomes to be eliminated by autophagy as occurs in hepatocytes. ³³ Increase of lipid droplets may reflect a declined lipid metabolism as seen in hibernation. ³⁴ Lipid droplets increase in association with accumulation of lipofuscin-like inclusions induced by antioxidant nutrient deficiency. ¹⁷ Whatever the cause of a decrease in the SER, it would result in a decline in the capacity for lipid metabolism, which, if sustained, may lead to cell dysfunction and finally cell death. It would be important to establish if this phenomenon is unique to this experimental model or common to other experimental conditions and diseases in which accumulation of lipofuscin or lipofuscin-like granules occurs. 

During the first week after the last injection, the RPE in which inclusions increased manifested disorganization of the basal infoldings and apical processes. Flattening and widening of basal infoldings occur in aging, ³⁵ in some diseases such as Chediak–Higashi syndrome in cats characterized by increased lipofuscin granules, ³⁶ and under certain experimental conditions such as photic insult ³⁷ and sodium iodate-induced retinopathy. ³⁸ All these features have been interpreted as indicating a relatively inactive state of the RPE cells, and presage atrophy. ^{36 38}  

Shortening and loss of the OS observed is more likely to be due to impaired morphogenesis of disc membranes rather than degeneration, because photoreceptor cell OS showed no degenerative changes such as tubular change of disc membranes. The ability of the RPE to recycle lipids has been well-illustrated. ^{39 40} Possible inability of the RPE to recycle materials due either to reduced phagosomal degradation or to compromised lipid metabolism might result in impaired morphogenesis of OS. If these conclusions are true, there is implied dependence on recycling of products derived from degradation of phagosomes to generate new OS membranes, and that acquisition of plasma-derived material may be insufficient to sustain this process fully. On the other hand, the possibility cannot be ruled out that inhibition of calpain II, a cytosol cysteine protease, by E-64 might interfere with OS morphogenesis because calpain II has proved to exist in photoreceptors and may be involved in disc membrane morphogenesis by proteolysis of myosin II. ⁴¹  

Inward folding of the outer nuclear layer has been reported in aged rats ⁴² and in certain feline disorders, ^{43 44} and has been attributed to the loss of Müller cells ⁴² or to retinal detachment followed by reattachment. ⁴³ Whatever the mechanism, the lesions may have been precipitated by focal loss of the spatial interaction between OS and RPE cells. 

The alterations of Bruch’s membrane and the choriocapillaris observed in this study also appear to have resulted from the dysfunction of the RPE, as these changes were found in proximity to apparently degenerate RPE cells. Choroidal alterations associated with atrophy and degeneration of the RPE have been found in human age-related macular disease ⁴⁵ and under experimental conditions such as light insult ³⁵ and sodium iodate injection. ³⁷ These changes are characterized by reduced fenestrae and reduced lumina, and invasion of Bruch’s membrane by pericytes and fibroblasts. Our results are consistent with these findings and provide additional evidence supporting the view that the RPE modulates the structure and function of the choriocapillaris. ³⁸  

The possibility that the changes observed in RPE and other cells might result from effects of E-64 on some nonlysosomal cysteine proteases, such as calpains and caspases cannot be ignored. Calpains, when activated by Ca²⁺, function in the degradative pathway for cytoskeletons, resulting in interference with phagocytic process and disintegrative change of lens fibers. ⁴⁶ Predictably, inhibition of calpains by E-64 would stabilize those organelles. Also, they modify protein receptors for steroid hormones, and so forth. ³⁰ Calpain II has been detected in most layers of the bovine retina and may be involved in disc membrane morphogenesis. ⁴¹ In the present study, we did not see significant morphologic changes in the inner retina. Caspases are known to be resistant to E-64. ⁴⁷ Alkyl-dihydroxyacetonephosphate synthase, a peroxisomal enzyme involved in the biosynthesis of ether phospholipids, is processed by a cysteine protease, which is inhibited by leupeptin. ⁴⁸ If E-64 inhibits this protease in RPE cells, the effect may result in some morphologic and functional changes. 

Our results indicate that intravitreal injection in rat of a lysosomal protease inhibitor (E-64) induces not only abnormal accumulation of phagosomes and phagolysosomes, but also evidence of general cellular dysfunction in RPE cells, which leads to changes in both photoreceptor cells and choroid. The changes are similar in some respects to age changes in RPE, photoreceptor, and choroid, though we freely acknowledge that there are differences between an acute experiment such as ours and the consequences of lifelong metabolic activity, and species differences between rat and human. Because age-related maculopathy is thought to be an exacerbation of the normal aging process, this model may illustrate potential pathogenetic mechanisms of some components of age-related macular degeneration, and refinement of the model may lead to valuable new information. 

 Supported by Grants-in-Aid for Scientific Research 60295251 from the Ministry of Education, Science and Culture of Japan.

 Submitted for publication June 21, 1999; revised March 1 and August 3, 2000; accepted August 22, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Electron micrographs of rat retinal pigment epithelium (RPE) after 48 hours following one injection of E-64 or vehicle. (A) Control RPE cell shows a moderate number of phagolysosomes (arrows). (B) E-64-treated RPE cell shows numerous phagosomes (arrowheads) and phagolysosomes (arrows). OS, photoreceptor outer segments; BM, Bruch’s membrane. Bar, 1 μm.

Table 1.

 

Number of Phagosomes and Phagolysosomes per mm

Days after Single Injection	1	2	7
Experimental eyes (9)	1306 ± 221* (3)	1341 ± 276* (3)	450 ± 55, ** (3)
Control eyes (9)	364 ± 144* (3)	276 ± 25* (3)	323 ± 164, ** (3)

 Data are summarized as mean ± SD. Data in parentheses are number of eyes.

*  P < 0.05;

**  P = 0.28.

Figure 2.

 

RPE after 4 days following three daily injections of E-64 or vehicle. (A) Control RPE cell appears normal. Inset: the outlined portion at higher magnification, showing dense SER. (B) E-64-treated RPE cell shows numerous inclusions, many autophagic vacuoles (arrowheads), and reduced SER. Note disorganized apical processes (arrows) and shortened outer segments. Inset: the outlined area at higher magnification, showing sparse SER. Bar, 1 μm.

Table 2.

 

Ratio of SER Area to Cytoplasmic Area^{, †} in RPE

Experimental (3)	Control (3)
0.1419 ± 0.0156*	0.2580 ± 0.0140*

 Data are summarized as mean ± SD. Data in parentheses are number of eyes.

*  P < 0.05.

†  The cytoplasmic area does not contain nucleus, mitochondria, inclusions, etc.

Figure 3.

 

Portions of RPE cells after 4 days following three daily injections of E-64. (A, B) Autophagic vacuole (arrow) containing SER. (C) Autophagic vacuole (arrow) containing probable microproxisomes. Inset: The identical autophagic vacuole in a stepped serial section, showing that a double-layered isolation membrane is in the process of enclosing the particles. (D) Small lipid droplets (arrowheads). Pl, phagolysosome. Bar, 1 μm.

Figure 4.

 

RPE cells after 4 days following three daily injections of E-64. The apical processes and basal infoldings (arrow) have been disorganized. Granular material (arrowhead) is present between the plasma membrane and the basal lamina. Photoreceptor OS have been shortened or lost. The endothelial cells of the choriocapillaris have decreased number of fenestrations. Bar, 1 μm.

Figure 5.

 

RPE after 7 days following three injections on alternate days with E-64. Photoreceptor inner segments (IS) that lost OS have come into contact with degenerating RPE cells (RPE). A macrophage-like cell (M) exists between IS and RPE cell. Pericyte processes that have a basal lamina (arrowhead) and unidentified cells exist in Bruch’s membrane (BM). A cell (arrow) seems to be either a degenerative endothelial cell or a pericyte. Choriocapillaris (∗) has lost most of its fenestrations. Bar, 1 μm.

The authors thank Yoshiko Maeda for expert technical assistance.