Ultrastructural Localization of Light-Induced Lipid Peroxides in the Rat Retina

Abstract

purpose. Localization of light-induced lipid peroxides in the rat retina at an ultrastructural level as benzidine-reactive substances.

methods. Long–Evans rats with nondilated pupils were exposed to intense light of 6000 lux for 12 or 24 hours. Control animals were kept under physiological light conditions. Rats with dilated pupils were exposed to a light intensity of 50 lux or 150,000 lux for 1 hour. For ultrastructural localization the enucleated eyes were fixed in a 0.1-M cacodylate buffer (pH 7.4) containing 2% glutaraldehyde for 2 hours. Pieces of the superior part of the central eyecup were incubated overnight with tetramethylbenzidine (TMB; pH 3.0) at 4°C, postfixed with 1.5% OsO₄, and embedded for electron microscopy.

results. In animals exposed to intense light, electron-dense structures appeared exclusively throughout the rod outer segments after an irradiation of 6000 lux for 24 hours or 150,000 lux for 1 hour and were absent in animals with nondilated pupils kept at physiological light conditions. Dilation of the pupils leads to the appearance of electron-dense structures after just 1 hour of 50 lux, whereas rats with nondilated pupils withstand even a 12-hour irradiation with 6000 lux. No electron-dense structures were found when no TMB was used in incubation.

conclusions. The appearance of electron-dense structures in the rod outer segments depends on the incubation with TMB and intensive light exposure of the rat. Dilation of the pupils lowers the threshold for the emergence of electron-dense structures significantly. This strongly supports the view that light-induced lipid peroxides in the rat retina are localized at an ultrastructural level as benzidine-reactive substances. This protocol presents a tool for the generation and ultrastructural localization of lipid peroxides in rat retinas.

Effects of visible light on the rat retina have been extensively studied. ^{1 2 3 4 5 6} Photoreceptor rod outer segment (ROS) membranes contain the highest levels of long-chain polyunsaturated fatty acids of any tissue in the body. ⁷ Docosahexaenoic acid (22:6ω3), the major polyunsaturated fatty acid in the retina is particularly susceptible to lipid peroxidation reactions. ⁸ Intense light exposure leads to retinal damage, ^{1 9} elevation of retinal hydroperoxide levels, and a decrease in the levels of rod outer segment docosahexaenoic acid. ^{10 11 12 13} Lipid peroxidation causes retinal degeneration ^{14 15} and has been discussed in the pathogenesis of age-related macular degeneration (ARMD). ¹⁶ There is good evidence for the assumption that lipid peroxidation is the main cause of lipofuscin formation, ^{17 18} although a different genesis has also been discussed. ¹⁹ An accumulation of lipid peroxides in the ROS diminishes the susceptibility of the associated proteins to enzymatic degradation, ²⁰ increasing the amount of indigestible residual material in the retinal pigment epithelium that is associated with the formation of lipofuscin ¹⁸ and drusen ²¹ and is likely to promote the development of ARMD. ²² The commonly used methods to detect lipid peroxides are biochemical techniques ²³ or a light microscopic technique, ²⁴ which do not allow ultrastructural localization. Until now, no ultrastructural method has been available to localize lipid peroxides. 

The appearance of electron-dense structures was recently found in ROS of hamsters after incubation of the retinas with tetramethylbenzidine (TMB) in the presence of H₂O₂. ²⁵ This finding was not completely understood and was thought to be mediated by the action of a peroxidase. This topic was carefully reinvestigated, and it was found that TMB reacts with endogenous lipid peroxides to an electron-dense reaction product. ²⁶ These electron-dense structures are therefore called benzidine-reactive substances (BRS). These findings suggest the development of a method for an ultrastructural localization of lipid peroxides in the retina of the rat, a commonly used laboratory animal. 

Methods

Peroxidation of Rat Retinas by Light Exposure

Male Long–Evans rats of 3 months of age (approximately 250 g) were kept singly in cages under a 12-hour day–night cycle (light period: 7:00 AM to 7:00 PM at 50 lux, measured within the rearing cages with a photometer [Colormaster 3F; Gossen, Erlangen, Germany ]) with fluorescent light. For illumination, animals were exposed to constant light with a high-pressure mercury lamp (HPL-N 125 W; Philips, Eindhoven, The Netherlands) for 12 hours (10:00 PM to 10:00 AM) or 24 hours (10:00 AM to 10:00 AM). The spectrum of the light source consists of six peaks at 365 nm, 405 nm, 435 nm, 545 nm, 580 nm, and 625 nm (UV-B, 7.1 μW/lm; UV-A, 784.6 μW/lm; UV-total, 791.7μ W/lm; visible, 3092.6 μW/lm; total radiation, 3884.4 μW/lm). 

The distance of the lamp from the bottom of the cage was 35 cm, corresponding to a light intensity of 6200 lux at the center of the cage and 5400 lux at the periphery, with the photometer probe pointed toward the light source. The light was not diffused in any way. Rats so treated did not seem to take any precautions to avoid the light, such as closing eyes or hiding in a corner of the cage. 

Cages used for constant-light exposure were constructed of translucent plastic (36 cm × 21 cm × 15 cm) and were not sandblasted. The top of the cages was covered by a grate made of steel (2-mm bars and 7-mm spacing). Humidity was 60%, and because of the absence of infrared light in the spectrum of the light source, the temperature inside the cages during illumination rose to a maximum of 24°C, and there was no need for fans. 

For the second light protocol, male Long–Evans rats of 3 months of age, raised under the same conditions were anesthetized using intramuscular injection of ketamine hydrochloride (200 mg/kg). Pupils were dilated using a single application of phenylephrine hydrochloride (2.5%) and tropicamide (0.5%) and prevented from drying by a frequent application of a 0.9% sodium chloride solution. Eyelids were open without fixation. Light exposure was performed for 1 hour (9:00 AM to 10:00 AM), without additional illumination (approximately 50 lux) or using a cold light source (KL 1500 position 5; Schott, Wiesbaden, using a halogen reflector lamp HLX 64634 EFR 15 V, 150 W, 3200 K; Osram, München), pointed directly at the eye. The light intensity at the surface of the eye was approximately 150,000 lux. The spectrum of the light source was a bell-shaped curve with a maximum at 620 nm and negligible proportions of UV-A and infrared light. 

None of the rats was dark adapted before illumination, and all animals were killed immediately after illumination under common laboratory light conditions (approximately 500 lux) by cervical dislocation under CO₂ anesthesia at 10:00 AM. 

All procedures involving animals were performed according to policies set forth in the ARVO Resolution for the Use of Animals in Ophthalmic and Vision Research. 

Preparation and Fixation of the Eye

Preparation and fixation of the eyes was performed according to a modified protocol of Fisher et al. ²⁷ Eyes were enucleated, and the cornea was incised by a small slit just behind the iris, to allow fixative access into the inside, and incubated in 0.1 M cacodylate buffer (pH 7.4) containing 2% glutaraldehyde at 4°C. After 10 minutes, the cornea, iris, and lens were entirely removed, and fixation proceeded for another 2 hours. After fixation, the vitreous body was removed, and the eyecups were bisected through the optic nerve head along the posterior ciliary artery. ⁵ Retinal buttons with a diameter of approximately 2 mm and the center approximately 2 mm superior to the optic nerve head were dissected from each animal, washed in 0.1 M cacodylate buffer (pH 7.4) five times, and subjected to the TMB reaction. 

Localization of Benzidine-Reactive Substances in the Eye

For the TMB reaction 0.5 mg/ml 3,3′,5,5′-tetramethylbenzidine dihydrochloride (Sigma, Deisenhofen, Germany) were prepared in McIlvaine’s Na₂HPO₄-citric acid buffer (BRS mixture; pH 3.0). ²⁸ Tetramethylbenzidine (0.5 mg/ml) was dissolved first in four parts 0.1 M citric acid solution supported by sonification. By adding one part 0.2 M Na₂HPO₄, the hydrogen concentration was adjusted to pH 3.0. 

Aldehyde-fixed and washed pieces of eyecups were incubated with freshly prepared BRS mixture at 4°C overnight. Control tissue was treated the same, including incubation with buffer, but without using TMB or were fixed by glutaraldehyde overnight, without using the BRS mixture. The pieces of eyecups were washed in cold cacodylate buffer five times and were postfixed by 1.5% osmium tetroxide (Paesel & Lorei, Hanau, Germany) in 0.1 M cacodylate buffer (pH 7.4) for 2 hours, bloc stained with 1% uranyl acetate (Merck, Darmstadt, Germany) in 70% ethanol, dehydrated in a graded series of ethanol, and embedded in Spurr’s resin. Semithin sections (approximately 700 nm) were stained with toluidine blue and examined by light microscope (Axiophot; Zeiss, Oberkochen, Germany). Ultrathin sections (light gold, approximately 100 nm) were poststained with uranyl acetate and lead citrate and investigated by electron microscope (Model 902 A; Zeiss). 

Results

In rats exposed to approximately 6000 lux for 24 hours without dilation of the pupils, electron-dense structures appeared throughout the rod outer segments (Figs. 1 A, 1B, 2 A, 2B) and were absent in animals kept at physiological (50 lux) light conditions (Fig. 3 A) or exposed to 6000 lux for 12 hours (data not shown). No electron-dense structures were found after incubation with buffer at pH 3.0 but with no TMB (data not shown). In retinas of animals kept under physiological light conditions, incubation with the BRS mixture resulted in swollen and irregular disc membranes, resembling light damage (Fig. 3A) , whereas retinas of animals treated the same but without the BRS mixture did not show these aberrations (Fig. 3B) . 

The electron-dense structures appeared as electron-dense sites of the disc membranes, bubblelike structures or spherical structures that located in spaces between individual discs. The continuity of the discs was interrupted at sites where the electron-dense structures appeared. The electron-dense sites seemed to extrude from the disc membranes. Several discs contributed to one electron-dense bubblelike structure (Figs. 2A 2B) . 

Illumination of rats with dilated pupils at a light intensity of 50 lux for 1 hour led to the appearance of small electron-dense sites of disc membranes throughout the ROS (Fig. 4) . However, bubblelike structures or spherical structures were not detected. 

Irradiation for 1 hour to a light intensity of 150,000 lux led to a picture similar to the 24 hours irradiation of 6000 lux (Figs. 5 A, 5B, 5C), comprising the formation of bubblelike structures or spherical structures that located in spaces between individual discs (Figs. 5A 5B) . The emergence of bubblelike and spherical electron-dense structures after light-induced peroxidation and incubation with TMB was strongly restricted to the ROS (Figs. 6A 6B 6C 6D) and was not seen in the retinal pigment epithelium (Fig. 5C) , in the inner segments (Figs. 6A 6B) , in the nuclear layer (Fig. 6C) or in the nerve fiber layer (Fig. 6D) . 

Discussion

The idea of detecting lipid peroxides by their ability to oxidize an indicator substance was discussed by Sehrt in 1927. ²⁹ Tetramethylbenzidine has been oxidized by lipid peroxides in vitro ^{30 31} and in fixed tissue. ^{26 32} The ability of TMB to react with lipid peroxides was shown spectrophotometrically by Thomas et al. ³¹ and in our laboratory ²⁶ by a staining of synthesized lipid peroxides with TMB. Lipid peroxides have been generated by the enzyme lipoxygenase ³³ and isolated from the unreacted fatty acids by thin-layer chromatography. The staining reaction is selective for peroxidized lipids. In the same publication it was shown that biochemically generated lipid peroxides in pig retinas using lipoxygenase and light-induced lipid peroxides ^{10 11 12} in hamster retinas are detected as electron-dense structures, whereas in untreated animals, these structures are absent. ²⁶  

A high intracellular oxygen concentration ³⁴ and exposure to light and high levels of polyunsaturated fatty acids ⁷ render the ROS particularly susceptible to lipid peroxidation. ³⁵ In albino rats, Wiegand et al. ¹⁰ detected light-induced lipid peroxides in the retinas after an irradiation of approximately 1000 lux for 24 hours, and Kagan et al. ¹¹ obtained the same result using a light protocol of approximately 10,000 lux for 3.5 hours. Organisciak et al. ³⁶ describe a 13% increase of hydroperoxides in the retinas of adult albino rats after an irradiation of approximately 3500 lux for only 1 hour, but the increase was not rated as significant. 

In the present study, the light regimen that led to the bubblelike or spherical electron-dense structures was an irradiation of pigmented rats with nondilated pupils using approximately 6000 lux for 24 hours or an illumination of eyes with dilated pupils using 150,000 lux for 1 hour. Considering the detection of lipid peroxides in albino rats after light exposure by Wiegand et al. ¹⁰ and Kagan et al. ¹² and other evidence ^{11 12 36 37 38 39 40 41 42 43} for an involvement of peroxidation in light-induced retinal damage, the light protocols applied in the present study probably generated lipid peroxides in the rat retina. 

The electron-dense bubblelike and spherical structures were seen exclusively in the ROS and only in rats exposed to an intense illumination (either 6,000 lux or 150,000 lux) and treated with TMB. The appearance of BRS exclusively in the ROS is consistent with findings of Organisciak et al., ³⁶ who state that after light-induced peroxidation in vivo for 1 hour (3000–4000 lux), within the retina of albino rats, the photoreceptor cell contains at least a twofold higher concentration of hydroperoxides than the remaining retina. They conclude that the photoreceptor cell is a major site of hydroperoxide formation in the retina. 

Electron-dense sites of disc membranes that appeared to represent early stages of bubblelike structures emerged after soft illumination (50 lux) for 1 hour, if the pupils of the anesthetized animals were dilated rendering the retina unprotected from the deleterious effect of light (Fig. 4) . In animals with nondilated pupils kept under a cycling light of 50 lux, no such electron-dense structures were seen, whereas the intense illumination of eyes with dilated pupils resulted in the emergence of the described bubblelike electron-dense structures. These findings emphasize the dose–response relationship between the light protocol and the detection of BRS. Because ROS are especially susceptible to lipid peroxidation it is useful to work on the assumption that, by far, most of the hydroperoxides generated under light-induced peroxidate conditions consist of lipid peroxides. Taken together, it is likely that these electron-dense structures, detected in the ROS as BRS, represent lipid peroxides, although it cannot be ruled out that peroxidized proteins and carbohydrates additionally appear as BRS. 

The susceptibility to light damage depends on various factors, ^{2 44 45 46} and earlier studies have indicated that the distribution of photoreceptor cell destruction after light-induced damage was not uniform throughout the retina, and the central regions of the retina seem to be more susceptible compared with the far periphery. ⁴⁶ In albino and pigmented rats, Rapp and Williams ^{46 47} found the superior hemisphere to be more sensitive to light damage than the inferior region, with a particularly sensitive region 1 to 2 mm superior to the optic nerve head. This is the area from which the samples for the present study were taken. 

The detection of BRS in eyes with dilated pupils after illumination of 50 lux for 1 hour is consistent with the findings of Williams et al., ⁴⁸ who demonstrated that light intensities as low as 100 lux are sufficient to cause photoreceptor cell death in the retina of pigmented animals with dilated pupils and in albino rats with nondilated pupils. However, the present study shows that pigmented rats with nondilated pupils resisted even an irradiation of 6000 lux for 12 hours without generating BRS. 

The irregularities of the disc membranes in the control retina of animals kept under a cycling light of 50 lux without dilation of the pupils or additional irradiation (Fig. 3A) is not because of light damage but because of the overnight incubation with TMB (pH 3.0) before the fixation with osmium tetroxide. In rats kept under the same light conditions and treated the same but without overnight incubation with the BRS mixture, these irregularities were reduced to a minimum (Fig. 3B) . The specimens of the present study were fixed by immersion fixation using glutaraldehyde. ²⁷ The remaining aberrations of the disc morphology depicted in Figure 3B may be abolished by fixation through cardiac perfusion. ²⁷  

The typical signs of light damage are edema, disruption and fragmentation of ROS; and shrinking of inner segments and nuclear pyknosis, leading to complete degeneration and photoreceptor cell death evidenced by the disappearance of the cell bodies. Even the entire receptor layer may be missing. ⁴⁶ Despite even extreme light conditions, however, almost none of these indications of light damage were present in the recent study. This is consistent with both Li et al., ⁴⁹ who state that cell death because of constant light takes some time to manifest itself after the initial light insult, and Shvedova et al., ⁵⁰ who state that no ultrastructural changes were apparent after an incubation of frog retinas with Fe^II and ascorbate for 20 minutes, decreasing the electroretinogram amplitude and correlating with the appearance of lipid peroxides. 

BRS appeared as bubblelike electron-dense structures that seemed to extrude from the disc membrane or as spherical structures that located in spaces between individual discs. Possibly, both appearances represented the same structures and differed only in the plane of the section. The continuity of the discs was interrupted at sites at which the electron-dense structures appeared, and several discs seemed to contribute to one electron-dense bubblelike structure. These observations suggest that the membrane integrity of the disc membranes is affected by the peroxidation and subsequent incubation with TMB at pH 3.0. Nevertheless, this effect was specific for irradiated retinas and did not occur in untreated animals. Furthermore, these structures appeared only after the incubation with the BRS mixture and were not seen in irradiated eyes without incubation with TMB. 

A possible explanation for these changes in the structure of the disc membranes could be the fatty acid composition of the membranes. The properties of a lipid bilayer depend on its composition, ⁵¹ and lipid peroxidation has damaging effects on membrane integrity. ^{52 53} Peroxidative stress is known to increase lipid peroxides and to decrease the amount of polyunsaturated fatty acids in lipid membranes, ^{10 11 12 13} both resulting in an increased osmotic fragility of lipid membranes ^{54 55} and particularly in the loss of the structural integrity of rod outer segment membranes. ³⁵ However, these considerations appear not to be sufficient in explaining the findings, because the emergence of the described electron-dense structures depends strongly on the use of the BRS mixture comprising TMB. And an incubation of irradiated retinas with the BRS buffer (pH 3.0) without using TMB does not result in the formation of these electron-dense bubblelike structures. Because TMB is able to react with hydroperoxides, ^{26 30 31 32} it seems likely that the peroxidative property of the detected structures causes this effect. Even the involvement of enzymes cannot be ruled out, because the incubation with the BRS mixture is performed after a merely soft fixation by 2% glutaraldehyde for 2 hours, rendering at least part of the cellular enzymes still functional. ⁵⁶ With an increasing time of fixation, the appearance of BRS vanishes (data not shown). Alternatively, this can be explained by a sole physical process, requiring a certain flexibility of the lipid membranes, decreasing with prolonged fixation. 

The applied procedure was performed after an improved protocol to localize lipid peroxides in pig and hamster retinas. ²⁶ Yet, this method failed to detect light-induced lipid peroxidation in the retina of rats, the most commonly used laboratory animal. The main improvements to the original procedure relate to the light exposure protocol and to the buffer system. In the present study it was shown that the light intensity (1000 lux) and the illumination time (12 hours) applied in the previous study ²⁶ were insufficient for rats with nondilated pupils to generate a peroxidation level in the retina that could be detected as BRS. A further improvement was the change in the buffer system for the BRS mixture. McIlvaine’s Na₂HPO₄–citric acid buffer allowed TMB to dissolve in pure citric acid, because it dissolved only poorly at pH 3.0. The pH was then adjusted to 3.0 by adding Na₂HPO_4; 3.0 is the optimal pH for the localization of lipid peroxides as BRS, using TMB (Peter Kayatz unpublished pilot tests, 1997). The modification of the buffer system resulted in an easier preparation of the BRS mixture and a more sensitive detection of BRS. 

The presented protocol renders an uncomplicated tool for generation and ultrastructural localization of lipid peroxides in the rat retina. The application of the BRS method to the rat offers, for the first time, the possibility of investigating the formation, decomposition, or transportation of lipid peroxides at an ultrastructural level in a commonly used laboratory animal. 

 Supported in part by Deutsche Forschungsgemeinschaft Grants He/840/5-2 and Es 82/5-3, the Retinovit Foundation, Propter Homines Foundation, and the Koeln Fortune Program.

 Submitted for publication January 21, 1999; revised April 6, 1999; accepted April 20, 1999.

 Proprietary interest category: N.

 Corresponding author: Peter Kayatz, Department of Vitreoretinal Surgery, University of Cologne, Joseph Stelzmann Str. 9, 50931 Cologne, Germany. E-mail: [email protected]

 

Figure 1.

 

(A) Overview of choroid, RPE, and the photoreceptor cell layer. In rats with nondilated pupils exposed to intense light (approximately 6000 lux) for 24 hours, BRS (arrows) appeared after incubation with TMB as electron-dense structures exclusively throughout the ROS but were missing in choroid, RPE, RIS, and ONL. (B) Higher magnification of RPE and ROS of the same eye. BRS (arrows) were seen exclusively in the ROS. No BRS appeared in the RPE. BM, Bruch’s membrane; CH, choroid; E, erythrocyte; M, mitochondrion; MV, microvilli of the retinal pigment epithelium; N, nucleus of RPE cell; ONL, outer nuclear layer; RIS, rod inner segments; ROS, rod outer segments; RPE, retinal pigment epithelium.

Figure 2.

 

(A) BRS in the ROS appeared as electron-dense sites of disc membranes (arrowheads) or bubblelike structures (asterisks). The electron-dense sites seemed to extrude from the discs to form spherical structures that located in spaces between individual discs (large arrows), but possibly, both appearances represented the same structures and differed only in the plane of the section. The continuity of the discs was interrupted at sites where the electron-dense structures appeared (small arrows). (B) A high magnification of the BRS reveals that several discs seemed to contribute to one electron-dense bubblelike structure (arrows) and the interior of the bubbles appeared to be electron-lucent.

Figure 3.

 

(A) In rats kept at physiological light conditions (50 lux) with nondilated pupils, BRS were absent in the ROS after incubation with the BRS mixture. The ROS exhibited swollen, irregular disc membranes, resembling light damage, but these aberrations were caused by incubation with the BRS mixture and not by light damage. (B) In animals treated by the same light protocol without using the BRS mixture, the irregularities were reduced to a minimum. BM, Bruch’s membrane; MV, microvilli of the retinal pigment epithelium; N, nucleus; ROS, rod outer segments; RPE, retinal pigment epithelium.

Figure 4.

 

In rats with dilated pupils, a light intensity of 50 lux for 1 hour resulted in the formation of electron-dense sites of disc membranes throughout the ROS (arrowheads). These electron-dense sites seemed to represent early stages of bubblelike and spherical electron-dense structures that were not detectable using this soft-light regimen. MV, microvilli of the retinal pigment epithelium; ROS, rod outer segments.

Figure 5.

 

(A) In rats with dilated pupils, intense irradiation (150,000 lux) for 1 hour resulted in the formation of high numbers of bubblelike and spherical electron-dense structures (arrowheads) throughout the entire ROS. (B) A higher magnification reveals that these electron-dense structures resembled the structures that appeared after illumination of 6000 lux for 24 hours (Figure 2A) . (C) Electron-dense structures (arrowheads) in the tips of ROS adjacent to the microvilli of the RPE. The electron-dense structures emerged exclusively in the ROS. Other membranous structures, such as the microvilli of the RPE or the basal labyrinth of the RPE, did not contain electron-dense BRS. BL, basal labyrinth of the RPE; BM, Bruch’s membrane; E, erythrocyte; M, mitochondrion; MV, microvilli of the RPE; N, nucleus of an RPE cell; ROS, rod outer segments; RPE, retinal pigment epithelium;

Figure 6.

 

(A) Electron-dense structures appeared exclusively in the ROS. The basal part of the ROS comprises electron-dense structures (arrowheads). Rod inner segments (RIS) do not show these structures. (B) A higher magnification of the framed area in (A). Within the layer of the RIS, there was an isolated ROS, again comprising electron-dense, bubblelike, and spherical structures, whereas the surrounding RIS were free of electron-dense structures. (C) The outer nuclear layer (ONL) next to the outer plexiform layer (OPL). No such electron-dense structures were visible. (D) Part of the ganglion cell layer (GCL) and the nerve fiber layer (NFL) separated from the vitreous body (VB) by the inner limiting membrane (ILM). Here again, no such bubblelike or spherical electron-dense structures were detectable, confirming the specificity of these reaction products to the ROS. The dark grains visible in the GCL represent ribosomes. Burst mitochondria and other tissue damage was caused by overnight incubation with the BRS mixture before fixation with osmium tetroxide. E, erythrocyte; ILM, inner limiting membrane; L, lumen of a capillary; M, mitochondrion; N, nucleus; VB, vitreous body.

The authors thank Andrea Bieker for excellent technical assistance and Ian Edwards for language correction.