Delayed Dark-Adaptation and Lipofuscin Accumulation in abcr+/− Mice: Implications for Involvement of ABCR in Age-Related Macular Degeneration

Abstract

purpose. To examine the ocular phenotype in mice heterozygous for a null mutation in the abcr gene.

methods. Retinas and retinal pigment epithelia (RPE) were prepared from wild-type, abcr+/−, and abcr−/− mice. Fresh tissues were homogenized and analyzed by normal phase high-performance liquid chromatography (HPLC) for the presence of retinoids and phospholipids. In another study, fixed tissues were sectioned and analyzed by light and electron microscopy. Finally, anesthetized mice were studied by electroretinography (ERG) at different times after exposure to strong light.

results. A2E, the major fluorophore of lipofuscin, and its precursors, A2PE-H₂ and A2PE, were approximately fourfold more abundant in 8-month-old abcr+/− than in the wild-type retina and RPE. The levels of these substances in abcr+/− mice were approximately 40% those in abcr−/− mice. Lipofuscin pigment-granules were also visible in abcr+/− RPE cells by electron microscopy. Accumulation of A2PE-H₂ and A2E in abcr+/− retina and RPE, respectively, was strongly dependent on light exposure. Heterozygous mutants also exhibited delayed recovery of rod sensitivity by ERG. This delay was correlated with elevated levels of all-trans-retinaldehyde (all-trans-RAL) in retina after a photobleach and was not caused by a reduction in quantum-catch due to depletion of 11-cis-retinaldehyde (11-cis-RAL).

conclusions. Partial loss of the ABCR or rim protein is sufficient to cause a phenotype in mice similar to recessive Stargardt’s disease (STGD) and age-related macular degeneration (AMD) in humans. These data are consistent with the suggestion that the STGD carrier-state may predispose to the development of AMD.

Mutations in the ABCR gene are responsible for STGD, ^{1 2} a blinding disorder of children characterized by delayed dark-adaptation and accelerated deposition of lipofuscin in the RPE. ^{3 4} A similar pattern is seen in AMD, a common cause of visual loss in the elderly. ^{5 6 7 8 9} Heterozygous mutations in ABCR have been associated with AMD in several studies, ^{10 11 12} but the significance of this association has been challenged. ^{2 13 14}  

The ABCR gene encodes rim protein (RmP), an ATP-binding-cassette transporter in the rims of photoreceptor outer-segment (OS) discs. ^{15 16 17} The transported substrate for RmP is unknown. Based on the results of reconstitution studies and the biochemical phenotype in abcr−/− mice, it has been suggested that RmP functions as a flippase for N-retinylidene-phosphatidylethanolamine (APE), the normally occurring Schiff-base conjugate of phosphatidylethanolamine with all-trans-RAL. ^{18 19 20} RmP may accelerate recovery of rod sensitivity after light exposure by removing all-trans-RAL from the disc interior. ¹⁹  

Accumulation of lipofuscin in cells of the RPE is observed in several forms of macular degeneration including STGD and AMD. ^{4 7} Slow accumulation of lipofuscin is also seen during normal aging. ²¹ A major fluorophore of lipofuscin is the bis-retinoid, N-retinylidene-N-retinylethanolamine (A2E). ^{22 23} A2E and its precursors, dihydro-N-retinylidene-N-retinyl phosphatidylethanolamine (A2PE-H₂) and N-retinylidene-N-retinyl phosphatidylethanolamine (A2PE), are present at dramatically higher levels in ocular tissues from abcr−/− mice and humans with STGD than in age-matched controls. ^{19 24} Thus, an additional role of RmP may be to prevent A2E deposition in RPE cells by eliminating its precursors from photoreceptor OS. 

In the current work, we examined the ocular phenotype in mice heterozygous for a null allele of abcr. We examined abcr+/− mice biochemically, for accumulation of A2E and its precursors in ocular tissues, by ERG, for evidence of delayed dark-adaptation, and histologically, for evidence of photoreceptor degeneration and lipofuscin accumulation in the RPE. 

Methods

Mice

Wild-type (strains B6 × D2F1 and B6 × 129F1), abcr+/− (strain B6 × 129F1), and abcr−/− (strain B6 × 129F1) mice were raised from birth under 12-hour cyclic illumination (25–30 lux) or under total darkness in a ventilated cabinet for the indicated times. Genotypes of the mice were determined by Southern blotting as described. ¹⁹ All studies were conducted in accordance with the NIH guidelines and the ARVO statement on the Care and Use of Animals in Ophthalmic and Vision Research. 

Tissue Preparation and Extraction

Mice were anesthetized with intraperitoneal ketamine (200 mg/kg) plus xylazine (10 mg/kg) and killed by cervical dislocation. Eyes were immediately enucleated and hemisected, and the posterior segments were placed in ice-cold PBS (pH 7.2). Retinas and remaining RPE/eyecups were trimmed of excess tissue and homogenized separately in 1 ml of PBS. For analysis of phospholipids, 1 ml of chloroform/methanol (2:1, v/v) was added to each homogenate and the samples were re-homogenized. APE, A2PE-H₂, A2PE, and A2E were extracted from the samples after addition of 4 ml of chloroform and 3 ml water. The samples were centrifuged at 1500g for 10 minutes and the organic phases were removed. Extraction was repeated and the pooled organic phases were dried under a stream of argon. For analysis of retinaldehydes, tissues were homogenized in 0.1 M KH₂PO₄ (pH 7.0) containing 6.0 M formaldehyde. Two ml of methylene chloride was added to the homogenates followed by incubation at 30°C for 10 minutes and extraction with methylene chloride-hexane. After evaporation, sample residues were resuspended in 200 μl hexane and analyzed by HPLC. 

HPLC Analysis

APE, A2PE-H₂, A2PE, and A2E were analyzed by normal-phase HPLC as previously described. ²⁴ 11-cis-RAL and all-trans-RAL were analyzed by normal phase HPLC as described. ¹⁹ Spectral data were obtained (210–450 nm) for all eluted peaks. Quantitation of sample peaks was performed by area-unit versus concentration-slope coefficients, determined with authentic standards immediately before sample analysis. 

ERG Analysis

Mice were dark-adapted overnight and anesthetized with ketamine plus xylazine, and pupils were dilated by topical application of 1.0% atropine sulfate. Anesthetized mice were kept on a heating pad at 37°C during recordings. Full-field ERGs were obtained in a Ganzfeld dome using a gold coil wire on the corneal surface overlaid with 1% methylcellulose, a reference electrode of the same material in the mouth, and a needle electrode in the tail to serve as a ground. A high-intensity flash unit (Novatron, Dallas, TX) provided short-wavelength flashes (Kodak Wratten 47B, Sigma Chemical Co., St. Louis, MO) from 1 to 3.4 log scot-td · sec in 0.3 log unit steps. Initially, a-wave responses were obtained in the dark-adapted state. Mice were then exposed to white light at an intensity of 400 lux in the Ganzfeld dome for 5 minutes. After this photobleach, mice were returned to darkness and analyzed by ERG to measure recovery of rod sensitivity. The leading edge of the a-waves was fit (as an ensemble) by the Lamb and Pugh model for the activation phase of the phototransduction. ²⁵ The a-wave maximal responses (Rm_P3) and the amplification constants (S) were calculated from this model. 

Light and Electron Microscopy

Mice were anesthetized with ketamine plus xylazine and perfused through the heart with 1% glutaraldehyde and 2% paraformaldehyde in PBS (pH 7.4). Fixed eyes were removed and sectioned along the ora serrata, and eyecups were immersed in 2% glutaraldehyde and 2% paraformaldehyde in 100 mM cacodylate buffer (pH 7.4) overnight at 4°C. Eyecups were dehydrated in an ethanol series to 100%, embedded in Poly/Bed 812 media (Polysciences, Inc., Warrington, PA), and polymerized at 60°C for 48 hours. For light microscopy, 0.5-μm sections were stained with 1% toluidine blue. For electron microscopy, 60-nm sections were stained with 5% uranyl acetate and lead citrate before examination. For quantitation of photoreceptor nuclei, 0.5-μm sections of retina from 15-month-old wild-type, abcr+/−, and abcr−/− mice were scanned by light microscopy with a digital camera. Photoreceptor nuclei were counted in the central retina (400 μm from the optic nerve) using Metamorph software (Universal Imaging Corp., West Chester, PA). The numbers of nuclei were normalized to a width of 100 μm along the outer nuclear layer (ONL). 

Results

Accumulation of A2E and Its Precursors

Because lipofuscin accumulation is a pathologic feature of AMD, ^{5 9 26} we analyzed retina and RPE from wild-type, abcr+/−, and abcr−/− mice for presence of the lipofuscin fluorophores: A2PE-H₂, A2PE, and A2E (Figs. 1A 1B 1C 1D) . In both mutants, A2PE-H₂ was present in retina and RPE whereas A2PE and A2E were only detectable in RPE. The level of APE in light-adapted abcr+/− retinas was twofold higher than in wild-type retinas (not shown), in contrast to 2.6-fold higher in abcr−/− retinas. ²⁴ The levels of A2E and its bis-retinoid precursors in abcr+/− mice were generally approximately 40% those of abcr−/− mutants and several-fold higher than in wild-type mice. A2PE-H₂ in RPE was an exception, with a level approximately sixfold higher in abcr−/− than in abcr+/− mutants. This suggests that the rate of A2PE-H₂ conversion to A2E is slower than the rate of its accumulation in phagolysosomes. The accumulation of both A2PE-H₂ in retina and A2E in RPE was dramatically higher in mice raised under 12-hour cyclic lighting compared with mice raised in total darkness (Figs. 1E 1F) . Thus, photoisomerization of visual pigment may be required for the formation of A2PE-H₂ and A2E. 

Delayed Dark Adaptation in abcr+/− Mice

Full-field ERGs were performed on 6-month-old wild-type and abcr+/− mice. No significant differences in Rm_P3 were observed between dark-adapted mice of the two genotypes. To test the rate of recovery after a photobleach, we exposed mice of both genotypes to 400 lux illumination for 5 minutes. Mice were then returned to darkness and ERGs were performed at 10-minute intervals for up to 1 hour. Full recovery of rod sensitivity was observed after 40 minutes in wild-type mice (Fig. 2) . In contrast, age-matched abcr+/− mice recovered only∼ 75% of prebleach sensitivity at 40 minutes. Although abcr+/− mice did not recover full sensitivity before awakening from anesthesia (∼60 minutes after the bleach), full restoration of sensitivity was observed in similarly treated mice after overnight dark adaptation (not shown). At 10 minutes after the photobleach, the phototransduction gain parameter (S) was reduced∼ 50% in both wild-type and abcr+/− mice, and returned to prebleach levels by 50 minutes. No significant differences in S were observed between wild-type and abcr+/− mice at any time points studied. 

To address the biochemical cause of delayed dark adaptation in abcr+/− mice, we measured the levels of 11-cis-RAL and all-trans-RAL by HPLC analysis in retinas from wild-type and abcr+/− mice after similar light exposure. No significant difference in the levels of 11-cis-RAL were observed between dark-adapted wild-type and abcr+/− mice at any time points (Fig. 3A) . However, we observed significantly higher levels of all-trans-RAL in abcr+/− than in wild-type retinas at all time points after the photobleach from 5 to 60 minutes (Fig. 3B) . This pattern is similar to that observed previously in abcr−/− mice. ¹⁹  

Accumulation of Lipofuscin in abcr+/− and abcr−/− RPE Cells

Retina sections from 6-month-old wild-type, abcr+/−, and abcr−/− mice were examined histologically (Figs. 4A 4B 4C) . No significant differences were observed between animals in the number of photoreceptor nuclei. However, thickening of the RPE cell layer was observed in abcr+/− and abcr−/− retinas (Figs. 4B 4C) . Also, OS were shorter in the abcr−/− retina (Fig. 4C) . To test for possible photoreceptor degeneration, we counted photoreceptor nuclei along a 100-μm width of ONL from 15-month-old wild-type, abcr+/−, and abcr−/− retinas. No significant difference was observed in the numbers of nuclei between mice of the three genotypes (Fig. 4D) , indicating no photoreceptor degeneration. 

We examined retinal sections from 6-month-old mice of the same genotypes by electron microscopy (Figs. 5A 5B 5C) . The most prominent ultrastructural change in abcr+/− and abcr−/− RPE was the presence of numerous, irregularly shaped dense bodies in the basal region of the cells. These structures resemble lipofuscin granules in postmortem RPE tissue from patients with STGD and AMD. ^{4 21} Disorganization of the basal processes adjacent to Bruch’s membrane was also seen in mutant RPE cells. Another ultrastructural change in abcr+/− and abcr−/− mice was the partial redistribution of melanosomes from apical processes to the cytoplasm of RPE cells. Finally, RPE cells were thicker in the mutants. These ultrastructural changes were slightly more severe in abcr−/− compared to abcr+/− mice. OS discs appeared normal in both mutants. 

Discussion

This article presents the phenotype in abcr+/− mice. One reason for studying these animals is that heterozygous mutations in the human ABCR gene have been associated with AMD in a subset of cases. A clinical feature of both AMD and STGD is delayed recovery of rod sensitivity after light exposure. ^{3 6 8 27} Here, we observed significantly delayed dark adaptation in abcr+/− mice. Our analysis of the retinoid profiles in wild-type and abcr+/− mice after a photobleach offers clues about the etiology of this delayed dark adaptation. First, we can rule-out reduced quantum-catch due to depletion of 11-cis-RAL as a possible explanation, because the levels of 11-cis-RAL were similar in wild-type and abcr+/− retinas (Fig. 3A) . On the other hand, clearance of all-trans-RAL was significantly delayed in abcr+/− retinas after a photobleach (Fig. 3B) . A similar pattern was observed in abcr−/− mice. ¹⁹ All-trans-RAL has been shown to interact spontaneously with opsin apoprotein to form a noncovalent complex that activates the transduction cascade with at least 10⁵-fold greater efficiency than opsin alone and nearly 10% the efficiency of metarhodopsin II. ^{28 29 30 31} The presence of the opsin/all-trans-RAL complex may explain delayed dark adaptation in abcr+/− mice due to a desensitizing background of “equivalent light.” The similar kinetics of delayed dark adaptation (Fig. 2) and delayed clearance of all-trans-RAL (Fig. 3B) corroborate this explanation, if we assume that it takes several minutes after a photobleach for the“ noisy” photoproduct to accumulate. 

Another aspect of the phenotype in abcr+/− mice is age-dependent accumulation A2E within the RPE. A2E, the major fluorophore of lipofuscin, forms in a four-step process involving condensation of all-trans-RAL with phosphatidylethanolamine to yield APE, secondary condensation of APE with another all-trans-RAL to yield the bis-retinoid, A2PE-H₂, oxidation of A2PE-H₂ to A2PE, and final hydrolysis of the phosphate ester to yield A2E. ^{24 32} Elevations in the A2E precursors: all-trans-RAL, APE, A2PE-H₂, and A2PE were also observed in abcr+/− retina and RPE, consistent with this scheme. Accumulation of A2E was almost completely suppressed in abcr+/− mice raised in total darkness, suggesting dependence of A2E formation on the presence of all-trans-RAL produced by photoisomerization. A2E has been shown to inhibit lysosomal proteolysis in RPE cells. ^{33 34} At high concentrations, A2E acts as a cationic detergent dissolving cellular membranes. ^{35 36 37}  

A possible mechanism for the degeneration of photoreceptors and resulting blindness in STGD is that the RPE degenerates due to accumulation of A2E,¤ and that photoreceptors die secondarily because of loss of the RPE support-role. An observation that conflicts with this model is that virtually no photoreceptor degeneration was observed in abcr+/− or abcr−/− mice up to 15 months of age. Given the observed RPE changes, why are photoreceptors not degenerating? An important difference between mouse and human retinas is the presence of a macula in humans. The density of rod photoreceptors is several-fold higher in the perifoveal macula compared with the peripheral retina. ³⁸ Also, in a study of aged postmortem retinas, the concentration of lipofuscin was highest in RPE cells overlying the perifovea. ³⁹ Thus, the rate of lipofuscin accumulation is correlated with the ratio of OS to RPE cells. Further evidence for heightened vulnerability of the macula is that degeneration of the entire retina is seen with more severe alleles of ABCR, in retinitis pigmentosa and cone-rod dystrophy, whereas milder alleles are associated with more limited degeneration of the macula, in STGD. ^{40 41 42} Thus, the absence of photoreceptor degeneration in mice may be related to the lack of a macula. Another consideration is that in even the most severe of ABCR-mediated diseases, photoreceptor degeneration only becomes clinically significant after years to decades of life, far longer than the 15 months examined here. 

The data presented in this study establish that a partial reduction in the level of RmP is sufficient to cause a retinal phenotype in mice. This phenotype bears similarities to AMD in humans, including delayed dark adaptation and lipofuscin accumulation by the RPE. Given the very slow rate of photoreceptor loss in AMD, the absence of photoreceptor degeneration by 15 months in abcr+/− mice might be expected. The earliest histopathologic change in AMD is the development of basal deposits (drusen) between the RPE and Bruch’s membrane. ^{7 43} Ultrastructurally, we observed changes in the basal RPE adjacent to Bruch’s membrane in both abcr+/− and abcr−/− mice, but no drusen (Fig. 5) . Although the origin of drusen is unknown, these deposits contain lysosomal and cytoplasmic debris from RPE cells. ^{44 45} In a recent study of AMD by scanning laser ophthalmoscopy, drusen were shown to exhibit autofluorescent properties similar to those of lipofuscin. ⁹ Thus, drusen may represent lipofuscin-containing debris after degeneration of RPE cells. Lipofuscin was abundantly present in RPE from abcr+/− and abcr−/− mice. The absence of drusen in abcr+/− mice may reflect the large difference in time scales (months versus decades) over which the disease process develops in mice compared with humans. Alternatively, it may reflect an altogether different disease process. Choroidal neovascularization (invasion of choroidal vessels through the RPE into the retina) is another pathologic feature of AMD not seen in abcr+/− mice. However, because choroidal neovascularization is seen in <10% of younger patients with AMD, ⁷ its absence in abcr+/− mice also may not be important. 

In summary, our results suggest that heterozygous-null mutations in the human ABCR gene may cause a clinical picture that resembles STGD but with slower progression. Given the similarity between STGD and AMD, these results are consistent with the proposal that the STGD carrier-state predisposes to the development of AMD. ^{10 11 12} However, the results do not speak to the prevalence of this association in humans. If mutations in ABCR are responsible for a subset of AMD, this would represent another instance where a homozygous state causes severe recessive disease in children, and the heterozygous state predisposes to a milder disease of the aged. The abcr+/− mouse may be a useful animal model to develop new therapies for AMD, especially pharmacologic interventions that suppress lipofuscin accumulation in RPE cells. 

 Supported by grants from the National Eye Institute, the Foundation Fighting Blindness, the Macula Vision Research Foundation, and the Steinbach Fund.

 Submitted for publication August 9, 2000; revised January 11, 2001; accepted February 16, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Distribution and light-dependent accumulation of A2E and A2E precursors in wild-type, abcr+/−, and abcr−/− retina and RPE. (A) Representative chromatogram of a phospholipid extract from 8-month-old abcr+/− RPE in peak-height absorbance units × 10⁻³ (mAbs) at detection wavelength 450 nm. Labeled peaks corresponding to A2PE, A2PE-H₂, and A2E were identified spectrally. ^{19 24} (B through D) Levels of A2PE-H₂ (B), A2PE (C), and A2E (D) in retina (white bars) and RPE (black bars) from 8-month-old wild-type (wt), abcr+/− (+/− ), or abcr−/− (−/−) mice are shown as mean area units × 10⁻³ (mAU) per eye (A2PE-H₂ and A2PE) or picomoles per eye (A2E) ± SDs (n = 3–4). (E) A2PE-H₂ levels (mAU per eye) in abcr+/− retinas from mice of the indicated ages raised under 12-hour cyclic light (▵) or total darkness (▴). (F) A2E levels (picomoles per eye) in abcr+/− RPE from mice of the indicated ages raised under cyclic light (▵) or total darkness (▴). Values for (E) and (F) are shown as the mean ± SD (n = 3). *Significant difference between the cyclic-light and dark-reared values (Student’s t-test, P < 0.05).

Figure 2.

 

ERG analysis showing recovery of rod sensitivity in 6-month-old abcr+/− compared with wild-type mice after 5-minute exposure to 400-lux illumination. Data are plotted as the mean ratio of observed to dark-adapted Rm_P3 values (normalized Rm_P3 amplitude) ± SE. *Significant difference between the values for abcr+/− (○) and wild-type (•) mice (Student’s t-test, P < 0.05). The dashed line indicates full recovery of dark-adapted rod sensitivity.

Figure 3.

 

Retinoid levels in 6- to 8-month-old wild-type and abcr+/− retinas after a photobleach. 11-cis-RAL (A) and all-trans-RAL (B) are shown in picomoles per eye ± SD (n = 4). Determinations were made in dark-adapted (DA) mice, immediately after a 5-minute 400-lux photobleach (BL) and at the indicated times in darkness after the photobleach. *Significant difference between the abcr+/− and wild-type values (Student’s t-test, P < 0.05).

Figure 4.

 

Light microscopic analysis of outer retinas from 6-month-old wild-type (A), abcr+/− (B), and abcr−/− (C) mice. RPE, OS, inner segment (IS), and ONL are indicated. Scale bar, (A) 20 μm. Micrographs were obtained at the same magnification. Note the similar ONL thickness in all three panels. Also note the thickening of RPE cell-bodies and the slight shortening of OS in (C). (D) Histogram showing the average number of photoreceptor nuclei per 100 μm of ONL from the central retinas of 15-month-old wild-type (n = 4), abcr+/− (n = 3), and abcr−/− (n = 4) mice.

Figure 5.

 

Electron microscopic analysis of the RPE in 6-month-old wild-type (A), abcr+/− (B), and abcr−/− (C) mice. The RPE and OS layers are indicated. Bruch’s membrane is visible immediately above the basal surface of the RPE layer. Scale bar, (A) 2.0 μm. All micrographs were obtained at the same magnification. Note the (i) predominantly apical distribution of the large oval melanosomes in wild-type and predominantly cytoplasmic distribution in abcr+/− and abcr−/− RPE; (ii) presence of small, irregular, dense bodies (white arrows) near the basal region of abcr+/− and abcr−/− RPE; (iii) thickening and disorganization of the basal RPE underlying Bruch’s membrane in abcr+/− and abcr−/− mice; and (iv) thickening of the RPE cell-bodies in abcr+/− and abcr−/− mice.

The authors gratefully acknowledge Roxana Radu for her outstanding technical assistance and Sassan Azarian and Wojciech Kedzierski for their valuable comments on the manuscript.