Acute Radiolabeling of Retinoids in Eye Tissues of Normal and Rpe65-Deficient Mice

Abstract

purpose. Mice with a targeted disruption of the gene encoding RPE65, a protein ordinarily highly expressed in the retinal pigment epithelium (RPE), accumulate abnormally high levels of all-trans retinyl ester in the RPE and exhibit very little 11-cis retinal in the retina. The present study was undertaken to determine whether the Rpe65-deficient mouse exhibits an abnormal flux of retinoid between the systemic circulation and the eye tissues.

methods. Dark-adapted Rpe65-deficient mice (Rpe65 ^−/−) and wild-type control mice (Rpe65 ^+/+) of approximate ages 1 and 3 months received an intraperitoneal injection of all-trans (³H)retinol. The mice were maintained in darkness for a defined period (≈1.5, 4.5, 24, or 48 hours) and then anesthetized, exsanguinated, and killed. Retinoids contained in the retina, RPE, serum, and liver were extracted and analyzed for (³H) radioactivity and molar level.

results. The specific activity (SA, in counts per minute per nanomole) of serum all-trans (³H)retinol in all mice exhibited a peak at postinjection times of 1.5 or 4.5 hours, and by 48 hours declined to approximately 7% or less of the peak. In Rpe65 ^+/+ mice, the average SA of RPE (³H)retinyl ester similarly exhibited an early peak (4.5 hours) and by 48 hours declined to approximately 6% to 10% of the peak. By contrast, the average SA of RPE (³H)retinyl ester in Rpe65 ^−/− mice exhibited a peak at 24 or 48 hours. Radioactivity and molar data for serum all-trans retinol and RPE retinyl ester obtained at 4.5 hours were analyzed to infer the molar influx of all-trans retinol from the circulation into the RPE. Levels of all-trans retinol influx derived from this analysis (mean ± SD: 0.014 ± 0.004 nmol in 1-month Rpe65 ^+/+ mice; 0.021 ± 0.009 nmol in 1-month Rpe65 ^−/− mice; 0.016 ± 0.013 nmol in 3-month Rpe65 ^+/+ mice; 0.026 ± 0.018 nmol in 3-month Rpe65 ^−/− mice) did not differ significantly from one another (P > 0.169). However, the inferred fractional influx (molar amount of entering all-trans retinol divided by the molar amount of RPE retinyl ester) in Rpe65 ^+/+ animals (0.34 ± 0.04 and 0.10 ± 0.03, respectively, for 1- and 3-month mice) substantially exceeded that for Rpe65 ^−/− animals (0.055 ± 0.023 and 0.015 ± 0.006, respectively, for 1- and 3-month mice). Significant levels of (³H)retinaldehydes were detected in the retinas of Rpe65 ^+/+ mice, but not in those of Rpe65 ^−/− mice, after the longer postinjection periods.

conclusions. The results indicate preservation of a substantial inward flux of all-trans retinol from the circulation into the RPE of Rpe65 ^−/− mice, despite the presence of abnormally high molar levels of RPE retinyl ester. They further imply the occurrence of a robust outward movement of all-trans retinol from the RPE into the circulation in Rpe65 ^+/+ mice, and substantial impairment of this efflux process in Rpe65 ^−/− mice. These findings raise the hypothesis that in normal RPE, 11-cis retinal and/or 11-cis retinol stimulate the efflux of all-trans retinol at the RPE basolateral membrane. In 3-month Rpe65 ^+/+ mice, the observed relationship between the SAs of retinaldehydes in the retina and of RPE retinyl ester is consistent with a last-in/first-out processing of all-trans retinol to 11-cis retinal within normally functioning RPE.

A major role of the retinal pigment epithelium (RPE) is the operation of the retinoid metabolic pathway required for the regeneration of visual pigment in the photoreceptors. This pathway, termed the visual cycle, involves the isomerization in darkness of all-trans retinoid produced by photoisomerization of the 11-cis retinal chromophore bound to opsin (for recent reviews, see Refs. ^{1 2} ). Perturbation of the visual cycle by targeted disruption of essential genes leads to the accumulation and/or impaired formation of retinoid intermediates, and determining the metabolic consequences of such disruptions can lead to new insights into the nature of this critical pathway. 

One model of genetically altered visual cycle function is that produced by disruption of the Rpe65 gene in the mouse. ³ RPE65, a protein preferentially expressed at high level in the RPE, ⁴ is thought to play a central role in the visual cycle. ^{3 5} Moreover, RPE65 is a known genetic locus for a group of severe early-onset retinal dystrophies in humans, including a subset of Leber congenital amaurosis. ^{6 7 8} Understanding the role of RPE65 in the visual cycle is thus of importance in developing possible therapeutic strategies. ^{9 10 11} The Rpe65-deficient (Rpe65 ^−/−) mouse exhibits pronounced visual cycle abnormalities, including a virtually complete lack of formation of 11-cis retinoids and a large accumulation of all-trans retinyl ester, ³ the putative substrate of isomerohydrolase-catalyzed all-trans to 11-cis isomerization. ^{12 13} The impairment of 11-cis retinal formation furthermore leads to a severe (>1000-fold) loss of rod photoreceptor sensitivity. ¹⁴  

All-trans retinol released from hepatic stores into the circulation and associated with serum retinol-binding protein (RBP) and transthyretin (TTR) is thought to serve as the primary source of retinoid for the RPE. ^{15 16} However, many aspects of the uptake of all-trans retinol from the systemic circulation and its metabolism within the RPE remain obscure. For example, it is not known whether this retinol uptake is receptor mediated ¹⁷ or occurs by the direct transfer of retinol across the RPE basolateral plasma membrane. ^{18 19} In addition, RPE itself expresses and secretes both RBP and TTR, ^{20 21 22} but the relationship of these processes to a possible efflux of all-trans retinol from the RPE back into the circulation remains unclear. Furthermore, relatively little quantitative information is available concerning the flux of all-trans retinol into the eye. The present study was undertaken to examine the kinetics of retinoid movement into and within the eye tissues of Rpe65 ^−/− mice and wild-type (Rpe65 ^+/+) control mice, by tracking the fate of systemically introduced, radiolabeled all-trans retinol. Preliminary results have been reported (Qtaishat NM, Redmond TM, Pepperberg DR, ARVO Abstract 1046, 2000). 

Methods

All-trans (³H)retinol [(11,12-³H)retinol, 30–50 Ci/mmol; or (20-³H)retinol, 72 Ci/mmol] in ethanolic solution supplemented with 1 mg/mL of α-tocopherol was purchased from NEN Life Science Products (Boston, MA) and stored at −20°C. Preparation of the all-trans (³H)retinol for systemic administration adhered to procedures similar to those described by Katz et al. ²³ Under dim red light, a defined volume of the commercially obtained all-trans (³H)retinol solution (3 μL per 200 μL final volume of the preparation) was combined in a test tube with Tween 40 (6 μL per 200 μL final volume; Sigma Chemical Co., St. Louis, MO) and phosphate-buffered saline (PBS, 191 μL per 200 μL final volume; product 70011-044; Life Technologies, Inc. [Gibco], Rockville, MD) and mixed by repeated passage through a pipette. 

All procedures conformed to the principles embodied in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Rpe65 knockout mice (Rpe65 ^−/−; abbreviated in the Results section as −/−) and controls (Rpe65 ^+/+; abbreviated as +/+) ³ were maintained on a 12-hour light-12-hour dark cycle and dark-adapted overnight before the experiment. Ages of the investigated animals fell within three groups: 3 to 5 weeks (average ± SD: 26 ± 3 days; n = 35), 2 to 3 months (86 ± 15 days; n = 42), and 10 to 13 months (371 ± 34 days; n = 14). For brevity, the ages of the mice in these groups will be referred to below as 1 month, 3 months, and 1 year. C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME) of age 2 to 3 months served as additional control animals. Under dim red light, the mouse received a single intraperitoneal injection of the mixture containing all-trans (³H)retinol. The injected volume was 133 μL for mice of 1 month of age, or 200 μL for mice of 3 months and 1 year of age. The time of injection defined time 0 in the experiment. The mouse was maintained in darkness for a defined period of 1 to 2 hours (termed 1.5 hours), 4 to 5 hours (termed 4.5 hours), 23 to 25 hours (termed 24 hours), or 47 to 49 hours (termed 48 hours), then anesthetized by an intraperitoneal injection of ketamine (0.15–0.18 mg/g body weight) and xylazine (0.004–0.006 mg/g body weight). A defined volume of blood (500 μL) was collected by cardiac puncture, and the mouse was killed by cervical dislocation. Data were obtained from 48 Rpe65 ^−/− and 43 Rpe65 ^+/+ mice, and from 8 C57BL/6J mice. 

Retinoids contained in the retina, RPE/choroid, serum, and liver were extracted and analyzed by normal-phase high-performance liquid chromatography (HPLC) in a system equipped with a flow scintillation analyzer (model 500TR; Packard Instrument Co., Meriden, CT), and the accompanying analysis software (Flow-One; Packard). The efficiency of counting ranged from 56% to 58%. Radioactivity data are quoted as measured counts per minute—that is, data are not corrected for the efficiency of counting. Samples obtained from the eyes of a given animal were combined, and results for ocular retinoids are quoted as the total amount in both eyes. Retinoids in the retina and RPE were determined by methods similar to those described by Qtaishat et al. ²⁴ (also see Refs. ^{25 26} ). The two retinas obtained from a given animal were placed in 250 μL of PBS and homogenized (1-mL manual tissue grinder; Wheaton Co., Millville, NJ). A 200-μL aliquot of this homogenate was withdrawn and added to 400 μL formaldehyde. To this mixture was added 800 μL isopropanol and 500 μL of water. The mixture was vortexed and then extracted three times with 2 mL hexane. The three hexane extracts were combined and evaporated under nitrogen and the residue redissolved in 200 μL of hexane. The entire 200-μL volume of this extract was analyzed by normal-phase HPLC (Adsorbosphere HS 5-μm silica column, 250 × 4.6 mm; Alltech, Deerfield, IL). RPE/choroid (henceforth termed RPE) obtained from both eyes were similarly added to 250 μL PBS and homogenized. A 200-μL aliquot of this homogenate was combined with 800 μL isopropanol and 500 μL of water, vortexed, and extracted three times with 2 mL hexane. The combined extract was evaporated, redissolved in hexane, and analyzed by HPLC. For retinoids obtained from the eye tissues as well as other compartments examined (described later), the molar limit of detection for all retinoids was 2 pmol or less. The routinely used elution program resolved 11-cis, all-trans, 9-cis and 13-cis retinaldehydes, but did not distinguish 11-cis from all-trans retinyl esters. Retinaldehydes observed in the retina consisted virtually entirely of 11-cis and all-trans retinal. The all-trans isomer, typically present at relatively minor levels (see the Results section), was interpreted as arising from 11-cis retinal (thermal isomerization of extracted 11-cis retinal, or photoisomerization of rhodopsin chromophore by the red light used for dissection and preparative procedures), and retinaldehyde data are quoted as total retinaldehyde levels. 

Serum (typically, approximately 300 μL) contained in the collected 500 μL volume of blood was isolated by centrifugation (Microtainer tube, type 365972; BD Biosciences, Franklin Lakes, NJ). A 200-μL aliquot of the serum was removed and extracted with ethanol/hexane. ²⁴ The entire extract from this 200 μL was analyzed by HPLC. Serum data are quoted as amounts contained in the total volume of serum in the animal. Total blood volume was calculated by assuming the volume of blood in mL to be equal to 6% of the body weight in grams. ²⁷ The total serum volume was then calculated from the total volume of blood. For determinations of retinoid in the liver, the entire liver was removed and stored at −20°C for subsequent analysis. The stored liver was thawed and homogenized (Brinkmann Polytron, Model PT 10/35; Fisher Scientific Co., Itasca, IL) in 6 mL of PBS. Aliquots of the homogenate were then extracted with ethanol/hexane. ²⁴ Liver data are quoted as amounts normalized to the entire liver. 

In each experiment, an aliquot of the all-trans (³H)retinol/Tween 40/PBS mixture that had been prepared for injection was extracted with ethanol/hexane and analyzed for radioactivity. The HPLC-isolated peak of (³H) radioactivity corresponding with the retention time of all-trans retinol was typically the major peak. For the experiments reported herein, the counts per minute per microliter of the all-trans (³H)retinol/Tween 40/PBS associated with this all-trans (³H)retinol amounted to 4963 ± 1550 cpm/μL (mean ± SD; n = 24) and ranged from 2028 to 7940 cpm/μL. The average counts per minute per microliter of the mixture was small by comparison with 1.9 × 10⁴ cpm/μL, the estimated value predicted from the quoted volumetric radioactivity of the commercially obtained all-trans (³H)retinol (∼1 μCi/μL), the efficiency of counting, and the known volumetric dilution of this all-trans (³H)retinol in the mixture with Tween 40 and PBS, as described earlier. This difference, and the large range of values among differing preparations of the mixture, may have resulted from differing losses of the all-trans (³H)retinol due to adhesion to the walls of the mixing tube, to a variable efficiency of extraction from micelles formed by the Tween 40, and/or to degradation of the retinol. To compensate for this variability in all-trans (³H)retinol contained in the solutions prepared for injection, radiolabeling data for (³H)retinoids in the serum, RPE, and retina of a given animal were normalized to the sum of (³H) levels determined for (³H)retinyl ester and all-trans (³H)retinol in the liver of the same animal (see the Results section). Based on the specific activity (SA) of all-trans (³H)retinol in the commercially obtained stock solution and incorporating the loss of all-trans (³H)retinol evident from the calculations just noted, the estimated molar amount of all-trans (³H)retinol injected was approximately 0.01 nmol for 1-month animals (133-μL injection volume) and approximately 0.01 to 0.03 nmol for 3-month and 1-year animals (200-μL injection volume). 

Aliquots of the all-trans (³H)retinol/Tween 40/PBS mixture prepared for injection typically contained two unidentified peaks with retention times of approximately 16 (peak 1) and 18 (peak 2) minutes—that is, less than that of all-trans retinol (≈22 minutes). The ratio of ³H radioactivity contained in these two peaks [(peak 1)/(peak 2)] was relatively constant among preparations (0.56 ± 0.05; n = 24), and their combined amount of ³H represented 0.82 ± 0.82 that of the all-trans (³H)retinol. Peaks 1 and 2 were similarly observed in extracts obtained from the serum. Typically, these peaks were also present in relatively small amounts in extracts of the liver and were absent from extracts obtained from the eye tissues. 

Results

Liver

Retinoids detected in the livers of both +/+ and −/− mice consisted virtually entirely of retinyl ester and all-trans retinol. Column A of Tables 1 and 2 show, respectively, molar amounts of retinyl ester determined for the investigated groups of +/+ and −/− mice. Here, as in the case of the other data in Tables 1 and 2 to be described, entries in column A indicate results obtained from animals of a given age (1 month or 3 months) and with a given period of darkness after the intraperitoneal injection of all-trans (³H)retinol (1.5, 4.5, 24, or 48 hours). In this and the immediately following sections, the text will focus on results obtained from these 1- and 3-month mice. Results were also obtained from a smaller number of 1-year mice (lower parts of Tables 1 and 2 ), and these data will be further considered below. 

In mice of a given genotype, levels of liver retinyl ester in 3-month animals (1884 ± 514 nmol [n = 20] for +/+; 1456 ± 749 nmol [n = 22] for −/−) markedly exceeded those in 1-month animals (180 ± 66 nmol [n = 16] for +/+; 155 ± 57 nmol [n = 19] for −/−). The determinations for 1-month +/+ mice did not differ significantly from those for 1-month −/− mice (t-test; P = 0.242), but the determinations for 3-month +/+ versus −/− animals differed significantly (P = 0.039). Large differences were observed in some cases among data obtained from mice of a given genotype and age but differing postinjection period. For example, among 3-month −/− mice, retinyl ester levels determined at 48 hours after injection (2565 ± 427 nmol) differed substantially from those determined at 1.5 hours (953 ± 239 nmol). In addition, among 1-year −/− mice, liver retinyl ester determined at 1.5 hours (17,489 ± 1,644 nmol) was significantly greater (P < 10⁻²) than that determined at 4.5 and 24 hours (7,690 ± 1,097 nmol and 6,058 ± 568 nmol, respectively). The possibility that these differences reflect a dependence of the determined liver retinyl ester level on postinjection period cannot be ruled out. For example, the injection procedure may somehow have induced a time-dependent change in the capacity of the liver to store retinoid, or a change in this tissue that affects the extractability of retinoid. 

Molar amounts of all-trans retinol in the liver (Tables 1 2 ; column B) ranged from a few percent to approximately 35% of those for liver retinyl ester. Molar levels of all-trans retinol in mice of a given genotype were much greater in 3-month animals than in 1-month animals. For 1- and 3-month mice, all-trans retinol levels determined for +/+ versus −/− animals of a given age differed significantly from one another (for 1-month mice, 18.2 ± 10.8 nmol for +/+, 11.2 ± 5.3 nmol for −/−, and P = 0.018; for 3-month mice, 306 ± 148 nmol for +/+, 194 ± 81 nmol for −/−, and P < 10⁻²). 

Figure 1 shows levels of ³H radioactivity for (³H)retinyl ester (Figs. 1A 1B) and all-trans (³H)retinol (Figs. 1C 1D) contained in the livers of +/+ and −/− mice, as a function of the postinjection period. The data illustrated in this and the following figures were obtained from the same groups of animals as those described in Tables 1 and 2 . Figures 1A and 1B describe the time course of radiolabeling of (³H)retinyl ester in 1- and 3-month mice, respectively. Among 1-month +/+ animals, the average level of (³H)retinyl ester grew to a peak of 6.0 × 10⁵ cpm at 4.5 hours and then decreased somewhat at later times (Fig. 1A , open circles). Among −/− mice of the same age, the average level of (³H)retinyl ester at 4.5 hours (3.5 × 10⁵ cpm) was lower than that for +/+ mice. However, at 48 hours, (³H)retinyl ester in the −/− animals (5.7 × 10⁵ ± 9.1 × 10⁴ cpm) significantly exceeded that in +/+ animals (3.9 × 10⁵ ± 2.6 × 10⁴ cpm; P = 0.010). A somewhat similar pattern was evident in the (³H)retinyl ester data obtained from 3-month mice. Again, for example, (³H)retinyl ester in −/− mice at 48 hours was significantly greater than that in +/+ mice (P = 0.017). The occurrence of generally higher levels of (³H)retinyl ester in the 3-month mice was consistent with the larger aliquot of all-trans (³H)retinol injected into the 3-month animals (see the Methods section). Radiolabeling of liver all-trans retinol was evident in both 1- and 3-month animals of both genotypes (Figs. 1C 1D) . Within a given group of mice (i.e., same genotype, age, and postinjection period), levels of radioactivity of this all-trans (³H)retinol were substantially greater than (³H)retinoid levels in the serum, RPE, and retina (described later) but represented only a small fraction of the radioactivity due to liver (³H)retinyl ester (averages of 13% and 11%, respectively, for 1-month +/+ and −/− animals; averages of 21% and 23%, respectively, for 3-month +/+ and −/− animals). For both retinyl ester and all-trans retinol in the liver, as well as for serum all-trans retinol, serum retinyl ester, RPE retinyl ester, and retinaldehydes in the retina (see below), the fractional molar amount of retinoid due to (³H)retinoid was in all cases less than 0.005. The systemic administration of all-trans (³H)retinol thus produced only a tiny perturbation of the molar levels of retinoid in the investigated tissues and compartments. 

Consistent with the fact that the liver is a major site of retinoid storage, ^{16 28} the radioactivity of liver (³H)retinyl ester and all-trans (³H)retinol together represented more than 91% of total (³H)retinoid radioactivity determined in the investigated tissues of a given animal. Accordingly, to compensate for variability in the delivery of all-trans (³H)retinol and its incorporation into the systemic circulation, radiolabeling data from a given animal in the following sections are presented on normalization to the combined levels of liver (³H)retinyl ester and liver all-trans (³H)retinol for that animal. Levels of radioactivity expressed in this normalized form will be referred to below as counts per minute-normalized (cpm-normalized). 

Serum

Column C of Tables 1 and 2 show molar amounts of all-trans retinol contained in the serum of +/+ and −/− mice. Group-averaged levels of serum all-trans retinol in 1-month −/− mice (0.129 ± 0.044 nmol) differed significantly from those of 1-month +/+ mice (0.188 ± 0.085 nmol; P = 0.013). In addition, the amount of serum all-trans retinol in 3-month −/− mice differed significantly from that in both 1-month −/− animals (P < 10⁻⁴) and in 3-month +/+ animals (P = 0.023). The serum of both +/+ and −/− mice of both investigated ages also typically contained significant molar amounts of retinyl ester (Tables 1 2 ; column D). 

Figures 2A and 2B show the cpm-normalized radioactivity of all-trans (³H)retinol contained in the serum. In both +/+ and −/− mice of age 1 month (Fig. 2A) , there was a rapid initial increase in the average value of cpm-normalized all-trans (³H)retinol to a peak at 1.5 or 4.5 hours, and a subsequent decline at later postinjection times. A similar trend was observed for 3-month +/+ and −/− mice (Fig. 2B) . In both 1- and 3-month mice, cpm-normalized levels of serum (³H)retinyl ester similarly exhibited peak average values at early postinjection times (Figs. 2C 2D) . Typically, however, the radioactivity of serum (³H)retinyl ester was smaller than that of serum all-trans (³H)retinol. 

In addition to all-trans (³H)retinol and (³H)retinyl ester, serum extracts from both +/+ and −/− animals showed a prominent absorbance peak with retention times longer than that of all-trans retinal but shorter than that of all-trans retinol. This unidentified absorbance peak did not contain substantial (³H) radioactivity. 

Retinal Pigment Epithelium

Consistent with previous observations, ^{3 9 11} molar levels of RPE retinyl ester in −/− mice increased with age and substantially exceeded the amounts determined in +/+ mice of the same age (Tables 1 2 ; column E). In the present experiments this molar difference was approximately sixfold among 1-month animals (0.589 ± 0.214 nmol for −/−; 0.089 ± 0.035 nmol for +/+) and approximately 15-fold among 3-month animals (2.217 ± 0.991 nmol for −/−; 0.147 ± 0.112 nmol for +/+). The RPE of both +/+ and −/− mice contained only tiny molar amounts of all-trans retinol (0.003 ± 0.001 nmol for 1-month +/+; 0.005 ± 0.002 nmol for 1-month −/−; 0.003 ± 0.002 nmol for 3-month +/+; 0.013 ± 0.008 nmol for 3-month −/−). 

For both +/+ and −/− mice of both investigated ages, virtually all of the ³H radioactivity detected in extracts of the RPE corresponded with retinyl ester. Figures 3A and 3B show cpm-normalized levels of RPE (³H)retinyl ester determined for the 1- and 3-month animals, respectively. For 1-month +/+ mice (Fig. 3A) , normalized (³H)retinyl ester exhibited a peak average of 0.0030 at 4.5 hours and then declined to 0.0017 and 0.0011 at 24 and 48 hours, respectively. Very different results were obtained with 1-month −/− mice. Here, the average level of normalized (³H)retinyl ester at 4.5 hours (0.0045) exceeded that determined in +/+ animals. Furthermore, by contrast with the +/+ results, the average level in −/− mice increased with time to 0.0089 and 0.0105 at 24 and 48 hours, respectively. Averages determined for 3-month +/+ animals (Fig. 3B) were lower than corresponding values determined for 1-month +/+ animals, but exhibited a similar variation with postinjection period. Those determined for 3-month −/− mice were smaller than corresponding averages among 1-month −/− mice at the same postinjection time. However, for both investigated ages, average levels in −/− animals at 24 and 48 hours substantially exceeded those in +/+ animals. In both +/+ and −/− mice of both investigated ages, the radioactivity of RPE all-trans (³H)retinol was in most cases below detection and in all cases amounted to a cpm-normalized level of less than 0.002. We did not detect substantial amounts (counts per minute, or nanomoles) of 11-cis retinol in the RPE. Occasionally, small levels of retinaldehydes (counts per minute and nanomoles) were detected in the RPE of +/+ mice, possibly due to contamination by retinaldehydes in the retina (however, see the Discussion section). Retinaldehydes were not detected in RPE extracts obtained from −/− mice. 

Retina

Column F in Tables 1 and 2 shows molar data for total retinaldehydes in the retina. Group-average retinaldehyde levels in 3-month +/+ animals (0.402 ± 0.076 nmol) did not differ significantly from those of 1-month +/+ animals (0.358 ± 0.065 nmol; P = 0.087). Control experiments performed on C57BL/6J mice of ages 2 to 3 months indicated the presence (as a combined total in both eyes) of 0.546 ± 0.117 nmol of retinaldehyde in the retina (n = 8). Thus, the relatively low molar levels noted earlier for retinaldehyde in the +/+ animals is a property of this strain and does not reflect a low efficiency of extraction (also cf. Ref. ²⁹ ; Crouch RK, Moiseyev G, Goletz P, Bealle G, Redmond TM, Ma JX, ARVO Abstract 3525, 2001). In +/+ mice, the molar amount of all-trans retinal relative to that of 11-cis retinal was 0.101 ± 0.091 in 1-month animals and 0.163 ± 0.042 in 3-month animals. Retinaldehydes in retina extracts from −/− mice were barely detectable both at 1 (0.002 ± 0.003 nmol) and 3 (0.010 ± 0.009 nmol) months, consistent with previous findings. ^{3 11}  

Figures 4A and 4B show cpm-normalized levels of (³H)retinaldehydes determined for the retinas of +/+ and −/− mice. In 1-month +/+ mice, there was an increase in (³H)retinaldehyde with increasing postinjection period. A similar trend, but with lower levels, was observed for 3-month +/+ animals. The only isomers detected at significant level in these analyses of +/+ retinas were 11-cis and all-trans retinal, and each of the determinations in Figures 4A and 4B represent the sum of contributions from these two isomers. As shown by filled circles in Figures 4A and 4B , determinations for 1- and 3-month −/− animals indicated no substantial amount of (³H)retinaldehydes in the retina at any postinjection period. 

SAs (in counts per minute per nanomole) of 11-cis and all-trans retinal recovered from the retina (SA _cis and SA _trans , respectively) were compared in each of the 1- and 3-month +/+ mice investigated. This analysis yielded, for the ratio SA _cis /SA _trans , 0.620 ± 0.206 in the 1-month animals and 0.763 ± 0.278 in the 3-month animals. The occurrence of average levels below unity for these determinations is unexplained. On the assumption that radiolabeled retinaldehyde in the retina represented 11-cis (³H)retinal transferred from the RPE and bound by newly synthesized opsin ³⁰ (see the present Discussion section), this result could reflect a greater susceptibility of recently formed rhodopsin to photo- or thermal isomerization of the 11-cis retinal chromophore. 

Typically, no retinoids other than retinaldehydes were detected in the retinas of +/+ mice. Occasionally, small amounts of all-trans retinol and/or retinyl ester were detected in the +/+ retina extract. Retina extracts obtained from −/− animals showed variable and sometimes substantial molar amounts of retinyl ester, perhaps due to contamination by the high level of retinyl ester in the RPE (discussed earlier). The tabular data reporting molar amounts of RPE retinyl ester do not include contributions from this retinyl ester. In both 1- and 3-month −/− mice analyzed at 48 hours after injection, radioactivity due to (³H)retinyl ester in the retina represented, on average, 6% of the radioactivity of RPE (³H)retinyl ester; the average molar amount of retinyl ester in the retina represented 8% of that in the RPE. 

Specific Activities of (³H)retinoids

Tables 3 and 4 summarize results obtained for the determinations of SAs of the investigated retinoids. These data exhibit three noteworthy patterns. First, in +/+ and −/− animals of both investigated ages, the average SA of serum all-trans retinol was highest at 1.5 or 4.5 hours and then declined with increasing postinjection period to levels representing approximately 7% or less of the peak (Tables 3 4 ; rows 3 and 10). Furthermore, in all cases the average SA of serum all-trans retinol at 48 hours differed by 40% or less from those of liver retinyl ester and all-trans retinol (rows 1–3, 8–10). A second point concerns the kinetics of the SA of RPE retinyl ester. In +/+ animals of both investigated ages, the average SA of RPE retinyl ester exhibited a peak at 4.5 hours and by 48 hours declined to approximately 6% to 10% of the peak (Tables 3 4 , row 5). By contrast, in both 1- and 3-month −/− mice, the peak average SA occurred at 24 or 48 hours (row 12). Third, in +/+ mice the SA of retinaldehydes in the retina was zero at 1.5 hours and then increased with postinjection period (row 6). At 24 hours, the average SA of these retinaldehydes in 1- and 3-month +/+ animals represented, respectively, 34% and 59% of the average SA of RPE retinyl ester. At 48 hours, the corresponding percentages were 60% and more than 146% (row 7). 

As noted above (Tables 1 2) , the present study included analysis of 1-year +/+ and −/− mice. For a given compartment, and among animals of a given genotype, the pattern of SA results obtained from these 1-year mice at differing postinjection times was in most cases qualitatively similar to those shown in Tables 3 and 4 for the younger animals. For example, in 1-year +/+ mice, as in 1- and 3-month +/+ mice, the SA of serum all-trans retinol and of RPE retinyl ester exhibited a peak at 1.5 or 4.5 hours. In addition, as in the case of 3-month +/+ mice, data obtained from 1-year +/+ mice at 24 hours indicated a significant labeling of retinaldehyde in the retina; the SA of this retinaldehyde (1173 ± 197 cpm/nmol; n = 2) greatly exceeded those of retinyl ester and all-trans retinol in the liver (96 ± 35 cpm/nmol [n = 2] and 144 ± 22 cpm/nmol [n = 2], respectively). However, both +/+ and −/− mice of age 1 year differed considerably from those of the younger animals, in that the molar amounts of retinoid in the liver far exceeded those determined for the younger animals (cf. Tables 1 2 ). In addition, the molar amount of RPE retinyl ester in 1-year −/− mice (5.170 ± 1.795 nmol) significantly exceeded that of 3-month as well as 1-month −/− mice (P < 10⁻⁵). 

Inferred Flux of Retinol into RPE

Based on previous evidence indicating that all-trans retinol bound to RBP in the bloodstream is the primary source of retinoid uptake by the RPE ^{15 16} and on the present finding that the RPE contains little if any unesterified retinol, it is reasonable to believe that radiolabeled serum all-trans retinol is the principal source of radiolabeled retinyl ester appearing in the RPE. We furthermore assume in the following analysis that, during a period of approximately 4.5 hours immediately after the systemic administration of all-trans (³H)retinol, the absolute SA of serum all-trans (³H)retinol was roughly constant (cf. Fig. 2 ), and hydrolysis of the RPE (³H)retinyl ester formed was negligible. With these assumptions, the molar influx of all-trans retinol from the serum into the RPE is approximately given by the relation (equation 1)  

\[(\mathrm{nmol\ of\ retinol\ influx})\ {\approx}\ (\mathrm{cpm\ of\ RPE\ retinyl\ ester})/\mathrm{SA}_{\mathrm{serumOL,T}}\]

where SA_serumOL,T is the SA of serum all-trans retinol (in counts per minute per nanomole) determined at the conclusion time T of this period. Column A of Table 5 shows determinations of this inferred retinol influx over a 4.5-hour postinjection period in +/+ and −/− animals of both investigated ages. As indicated by the summary data obtained for each group (Table 5 , column A, group average rows 1–4), the determined average influx over 4.5 hours ranged from 0.014 to 0.026 nmol, with no significant differences among groups (P > 0.169). Column B of Table 5 reproduces data summarized in Tables 1 and 2 for the molar amount of retinyl ester in the mice under consideration. The inferred fractional influx of all-trans retinol during the 4.5-hour period (column C) was obtained by dividing the column A determinations of influx by the column B levels of retinyl ester. As shown by the group-averaged data (Table 5 , column C, rows 1–4), the fractional influx in +/+ mice greatly exceeded that in −/− mice of the same age. Moreover, in animals of given genotype, the fractional influx for 1-month mice considerably exceeded that for 3-month mice. 

Table 6 shows the results for similar determinations of fractional retinol influx based on measurements made at 1.5 hours after injection. In column D of Table 6 the group-average influx data at 1.5 hours (column C) are normalized to the 4.5-hour data in Table 5 , column C. If both the SA of serum all-trans retinol and the instantaneous rate of retinol influx were strictly constant throughout the longer period (4.5 hours) considered in Table 5 , each entry in column D of Table 6 would equal 0.33 (=1.5/4.5). The determined average ratios, which range from 0.13 to 0.49, in all cases differed from 0.33 by less than a factor of 3. 

Discussion

Passage of Retinoid into and out of the RPE

This study has examined the fate of systemically introduced all-trans (³H)retinol in Rpe65 ^−/− mice and Rpe65 ^+/+ control mice as a function of postinjection time. A main finding, supported by the cpm-normalized data of Figure 3 and the SA data of Tables 3 and 4 , is that Rpe65 ^−/− animals of both 1 and 3 months of age exhibit a progressive radiolabeling of RPE retinyl ester within the investigated postinjection periods. On the assumptions noted in text accompanying equation 1 above, the molar amount of serum retinol converted to RPE retinyl ester within a 4.5-hour period is substantial, representing 0.021 ± 0.009 and 0.026 ± 0.018 nmol, respectively, in 1- and 3-month Rpe65 ^−/− mice. Furthermore, these values of inferred molar influx do not differ significantly from those determined at 4.5 hours in Rpe65 ^+/+ mice (0.014 ± 0.004 and 0.016 ± 0.013 nmol, respectively, in 1- and 3-month animals). Thus, despite the abnormally large molar quantity of RPE retinyl ester in the Rpe65 ^−/− (Refs. ^{3 31} ; and present Table 2 ), the absence of the Rpe65 gene product appears not to affect substantially the inward movement of retinoid into the RPE. This point is of particular interest, as RPE65 (also known as p63) was initially hypothesized as a component of the RPE receptor for serum retinoid-binding protein at the basolateral membrane. ^{32 33 34}  

The finding in Rpe65 ^+/+ mice that the inferred molar influx of all-trans retinol from the circulation over a period of approximately 4.5 hours represented a significant fraction of the prevailing molar level of RPE retinyl ester (Table 5) suggests that this influx is accompanied by an outflow of retinoid, presumably as all-trans retinol, from the RPE into the circulation. Such a possibility is consistent with the results shown in Figure 3 , which indicate a transience of RPE retinyl ester radiolabeling in the Rpe65 ^+/+. A rough indication of the activity of this putative retinol efflux can be obtained from the data in Tables 5 and 6 . The determined average fractional influx of all-trans retinol for 1-month Rpe65 ^+/+ mice (0.34 per 4.5 hours in Table 5 ; 0.10 per 1.5 hours in Table 6 ) is approximately 0.07 per hour. Assuming that retinoid exchange between the RPE and retina under the present dark-adapted conditions (passage of 11-cis retinal to the retina, and uptake of rhodopsin-containing shed discs by the RPE) has a negligible effect on RPE retinoid level, a fractional influx of 0.07 per hour predicts an approximately 50-fold increase per month in the molar amount of RPE retinoid in the absence of retinoid efflux to the circulation. As the observed molar increase in RPE retinyl ester of Rpe65 ^+/+ mice between approximately 1 and 3 months of age was, on average, only approximately two- to fourfold (Tables 1 5 6) , the efflux is robust. 

A calculation similar to that just described can be applied to the data from Rpe65 ^−/− mice. Taking the average fractional influx for 1-month animals as approximately 0.015 per hour (0.055 per 4.5 hours in Table 5 ; 0.027 per 1.5 hours in Table 6 ), the molar increase predicted in the absence of efflux is approximately 11-fold per month. For Rpe65 ^−/− mice the observed molar increase in RPE retinyl ester between approximately 1 and 3 months of age was on average approximately three- to fourfold (Tables 2 5 6) . The magnitude of the difference between the projected increase and the observed increase in molar amount of RPE retinyl ester is thus much less for the Rpe65 ^−/− than for the Rpe65 ^+/+ mice. On the basis of this difference and from the similarity among Rpe65 ^+/+ and Rpe65 ^−/− mice in absolute molar values of inferred retinol influx (see above), we conclude that Rpe65 deficiency leads selectively to an impairment of the efflux of retinol from the RPE to the circulation. 

The evidence that Rpe65 deficiency impairs both 11-cis retinoid formation ^{3 9 11} and retinoid efflux (present data) raises the question of whether these two effects are causally linked. Two recent studies bear closely on this question. First, Sieving et al. ³⁵ have shown in normal rats that 13-cis retinoic acid (isotretinoin), an agent known to inhibit the oxidation of 11-cis retinol to 11-cis retinal by 11-cis retinol dehydrogenase, ^{36 37} promotes the build-up of both all-trans and 11-cis retinyl ester in the RPE. Second, Driessen et al. ³⁸ have shown that mice deficient in 11-cis retinol dehydrogenase and therefore unable to generate 11-cis retinal in the RPE also accumulate abnormally high levels of both all-trans and 11-cis retinyl ester. 

Considered in light of these two studies and others, the present results suggest a simple hypothesis: namely, that 11-cis retinal and/or 11-cis retinol in the RPE specifically promote(s) all-trans retinol efflux at the RPE basolateral membrane. This hypothesis is illustrated by the scheme shown in Figure 5 . Reactions 1–3 in the figure denote the RPE processing of all-trans retinol to 11-cis retinal in the normally functioning visual cycle. (The light-dependent conversion of all-trans retinal to 11-cis retinal mediated by retinal G-protein-coupled receptor [RGR] ³⁹ is omitted for brevity.) Also shown are the movements of retinoid across the RPE apical membrane (inward passage of all-trans retinol and outward passage of 11-cis retinal) involved in the visual cycle, and the interconversion of 11-cis retinol and 11-cis retinyl ester within the RPE (reactions 4). Reactions 5 and 6 denote the influx and efflux of all-trans retinol at the RPE basolateral membrane, and reactions 7 denote the interconversion of this retinol and all-trans retinyl ester. According to the present hypothesis, 11-cis retinal and/or 11-cis retinol in the RPE of the normal, fully dark-adapted eye are present at (low) levels that suffice, through a mechanism as yet unknown, to stimulate the efflux of all-trans retinol at the basolateral membrane (Fig. 5 , dashed arrow A, and ⊕ at reaction 6). The hypothesis further supposes that Rpe65 deficiency impairs the reaction(s) immediately responsible for the formation of 11-cis retinal and/or 11-cis retinol (Fig. 5 reactions labeled by asterisks). This impairment, presumably in concert with the delivery of available 11-cis retinal to the rods and its binding to opsin, causes the RPE level of one or both of these 11-cis retinoids to fall, thereby reducing or eliminating the stimulatory action on all-trans retinol efflux. The abnormal build-up of RPE all-trans retinyl ester in Rpe65 ^−/− mice is thus explained as the consequence of reduced retinol efflux with essentially maintained retinol influx at the RPE basolateral membrane. 

The present hypothesis is consistent with the evidence that the reaction(s) affected by Rpe65 deficiency is(are) subsequent to all-trans retinyl ester formation in the visual cycle ^{3 40} and with recent evidence that the administration of large doses of 9-cis retinal (a functional analogue of 11-cis retinal) to young Rpe65 ^−/− mice attenuates the accumulation of abnormally high levels of RPE retinyl ester. ¹¹ As noted, both reaction 2 (mediated by isomerohydrolase ¹² ) and reaction 3 (the dehydrogenase-mediated step) in Figure 5 are immediate candidates for the site(s) of RPE65 action. The absence of measurable 11-cis retinyl ester accumulation in Rpe65 ^−/− mice ^{3 9} suggests impairment of the reaction catalyzed by isomerohydrolase. However, the possibility that the dehydrogenase-catalyzed step is a second site of impairment cannot be ruled out. ⁴⁰  

An immediate prediction of the present hypothesis is that genetic or pharmacological alterations that decrease the RPE’s capacity to generate 11-cis retinoid should reduce the rate of all-trans retinol efflux at the RPE basolateral membrane. Potential approaches for testing this hypothesis and for comparing the putative efflux-stimulating activities of 11-cis retinal and 11-cis retinol include treating wild-type animals with isotretinoin, ³⁵ disrupting the gene for cellular retinaldehyde-binding protein (CRALBP) ⁴¹ and treating Rpe65 ^−/− mice with 9-cis retinal. ¹¹ A further interesting question raised by these considerations is whether the RPE’s provision of 11-cis retinal to the rods in the normally functioning visual cycle after rhodopsin bleaching, by transiently reducing the RPE levels of 11-cis retinoids, may transiently downregulate retinol efflux at the RPE basolateral membrane. 

Movement of (³H)retinoid into Retina of Rpe ^+/+ Mice

Figure 4 shows an increase, with postinjection time, in the normalized counts per minute of (³H)retinaldehydes in the retinas of 1- and 3-month Rpe65 ^+/+ mice. Because the present experiments involved maintenance of a dark-adapted condition throughout the postinjection period, it is likely that this (³H)retinaldehyde represents the supply, to the photoreceptors, of 11-cis retinal chromophore for recently synthesized opsin. ³⁰ (By contrast, there was little if any passage of (³H)retinaldehydes into the retinas of Rpe65 ^−/− mice, consistent with the tiny level of native retinaldehyde in the Rpe65 ^−/− observed in this and previous studies.) In 3-month Rpe65 ^+/+ mice at postinjection times of 24 and 48 hours, the average SA of retinaldehydes in the retina (Table 4 ; average of 1803 and 1459 cpm/nmol) represented approximately 40% of the peak average SA of RPE retinyl ester (3817 cpm/nmol, determined at 4.5 hours) and was approximately equal to the average SA of RPE retinyl ester at 24 and 48 hours (average of 3051 and 216 cpm/nmol). This result is striking when considered in light of the evidence that the fully dark-adapted mammalian retina contains relatively little free (i.e., nonchromophoric) 11-cis retinal, ⁴² and that the molar amount of 11-cis retinal transferred per day from the RPE to opsin represents approximately 13% of the opsin content of the rods. This approximately 13% is based on an estimated transfer of approximately 10% per day representing the supply of chromophore to newly synthesized opsin, ^{43 44} and an additional approximately 3% per day due to the exchange of chromophore in darkness. ⁴⁵ (The 3% per day estimate derives from a chromophore exchange rate of approximately 2 × 10³ per rod per minute ⁴⁵ and an overall titer of approximately 10⁸ rhodopsin molecules per rod. ⁴⁶ ) Thus, (³H)retinaldehyde delivered to the retina during the postinjection periods of 24 and 48 hours represented a molar amount of 11-cis retinal equal only to approximately 13% to 26% of the molar retinaldehyde level underlying the Table 4 SAs for retinaldehydes in the retina. This argument leads to the conclusion that the SA of 11-cis retinal formed and delivered to the retina during the 24- and 48-hour postinjection periods was at least approximately 60% greater than the determined peak average SA of RPE retinyl ester and at least twice that of RPE ester prevailing at 24 and 48 hours. A similar comparison of data obtained from 1-month Rpe65 ^+/+ mice (Table 3 ; average SA of retinaldehydes in the retina at 24 and 48 hours = 2848 cpm/nmol; peak average SA of RPE retinyl ester = 47618 cpm/nmol) implies, for the SA of 11-cis retinal formed and delivered to the retina over 24 to 48 hours, an amount equal only to approximately 23% to 46% of the peak SA of RPE retinyl ester. 

Previous experiments involving the supply of radiolabeled all-trans retinol to the RPE of amphibian RPE-eyecup preparations have shown that the SA of 11-cis (³H)retinal formed in the RPE exceeds that of its presumed precursor, all-trans (³H)retinyl ester. This finding was taken to suggest the occurrence in the RPE of a last-in/first-out processing of all-trans retinol in the visual cycle, in which recently entering and thus recently esterified all-trans retinol is favored over previously formed (“old”) retinyl ester for conversion to 11-cis retinal. ²⁵ Reactions 8 in Figure 5 denote the presumed slow interconversion of these two pools of all-trans retinyl ester. The inferred high SA of 11-cis retinal delivered to the retina in the present experiments on 3-month Rpe65 ^+/+ mice is consistent with the occurrence of a similar last-in/first-out processing of all-trans (³H)retinol entering the RPE from the systemic circulation. That is, as illustrated in Figure 5 , the esterification of all-trans retinol that has recently entered the RPE at the basolateral membrane contributes to the pool of “recent” trans ester and is similarly favored over previously formed ester for processing to 11-cis retinal. Also consistent with the inferred high SA of 11-cis retinal formed in the RPE are the findings of Stecher et al., ⁴⁷ which suggest that nonesterified all-trans retinoid may be the immediate substrate of the isomerohydrolase reaction. 

 Supported by Grants EY05494 and EY01792 from the National Eye Institute and an unrestricted award from Research to Prevent Blindness.

 Submitted for publication July 8, 2002; revised September 24, 2002; accepted November 10, 2002.

 Disclosure: N.M. Qtaishat, None; T.M. Redmond, None; D.R. Pepperberg, None

 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.

 

Table 1.

 

Molar Amounts of Retinoid Determined in Rpe65 ^+/+ Mice

Age	Postinjection Period*	A	B	C	D	E	F
				Liver Retinyl Ester	Liver All-trans Retinol	Serum All-trans Retinol	Serum Retinyl Ester	RPE Retinyl Ester	Retina Retinaldehydes, †
1 month	1.5 h (4)	191 ± 60	16.4 ± 5.5	0.107 ± 0.051	0.047 ± 0.037	0.092 ± 0.030	0.365 ± 0.047
	4.5 h (4)	140 ± 7.9	8.6 ± 3.5	0.126 ± 0.019	0.052 ± 0.012	0.042 ± 0.008	0.353 ± 0.096
	24 h (4)	240 ± 95	30.4 ± 13.8	0.290 ± 0.032	0.068 ± 0.021	0.118 ± 0.015	0.349 ± 0.035 (3)
	48 h (4)	148 ± 23	17.2 ± 4.7	0.229 ± 0.047	0.077 ± 0.052	0.105 ± 0.018	0.363 ± 0.087
	Group (16)	180 ± 66	18.2 ± 10.8	0.188 ± 0.085	0.061 ± 0.033	0.089 ± 0.035	0.358 ± 0.065 (15)
3 months	1.5 h (6)	1668 ± 608	218 ± 96	0.146 ± 0.054	0.114 ± 0.075	0.203 ± 0.172	0.434 ± 0.048 (5)
	4.5 h (6)	1753 ± 646	219 ± 71	0.137 ± 0.089	0.292 ± 0.322	0.149 ± 0.098	0.388 ± 0.106 (5)
	24 h (4)	2130 ± 173	388 ± 172	0.140 ± 0.045	0.712 ± 0.654	0.086 ± 0.025	0.361 ± 0.091
	48 h (4)	2160 ± 133	485 ± 46	0.374 ± 0.066	0.693 ± 0.896	0.118 ± 0.038	0.421 ± 0.043
	Group (20)	1884 ± 514	306 ± 148	0.188 ± 0.114	0.403 ± 0.540	0.147 ± 0.112	0.402 ± 0.076 (18)
1 year	1.5 h (3)	9971 ± 2046	734 ± 262	0.432 ± 0.243	0.123 ± 0.033	0.385 ± 0.129	0.318 (1)
	4.5 h (2)	9938 ± 3162	738 ± 64	0.257 ± 0.056	5.113 ± 6.527	0.205 ± 0.137	0.702 ± 0.084
	24 h (2)	7522 ± 1187	680 ± 52	0.238 ± 0.069	0.076 ± 0.002	0.585 ± 0.120	0.215 ± 0.085
	Group (7)	9262 ± 2170	720 ± 157	0.326 ± 0.176	1.535 ± 3.616	0.391 ± 0.188	0.430 ± 0.259 (5)

Table 2.

 

Molar Amounts of Retinoid Determined in Rpe65 ^−/− Mice

Age	Postinjection Period	A	B	C	D	E	F
				Liver Retinyl Ester	Liver All-trans Retinol	Serum All-trans Retinol	Serum Retinyl Ester	RPE Retinyl Ester	Retina Retinaldehydes
1 month	1.5 h (6)	144 ± 51	10.4 ± 4.8	0.117 ± 0.039	0.085 ± 0.049	0.562 ± 0.134	0.003 ± 0.004
	4.5 h (5)	129 ± 20	8.2 ± 3.0	0.120 ± 0.035	0.055 ± 0.028	0.415 ± 0.143	0.000 ± 0.000
	24 h (4)	186 ± 51	11.6 ± 2.5	0.161 ± 0.050	0.064 ± 0.020	0.829 ± 0.273	0.002 ± 0.003
	48 h (4)	174 ± 93	15.7 ± 7.7	0.127 ± 0.059	0.109 ± 0.072	0.607 ± 0.111	0.002 ± 0.002
	Group (19)	155 ± 57	11.2 ± 5.3	0.129 ± 0.044	0.078 ± 0.047	0.589 ± 0.214	0.002 ± 0.003
3 months	1.5 h (6)	953 ± 239	157 ± 105	0.320 ± 0.085	0.147 ± 0.115	1.959 ± 0.739	0.007 ± 0.007
	4.5 h (8)	1087 ± 342	197 ± 91	0.257 ± 0.110	0.202 ± 0.282	1.772 ± 0.942	0.003 ± 0.005
	24 h (4)	1841 ± 817	228 ± 38	0.310 ± 0.227	1.541 ± 2.803	3.081 ± 1.257	0.019 ± 0.007
	48 h (4)	2565 ± 427	210 ± 42	0.207 ± 0.037	0.478 ± 0.470	2.632 ± 0.652	0.021 ± 0.003
	Group (22)	1456 ± 749	194 ± 81	0.275 ± 0.123	0.481 ± 1.208	2.217 ± 0.991	0.010 ± 0.009
1 year	1.5 h (3)	17489 ± 1644	341 ± 166	0.412 ± 0.177	5.614 ± 5.054	6.192 ± 1.875	0.023 ± 0.005
	4.5 h (2)	7690 ± 1097	458 ± 4.0	0.166 ± 0.081	15.749 ± 21.568	4.050 ± 2.475	0.024 ± 0.012
	24 h (2)	6058 ± 568	362 ± 129	0.243 ± 0.053	0.120 ± 0.034	4.755 ± 0.431	0.010 ± 0.006
	Group (7)	11422 ± 5813	380 ± 122	0.293 ± 0.159	6.940 ± 11.327	5.170 ± 1.795	0.019 ± 0.009

Figure 1.

 

Liver retinoids: (³H)retinyl ester in (A) 1- and (B) 3-month mice, and all-trans (³H)retinol in (C) 1- and (D) 3-month mice. Here and in Figs. 2 3 4 , open and filled symbols in each panel indicate, respectively, results obtained from +/+ and −/− mice. The data illustrate the mean ± SD of results from the indicated group of mice. In this and all later figures, the number of animals investigated under a given experimental condition is as indicated in Tables 1 and 2 .

Figure 2.

 

Normalized radioactivity (normalized counts per minute [cpm]) for (³H)retinoids in the serum. ³H radioactivity levels for a given animal normalized to the sum of liver (³H)retinyl ester plus liver all-trans (³H)retinol for that animal. Serum all-trans (³H)retinol in (A) 1- and (B) 3-month mice, and serum (³H)retinyl ester in (C) 1- and (D) 3-month mice. In this and later figures, the data illustrate the mean ± SD for normalized counts-per-minute determinations.

Figure 3.

 

Normalized radioactivity (normalized counts per minute [cpm]) for RPE (³H)retinyl ester. ³H radioactivity levels in (A) 1- and (B) 3-month mice were normalized to the sum of liver (³H)retinyl ester plus liver all-trans (³H)retinol, as in Figure 2 .

Figure 4.

 

Normalized radioactivity (normalized counts per minute [cpm]) for (³H)retinaldehydes in the retina. Data from (A) 1- and (B) 3-month mice were normalized to the sum of liver (³H)retinyl ester plus liver all-trans (³H)retinol, as in Figure 2 .

Table 3.

 

Specific Activity of Investigated Retinoids in 1-Month Mice

Row	Genotype	Compartment and Retinoid	1.5 h	4.5 h	24 h	48 h
1	Rpe65 +/+	Liver ester	1289 ± 700	4260 ± 792	2346 ± 1036	2710 ± 515
2		Liver retinol	2394 ± 1217	4456 ± 643	2797 ± 772	3004 ± 571
3		Serum retinol	93348 ± 25898	143520 ± 60380	6575 ± 2375	3386 ± 676
4		Serum ester	31002 ± 15640	44264 ± 19019	2491 ± 2877	0
5		RPE ester	5752 ± 1150	47618 ± 17876	8323 ± 1669	4590 ± 203
6		Retina aldehydes	0	1606 ± 575	2938 ± 938	2759 ± 422
7		(Ret. al.)/(RPE ester)	0	0.0347 ± 0.0119	0.337 ± 0.044	0.603 ± 0.104
8	Rpe65 −/−	Liver ester	1629 ± 413	2726 ± 1321	1922 ± 645	4264 ± 2546
9		Liver retinol	3227 ± 560	3428 ± 1483	2790 ± 588	4217 ± 2218
10		Serum retinol	64012 ± 27570	75908 ± 24182	6103 ± 3134	5564 ± 3124
11		Serum ester	29629 ± 21147	16669 ± 6562	2607 ± 2151	3782 ± 5017
12		RPE ester	1429 ± 363	4514 ± 3110	4044 ± 572	11072 ± 5648

Table 4.

 

Specific Activity of Investigated Retinoids in 3-Month Mice

Row	Genotype	Compartment and Retinoid	1.5 h	4.5 h	24 h	48 h
1	Rpe65 +/+	Liver ester	274 ± 84	328 ± 80	364 ± 53	303 ± 18
2		Liver retinol	474 ± 238	464 ± 96	390 ± 77	298 ± 36
3		Serum retinol	80347 ± 30743	40139 ± 16895	5025 ± 2630	414 ± 827
4		Serum ester	20487 ± 17812	10095 ± 7694	1314 ± 1744	178 ± 356
5		RPE ester	1183 ± 1337	3817 ± 1004	3051 ± 466	216 ± 432
6		Retina aldehydes	0	195 ± 183	1803 ± 374	1459 ± 228
7		(Ret. al.)/(RPE ester)	0	0.0479 ± 0.0455	0.588 ± 0.042	>1.46*
8	Rpe65 −/−	Liver ester	509 ± 284	579 ± 322	368 ± 121	378 ± 88
9		Liver retinol	852 ± 325	818 ± 331	496 ± 138	444 ± 155
10		Serum retinol	67301 ± 18700	54460 ± 16072	8383 ± 2794	507 ± 1014
11		Serum ester	18696 ± 9917	14558 ± 11768	2614 ± 1451	117 ± 234
12		RPE ester	347 ± 184	796 ± 354	840 ± 222	716 ± 150

Table 5.

 

Influx of All-trans Retinol Inferred from Data Obtained at 4.5 h after-Injection

Genotype and Age	A	B	C
		Inferred OL Influx	RPE Ester	Fractional Influx
		(nmol)	(nmol)
Rpe65 +/+, 1-month	0.019	0.053	0.36
	0.010	0.033	0.30
	0.016	0.043	0.37
	0.012	0.038	0.31
	(1) 0.014 ± 0.004		0.34 ± 0.04
Rpe65 −/−, 1-month	0.018	0.338	0.053
	0.010	0.483	0.020
	0.025	0.381	0.065
	0.033	0.623	0.053
	0.021	0.250	0.083
	(2) 0.021 ± 0.009		0.055 ± 0.023
Rpe65 +/+, 3-months	0.040	0.316	0.12
	0.017	0.222	0.077
	0.008	0.096	0.082
	0.013	0.085	0.15
	0.007	0.094	0.078
	0.008	0.083	0.098
	(3) 0.016 ± 0.013		0.10 ± 0.03
Rpe65 −/−, 3-months	0.061	3.320	0.018
	0.030	2.782	0.011
	0.009	1.137	0.0083
	0.013	2.204	0.0061
	0.010	0.554	0.017
	0.042	1.821	0.023
	0.022	1.105	0.020
	0.024	1.253	0.019
	(4) 0.026 ± 0.018		0.015 ± 0.006

Table 6.

 

Influx of All-trans Retinol Inferred from Data Obtained at 1.5 h after Injection

Genotype and Age	A	B	C	D
		Inferred OL Influx (nmol)	RPE Ester (nmol)	Fractional influx	Average Fractional Influx Ratio, †
Rpe65 +/+, 1-month	0.006	0.087	0.069
	0.008	0.133	0.062
	0.009	0.059	0.15
	0.012	0.090	0.13
	(1) 0.009 ± 0.002		0.10 ± 0.04	0.29
Rpe65 −/−, 1-month	0.021	0.414	0.051
	0.009	0.583	0.015
	0.008	0.559	0.014
	0.008	0.498	0.016
	0.023	0.809	0.029
	0.018	0.511	0.035
	(2) 0.014 ± 0.007		0.027 ± 0.015	0.49
Rpe65 +/+, 3-months*	0.008	0.352	0.022
	0.009	0.484	0.020
	0	0.095	0
	0	0.097	0
	0	0.088	0
	0.003	0.101	0.034
	(3) 0.003 ± 0.004		0.013 ± 0.015	0.13
Rpe65 −/−, 3-months	0.007	2.658	0.003
	0.007	3.054	0.002
	0.008	1.439	0.006
	0.006	1.225	0.005
	0.011	1.478	0.007
	0.012	1.897	0.006
	(4) 0.008 ± 0.002		0.005 ± 0.002	0.33

Figure 5.

 

Hypothesized scheme of visual cycle retinoid processing in the RPE and of regulated retinoid efflux at the RPE basolateral membrane. In the visual cycle, all-trans retinol (trans OL) entering the RPE cell from the interphotoreceptor matrix (IPM) undergoes enzymatic conversion, sequentially, to all-trans retinyl ester (trans ester), 11-cis retinol (cis OL), and 11-cis retinal (cis AL; reactions 1–3). Reactions 4 denote the interconversion of 11-cis retinol and 11-cis retinyl ester (cis ester). The pool of recently formed (“recent”) trans ester is presumed to undergo slow interconversion with previously formed (“old”) trans ester (reactions 8). ²⁵ Reactions 5 and 6 denote, respectively, the influx of all-trans retinol from the circulation and its efflux to the circulation across the RPE basolateral membrane. Reactions 7 denote the interconversion of the entering all-trans retinol and recently formed all-trans retinyl ester. Dashed arrow A and ⊕ indicate the hypothesized stimulatory effect of 11-cis retinal and/or 11-cis retinol on all-trans retinol efflux at the basolateral membrane. Asterisks at reactions 2 and 3 denote the hypothesized site(s) of action of RPE65.

The authors thank Rosalie K. Crouch for helpful discussions during the course of this study.