Synthesis and Release of Docosahexaenoic Acid by the RPE Cells of prcd-Affected Dogs

Abstract

purpose. Dogs affected with progressive rod-cone degeneration (prcd) have reduced levels of docosahexaenoic acid (DHA, 22:6n-3) in their plasma and rod photoreceptor outer segments (ROS). Dietary supplementation of DHA has failed to increase the ROS DHA levels to that of unaffected control dogs. The present study was undertaken to test the hypothesis that prcd-affected dogs have a reduced capacity for the synthesis and/or release of DHA in retinal pigment epithelial (RPE) cells.

methods. RPE cells (first passage cultures) from prcd-affected and normal dogs were incubated with [³H]eicosapentaenoic acid (EPA, 20:5n-3) for 24 and 72 hours. After incubation, the radiolabeled fatty acids in the cells and media were analyzed.

results. DHA and all its metabolic intermediates were detected in RPE cells from prcd-affected and normal dogs. No significant difference was found in the amount of products (including DHA) synthesized between normal and affected RPE cells at either time point. In the culture media, RPE cells from prcd-affected dogs released significantly more DHA than cells from normal dogs after 72-hour incubation, but not after 24-hour incubation.

conclusions. RPE cells from prcd-affected dogs can synthesize and release DHA at least as efficiently as cells from normal dogs. Therefore, synthesis of DHA from its precursor and its release from RPE cells does not appear to contribute to the reduction in ROS DHA levels found in prcd-affected animals.

Although gene mutations for photoreceptor-specific proteins are generally considered the primary causative factor for retinitis pigmentosa, abnormalities in systemic levels of essential long-chain polyunsaturated fatty acids, particularly docosahexaenoic acid (DHA, 22:6n-3), frequently have been reported to be associated with various forms of retinitis pigmentosa in humans. Dogs with progressive rod-cone degeneration (prcd) have reduced DHA levels in the plasma, ¹ similar to the findings in humans with retinitis pigmentosa, and also have lower DHA levels (approximately 20%) in their rod photoreceptor outer segments (ROS). ² Although ROS DHA levels in humans with retinitis pigmentosa have not yet been examined, pronounced reduction in DHA levels has been observed in the sperm (the only other body fluid or tissue in the body, besides retina and brain, that is enriched in DHA) of male patients affected with retinitis pigmentosa. ³ Thus, the prcd-affected dogs are an excellent animal model for the human disease to study the mechanisms underlying the lower DHA phenotype and its cause-and-effect relationship with the disease process. 

In an effort to correct or slow down the progressive loss of photoreceptor cells, we previously gave daily DHA supplements to prcd-affected dogs for 5 months and assessed subsequently the prcd disease phenotype. ² We found that, although DHA supplementation produced a sustained elevation in plasma and liver DHA levels, the ROS DHA levels in the affected animals remained lower than those in control dogs, and the prcd disease phenotype was not changed by this dietary manipulation. These results suggest that the synthesis of DHA-containing phospholipid molecular species in the retinas or the delivery of DHA from circulation to the retina could be downregulated in the affected animals. 

In the present study, we examined the possible contribution by retinal pigment epithelial (RPE) cells to the lower ROS DHA phenotype. Our study indicates that RPE cells from affected animals have a similar capability for synthesis of DHA from eicosapentaenoic acid (20:5n-3) and for release of DHA to the surrounding media, compared to those cells from normal animals. 

Materials and Methods

Animals

The dogs used in this study were derived from miniature poodles and were homozygous normal or affected with the prcd mutation; the dogs were bred and raised as part of a colony supported by an NEI/NIH grant (EY06855, “Models of Hereditary Retinal Degeneration”). All dogs were maintained in the same animal care facility (Retinal Disease Studies Facility, Kennett Square, PA) with controlled cyclic illumination (12-hour light–dark). Four prcd-affected animals (2–3 months old) and three age-matched normal dogs were used for the studies. The prcd-affected animals at this age show no morphologic evidence of retinal disease. All studies described were performed in compliance with ARVO Resolution on the Use of Animals in Ophthalmic and Vision Research and followed protocols approved by the animal protocol review committees of Cornell University and University of Oklahoma Health Sciences Center. The dogs were euthanatized with a barbiturate overdose, and the eyes were enucleated and transported to the laboratory in a chilled Puck F saline with calcium and antibiotics. 

RPE Cultures

RPE cells from normal and prcd-affected animals were isolated as previously described. ⁴ Briefly, after lens, vitreous, and retinas were removed aseptically, the RPE cells were released by repeated trypsinization. Dissociated cells were placed in DMEM medium containing 15% fetal bovine serum (FBS) and plated in 35-mm dishes at a density of 2 × 10⁵ cells/dish. The cells were allowed to attach to the plate and grow. At confluence, RPE cells were subcultured (1:3 ratio) by trypsinization. The confluent first passage (P1) cultures from both normal and affected RPE cells were used for the following isotope incubations. 

Incubation Conditions

Elongation and desaturation of [³H]20:5n-3 (American Radiolabeled Chemicals, St. Louis, MO) were studied in confluent P1 RPE cultures grown in HL-1 medium (Hycor Biomedical, Irvine, CA) containing 10% FBS on 35-mm dishes. Incubations were carried out for 24 and 72 hours in duplicate with 2 ml culture medium containing 15 μCi (0.75 nmol) of [³H]20:5n-3 (conjugated with delipidated bovine serum albumin [BSA] at a molar ratio of 2:1). The cells were maintained at 37°C in an incubator containing 5% CO₂. At the end of the incubation period, the media were removed from the cells and centrifuged to remove cellular debris; the cells were rinsed twice with ice-cold HL-1 medium containing 10 μM fatty acid–free BSA and harvested by trypsinization. Both cells and clarified media were stored at –80°C until analysis. 

Lipid Extraction and Saponification

Before lipid extraction, the culture media were centrifuged at 100,000g for 1 hour to remove subcellular membrane contaminants. Total lipids from the cells and media were extracted twice with chloroform/methanol. Lipid extracts were dried under nitrogen and suspended in a known volume of ethanol. Aliquots were taken for total radioactivity determination; the remaining lipid extracts were suspended in 1 ml of 2% KOH (wt/vol) in ethanol and saponified at 100°C for 30 minutes. After cooling to room temperature, 1 ml water and 75 μl concentrated HCl were added, and the released free fatty acids were extracted three times with 2 ml hexane. 

HPLC Analysis of Fatty Acids

The extracted free fatty acids were converted to phenacyl esters and separated by high-performance liquid chromatography (HPLC) on a Sulpeco LC-18 column (25 cm × 4.6 mm). ⁵ The fatty acid phenacyl esters were eluted at a flow rate of 2 ml/min with a linear gradient of acetonitrile/water from 80/20 (vol/vol) to 92:8 (vol/vol) for 45 minutes, followed by holding at 92:8 (vol/vol) for 10 minutes and returning to 80:20 (vol/vol) for 5 minutes. The radioactivity profile was monitored by online scintillation counting (Flo-one A250; Radiomatic, Tampa, FL) using Ultima-Flo M (Packard Instrument, Downers Grove, IL) at 2.5:1 (vol/vol) ratio of cocktail to mobile phase. The phenacyl esters were monitored at an absorbance of 242 nm. Identities of individual fatty acids have been previously determined ⁵ and further confirmed by catalytic hydrogenation and HPLC analysis of hydrogenated products. 

Results

RPE cells from normal and prcd-affected dogs grew well in culture and maintained a nice, distinguishable epithelial cell morphology. The affected cells appeared the same as normal cells, except that they grew and reached confluency slightly faster than normal cells. 

Labeling in the Cells

The labeling profiles from both normal and affected RPE cells after 24- and 72-hour incubation with[ ³H]20:5n-3 are shown in Figure 1 . Elongation and desaturation products of 20:5n-3, namely 22:5n-3, 24:5n-3, 24:6n-3, and 22:6n-3, were all detected. The identities of these labeled compounds were confirmed by HPLC analysis of their respective hydrogenation products (data not shown). 

The relative percentages of radioactivity in precursor 20:5n-3 and its products are shown in Figure 2 . No significant difference was found on the relative amounts of labeling in DHA and other metabolites between normal and affected cells at either time point, although affected cells tended to have more but variable labeling in DHA. With increasing incubation time, both RPE cells had more labeling in 22:6n-3 and proportionally less labeling in 20:5n-3. 

Labeling in the Media

Distributions of radioactivity in 20:5n-3 and its metabolic products from culture media are shown in Figures 3 and 4 . After 24-hour incubation, no significant difference was detected on the relative (Fig. 3) or total amount (Fig. 4) of labeling in any of the fatty acids from normal and affected RPE cells. However, after 72-hour incubation, there was significantly more labeling (both relative and total amount) in DHA from affected cells than from normal ones. With increasing incubation time from 24 to 72 hours, the total amount of radioactivity in DHA was increased 4 times from affected cells and only 2 times from normal ones, whereas changes for 22:5n-3 and other products were much smaller. 

It is worth noting that the labeling in the media did not simply reflect that in the cells. Compared to the radioactivity distribution in the cells (Fig. 2) , relatively more DHA and less 22:5n-3 was found in the media at either incubation time (Fig. 3) for both types of cells. As a result, the ratio of DHA/22:5n-3 was 3 times higher in the media than in the cells, indicative of preferential release of DHA over 22:5n-3 in these cells. 

Discussion

Dogs with prcd have abnormalities in long-chain polyunsaturated fatty acid metabolism, with reduced DHA levels (approximately 20%–25%) in plasma and ROS and slightly elevated levels in the liver, compared to unaffected dogs. ^{1 2} The mechanism for the reduction in ROS DHA levels in prcd-affected animals is not clear. It was initially thought that synthesis of DHA in the liver and/or transport to the target tissues could be affected. We tested this hypothesis previously and tried to bypass the possible defect in DHA production in the liver by providing preformed DHA to affected and normal dogs for up to 5 months. ² We found that supplementation caused a rapid and sustained elevation in plasma DHA levels in both groups, but had no effect on the progression of the disease in prcd-affected animals. The prcd-affected dogs fed a DHA-enriched diet still had significantly lower DHA levels in ROS compared to control dogs. ² In dogs supplemented with DHA for up to 24 days, levels of DHA were lower in ROS and higher in the liver from prcd-affected dogs. These studies indicate that the lower content of ROS DHA in prcd-affected dogs is not the result of a dietary deficiency of DHA and its precursor fatty acids. Besides a possible problem in packaging DHA into lipoproteins in the liver, ⁶ these results suggest the possibility that the problem in prcd-affected animals could be in DHA trafficking between the blood and the retina, possibly involving passage through RPE cells. 

In this study, we investigated a possible defect in RPE of prcd-affected dogs in the synthesis and release of DHA to the extracellular space, by comparing these two activities in the RPE of prcd-affected and normal animals. This is important because RPE cells could use 22:5n-3, which is slightly elevated in the plasma of affected animals, ¹ to synthesize DHA to compensate for the reduced DHA levels in ROS of prcd-affected animals. RPE cells of frogs have been shown to readily synthesize DHA from its precursors. ⁷ Also, because RPE cells are the major nutrient supplier for the photoreceptors and play central roles in recycling of DHA from daily shed ROS tips back to photoreceptors, ^{8 9} it is conceivable that alterations in DHA metabolism in the RPE could affect photoreceptor DHA content. We found that RPE cells from both prcd-affected and normal dogs can elongate 20:5n-3 to 22:5n-3 and further convert 22:5n-3 to 22:6n-3 through Sprecher’s Δ4-desaturase independent pathway ¹⁰ (22:5n-3 → 24:5n-3 → 24:6n-3 → 22:6n-3). Rates for DHA synthesis appeared only slightly higher for affected RPE compared to normal cells. Thus, in agreement with our previous in vivo studies, ¹¹ there is no defect in DHA synthesis in the RPE of prcd-affected dogs. Furthermore, under our experimental conditions, RPE cells from prcd-affected animals showed no apparent defect in releasing DHA, inasmuch as the level of DHA in the culture media from affected cells was clearly not lower than that from normal cells. 

It is noteworthy that the labeling pattern in the culture media did not simply reflect that of the cells. When the distribution of radioactivity was compared, clearly more DHA and less 22:5n-3 was found in the media than in the cells. Thus, dog RPE cells appeared to favor the release of DHA over 22:5n-3, although both are similar in structure, with 22:5n-3 having one less cis double bond. Evidently the types of fatty acids in the media are not the result of equilibration, but rather of cellular regulation, which also has been demonstrated in retinal and cerebral microvascular endothelial cells. ^{12 13}  

In light of our earlier ^{2 11 14} and present studies on prcd-affected dogs, it is possible that lower ROS DHA levels could be due to cellular upregulation of DHA catabolism and/or downregulation of DHA incorporation into membranes to reduce disease-originated metabolic (perhaps oxidative) stress. DHA provides optimal lipid environment for rhodopsin and probably other membrane proteins in photoreceptors and thus is essential for optimal photoreceptor function. On the other hand, DHA is highly unsaturated (six double bonds) and easily peroxidized, especially in the photoreceptor cells, where high levels of oxygen and unsaturated fatty acid content and light exposure provide the ideal environment for lipid peroxidation. Toxic factors from lipid hydroperoxides derived from DHA could affect photoreceptor enzyme activities and damage photoreceptor membranes, as has been reported for light damage in albino rats. ^{15 16 17} Biological adaptation, such as shortened ROS length, reduced levels of rhodopsin and DHA-containing glycerolipids, and increased antioxidants vitamin E and vitamin C and glutathione-dependent enzyme activities, as well as neurotrophic factors, in response to light stress has been shown in the retinas of rats raised in bright cyclic light compared to controls raised in dim cyclic light. ^{16 17 18 19} Similarly, oxidative stress achieved through the stimulation of endogenous oxidant generation in human spermatozoa can cause DNA fragmentation and loss in their capacities for movement and oocyte fusion. ²⁰ Therefore, it is possible that reduction in DHA levels in retinas and plasma could be part of a biological adaptation to metabolic stress, possibly oxidative, caused by different mutations in retinitis pigmentosa. We currently are testing the hypothesis that mutations in a number of genes encoding retina-specific proteins can cause lower DHA phenotypes. 

 Supported by the National Institutes of Health (Grants EY06855, EY00871, and EY04149); Research to Prevent Blindness, Inc.; The Foundation Fighting Blindness; Samuel Roberts Noble Foundation; and Presbyterian Health Foundation. REA is a Senior Scientific Investigator for Research to Prevent Blindness, Inc.

 Submitted for publication December 10, 1998; revised April 7, 1999; accepted May 5, 1999.

 Proprietary interest category: N.

 Corresponding author: Robert E. Anderson, Dean A. McGee Eye Institute, 608 Stanton L. Young Boulevard, Room 409, Oklahoma City, OK 73104. E-mail: [e-mail]

 

Figure 1.

 

HPLC elution profiles of radiolabeled 20:5n-3 and its metabolic products from normal and prcd-affected dog RPE cells after incubation for 24 and 72 hours. Confluent dog RPE cells (3 normal and 4 affected) were incubated in duplicate at 37°C for 24 and 72 hours with 2 ml of HL-1 medium containing 10% fetal bovine serum and 15 μCi [³H]20:5n-3. Total lipids from the cells were saponified, and the released free fatty acids were converted to phenacyl esters and analyzed by HPLC and flow-through scintillation counting.

Figure 2.

 

Relative distribution of radioactivity in 20:5n-3 and its metabolic products from normal and prcd-affected dog RPE cells after incubation for 24 and 72 hours. Confluent dog RPE cells (3 normal and 4 affected) were incubated in duplicate at 37°C for 24 and 72 hours with 2 ml of HL-1 medium containing 10% fetal bovine serum and 15 μCi [³H]20:5n-3. Total lipids from the cells were saponified, and the released free fatty acids were converted to phenacyl esters and analyzed by HPLC and flow-through scintillation counting. Values are means ± SD.

Figure 3.

 

Relative distribution of radioactivity in 20:5n-3 and its metabolic products from culture media of 24- and 72-hour incubations with normal and prcd-affected dog RPE cells. Confluent dog RPE cells (3 normal and 4 affected) were incubated in duplicate at 37°C for 24 and 72 hours with 2 ml of HL-1 medium containing 10% fetal bovine serum and 15 μCi [³H]20:5n-3. Total lipids were extracted from the media and saponified, and the released free fatty acids were converted to phenacyl esters and analyzed by HPLC and flow-through scintillation counting. Values are means ± SD.* Statistically significant difference (P < 0.05) as assessed by Student’s two-tail t-test.

Figure 4.

 

Total radioactivity in 20:5n-3 and its metabolic products from the culture media of 24- and 72-hour incubations with normal and prcd-affected dog RPE cells. Confluent dog RPE cells (3 normal and 4 affected) were incubated in duplicate at 37°C for 24 and 72 hours with 2 ml of HL-1 medium containing 10% fetal bovine serum and 15 μCi [³H]20:5n-3. Total lipids were extracted from the media and saponified, and the released free fatty acids were converted to phenacyl esters and analyzed by HPLC and flow-through scintillation counting. Values are means ± SD. *Statistically significant difference (P < 0.05) as assessed by Student’s two-tail t-test.