Dietary Docosahexaenoic Acid Supplementation Prevents Age-Related Functional Losses and A2E Accumulation in the Retina

Abstract

Purpose: With age, retina function progressively declines and A2E, a constituent of the toxin lipofuscin, accumulates in retinal pigment epithelial (RPE) cells. Both events are typically exacerbated in age-related retina diseases. We studied the effect of dietary docosahexaenoic acid (DHA, C22:6n-3) supplementation on these events, using a transgenic mouse model (mutant human ELOVL4; E4) displaying extensive age-related retina dysfunction and massive A2E accumulation.

Methods: Retina function was assessed with the electroretinogram (ERG) and A2E levels were measured in E4 and wildtype (WT) mice. Dietary DHA was manipulated from 1 to 3, 1 to 6, 6 to 12, and 12 to 18 months: 1% DHA over total fatty acids (E4+, WT+) or similar diet without DHA (E4−, WT−).

Results: Increased omega-3/6 ratios (DHA/arachidonic acid) in E4+ and WT+ retinas were confirmed for the 1- to 3-month and 1- to 6-month trials. Although 1- to 3-month intervention had no effects, when prolonged to 1 to 6 months, RPE function (ERG c-wave) was preserved in E4+ and WT+. Intervention from 6 to 12 months led to maintained outer and inner retina function (ERG a- and b-wave, respectively) in E4+. At 12 to 18 months, a similar beneficial effect on retina function occurred in WT+; A2E levels were reduced in E4+ and WT+.

Conclusions: DHA supplementation was associated with: preserved retina function at mid-degenerative stages in E4 mice; prevention of age-related functional losses in WT mice; and reduced A2E levels in E4 and WT mice at the oldest age examined. These findings imply that dietary DHA could have broad preventative therapeutic applications (acting on pathologic and normal age-related ocular processes).

Introduction

With age, the eye undergoes a decline in retina function as assessed with the electroretinogram (ERG), ^1–4 and toxic substances, such as the vitamin A–derived pyridinium bisretinoid isomer A2E (a constituent of lipofuscin), accumulate in the retinal pigment epithelium (RPE). ^5,6 Both of these age-related changes are already detectable in young adults affected by Stargardt macular dystrophy, for both the recessive (STGD1) ^7,8 and the dominant (STGD3) ^9–11 forms. STGD3 is caused by mutations in the ELOVL4 gene, ¹² which encodes an enzyme involved in the elongation of very long chain fatty acids (at least 28 carbon chains). ¹³ Hubbard et al. ¹⁴ reported that the severity of STGD3 is inversely proportional to the blood level of docosahexaenoic acid (DHA). Chiefly derived from dietary intake, DHA is the major fatty acid in photoreceptor outer segments, where phototransduction occurs. ^15,16 The progressive decline in DHA levels with normal aging as well as with disease ^17,18 suggests a potential role of dietary DHA supplementation in preserving retina function. Support for a direct link between higher dietary intake of DHA and milder STGD3 phenotype comes from the study of MacDonald et al., ¹⁹ in which retina function (including assessment with the ERG) was preserved following dietary DHA supplementation in a 15-year-old STGD3 patient with significant macular dysfunction. A similar supplementation regimen in seven additional patients with ELOVL4 mutations confirmed the correlation between higher plasma DHA levels and preserved visual function (MacDonald IM, et al. IOVS 2004;45:ARVO E-Abstract 1768). 

We relied on a model of STGD3, the mutant human ELOVL4 transgenic mouse, ²⁰ to examine whether DHA nutritional supplementation might delay decline in ERG responsiveness and A2E accumulation, which are landmark phenotypes in this murine model. The data presented here support that DHA nutritional supplementation is associated with a preventative effect not only in a mouse model of STGD3 but also in wildtype (WT) mice, which naturally display age-related declines in ERG responsiveness and accumulation of A2E deposits in the RPE. 

Materials and Methods

Animals

Heterozygous females from the E4 transgenic (line TG2) mouse model of STGD3²⁰ were bred with C57BL/6N male mice (Charles River Laboratories, Wilmington, MA) in a colony maintained at the University of Alberta. Litters were genotyped by PCR as previously described. ²¹ All animals were maintained on a 12:12 light–dark cycle (to ensure a normal production of the interphotoreceptor retinoid binding protein mRNA²²), temperature 21°C, relative humidity approximately 50%, and supplied with water and food without restriction (see dietary manipulation for details). Experiments were carried out in accordance with the Institutional Animal Care and Use Committee (University of Alberta) and the ARVO (Association for Research in Vision and Ophthalmology) Statement for the Use of Animals in Ophthalmic and Visual Research. 

Primary Antibodies

Bipolar cells were labeled with a polyclonal anti–protein kinase C subunit-alpha (PKCα) antibody (raised in rabbit against an immunizing peptide containing amino acids 650–690 from the C-terminus of human PKCα; catalog #sc-208, 1:200; Santa Cruz Biotechnology Inc., Santa Cruz, CA); on Western blots from mouse retina proteins, it detects both major (∼80 kDa) and minor bands corresponding respectively to the βI and βII isoforms (manufacturer's datasheet). This antibody was shown to label both rod bipolar cells and a subset of amacrine cells as efficiently as PKCα monoclonal antibodies. ²³ Cone photoreceptors were labeled using an anti–γ-transducin polyclonal antibody (raised in rabbit against recombinant human Gγc protein), which recognizes the 10-kDa retinal cone Gγ8 subunit (catalog # PAB-00801G, 1:1000; CytoSignal, Irvine, CA). As formerly shown in other rodents, ^24,25 this antibody did not stain rods. In addition, the two following polyclonal antibodies were used: anti-M/L-opsin (raised in rabbit against an immunizing peptide containing the last 42 amino acids of recombinant human red/green opsin; catalog # AB5405, 1:500; Chemicon, Eugene, OR) and anti–S-opsin (raised in goat against an immunizing peptide containing the amino acid sequence EFYLFKNISSVGPWDGPQYH located at the N-terminal of the human blue-sensitive opsin, catalog # sc-14363, 1:200; Santa Cruz Biotechnology). These two antibodies label the outer segments and cell membranes of specific types of cones in mouse retinas. ²¹  

Dietary Manipulation

Dietary DHA intake was manipulated at four time windows: 1–3 months; 1–6 months; 6–12 months; and 12–18 months. Both E4 and WT mice were fed a nutritionally complete, semipurified diet containing either 0% DHA (control diet; E4−, WT−) or 1% DHA (wt/wt, total fatty acids; E4+, WT+). DHA was provided as triglycerides isolated from single-cell algal source (DHASCO; Martek Biosciences Corporation, Columbia, MD). Previous studies have shown that the same 0% DHA diet does not cause DHA deficiency in the retina. ^27–29 In the present study, we confirmed that the ratios of DHA/arachidonic acid (DHA/AA; a reliable indicator of retina degeneration³⁰) in E4− and WT− at the end of the 1- to 6-month trial (see values in the following text) were analogous to those previously measured ²¹ in 6-month-old E4 (3.1 ± 1.0) and WT (4.1 ± 0.8) mice fed commercial chow (LabDiet #5001; PMI Nutrition Intl., Richmond, IN; routinely used for the maintenance of our E4 and WT mouse colonies^21,26) from weaning age 21 days onward. We analyzed the commercial chow (LabDiet #5001 chow) and found the proportion of DHA over total fatty acids to be about six times lower (0.17 ± 0.03%) than the DHA+ diet (1%). Finally, our studies of retina function showed no differences between chow-fed (LabDiet #5001) and 0% DHA fed mice: the amplitudes of both a- (outer retina activity) and b- (inner retina activity) ERG waves in 6-month-old E4−/WT− mice and E4/WT mice fed chow are comparable. For instance, b/a ratios in dark-adapted conditions are respectively 1.8 ± 0.2 in WT− and 1.6 ± 0.2 in WT-fed chow. Table 1 provides the diet content for both experimental groups; total fatty acid levels were the same in both diets. The custom-made food was stored at −20°C and replaced in clean food jars every 3 days. 

Table 1.

 

Diet Composition for the Two Experimental Groups: Control (0% DHA) and DHA (1% DHA)

Ingredients	Control (g/kg)	DHA (g/kg)
Basal
Casein, high protein	270.0	270.0
l-Methionine	2.5	2.5
Dextrose monohydrate	208.485	208.485
Corn starch	200.0	200.0
Cellulose	50.0	50.0
Mineral mixture	50.85	50.85
Sodium selenite	0.3	0.3
Manganese sulfate	0.24	0.24
Vitamin mix (AOAC)	10.0	10.0
Inositol	6.25	6.25
Choline chloride	1.375	1.375
Fat blend
Corn oil	42.0	41.076
Canola oil	92.0	89.976
Coconut oil	46.0	44.988
Oleo oil	20.0	19.560
DHASCO	0.0	4.40

 AOAC, Association of Analytical Communities (http://www.aoac.org/); DHASCO, oil extracted from the unicellular alga Crypthecodinium cohnii mixed with high oleic sunflower oil (HOSO). DHASCO contains 46% DHA and 44% oleic acid.

Immunohistochemistry

To assess the transgenic phenotype, the retinal morphologies of both E4 (n = 15) and WT (n = 10) littermate chow-fed mice were studied immunohistochemically as previously described ^21,26 at 1, 3, 6, 12, and 18 months. In brief, after euthanasia, the eyes were enucleated between 09:00 and 11:00 hours in the morning, cornea and lens were removed, and the eyecups were lightly postfixed in 4% paraformaldehyde. Following cryoprotection in graded sucrose, the eyecups were embedded and frozen (Shandon Cryomatrix; Anatomical Pathology USA, Pittsburgh, PA). Cross-sections were cut at 20-μm thickness, parallel to the temporonasal axis through the optic nerve head and mounted on slides. After hydration in PBS (pH 7.3), sections were blocked for 1 hour in PBS + 0.3% Triton X-100 + 10% goat (or horse) serum (same species as secondary antibody), and reacted overnight in a humid container with the primary antibodies, either alone or in combination, and diluted in a 1:10 solution of the previous blocking medium. After extensive washing with PBS, sections were reacted for 1 hour with species-appropriate secondary antibodies conjugated to fluorescent dyes (Alexa Fluor series; Molecular Probes, Eugene, OR), diluted 1:1000 in a 1:10 solution of the blocking medium, washed again in PBS, coated with antifade reagent with DAPI (Prolong Gold, catalog # P36931; Molecular Probes, Eugene, OR), and coverslipped. All reactions were performed at room temperature. For cone counts, retina flatmounts ²⁵ were used instead of cross-sections. Images were captured on a confocal microscope (Zeiss LSM510 META; Carl Zeiss MicroImaging GmbH, Jena, Germany) using a ×40/1.3 oil objective (Plan-Neofluar; Carl Zeiss Microscopy, Oberkochen, Germany). Brightness and contrast levels were adjusted as necessary using commercial software (Adobe Photoshop CS2 software Version 9.0.2; Adobe, San Jose, CA). 

ERG Recordings

To examine the impact of DHA intake on retina function, dark- and light-adapted ERG responses were respectively recorded ²¹ in four groups: E4−, WT−, E4+, and WT+ mice. In brief, mice were dark-adapted for 1 hour prior to anesthesia with a mixture of ketamine (62.5 mg/kg, administered intraperitoneally [IP]) and xylazine (12.5 mg/kg, IP) and pupil dilation with 1% tropicamide. Body temperature was monitored with a rectal thermometer and maintained at 38°C with a homeothermic electrical blanket. Simultaneous bilateral recording was achieved with active gold loop electrodes (placed on each cornea) and a subdermal platinum reference electrode (placed behind each eye); a subdermal ground platinum electrode was inserted in the mouse's scruff. Light stimulation (10-μs duration flashes), signal amplification (0.3–300 Hz bandpass), and data acquisition were provided by a full field ERG system (Espion E ² system; Diagnosys LLC, Littleton, MA). Analysis of RPE function was derived from estimating the c-wave amplitude by subtracting the ERG amplitude at 800 ms poststimulus from the negative peak (also referred to as the tail of PIII³¹) that immediately follows the b-wave. The rationale for using 800 ms was that this time point gave the highest positive peak in WT mice (the same value was also obtained by an independent lab, under similar conditions³²), and that it avoided having to make statements on multiple positive-going peaks elicited from transgenic mice for which no clear c-wave was apparent between the tail of PIII and 2000 ms poststimulus. For each animal, only one eye was considered for statistical comparisons: it corresponded to the eye associated with the dark-adapted a-wave of highest maximal amplitude. 

DHA Measures

DHA measures were taken at 1, 3, and 6 months (i.e. at the beginning and the end of the two first DHA supplementation windows). A total of five retinas were pooled per individual measure; each group (E4+, E4−, WT+, and WT−) was measured either in quadruplicate (1 and 3 months) or in duplicate (6 months). The percentage of DHA over total fatty acids in retinas was determined as previously described. ³³ In brief, membrane lipids were extracted and phospholipids were isolated on thin-layer chromatography plates (Analtech silica-gel G; Mandel Scientific Co., Guelph, Ontario, Canada) with petroleum ether/diethyl ether/formic acid (60:40:1.6, v/v) used as the solvent system. Fatty acid methyl esters were prepared using BF₃/methanol reagent and separated by automated gas–liquid chromatography (Vista 6010 G.L.C. and Vista 402 data system; Varian Instruments, Mississauga, Ontario, Canada). Chromatography was performed using a fused silica capillary column (BP20 GC column: 25 m × 0.24 mm ID; Varian Instruments). Helium was used as the carrier gas at a flow rate of 1.8 mL/min using a splitless injection mode. The initial oven temperature was 150°C, increased to 190°C at 20°C/minute and held for 23 minutes, then increased to 220°C at 2°C/minute, for a total analysis time of 40 minutes. These analytical conditions separated all saturated, mono-, di-, and polyunsaturated fatty acids from C₁₂ to C₂₄ in chain length. 

A2E Measures

A recent study ²¹ from this laboratory showed that age-related retinal accumulation of A2E was exacerbated in E4 compared with WT mice. Differences were significant as early as 6 months of age. We therefore prioritized the use of nutrition trials at older ages (6–12 and 12–18 months) for determining retina A2E levels as a potential outcome measure of the effect of DHA intake on retina integrity. Since A2E measures are terminal, these were taken at the end of the nutrition trials. A total of six eyecups were pooled per measure; triplicate measures were performed at each time point. A2E and its isomer iso-A2E (referred to collectively as A2E in the text) were measured as previously described. ^34,35 In brief, mouse posterior eyecups were homogenized, extracted with chloroform/methanol (2:1), and analyzed on an HPLC system (Alliance; Waters Corp., Milford, MA) with a universal, silica-based, reversed-phase C18 column (Atlantis dC18 column; 3 μm, 4.6 × 150 mm) and monitored by photodiode array. For the mobile phase, gradients of water and acetonitrile with 0.1% trifluoroacetic acid were used with a flow rate of 0.8 mL/min. The bisretinoid lipofuscin compounds A2E and iso-A2E were identified on the basis of UV-visible absorbance spectra and elution times that correspond to authentic synthetic compounds. ^34,36,37 Molar quantities per mouse eye were determined using standard curves, constructed from known concentrations of purified external standards, and summed for data and graphs. 

Statistics

Unless otherwise noted, differences between groups were tested for significance using the nonparametric Mann–Whitney U test as required for small sample sizes. Comparison between traces (see Figs. 4G and 5A) was achieved using repeated-measures ANOVA with the Greenhouse–Geisser correction for sphericity (SPSS Inc., Chicago, IL). Significance was set at P < 0.05. Values in text and data points on graphs represent mean ± SD. 

Results

Age-Related Anatomic Changes in the Retina of E4 and WT Mice

The anatomic state of the retina in E4 and WT mice was assessed at the time points corresponding to the beginning and end of all the nutritional interventions undertaken in this study (Fig. 1). In E4 mice, the first anatomic changes occurred at 2 months ²⁶ and consisted of rod loss in the central retina, followed at 6 months ²¹ by the accumulation of autofluorescent debris in the subretinal space (arrow in Fig. 1) and by the truncation of rod outer segments. ²⁰ The specific loss of rods in this model can be deduced from the fact that the outer nuclear layer (ONL) undergoes a progressive thinning despite a stable cone population (labeled with S and M/L opsin antibodies) until 12 months. Counts performed at this time point in retinal flat mounts stained for γ-transducin yielded a total population of 189,600 ± 26,600 and 186,150 ± 36,400 cone photoreceptors in E4 and WT mice, respectively. These values are in agreement with those reported previously in the C57BL/6 mouse strain. ^38,39 Cone loss became apparent in E4 mice by 18 months; in such animals, the ONL consisted of a monolayer of cones (≈70% the normal density) expressing opsins in their cell bodies and in aberrant processes (Figs. 1B, 1C). In WT mice, there was no significant loss of photoreceptors up to 18 months. The only visible age-related anatomic change consisted of rod bipolar cell dendrites sprouting in the ONL ²¹ (Fig. 1A); such sprouting was never seen in E4 mice. ²¹  

Figure 1.

 

Anatomy of E4 and WT mouse retinas at key ages for dietary manipulations (1, 3, 6, 12, and 18 months). (A) Rod bipolar cells labeled with anti-PKC antibody (red) and nuclei stained with DAPI (blue), showing the progressive loss of nuclei in the outer nuclear layer (ONL, located above the bipolar cell bodies); there was a 2/3 reduction of ONL thickness from 1 to 6 months. At 18 months, only a single layer of ONL remained in E4. (B) Double labeling with S-opsin (green) and M/L-opsin (red); although actual loss of cones occurred after 12 months, there was a progressive shortening of cone photoreceptor outer segments beginning at 6 months in E4. By 18 months, there were no detectable outer segments and opsins accumulated in the remaining cones of the monolayered ONL, revealing their cell bodies and anomalous processes; arrow points to double-labeled subretinal deposit at 6 months in E4. (C) Same staining as in B, showing examples of cone photoreceptors labeled with S-opsin; 18-month cones displaying severe morphologic distortions. For all panels, outer retina is upward. Scale bar, 20 μm.

Histologic analysis of retinas of mice fed custom diets was not performed. First, it is well established that changes in DHA levels induced by dietary intake or genetic manipulation do not affect the morphology or number (inferred from counting ONL rows) of photoreceptors in either E4 or WT mice. ^40–44 Second, potential effects of DHA intake on outer retina anatomy in E4 mice would be difficult to assess in view of the large variability in the number of ONL rows observed between age-matched transgenic mice. ^21,26  

Increased Retina DHA/AA Ratios in DHA Supplemented Mice

Initial studies performed at 1 month of age on E4 and WT pups fed breast milk (followed by standard chow for 1 week) revealed a larger proportion of DHA over total fatty acids in E4 (22.03 ± 2.57%) versus WT (18.9 ± 1.86%), but a nearly equal proportion of AA in both groups (E4: 7.33 ± 0.47%; WT: 7.5 ± 0.26%). As a result, the DHA/AA ratio in E4 mice (3.02 ± 0.53) was significantly (P = 0.038) higher than that in WT mice (2.51 ± 0.17). At the end of the 1- to 3-month trial, both E4− and WT− control mice displayed equivalent proportions of DHA (25.5 ± 3.0% and 26.0 ± 2.2%, respectively) and AA (7.18 ± 0.67% and 7.01 ± 1.06%, respectively), thus roughly similar DHA/AA ratios (3.60 ± 0.67 and 3.77 ± 0.61). A similar outcome was observed with a different approach. ⁴⁴  

The goal of DHA supplementation was to prevent a decline in the DHA/AA ratio, and thus its measure as an indicator of the effectiveness of the nutritional supplementation. As expected, for the 1- to 3-month trial, DHA/AA ratios were increased in E4+ versus E4− and in WT+ versus WT− (Fig. 2). Nutritional supplementation had similar outcomes on DHA/AA ratios regardless of the genotypes. Finally, as observed in humans (MacDonald IM, et al. IOVS 2004;45:ARVO E-Abstract 1768), there was a significant (P < 0.05) increase in plasma DHA levels in supplemented (12–14.5 μg/mL) compared with control (≈5.4 μg/mL) E4 and WT mice. Measures done at the end of the 1- to 6-month trial showed that retina DHA/AA ratios were lower in E4 (3.3 and 3.0 for E4+ and E4−, respectively) compared with WT retina (4.3 and 3.8 for WT+ and WT−, respectively). Although the smaller number of animals used for this trial (see the Materials and Methods section) precluded statistical comparisons, the trend was the same as that for the 1- to 3-month trial. A trend analysis (R ²) was performed on the two independent data sets (from E4−, E4+, WT−, and WT+, respectively), yielding values > 0.99. 

Figure 2.

 

The DHA/AA ratio is higher in supplemented (DHA+) versus nonsupplemented (DHA−) groups; two-way ANOVA returned a significant diet effect (F _1,15 = 5.33; P = 0.035). Although there was a trend for the DHA/AA ratio to be lower in E4 than that in WT mice, this trend was not significant (F _1,15 = 2.64; P = 0.125). Left panel: 1- to 3-month DHA supplementation. Right panel: 1- to 6-month DHA supplementation. Error bars represent SD.

Dietary DHA Supplementation Has No Effect on Retina Function for the 1- to 3-Month Intervention

Comparison between E4 and WT at 3 months did reveal a significant amplitude reduction in dark-adapted a-wave, b-wave, c-wave, and double-flash–isolated dark-adapted cone b-wave in E4 versus WT groups. Pure cone responses recorded under light adaptation were similar between E4 and WT groups. 

Nutritional supplementation with 1.0% DHA from 1 to 3 months (at which point anatomic and functional rod loss had already begun in E4 mice) did not cause any changes in ERG responsiveness between E4+ and E4− or between WT+ and WT−. 

Dietary DHA Supplementation Preserves RPE Function for the 1- to 6-Month Intervention

E4 at 6 months showed declining amplitudes for all dark-adapted ERG components, when compared with that at 1 month. In WT, only one component (dark-adapted c-wave) underwent amplitude reduction between 1 and 6 months of age (Fig. 3). 

Figure 3.

 

DHA supplementation from 1 to 6 months preserves RPE function (ERG c-wave). Regardless of the diet, E4 mice had significantly lower c-wave amplitudes than those of WT mice (P < 0.05). DHA supplementation resulted in significantly higher ERG c-wave amplitudes in both E4 and WT mice (asterisks; P < 0.05). Error bars represent SD.

DHA supplementation from 1 to 6 months was not associated with preservation of a- or b-wave amplitudes. DHA supplementation, however, showed a preventative effect on the declining c-wave in both E4 and WT (Fig. 3). 

Preventative Effect of Dietary DHA Supplementation on Retina Function for the 6- to 12-Month Intervention

Further declining ERG responsiveness (now including light-adapted cone-driven a- and b-waves) was observed in E4 mice at 12 months when compared with that at 6 months. In WT, dark-adapted c-wave amplitude reduction was even more pronounced than that at 6 months, without changes in any other ERG components. 

ERG recordings in DHA supplemented versus control animals showed maintained retina function from 6 to 12 months in E4 mice. Cone photoreceptor contribution to the ERG, reflected by light-adapted a-wave amplitude, was preserved in E4+ versus E4− (Figs. 4A, 4B). Further support for an effect of DHA supplementation on cone function comes from double-flash–isolated cone-driven responses under dark adaptation: b-wave amplitudes (Figs. 4C, 4D) and implicit times (Fig. 4E) were maintained in E4+ versus E4−. Finally, the light-adapted flicker (a well-established test of cone function) showed higher amplitudes in E4+ versus E4− over the entire range of frequencies tested, 3 to 35 Hz (Figs. 4F, 4G). There was no effect of DHA supplementation on ERG responsiveness for the WT mice supplemented between 6 and 12 months. 

Nutritional manipulation was not associated with any statistically significant changes in A2E levels (measured at 12 months) within WT and E4 groups (n = 3 measures of six pooled retinas per measure). However, the lowest levels of A2E were found in DHA supplemented WT+ and E4+ mice, respectively, when compared with control animals. 

Preventative Effect of Dietary DHA Supplementation on Retina Function and A2E Levels for the 12- to 18-Month Intervention

Figure 4.

 

DHA supplementation preserves cone-driven function in E4 mice: 6- to 12-month intervention (E4− n = 6; E4+ n = 7; WT− n = 6; WT+ n = 5). (A) Representative traces of light-adapted a-wave elicited by low (−0.42 log cds/m², top row) and high (1.89 log cds/m², bottom row) intensity flashes. (B) a-wave amplitudes were higher in the DHA supplemented (+) versus control (−) groups at the high intensity stimulus (P < 0.05). (C) Representative traces of double-flash–isolated cone-driven responses at low (−0.42 log cds/m², top row) and high (1.89 log cds/m², bottom row) intensity stimuli. (D) Note the higher b-wave amplitudes in the (+) group for the low intensity stimulus (P < 0.05). (E) Normalized light-adapted b-waves (elicited by a flash of 2.39 log cds/m²) averaged between (+) and (−) E4 groups. Note the shorter implicit time for the (+) group (P < 0.05). (F) Representative traces of white flicker responses from low (3 Hz, top row) to high (35 Hz, bottom row) frequency flashes. (G) Peak-to-peak amplitudes for flicker stimuli elicited at increasing frequencies. Statistical comparisons showed significantly higher amplitudes for (+) versus (−) groups from low to high frequency stimuli (P < 0.05; Bonferroni repeated-measures ANOVA). Error bars represent SD.

Figure 5.

 

DHA supplementation preserves cone-driven function in WT mice: 12- to 18-month supplementation ERG results (E4− n = 4; E4+ n = 6; WT− n = 6; WT+ n = 5). (A) Normalized mixed dark-adapted a-waves (elicited by a flash of 1.37 log cds/m²) averaged between DHA supplemented (+) and control (−) WT groups. Inset shows a significantly faster a-wave leading edge for the (+) versus (−) group (from 4.0 to 6.0 ms poststimulus, P < 0.05). (B) Representative traces of light-adapted a-wave elicited by low (0.38 log cds/m², top row) and high (2.39 log cds/m², bottom row) intensity flashes. (C) a-Wave amplitudes were higher in the (+) versus (−) groups at the high intensity stimulus (P < 0.05). (D) Representative traces of double-flash–isolated cone-driven responses at low (−0.38 log cds/m², top row) and high (2.39 log cds/m², bottom row) intensity stimuli. (E) b-wave amplitude was larger in the (+) versus (−) group for the high intensity stimulus (P < 0.05). (F) Representative traces of flicker responses elicited at 5 Hz (top row) and 30 Hz (bottom row). (G) Peak-to-peak amplitudes for flicker stimuli. Statistical comparisons showed significantly higher amplitudes for (+) versus (−) groups for both the low and the high frequencies (P < 0.05; Bonferroni repeated-measures ANOVA). Error bars represent SD.

By 18 months, there was a further decline in responsiveness of all ERG components for E4 mice when compared with those at 12 months. Although much milder, a similar age-related decline in ERG responsiveness was observed in WT. 

DHA supplementation was associated with the preservation of several aspects of ERG responsiveness in WT but not in E4 mice (Figs. 5A–G). Under dark adaptation, phototransduction activation kinetics for mixed rod–cone responses was faster in WT+ compared with control WT− mice, as reflected by larger normalized a-wave leading edge amplitudes between 4 and 6 ms after flash presentation (Fig. 5A). Under light adaptation, DHA supplementation was associated with preserved cone-driven a-waves in WT mice (Figs. 5B, 5C). Dark-adapted cone-driven responses (isolated with a double-flash protocol) showed higher amplitude b-waves in WT+ versus WT− mice (Figs. 5D, 5E). Finally, preserved flicker responsiveness provided additional confirmation of the preventative effect of DHA supplementation on cone function (Figs. 5F, 5G). 

Intervention during this period showed that E4+ and WT+ groups had lower A2E levels in the retina when compared with control E4− and WT− groups, respectively (Fig. 6). Values were as follows: 2.4 ± 0.3 and 10.1 ± 0.9 pmol/eye for WT+ and WT−, respectively; 10.8 ± 1.2 and 14.0 ± 1.1 pmol/eye for E4+ and E4−, respectively. The effect of DHA supplementation on A2E reduction was lower in E4 (23% reduction) than that in WT (76% reduction) mice. A2E levels in WT+ mice compared with those seen at the youngest ages. ²¹  

Figure 6.

 

DHA supplementation from 12 to 18 months prevents A2E accumulation in E4 and WT mice. Animals used in the present work carried the Leu450Met substitution in the Rpe65 gene classically associated with reduced A2E accumulation. Total A2E levels (measured at 18 months) were lower in both DHA supplemented (+) groups compared with control (−) groups (P < 0.05). Error bars represent SD.

Discussion

We tested the hypothesis that dietary DHA supplementation might exert a preventative effect on deleterious age-related changes in degenerating and healthy retinas. To this extent, we relied on a mouse model of Stargardt-like retina degeneration (transgenic E4 mouse) and WT littermates. The main outcome measures showed delays in both loss of function and A2E accumulation in the retinas of E4 and WT mice. Optimal effects were achieved with interventions done at earlier time points in E4 compared with WT mice. The degree of degeneration in E4 retinas at later ages is too advanced for DHA to be of any benefit and, conversely, beneficial effects of DHA on normal retinal aging are not apparent until advanced age. Comparison of ERG amplitude data obtained at time points corresponding to the beginning of each trial (obtained in a previous study²¹) with time points at the end of the trials confirmed that retina function was not improved by the trials but preserved. These data support our initial hypothesis that DHA supplementation prevents the progression of deleterious age-related processes in degenerating and healthy retinas. 

The DHA/AA ratio declines as a function of healthy aging and pathology in the retina. ^17,18 Maintaining this ratio, as was achieved in E4 and WT mice in this study, is crucial for retina cell survival and function. An example of the consequence of lower DHA/AA ratios includes a study in which mice, fed abnormally high AA levels, had a higher risk of developing microphthalmia. ⁴⁵ A recent report suggesting that this ratio is decreasing in the Western diets, due in part to increased soybean oil consumption, ⁴⁶ does raise concerns about one of the many factors that could be related to the higher incidence of age-related macular degeneration in the Western World. ⁴⁷ Increasing the DHA/AA ratio via fish consumption has been shown to reduce risks of AMD. ⁴⁸ However, abnormal increases in DHA/AA retina ratio, as achieved in transgenic fat-1 mice, lead to retina dysfunction associated with increased gliosis and oxidative stress in photoreceptors. ⁴³ The fat-1 gene (fatty acid metabolism 1) from Caenorhabditis elegans , encodes for an n-3 fatty acid desaturase, absent in mammals, which transforms n-6 PUFA into n-3 PUFA, leading to extremely high DHA/AA ratios. ^43,49 . This likely explains the failed attempts at preventing retina degeneration in E4 mice by breeding them with fat-1 mice. ⁴⁴  

Preservation of DHA levels (while maintaining a normal DHA/AA ratio as described earlier) is essential for photoreceptor survival and function. ⁵⁰ Mechanisms through which tissue DHA status might have an impact on the function of photoreceptors include the control of permeability, fluidity, thickness, and lipid phase properties of their outer segment membranes (reviewed in SanGiovanni and Chew⁵¹). DHA has also been implicated in phototransduction signaling mechanisms such as G-protein coupled signaling ⁵² and in rhodopsin regeneration. ⁵³ Furthermore, DHA, which was shown to act as a cochaperone for heat shock protein-70, ^54,55 may prevent unfolded protein responses ⁵⁶ that occur in vitro in photoreceptors of the E4 mouse model. ⁵⁷ Heat shock protein-70 was shown to prevent photoreceptor degeneration ⁵⁸ and RPE autophagy ⁵⁹ occurring both with healthy aging of the retina and in AMD. ⁶⁰ Therapeutic mechanisms other than preventing the unfolded response, which has yet to be shown in STGD3 models in vivo, ^26,61 should be considered. First, DHA could prevent age-related retina A2E accumulation. Despite a lack of evidence that DHA can directly attenuate bisretinoid formation (such as A2E), ⁶ it could achieve a similar outcome by thwarting the activation of Toll-like receptors 1–6 and NF-κB (as was shown in vitro^62,63). Second, DHA could exert an antiapoptotic effect on photoreceptors via its metabolite, neuroprotectin D1 (NPD1). ^64,65 NPD1 is also required for RPE cell functional integrity and has been shown to prevent the apoptosis of RPE cells induced by A2E. ^66,67  

The earliest effect of DHA supplementation on retina function (observed for both E4 and WT mice at 6 months with onset of supplementation at 1 month) consisted of higher ERG c-wave amplitudes, an ERG component generated by light-induced changes in RPE transepithelial potential. Although increased c-wave amplitude typically implies functional preservation of RPE cells, ⁶⁸ this increase could also be a consequence of higher resistance of RPE cells via their incorporation of DHA as part of its recycling process. ⁶⁹ However, the latter is unlikely since the tight junctions established between cells forming this monolayer create most of the RPE resistance. ⁷⁰ Independently from the c-wave results, our data do provide direct evidence for the preventative effect of DHA on RPE integrity: DHA supplementation was associated with lower levels of A2E at 18 months in both E4 and WT mice when compared with age-matched control animals. The mechanism via which DHA supplementation might lead to lower retina A2E levels remains to be explored. 

Conclusion

Since the enzyme ELOVL4 is not involved in DHA synthesis (but instead of C28 and longer carbon chain fatty acids), DHA supplementation unlikely corrects the primary defect in E4 mice; however, it does delay retina degeneration in this mouse model of STGD3. The likely mechanisms that could account for the therapeutic effect of DHA nutritional supplementation, not only in E4 but also in aging WT mice, include the diminution of A2E combined with the neuroprotective effect of DHA's metabolite NPD1 on RPE and photoreceptors. Future studies will investigate the mechanisms preventing A2E accumulation and examine NPD1 activity in DHA supplemented versus nonsupplemented retinas. 

Acknowledgments

The authors thank Janet R. Sparrow, PhD (Columbia University) for having measured A2E levels in eyecups, Catherine Field, PhD (University of Alberta) for having measured DHA levels in diets used for this study, and Silvina C. Mema, MD (University of Alberta) for her technical assistance. 

Footnotes

 Supported in part by Canadian Institutes of Health Research Grant 151145, Alberta Heritage Foundation for Medical Research, Barbara Tuck/MacPhee Family Vision Research Award in Macular Degeneration, Canadian National Institute for the Blind, the Olive Young Foundation, the Lena McLaughlin Foundation (Mona and Rod McLennan), and the Foundation Fighting Blindness Canada (YS).

Footnotes

 Disclosure: B. Dornstauder, None; M. Suh, None; S. Kuny, None; F. Gaillard, None; I.M. MacDonald, None; M.T. Clandinin, None; Y. Sauvé, None