Enhancing the efficacy of AREDS antioxidants in light-induced retinal degeneration

Antioxidant supplements

Weanling rats were acclimated to their respective light environments for 1 day and then given daily antioxidant supplements for a period of 6 weeks. Every morning, each animal (n = 4–6 per treatment) was removed from its cage and placed in a clean cage without bedding but with a single piece of cracker infused with a small volume (about 50 μl) of antioxidant. The Group 1 antioxidant supplements consisted of AREDS complete (ICAPS soft-gels, Alcon Ltd, Fort Worth, TX), which contains the AREDS1 recommended doses of vitamins A (as beta carotene), C and E, zinc, and copper [6], AREDS antioxidants minus zinc, aqueous zinc oxide [25] (Alfa Aesar, Ward Hill, MA), or vehicle (1% Tween-80/16% ethanol/25% ROPUFA). The omega 3 polyunsaturated fatty acid ROPUFA mixture (75N-3 EE) was from DSM (Kaiseraugst, Switzerland). The Group 2 antioxidant supplements consisted of AREDS plus rosemary (17 mg/kg), zinc oxide plus rosemary, rosemary alone, or vehicle (1% Tween-80/16% ethanol/ 25% ROPUFA). Rosemary powder was obtained from LycoRed Ltd. (Beer Sheba, Israel) and emulsified as described [26]. A third group of animals (Group 3) was fed AREDS alone, AREDS plus rosemary (34 mg/kg), rosemary alone, or AREDS plus carnosic acid (Enzo Life Sciences, Farmingdale, NY), or AREDS plus ursolic acid (MP Biochemicals, Solon, OH) for a period of 1 week. Carnosic acid and ursolic acid were dissolved in 50% ethanol/ 50% ROPUFA and given at or above the concentrations found in 34 mg/kg of rosemary powder [26]. Other rats received antioxidant supplements as a single intraperitoneal (i.p.) injection 1 h before exposure to acute intense light. All rats were weighed every 2 days, and the volume of antioxidant was adjusted to provide the AREDS1 recommended daily dose. Daily feeding is less stressful than long-term oral gavage or multiple i.p. injections. The animals eagerly ingested the antioxidant-infused crackers, and their growth rates were identical to those fed only the house diet. Figure 1 describes the dietary paradigms used in this study, and Table 1 contains a summary of the antioxidant supplements, their daily doses, the total number of animals (n) for each antioxidant treatment, and the vehicles used.

Photoreceptor cell survival and antioxidant efficacy

The extent of visual cell loss was determined by measuring residual DNA in retinas from rats exposed to intense light, in comparison to retinal DNA measured in unexposed animals kept in darkness for the same 2-week post-light exposure period. DNA fluorescence was determined for individual retinas with the Hoechst (Calbiochem-Behring, La Jolla, CA) dye-binding assay, as described [26]. Photoreceptor cell DNA was calculated by subtracting the DNA content of the inner nuclear layers in 6-month old Royal College of Surgeons rat retinas from total retinal DNA. The relative level of protective efficacy was then determined by subtracting photoreceptor cell DNA in animals treated with vehicle and exposed to intense light from the DNA levels in rats treated with antioxidants and exposed to light, and from the DNA levels in unexposed-antioxidant-treated rats [25].

DNA damage

Rat retinas were excised 2 days after exposure to intense light, and DNA was isolated by using a Sigma Gene Elute kit. DNA was then pooled, and 1 μg aliquots were separated on 1.5% agarose neutral (pH 7.0) electrophoretic gels. Ethidium bromide and ultraviolet (UV) light were used to visualize apoptotic DNA ladders and higher molecular weight DNA fragments. These gels were run with DNA extracts from three to five retinas, isolated from different rats, whereas the fellow eyes were used for western blotting of retinal proteins.

Histology

Eyes were enucleated from rats treated with intense light, and unexposed rats, 2 weeks after light exposure and then fixed in 50% Karnovsky’s solution for 24 h. Before fixation, each eye was marked to identify the superior hemisphere and, after 10 min of fixation, the lens was removed. The next day, the eyecups were transferred to 0.1 M sodium cacodylate buffer (pH 7.4) and stored in that solution at 4 °C. For histology, the eyes were embedded in paraffin, sectioned vertically at 4 μm, and then stained with hematoxylin and eosin. Retinal tissue sections in the mid-superior hemisphere were examined.

Western blotting

Retinal homogenates were prepared in PBS (Cat #161-0780, Bio-Rad Inc. Hercules, CA) containing a protease inhibitor cocktail (P8340; Sigma, St. Louis, MO), 1 mM EDTA, 1 mM diethylenetriaminepentaacetic acid (DTPA), and 100 μM butylated hydroxytoluene (BHT; Sigma). The protein contained in three to five retinas from different rats was combined, measured with the Bradford technique (Bio-Rad Laboratories, Hercules, CA), and then extracted with 0.4% sodium dodecyl sulfate (SDS) in PBS (pH 6.8) at 4 °C for 30 min. Protein samples (20 μg/lane) were applied to 4% stacking/12.5% SDS-polyacrylamide running gels and separated in a Mini Protean 3 gel electrophoresis apparatus (Bio-Rad Laboratories). Following electrophoresis, the proteins were transferred to polyvinylidene fluoride (PVDF) membranes at 200 mA overnight, using 10 mM N-cyclohexyl-3-aminopropanesulfonic acid (CAPS) buffer (pH 11) containing 20% methanol. The PVDF membranes were blocked in PBS buffer, containing 2% non-fat dry milk and 0.2% Tween-20 for 1 h at 37 °C, and then incubated with antibodies. Subsequently, the PVDF membranes were stripped and reprobed with other antibodies. Western blot staining was visualized with chemiluminescence and exposure to X-ray film. Relative protein staining intensity was determined with Image J (NIH, Bethesda, MD). A summary of the antibodies used in this study appears as supplemental data in Appendix 1.

Rhodopsin regeneration

To assess the potential effects of antioxidants on the retinoid visual cycle [47], rats were given a single dose of rosemary extract (34 mg/kg, i.p.) 1 h before exposure for 1 h to bleaching light. After the light treatment, the animals were in darkness for various periods before being euthanized in a saturated CO₂ atmosphere, and then both eyes were enucleated for rhodopsin measurement [26]. For each rat (n = 2–5), the rhodopsin values were averaged and the time course of regeneration was determined.

Retinal gene array analysis

Gene expression profiles were established for different groups of animals reared under high- or low-cyclic light conditions while being fed a dietary supplement consisting of rosemary, AREDS plus rosemary, zinc plus rosemary, or vehicle (R, A+R, Z+R, or V, respectively; see Figure 1). Retinas, with the vitreous removed, were excised from four to six rats after 6 weeks of daily dietary supplementation and stored in RNAlater (Qiagen, Inc., Valencia, CA) at −20 °C. Three independent groups of animals were used for each treatment condition to provide triplicate gene expression profiles. The fellow retinas were used for gel electrophoresis. RNA was extracted, using the E.Z.N.A. tissue RNA kit by Omega Bio-Tek (Norcross, GA), with some modifications to the manufacturer’s protocol. Briefly, the retinas were thawed on ice for approximately 5 min. RNAlater was then aspirated, and 500 μl TRK lysis buffer, supplemented with 50% more beta-mercaptoethanol than called for (30 μl/ml final), was added to each retina to prevent RNAse activity. The tissue was homogenized with a clean rotor-stator homogenizer and the lysate centrifuged at 13,000 ×g for 4 min. Supernatant was removed and then combined with 500 μl of 70% ethanol. Column purification was conducted according to the manufacturer’s instructions; the elution volume was 40 μl. RNA was quantified on a NanoDrop spectrophotometer (ThermoFisher Scientific, Waltham, MA) and an Agilent RNA Nano Bioanalyzer Chip (Agilent, Santa Clara, CA). The mean RNA integrity number (RIN) score was 7.01. One hundred ng total RNA was used as the input to prepare fragmented, single-stranded cDNA using the Ambion WT Expression Kit (ThermoFisher Scientific) according to the manufacturer’s directions. This cDNA was labeled using the GeneChip WT Terminal Labeling Kit from Affymetrix (Santa Clara, CA). The Affymetrix GeneChip Hybridization, Wash, and Stain Kit was used to hybridize against the Affymetrix Rat Gene 1.0 ST microarrays. Fluidics were performed on an Affymetrix Fluidics Station 450 using script FS450_0007. Arrays were scanned on an Affymetrix GeneChip Scanner 3000 7G. For a given treatment, each of the three biologic replicates was used to screen an independent Affymetrix GeneChip.

Gene array data analysis

CEL files generated from the scanned arrays were analyzed first in Affymetrix Expression Console build 1.3.1.187. CEL files were imported into Partek Genomics Suite version 6.6 (Partek Inc., St. Louis, MO) using the RMA import settings. Differentially expressed genes were detected using ANOVA, using the treatments as contrasts (light rearing condition (High and Low), and dietary supplement. A second analysis of the gene array data was performed using R version 3.0.1 using the BioConductor statistical analysis packages. The oligo package [48] was used for import, quality assessment, probe level modeling, and RMA normalization. Limma was used to produce linear models for identifying differential genes [49], and the annotate package was used to link gene names and symbols to microarray probe IDs. Results from both analyses were combined, and annotations were manually inspected to reconcile missing gene names where possible.

Statistical analysis

Differentially expressed genes for each condition were filtered to remove any genes that did not have a Benjamini-Hochberg false discovery rate (FDR)–adjusted p value of less than 0.05 (noted as p<0.05 hereafter) [50]. Even small differences in gene expression levels can be statistically significant if the difference is reproducible; however, the smaller the difference, the more difficult it is to differentiate it from background noise. Gene markers showing differential expression levels of greater than 1.3 fold-change or less than [–1.3] with p<0.05 were used for general bioinformatics and Gene Ontology analyses in Ingenuity Pathway Analysis (IPA, Qiagen, Redwood City, CA), or as otherwise noted in the text. For any pathway analysis or functional enrichment analysis, p values were established using a Fisher’s exact test. A standard goodness of fit test was used to determine whether a statistically significant variation existed between any two variables. Biochemical data are presented as the mean ± standard deviation (SD), or standard error of the mean (SEM), for four to eight retinas from different rats. Using a two-tailed t test, a p value of less than 0.05 was considered to represent a statistically significant difference.

Antioxidant efficacy

To compare protective efficacy during photooxidative retinal damage, one sub-group of rats was fed an AREDS1 antioxidant and mineral mixture, with or without a rosemary supplement, and another sub-group was given zinc oxide plus rosemary. After 6 weeks of daily supplements, the rats were dark-adapted and then exposed to intense white light for 6 h. Other animals received the daily AREDS1 recommended dose, with or without rosemary, in a single 1.p. injection 1 h before light treatment. All rats were reared in the 20 lux cyclic light environment and allowed to recover for 2 weeks in darkness after exposure to intense light. Then, retinal DNA levels were determined to assess the extent of visual cell survival. Figure 2 shows that feeding rosemary alone, zinc oxide alone, or AREDS alone (fed or injected) did not prevent photooxidative damage. Protective efficacy for these treatments was less than 10%. Rosemary extract (17 mg/kg) added to the AREDS mixture or combined with zinc oxide (1.4 mg/kg) increased protective efficacy. For the fed and i.p. injected animals, efficacy was about 20%. AREDS supplemented with additional rosemary led to even greater visual cell survival. Rats fed or injected i.p. with AREDS plus 34 mg/kg rosemary exhibited 35–40% protective efficacy, indicating that bioavailability was nearly identical. Approximately 55% efficacy was found when the AREDS antioxidant and mineral mixture was given i.p. at five times the usual daily dose or when rosemary (34 mg/kg) was combined with zinc oxide at the dose normally present in a typical AREDS mixture (Table 1).

Western blotting of retinal proteins

As shown in Figure 3, chronic high light rearing conditions had a dramatic effect on cone cell opsin. Rats kept for 6 weeks in the 200 lux cyclic light environment had approximately 50% of the cone opsin staining found in rats from the 20 lux environment (Figure 3A). The medium wavelength cone arrestin (mCAR) was also dramatically lower in the animals supplemented with AREDS and kept in 200 lux cyclic light. Antioxidant feeding did not appreciably alter the relative staining of the more abundant rhodopsin, or of rod cell S-antigen (S-ag). To assess the effects of chronic light rearing conditions on protein oxidation, CEP immunoreactivity was determined. A higher level of CEP-protein staining was seen in the retinal extracts from 200 lux–reared rats than for those from the 20 lux environment. Supplementation of rats with either AREDS complete or with AREDS antioxidants (see Group 1, in Table 1) did not lead to a decrease in the CEP-protein adduct levels. However, CEP adduct staining was decreased by supplemental zinc oxide in animals from the high light environment.

A second set of Group 1 animals was exposed to intense light for 6 h followed by recovery for 2 weeks in darkness. Figure 3B shows that cone opsin staining was undetectable in rats previously reared in the 20 lux light environment, whereas those from the 200 lux condition recovered substantial amounts. Compared with vehicle or zinc treatment, AREDS complete and the AREDS antioxidants fraction enhanced the level of cone opsin staining in rats from the 200 lux environment. mCAR levels were similar in rats reared in 20 lux or 200 lux light, while the 200 lux zinc-treated animals had slightly lower mCAR levels. Rhodopsin staining was greater for the rats reared in 200 lux light, including the monomer and higher molecular weight forms, whereas rhodopsin staining was appreciably lower in those reared in 20 lux light. Lower molecular weight degradation fragments of rhodopsin were also present. Likewise, retinal S-ag levels were slightly higher in rats reared in high light than in those reared in low light. Photoreceptor cell DNA was lower in rats originally from the 20 lux light environment (Figure 3C). The vehicle-treated rats recovered 55 μg DNA/retina compared with an average of 122 μg for those reared under the high light condition (p<0.05). Antioxidant-fed rats from the high light condition also had significantly higher retinal DNA than found in rats from the 20 lux environment, but there were no statistically significant differences among the various antioxidants. When compared to the DNA values in rats reared in low light, a modest decrease in photoreceptor cell DNA (10%) was found for the unexposed controls reared in higher intensity cyclic light.

The same high and low environmental light rearing effects were found for animals fed supplemental rosemary, AREDS plus rosemary, or zinc oxide plus rosemary (Group 2, Table 1; Figure 3D). Cone opsin staining was nearly undetectable in rats maintained in the high light condition. mCAR levels were lower in the rosemary-fed rats but appeared to be largely unaffected in rats fed rosemary combined with AREDS or with zinc oxide. Retinal rhodopsin and S-ag levels also appeared to be unaffected by the high light rearing environment. Major changes in retinal proteins occurred after intense photooxidative light treatment of rats reared in 20 lux cyclic light (Figure 3E). Cone opsin was undetectable in all animals, whereas mCAR staining was present in rats fed AREDS plus rosemary and in those fed zinc plus rosemary. The levels of rhodopsin (monomer) and retinal S-ag were also much lower. In addition, lower molecular weight rhodopsin fragments were present, indicating extensive rod cell damage, although these were somewhat reduced by the zinc plus rosemary treatment.

Short-term supplementation

To determine whether carnosic acid and ursolic acid, two major rosemary antioxidants [26], were effective, we fed AREDS alone or rosemary alone (34 mg/kg) to rats for 1 week. Other animals received AREDS plus rosemary, or AREDS plus carnosic acid at the concentration found in 34 mg of rosemary (7.9 mg/kg) or AREDS plus 5.3 mg/kg ursolic acid (Group 3, Table 1). The rats were dark adapted overnight and then treated with acute intense white light followed by 2 days in darkness. Previously, all rats had been in the 20 lux cyclic light rearing environment. Western blot analysis revealed that rosemary alone and AREDS alone were ineffective. Cone opsin staining was approximately 50% of that in unexposed control animals (Figure 4A). Treatment with AREDS plus rosemary or with AREDS plus carnosic acid resulted in cone opsin levels that were 75–80% of the level present in unexposed rats, while AREDS plus ursolic acid provided only modest protection. AREDS plus carnosic acid or ursolic acid also led to a slightly reduced level of Rac 1, a small GTPase induced by intense light [43,44] (Appendix 2).

Another sub-set of Group 3 rats received AREDS alone or AREDS plus a higher level of carnosic acid (15.8 mg /kg) for 1 week. Following exposure to intense light, one-half of these animals recovered in darkness for 2 days, while the other half remained in darkness for 2 weeks. As shown in Figure 4B, AREDS plus the higher concentration of carnosic acid effectively reduced the loss of cone cell opsin. Two days after intense light treatment, cone opsin staining was 80% of the value in unexposed rats, whereas the rats treated with vehicle or with AREDS had approximately 60% of the opsin staining present in unexposed controls. Retinal mCAR staining was less impacted 2 days after light treatment. Figure 4B also shows that light-induced visual cell damage continued to progress during the dark recovery period. Compared with the level of cone opsin staining after 2 days, immunoreactivity was much lower 12 days later. Vehicle treatment or AREDS alone led to cone opsin levels that were approximately 25% of those in the unexposed control rats, and mCAR staining that was approximately 35% of control. However, rats treated with the AREDS and carnosic acid combination retained almost 65% of both proteins. Rod photoreceptors exhibited the same relative effects following photooxidative damage. Total rhodopsin staining was similar in all samples 2 days after light exposure, while S-ag staining was higher in the rats treated with AREDS plus carnosic acid than in rats treated with vehicle or with AREDS alone. The loss of rhodopsin was especially noticeable 2 weeks after intense light treatment. In the animals treated with vehicle or with AREDS, the higher molecular weight polymeric forms of rhodopsin were markedly decreased, whereas rats given AREDS plus carnosic acid exhibited higher levels of rhodopsin staining. Retinal S-antigen staining was also higher for rats treated with AREDS plus carnosic acid than for those given either vehicle or AREDS alone.

We determined retinal DNA damage by using neutral pH, agarose gel electrophoresis 2 days after intense light treatment, and then measured residual DNA 2 weeks later (Figure 4C). A typical low molecular weight apoptotic DNA ladder was present in rats treated with vehicle or with AREDS, while the DNA fragmentation pattern for the AREDS plus carnosic acid animals was less intense. As shown by the diminished staining for high molecular weight DNA, single strand breaks also occurred in all light-exposed rat retinas, but DNA loading was uniform in all samples (Appendix 3). Two weeks after exposure to intense light, visual cell DNA in rats fed vehicle or AREDS was approximately 50 μg per retina, while rats treated with AREDS plus carnosic acid retained an average of 100 μg DNA/retina, a value that is significantly higher (p<0.02) than for vehicle treatment. Staining for heme-oxygenase 1 (HO-1), a protein marker of oxidative stress, was absent in non-light-exposed rat retinas and in those from light-treated animals 2 weeks after exposure (Appendix 4). Two days after photooxidative damage, HO-1 staining was elevated in all light-exposed rat retinas, but the level was lower in AREDS plus carnosic acid tissues. To confirm the protective effect of AREDS plus carnosic acid, we performed retinal histology 2 weeks after intense light exposure. As shown in Figure 4D, only one to three rows of photoreceptor cell nuclei were present in the outer nuclei layer (ONL) of rats fed vehicle or AREDS. However, the AREDS plus carnosic acid (A+C) treated rat retina retained seven to eight rows of visual cell nuclei, resulting in an ONL nearly as thick as that in the non-light-treated control rat retina.

Rhodopsin regeneration

Because antioxidant treatment may interfere with the vitamin A visual cycle [47], we measured rhodopsin recovery at various times after a 1 h exposure to bleaching light. Regeneration of rhodopsin in the vehicle- and rosemary-treated rats was nearly identical and more than 75% of control following 90 min of darkness (Appendix 5).

Retinal gene array analysis

Pretreatment of animals with antioxidants, such as ascorbic acid, dimethylthiourea (DMTU), and others, ameliorates the effects of intense light treatment [31], presumably by reducing oxidative stress during the process of photooxidative visual cell death. In this study, we demonstrated that rosemary dietary supplementation also leads to amelioration of light-induced retinal degeneration. We examined the gene profiles of antioxidant- and vehicle-treated rats to determine whether rosemary-based antioxidant treatments also condition or alter the expression of retinal genes. By examining a high versus a low cyclic light rearing condition (which in itself does not compromise the retina significantly), we were able to examine early stages in retinal light damage and to determine whether environmental light preconditioning affects the retina’s susceptibility to subsequent intense damaging light. Gene expression profiles were established for different groups of animals reared in high (H) or low (L) cyclic light and fed one of the antioxidant dietary supplements (rosemary (R), zinc plus rosemary (Z+R), AREDS plus rosemary (A+R)), or the vehicle control (V) as described in the Methods section.

Differential analysis of the retinal gene profiles were performed in three sets of comparisons (Figure 5). MA (log ratio and mean average) plots/volcano plots for each set of gene array comparisons are shown in Appendix 6. Comparisons of antioxidant (AO) dietary treatment groups with the vehicle-treated (V) animals, all reared under low cyclic light (LAO/LV), identified one differentially expressed gene (Figure 5A). The A+R supplement comparison identified Oprk1, opioid receptor, Kappa 1, as a differentially expressed gene (FC = –1.48, p<0.05). Comparisons of the antioxidant treatment groups with the vehicle-treated animals, all high cyclic light-reared (HAO/HV), identified 143 differentially expressed gene markers (Figure 5B). R treatment lead to the identification of eight differentially expressed genes (three R specific), Z+R to 80 differentially expressed genes (20 Z+R specific), and A+R to 120 genes (60 A+R specific). For five of these gene markers (Egr1, Gng11, RGD1564999, RGD1564999, and Scn7a), all three antioxidant treatments resulted in higher relative expression levels than vehicle treatment alone (|FC|>1.3, p value<0.05). Differential comparisons of gene profiles derived from animals fed different rosemary-based antioxidants demonstrated that these treatments have an effect on retinal gene expression. Animals reared under high cyclic light conditions are more sensitive to antioxidant supplement associated changes in gene expression than animals raised in low cyclic light.

Comparisons between each treatment group (T) reared in high cyclic light with its respective dietary group reared in low cyclic light (HT/LT) provides a broader perspective to help define the effects of light rearing conditions and dietary supplementation (Appendix 7, Appendix 8, Appendix 9, Appendix 10 and Appendix 11). We established that the expression level of a given gene (with the exception of Oprk1) is such that gene levels at LV = LR = LZ + R = LA + R. Augmentation of differential status (H/L) for most genes by a given treatment is therefore co-mediated by a high light rearing environment. For any differential gene expression comparison, three trends are possible: The fold change goes up (U), the fold change goes down (D), or the fold change is not statistically significant (N). In total, we found 352 gene markers, defining 336 differentially expressed genes (U or D) across the four HT/LT analyses.

The count of differential gene markers decreased from (HV/LV > HR/LR > (HZ+R/LZ+R) > (HA+R/LA+R); Figure 5C). For a given gene, its expression trend status across all four H/L comparisons (V, R, Z+R, and A+R) can be collated to generate an expression phenotype. Twenty-two different-differential expression phenotypes were thus derived (Appendix 7, Appendix 8, Appendix 9, Appendix 10, and Appendix 11). The most abundant differential expression phenotype observed was UNNN (91 gene markers, Appendix 10), indicating that the primary response to rosemary-based dietary treatments is suppression of the inherent retinal response to shifting from a low cyclic light environment to a high cyclic light rearing environment. Next, we performed a trend analysis to determine the degree of similarity between each antioxidant treatment in high light versus vehicle (H/L) with HV/LV (Figure 6). The transition of rats from low cyclic light to a high cyclic environment light resulted in the differential expression of 280 gene markers; R, Z+R, and A+R treatment suppressed 131, 241, and 266 of these gene responses (respectively) in rats reared in high light versus low light. Although all three antioxidant (H/L) scenarios were significantly different from (HV/LV), the HV/LV differential gene profile had a higher degree match (similarity) with HR/LR than with (HZ+R/LZ+R) or (HA+R/ LA+R).

The expression data were partitioned into vehicle effects (Figure 7; A ring, HV/LV) and antioxidant effects (Figure 7; B ring, HAO/LAO). From an overlapping comparison, three classes of differential genes become evident (Figure 7). 1) Nine gene markers with a similar fold change expression trend in all four analyses define class 1 genes (Appendix 7) corresponding to high light effects that could not be inhibited by antioxidant treatment. 2) 271 HV/LV differentially expressed gene markers, where the differential status is inhibited by at least one antioxidant treatment, corresponding to class 2 genes (light effects that can be modified by antioxidant treatment (Appendix 8, Appendix 9, and Appendix 10). 3) 72 class 3 genes, characterized by a HAO/LAO differential status in at least one of the antioxidant treatments, but not for HV/LV, thus defining purely antioxidant effects in the high light rearing environment (Appendix 11).

Cell function trends

Two independent cell function analyses confirmed that the sets of differential genes from the four high light/low light (H/L) comparisons were distinct, implying that each treatment alters the retinal transcriptome toward a different tissue microenvironment (Figure 8). A focused analysis was then performed to determine whether specific categories of genes, defining specific cell responses, were enriched by the selection process and whether this selection was the same or different depending on the dietary supplement. The transition to a high- from a low-cyclic light rearing environment resulted in a marked enrichment in the expression of inflammation, apoptosis, cytokine, innate immune response, and receptor-related genes, a much more modest enrichment of growth factor and cell stress–related genes, and no enrichment of transcription-related genes (Figure 9A). Comparatively, rosemary treatment selectively decreased the enrichment for apoptosis, receptor, and growth factor–related genes while increasing the level of enrichment for stress response–related genes, relative to vehicle-treated animals (HV/LV). Z+R and A+R supplementation lead to decreased levels of all gene subgroups, except transcription factors, relative to vehicle treatment. Moreover, apoptosis- and receptor-related genes were highly enriched in HV/LV, but markedly depleted by all three antioxidant treatments.

Retina meta-analysis

The expression of several known retinal cell expressed genes provided insight into the state of the retina under the different rearing conditions and antioxidant dietary regimens (Appendix 7, Appendix 8, Appendix 9, Appendix 10 and Appendix 11, Figure 9B). A direct analysis of the differential data (HT/LT) for these associated gene markers demonstrated that rosemary supplementation results in a conservation of three of the five “retinal-specific” gene responses observed with vehicle treatment (Figure 9B). In contrast, the retinal expression response induced by comparing HV/LV was greatly inhibited by Z+R (four of five gene markers) or by A+R supplementation (five of five gene markers). Opn1mw levels for HV/LV and HR/LR were significantly decreased in that fold change (FC) was less than –1.3 and p<0.05. In the case of HZ+R/LZ+R and HA+R/LA+R, fold changes did not go below −1.3 (the cut-off threshold) although the p values were less than 0.05 (and therefore deemed not statistically significant). In addition, thyroid hormone receptor beta 2 (a nuclear hormone receptor, Thrb) is known to regulate the topography of green opsin (encoded by Opn1mw) and UV/blue opsin (Opn1sw) in the retina [51,52]. Thrb gene expression levels were decreased in the vehicle-treated rats (HV/LV) by 1.17 fold (p<0.05). However, the levels were consistent for all dietary treatments (with an average FC = –1.12 ± 0.04, for 12 independent retina samples on 12 independent arrays) and in each instance, p<0.05, making the Thrb gene a putative class 1 gene. Arrestin (Arr3) codes for a molecule functioning in the deactivation of G protein-coupled receptors involved in color vision (171107). Arr3 expression levels were decreased 1.45 fold for HV/LV but did not show statistically significant levels of variation for H/L for any of the antioxidant treatments. A similar trend was found for Opn4, the visual pigment of phototransducing retinal ganglion cells involved in setting the circadian clock (192223). Moreover, glial fibrillary acidic protein message (Gfap), a Müller cell stress marker and a retinal response to wounding, was elevated (2.27 ± 0.68) under the high light condition for animals treated with V, R, and Z+R, but showed no statistically significant difference in H/L levels for A+R.

Opn1mw and Opn4 are G-protein coupled receptors (GPCRs), a large and diverse family of proteins whose primary function is to transduce extracellular stimuli into intracellular signals [53]. High cyclic light rearing, versus low cyclic light, (H/L) caused significant changes in the expression levels of eight additional GPCR coding genes (Rxfp1, Ednrb, P2ry12, Ghrhr, Cx3cr1, Emr1, Gpr34, and S1pr3; highlighted in blue in Appendix 8 and Appendix 9 and in Figure 9B). All ten are class 2 genes; six did not show significant expression changes when treated with R, nine did not show statistically significant changes in levels when treated with Z+R, and all ten lacked differential status when treated with A+R. In addition, the expression levels of a different set of six GPCR genes (Aplnr, Olr1425, Olr1565, Olr1566, Oprk1, and Vom2r65; highlighted in yellow in Appendix 11, and Figure 9B) were unresponsive to the change from the low to high light rearing environment when the animals were fed vehicle but showed specific changes when the rats were fed at least one of the antioxidants. Thus, specific dietary treatments can, for some genes, suppress the inherent retinal transcriptome response to high light rearing (compared to low light rearing) and in other cases mediate antioxidant-driven changes in gene expression.

Dietary HA+R/LA+R (but not the other dietary treatments) resulted in the differential expression of Olr1425, Vom2r65, and Oprk1. Oprk1 is a GPCR that plays a role in the perception of pain and addiction. It binds endogenous endorphins and various synthetic opioids that result in signaling, via guanine nucleotide-binding proteins (G proteins), and inhibition of adenylate cyclase activity (P41145). G protein subunit gamma 11, Gng11, was observed to be elevated (2.43 fold, p<0.05; Appendix 7) in high light versus low light rearing conditions when animals were fed A+R, the dietary treatment that resulted in the highest degree of protection from photooxidative damage (Figure 2). Oprk1 and Gng11 are two components of several known canonical pathways, including Gas signaling, Gai signaling, and interleukin-8 (IL-8) signaling ( IPA Database, Ingenuity® Systems). Gng11 promotes senescence in human fibroblasts [41]. It activates ERK1/2 but not Ras, and its expression can be induced by exposure to H₂O₂ [41]. An analysis of the proximal promoter for Gng11 revealed three separate binding sites for the transcription factor Pax4 and one for the related protein Pax6 (Table 2). We found that Pax4 was downregulated about 1.3 fold (p<0.05) by vehicle and rosemary treatment (H/L). No statistically significant variation occurred for the H/L reared animals treated with Z+R or A+R, implying that these two dietary supplements prevented the high light mediated downregulation of Pax4.