ω-3 and ω-6 long-chain PUFAs and their enzymatic metabolites in neovascular eye diseases

Abstract

Neovascular eye diseases, including retinopathy of prematurity, diabetic retinopathy, and age-related macular degeneration, threaten the visual health of children and adults. Current treatment options, including anti–vascular endothelial growth factor therapy and laser retinal photocoagulation, have limitations and are associated with adverse effects; therefore, the identification of additional therapies is highly desirable. Both clinical and experimental studies show that dietary ω-3 (n–3) long-chain polyunsaturated fatty acids (LC-PUFAs) reduce retinal and choroidal angiogenesis. The ω-3 LC-PUFA metabolites from 2 groups of enzymes, cyclooxygenases and lipoxygenases, inhibit [and the ω-6 (n–6) LC-PUFA metabolites promote] inflammation and angiogenesis. However, both of the ω-3 and the ω-6 lipid products of cytochrome P450 oxidase 2C promote neovascularization in both the retina and choroid, which suggests that inhibition of this pathway might be beneficial. This review summarizes our current understanding of the roles of ω-3 and ω-6 LC-PUFAs and their enzymatic metabolites in neovascular eye diseases.

INTRODUCTION

Neovascular eye diseases, such as retinopathy of prematurity (ROP) in premature infants, diabetic retinopathy (DR) in working-age adults, and age-related macular degeneration (AMD) in the elderly, are the most common causes of vision loss in developed countries (1, 2). Decades of research have led to a medical intervention to suppress neovascularization with the development of anti–vascular endothelial growth factor (anti-VEGF) therapy. However, blocking VEGF does not cure the underlying causes of VEGF production and may worsen retinal neural degeneration and suppress normal vascular development (3, 4). Other problems associated with anti-VEGF treatment include lack of response (5) and frequent intraocular injections, each of which carries the risk of endophthalmitis and suppression of VEGF systemically as intravitreally injected drug leaks into the circulation (6, 7). Therefore, it is important to understand the role of other factors that modulate neovascularization to develop less invasive and more effective therapies.

Lipids are important for cell signaling and metabolism as well as for membrane structure and energy storage (8). Specifically, lipids and their metabolites are potent regulators of neovascular eye diseases (9, 10). Two major classes of lipids that have been studied in this context are ω-3 and ω-6 long-chain PUFAs (LC-PUFAs) (11, 12). Both can be synthesized from linolenic acid through a multistage process requiring successive actions by elongase and desaturase enzymes. In humans, due to the lack of the necessary fatty acid desaturase (FADS) converting oleic acid into linolenic acid, LC-PUFAs are essential fatty acids that cannot be endogenously synthesized in adequate amounts and must therefore be obtained from the diet (13). Structurally, ω-3 and ω-6 LC-PUFAs differ solely in the location of their double bonds: ω-3 LC-PUFAs, such as DHA and EPA, have the first double bond after the third carbon atom, and ω-6 LC-PUFAs, such as arachidonic acid (AA), have the first double bond after the sixth carbon atom (Figure 1). EPA is also a precursor to DHA. ω-3 And ω-6 LC-PUFAs competitively bind desaturases and elongases for the addition of carbons or double bonds in their chains, respectively. Different ratios of ω-3 and ω-6 LC-PUFA intake may result in different inflammatory and angiogenic responses, which can be used as clinical indicators in various neovascular eye diseases (14).

FIGURE 1.

Chemical structures of AA, DHA, and EPA. In ω-6 LC-PUFAs (AA), the first double bond is located after the sixth-carbon atom from the ω end, whereas the third-carbon atom is followed by the first double bond in ω-3 LC-PUFAs (DHA and EPA). AA, arachidonic acid; LC-PUFA, long-chain PUFAs.

LC-PUFAs are essential for normal growth and neurodevelopment in the brain and eyes (15, 16). The primary ω-3 LC-PUFA in the retina, DHA, makes up ∼20% of retinal weight as well as ∼10–15% of tissue weight in sperm and brain cortex (17). The major ω-6 LC-PUFA, AA, comprises ∼10% of retinal weight. The membrane lipid composition rapidly changes with changes in ω-LC-PUFA supply (18). DHA influences both neuronal and retinal vascular cell survival and development (19). An increased intake of ω-3 LC-PUFAs is associated with a reduced risk of pathological ocular angiogenesis (12, 20, 21). Therefore, the manipulation of ω-3 and ω-6 LC-PUFAs and their metabolites may be clinically relevant in preventing and treating retinopathy and neovascular AMD. This review provides an overview of our current knowledge of the contribution of ω-3 and ω-6 LC-PUFAs and their enzymatic metabolites to both the pathogenesis and potential treatments of neovascular eye diseases, including ROP, DR, and AMD.

CORRELATIONS BETWEEN ω LC-PUFAs AND NEOVASCULAR EYE DISEASES

ROP

ROP is a proliferative retinopathy in premature infants with a first phase of cessation of retinal vessel growth or vessel loss followed by a compensatory but ultimately pathological second phase of hypoxia-driven vessel proliferation (Figure 2) (22). It is a leading cause of blindness in children worldwide because ∼10% of births are preterm. Severe ROP is reported in ∼10% of infants born with a gestational age of <29 wk (23). The number of very premature infants who survive is increasing in developing countries, as is the incidence of associated complications of preterm birth such as ROP. The higher oxygen supplementation guidelines (24) suggested by the SUPPORT (Surfactant, Positive Pressure, and Oxygenation Randomized Trial), which indicates slightly better survival with higher oxygen saturations but a doubling of ROP, have been adopted by most neonatal intensive care units, which is likely to lead to an increase in ROP.

FIGURE 2.

Progression of retinopathy of prematurity. Oxygen tension is low, and vascular growth is normal in utero. After premature birth, retinal vessel growth stops due to hyperoxia and loss of the nutrients, including ω-3 LCPUFAs. As the retina matures and metabolic demand increases, hypoxia results. The hypoxic retina stimulates expression of proangiogenic factors, which stimulate retinal neovascularization. LCPUFA, long-chain PUFAs. Reproduced from reference 22 with permission.

Factors other than oxygen also influence retinopathy. After premature birth, the loss of both ω-3 and ω-6 LC-PUFAs that are normally provided by the intrauterine environment in the third trimester of pregnancy contributes to the initiation and progression of ROP (25, 26). Because LC-PUFAs are not synthesized in adequate quantities, enteral or parenteral sources are required. Preterm infants have a DHA deficit of ≤44% shortly after birth due to inadequate enteral or parenteral sources (27). The total parenteral nutrition for the very premature infants who are unable to tolerate oral feeding is usually devoid of DHA and often of AA. In premature infants, serum DHA concentrations decrease by 1 wk after birth and remain low for ≥4 wk (28, 29) due to an inadequate postnatal nutritional supply but also due to inadequate fat stores and ineffective conversion from precursor fatty acids, such as α-linolenic acid (30). The FADS single nucleotide polymorphism (SNP) rs174602 modifies the effects of prenatal supplementation with DHA on birth weight, which is an important risk factor for retinopathy in preterm infants (31).

Premature infants who receive fish-oil fat emulsion supplementation show less severe retinopathy (32–34). In the mouse oxygen-induced retinopathy model, dietary intakes of ω-3 compared with ω-6 LC-PUFAs promote normal revascularization of avascular retina to reduce the area of nonvascularized retina that stimulates neovascularization and independently directly suppresses pathological retinal neovascularization by ∼50% (12, 13). The same phenotype was also observed in Fat-1 transgenic mice that express a Caenorhabditis elegans gene, which encodes an enzyme to convert ω-6 to ω-3 LC-PUFAs, resulting in a high ω-3–to–ω-6 ratio (12). The protective effects of ω-3 LC-PUFAs and their bioactive metabolites are mediated in large part through the activation of peroxisome proliferator–activated receptor (PPAR) γ (13) as well as through increased serum concentrations of bioactive high-molecular-weight adiponectin, which is produced and assembled by white adipose tissue (12, 25). The suppression of TNF-α secreted by microglia and macrophages in the retina may also play a part (13).

DR

DR is a major complication of diabetes that afflicts ≥93 million people worldwide, and the incidence is increasing rapidly with the global increases in obesity and type 2 diabetes (35). Clinically, DR can be classified into nonproliferative DR and proliferative DR. Diabetes is characterized by hyperglycemia and dyslipidemia, which are associated with multiple cellular alterations (36), endothelial cytotoxicity (37, 38), dysregulation of FADS and elongases (39, 40), and retinal vascular abnormalities and loss. When the vessel loss is great enough, the retina’s lack of nutrition and oxygen stimulates the release of growth factors, which induces proliferative angiogenesis (41). Dyslipidemia correlates with the progression of human DR (10). The Early Treatment Diabetic Retinopathy Study showed that higher concentrations of serum lipids are associated with an increased risk of the development of hard exudates in the macula and visual loss (42). Increased serum concentrations of apolipoprotein B (associated with LDL) represent a significant metabolic risk factor for proliferative DR (43). Serum apolipoprotein B concentrations correlate with systematic proinflammatory status, which may contribute to the presence and severity of DR in diabetic patients (44, 45).

Dietary ω-3 LC-PUFAs influence DR. In the PREDIMED (Prevención con Dieta Mediterránea) trial, a 6-y follow-up analysis in middle-aged and older individuals with type 2 diabetes adhering to a “healthy” Mediterranean diet showed that the subset of patients whose self-reported diet included ω-3 LC-PUFAs (≥500 mg EPA + DHA/d or ≥2 weekly servings of oily fish) at baseline show a 48% decreased incidence of sight-threatening DR (46). In murine models of diabetes, fish oil and a ω-3 LC-PUFA–enriched diet preserved retinal neuronal function (Figure 3) (47, 48). A genomewide association study showed that SNPs near FADS correlated with the plasma concentration of EPA and AA (49). Microarray analysis of insulin-resistant and insulin-sensitive individuals showed that FADS were differentially regulated in both adipose and muscle of insulin-resistant individuals, possibly through an association between neighboring SNPs and fasting glucose homeostasis (50, 51).

FIGURE 3.

ω-3 LC-PUFAs preserve retinal function in db/db mice. (A) Type 2 diabetic db/db mice fed defined rodent feed pellets with 10% (wt:wt) safflower oil containing either 2% ω-3 LC-PUFAs (1% DHA and 1% EPA) or 2% ω-6 LC-PUFAs (AA). (B) Representative electroretinographic wave from mice fed ω-3 or ω-6 LC-PUFAs. (C) Saturating amplitude of responses originating in photoreceptors and bipolar cells from db/db mice fed a ω-3 and ω-6 LC-PUFA–enriched diet. Light stimulation triggered the hyperpolarizaton of the photoreceptors (“a” wave), resulting in a decreased release of neurotransmitter, which subsequently led to the depolarization of bipolar cells (“b” wave). The amplitude of the “a” wave was from the baseline to the negative trough of the “a” wave; the amplitude of the “b” wave was from the trough of a wave to the peak of the “b” wave. The data were normalized by the values in age-matched heterozygotes fed normal feed pellets (gray shading). A significant attenuation was observed in the ω-6 LC-PUFA group. AA, arachidonic acid; LC-PUFA, long-chain PUFAs. Reproduced from reference 47 with permission.

In the diabetic rat retina, decreased incorporation of very-long-chain PUFAs (VLC-PUFAs) secondary to increased expression of elongases resulted in increased inflammation (52). Inflammation is strongly associated with the pathogenesis of clinical DR. High vitreous expressions of TNF-α, IL-1β, and IL-6 are reported in patients with proliferative DR (53). Although ω-6 LC-PUFAs induce adhesion molecule expression and leukocyte adhesion in human retinal endothelial cells, the ω-3 LC-PUFA DHA inhibits cytokine-induced nuclear transcription factor κB activation and nuclear translocation, as well as adhesion molecule expression (54–56). An increased intake of ω-3 LC-PUFAs decreases cholesterol-induced inflammation in mouse eye and brain, which is associated with neuroprotection (57) as well as protection against hyperglycemia-induced retinal vascular cell damage (58). Lipoxins, resolvins, and protectins derived from ω-3 LC-PUFAs are antiangiogenic and may be clinically useful to prevent diabetic macular edema and retinopathy (59). As noted, the dietary intake of ω-3 LC-PUFAs reduces neovascularization through activation of PPAR-γ (13). The Pro12Ala polymorphism in PPAR-γ is associated with a reduced risk of developing DR in type 2 diabetes (60, 61). PPAR-α shares a high structural homology with PPAR-γ (62), which may also be activated by ω-3 LC-PUFAs. Drugs that activate PPAR-α reduce the need for laser treatment of DR and slow the progression of clinical DR (63–66). Thus, PPAR-α may also mediate the protective effects of ω-3 LC-PUFAs on DR.

AMD

AMD is a leading cause of central vision loss in the population >50 y of age. AMD is classified into non-neovascular (dry or nonexudative) AMD and neovascular (wet or exudative) AMD (67). ω-3 LC-PUFAs are abundant in photoreceptors and are often deficient in many Western diets, leading to investigations of the correlation between dietary ω-3 LC-PUFA intake and the risk of AMD (68).

High dietary intakes of food containing ω-3 LC-PUFAs (100–800 mg DHA + 30–60 mg EPA/d) are associated with a lower prevalence and incidence of AMD (69–72). High plasma ω-3 LC-PUFA concentrations are associated with a 42% decreased incidence of AMD, as described in a large cohort study in 38,022 female health professionals in the United States. Specifically, women in the highest tertile of intake for DHA, compared with those in the lowest, had a multivariate-adjusted RR of AMD of 0.62 (95% CI: 0.44, 0.87). For EPA, women in the highest tertile of intake had an RR of 0.66 (95% CI: 0.48, 0.92). Consistent with the findings for DHA and EPA, women who consumed ≥1 servings of fish/wk, compared with those who consumed <1 serving/mo, had an RR of AMD of 0.58 (95% CI: 0.38, 0.87) (73). High plasma concentrations of ω-3 LC-PUFAs are associated with a decreased risk of late AMD, as observed in a prospective, population-based French study (74). The Blue Mountains Eye Study conducted in elderly Australians indicates that dietary intake of fish rich in ω-3 LC-PUFAs is associated with protection against early AMD (75). The population-based Rotterdam Study found that a high dietary intake of food rich in ω-3 LC-PUFAs is associated with a reduced risk in those at high genetic risk of AMD (76). The US Twin Study of Age-Related Macular Degeneration found that a dietary intake of ≥2 servings fish/wk also reduced the risk of AMD (77).

The Age-Related Eye Disease Study (AREDS) recruited 4757 participants, 55–80 y of age, in 11 US clinical centers with moderate-high risk of progression to advanced AMD. Participants with the highest self-reported intake of foods rich in ω-3 LC-PUFAs (DHA + EPA, 0.11% of total energy intake) were 30% less likely to develop central geographic atrophy and 50% less likely to develop neovascular AMD than their peers with the lowest ω-3 LC-PUFA intake (0.013% of total energy intake) (Table 1) (11). AREDS2 (in participants with a high risk of advanced AMD progression; median age: 74 y) was then designed to prospectively evaluate the impact of ω-3 LC-PUFA supplementation on AMD progression (78). Five years of ω-3 LC-PUFA supplementation (650 mg DHA + 350 mg EPA/d) plus the AREDS formulation did not further reduce the risk of progression to advanced AMD in this group of well-nourished participants (79). It is noteworthy that the baseline serum concentrations of ω-3 LC-PUFAs in the treated group was much higher than in the first retrospective AREDS, likely due to self-supplementation and a higher dietary intake of foods containing ω-3 LC-PUFAs, such as fish. However, even the subset of patients with the lowest quintile of fish intake saw no benefit with ω-3 LC-PUFA supplementation (80). In a Japanese population with a high baseline intake of fish oil, no significant association between variations in high serum ω-3 LC-PUFA concentrations and reduced risk of AMD was found (81). These results suggest that there may be a ω-3 LC-PUFA minimum required for the maintenance of retinal stability and supplementation above that threshold may not increase the benefit.

TABLE 1.

ORs (95% CIs) for the 12-y progression to CGA and neovascular AMD in patients with dietary intakes of ω-3 LC-PUFAs¹

		Models for progression to CGA			Models for progression to neovascular AMD
			Final model			Final model
Nutrient and quintile of intake	Energy intake, % TEI	CGA/no CGA, n/n	OR (95% CI)	P	Neovascular AMD/no neovascular AMD, n/n	OR (95% CI)	P
DHA
1	0.010	78/303	1.00		144/237	1.00
2	0.018	60/290	0.80 (0.56, 1.15)	0.225	111/239	0.89 (0.66, 1.20)	0.431
3	0.026	85/287	1.15 (0.83, 1.59)	0.417	104/268	0.75 (0.55, 1.02)	0.065
4	0.037	87/305	1.04 (0.75, 1.44)	0.832	138/254	0.89 (0.67, 1.19)	0.443
5	0.061	54/288	0.68 (0.47, 0.99)	0.042	86/256	0.66 (0.47, 0.92)	0.014
P-trend			0.099			0.026
EPA
1	0.000	78/286	1.00		129/235	1.00
2	0.009	77/311	0.90 (0.64, 1.25)	0.528	117/271	0.73 (0.54, 0.98)	0.038
3	0.015	81/285	1.04 (0.74, 1.45)	0.817	123/243	0.97 (0.72, 1.31)	0.834
4	0.024	65/306	0.74 (0.52, 1.00)	0.104	114/257	0.76 (0.55, 1.03)	0.079
5	0.044	63/285	0.70 (0.49, 1.00)	0.051	100/248	0.71 (0.51, 0.98)	0.039
P-trend			0.024			0.068
DHA + EPA
1	0.013	83/293	1.00		139/237	1.00
2	0.028	58/302	0.70 (0.49, 1.00)	0.051	110/250	0.79 (0.58, 1.06)	0.111
3	0.042	91/283	1.14 (0.83, 1.56)	0.417	115/259	0.85 (0.63, 1.15)	0.299
4	0.061	74/310	0.78 (0.56, 1.10)	0.157	129/255	0.83 (0.62, 1.13)	0.235
5	0.106	58/285	0.65 (0.45, 0.92)	0.016	90/253	0.68 (0.49, 0.94)	0.020
P-trend			0.028			0.044

¹The final models contained terms for age, sex, TEI, smoking history, AMD status at enrollment, and Age-Related Eye Disease Study clinical trial treatment. ORs (95% CIs) were calculated for ω-3 LC-PUFA intakes from logistic regression models comparing energy-adjusted LC-PUFA intake quintiles (quintile 1 is used as the reference). P values for trend were computed by using median values for intake quintiles. AMD, age-related macular degeneration; CGA, central geographic atrophy; LC-PUFA, long-chain PUFA; TEI, total energy intake. Reproduced from reference 11 with permission.

Inflammation and oxidative stress as well as lipid and carbohydrate metabolism are implicated in the complex process of AMD progression (82–84). ω-3 LC-PUFA supplementation lowers vitreal VEGF-A concentrations in patients with neovascular AMD (85). In a mouse model of laser-induced choroidal neovascularization, an AREDS2 diet (rich in lutein, zeaxanthin, DHA, and EPA) or a ω-3 LC-PUFA replete diet without any ω-6 LC-PUFAs compared with an otherwise matched diet with ω-6 LC-PUFAs without any ω-3 LC-PUFAs reduced choroidal neovascularization and restored retinal function, possibly by lowering the expressions of proinflammatory or angiogenic factors (21, 86, 87).

Overall, in experimental studies, the dietary components can be well controlled to compare the effects of ω-3 and ω-6 LC-PUFAs on retinopathy. In these studies, a diet containing only ω-3 LC-PUFAs (and no ω-6 LC-PUFAs) compared with a diet without any ω-3 LC-PUFAs and only ω-6 LC-PUFAs reduced retinopathy. Clinical investigations clearly show that intakes of fish (containing DHA and EPA as well as other lipids) as well as a diet with other foods containing many other fats reduce the risk of the progression of AMD and DR. Studies also showed that DHA and EPA serum concentrations correlate with less severe AMD. However, studies with DHA supplementation of an uncontrolled baseline diet are less clear (e.g., AREDS2), which suggests that there are other possible beneficial nutrients in fish, in addition to DHA and EPA. There may also be benefits to combinations of lipids that are difficult to study in long-term clinical trials in which baseline diets vary widely. More data are needed to more fully understand if DHA and EPA supplements are beneficial compared with increased fish intake, which more clearly improves AMD and DR.

LC-PUFA METABOLIC PATHWAYS AND PRO- AND ANTIANGIOGENIC METABOLITES

There is rapid turnover of membrane ω-3 and ω-6 LC-PUFAs, which are then metabolized by the same enzymes. Membrane-bound ω-3 and ω-6 LC-PUFAs are both released from the second-carbon group of glycerol by phospholipase A₂, and then further metabolized by 3 major pathways—the cyclooxygenases (COXs), lipoxygenases (LOXs) and cytochrome P450 oxidases (CYPs)—producing many LC-PUFA–derived metabolites with diverse biological functions.

COXs

COXs, also known as prostaglandin-endoperoxide synthases, are a group of enzymes that metabolize ω-3 and ω-6 LC-PUFAs to give rise to important biological mediators, prostanoids, including prostaglandins (PGs), resolvins, and thromboxanes (TXs). The 2 isozymes found in humans, COX-1 and COX-2, have 65% amino acid sequence homology and near-identical catalytic sites (88). COX-2, in particular, is among the early gene products that are increased in response to inflammatory stimuli and retinal ischemia (89, 90). In general, ω-3 LC-PUFA–derived COX-2 metabolites protect against neovascularization compared with ω-6 LC-PUFA–derived COX-2 products (Figure 4). Although ω-3 LC-PUFA–derived PGE₃ inhibits endothelial tubule formation by decreasing the production of angiopoietin 2 and matrix metalloproteinase 9 (91), increased production of ω-6 LC-PUFA–derived PGE₂ via COX-2 may stimulate the formation of pathological retinal neovessels through binding to its PGE receptor (89, 92). TXA₂, another COX-2 ω-6 LC-PUFA metabolite, can lead to time- and concentration-dependent death of retinal endothelial cells (93, 94). The resolvins, biosynthesized from ω-3 LC-PUFAs via COX-2, are locally acting mediators with potent anti-inflammatory properties including suppression of the production of proinflammatory factors (95).

FIGURE 4.

Schematic of the COX and LOX pathways metabolizing ω-6 and ω-3 LC-PUFAs. Liberated ω-6 and ω-3 LC-PUFAs are rapidly metabolized by COX and LOX enzymes and generated the pathway-specific lipid metabolites PGs, TXs, LTs, HETEs, resolvins, HDHAs, and neuroprotectins. The metabolites derived from ω-6 LC-PUFAs are proinflammatory and proangiogenic, whereas those derived from ω-3 LC-PUFAs are anti-inflammatory and antiangiogenic. COX, cyclooxygenase; HDHA, hydroxydocosahexaenoic acid; HETE, hydroxyeicosatetraenoic acid; LC-PUFA, long-chain PUFA; LOX, lipoxygenase; LT, leukotriene; PG, prostaglandin; TX, thromboxane.

Pharmacologic inhibitors of COX-1 and COX-2, such as aspirin and nonsteroidal anti-inflammatory drugs, are among the most frequently used over-the-counter medications. The resulting suppression of PG and TX production reduces inflammation and pain. The inhibition of COXs decreases the synthesis of both ω-3 and ω-6 LC-PUFA–derived metabolites, thus reducing the pool of both anti-inflammatory ω-3 and inflammatory ω-6 LC-PUFA mediators in the eye. Inhibiting COXs does not suppress retinopathy (91).

LOXs

Another major class of enzymes that contribute to LC-PUFA metabolite effects on retinopathy are the LOXs, a family of iron-containing enzymes that catalyze the dioxygenation of ω-3 and ω-6 LC-PUFAs. Mammalian LOXs, such as 5-LOX and 12/15-LOX, regulate chronic inflammation and oxidative stress and influence angiogenesis. In a mouse model of type I diabetes, 5-LOX but not 12/15-LOX regulated the degeneration of retinal capillaries in DR (96). LOX metabolites derived from ω-3 compared with ω-6 LC-PUFAs show different effects on neovascular eye diseases (Figure 4). The ω-6 LC-PUFA–derived LOX product 5-hydroxyeicosatetraenoic acid (5-HETE) is increased in the vitreous of patients with DR, and is involved in mediating both inflammatory and angiogenic processes (97, 98). Leukotriene B₄ (LTB₄), another 5-LOX ω-6 LC-PUFA metabolite, is involved in the pathogenesis of retinal light damage (99). The ω-3 LC-PUFA product of 5-LOX, 4-hydroxydocosahexaenoic acid (4-HDHA), regulates neovascularization in oxygen-induced retinopathy modeling ROP and neovascular DR (100). Parenteral administration of ω-3 LC-PUFA–derived LOX metabolites confers protection from retinal and choroidal neovascularization (101, 102). Neuroprotectin D₁, a metabolite derived from DHA, attenuates choroidal neovascularization by inducing a ramified, non–injury-inducing microglial phenotype (103) and reduces oxidative stress– or ischemia reperfusion–induced apoptosis in retinal pigment epithelial cell cultures (104, 105).

The relative contribution of the COX and LOX pathways to retinopathy was examined experimentally in mice (100). The overall protective effect of ω-3 compared with ω-6 LC-PUFA diets on oxygen-induced retinopathy was unchanged in 12/15-LOX, as well as in COX1 or COX2 knockout mice, which suggests that these enzymes did not contribute to the major ω-3 LC-PUFA–derived protective metabolites. However, in 5-LOX knockout mice, the overall protective effect of ω-3 LC-PUFAs compared with ω-6 LC-PUFAs was abolished, suggesting that the 5-LOX ω-3 LC-PUFA metabolite 4-HDHA mediates a majority of the general protective effect of ω-3 LC-PUFAs (100). 4-HDHA directly inhibits the sprouting and proliferation of endothelial cells, independently of its anti-inflammatory effects via PPAR-γ (100). The inhibition of 5-LOX may potentially lead to more severe ocular neovascularization, whereas inhibition of COX-1 or COX-2 with commonly used nonsteroidal anti-inflammatory drugs such as aspirin or ibuprofen, both COX inhibitors, will not block the overall protective effect of ω-3 LC-PUFAs on retinopathy.

CYPs

A third major metabolizing enzyme pathway for ω-3 and ω-6 LC-PUFAs includes specific CYPs. CYPs are a large and diverse group of enzymes that catalyze the oxidation of many drugs and organic substances, including lipids. The CYP enzymes CYP2C and CYP2J metabolize LC-PUFAs into bioactive epoxides (106), which are further hydrolyzed by soluble epoxide hydrolase (sEH) to less-active trans-dihydrodiols (107). In the mouse oxygen-induced retinopathy model, there is increased neovascularization in transgenic mice with Tie2-driven expression of human CYP2C8 fed a diet with either ω-3 or ω-6 LC-PUFAs. There is reduced neovascularization in Tie2-driven sEH-expressing mice, which is associated with reduced concentrations of CYP2C ω-3 and ω-6 LC-PUFA metabolites (108).

These studies suggest that CYP2C inhibition might benefit human pathological ocular angiogenesis, because CYP2C metabolites of both ω-6 LC-PUFAs and notably ω-3 LC-PUFAs promote retinopathy. As noted above, the metabolic products of ω-6 LC-PUFAs are generally proangiogenic, whereas ω-3 LC-PUFA metabolites of COX and LOX pathways show an antiangiogenic effect. However, because ω-3 LC-PUFA–derived CYP2C products are proangiogenic, as are ω-6 LC-PUFA–derived CYP2C products, they partially counteract the overall antiangiogenic effects of ω-3 LC-PUFAs. Our recent study showed that the inhibition of CYP2C activity increases the protective effects of dietary ω-3 LC-PUFAs on pathological retinal and choroidal angiogenesis in oxygen-induced retinopathy and laser-induced choroidal neovascularization (109, 110) associated with lower concentrations of CYP2C products derived from ω-3 and ω-6 LC-PUFAs. The addition of ω-3 LC-PUFA metabolites of CYP2C reverses the suppression of angiogenesis ex vivo by CYP2C inhibition. Dietary ω-3 LC-PUFA supplementation along with CYP2C inhibition is likely to benefit retinal neovascularization (Figure 5) (109, 110).

FIGURE 5.

The CYP pathway is a potential pharmaceutical target to treat neovascular eye diseases. The CYP2C metabolites derived from both ω-6 and ω-3 LCPUFAs promote pathological ocular angiogenesis. CYP2C inhibition adds to ω-3 LCPUFA protection against retinal and choroidal neovascularization. AA, arachidonic acid; CYP, cytochrome P450 oxidase; EDP, epoxydocosapentaenoic acid; EET, epoxyeicosatrienoic acid; GCL, ganglion cell layer; INL, inner nuclear layer; LCPUFA, long-chain PUFAs; ONL, outer nuclear layer; RPE, retinal pigment epithelium. Reproduced from references 109 and 110 with permission.

In other models, although ω-6 LC-PUFA–derived CYP2C metabolites, such as epoxyeicosatrienoic acids (EETs), are consistently reported to promote angiogenesis and endothelial cell migration (111, 112), other studies suggest that ω-3 LC-PUFA–derived CYP metabolites, such as DHA-derived epoxydocosapentaenoic acids (EDPs) and EPA-derived epoxyeicosatetraenoic acids (EEQs), suppress tumor angiogenesis and human umbilical vein endothelial cell migration (113). These findings with regard to CYP-mediated ω-3 LC-PUFA metabolites in angiogenesis may suggest cell- or tissue-specific functions.

Fenofibrate, which is known to inhibit DR and diabetic macular edema, may act through CYP2C inhibition as well as PPAR-α activation (110). Fenofibrate reduces the risk of proliferative DR by 35–40%, as noted in 2 intervention trials in >20,000 patients with type 2 diabetes: the FIELD (Fenofibrate Intervention and Event Lowering in Diabetes) and the ACCORD (Action to Control Cardiovascular Disease Risk in Diabetes) studies (63, 65). The mechanism underlying the protective effects of fenofibrate on DR was found to be independent of its PPAR-α–mediated lipid-lowering effect (114, 115). Fenofibrate is a more potent CYP2C antagonist than a PPAR-α agonist (116, 117). Mouse studies of laser-induced choroidal neovascularization and oxygen-induced retinopathy suggest that the suppression of fenofibrate on pathological choroidal and retinal angiogenesis is due to its inhibitory effects on CYP2C and reduction in ω-3 and ω-6 LC-PUFA–derived proangiogenic CYP2C metabolites (110).

ω-3 LC-PUFAs AS ESSENTIAL BUILDING BLOCKS IN PHOTORECEPTORS

Retinal VLC-PUFA is critical for photoreceptor function and survival (118). In human eyes, the VLC-PUFA–containing glycerophotidylcholine is detected in photoreceptors despite the overall low abundance, and DHA-containing glycerophotidylcholine is found solely in photoreceptors (119). In human AMD, VLC-PUFAs and ω-3:ω-6 ratios are significantly decreased (120, 121). Deficits in the synthesis of VLC-PUFAs lead to photoreceptor death and loss of vision in mice (122, 123). Deficiency in Mfs2a, a newly described DHA transporter from blood to the retina, leads to a lack of DHA in the eye and deficits in early postnatal photoreceptor development (124). Supplementation with DHA reduces the rate of progressive loss in rod function and visual field sensitivity in human retinitis pigmentosa (a retinal disease with aggressive rod loss first followed by cones) (125). Therefore, supplementation of ω-3 LC-PUFAs may also benefit the maintenance and protection of photoreceptors. In addition, serum and red blood cell lipids are excellent biomarkers for retinal LC-PUFA and VLC-PUFA series (120, 121), providing therapeutically applicable approaches to monitor the uptake and utilization of DHA in retinas.

ω-3 LC-PUFAs AND CHOLESTEROL-MEDIATED RETINAL ABNORMALITIES

Diets rich in cholesterol cause retinal degeneration in rabbits (relevant to AMD) (126). The injection of 7-ketocholesterol (the oxysterol formed by the autoxidation of cholesterol and cholesterol esters) into the anterior chamber of the rat eye induces ocular inflammation and angiogenesis in vivo (127). The modulation of cholesterol homeostasis protects against intravitreal N-methyl-d-aspartate–induced inner retinal damage and restores retinal pigment epithelial autophagy in mice (128, 129). Clinically, ω-3 LC-PUFA supplementation may modulate serum concentrations of HDL, LDL, VLDL, and total cholesterol (130, 131), although the impact of ω-3 LC-PUFAs on cholesterol concentrations is still inconclusive (132). Supplementation with the ω-3 LC-PUFA EPA reduces the incidence of coronary artery disease, particularly in patients with low HDL cholesterol (133). Therefore, there may be potential beneficial effects of ω-3 LC-PUFA supplementation on cholesterol-associated eye disorders. Further explorations are warranted.

CONCLUSIONS

Data from both clinical investigations and experimental studies suggest an overall protective effect of ω-3 LC-PUFAs on ocular inflammation and vascular eye diseases, although ω-3 LC-PUFA CYP2C metabolites promote retinopathy. Effective lipid-based treatments will likely depend on an adequate dietary intake of food or supplements containing ω-3 LC-PUFAs and selective use of beneficial lipid metabolites and/or inhibitors of enzymes producing proangiogenic metabolites from both ω-3 and ω-6 LC-PUFAs such as CYP2C. A thorough understanding of retinal lipid metabolism is needed to form the foundation of any lipid-based intervention. The promising findings obtained so far encourage more investigations in this field, because lipid-based therapies have obvious clinical potential, particularly in populations with an LC-PUFA–deficient diet.