Tissues with high metabolic rates often use lipid as well as glucose for energy, conferring a survival advantage during feast and famine.1 Current dogma suggests that high-energy consuming photoreceptors depend on glucose.2,3 Here we show that retina also uses fatty acids (FA) β-oxidation for energy. Moreover, we identify a lipid sensor Ffar1 that curbs glucose uptake when FA are available. Very low-density lipoprotein receptor (VLDLR), expressed in tissues with a high metabolic rate, facilitates the uptake of triglyceride-derived FA.4,5 Vldlr is present in photoreceptors.6 In Vldlr−/− retinas, Ffar1, sensing high circulating lipid levels despite decreased FA uptake5, suppresses glucose transporter Glut1. This impaired glucose entry into photoreceptors results in a dual lipid/glucose fuel shortage and reduction in the Krebs cycle intermediate α-ketoglutarate (KG). Low α-KG levels promote hypoxia-induced factor-1α (Hif1a) stabilization and vascular endothelial growth factor (Vegfa) secretion by starved Vldlr−/− photoreceptors, attracting neovessels to supply fuel. These aberrant vessels invading normally avascular photoreceptors in Vldlr−/− retinas are reminiscent of retinal angiomatous proliferation (RAP), a subset of neovascular age-related macular degeneration (AMD)7, associated with high vitreous VEGF levels in humans. Dysregulated lipid and glucose photoreceptor energy metabolism may therefore be a driving force in neovascular AMD and other retinal diseases.

Angiogenesis of the microvasculature is central to the etiology of many diseases including proliferative retinopathy, age-related macular degeneration and cancer. A mouse model of microvascular angiogenesis would be very valuable and enable access to a wide range of genetically manipulated tissues that closely approximate small blood vessel growth in vivo. Vascular endothelial cells cultured in vitro are widely used, however, isolating pure vascular murine endothelial cells is technically challenging. A microvascular mouse explant model that is robust, quantitative and can be reproduced without difficulty would overcome these limitations. Here we characterized and optimized for reproducibility an organotypic microvascular angiogenesis mouse and rat model from the choroid, a microvascular bed in the posterior of eye. The choroidal tissues from C57BL/6J and 129S6/SvEvTac mice and Sprague Dawley rats were isolated and incubated in Matrigel. Vascular sprouting was comparable between choroid samples obtained from different animals of the same genetic background. The sprouting area, normalized to controls, was highly reproducible between independent experiments. We developed a semi-automated macro in ImageJ software to allow for more efficient quantification of sprouting area. Isolated choroid explants responded to manipulation of the external environment while maintaining the local interactions of endothelial cells with neighboring cells, including pericytes and macrophages as evidenced by immunohistochemistry and fluorescence-activated cell sorting (FACS) analysis. This reproducible ex vivo angiogenesis assay can be used to evaluate angiogenic potential of pharmacologic compounds on microvessels and can take advantage of genetically manipulated mouse tissue for microvascular disease research.

Objective Regulation of angiogenesis is critical for many diseases. Specifically, pathological retinal neovascularization, a major cause of blindness, is suppressed with dietary ω3-long-chain polyunsaturated fatty acids (ω3LCPUFAs) through antiangiogenic metabolites of cyclooxygenase and lipoxygenase. Cytochrome P450 epoxygenases (CYP2C8) also metabolize LCPUFAs, producing bioactive epoxides, which are inactivated by soluble epoxide hydrolase (sEH) to transdihydrodiols. The effect of these enzymes and their metabolites on neovascularization is unknown. Approach and Results The mouse model of oxygen-induced retinopathy was used to investigate retinal neovascularization. We found that CYP2C (localized in wild-type monocytes/macrophages) is upregulated in oxygen-induced retinopathy, whereas sEH is suppressed, resulting in an increased retinal epoxide:diol ratio. With a ω3LCPUFA-enriched diet, retinal neovascularization increases in Tie2-driven human-CYP2C8–overexpressing mice (Tie2-CYP2C8-Tg), associated with increased plasma 19,20-epoxydocosapentaenoic acid and retinal epoxide:diol ratio. 19,20-Epoxydocosapentaenoic acids and the epoxide:diol ratio are decreased with overexpression of sEH (Tie2-sEH-Tg). Overexpression of CYP2C8 or sEH in mice does not change normal retinal vascular development compared with their wild-type littermate controls. The proangiogenic role in retina of CYP2C8 with both ω3LCPUFA and ω6LCPUFA and antiangiogenic role of sEH in ω3LCPUFA metabolism were corroborated in aortic ring assays. Conclusions Our results suggest that CYP2C ω3LCPUFA metabolites promote retinal pathological angiogenesis. CYP2C8 is part of a novel lipid metabolic pathway influencing retinal neovascularization.