Role of HIF1α and HIF2α in Cre Recombinase-Induced Retinal Pigment Epithelium Pathology and Its Secondary Effect on Choroidal Neovascularization

Abstract

Cre^Trp1 mice are widely used for conditional retinal pigment epithelium (RPE) gene function studies. Like other Cre/LoxP models, phenotypes in Cre^Trp1 mice can be affected by Cre-mediated cellular toxicity, leading to RPE dysfunction, altered morphology and atrophy, activation of innate immunity, and consequent impairment of photoreceptor function. These effects are common among the age-related alterations of RPE that feature in early/intermediate forms of age-related macular degeneration. This article characterizes Cre-mediated pathology in the Cre^Trp1 line to elucidate the impact of RPE degeneration on both developmental and pathologic choroidal neovascularization. Nonredundant roles of the two major components of the hypoxia-inducible factor (HIF) family of transcription regulators, HIF1α and HIF2α, were identified. Genetic ablation of Hif1a protected against Cre-induced degeneration of RPE and choroid, whereas ablation of Hif2a exacerbated this degeneration. Furthermore, HIF1α deficiency protected Cre^Trp1 mice against laser-induced choroidal neovascularization, whereas HIF2α deficiency exacerbated the phenotype. Cre-mediated degeneration of the RPE in Cre^Trp1 mice offers an opportunity to investigate the impact of hypoxia signaling in the context of RPE degeneration. These findings indicate that HIF1α promotes Cre recombinase-mediated RPE degeneration and laser-induced choroidal neovascularization, whereas HIF2α is protective.

Figure 1

Characterization of Cre-mediated recombination and HIF expression in RPE cells of the Cre^Trp1;Hif_x^f/f lines. A: Representative eye section from an adult Cre^Trp1;ROSA^mT/mG mouse, showing membrane-targeted Green fluorescent protein (mGFP) distribution in the different cellular layers. Cre-mediated excision in the neuroretina of Cre^Trp1 mice has been reported previously.B: Central and peripheral confocal magnified images show mGFP expression (and Cre-mediated DNA excision) to be confined to the RPE monolayer in the RPE/choroid/sclera ocular compartment. C: Representative confocal images of central and peripheral RPE/choroid flat mount from an adult Cre^Trp1 mouse indicate persistent Cre expression in adulthood. D: Representative images of central and peripheral portions of RPE/choroid flat mounts from adult Cre^Trp1;ROSA^mT/mG. tdTomato fluorescence not shown. E: Quantification of central and peripheral RPE cells stained positive for mGFP (ie, derived from cells that underwent Cre-mediated recombination). F and G: RT-qPCR analysis of whole adult RPE/choroid samples from the reported genotypes; targets analyzed are Hif1a (F) and Hif2a (G). Data are expressed as means ± SEM. n = 4 to 5 animals per genotype (F and G); n = 5 to 6 animals per genotype (E). ∗P < 0.05, ∗∗P < 0.01 versus controls in the same experiment (analysis of variance and Tukey test for multiple comparisons); ^††P < 0.01, ^†††P < 0.001 versus Cre^Trp1; ^‡‡P < 0.01 versus Cre^Trp1;Hif1a^f/f; ^§§P < 0.01 versus Cre^Trp1;Hif2a^f/f (all, versus group in the same experiment). Scale bars: 200 μm (A); 50 μm (B and C); 100 μm (D). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

Figure 2

Morphologic abnormalities of RPE monolayer in Cre^Trp1;Hifx^f/f lines. A: Bright-field images of adult choroid/RPE flat mounts; examples of depigmented areas are highlighted by arrowheads. B: Representative live autofluorescence scanning laser ophthalmoscopy images in adult animals, focusing on the subretinal space/RPE region and highlighting qualitative differences in diffused hyper-reflective regions and hyper-reflective foci (asterisks) among the Cre^Trp1 lines. C: Immunohistochemistry staining of phalloidin and ionized calcium-binding adaptor molecule 1 (Iba1) (both central and peripheral regions) show a particularly severe morphologic phenotype in adult RPE cells and ameboid-shaped inflammatory cells migrated in the subretinal space in correspondence to damaged RPE regions in Cre^Trp1 and Cre^Trp1;Hif2a^f/f mice. D–G: Quantification of cell density (D), cell area (E), cell perimeter (F), and Iba1⁺ cell density (G) in both the peripheral and central RPE portions. Data are expressed as means ± SEM. n = 5 to 30 animals per genotype. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus controls; ^†P < 0.05, ^††P < 0.01 versus Cre^Trp1; ^‡P < 0.05, ^‡‡P < 0.01, and ^‡‡‡P < 0.001 versus Cre^Trp1;Hif1a^f/f; ^§P < 0.05, ^§§§P < 0.001 versus Cre^Trp1;Hif2a^f/f (all, versus groups in the same RPE region, by analysis of variance and Tukey test for multiple comparisons). Scale bars: 1 mm (A and B); 100 μm (C).

Figure 3

Choroidal vascular phenotype in Cre^Trp1;Hif_x^f/f lines. A: Confocal images of isolectin B4–fluorescein isothiocyanate (iB4-FITC)-perfused adult choriocapillaris in the phenotypes under analysis. Representative examples in central and peripheral portions are reported. B and C: Quantification of vascularized area (B) and vessel thickness (C). D: Between-genotype comparisons of mean RPE cell area and choroidal vascularized area, showing negative linear correlation. Linear interpolation equations: central, y = –0.0101x + 86.001; peripheral, y = –0.0214x + 90.148. Dashed lines represent the calculated data interpolation lines. Coefficients of determination (R²) are shown (the closer to 1.0, the more reliable the fit of the model). E: RT-qPCR analysis of Vegfa expression in whole P0–P1 and adult RPE/choroid samples. Genotypes in E are applicable to B, C, and G. F:In vitro choroidal sprouting assay experiment and representative examples of choroidal endothelium growth at 3 and 6 days in vitro. P18-P21 eyes were dissected in 1 × 1-mm explants, embedded in Matrigel, and placed in culture for 6 days under deep hypoxic conditions. Sprouting was assessed microscopically on a daily basis from day 3. G: Quantification of sprouting area over time. For simplicity, data from days 0 to 2 (no detectable sprouting) are not shown. Data are expressed as means ± SEM. n = 3 to 5 explants per animal (G); n = 4 to 7 animals per genotype (G); n = 4 to 14 animals per genotype (E); n = 5 to 18 animals per genotype (D). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001 versus controls; ^†P < 0.05, ^†††P < 0.001 versus Cre^Trp1; ^‡P < 0.05, ^‡‡P < 0.01, and ^‡‡‡P < 0.001 versus Cre^Trp1;Hif1a^f/f; ^§P < 0.05, ^§§§P < 0.001 versus Cre^Trp1;Hif2a^f/f (C, versus groups in the same choroidal region, by analysis of variance and Tukey test for multiple comparisons; E, versus groups of the same age, by analysis of variance and Tukey test for multiple comparisons; G, versus groups cultured for the same number of days in vitro, by analysis of variance and Bonferroni test for multiple comparisons). Scale bars: 25 μm (A); 1 mm (F).

Figure 4

Differences in laser choroidal neovascularization (CNV) lesions in Cre^Trp1;Hifx^f/f lines at days 7 and 14 after laser-induced injury. A: Representative images of laser vascular injuries visualized using fundus fluorescein angiography. B: Quantification of CNV lesion area. C: Representative confocal images of laser vascular injuries stained with isolectin B4 at day 14 after laser injury and three-dimensional (3D) reconstruction. D: Quantification of CNV lesion volumes at day 14 using Imaris software version 8. Data are expressed as means ± SEM. n = 14 to 45 laser burns per genotype. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus controls; ^†P < 0.05, ^†††P < 0.001 versus Cre^Trp1; ^‡P < 0.05, ^‡‡‡P < 0.001 versus Cre^Trp1;Hif1a^f/f; ^§§§P < 0.001 versus Cre^Trp1;Hif2a^f/f (all, versus groups analyzed at the same number of days after laser injury, by analysis of variance and Tukey test for multiple comparisons). Scale bars: 2 mm (A); 100 μm (C). iB4-FITC isolectin B4–fluorescein isothiocyanate.