Arginase 2 deficiency prevents oxidative stress and limits hyperoxia-induced retinal vascular degeneration

Abstract

Background: Hyperoxia exposure of premature infants causes obliteration of the immature retinal microvessels, leading to a condition of proliferative vitreoretinal neovascularization termed retinopathy of prematurity (ROP). Previous work has demonstrated that the hyperoxia-induced vascular injury is mediated by dysfunction of endothelial nitric oxide synthase resulting in peroxynitrite formation. This study was undertaken to determine the involvement of the ureahydrolase enzyme arginase in this pathology.

Methods and findings: Studies were performed using hyperoxia-treated bovine retinal endothelial cells (BRE) and mice with oxygen-induced retinopathy (OIR) as experimental models of ROP. Treatment with the specific arginase inhibitor 2(S)-amino-6-boronohexanoic acid (ABH) prevented hyperoxia-induced apoptosis of BRE cells and reduced vaso-obliteration in the OIR model. Furthermore, deletion of the arginase 2 gene protected against hyperoxia-induced vaso-obliteration, enhanced physiological vascular repair, and reduced retinal neovascularization in the OIR model. Additional deletion of one copy of arginase 1 did not improve the vascular pathology. Analyses of peroxynitrite by quantitation of its biomarker nitrotyrosine, superoxide by dihydroethidium imaging and NO formation by diaminofluoroscein imaging showed that the protective actions of arginase 2 deletion were associated with blockade of superoxide and peroxynitrite formation and normalization of NOS activity.

Conclusions: Our data demonstrate the involvement of arginase activity and arginase 2 expression in hyperoxia-induced vascular injury. Arginase 2 deletion prevents hyperoxia-induced retinal vascular injury by preventing NOS uncoupling resulting in decreased reactive oxygen species formation and increased nitric oxide bioavailability.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Inhibiting arginase inhibits hyperoxia-induced cell death and preserves nitrite accumulation in vitro and vaso-obliteration in vivo.

For in vitro studies, BRECs were treated with hyperoxia (40% O2, 5% CO2) or normoxia (21% O2, 5% CO2) with and without ABH (100 µM) for 48 hours. Groups of cells were stained with propidium iodide and the percentage of apoptotic nuclei (hypodiploid, M1) and normal nuclei (diploid, M2) were quantified by flow cytometry (A, B). n = 4–5, * P≤0.05 vs normoxia and hyperoxia treated with ABH, # P≤0.05 vs hyperoxic and normoxic treatment. NO release was detected using chemiluminescence to measure nitrite levels in conditioned media from the treated cells (C. n = 6–9. * P≤0.05 vs normoxia with or without ABH, # P≤0.05 vs hyperoxia without ABH. For in vivo analyses, wild type mice were treated with daily i.p. injections of vehicle (saline) or ABH (15 mg/kg) from P7 to P9 or P12. Retinal vessels were visualized by lectin labeling (D) and the area of capillary dropout (yellow) was quantified in fluorescence micrographs using ImageJ (E). n = 7–9, *P≤0.05 vs saline.

Figure 2. Arginase 2 deletion limits hyperoxia-induced retinal vaso-obliteration.

Wild type (WT), arginase 2-deficient mice (A1+/+A2−/−) and arginase-deficient mice lacking one copy of arginase 1 (A1+/−A2−/−) were placed in 70% oxygen on P7 and prepared for analysis on P8, P9 or P12. Retinal vessels were visualized by lectin labeling (A) and the area of capillary dropout (yellow) was quantified in fluorescence micrographs using ImageJ (B, C). n = 5–9, *P≤0.05 vs WT. Images of hematoxylin and eosin stained cryostat sections from adult mice show comparable retinal morphology in WT, A2−/− and A1+/−A2−/− retinas (D, GCL: ganglion cell layer, IPL: inner plexiform layer, INL: inner nuclear layer, OPL: outer plexiform layer, ONL: outer nuclear layer, RPE: retinal pigment epithelium, scale bar  = 50 µm).

Figure 3. Arginase 2 deletion limits pathological vitreo-retina neovascularization while enhancing intra-retinal neovascularization.

Wild type (WT), arginase 2-deficient mice (A1+/+A2−/−) and arginase 2-deficient mice lacking one copy of arginase 1 (A1+/−A2−/−) were maintained in 70% oxygen from P7 to P12, returned to normoxia for 5 days and prepared for analysis on P17. Retinal vessels were visualized by lectin labeling (A) and areas of vitreoretinal neovascular tufts (arrows) and capillary dropout (yellow) were quantified in fluorescence micrographs using ImageJ (B,C). n = 11–13, *P≤0.05 vs WT and A1+/−A2−/− in B and WT in C.

Figure 4. Hyperoxia treatment and Arginase 2 expression in OIR retina.

Wild type mice (WT) and mice wild type for arginase 1 and deficient in arginase 2 (A2−/−) were placed in 70% oxygen (OIR) or room air (RA) on P7 and prepared for analysis on after 24 (A) or 48 hr (B). Immunofluorescence imaging (A) of arginase 2 and calbindin in retinal cryosections shows that arginase 2 (red) is highly expressed in horizontal cells (green). Scale bar  = 50 µm, n = 3. Western blot analysis (B) showing arginase 1 and 2 levels following normoxia and hyperoxia treatment, n = 5.

Figure 5. Arginase 2 deletion prevents hyperoxia-induced increases in retinal nitrotyrosine (A) and superoxide (C) levels and preserves NO formation (E).

Wild type (WT) and mice wild type for arginase 1 and deficient in arginase 2 (A2−/−) were placed in 70% oxygen (OIR) or room air (RA) on P7 and prepared for analysis on P8. Nitrotyrosine levels in retina samples were detected by slot blot (A) and quantified using ImageJ software (B). n = 6, *P≤0.05 vs WT-RA, # P≤0.05 vs WT-OIR. DHE imaging of superoxide formation was performed using flash frozen retinal sections (C). Results were quantified using Metamorph Imaging System (D). n = 3–5, scale bar  = 50 µm, *P≤0.05 WT-OIR vs all groups; # = P≤0.05 vs WT-OIR. † P