Fig 1: UNC5B silencing in endothelial cells disrupts blood-retinal barrier homeostasis in DR model mice and abrogates the protective effect of high-concentration netrin-1. At week 4 after successful DR model induction, different concentrations of netrin-1 (5, 50, 500, 1,000 and 5,000 ng/ml) were injected intravitreally. Retinal tissues were collected after ≥12 weeks of DR model induction for subsequent analysis. (A) EB staining was used to observe retinal vascular leakage in the different groups. Scale bar, 500 μm. At week 4 after DR model induction, control shRNA-AAV carrying an endothelial cell-specific promoter sequence (DR + shC), 1,000 ng/ml netrin-1 (DR + Netrin-1), UNC5B shRNA-AAV carrying an endothelial cell-specific promoter sequence (DR + shUNC5B) or a combination of 1,000 ng/ml netrin-1 and UNC5B shRNA-AAV (DR + shUNC5B + Netrin-1) were injected. After ≥12 weeks of DR model induction, retinal tissues were collected for analysis. (B) EB staining showed retinal vascular leakage in the different groups. Scale bar, 500 μm. (C) Periodic acid-Schiff staining revealed the formation of acellular capillaries (arrows) and the number of pericytes ('P') in the retinal tissues. Scale bar, 50 or 20 μm. (D) Whole-mount retinal immunofluorescence staining showed the pericyte coverage in the retinas of the different groups (red, NG2; green, IB4). Scale bar, 200 μm. The data are presented as the mean ± SD. n=5 per group. All statistical comparisons in this figure were performed among all groups using one-way ANOVA followed by Tukey's multiple comparisons test. *P<0.05 vs. WT. #P<0.05 vs. DR/DR + shC. AAV, adeno-associated virus; DR, diabetic retinopathy; EB, Evans Blue; IB4, isolectin B4; NG2, neuron-glial antigen 2; ns, not significant; shRNA, short hairpin RNA; shC, scramble control shRNA; shUNC5B, UNC5B shRNA; UNC5B, unc-5 netrin receptor B; WT, wild-type.