OBJECTIVETo investigate whether propranolol application as collyrium or intraperitoneal (IP) injection can promote the recovery of oxygen-induced retinopathy (OIR).

METHODThirty-six 7-day-old mice were divided into the following 6 groups: normal control, propranolol eye drops, propranolol IP injection, eye drops negative control, IP injection negative control, and pathological model with 6 mice in each. In a typical model of OIR, litters of mice pups with their nursing mothers were exposed to an infant incubator to high oxygen concentration (75 ± 5)% between postnatal day (PD) 7 and PD12, prior to returning to room air. Two routes of propranolol treatment were assessed from PD12 to PD17: IP injection and eye drop, with doses 2 mg/(kg·time), three times a day. Another three groups were given citric acid buffer eye drops, IP injection of citric acid buffer, and negative control were not treated with any drug. Neonatal mice fed in normal conditions served as normal control. Mice were sacrificed at PD17 to evaluate the morphological changes of retinal vessels by fluorescein isothiocyanate-dextran perfusion and retinal whole mount. The retinal neovascularization was evaluated by counting the number of nuclei of the endothelial cell breaking through the internal limiting membrane (ILM).

RESULTCompared with the oxygen-exposed group, the branches of retinal vessels went normal with a less un-perfused area in the propranolol eye drops and propranolol IP injection groups [(38.9 ± 9.9)% and (5.6 ± 2.3)% vs. (16.2 ± 10.0)% and (2.2 ± 0.8)%, (25.9 ± 5.0)% and (2.1 ± 2.7)%, F=36.12 and 14.55, P both<0.001]. The number of nuclei of endothelial cells breaking through the ILM on the retinal cross-section in the propranolol eye drops group decreased (14.2 ± 5.1) per slide, which was less than that in the oxygen-exposed group (49.1 ± 8.9) per slide and the propranolol IP injection group (18.0 ± 5.9) per slide; it was also less than that in the eye drops negative control group (47.4 ± 8.1) per slide (F=187.60, P<0.05). Moreover, the number of nuclei of endothelial cells breaking through the ILM on the retinal cross-section in the propranolol IP injection group was less than that in the IP injection negative control group (49.9 ± 7.1) per slide (P<0.05).

CONCLUSIONPropranolol could effectively inhibit the formation of retinal neovascularization in mice; the eye drops was more effective than the IP injection.