Quantification of Retinal Vessel Coverage, Tortuosity, and Width in an In vivo Mouse Model – Implications for Retinopathy of Prematurity

Background Retinopathy of prematurity (ROP) is a vaso‐proliferative disease in premature infants resulting from their immature retinas’ susceptibility to insults that disrupt angiogenesis and neurovascular growth, most significantly oxidative stress. Clinically, ROP is subjectively diagnosed by dilated retinal exams to monitor retinal vessel appearance and extent of peripheral growth. Abnormally tortuous and dilated retinal vessels define plus disease, an indicator of disease severity that mandates treatment within 72 hours to prevent retinal detachment. There is no known way to quantify plus disease, leading to discrepancies of diagnosis amongst experts, missed or delayed diagnosis. Despite current management strategies, there is no cure for ROP, which remains a major cause of childhood blindness worldwide. Animal models of oxygen‐induced retinopathy (OIR) utilize ex vivo methods to study retinal vascular growth, such as enucleation for histology and retinal flat mounts. This study's objective was to use fluorescein angiography (FA) to characterize and quantify retinal vessels in live OIR mice. Methods 28 wildtype mice were exposed to 77% oxygen from postnatal day (P) 7 till P12. 27 age‐matched mice were raised in room air (RA). At early (P17 and 18) and late (P33 and 34) phases of retinal vascular development, mice were anesthetized, their pupils dilated, and an intraperitoneal injection of fluorescein given. The retinal blood vessels were directly visualized using a retinal imaging system. A customized software program was developed to quantify retinal vessel coverage (RVC), retinal vein width (RVW) and retinal arterial tortuosity (RAT) in the FA images acquired. Results OIR mice showed areas of capillary obliteration in early phase with an increase in RVC by late phase, while RA mice had uniform RVC in both early and late phases. Retinal veins were more dilated in early phase in OIR (52.20±8.76μm, N=14) compared to RA mice (39.20±5.47μm, N=12, p=0.0002), but by early phase OIR mice RVW had decreased to become equivalent to RVW in RA mice (p=0.39). RA mice maintained uniform vessel width in all phases. Retinal arteries were more tortuous in early phase in OIR (1.22±0.12, n=15) compared RA mice (1.00±0.003, N=12, p<0.0001); and still remained more tortuous in late phase in OIR (1.06±0.03, n=12) than RA mice (1.00±0.002, n=16), (p<0.0001). Conclusions This study has shown the novel use of FA in live, anesthetized mice to differentiate arteries from veins and quantify retinal vascular features. Arterial tortuosity persisted in OIR mice, despite full peripheral retinal vascular growth, a feature not described in existing ex vivo techniques. RAT is a more reliable indicator of abnormal vascular recovery following oxidative stress than RVC and RVW. Further in vivo imaging studies could advance clinical translation of objective methods of ROP diagnosis, allow physiologic studies and testing of therapeutic options to prevent adverse visual outcomes in ROP. Support or Funding Information University of Wisconsin, Department of Pediatrics; UW Centennial Scholars Program