Retinal vascular PO
2 was measured using an established optical section phosphorescence lifetime imaging system.
14,15 A laser beam was focused to a line and projected at an oblique angle on the retina. Using an intensified charge-coupled device camera attached to a slit-lamp biomicroscope (Carl Zeiss, Oberkochen, Germany), optical section phosphorescence images were acquired nasal and/or temporal to the optic disc. Since the incident laser and imaging axis were not coaxial, retinal vessels were displaced laterally from the choroidal vessels in the section image according to depth. Additionally, due to the binding of the oxygen-sensitive molecular probe to albumin, measurements were obtained only within the retinal blood vessels and not tissue. Images were analyzed to measure phosphorescence lifetimes, which were converted to PO
2 using the Stern-Volmer relationship.
16 In each mouse, two or three repeated PO
2 measurements in the same retinal blood vessel were averaged. Overall, PO
2 measurements were obtained in 6 ± 2 blood vessels in one eye of each mouse. The number of vessels in which PO
2 was measured varied due to anatomic differences among mice. In addition, small pupil size also limited phosphorescence imaging of some vessels. In OIR mice, abnormalities in the retinal vasculature often precluded a clear distinction between arteries and veins based on FA. Since the blood vessel with the highest PO
2 had to be an artery and the blood vessel with the lowest PO
2 had to be a vein, in each mouse, the maximum and minimum of all vascular PO
2 measurements were designated as retinal arterial and venous PO
2, respectively. The arteriovenous PO
2 difference was calculated as the difference between maximum and minimum PO
2 measurements. Although retinal arterial and venous PO
2 are related to retinal oxygenation, the arteriovenous PO
2 difference is of particular interest, because it is the vascular parameter related directly to the amount of oxygen consumed by the retinal tissue.
17,18 Therefore, comparison of arteriovenous PO
2 difference between control and OIR mice will determine the presence of abnormal retinal oxygen extraction. Due to the small size of the mice, it was not feasible to measure systemic arterial PO
2 directly. Therefore, to reduce variability due to systemic physiologic condition and anesthesia, and ensure that control and OIR mice had similar retinal arterial oxygen levels, data were included from 11 control and 19 OIR mice that had retinal arterial PO
2 within one SD of the mean in control mice (35 ± 11 mm Hg,
N = 13). Data from 2 outliers from each of the control and OIR groups were excluded.