Rats were anesthetized as described above, with a femoral vein cannula inserted for infusion of 2 × 10
6 molecular weight fluorescein isothiocyanate (FITC)-dextran. The dye was injected as a bolus while the retina was video recorded through a Coolsnap ES camera (Photometrics, Tucson, AZ) attached to a Nikon Eclipse microscope (Nikon Instruments, Melville, NY) using a 10× objective. Two infusions in the amount of 5 mg/kg were performed so that measurements could be made for both eyes: the dye was administered in a volume of 100 μL saline for the first bolus, and then after an equilibration period, the dye infusion was concentrated 2× (i.e., using a volume of 50 μL instead of 100 μL) for the second eye. The vessels filling first with dye were the arterioles (
Fig. 1A), with the arteriolar fluorescent intensity reaching a peak, and beginning to decline as the dye made its transit through the capillaries and into the venules (
Fig. 1B). The fluorescent intensity curves of the arterioles and venules were used to calculate the mean transit times of each as described previously,
5,17,18 with the difference in mean transit times defined as the mean circulation time (MCT). In this technique, the tail end of the intensity curves are obscured by the recirculating dye in the bloodstream, but are estimated conventionally by a logarithmic fit of the declining phase of the curve, as shown in
Figure 1C. Data from the right and left eyes were averaged together, whenever data were obtained for both. A flow rate index (retinal blood volume divided by the MCT) can be calculated by dividing the sum of the squares of the arteriolar and venular diameters (D
art and D
ven) by the MCT, that is, (D
art 2 + D
ven 2)/MCT, with the assumption that the numerator is proportional to the effective circulating retinal blood volume.
17 Diameters were measured from the video recordings using NIS Elements software (Nikon Instruments): the 10× objective coupled with 2 × 2 binning of the camera pixels (to enhance fluorescent contrast) provided a resolution of 1.29 μm per measurement pixel. The numbers of rats included in these measurements were 12 controls and 9 STZ.