Abstract
Purpose:
Diabetic retinopathy is commonly characterized as a microvasculopathy that results in retinal capillary loss with increasing severity of disease. Recent studies by us and others suggest that ganglion cell and nerve fiber loss is progressive in diabetic eyes with little to no detectable diabetic retinopathy on fundoscopy. We sought to determine whether neurodegeneration or vascular changes occur first in diabetic eyes with no or minimal retinopathy.
Methods:
Donor eyes from the Iowa Lions Eye bank were used in this study. Whole mount retinal punches (4mm from inferior macula) were immunostained with lectin and anti-SNCG antibody to label vessel walls and ganglion cells respectively and the entire tissue was assessed for capillary density; ganglion cells counts averaged in 6-7 regions of each piece of retina using confocal imaging, while capillary density over the entire punch was measured using a custom macro in Image J as a fraction of the punch area, Comparisons of variables were performed using two-tailed paired Student t-test, all reported as mean+/-SD with a 95% confidence interval (95% CI).
Results:
6 diabetic eyes (mean age 77.8 years) with no to mild NPDR and no history of ocular treatment for retinopathy were compared to 3 control eyes with no history of glaucoma or disease affecting the macula (mean age 76.3 years). The mean density of ganglion cells was reduced in the diabetic eyes (1.56/1000µm2; 95% CI: 1.16-1.96) compared to the control group (2.23/1000µm2; 95% CI: 1.59-2.87). Capillary density was higher in the diabetic eyes (0.182; 95% CI: 0.162-0.202) compared to controls (0.154; 95% CI: 0.153-0.155). There was a modest negative correlation between ganglion cell density and capillaries (R2=0.36) in all eyes.
Conclusions:
This study of donor eyes with minimal diabetic retinopathy suggests that loss of ganglion cells occurs in the absence of capillary dropout, but more eyes are needed to confirm this hypothesis. There is a modest negative correlation between numbers of ganglion cells and retinal capillary density. These results strengthen our hypothesis of a diabetic neural retinopathy that is neurotoxic in origin, rather than ischemic.