Abstract
Purpose:
Retinal ischemia is common in eye disorders, like diabetic retinopathy or retinal vascular occlusion. The goal of this study was to evaluate the effects of an intravitreally injected vascular endothelial growth factor (VEGF) inhibitor (Ranibizumab) on retinal cells in an ischemia-reperfusion animal model (I/R).
Methods:
I/R was induced in adult Brown Norway rats and one half was treated with Ranibizumab (10 mg/ml) three days later. The control group did not undergo I/R or treatment (n=6-8/group). After 3 weeks aqueous humor VEGF levels were measured via ELISA. Retinal cross-sections were stained with H&E, Brn-3a (retinal ganglion cells), rhodopsin (photoreceptors), ChAT (amacrine cells), GFAP (macroglia), and VEGF-R2 (VEGF-receptor). Cell counts and staining area evaluation was performed using ImageJ followed by statistical analysis (Statistica).
Results:
A significant VEGF increase was detected in aqueous humor of I/R eyes (p=0.02), but not in Ranibizumab ones (p=0.9). I/R retinas showed significantly lower RGC numbers (16.7±3.0 cells/mm; p=0.001) compared to controls (34.5±1.7 cells/mm). Numbers in the Ranibizumab group (31.4±3.3 cells/mm) were comparable to controls (p=0.2). The rhodopsin+ area was significantly reduced in the I/R group (I/R: 9.1±1.2%; control: 12.9±0.9%; p=0.02). The Ranibizumab group (15.2±1.1%) had a significantly greater rhodopsin+ area than the I/R group (p<0.001). In the I/R and Ranibizumab group a decreased ChAT score was observed (p<0.001). Also, fewer ChAT+ cells were noted in these groups (p<0.001). VEGF-R2 was highly expressed in I/R retinas (p=0.006), but not in Ranibizumab ones (p=0.4). Also, a stronger GFAP response was observed in I/R (p=0.01) than in Ranibizumab retinas (p=0.4).
Conclusions:
Ranibizumab treatment protects retinal cells, like RGCs and photoreceptors against ischemic damage. Only amacrine cells could not be rescued. These cells seem to be particularly sensitive to this damage. Possibly, an earlier intervention is necessary to protect amacrine cells in this model.