A behavioral model of excitotoxicity: retinal degeneration, loss of vision, and subsequent recovery after intraocular NMDA administration in adult rats.

To establish a new behavioral animal model of excitotoxicity, we injected adult rats intraocularly with a single dose of 2, 20, or 100 nmol of N-methyl-d-aspartate (NMDA). We quantified visual impairment by plotting the size of the visual field in which the rats successfully oriented towards a small, moving target. In comparison to the saline-injected (contralateral) control side, the side injected with 2 nmol of NMDA was not significantly impaired. When injected with higher doses, the rats were nearly blind immediately after surgery, with only about 20% (20 nmol NMDA) or 10% (100 nmol NMDA) of residual vision. Within about 3 weeks, however, visual performance returned to near-normal levels. Simultaneous intraocular administration of a non-competitive NMDA-antagonist, MK-801 (1 nmol), resulted in complete behavioral protection. NMDA administration led to a dose-dependent loss of cells within the ganglion cell layer, as assessed in whole-mounted retinae which were retrogradely labelled with horseradish peroxidase (HRP). Whereas 2 nmol of NMDA led to the loss of about 30% of retinal ganglion cells (RGCs), at higher NMDA doses only 13% of the RGCs survived. After the injection of 20 nmol of NMDA, large-diameter RGCs (>22 μm) survived the lesion to a greater extent than small diameter cells (8–21 μm); at 100 nmol cells of all diameters were equally affected. The number of Nissl-stained cells with small diameters (<11 μm), presumed to be displaced amacrine cells, was also affected by NMDA, although to a lesser degree. Analysis of behavioral performance (vision score) and the number of cells in the retina revealed a correlation of r=0.76 between visual performance and the number of HRP-filled RGCs immediately after surgery. Lower correlations were found between visual performance and cells stained with Nissl of diameters smaller than 11 μm (presumably displaced amacrine cells) or larger than 11 μm (presumed RGCs without retinofugal connections; r=0.55 and r=0.58, respectively). Because of the spontaneous recovery of vision, all correlations declined to values near 0 after 3 weeks. Thus, despite a dramatic loss of RGCs following NMDA administration, visual deficits recover significantly in adult rats within 2–3 weeks.