Pathfinding and target selection of goldfish retinal axons regenerating under TTX‐induced impulse blockade

To define the extent to which impulse blockade interferes with the morphological changes of regenerating retinal axons during their growth through the tectum, axons were deprived of activity by repeated intraocular injections of TTX. At intervals between 24 and 189 days after optic nerve section (ONS), a defined group of TTX‐silenced axons and of axons with normal activity (controls) were labeled by applications of HRP to the ventro‐ or dorsotemporal retina. The trajectories of these labeled axons were traced in DAB processed tectal wholemounts. As in controls, TTX‐blocked axons went through a phase of exploratory growth at early regeneration stages (24 to 80 days after ONS). Coursing in abnormal routes, the axons initially distributed their growing endings widely over the tectum. Axons with and without activity extended side branches with growth cones and filopodia over all regions of the tectum. These ramifications were of similar dimensions for the TTX‐blocked and control axons. Despite abnormal routes and branching over inappropriate territories, axons showed a preference for the rostral tectum. At late regeneration stages (120–189 days after ONS), axons had lost their side branches and their growth cones. Their preterminal segments exhibited striking bends, suggesting that they had undergone course corrections to achieve access to the retinotopic target. Axonal processes had disappeared from the caudal tectum, and the preferential accumulation of axons over the rostral tectum had increased. The majority of the TTX‐blocked and control axons ended in terminal arbors at retinotopic regions. The labeled arbors of the TTX‐group were no larger than those of the control group. The arbors of each group lay close together in a continuous cluster in the TTX‐group as well as in two‐thirds of the control group. In the other one‐third of the control group, however, terminal arbors were aggregated into separate patches. The clusters of the TTX‐blocked axons covered between 2.2 and 3.9% (mean 2.95%) of the tectal surface and the clusters and/or patches of active axons between 1.9 and 3.4% (mean 2.7%). Thus the terminal arbor clusters of the TTX‐silenced axons were not significantly larger than those of the active axons. These data show that retinal ganglion cell impulse activity is required for neither the extension of side branches in the early exploratory phase of regeneration nor for the withdrawal of these branches nor for the establishment of target‐directed routes and the deployment of normal‐size terminal arbors at retinotopic loci. Our data further suggest that the retinotopic map is refined considerably with time even in the absence of activity, a finding that is consistent with an abstract by Olsson and Meyer ('87).