Activity-dependent retinotopic refinement in a low-density retinotectal projection in the goldfish: evidence favoring...

During optic nerve regeneration in goldfish, the label from a small retinal spot injection of WGA-HRP has been previously reported to be initially widely dispersed in the tectum and subsequently to condense into a small patch in the retinotopically appropriate location of tectum. This refinement involves two separate processes: one is activity independent and generates gross retinotopy; the other is activity dependent and mediates the formation of fine retinotopy. Since the number of synapses remains constant during this refinement, one or both of these processes may involve some form of competition for a limited number of synaptic sites. To clarify the role of synaptic competition, we created a low-density retinotectal projection in goldfish by deflecting about 20% of the optic fibers from one tectum into the opposite tectum, which was denervated of all other optic fibers. Under this condition, it was previously shown that less than half the normal density of synapses is formed. If competition for synaptic sites is a requirement of refinement, refinement should be prevented or significantly hindered. To monitor refinement during regeneration, 2 nl spot injections of WGA-HRP were made into the retina at various times after deflection. To distinguish between activity-dependent and activity-independent refinement, retinal impulse activity was blocked in some fish with repeated injections of TTX into the eye for the duration of the experiment. It was found that considerable activity-independent refinement occurred under continuous TTX blockade although the fibers remained more dispersed than in previous TTX studies when normal numbers of fibers were present. Surprisingly, in fish with normal impulse activity, the degree of activity-dependent refinement was almost normal. Labeled fibers condensed into a small area roughly comparable in size to that observed when the full complement of fibers was regenerating into tectum. These results suggest that competition for limited synaptic sites is not essential for activity-dependent refinement, which may, instead, be mediated by a cooperative process that actively promotes convergence. The findings further suggest that if synaptic competition plays a role in this system, it is in regulating activity-independent mechanisms that determine the large-scale distribution of fibers within tectum.