Selective retinal reinnervation of a surgically created tectal island in goldfish. II. Electron microscopic analysis.

In the preceding study (Edwards et al., '85), we showed that regenerating optic axons reestablish a topographically restricted projection to a caudal tectal island created by surgical removal of a 1-mm-wide strip of caudal tectum in goldfish. In the present ultrastructural study, we evaluated the dependence of this axonal outgrowth on the presence of tectal target tissue caudal to the gap. Axon counts in the lesion zone were compared between cases with complete caudal tectal ablation and cases with ablation sparing a caudal tectal island (with and without optic nerve crush). During the postoperative interval of 20-50 days (early period), up to about 1,000 unmyelinated axons with features characteristic of optic axons were present in numerous small subpial bundles in both preparations. In the subsequent interval of 50-110 days (middle period), less than 200 axons were counted caudal to simple half-tecta, whereas 4,000-14,000 myelinated and unmyelinated axons were present in a few large bundles which crossed the lesion zone of tectal island cases. In this period, optic terminals could be demonstrated in the tectal island using the anterograde horseradish peroxidase method. At 170-300 days after surgery (late period), bridging bundles contained between 2,000 and 6,000 largely myelinated axons. We conclude that caudal tectal tissue is not necessary for the initial outgrowth of a small number of axons beyond a rostral half-tectum. The target is essential, however, for the maintenance of these axon fascicles and for the subsequent massive outgrowth of axons to the island. The contributions of glial guidance, diffuse exploratory outgrowth, and target-produced trophic factors to the formation of an initially exuberant projection to the island are discussed. A process of selective axon collateral withdrawal is proposed to account for the decrease in axon numbers within bridging bundles in the late period and for the late restriction in the retinal origin of the island projection indicated by results in the preceding study (Edwards et al., '85).