Abnormal visual input leads to development of abnormal axon trajectories in frogs

Throughout the normal vertebrate brain, visual maps from the left and right eyes overlap and are in register with one another. Visual input has a major role in the development of the pathways which mediate these binocular projections1–3. A dramatic example of the developmental role of sensory input occurs in the isthmo-tectal projection, which is part of the polysynaptic relay from the eye to the ipsilateral tectum of the frog, Xenopus laevis4–8. If one eye is rotated when the animal is still a tadpole, the isthmic axons respond by changing the topography of their terminations in the tectum; for example, a given isthmo-tectal axon which normally would connect with medial tectum can be induced to terminate in lateral tectum. Such rearrangements bring the ipsilateral visual map into register with the contralateral retinotectal map, even though one eye has been rotated. Indirect evidence8 has suggested that after early eye rotation, isthmo-tectal axons do not grow directly to their new tectal targets but instead reach those targets by routes which pass through their normal termination zones. Here I have used anterograde horseradish peroxidase labelling of isthmo-tectal fibres to show the trajectories of such axons and to compare them with the routes which axons take when allowed to develop normally. Tracings of individual axons in flat-mounted, unsectioned tecta show that most axons in normal Xenopus follow fairly straight paths in the tectum. In contrast, early eye rotation causes many isthmo-tectal axons to follow crooked, circuitous pathways before they terminate.