Propriospinal projections to the ventral horn of the rostral and caudal hindlimb enlargement in turtles.

In limbed vertebrates, the capacity to generate rhythmic motor patterns for locomotion and scratching is distributed over spinal cord segments of the limb enlargement (e.g., lumbosacral segments), but within this region, rostral segments are more rhythmogenic than caudal segments. The underlying reasons for this rostrocaudal asymmetry are not clear. One possibility is that rostral and caudal segments receive distinct sets of propriospinal projections. To test this hypothesis, I injected horseradish peroxidase (HRP) into the ventral horn unilaterally in a rostral or caudal segment of the turtle hindlimb enlargement. I quantitatively assessed the distributions of retrogradely labeled neurons in six hindlimb enlargement and pre-enlargement segments. The cross-sectional distribution did not depend on which segment was injected. Ipsilateral labeling occurred predominantly in the deep dorsal horn, the lateral part of the intermediate zone, and the dorsal two-thirds of the ventral horn, while contralateral labeling occurred mainly in the medial part of the ventral horn and the lateral part of the intermediate zone. This cross-sectional distribution is similar to what has been seen in mammals. The rostrocaudal distribution of labeled cells, however, depended on which segment was injected. Rostral injections gave rise to rostrally skewed distributions, dominated by descending propriospinal neurons. Caudal injections gave rise to caudally skewed distributions, dominated by ascending propriospinal neurons. Thus, rostral segments of the hindlimb enlargement received more propriospinal inputs from immediately rostral than immediately caudal segments, while the reverse was true for inputs to caudal segments. This anatomical asymmetry may contribute to known functional asymmetries within the enlargement.