Axonal sprouting and frank regeneration in the lizard tail spinal cord: correlation between changes in synaptic circuitry and axonal growth

In our previous studies, we found that the number of supraspinal neurons projecting to the level of tail spinal cord increases by 74% during tail regeneration and that the number of local spinal neurons with descending projections increases 233%. However, only a small fraction of the supraspinal axons (less than 4%) and half of the local spinal axons actually enter the regenerated spinal cord. We suggested that this may be the result of "synaptic capture" in which regrowing axons make synapses on denervated targets rostral to the transection, aborting further regeneration. To examine this hypothesis, morphometric analysis of electron microscope (EM) photomontages was used to test for changes in synaptic distribution on ventral horn neurons rostral to regenerating tail spinal cord. In addition, 3H-thymidine and retrograde markers were used to determine whether the regenerate axons arose from cut axons, neurogenesis, or sprouting from uninjured neurons. 3H-thymidine injections during regeneration, combined with retrograde HRP pathway tracing, did not reveal the production of new neurons in the tail spinal cord. To test whether cut axons regenerate, fluorescein isothiocyanate conjugated latex beads were applied to the exposed end of the tail spinal cord. After tail regeneration, HRP was applied to the new spinal cord in the regenerated tail. Examination of local spinal neurons (the primary source of axons that enter the regenerated tail spinal cord) revealed that 28% of the neurons contained both labels. This indicated that cut axons successfully regrew into the new tail spinal cord. The regenerated axons that fail to enter the new tail spinal cord can be found in the normal spinal cord immediately rostral to the regenerated tail. To determine whether these axons were making synaptic contacts, lamina IX ventral horn neurons were examined. EM photomontages of the spinal cord rostral to the regenerate tail revealed the following properties: (1) neurons rostral to regenerated tails are larger in area compare to non-regenerates (mean increase = 112%); (2) axosomatic contacts cover a greater percentage of the neuronal soma following regeneration compared to normal (mean increase = 23%); and (3) this increased innervation is the result of an increase in the number of synaptic boutons rather than larger boutons. The number of synaptic contacts in regenerated lizards returned to normal following lumbar transection, indicating that supraspinal and/or long descending propriospinal afferents were the major source of the increased synaptic contacts.(ABSTRACT TRUNCATED AT 400 WORDS)