Homotypic fetal transplants into an experimental model of spinal cord neurodegeneration

Many neurotransplantation studies have dealt with the ability of solid fetal spinal grafts to develop in the previously traumatized spinal cord of a host. In neurodegenerative spinal diseases, however, motoneuronal death occurs in the absence of a trauma, i.e., in the absence of axotomy of afferent fibers. Lesioning the spinal cord with an excitotoxic agent may provide a useful neurodegenerative model. The present study has been undertaken to determine whether homotypic fetal neurons transplanted as a cell suspension are able to rebuild a neural circuitry. Emphasis is given here to the analysis of the development of transplanted motoneurons and host-graft connectivity. The lesion was made by kainic acid on the right side of the lumbar enlargement 1 week before transplantation. The fetal spinal cords were taken from rat embryos (gestational day E12-13) and transplanted as cell suspensions. Light- and electron-microscopic analysis demonstrated that the excitotoxic lesion extended over the entire spinal segment and was confined primarily to the ventral and intermediate horns, implying the death of all motoneurons with consequent paralysis and muscular atrophy of corresponding hindlimb. The lesion was characterized by a lack of neurons, glial proliferation, and sparing of fibers of passage and afferents. Two to fourteen months after surgery, the transplants were generally large, occupying most of the neuron-depleted area. The boundaries between the transplant and host tissue were clearly delineated by the higher cellular density of the graft and the particular cytoarchitecture, i.e., the cell suspension grafts did not display a laminar organization. Among the different neuronal populations within the transplant, one resembled motoneurons: large, typically Nissl-stained and immunoreactive for calcitonin gene-related peptide (CGRP). No grafted neuron, however, extended an axon into the host ventral roots. Monoaminergic afferents from the host were studied using immunostaining for serotonin, noradrenaline, and tyrosine hydroxylase. These afferent fibers, thin and varicose, grew for a long distance and formed a network within transplants. Similarly, primary sensory CGRP-immunoreactive fibers (entering the graft from the dorsal host-graft interface) penetrated deeply into transplants. The response of cortico-and rubro-spinal afferents to the implantation of fetal tissue was different. After injection of WGA-HRP, a few anterogradely labeled cortical and rubral fibers entered only the most peripheral portion of transplants. In conclusion, our results indicate that fetal spinal neurons can be successfully transplanted into the adult neuron-depleted spinal cord. Host-to-graft connections can be formed, although their spatial extent in the transplants may depend upon structural features of the afferent fiber systems. In contrast, axonal growth of transplanted neurons into the depopulated ventral roots was not observed. Such a failure may be due to the nonpermissive substrate formed by mature oligodendrocytes and myelin of the spinal white matter.