Multipotent neural stem cells (NSCs) are operationally defined by their ability to self-renew, to differentiate into cells of all glial and neuronal lineages throughout the neuraxis, and to populate developing or degenerating CNS regions. Thus their use as a graft material can be considered analogous to hematopoietic stem cell-mediated reconstitution and gene transfer. The recognition that NSCs propagated in culture could be reimplanted into mammalian brain, where they might integrate appropriately throughout the mammalian CNS and stably express foreign genes, has unveiled a new role for neural transplantation and gene therapy and a possible strategy for addressing the CNS manifestations of diseases that heretofore has been refractory to intervention. We have tracked the response of host and transplanted NSCs to brain or spinal cord injury and explored the therapeutic potential of NSCs injected into the animal CNS subjected to focal hypoxic-ische-mic (HI) brain or spinal cord injury. Such cells integrated appropriately into the degenerating CNS, showed robust engraftment and foreign gene expression within the region of CNS injury, and appeared to have migrated preferentially to the site of injury, experienced limited proliferation, and differentiated into neural cells lost to injury, trying to repopulate the damaged CNS area. The transplantation of exogenous NSCs may, in fact, augment a natural self-repair process in which the damaged CNS "attempts" to mobilize its own pool of stem cells. Providing additional NSCs and trophic factors may optimize this response. Therefore, NSCs may provide a novel approach to reconstituting CNS damaged by HI brain or spinal cord injury. Preliminary data in animal models of hypoxic-ischemic brain injury or contusive spinal cord injury lend support to these hypotheses.