idence of periventricular leukomalacia sustained injury prior to birth. This is consistent with the results of the National Institutes of Health Perinatal Collaborative Study showing that most cerebral palsy originates prenatally.' The paper by Menkes and Curran' cited in the editorial' also reflects this view. The authors state that periventricular leukomalacia recognized in term infants is believed to reflect intrauterine "late second-or early third-trimester injury." In his recent editorial concerning the potential neurotoxicity of levodopa,' Fahn cites in vitro data indicating that in coculture paradigms, the presence of glial cells confers significant cytoprotection to neurons in the face of levodopa chall e x~g e .~,~ O n the basis of these observations, he concludes that because "glia are present in situ, the question of in vivo neurotoxicity from levodopa remains more doubtful than previously thought."' Although data attesting to the beneficial role of astrocytes in neuronal survival and neuritic outgrowth are l e g i~n ,~ virtually all the effects have been demonstrated in studies (including the papers cited by Fahn2.3) using immature healthy astroglia. Far less is known regarding the behavior (or possible misbehavior) of "stressed" astroglia in the senescent and diseased nervous system. For example, although adenosine triphosphate-dependent uptake of extracellular glutamate by healthy astroglia is an "asset," failure of this glial mechanism in ischemic brain may constitute a serious "liability" predisposing to excitotoxic neuronal i n j~r y .~ Similarly, our laboratory and others have shown that progressive mitochondria1 damage, sequestration of redox-active transition metals, and augmentation of H,O,-generating monoamine oxidase-B activity in subpopulations of astroglia may render the senescent nervous system prone to oxidative injury, whereas astrocytes in young healthy brain exhibit properties that are purely adaptive in nature. Regarding the potential for catecholamine toxicity, we have demonstrated that (1) ferrous iron deposited in senescent astroglial mitochondria behaves as a nonenzymatic peroxidase activity capable of oxidizing dopamine and other catechols to potentially neurotoxic ortho-semiquinone radicals6 and (2) catecholamine-secreting PC 12 cells grown atop monolayers of astroglia bearing this senescent phenotype are far more susceptible to dopamine/H,O,-related killing than PCI 2 cells cocultured with immature control astroglia.' (A similar argument can be made for microglial cells which may subserve adaptive immunomodulatory functions in normal brain but release cytotoxic quantities of superoxide, nitric oxide, and tumor necrosis factor under pathological conditions.x) Our understanding of the precise role(s) that neuroglia play in promoting or ameliorating neuronal degeneration in Parkinson's disease and other aging-related neurodegenerative disorders is in its infancy. Further delineation of these glial functions may implicate these cells as important targets for therapeutic intervent...