Te n years ago, a glial cell line‐derived neurotrophic factor (GDNF) that has prominent actions on nigral dopaminergic neurons, both in vitro and in animal models of Parkinson disease (PD), was discovered. A recently published open-label clinical trial now reports that long-term intracerebral delivery of GDNF may also markedly improve symptoms in patients with PD. Here we review the experimental data underlying the current clinical trial and discuss the types of structural and functional changes induced by GDNF that may provide symptomatic benefit in PD patients. Data obtained in rodent and primate models of PD highlight the importance of how and where the factor is administered, supporting the view that GDNF has to be delivered locally in the brain parenchyma, at the receptive target site, to provide therapeutic benefit in PD patients. The cardinal symptoms of PD, including a difficulty in initiating movement, slowness of movement and stiffness and shaking at rest, are to a large extent caused by the progressive degeneration of the dopamine-producing neurons in the substantia nigra. Nigral cell loss proceeds over many years, during the early symptomatic stage, during manifest PD and during severe, end-stage disease. At the onset of disease, about 50% of dopaminergic neurons have been lost, and there is on average a further loss of 45% within the first decade, accompanied by a profound striatal dopaminergic denervation. It is this slow and protracted degenerative process that creates opportunities for disease intervention, such as blocking nigral cell loss and promoting recovery by improved function—and possibly by inducing regeneration and sprouting—of the surviving nigral dopaminergic neurons. Results obtained in animal models of PD indicate that GDNF may possess the desired properties to be used as a disease-modifying therapeutic factor for PD. Neurotrophic factors, by virtue of their neuroprotective properties, have attracted considerable interest as potential therapeutic agents in neurodegenerative diseases. Attempts to apply these factors clinically, however, have so far been disappointing because of their poor efficacy and induction of troublesome side effects. In these clinical trials, the recombinant protein was delivered either systemically or into the cerebrospinal fluid (intraventricularly or intrathecally) in patients suffering from amyotrophic lateral sclerosis, peripheral neuropathy, PD or Alzheimer disease 1,2 .R esults from these studies indicate that the neurotrophic factors, whose receptors are widely distributed, are prone to inducing pronounced side effects when delivered by these routes. The poor penetration across the blood‐brain barrier, as well as the limited passage of proteins from the cerebrospinal fluid into the brain tissue, has made it necessary to administer the factors at doses that are likely to induce side effects. These effects may not be so evident in small-sized experimental animals. For this reason, they may have gone unnoticed in the preclinical studies and may have become apparent in some cases only at the phase II/III stage of the clinical trails when larger numbers of patients were included. The therapeutic value of neurotrophic-factor delivery, therefore, may not be possible to achieve unless the factors are delivered locally at the receptive target sites within the central nervous system. Steven Gill and collaborators 3 have, for the first time, tested this mode of delivery in patients with advanced PD using continuous intracerebral infusion of GDNF. Although quite promising, the results of this initial open-label trial should be interpreted cautiously because the study was based on a small number of patients who were monitored over a relatively short follow-up period. Nevertheless, the data reported indicate that pronounced clinical benefit, in the absence of any serious side effects, may be possible to obtain by GDNF using intrastriatal delivery. Neuroprotective effects of GDNF in animal models of PD