Summary:After nearly 20 years of preclinical experimentation with various gene delivery approaches in animal models of Parkinson's disease (PD), clinical trials are finally underway. The risk/benefit ratio for these procedures is now generally considered acceptable under approved protocols. The current vehicle for gene delivery to the human brain is recombinant adeno-associated viral vector, which is nonpathogenic and non-self-amplifying. Candidate genes tested in PD patients encode 1) glutamic acid decarboxylase, which is injected into the subthalamic nucleus to catalyze biosynthesis of the inhibitory neurotransmitter ␥-aminobutyric acid and so essentially mimic deep brain stimulation of this nucleus; 2) aromatic Lamino acid decarboxylase, which converts L-dopa to dopamine; and 3) neurturin, a member of the glial cell line-derived neurotrophic factor family. Unraveling the genetic underpinnings of PD could allow gene therapy to go beyond modulating neurotransmission or providing trophic effects to dopaminergic neurons by delivering a specific missing or defective gene. For example, the parkin gene (PARK2) is linked to recessively inherited PD due to loss of function mutations; it prevents ␣-synuclein-induced degeneration of nigral dopaminergic neurons in rats and nonhuman primates. On the other hand, for dominantly inherited Huntington's disease (HD), in which an expanded polyglutamine tract imparts to the protein huntingtin a toxic gain of function, repressing expression of the mutant allele in the striatum using RNA interference technology mitigates pathology and delays the phenotype in a mouse model. Here we review the current state of preclinical and clinical gene therapy studies conducted in PD and HD.