PD is ultimately associated with gradual loss of dopaminergic neurons (DAN) in the substantia nigra pars compacta. These neurons provide dopaminergic input to the corpus striatum. The decrease in production and release of dopamine (DA) results in abnormal motor function. Only 10% of PD patients show direct heritability of or family linkage to PD; the great majority of cases are thus sporadic and idiopathic. Reasons for the loss of DAN, in addition to the few defined genetic mutations and the gradual loss associated with age, include environmental exposures to toxins. Currently the most effective pharmacologic treatment for PD, replacement of DA via precursor (levodopa) administration, is temporarily able to replenish DA levels and improve the motor symptoms of PD, but does not slow the progressive loss of DAN that accounts for the deficiency in the first place. A potentially more effective therapeutic approach would be to target the pathogenetic mechanisms that lead to neurodegeneration, and thereby support retention and replacement of DAN (Figure 1). Neuron loss silently accumulates before clinical signs become evident; therapy should therefore start as early as possible in confirmed PD cases to retain existing DAN. On the other hand, PD is a gradually developing disease, and the potential window of treatment appears to be large. Symptoms develop when approximately 50% of DAN are still present; neuroprotective therapy at this point should have a good chance of restoring the DA level to above the symptomatic threshold. Support for missing DAN function may be provided directly, by replacing DAN product, or indirectly, by stimulating astroglial or microglial cells to secrete factors that support protection and regeneration of DAN or by inhibiting inflammatory and oxidative damage pathways that lead to their demise (Figure 1).

Some forms of gene therapy (GT), the introduction of heterologous functional genes, target neurotransmitter synthesis, with the goal of relieving motor symptoms of PD, whereas another approach targets the underlying mechanism of DAN loss, with the goal of arresting or reversing progression of the disease. This second approach would be particularly applicable for some cases of familial PD, where delivery and consequent expression of a wild-type version of a recessive mutant gene would be reasonable. A direct test of this notion in a rodent model of PD, however, showed that although viral transduction and expression of the E3 ligase parkin (PARK2) reestablished the nerve signaling pathways and increased production of tyrosine hydroxylase (TH) and DA level, it did not serve to protect the nigrostriatal DA system against the toxic challenge that induced the PD-like condition.^1,2

GT as a means of treating PD has in fact been studied for years, initially in a variety of animal models,^3,4 and most recently in several phase I and II trials in humans. Rather than the simple expression of missing genes, the current approaches to GT for PD⁵ take advantage of the wealth of knowledge of the CNS circuitry relevant to PD, the anatomically restricted region of involvement, DAN biology, and the technical development of ways to selectively target heterologous genes to DAN and surrounding cells and have the genes expressed appropriately.^6,7 Viruses with neuronal tropism, such as herpes simplex (HSV-1), lentivirus (LV), adenovirus (AdV), and adeno-associated viruses (AAV), have all been developed as vectors for delivering genes to CNS tissues (transfection).³ Replication- and integration-deficient forms of these viruses can transfect neurons in regions of the brain into which they are injected and, by way of tissue-specific promoters, then direct the expression of heterologous genes that they carry into these neurons. Of the various viral vectors, AAV2 has received the most analysis in rodent and primate models.^8-10 Consistent with these animal studies, which have shown no evidence of direct toxic effects of AAV, no immune responses giving rise to inflammation or neuronal damage, and no chromosomal integration of viral sequences leading to oncogene generation, AAV2 has so far been associated with no serious adverse events in the human trials discussed below.

A variation of direct administration of GT virus expression vectors has been explored using a rodent model of PD.^11,12 In this case, an AdV expression virus encoding a differentiation-inducing transcription factor, NURR1, was used to transfect neural stem cells (NSC). The AdV–NURR1-treated stem cells differentiated into DAN-like neurons at a greater rate than nontransfected cells in vitro. When the NSC were then implanted into rats lesioned by 6-hydroxydopamine (6-OHDA), those receiving nontransfected and those receiving transfected stem cells both showed improvement in apomorphine-induced rotation rate, but the recipients of the transfected cells showed greater improvement.

Three Phase I trials (www.clinicaltrials.gov) of direct administration of recombinant expression viruses either are ongoing or have been completed in humans, each involving the transduction of a distinct gene and targeting a different aspect of PD etiology.¹³ The first human GT trial for PD was designed to mimic a common surgical therapy for advanced PD, electrical stimulation of the subthalamic nucleus (STN). The motor abnormalities associated with PD result in part from lack of DA in the striatum, allowing disinhibition of the STN and excess activity of output nuclei. Luo et al used an animal model to test whether glutamatergic neurons of the STN could be converted from excitatory to inhibitory output by overexpression of glutamic acid decarboxylase (GAD), which not only degrades glutamate but is the rate-limiting enzyme for synthesis of the inhibitory neurotransmitter GABA.¹⁴ An AAV–GAD65 expression construct was transduced into neurons in the STN, resulting in a 4.0±1.5-fold increase in GABA release upon electrical stimulation, accompanied by marked increases in inhibitory responses in the substantia nigra pars reticulata. The increased inhibitory response also served to protect DA neurons from neurotoxic lesioning with 6-OHDA and allowed retention of the majority of the locomotor activity that was lost in lesioned but nontransduced animals. This treatment rationale was applied in a phase I study of 12 PD patients who received unilateral delivery of AAV–GAD and were followed for 12 months postsurgery.¹⁵ Although not designed to test efficacy, the study reported significant improvements in the UPDRS [Unified Parkinson’s Disease Rating Scale] and in activities of daily living at 6 and 12 months, with no significant adverse effects. More complete efficacy determinations await results of a phase II trial that will employ bilateral administration of AAV–GAD.