Abstract
Glial-derived neurotrophic factor (GDNF), a transforming growth factor (TGF)-β family member, is a potent neurotrophic factor that is crucial for the outgrowth and survival of dopaminergic neurons (Sauer et al., 1995), making it a good candidate for protecting existing neurons against environmental insults and restoring function in deteriorating affected neurons. However, GDNF is rapidly degraded in the human body, and does not cross the blood–brain barrier efficiently (Deierborg et al., 2008); therefore, the only effective mode of delivery is directly into the desired brain area. Initial clinical trials delivering bolus injections of GDNF into the lateral ventricle resulted in no clinical improvement, severe side effects (Nutt et al., 2003), and in some patients the development of antibodies against GDNF was observed. The main conclusion was that a more effective delivery method was required and, until then, all clinical trials involving infusion of GDNF have been withdrawn.
The gene therapy approach seems to be an adequate alternative, that it would enable the delivery of GDNF-encoding expression vectors in a relatively small volume by a procedure permitting a prolonged effect. Given that the majority of the cells in the CNS are postmitotic, this approach requires vectors that are capable of transfecting quiescent cells. Lentivirus (LV) and adeno-associated virus (AAV) have been successfully tested in producing continuous expression of GDNF in a rat model of PD when injected into the striatum (AAV) (Kirik et al., 2000) and into the substantia nigra (LV) (Bensadoun et al., 2000), and LV-GDNF was shown to be effective in the reversal of symptoms in rhesus monkeys in which PD was induced by the neurotoxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (Kordower et al., 2000).
In this context, two papers in this issue from the group of Bankievicz, Eberling et al. (2009) and Johnston et al. (2009), illustrate the use of the AAV2-GDNF vector, delivered by a convection-enhanced method into the putamens of two different rhesus monkey models. They report clinical improvement and show evidence of functional recovery in the nigrostriatal pathway, absence of adverse effects, and prolonged GDNF expression over a 12-month period in MPTP-lesioned parkinsonian monkeys. Similarly, AAV2-GDNF was introduced into aged monkeys that exhibited naturally occurring slowing of motor activity, resulting in increased uptake of the dopamine analog and enhanced locomotor activity. The density of dopaminergic terminals in the putamen, but also the number of nonpigmented neurons in the SN, increased upon delivery of high-dose vector particles, with no signs of adverse effects.
The lack of appropriate animal models that reflect accurately the disease is a significant obstacle. The use of neurotoxins (e.g., MPTP and 6-hydroxydopamine [6-OHDA]) can result in partial or complete lesion and acute loss of dopamine neurons and thus recapitulate some of the symptomatic features, but does not mimic the true etiology of PD. In this respect, the study of GDNF delivery in aged monkeys is valuable in demonstrating the beneficial effects in a condition quite similar to that observed in aged humans.
In both papers, a unique convection-enhanced delivery (CED) method was used. This method, which was initially developed by Oldfield and colleagues (1994) for delivery of macromolecules into the brain, has been extensively used by this group for the robust delivery of viral vector particles in a large volume. CED has a clear advantage over methods such as magnetic resonance imaging (MRI)-assisted microinjection, which is prone to reflux. More importantly, it enables diffusion over a broad area while preventing the formation of “hot spots”—focal concentration of dopaminergic activity that may account for the side effects of dyskinesias. The authors reported that the administration of high-volume infusate was well tolerated in aged monkeys and provided useful data for prospective clinical trials.
The accumulating data from preclinical large-animal intraparenchymal rAAV administration seem sufficient to proceed with phase I studies. rAAV2 vector administration to the human CNS appears well tolerated (McPhee et al., 2006), although low levels of immune response to AAV2 were detected in subjects, suggesting that this approach is relatively safe.
The presented results from both studies are encouraging and provide important data to support progression into clinical trials. Yet, several issues still remain to be addressed. Although GDNF was meant to augment dopaminergic activity, the possibility that other nondopaminergic neurons as well as non-neuronal cells were also transduced is plausible. In fact, the authors report that the treatment significantly increased the levels of 5-HT (serotonin) and norepinephrine in the putamen and the SN (Johnston et al., 2009). It is not clear whether long-term GDNF-induced boosting of these neurotransmitter levels is beneficial or might become a source of complications. Another question still pending concerns the long-term effect of unregulated GDNF overexpression. The papers indicate that GDNF vector distribution is not limited to the site of delivery but rather occupies a large tissue volume (Eberling et al., 2009; Johnston et al., 2009). This raises another concern that viral particles might find their way via the cerebrospinal fluid to remote locations, and elicit ectopic GDNF expression. Most importantly, the finding of an increased number of nonpigmented neurons in the SN raises the question about the possible origin of these cells. Neural stem cells in the adult brain are quite rare. Do these cells originate from pools of neural stem cells that were induced to proliferate and migrate into the SN? Were spots of active proliferation seen anywhere in the AAV2-GDNF-treated brains? Can the possibility that cell proliferation was induced by the viral vector itself be excluded? Perhaps control animals treated with empty vector could give some clue about vector-related effects. The results of the reports in this issue of Human Gene Therapy spark enthusiasm and could pave the way for clinical assessment of rAAV2-GDNF, and yet, adverse effects should be carefully considered. It was reported that bilateral continuous hypothalamic GDNF overexpression via rAAV could result in a significant failure to gain weight in young rats and induce weight loss in aged rats. This metabolic effect may possibly be attributed to the activation of hypothalamic corticotrophin hormone releasing neurons. While conducting a dose escalation phase I study, measurable related end-points should be monitored. Thus, preclinical studies such as the ones reported in this issue are extremely important in generating sufficient empirical data to set the ground for an intrepid big leap in regenerative treatment for PD.