I understand that further animal toxicity studies are in progress. However, over 100 patients have received intracerebral GDNF infusion by one route or another with no clinical toxicity and I can't believe that GDNF treatment won't be available again, at least in some form, in the medium term. To date, the immunogenicity of the recombinant GDNF has not proven to be of clinical significance, but could in any case be circumvented by implantation of autologous or encapsulated eukaryotic cells, genetically modified to secrete GDNF. Stimulating metabolic pathways that induce the synthesis of GDNF sounds attractive but poses problems of targeting, delivery, and specificity. I suggest that this is a less promising option, but would be happy to be proven wrong.

View all comments by Seth Love

The report by Love and colleagues in Nature Medicine provides intriguing preliminary evidence for a biological effect of GDNF in humans with Parkinson disease. The greater area of staining for tyrosine hydroxylase in the striatum on the side previously most affected by Parkinson disease suggests that GDNF stimulated neuronal sprouting and that this accounted for the increase in fluorodopa uptake seen on positron emission tomography. These observations are exciting but leave many unanswered questions. Is the change in striatal tyrosine hydroxylase and fluorodopa PET sufficient to account for the 38 percent reduction (i.e., improvement) in motor scores? Even more impressive changes in both of these parameters are seen following fetal nigral transplantation, but clinical benefit has been disappointing in double-blind placebo-controlled trials. Love and colleagues present additional results of GFAP and GAP43 immunohistochemistry; however, similar control data from normal and untreated parkinsonian brains were not provided for comparison. Finally, although the results are potentially...  Read more

The report by Love and colleagues in Nature Medicine provides intriguing preliminary evidence for a biological effect of GDNF in humans with Parkinson disease. The greater area of staining for tyrosine hydroxylase in the striatum on the side previously most affected by Parkinson disease suggests that GDNF stimulated neuronal sprouting and that this accounted for the increase in fluorodopa uptake seen on positron emission tomography. These observations are exciting but leave many unanswered questions. Is the change in striatal tyrosine hydroxylase and fluorodopa PET sufficient to account for the 38 percent reduction (i.e., improvement) in motor scores? Even more impressive changes in both of these parameters are seen following fetal nigral transplantation, but clinical benefit has been disappointing in double-blind placebo-controlled trials. Love and colleagues present additional results of GFAP and GAP43 immunohistochemistry; however, similar control data from normal and untreated parkinsonian brains were not provided for comparison. Finally, although the results are potentially important in demonstrating a biological effect of this treatment, they also raise questions about the early and bilaterally symmetrical clinical benefit reported following open-label unilateral infusion by Slevin et al., since Love’s patient showed very clear progression on the non-infused side.

The double-blind placebo-controlled trial failed to demonstrate significant efficacy of bilateral intraputamenal GDNF infusion. Importantly, this trial utilized a different catheter and somewhat different doses than were used in the patient reported by the Bristol group. It is not known whether these differences could account for the contrasting results of the open-label and double-blind studies. In addition, a similar change in fluorodopa PET to that originally reported by Gill and colleagues) was obtained in the double-blind trial despite the lack of benefit. During this trial and its open-label extension, 10 percent of patients developed blocking antibodies to GDNF, and subsequently, studies in primates demonstrated evidence for an unusual cerebellar toxicity. The clinical implications of these two findings for humans with Parkinson disease are unknown. In the face of a negative double-blind clinical trial and the development of these safety issues, Amgen chose to discontinue further use of this treatment in Parkinson disease. Subsequently, two patients from New York took the company to court demanding that continued treatment with GDNF be made available to them. The court found in favor of Amgen and the lawsuit was dismissed. Unfortunately, the current formulation of recombinant GDNF as manufactured by Amgen will probably not be used again in patients with Parkinson disease unless further basic studies can resolve these potentially important safety issues. It is hoped that other trophic factors or other methods of applying GDNF (e.g., gene therapy or cell-based therapies) will fulfill the promise of this approach in Parkinson disease. Finally, it should be emphasized that many of the problems we face in managing late-stage Parkinson disease do not stem from striatal dopamine deficiency, and therefore would not be expected to respond to even the most effective rejuvenation or replacement of the nigrostriatal dopamine system.

References:
Slevin JTGGA, Smith CD, Gash DM, Kryscio R, Young AB. Improvement of bilateral motor functions in patients with Parkinson disease through the unilateral intraputamenal infusion of glial cell line-derived neurotrophic factor. J Neurosurg. 2005 Feb;102(2):216-22. Abstract

View all comments by Anthony Lang

The findings reported by Dr. Love, while having some overlap with the tissue transplant studies, show significant differences, particularly with regard to neuronal resprouting and also to the involvement of the substantia nigra. It seems likely that GDNF does not generate new neurons but restores existing neurons and their extensive arborization. Since these existing neurons are capable of making postsynaptic connections, the dopamine can get where it needs to. Presumably, fetal tissue is unable to make synaptic connections, possibly because of a lack of signaling proteins like GDNF.

However, two enduring myths continue to surface.

First, that the double-blind trial of GDNF was "negative." It was not. While it is correct to say that it failed to meet its preset endpoints, nevertheless, unlike the tissue transplant study, there was a strong signal suggestive of drug efficacy, which is why Amgen continued to prepare for a proper phase III study even after announcing the phase II results.

Second, that there continue to be "safety issues," specifically regarding the...  Read more

The findings reported by Dr. Love, while having some overlap with the tissue transplant studies, show significant differences, particularly with regard to neuronal resprouting and also to the involvement of the substantia nigra. It seems likely that GDNF does not generate new neurons but restores existing neurons and their extensive arborization. Since these existing neurons are capable of making postsynaptic connections, the dopamine can get where it needs to. Presumably, fetal tissue is unable to make synaptic connections, possibly because of a lack of signaling proteins like GDNF.

However, two enduring myths continue to surface.

First, that the double-blind trial of GDNF was "negative." It was not. While it is correct to say that it failed to meet its preset endpoints, nevertheless, unlike the tissue transplant study, there was a strong signal suggestive of drug efficacy, which is why Amgen continued to prepare for a proper phase III study even after announcing the phase II results.

Second, that there continue to be "safety issues," specifically regarding the antibodies and the cerebellar lesions seen in four monkeys. There is now overwhelming evidence that the lesions were caused not by direct toxicity, but by abrupt withdrawal from very high concentrations of GDNF. Regarding the antibodies, these occur with the injection of just about any protein, and were expected in this study. Life-threatening complications are extremely rare. Indeed, we have it on good authority that senior officials at Amgen, who were in a position to speak authoritatively on this topic, did not consider them at all problematic.

Thus, the present formulation of GDNF is safe and also very probably effective.

View all comments by Michael Hutchinson