While the germline transgenic approach is the gold standard for studying gene function in the brain, it is also labor-intensive, expensive, and requires long periods of time to screen and raise the animals. We have adopted an alternative somatic transgenic strategy (historically applied for gene therapy) where the mutant genes can be expressed directly in the brains of adult animals, a method that by comparison, is inexpensive and fast.

The application of a somatic-cell transgenic approach to modeling neurodegenerative diseases will be useful for several reasons, including: (1) temporal control of expression to avoid developmental effects and also to study aging effects, i.e. similar periods of expression in young or aged subjects; (2) spatial control of expression to target brain regions associated with specific disease states; and (3) transgene combinations.

Further, the vector-based approaches can be applied to rats or monkeys and also take advantage of internal controls, i.e. comparisons to the contralateral, untreated hemisphere. Despite some real limitations of this...  Read more

While the germline transgenic approach is the gold standard for studying gene function in the brain, it is also labor-intensive, expensive, and requires long periods of time to screen and raise the animals. We have adopted an alternative somatic transgenic strategy (historically applied for gene therapy) where the mutant genes can be expressed directly in the brains of adult animals, a method that by comparison, is inexpensive and fast.

The application of a somatic-cell transgenic approach to modeling neurodegenerative diseases will be useful for several reasons, including: (1) temporal control of expression to avoid developmental effects and also to study aging effects, i.e. similar periods of expression in young or aged subjects; (2) spatial control of expression to target brain regions associated with specific disease states; and (3) transgene combinations.

Further, the vector-based approaches can be applied to rats or monkeys and also take advantage of internal controls, i.e. comparisons to the contralateral, untreated hemisphere. Despite some real limitations of this approach, such as the volume of brain that can be affected and the size of the DNA that can be incorporated, we predict that somatic transgenic models of neurodegenerative diseases will be useful to study the disease process.

Somatic gene transfer offers several specific advantages for neurodegenerative disease-modeling. The neurodegenerative diseases are typically associated with specific parts of the brain and are age-related. The spatio-temporal control of expression that is possible with the somatic gene transfer is therefore a powerful way to model these features of the disease process. Further, these diseases are often multi-factorial, and there is greater facility to combine genetic manipulations in the somatic transgenic approach compared to germ?line transgenic mice.

It is unclear why the AAV vector system was so effective in our study, although it could be due to its ability to express high levels of the gene product selectively in this brain region, and perhaps also to the application to adults, avoiding some form of developmental compensation. Due to the complex nature of these models and hypotheses tested, the similar results obtained from Dr. Bjorklund and colleagues, a foremost laboratory in Parkinson's disease research, are an important confirmation of our work and the significance of the approach.

We are currently studying other factors linked to Parkinson's disease that may block or exacerbate the cell loss, and hope that these approaches will help us understand why dopamine neurons are vulnerable during Parkinson's disease, as well as lead toward novel therapies, including gene therapy.

View all comments by Ronald Klein

I believe the most important aspect of our study is that we describe a new approach to study protein dysfunction in appropriate animal models with a strong basis in morphological, biochemical as well as behavioral assessments. Our model offers numerous advantages over transgenic mouse models. For example, it can easily be generated using wild-type animals in sufficient numbers with a very simple surgical intervention; the animals can be rendered transgenic at any time during their lifetime; it can be applied unilaterally leaving the contralateral side as an internal control; detailed functional assessments can better be done using rats as compared with mice; the model can be applied to primates to answer certain question that cannot be satisfactorily addressed in rodents.

The method is very reproducible. We have now done (including new experiments not included in this paper) over 500 surgeries, and in all cases we can hit nearly all of the nigral dopamine cells with very high precision. This model is unique because the AAV vectors have a high affinity to these cells.

As...  Read more

I believe the most important aspect of our study is that we describe a new approach to study protein dysfunction in appropriate animal models with a strong basis in morphological, biochemical as well as behavioral assessments. Our model offers numerous advantages over transgenic mouse models. For example, it can easily be generated using wild-type animals in sufficient numbers with a very simple surgical intervention; the animals can be rendered transgenic at any time during their lifetime; it can be applied unilaterally leaving the contralateral side as an internal control; detailed functional assessments can better be done using rats as compared with mice; the model can be applied to primates to answer certain question that cannot be satisfactorily addressed in rodents.

The method is very reproducible. We have now done (including new experiments not included in this paper) over 500 surgeries, and in all cases we can hit nearly all of the nigral dopamine cells with very high precision. This model is unique because the AAV vectors have a high affinity to these cells.

As to the applicability of these methods to Alzheimer's research, to my knowledge there are no data with AD genes such as AβPP, presenilins, or tau. We are very interested in doing such work and have recently initiated efforts in this direction.

The vectors Dr. Klein and colleagues used were prepared with essentially the same procedures as ours, except that they used the A30P mutant form of the human gene, whereas we used A53T mutation and the wild-type human genes. I should mention, however, that our interpretation of these data and some others (yet unpublished) is that the mutation does not seem to augment the pathology. Note also that wild-type human a-synuclein is different from the wild-type rat α-synuclein.

Klein and colleagues report a similar cell loss of the TH+ cells in the substantia nigra as in our study, but their morphological analysis of the data is limited to this. No studies have been included to look at biochemical changes or the time course of the degeneration, and no tracing studies were performed. Documentation of α-synuclein-induced pathology is superficial; the descriptions of Lewy-like axonal pathologies are based solely on marker protein GFP.

Klein and colleagues have concluded that, in their animals, they did not see behavioral impairments. This can easily be explained in two ways: First, the behavioral testing paradigms applied in their study are insufficient to pick up impairments in animals with mild-to-moderate degeneration in the ascending dopamine system. Second, they have tested a very small group of animals in their study. This dramatically reduces the power of their analysis.

We observed that about 25 percent of the animals manifest behavioral impairments, while the majority seems to have compensated for their neuron loss. The impairments could, however, be revealed by use of sub-threshold blockade of TH enzyme in the seemingly normal animals. If we come back to the Klein experiment, it is natural for them to conclude that there was no apparent behavioral deficit in their animals as they have looked at only 6-8 of them. So, briefly, I believe that their data is essentially supporting and replicating a subset of ours, but the difference is that it has not been analyzed as carefully and with appropriate detail.

View all comments by Deniz Kirik

Use of viral vector for targeting transgene expression in a region- and time-specific manner was first demonstrated to model a neurodegenerative disorder, namely trinucleotide repeat disease, by Fred Gage's lab (Senut et al., 2000). This approach offers a unique opportunity for the development and modeling of neurodegenerative disorders in a rapid and reliable manner. It also offers a unique opportunity of testing in vivo selective neuronal vulnerability and the differential effects of mutations.

A similar approach is currently under development by our group in collaboration with Fred Gage and Inder Verma to model Alzheimer's disease and other neurodegenerative disorders using lentiviral vectors. This approach also holds great promise for the development of new treatments for PD.

View all comments by Eliezer Masliah