Diffusion tensor imaging in a rat model of Parkinson's disease after lesioning of the nigrostriatal tract.

Publication Type:

Journal Article

Source:

NMR in biomedicine, Volume 22, Issue 7, p.697-706 (2009)

Keywords:

Animals, Behavior, Animal, Diffusion Magnetic Resonance Imaging, Disease Models, Animal, Female, Imaging, Three-Dimensional, Immunohistochemistry, Parkinson Disease, Positron-Emission Tomography, Rats, Rats, Wistar, Substantia Nigra

Abstract:

Parkinson's disease (PD) is characterised by degeneration of the nigrostrial connection causing dramatic changes in the dopaminergic pathway underlying clinical pathology. Till now, no MRI tools were available to follow up any specific PD-related neurodegeneration. However, recently, diffusion tensor imaging (DTI) has received considerable attention as a new and potential in vivo diagnostic tool for various neurodegenerative diseases. To assess this in PD, we performed DTI in the acute 6-hydroxydopamine (6-OHDA) rat model of PD to evaluate diffusion properties in the degenerating nigrostriatal pathway and its connecting structures. Injection of a neurotoxin in the striatum causes retrograde neurodegeneration of the nigrostriatal tract, and selective degeneration of nigral neurons. The advantage of this model is that the lesion size is well controllable by the injected dose of the toxin. The degree of functional impairment was evaluated in vivo using the amphetamine rotation test and microPET imaging of the dopamine transporter (DAT). Despite a nearly complete lesion of the nigrostriatal tract, DTI changes were limited to the ipsilateral substantia nigra (SN). In this study we demonstrate, using voxel-based statistics (VBS), an increase in fractional anisotropy (FA), whereas all eigenvalues were significantly decreased. VBS enabled us to visualise neurodegeneration of a cluster of neurons but failed to detect degeneration of more diffuse microstructures such as the nigrostriatal fibres or the dopaminergic endings in the striatum. VBS without a priori information proved to be better than manual segmentation of brain structures as it does not suffer from volume averaging and is not susceptible to erroneous segmentations of brain regions that show very little contrast on MRI images such as SN.