Abstract
In this thesis, Scrapie-infected tissue was investigated by
Fourier-transform infrared (FTIR) spectroscopy, a vibrational
spectroscopic method. The aim of the work was to gain more information
on molecular aspects of TSE pathogenesis. The study concentrated on
important anatomical structures of the brain stem and cerebellum, as
well as on the investigation of dorsal root ganglia of 263K
scrapie-infected Syrian hamsters (Mesocricetus auratus). IR spectra
were acquired from cryotomized tissue sections by IR
microspectroscopic mapping. Adjacent tissue sections were stained for
comparison using immunohistochemical and standard histological
procedures. To compare identical anatomical structures in
scrapie-infected and control brains, new methods suited for the
analysis of large amounts of data and for the identification of
specific anatomical structures were developed and optimized.
Anatomical structures were identified by applying a combination of
univariate and multivariate imaging methods. Based on overview maps
which were acquired using a lateral resolution of 100µm and
constructed using the protein/lipid ratio, areas containing the
hypoglossal nucleus (HypN), the dorsal motor nucleus of the vagus
nerve (DMNV) and parts of the solitary tract nucleus (SolN) in the
medulla oblongata, as well as an area containing the interposed
cerebellar nucleus (IntN) were identified and examined in depth in
detailed measurements applying a resolution of 50µm. Spectral
classification according to the known histological structures was
achieved by employing the spectral information over the fingerprint
spectral region of 1480-950cm-1 in hierarchical cluster
analyses and pattern-based image reconstruction. Based on the results
of these cluster analyses, spectra of DMNV/SolN, HypN and IntN were
extracted from the data sets and used for systematic comparison
between identical structures in scrapie-infected and control
tissue.
The spectra obtained from these structures in infected and control
animals were compared at three stages of scrapie (90 d.p.i., 120
d.p.i., and at the terminal stage). At terminal stage, differences
between spectra were found for all investigated structures throughout
the whole spectral range. At 120 d.p.i. and 90 d.p.i. changes were
confined to the spectral region 1300-1000cm-1. The spectral
alterations reflected complex, fingerprint-like changes of membrane
components, carbohydrates, nucleic acids, and proteins. They differed
qualitatively and quantitatively between the different stages of
infection and brain structures. It could be shown that the data were
in good agreement with results from immunocytochemical investigations
of scrapie pathogenesis. The majority of spectra from diseased and
control tissue could be clearly separated by cluster analysis in the
terminal stage and at 120 d.p.i. Application of artificial neural
networks based on feature selection by variance analysis yielded an
identification accuracy of 80% at 90 d.p.i., indicating that based on
molecular changes scrapie-infected brain tissue can be identified even
in the pre-clinical stage. In a pilot study, high quality IR spectra
could be collected from single neurons of dorsal root ganglia. The
spatial resolution could be increased to the diffraction limit by
using a synchrotron source. Very small, localized areas with a high
concentration of ß-sheet secondary structures were detected in the
scrapie-affected dorsal root ganglia by this method, hinting at the
presence of PrPSc aggregates.
In this work, for the first time infrared (IR) microspectroscopy
was successfully applied to the investigation of molecular aspects of
transmissible spongiform encephalopathies. The microspectroscopic
approach opens up the opportunity of direct and selective
investigation of specific structures or substructures in the tissue.
The knowledge about scrapie-specific molecular alterations obtained
from the spectroscopic studies could provide a basis for the
development of new methods for the rapid post mortem identification of
infectious tissue. |