| Summary: | The prototypical central nervous system (CNS) inflammatory demyelinating disease (IDD) is multiple
sclerosis (MS), which is a complex, immune-mediated disease involving genetic and environmental
factors, both contributing and interacting, to produce the underlying immunopathogenesis. The
inherent pathogenic heterogeneity within MS makes it challenging for a single candidate biofluid
marker to reflect all the underlying disease processes, both neuroinflammatory and
neurodegenerative, at the same time. Thus, the combination of biofluid markers into a composite
biomarker, i.e. a biomarker signature, is a feasible approach to provide a summative representation
of various disease processes.
<br></br>
Metabolomics is well-placed for this type of summative approach as a single metabolomics
experiment provides hundreds (if not thousands) of metabolites at one go, i.e. a metabolic snapshot,
which represents the downstream products of all neuroinflammatory and degenerative processes.
Against this background, this thesis explores whether the combination of 1H nuclear magnetic
resonance (NMR) spectroscopy-based metabolomics with multivariate dimensionality reduction
modelling can be used to construct discriminatory models to inform on the diagnosis, disease
activity and progression of CNS IDD, in particular MS, with inference to the underlying
pathophysiological processes.
<br></br>
The results in this thesis demonstrated that it is possible to use blood metabolomics to provide
molecular validation to phenotypic clusters, derived from pattern recognition modelling, within
antibody-negative CNS IDD patients, and in so doing, identified individuals who likely have MS and
those with antibody-mediated CNS pathology. In a focal rat model of MS, both the blood and CSF
metabolome were significantly perturbed during acute neuroinflammation and these changes
evolved over time. As the model does not produce overt clinical signs, these metabolic signatures,
and indeed individual metabolites, provide pathogenic insights that are solely attributable to MS-like
lesion formation. Drawing from these observations, the same methodology was applied to MS
patients who are in relapse. Using serum metabolomics, it is possible to differentiate patients in
relapse from patients who have no evidence of clinical inflammatory activity. Similar to that
observed in rodents, the metabolic differences decreased with time away from relapse and specific
metabolite biomarkers could be identified. These were then found to be applicable in an
individualised manner, differentiating between relapse and remission status. The results reported in
this thesis also illustrate that serum metabolomics was able to distinguish between relapsing-remitting
MS (RRMS) from its secondary progressive (SPMS) phase, validating the previous work
done by the Oxford group. In addition to the patient-based studies, this thesis also describes studies
performed to validate the methodologies used in relation to sample-handling variations that are
commonly encountered in the clinic. It was found that the accuracy of this RRMS vs. SPMS
metabolomics test was maintained despite these variations, supporting its clinical applicability.
<br></br>
In summary, these results show that metabolomics has substantial potential in precision medicine in
CNS IDD by: (1) aiding in CNS IDD diagnosis, (2) providing early detection of active
neuroinflammation, and (3) allowing for the objective definition of SPMS.
|
|---|