Use this URL to cite or link to this record in EThOS:
Title: Alzheimer's disease biomarkers discovery using metabolomics approach
Author: Kim, Min Gyu
ISNI:       0000 0004 7232 0767
Awarding Body: King's College London
Current Institution: King's College London (University of London)
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
Alzheimer's disease (AD) is a major debilitating disease with both cognition and independence gradually lost over time. AD can be diagnosed post-mortem when the accumulation of amyloid-β (Aβ) peptides and tau proteins in the brain are visible. The best diagnostic markers at present are cerebrospinal fluid (CSF) Aβ and tau levels. However, the diagnostic accuracies of CSF markers are limited and all clinical trials targeting Aβ and tau pathologies have failed to show promising results. Therefore, there is an urgent need for new molecular leads associated with AD which can potentially be utilised as disease-modifying therapy agents and as AD biomarkers. This thesis investigates the possibility of using metabolomics to discover AD associated metabolites. Initially in chapter two, CSF samples (20 AD, 20 age-matched healthy controls) were analysed by Nuclear Magnetic Resonance (NMR). Results showed that metabolic fingerprints were non-differentiable between the two cohorts. One metabolite was found decreased, pyruvate, suggesting energetic hypometabolism in AD (p < 0.05). Acetate levels were increased in the AD group (p < 0.05), suggesting acetate being used as an alternative energy source to pyruvate. In the next chapter, in order to scan for more metabolites, Liquid Chromatography- Mass Spectrometry (LC-MS) was applied on the same cohorts. The experiments allowed for the detection of 4426 metabolic features. Although the fingerprints were non-differentiable between the cohorts, 2 unidentified metabolic features were found to be able to discriminate AD from age-matched controls with a Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) value of 70%. After this, work focused in blood metabolites because they are easily accessible. In chapter four we wanted to test whether lipids that had been previously implicated with AD would reproduce with similar trends. Semi-targeted analysis of 9 lipids was carried out (3 phosphatidylcholines (PC) and 6 ceramides) in AD (n=205) and Controls (n=207). Elevated levels of three ceramides and diminished levels of 2 PC molecules were found in AD blood (p < 0.05), one PC associated with hippocampal atrophy. In the following chapter, these 9 lipids along with cholesterol and absolute cholesteryl esters levels were measured on pre-conversion serum samples (112 eventual AD participants and 113 Control participants at 3 time-points). At timepoints 1 and 2 (pre-clinical stage), no lipids showed significant differences between the pre-converters and the stable cohorts. One PC and total cholesteryl ester levels were diminished at the symptomatic stage, time point 3 (p < 0.01). When using lipidomics, the whole fingerprints and 1207 metabolic features, one feature (annotated as a PC) was found to be significantly different at symptomatic stage (q < 0.05). Finally in chapter six, untargeted lipidomic analysis was performed on 148 AD and 152 control plasma samples in search of AD blood biomarkers. Application of random forest showed a combination of 24 lipid features consisting of cholesteryl ester, triglycerides and phosphatdiylcholines being able to discriminate the cohorts with AD classification accuracy of > 70%. In addition, other lipid signatures were found to predict disease progression (R2=0.10) and brain atrophy in all brain regions except for left entorhinal cortex (R2 > 0.14).
Supervisor: Mason, Andrew James ; Legido Quigley, Cristina Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available