Blood-based diagnosis of Alzheimer's disease using fingerprinting metabolomics based on hydrophilic interaction liquid chromatography with mass spectrometry and multivariate statistical analysis

• Evaluation of HILIC separation and repeatability for metabolomics approach.
• Challenge of blood based fingerprinting metabolomics for Alzheimer's disease.
• Multivariate statistical analysis of Alzheimer's postmortem plasma by HILIC/MS.
• Multivariate statistical evaluation for showing 62.5% using four increased MRM signals.

Early and definitive diagnosis of Alzheimer's disease (AD) can lead to a better and more-targeted treatment and/or prevention for patients. In the diagnostic biomarkers of AD, the blood sample represents a more non-invasive, inexpensive and acceptable sources for repeated measurements than the cerebrospinal fluid. In this study, the fingerprinting metabolomics was proposed for the challenge of the blood-based diagnosis of defined AD by hydrophilic interaction liquid chromatography mass spectrometry (HILIC/MS). These plasma samples were selected from postmortem specimens based on these pathological examinations. Firstly, we compared these HILIC columns for the non-targeted metabolic assay using pooled plasma. The principal component analysis plot of these seven columns was performed using the repeatability of these chromatograms, and can be used to visualize trends in data sets by three-dimensional dispersion, contributory standard deviation and the number of detections. Based on these results, TSK-Amide 80 and TSKgel-NH2 columns are used as a reliable HILIC/MS assay of blood-based AD metabolomics that showed metabolic profiling of the AD pathology in MS chromatograms that ranged from 1182 to 2284 compounds. A total of 54 peaks were evaluated in order to identify useful ion signal candidates using an orthogonal partial least-squares-discriminant analysis. These peaks were then specifically analyzed using the HILIC–tandem MS assay by a receiver operating characteristic curve and linear discriminant analysis for the diagnosis of the defined AD. The fingerprinting metabolomics can overcome the limitations of previous challenging blood-based diagnosis of AD, and directly evaluates the specific comparative statistical values from the raw data.