Discovery of a metabolic signature predisposing high risk patients with mild cognitive impairment to converting to alzheimer’s disease

Yi Long Huang, Chao Hsiung Lin, Tsung Hsien Tsai, Chen Hua Huang, Jie Ling Li, Liang Kung Chen, Chun Hsien Li, Ting Fen Tsai*, Pei Ning Wang

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

Assessing dementia conversion in patients with mild cognitive impairment (MCI) remains challenging owing to pathological heterogeneity. While many MCI patients ultimately proceed to Alzheimer’s disease (AD), a subset of patients remain stable for various times. Our aim was to characterize the plasma metabolites of nineteen MCI patients proceeding to AD (P-MCI) and twentynine stable MCI (S-MCI) patients by untargeted metabolomics profiling. Alterations in the plasma metabolites between the P-MCI and S-MCI groups, as well as between the P-MCI and AD groups, were compared over the observation period. With the help of machine learning-based stratification, a 20-metabolite signature panel was identified that was associated with the presence and progression of AD. Furthermore, when the metabolic signature panel was used for classification of the three patient groups, this gave an accuracy of 73.5% using the panel. Moreover, when specifically classifying the P-MCI and S-MCI subjects, a fivefold cross-validation accuracy of 80.3% was obtained using the random forest model. Importantly, indole-3-propionic acid, a bacteria-generated metabolite from tryptophan, was identified as a predictor of AD progression, suggesting a role for gut microbiota in AD pathophysiology. Our study establishes a metabolite panel to assist in the stratification of MCI patients and to predict conversion to AD.

Original languageEnglish
Article number10903
JournalInternational Journal Of Molecular Sciences
Volume22
Issue number20
DOIs
StatePublished - 1 Oct 2021

Keywords

  • Alzheimer’s disease
  • Feature selection
  • Mild cognitive impairment
  • Plasma
  • Untargeted metabolomics

Fingerprint

Dive into the research topics of 'Discovery of a metabolic signature predisposing high risk patients with mild cognitive impairment to converting to alzheimer’s disease'. Together they form a unique fingerprint.

Cite this