Abstract
Introduction
The biomarkers and the mechanisms underlying the cognitive decline in ageing are not fully clarified although they might be effective for dementia prevention. We comprehensively explore metabolites associated with cognitive function by conducting metabolomics analysis in a cohort study of general Japanese men.
Methods
The study population was 672 Japanese men aged 46-83 years without stroke, who were randomly selected for the SESSA study from Kusatsu City, Japan. Metabolomics using liquid chromatography mass spectrometry (LC-MS) was conducted on plasma samples, focusing on lipid metabolism. The 442 metabolites were identified. The cognitive function was assessed by Cognitive Abilities Screening Instrument (CASI), and CASI score less than 82 was defined as having mild cognitive impairment (MCI). We examined the metabolites among groups with and without MCI using principal component analysis and partial least squares regression (PLS) or PLS rank order of groups. Multivariable adjusted logistic regression estimated the odds ratios (ORs) and 95%CI of MCI for the extracted metabolites per standard deviation. The 9 domain specific MCI using median value of each domain specific CASI score as cutoff points were also examined. KEGG pathway analysis was done to clarify the linked pathway for the extracted metabolites.
Results
The prevalence of MCI was 5.5%. The 22 metabolites were statistically significantly associated with CASI score (q <0.05 in PLS, 21 metabolites in the positive direction and 1 metabolite in the negative direction). Phosphatidylcholines (PC) and Lysophosphatidylcholines (LPC) were associated with low risk of MCI (OR(95%CI): PC16:0/20:3; 0.51(0.32-0.51), LPC18:2/0.0; 0.59 (0.36-0.96), LPC20:3/0.0; 0.48(0.28-0.82)), phenacetylcarnitine was associated with the high risk of MCI (OR (95%CI); 1.33 (1.01-1.75)) (q <0.05). The analysis for domain specific MCI extracted the metabolites for short term memory, language abilities and category fluency. The Linoleic acid metabolism pathway was statistically significantly associated with CASI score.
Conclusion
Glycerophosphocolines such as PC16:0/20:3, LPC18:2/0.0 and LPC20:3/0.0 were reduced in participants with MCI, and those metabolites associated with low risk of MCI. Linoleic acid metabolism might be important for prevention of dementia among Japanese.
