Metabolic signature of insulin resistance and risk of Alzheimer's disease.

Background: Substantial evidence supports the relationship between peripheral insulin resistance (IR) and the development of Alzheimer's disease (AD)-dementia. However, the mechanisms explaining these associations are only partly understood. We aimed to identify a metabolic signature of IR associated with the progression from mild cognitive impairment (MCI) to AD-dementia.

Methods: This is a case-control study on 400 MCI subjects, free of type 2 diabetes, within the ACE cohort, including individuals ATN+ and ATN-. After a median of 2.1 years follow-up, 142 subjects converted to AD-dementia. IR was assessed using the HOMA-IR. A targeted multi-platform approach profiled over 600 plasma metabolites. Elastic net penalized linear regression with 10-fold cross-validation was employed to select those metabolites associated with HOMA-IR. The prediction ability of the signature was assessed using support vector machine and performance metrics. The metabolic signature was associated with AD-dementia risk using a multivariable Cox regression model. Using counterfactual-based mediation analysis we investigated the mediation role of the metabolic signature between HOMA-IR and AD-dementia. The metabolic pathways in which the metabolites were involved were identified using MetaboAnalyst.

Results: The metabolic signature comprised 18 metabolites correlated with HOMA-IR. After adjustments by confounders, the signature was associated with increased AD-dementia risk (HR 1.234; 95%CI 1.019-1.494; p<0.05). The metabolic signature mediated 35% of the total effect of HOMA-IR on AD-dementia risk. Significant metabolic pathways were related to glycerophospholipid and tyrosine metabolism.

Conclusions: We have identified a blood-based metabolic signature that reflects IR and may enhance our understanding of the biological mechanisms through which IR affects AD-dementia.