Peripheral serum metabolomic profiles inform central cognitive impairment

Area of Science:

• Neuroscience
• Metabolomics
• Biochemistry

Background:

• Alzheimer's disease (AD) incidence rises with age, posing a significant global health challenge.
• The underlying metabolic disturbances contributing to AD onset remain largely unknown.
• Identifying early metabolic markers is crucial for timely intervention and disease management.

Purpose of the Study:

• To identify specific metabolites and metabolic pathways associated with AD neuropathology and cognitive decline.
• To develop machine learning models for predicting cognitive impairment risk using metabolomic data.
• To assess the potential for early detection of cognitive impairment before clinical symptoms manifest.

Main Methods:

• Performed comprehensive metabolite profiling in human brain and matched serum samples.
• Analyzed abundances of 6 key metabolites (e.g., glycolithocholate) and the deoxycholate/cholate ratio.
• Assessed dysregulation scores for 3 metabolic pathways: primary bile acid biosynthesis, fatty acid biosynthesis, and unsaturated fatty acid biosynthesis.
• Utilized machine learning models (metabolite and pathway levels) to differentiate between cognitively impaired and unimpaired individuals.
• Applied models to a baseline control group to predict future cognitive decline.

Main Results:

• Significant differences in 6 metabolites and 3 metabolic pathways were observed between diagnostic groups (P < 0.05).
• Metabolite and pathway levels strongly correlated with cognitive performance, neurofibrillary tangles, and neuritic plaque burden.
• Machine learning models achieved notable accuracy in differentiating cognitive impairment (AUCs ranging from 0.731 to 0.772).
• Models successfully predicted future cognitive decline in a baseline control group (AUCs up to 0.804).

Conclusions:

• Metabolomic profiling can identify key metabolic alterations linked to Alzheimer's disease progression.
• Machine learning models based on metabolomic data show promise for early detection of cognitive impairment.
• These findings support the potential for metabolomics-driven early intervention strategies to mitigate AD progression.