Multi-omic ADNI CSF and plasma data integration identifies distinct metabolic transitions in disease progression in Alzheimer's Disease

Abstract

Age and APOE genotype are the strongest known risk factors for late-onset Alzheimers disease (AD), but the mechanisms linking them to neuronal loss remain incompletely defined. Using multiomic data from the Alzheimers Disease Neuroimaging Initiative (ADNI), we propose a unified hypothesis in which two interdependent failure modes--saturation of microglial lipid flux capacity and disruption of the astrocyte-neuron lactate shuttle (ANLS) due to excess astrocytic membrane cholesterol--drive disease progression upstream of amyloid and tau pathology. Stratifying participants by cognitive score quartiles, we find consistent associations linking impaired lipid clearance, metabolic stress, and genetic variants regulating cholesterol handling. These processes appear to reinforce each other, resulting in accelerating neurodegeneration. Our hypothesis reframes AD as a systems-level collapse in metabolic coordination, rather than a purely linear pathological cascade. These insights emerged during the development of digital twin models for personalized interventions, highlighting the power of systems approaches to reveal hidden drivers of neurodegeneration.