Introduction: We sought biological pathways that explained discordance between Alzheimer's disease (AD) pathology and symptoms.
Methods: In 306 Alzheimer's Disease Neuroimaging Initiative (ADNI)-1 participants across the AD clinical spectrum, we investigated association between cognitive outcomes and 23 cerebrospinal fluid (CSF) analytes associated with abnormalities in the AD biomarkers amyloid β1-42 and total-tau. In a 200-person "training" set, Least Absolute Shrinkage and Selection Operator regression estimated model weights for the 23 proteins, and for the AD biomarkers themselves, as predictors of ADAS-Cog11 scores. In the remaining 106 participants ("validation" set), fully adjusted regression models then tested the Least Absolute Shrinkage and Selection Operator-derived models and a related protein marker summary score as predictors of ADAS-Cog11, ADNI diagnostic category, and longitudinal cognitive trajectory.
Results: AD biomarkers alone explained 26% of the variance in validation set cognitive scores. Surprisingly, the 23 AD-related proteins explained 31% of this variance. The biomarkers and protein markers appeared independent in this respect, jointly explaining 42% of test score variance. The composite protein marker score also predicted ADNI diagnosis and subsequent cognitive trajectory. Cognitive outcome prediction redounded principally to ten markers related to lipid or vascular functions or to microglial activation or chemotaxis. In each analysis, apoE protein and four markers in the latter immune-activation group portended better outcomes.
Discussion: CSF markers of vascular, lipid-metabolic and immune-related functions may explain much of the disjunction between AD biomarker abnormality and symptom severity. In particular, our results suggest the hypothesis that innate immune activation improves cognitive outcomes in persons with AD pathology. This hypothesis should be tested by further study of cognitive outcomes related to CSF markers of innate immune activation.
Keywords: Alzheimer's disease; Biomarkers; Immunity; Prevention; Resilience.
Copyright © 2019 The Authors. Published by Elsevier Inc. All rights reserved.