Using Genetic Diversity to Manage Neurological Disease

Project Summary/Abstract: This one-year Administrative Supplement request in response to NOT-AG-22-025/PA-20-272 is a proposal to use our state-of-the-art Gaussian Process (GP) machine learning technologies we have pioneered to understand from a residue-by-residue covariance perspective cholesterol mismanagement by Niemann-Pick C1 (NPC1), as

described in parent grant AG070209, to address the poorly understood world-wide population of natural variants found in the genetic risk factors APOE, ABCA7, and TREM2 involved in cholesterol homeostasis in the brain. NPC1 is a protein that is required for the management of the intracellular distribution of cholesterol in all cells in

the human body through its central lysosomal localization along the exocytic and endocytic pathways. Mutations in NPC1 lead to an early onset (2-10-years of age) severe neurological disease resembling AD due to loss of intracellular cholesterol management critical for cell growth and function. In contrast, APOE, ABCA7 and TREM2

is a team of genetic risk factors that are thought to contribute as a collective to management of extracellular cholesterol homeostasis in the brain and that, when mutated, lead to late-onset (LO) Alzheimer’s Disease (AD) (LOAD). The primary purpose of this one-year Administrative Supplement is to use the approach described in

AG070209 build an unprecedented set of new molecular genetic tools to generate a comprehensive GP-based covariance view of the genetic risk factors APOE, ABCA7 and TREM2 that, when mutated, contribute to mismanagement of cholesterol leading to LOAD. Goal 1 will be to generate molecular tools to study 75 natural

variants impacting APOE function in the population that contribute to extracellular cholesterol mismanagement with the ApoE4 variant being a well-established genetic risk factor for LOAD; to study the impact of 75 variants of ABCA7, a cell surface transmembrane lipid/cholesterol channel involved in management of cholesterol

transfer at the extracellular interface through interaction with APOE containing lipid particles; and to generate 75 variants in the single pass transmembrane protein TREM2, an APOE signaling receptor in microglia involved in Ab42 aggregate clearance. In Goal 2, for each member of the cholesterol management team, we will

experimentally address the impact of natural variants on protein folding and stability by analysis of their trafficking through the exocytic pathway for secretion (APOE) or localization at the cell surface (ABCA7 and TREM2). In Goal 3, these membrane trafficking data will be used to build GP based ‘phenotype landscapes’, as described

in AG070209, in the context of the thermodynamic stability of each residue and clinical impact using ClinVar/ALZFORUM databases. These phenotype landscapes will describe quantitatively the covariance relationships between all residues in each protein sequence with defined uncertainty, linking each residue to the

thermodynamic properties governing its fold and clinical metrics that score their impact on LOAD. These analyses will provide a baseline for future studies of the therapeutic response of APOE, ABCA7 and TREM2 variants to proteostasis modulators as described in parent grant AG070209.