UC Davis

The purpose of this dissertation is to update the current methodology in understanding the function and structure of high-density lipoprotein (HDL), to describe improved methods for HDL isolation and deep-learning-assisted particle size measurement, and to show how HDL size and function alter in in vitro treatments with antioxidant carotenoids and environmental insults, as well as in clinical samples from study participants with Alzheimer’s disease (AD), mild cognitive impairment (MCI), or normal cognition. The dissertation is dived into four chapters with the corresponding objectives: 1. The first chapter is a literature review on recent updates and advancement of methods in understanding HDL characteristics, including physical properties, biological functions, and metabolism. 2. The second chapter reports an improved method for HDL isolation using ultracentrifugation followed by size exclusion chromatography. The isolated HDL fractions were evaluated on yield and purity, showing high HDL recovery and low levels of protein markers that are not HDL-associated. HDL fractions isolated using this method were further analyzed for HDL-associated protein markers, and we identified specific HDL-associated proteins that have preference for HDL subpopulations of specific size ranges. 3. The third chapter reports experiments to evaluate HDL’s size and functional alterations under direct treatment with lutein, zeaxanthin, oxidants (CuSO4, H2O2, HOCl), glycation, fructosylation, glycosylation removal (N-glycan, sialic acid), or acidic pH, as well as HDL pre-treated with antioxidant carotenoids (lutein or zeaxanthin) and then treated with H2O2 oxidant. We showed that HDL functions, including cholesterol efflux capacity (CEC), lecithin-cholesterol acyltransferase (LCAT) activity, and paraoxonase-1 (PON1) activity were altered by the treatment. Pre-incubating HDL with lutein attenuated the negative effects on HDL function caused by H2O2 treatment. 4. The fourth chapter reports an innovative method to use transmission electron microscopy followed by deep-learning-assisted particle detection for HDL particle size measurement of isolated HDL particles. The method showed improved precision and repeatability compared to a common image analysis software ImageJ on the specific task of HDL particle size measurement. We further use the method to characterize HDL particle size distribution from a large-scale data set from 183 study participants with AD, MCI, or normal cognition (age matched) and discovered a specific shift of HDL size from large to small in participants with dementia, with APOE genotype-dependent effects.