Effects of amyloid and APOE4 on medial temporal lobe subregions in cognitively unimpaired elderly

Medial temporal lobe (MTL) sub-structures are differentially affected in early Alzheimers disease (AD), with a specific involvement of the entorhinal cortex (ERC), the perirhinal cortex (PRC) and CA1. However, the impact of amyloid (A{beta}) pathology and APOE {varepsilon}4 on MTL subregional atrophy remains relatively unknown. Our aim was to uncover these effects to further our understanding of the mechanisms underlying MTL atrophy in a population at-risk for AD. We used baseline data from 130 unimpaired older adults (mean age: 68.9 {+/-} 3.8 years) from the Age-Well randomized controlled trial for whom high-resolution structural MRI (T2-weighted; 0.4x0.4x2.5mm3), amyloid-PET (Florbetapir) and APOE genotype were available. Participants were dichotomized into amyloid positive (A{beta}+, n=27) and negative (A{beta}-, n=103), and APOE {varepsilon}4 carrier ({varepsilon}4+, n=35) and non-carriers ({varepsilon}4-, n=95). Hippocampal subfield (CA1, CA2, CA3, dentate gyrus [DG], subiculum [SUB]) and extra-hippocampal region (ERC, Brodmann area [BA] 35 and 36, and parahippocampal cortex [PHC]) volumes were estimated using ASHS and normalized by total intracranial volume. For each subregion, group comparisons were performed (A{beta}+ vs A{beta}- and {varepsilon}4+ vs {varepsilon}4-) using ANCOVAs, including age, sex and education as covariates. Interactions with age (i.e., A{beta} status * age and APOE {varepsilon}4 status * age) were also investigated for each subregion. No significant differences were observed between A{beta}+ and A{beta}-, nor between {varepsilon}4+ and {varepsilon}4-. However, a significant A{beta} status * age interaction were observed for CA1 (p<0.05), where volumes were negatively associated with age in the A{beta}+ group only. In addition, significant APOE {varepsilon}4 status * age interactions were found for CA1, SUB, ERC, DG and the whole hippocampus (p<0.05), where volumes were negatively associated with age in the {varepsilon}4+ group only. Overall, our analyses showed that both A{beta} and APOE {varepsilon}4 status interact with age on CA1, which is known to be specifically atrophied in early AD. In addition, APOE {varepsilon}4 status mediated the effects of age on other subregions (SUB, ERC, DG), suggesting a more important contribution of APOE {varepsilon}4 than amyloid to MTL atrophy in cognitively unimpaired population. These results are particularly important to develop MRI-based biomarkers to detect early AD and further our understanding of the mechanisms underlying MTL atrophy.