Associations of endogenous estrogens, plasma Alzheimer's disease biomarkers, and APOE4 carrier status on regional brain volumes in postmenopausal women.

ABSTRACT

Background: Women carrying the APOE4 allele are at greater risk of developing Alzheimer's disease (AD) from ages 65-75 years compared to men. To better understand the elevated risk conferred by APOE4 carrier status among midlife women, we investigated the separate and interactive associations of endogenous estrogens, plasma AD biomarkers, and APOE4 carrier status on regional brain volumes in a sample of late midlife postmenopausal women. Methods: Participants were enrolled in MsBrain, a cohort study of postmenopausal women (n = 171, mean age = 59.4 years, mean MoCA score = 26.9; race = 83.2% white, APOE4 carriers = 40). Serum estrone (E1) and estradiol (E2) levels were assessed using liquid chromatography-tandem mass spectrometry. APOE genotype was determined using TaqMan SNP genotyping assays. Plasma AD biomarkers were measured using single molecule array technology. Cortical volume was measured and segmented by FreeSurfer software using individual T1w MPRAGE images. Multiple linear regression models were conducted to determine whether separate and interactive associations between endogenous estrogen levels, plasma AD biomarkers (Aß42/Aß40, Aß42/p-tau181), and APOE4 carrier status predict regional brain volume (21 regions per hemisphere, selected a priori); and, whether significant interactive associations between estrogens and AD biomarkers on brain volume differed by APOE4 carrier status. Results: There was no main effect of APOE4 carrier status on regional brain volumes, endogenous estrogen levels, or plasma AD biomarkers. Estrogens did not associate with regional brain volumes, except for positive associations with left caudal middle frontal gyrus and fusiform volumes. The interactive association of estrogens and APOE4 carrier status on brain volume was not significant for any region. The interactive association of estrogens and plasma AD biomarkers predicted brain volume of several regions. Higher E1 and E2 were more strongly associated with greater regional brain volumes among women with a poorer AD biomarker profile (lower Aß42/40, lower Aß42/p-tau181 ratios). In APOE4-stratified analyses, these interactions were driven by non-APOE4 carriers. Conclusion: We demonstrate that the brain volumes of postmenopausal women with poorer AD biomarker profiles benefit most from higher endogenous estrogen levels. These findings are driven by non-APOE4 carriers, suggesting that APOE4 carriers may be insensitive to the favorable effects of estrogens on brain volume in the postmenopause.