Genetic risk for Alzheimer disease predicts hippocampal volume through the human lifespan

Abstract

Objective: To test the hypothesis that genetic risk for Alzheimer disease (AD) may represent a stable influence on the brain from early in life, rather than being primarily age dependent, we investigated in a lifespan sample of 1,181 persons with a total of 2,690 brain scans, whether higher polygenic risk score (PGS) for AD and presence of APOE ε4 was associated with lower hippocampal volumes to begin with, as an offset effect, or possibly faster decline in older age.

Methods: Using general additive mixed models, we assessed the relations of PGS for AD, including variants in APOE with hippocampal volume and its change in a cognitively healthy longitudinal lifespan sample (age range: 4-95 years, mean visit age 39.7 years, SD 26.9 years), followed for up to 11 years.

Results: AD-PGS and APOE ε4 in isolation showed a significant negative effect on hippocampal volume. The effect of a 1 sample SD increase in AD-PGS on hippocampal volume was estimated to -36.4 mm³ (confidence interval [CI]: -71.8, -1.04) and the effect of carrying ε4 allele(s) -107.0 mm³ (CI: -182.0, -31.5). Offset effects of AD-PGS and APOE ε4 were present in hippocampal development, and interactions between age and genetic risk on volume change were not consistently observed.

Conclusions: Endophenotypic manifestation of polygenic risk for AD may be seen across the lifespan in cognitively healthy persons, not being confined to clinical populations or older age. This emphasizes that a broader population and age range may be relevant targets for attempts to prevent AD.

Figure 1. Hippocampal volume in relation to age and genetic risk

Hippocampal volume (across hemispheres, shown in mm³ on the y-axis) and change in relation to (A) age (in years, x-axis) plotted with individual trajectories overlaid, (B) polygenic risk score (PGS) for Alzheimer disease (AD) (x-axis, continuous scale 0–1), (C) age (in years, x-axis) with PGS for AD set to the first (red line) and fourth quartiles (blue line) of the sample, and (D) trajectories for carriers (blue line) and noncarriers (red line) of the APOE ε4 allele. PGSs for AD shown here were constructed using single nucleotide polymorphisms with association p value <0.5 from Lambert et al.

Figure 2. Hippocampal volume and PGS for AD limited to SNPs showing genome-wide significant association

Upper panel: including APOE. Hippocampal volume (across hemispheres, shown in mm³ on the y-axis) and change in relation to (A) AD-PGS including the APOE region as calculated based on effect sizes from Lambert et al. at p < 5e-08 (x-axis, continuous scale 0–1). (B) Age (in years, x-axis) with AD-PGS at p < 5e-08 including the APOE region set to the first (red line) and fourth (blue line) quartiles of the sample. Lower panel: excluding APOE. Hippocampal volume (across hemispheres, shown in mm³ on the y-axis) and change in relation to (C) AD-PGS excluding the APOE region as calculated based on effect sizes from Lambert et al. at p < 5e-08 (x-axis, continuous scale 0–1). (D) Age (in years, x-axis) with AD-PGS at p < 5e-08 excluding the APOE region set to the first (red line) and fourth (blue line) quartiles of the sample. AD = Alzheimer disease; PGS = polygenic risk score; SNP = single nucleotide polymorphism.