The epsilon-2 allele of the APOE gene (APOE2) has been shown to reduce the risk for late-onset (sporadic) Alzheimer’s disease (AD) dementia. Little is known about APOE2-related brain changes that may underlie this protective effect.


      We used multimodal neuroimaging data to comprehensively examine potential protective brain effects of the APOE2 genotype compared to the risk-neutral homozygous APOE3 genotype in a large sample of non-demented older individuals (cognitively normal subjects and those with mild cognitive impairment). Imaging data was obtained from a total of 572 APOE genotyped individuals enrolled in the ADNI study and included assessments of regional amyloid load using AV45-PET, glucose metabolism using FDG-PET, and gray matter volume using structural MRI. Group differences in imaging markers were assessed using region-of-interest (ROI) and voxel-based analyses, controlled for age, sex, education, and clinical diagnosis. Additional linear regression models examined genotype-specific effects of age on the distinct imaging markers. In secondary analyses, cerebrospinal fluid (CSF) markers of amyloid and tau pathology were examined to assess the reproducibility of the main imaging findings using fluid biomarkers.


      In region-of-interest based analyses, APOE2 carriers had less precuneal amyloid pathology (p = 0.009) and did not show the typical age-related increase in amyloid load (β = 0.10, p = 0.49; Figure 1). By contrast, parietal metabolism and hippocampal volume did not differ between APOE2 and APOE3 genotypes, and both groups showed comparable negative effects of age on these markers. The amyloid-specificity of APOE2-related brain changes was corroborated in spatially unbiased voxel-wise analyses (Figure 2), as well as by a significant APOE2 effect on CSF markers of amyloid (p = 0.02), but not tau, pathology.


      The reduced dementia risk associated with the APOE2 allele may be mediated by a relatively specific reduction of amyloid accumulation. Whether amyloid-lowering therapeutic interventions may yield similar decreases in dementia risk remains to be established in controlled clinical trials.
      Figure thumbnail fx1
      Figure 1Effects of age and APOE genotype on multimodal imaging markers. Precuneal AV45-SUVR (top), temporoparietal FDG-SUVR (middle), and hippocampal volume (bottom) are plotted against age for APOE3 (black) and APOE2 (red) genotypic groups. APOE4 carriers (blue) are included for comparison. Separate linear regression lines are fitted for each group.
      Figure thumbnail fx2
      Figure 2Voxel-wise analysis of reduced amyloid load in non-demented older APOE2 carriers
      Significant effects of a voxel-wise two-sample t-test assessing reduced AV45-SUVR in APOE2 carriers compared to the APOE3 control group, controlled for age, gender, education, and clinical diagnosis. Statistical map was thresholded at p(FDR) < 0.05. Analogous analyses for FDG-SUVR or gray matter volume did not reveal any significant effects