Differential effects of APOE2 and APOE4 alleles on PET‐measured amyloid‐β and tau deposition in older individuals without dementia: Neuroimaging / imaging and genetics

Abstract

Background: The APOE 4 allele is the best‐established genetic risk factor for sporadic Alzheimer’s disease (AD), whereas the APOE2 allele has been linked to decreased AD risk. Although the most prominent neurobiological effect of the APOE polymorphism appears to relate to altered amyloid‐β (Aβ) deposition, recent PET imaging evidence has also pointed to Aβ‐independent APOE4 effects on increased medial temporal tau deposition. Here, we investigated whether APOE2 has a reversed (i.e. protective) association with PET‐measured tau deposition. Method: From the ADNI cohort we identified 467 APOE‐genotyped individuals without dementia (31% with mild cognitive impairment (MCI); 53% female) who underwent tau ([18 F]flortaucipir) and Aβ ([18 F]florbetapir or ([18 F]florbetaben) PET as well as a structural MRI scan. 257 were APOE3/3 homozygotes, 45 had an APOE2/3 genotype, 160 were APOE3/4 or APOE4/4. Five individuals with a APOE2/4 genotype were excluded from the analysis. Linear models, controlled for age, sex, education, and diagnosis, were used to compare genotype groups on multimodal imaging biomarkers, including entorhinal cortex (ERC) flortaucipir standard uptake value ratio (SUVR, inferior cerebellum reference), cortical Aβ load expressed as centiloid values, and hippocampal volume on MRI. On the basis of an emerging literature highlighting female‐specific risks for developing AD, we also studied sex*APOE interactions on biomarker outcomes. Result: Compared to APOE3/3, APOE2/3 carriers had significantly reduced Aβ burden (14.2 vs 25.5 centiloids, p=0.021), but did not differ in ERC tau (1.166 vs 1.148 SUVR, p=0.52) or hippocampal volume (p=0.82)(Fig‐1). These effects did not differ between males and females, and similar results were obtained when using alternative medial/inferior temporal tau regions‐of‐interest. By contrast, APOE4 carriers showed both significantly increased Aβ (48.3 centiloids, p<0.001) and ERC tau burden (1.249 SUVR, p<0.001), but no difference in hippocampal volume (p=0.33). In combined regression models, Aβ load fully mediated the APOE4 effect on ERC tau in males (APOE4: t(188)=0.6/p=0.52; Aβ: t(188)=5.3/p<0.001), whereas in females APOE4 retained a minor direct effect (t(217)=2.2/p=0.03) in addition to the prominent effect of Aβ (t(217)=10.6/p<0.001). Conclusion: We demonstrate a relatively minor Aβ‐independent effect of APOE4 on medial temporal tau deposition, which was observed in females only. Our data do not support a corresponding protective effect of APOE2.