Differential associations between regional amyloid PET and Alzheimer’s disease polygenic risk scores

Abstract

AbstractBackgroundWe previously described three spatio‐temporal subtypes of amyloid accumulation trajectories, measured with PET (Collij et al., 2022). APOE4 carriers tended to have initial amyloid deposition in frontal regions compared to those showing initial deposition in parietal and occipital regions, suggesting different underlying disease mechanisms. Here, we investigated the relationship between global and regional amyloid burden and Alzheimer’s disease polygenic risk scores (AD‐PRS).MethodAD‐PRS were calculated for 458 non‐demented individuals participating in EPAD and AMYPAD, based on 85 identified risk loci, including and excluding the two major APOE variants (Bellenguez et al., 2022). After exclusion of APOE, the remaining 83 risk loci were functionally annotated and classified into six pathway sets (annotation tool as defined in Tesi et al., (2020)): path‐PRSimmune‐activation, path‐PRSsignal‐transduction, path‐PRSinflammatory‐response, path‐PRSmigration, path‐PRSamyloid‐production, path‐PRSclearance. 18F‐Florbetaben or 18F‐Flutemetamol PET were acquired and quantified using the Centiloid pipeline to obtain global and regional (frontal, precuneus, lateral parietal, lateral temporal, occipital) amyloid burden (Klunk et al., 2015). Linear models assessed the association between amyloid burden and each (path‐)PRS, corrected for age, sex, and PET‐tracer.ResultParticipants were 67.5±7.03 years of age, 57.6% female, and 109 were considered amyloid‐positive (CL>21). Global amyloid burden was highly associated with AD‐PRSAPOE and, to a lesser extent, AD‐PRSnoAPOE (β = 14.75,p<5.9e‐15 and β = 7.99,p = 0.03, respectively, Figure 1). Similar results were obtained for regional burden (Table 1). After exclusion of APOE, both path‐PRSmigration (β = 51.92,p = 0.02) and path‐PRSclearance (β = 39.87,p = 0.02) were significantly associated with global amyloid burden, where path‐PRSclearance showed the strongest association with amyloid burden across all regions (frontal: β = 38.96,p = 0.009; precuneus: β = 50.11,p = 0.006; lateral parietal: β = 44.10,p = 0.003; lateral temporal: β = 26.59,p = 0.04; occipital: β = 17.17,p = 0.049). Path‐PRSmigration and Path‐PRSamyloid‐production were significantly associated with only the precuneus (β = 58.15,p = 0.02; β = 19.03,p = 0.04, respectively) and lateral parietal cortex (β = 47.12,p = 0.02; β = 16.41,p = 0.03, respectively).ConclusionResults illustrate that genetic predisposition beyond APOE is associated with global and regional amyloid burden. Path‐PRS results implicate clearance mechanisms in early disease pathogenesis, though strongest with the precuneus and lateral parietal cortex, which are early amyloid‐accumulating regions. Since AD‐PRS were associated with longitudinal changes in amyloid (Luckett et al., 2022), we aim to further investigate the association of path‐PRS with longitudinal amyloid within EPAD and AMYPAD.