Using amyloid PET as a biomarker to detect progression of early Alzheimer disease

Abstract Background Alzheimer’s disease (AD), the most common form of dementia, is a complex polygenic disease with genetic, cellular, pathologic, and clinical heterogeneity. Recently, significant attempts have been made for identifying AD biomarkers for reliably tracking disease progression in its early asymptomatic stages. To this end, amyloid PET imaging has provided useful information tracking accumulation of parenchymal amyloid beta (Aβ) deposits in the brain. Previous investigations have conducted genetic association studies of CSF Aβ42, both continuously as well as using a threshold for amyloid positivity, and identified novel genes and loci potentially contributing to the preclinical stage in AD. In a recent case‐control based study using amyloid PET as a quantitative trait, Raghavan et al., identified a novel locus for amyloidosis within RBFOX1 gene in 4,314 participants, however, further investigations are needed to replicate and expand on these findings. Method In order to investigate the underlying genetic basis for brain amyloidosis in AD, we systematically analyzed the largest collection of amyloid imaging data (N = 6,320), across multiple ethnicities from multicenter cohorts (ADRC, A4, DIAN, ADNI, ADNIDOD, UPitt, and HABS) as a quantitative trait to identify the functional variants and genes driving the association of AD. Furthermore, we have conducted diagnosis‐, gender‐, and APOE‐stratified analyses to investigate the effect of these variables on the brain amyloidosis. Result Preliminary results found a strong APOE signal in chr19 (min P = 1e‐185) and a genome‐wide significant hit (P = 2e‐08) in the chromosome 7 (potentially, associated to TMEM106B gene) warranting further investigation (Fig. 1). Moreover, some suggestive signals were detected in chr10 (P = 1e‐07) and chr12 (P = 5e‐07). Consistent with previous finding, we found RBFOX1 to be associated with increased amyloid burden (β = 0.07), however, the signal was only nominally significant (P = 0.03). Additional analyses including multi‐ancestry meta‐analysis using FHS, AIBL, and EISAI cohorts and post‐GWAS analyses are ongoing. Conclusion The identification of pre‐clinical AD‐specific molecular signatures and pathways will enable the characterization of appropriate therapeutic targets for the prevention and/or treatment of AD. Results generated from our investigation will further serve to generate prediction models for amyloid positivity and Mendelian Randomization (MR) analyses.