Systematic validation of variants of unknown significance in APP, PSEN1 and PSEN2

• Published in: Neurobiology of disease 2020-06, Vol.139, p.104817-104817, Article 104817
• Main Authors: Hsu, Simon, Pimenova, Anna A., Hayes, Kimberly, Villa, Juan A., Rosene, Matthew J., Jere, Madhavi, Goate, Alison M., Karch, Celeste M.
• Format: Article
• Language: English
• Subjects: Alzheimer Disease - genetics; Alzheimer's disease; Amyloid beta-Protein Precursor - genetics; APP; Cell-based assays; Humans; Mutation; Pathogenicity algorithm; Presenilin-1 - genetics; Presenilin-2 - genetics; PSEN1; PSEN2
• ISSN: 0969-9961; 1095-953X
• DOI: 10.1016/j.nbd.2020.104817

Alzheimer's disease (AD) is a neurodegenerative disease that is clinically characterized by progressive cognitive decline. More than 200 pathogenic mutations have been identified in amyloid-β precursor protein (APP), presenilin 1 (PSEN1) and presenilin 2 (PSEN2). Additionally, common and rare variants occur within APP, PSEN1, and PSEN2 that may be risk factors, protective factors, or benign, non-pathogenic polymorphisms. Yet, to date, no single study has carefully examined the effect of all of the variants of unknown significance reported in APP, PSEN1 and PSEN2 on Aβ isoform levels in vitro. In this study, we analyzed Aβ isoform levels by ELISA in a cell-based system in which each reported pathogenic and risk variant in APP, PSEN1, and PSEN2 was expressed individually. In order to classify variants for which limited family history data is available, we have implemented an algorithm for determining pathogenicity using available information from multiple domains, including genetic, bioinformatic, and in vitro analyses. We identified 90 variants of unknown significance and classified 19 as likely pathogenic mutations. We also propose that five variants are possibly protective. In defining a subset of these variants as pathogenic, individuals from these families may eligible to enroll in observational studies and clinical trials. •We sought to classify variants of unknown significance in APP, PSEN1, and PSEN2.•We applied a pathogenicity algorithm using genetic, bioinformatic, and in vitro data.•90 variants of unknown significance were evaluated.•19 variants were classified as likely pathogenic mutations.•5 variants were classified as possibly protective.