Alzheimer’s disease (AD) is a major cause of dementia that affects more than 44 million people worldwide. Mutations in APP and catalytic subunits of the γ-secretase complex (PSEN1, PSEN2) have been identified as major causes of familial early-onset AD. However, genetic factors that contribute to late-onset AD (LOAD) are mostly unknown. Through an exome-wide genotyping microarray approach, the CHARGE (Cohorts for Heart and Aging Research in Genomic Epidemiology) Consortium identified a novel LOAD-associated risk variant in the TM2D3 locus within an Icelandic population. Interestingly in Drosophila, mutations in almondex (amx), the fly homolog of TM2D3, cause a maternal effect neurogenic phenotype that phenocopy loss-of-function phenotypes of Notch and Presenillin. However, multiple bioinformatics programs (PolyPhen, SIFT, CADD) predicted that this LOAD associated variant in TM2D3 is not damaging, raising the question of whether or not this is a true functional variant.
To determine if the LOAD-associated variant in TM2D3 is functional, we “humanized” the amx gene by rescuing the Drosophila amx mutants with a genomic rescue construct that encodes human TM2D3. Interestingly, while a genomic rescue construct of amx that expresses wild-type human TM2D3 can significantly suppress the maternal effect neurogenic phenotype, the LOAD-variant failed to do so. These results suggest that the LOAD-associated variant in TM2D3 acts as a loss-of-function mutation in the context of embryonic Notch signaling. Considering that amx has been previously implicated to function at the γ-secretase mediated activation step of Notch, we hypothesize that TM2D3 plays a role in AD through modulation of the β-amyloid pathway.