Comparative characterization of 5XFAD and hAbetaSAA mouse models of familial AD

Abstract

AbstractBackgroundThe failure of most clinical Alzheimer’s disease (AD) trials has been partially attributed to the lack of translatability of current AD mouse models to human patients. A recently developed model of familial AD (fAD), expressing Swedish, Arctic and Austrian mutations in App (hAbetaSAA), has been shown to be a useful amyloidogenic model which recapitulates many aspects of human AD, including plaque distribution and microglial transcriptional changes. Our aim is to further characterize the hAbetaSAA model and compare it to the widely used 5xFAD transgenic model.MethodMotion Sequencing (MoSeq) software was used to model the underlying structure of spontaneous behaviors recorded in an open field in hAbetaSAA and 5xFAD mice longitudinally from 10 to 18 months of age. Aged hAbetaSAA and 5xFAD brain tissue was used for spatial transcriptomic/proteomic profiling, performed by Nanostring’s GeoMx®, which allowed for quantification of gene and protein expression in plaque‐associated and non‐plaque‐associated regions of interest. A fluorophore‐conjugated amyloid antibody (Methoxy‐X04) was administered to cohorts of hAbetaSAA and 5xFAD prior to harvest at various ages. An additional fluorophore‐conjugated antibody (lectin Dylight®594) was administered to 19‐month‐old cohorts permitting visualization of cerebral amyloid angiopathy (CAA). Whole brains were sectioned/imaged using Serial Two‐Photon Tomography on the TissueCyte (TissueVision), creating indexed tissue sections and high‐resolution 3D models of each brain. Subsequent rounds of staining permitted characterization of disease‐associated microglia (DAM) and dystrophic neurites. An independent cohort was evaluated for EEG telemetry.ResultMoSeq revealed divergent behaviors of 5xFAD and hAbetaSAA mice suggesting differences in behavioral phenotypes of these two models. Preliminary GeoMx data shows upregulation of DAM genes localized to plaques in hAbetaSAA homozygotes as identified by RNA‐seq; correlating 5xFAD data is in progress. Further assessments of bulk RNA‐Seq, amyloid distribution, CAA burden and cortical EEG spectral analysis are underway.ConclusionComparison of hAbetaSAA and 5xFAD using innovative modes of assessment showcase hAbetaSAA as an amyloidogenic mouse model of fAD that aligns more closely with human than 5xFAD. This model is available for preclinical research with no licensing restrictions and is devoid of artifacts related to transgenic overexpression, positioning it as an improved mouse model for studying fAD mutations.