Amyloid precursor protein (APP) knock‐in mouse model recapitulates transciptomic signature in human late‐onset Alzheimer’s disease

Abstract

AbstractBackgroundOver the years, improved rodent models have been generated aiming at recapitulating Alzheimer’s disease (AD) phenotype. However, the majority of these models harbor pathological mutations that are present in early‐onset AD, which represents only 5% of all cases. On the other hand, new models aiming at recapitulating late‐onset AD (LOAD) are of great interest. In this context, the knock‐in mouse model (hAβ‐KI line), in which murine amyloid precursor protein (APP) gene was humanized at the Aβ locus, was recently introduced as a potential LOAD model. Here, compare hippocampal transcriptomic profiles of human LOAD individuals and hAβ‐KI mouse model. We hypothesized that hippocampal transcriptomic profile of hAβ‐KI mouse model will be similar to the human LOAD.MethodPublicly available human AD/cognitively unimpaired (CU) transcriptomic profiles of hippocampus were collected from GEO (https://www.ncbi.nlm.nih.gov/geo/), merged and submitted to differential expression analysis and master regulator analysis (MRA) using R. RNAseq expression profile of hAβ‐KI mouse model was obtained at the AMP‐AD Knowledge Portal (https://www.synapse.org/) and also submitted to differential expression analysis and MRA.ResultWe observed that the hippocampus of human LOAD individuals and hAβ‐KI mice model share over 100 differentially expressed genes (DEGs). Functional enrichment of Gene Ontology terms using these genes revealed several biological processes such as gliogenesis, glial cell differentiation, axon development and ensheathment of neurons. We also found shared transcription factors between human LOAD individuals and the hAβ‐KI mice model (Figure 1).ConclusionThe novel hAβ‐KI mice, which expresses an humanized APP, has been proposed as a better model to mimic sporadic LOAD. Our early findings indicate that at the transcriptomic level the hAβ‐KI mice presents molecular alterations compatible to human LOAD condition. Thus, studies using this model may reveal important insights to understanding sporadic LOAD and lead to better translational concordance.