The amyloid-β (Aβ) peptide, the primary constituent of amyloid plaques found in Alzheimer's disease (AD) brains, is derived from sequential proteolytic processing of the Amyloid Precursor Protein (APP). However, the contribution of different cell types to Aβ deposition has not yet been examined in an in vivo, non-overexpression system. Here, we show that endogenous APP is highly expressed in a heterogeneous subset of GABAergic interneurons throughout various laminae of the hippocampus, suggesting that these cells may have a profound contribution to AD plaque pathology. We then characterized the laminar distribution of amyloid burden in the hippocampus of an APP knockin mouse model of AD. To examine the contribution of GABAergic interneurons to plaque pathology, we blocked Aβ production specifically in these cells using a cell type-specific knockout of BACE1. We found that during early stages of plaque deposition, interneurons contribute to approximately 30% of the total plaque load in the hippocampus. The greatest contribution to plaque load (75%) occurs in the stratum pyramidale of CA1, where plaques in human AD cases are most prevalent and where pyramidal cell bodies and synaptic boutons from perisomatic-targeting interneurons are located. These findings reveal a crucial role of GABAergic interneurons in the pathology of AD. Our study also highlights the necessity of using APP knock-in models to correctly evaluate the cellular contribution to amyloid burden since APP overexpressing transgenic models drive expression in cell types according to the promoter and integration site and not according to physiologically relevant expression mechanisms.

Neuronal cell loss is a defining feature of Alzheimer's disease (AD), but it remains unclear how neurons die and how this relates to other defining characteristics of the disease. Existing in vivo AD models only partially recapitulate the neuropathology of AD with very mild or no neuronal cell loss. Here we demonstrate that human neurons xenografted in mouse brain exposed to amyloid pathology develop sarkosyl-insoluble tau filaments, positive Gallyas silver staining, release phosphorylated tau (P-tau181) into the blood, and display considerable neuronal cell loss, providing a model for the induction of full Tau pathology by simple exposure to amyloid pathology in AD. The alterations are specific to human neurons and contrast with the mild effects exhibited in mouse neurons. A core transcriptional program in the human neurons is characterized by strong upregulation of MEG3, a neuron-specific long noncoding RNA. MEG3 is also strongly upregulated in neurons from AD patients in situ. MEG3 expression alone is sufficient to induce necroptosis in human neurons in vitro. Orally administered small molecule receptor-interacting protein (RIP) kinase -1 and -3 inhibitors rescued the neuronal cell loss in this novel AD model. Thus, xenografted human neurons are uniquely sensitive to amyloid pathology, recapitulate all the defining neuropathological features of AD, and ultimately die by necroptosis.