Abstract
Despite ongoing debate, the amyloid β-protein (Aβ) remains the prime therapeutic target for the treatment of Alzheimer's disease (AD). However, rational drug design has been hampered by a lack of knowledge about neuroactive Aβ. To help address this deficit, we developed live-cell imaging of iPSC-derived human neurons (iNs) to study the effects of the most disease relevant form of Aβ-oligomeric assemblies (oAβ) extracted from AD brain. Of ten brains studied, extracts from nine caused neuritotoxicity, and in eight cases this was abrogated by Aβ immunodepletion. Here we show that activity in this bioassay agrees relatively well with disruption of hippocampal long-term potentiation, a correlate of learning and memory, and that measurement of neurotoxic oAβ can be obscured by more abundant non-toxic forms of Aβ. These findings indicate that the development of novel Aβ targeting therapeutics may benefit from unbiased activity-based discovery. To test this principle, we directly compared 5 clinical antibodies (aducanumab, bapineuzumab, BAN2401, gantenerumab, and SAR228810) together with an in-house aggregate-preferring antibody (1C22) and established relative EC50s in protecting human neurons from human Aβ. The results yielded objective numerical data on the potency of each antibody in neutralizing human oAβ neuritotoxicity. Their relative efficacies in this morphological assay were paralleled by their functional ability to rescue oAβ-induced inhibition of hippocampal synaptic plasticity. This novel paradigm provides an unbiased, all-human system for selecting candidate antibodies for advancement to human immunotherapy.