By combining microelectrode array (MEA) and human-induced pluripotent stem cell (hiPSC) technology, this study sought to validate the use of hiPSC-cortical neuron network long-term potentiation (LTP) for drug efficacy and safety studies for Alzheimer’s disease (AD) and anticholinergic cognitive burden (ACB), respectively.

There are current limitations on assessing disease pathology to improve the prediction of compound efficacy and safety for neurodegenerative disorders. With the recent advancements in hiPSC technology, the differentiation of cortical neurons has enabled biologically accurate modeling of difficult-to-source, complex neural networks, leading to accelerated timelines of clinically relevant preclinical drug assessment at a fraction of the cost of developing and testing animal models.

An AD phenotype was induced through the exogenous administration of amyloid-beta42 (Aβ42) or tau oligomers. ACB was modeled at various severity levels by administering concentrations of ACB drugs spanning two log scales to generate dose-response curves. This study utilized MEAs to evaluate electrically-induced LTP mimicking in vivo high-order cognitive function in a validated system.

Our system has successfully reproduced drug efficacy of four approved AD therapeutics following the induction of late-stage AD pathology from acute exposure to supraphysiological concentrations of Aβ42. It was found that only drugs mechanistically targeting pathology at the synapse were effective at ameliorating functional deficits from exposure to exogenous tau oligomers, which are typically onset in the hyperphosphorylated form downstream of Aβ42 plaque formation. Dose-response curves for ACB-induced neurocognitive detriment were established between ACB drug concentration and acute LTP maintenance.

The ability of our system to reproduce Aβ42 and tau disease pathology and differential drug efficacy for AD in addition to developing safety profile dose-response relationships of ACB drugs using LTP in a system sensitive enough to distinguish the ACB drug classes highlights to capabilities of hiPSC-derived neural networks for preclinical modeling of clinically relevant functional readouts.