Selective estrogen receptor modulators (SERMs) and aromatase inhibitors (ARIs) are currently used for treating estrogen receptor (ER) positive breast cancers. Interestingly, some of those SERM and ARI drugs have beneficial effects in preventing Alzheimer's via their estrogenic effects in the brain. However, the impact of those anti-breast cancer drugs on early tau aggregation in Alzheimer's is not –studied. Using AI-assisted virtual drug screening, we aim at discovering multi-target compounds (MTCs) that can simultaneously bind to three molecular targets associated with breast cancers and early tauopathies. These targets are ER-α, GSK-3β, and tau-k18 dimer. Here, ER-α is a major subtype of ER, the kinase GSK-3β is responsible for creating misfolded tau, and the tau-k18 dimer is a model of early tau aggregates. Our first step involves generating ∼800 chemical analogs of experimentally derived inhibitors of ERα, GSK-3β, and tau aggregation using chemical similarity search tools. We then successfully performed high-throughput molecular docking of those chemical analogs and their parent compounds against our three molecular targets via AutoDock Vina. Our molecular docking results comprised 20 docked conformational poses for each compound-target complex. Finally, we employed a recently developed supervised machine-learning rescoring tool, SCORCH, followed by the application of a dynamic threshold algorithm, to identify four MTCs that are top binders to our three protein targets. Our preliminary screening results indicate that exemestane, an exemestane analog, a raloxifene analog, and an apomorphine analog are potential MTC candidates for inhibiting ER and tau aggregation. Following these results, all-atom molecular dynamics simulations under physiological conditions will be performed to evaluate the free-energy binding affinity and protein disaggregation behaviors of the newly discovered MTCs in tau aggregates.
Preston et al. (2026) studied this question.