New Paths to LSD1 Inhibitor Design

Scientists used computer models that predict how proteins fold together with potential drug molecules. They focused on LSD1, an enzyme that changes how DNA is read by removing methyl groups from histones. Because LSD1 also serves as a scaffold for other proteins, blocking it could treat diseases like cancer. The team first collected all known reversible inhibitors of LSD1 and grouped them by chemical structure. They found that some groups were well represented, while large gaps existed where no inhibitors had been tested. These empty spots hint at hidden opportunities for new drugs that have not yet been explored. Next, the researchers fed the data into three advanced folding algorithms. Each algorithm tried to guess how a new inhibitor would fit into LSD1’s pocket.The predictions agreed strongly when the molecules resembled ones already in the training set and when they showed real laboratory activity. One model, AlphaFold3, performed best when only a few known inhibitors were available. All models captured the shape change of LSD1 that happens when a drug binds—a phenomenon called induced‑fit. However, they could not fully explain changes that occur far from the binding site, known as allosteric effects. The outcome is a detailed map of LSD1 inhibitors, showing which chemical shapes are promising and how they bind. The map also suggests new scaffolds that could be turned into potent drugs with nanomolar effectiveness. Researchers can use this guide to design better LSD1 inhibitors and apply the same approach to other proteins.