A Hybrid Experimental and in silico Platform for ITPK1 Chemical Probe Discovery

Inositol-tetrakisphosphate 1-kinase (ITPK1) is a pivotal enzyme in the inositol phosphate signaling pathway that functions to maintain the balance of inositol phosphate (IP) species. Dysregulation of this pathway has been linked to human disease, making ITPK1 an attractive therapeutic target. While high-throughput screening (HTS) is a traditional strategy for identifying small molecule inhibitors, integrating computational approaches can significantly speed up and enhance hit rates. Here, we developed a hybrid experimental and virtual approach towards the identification of ITPK1 chemical probe candidates. We first miniaturized an ITPK1 enzymatic assay to 1536-well format and screened ∼19,000 annotated and chemically diverse compounds. We then utilized the resulting dataset to develop Machine Learning (ML) and Pharmacophore (PH4) models to virtually screen a larger library of 120,000 compounds to expand the chemical diversity of the screening set. Importantly, our screening platform included a selectivity assay against PPIP5K2, the closest structural relative of ITPK1. The identified hits were evaluated for ITPK1 binding, including via a novel high-throughput Structure Dynamic Response (SDR) target engagement assay. Hits underwent further confirmation through orthogonal assays and mechanistic investigation, including obtaining a co-crystal structure for one of the hits. This integrated workflow-combining physical HTS with computational modeling-led to the identification of two novel candidate inhibitors. This study demonstrates an efficient, scalable strategy for targeting ITPK1 and offers a promising platform for drug discovery efforts in diseases linked to perturbed inositol phosphate pathways.