Combining large scale in silico analysis with fragment screening to identify novel, ligandable secondary sites in cancer-associated proteins

Abstract

Proteins often have multiple binding sites involved in interactions with other molecules. The majority of currently approved drugs bind a protein's primary site, the major functional site in the protein. Targeting the primary site can be challenging. Secondary site binders can allow for efficient inhibition of difficult-to-drug protein targets and there are now multiple examples of secondary site inhibitors in the clinic. However, the majority of known secondary sites were discovered through serendipity and the systematic identification of ligandable secondary sites remains challenging. This thesis integrates high throughput in silico analysis of publicly available protein structures based on canSAR3D with fragment screening to identify novel, ligandable and functionally relevant secondary sites in clinically relevant protein targets. Following the analysis, triaging of identified sites for functional relevance, clinical impact and technical feasibility identified a short-list of four targets - p53, ESR1, PIK3CA and IDH1. The novel secondary site in isocitrate dehydrogenase 1 (IDH1) was selected for validation. The tumour-promoting IDH1-R132X mutation is found in up to 80% of glioma patients and 15% of acute myeloid leukaemia patients. Fragment screening identified 19 fragments binding specifically to the novel secondary site in IDH1-R132H. Following up these fragments in biochemical assays confirmed that binding to this pocket inhibits enzyme activity. My work shows that the newly discovered secondary pocket of IDH1-R132H is both ligandable and functionally relevant, and that my in silico analysis can be used to identify novel secondary sites in therapeutically relevant proteins.