A pathway profile-based method for drug repositioning

Finding new applications for existing pharmaceuticals, known as drug repositioning, is a validated strategy for resolving the problem of high expenditure but low productivity in drug discovery. Currently, the prevalent computational methods for drug repositioning are focused mainly on the similarity or relevance between known drugs based on their features, including chemical structure, side effects, gene expression profile, and/or chemical-protein interactome. However, such drug-oriented methods may constrain the newly predicted functions to the pharmacological functional space of the existing drugs. Clinically, many drugs have been found to bind off-target (i.e. to receptors other than their primary targets), which can lead to undesirable effects. In this study, which integrates known drug target information, we propose a disease-oriented strategy for evaluating the relationship between drugs and disease based on their pathway profile. The basic hypothesis of this method is that drugs exerting a therapeutic effect may not only directly target the disease-related proteins but also modulate the pathways involved in the pathological process. Upon testing eight of the global best-selling drugs in 2010 (each with more than three targets), the FDA (Food and Drug Administration, USA)-approved therapeutic function of each was included in the top 10 predicted indications. On average, 60% of predicted results made using our method are proved by literature. This approach could be used to complement existing methods and may provide a new perspective in drug repositioning and side effect evaluation.