Computational polypharmacology comes of age.

Abstract

In the last years, the “one target, one drug” paradigm that has traditionally dominated drug discovery has been deeply challenged by the evidence that small molecules interact simultaneously with multiple targets, a phenomenon known as polypharmacology. Today, polypharmacology is recognized as a new valuable opportunity for drug discovery and development. It is now well established that drug molecules typically bind to several targets, and that their efficacy and safety is mostly dependent on their polypharmacological profile (Jalencas and Mestres, 2012; Peters, 2013; Anighoro et al., 2014). Indeed, one of the most common reasons for terminating a drug discovery program has been promiscuity or lack of selectivity of the developed compounds. This leads to important considerations regarding the polypharmacology inherent in chemical structures and its possible exploitation for drug discovery. First, side effects caused by drug binding to unwanted off-targets (adverse polypharmacology) should be identified as early as possible in the drug discovery pipeline. Second, potential synergistic effects arising from hitting multiple targets (beneficial polypharmacology) should be taken into consideration and thoroughly incorporated in the drug design strategy. Third, polypharmacological approaches have the potential to redirect stalled drug discovery projects and to reposition valuable hits or leads (drug repositioning). Finally, prediction of polypharmacological profiles can be used to uncover new macromolecular targets for already known or new developing drugs (target identification and deconvolution). In all these areas, computational polypharmacology is gaining a foothold in drug discovery, as witnessed by the increasing number of publications reporting theoretical approaches and methods specifically put forward to address these needs.