Molecular topology, a novel descriptor for compound quality assessment

Abstract: Abstract The pharmaceutical industry is currently facing a high clinical attrition rate. In order to prevent the late-stage clinical failure, many investigations on compound quality and drug-likeness of compounds have been carried out. It has been widely accepted that molecular size and lipophilicity plays an important role in compound quality. Many attempts have been done to find out other factors which can influence compound quality beyond size and lipophilicity. Recently, a molecular topology concept has been put forward and its influence on compound quality has been investigated. It has been shown that drugs have higher fraction of compounds with only one ring system compared to clinical candidate and bioactive compounds. As an extension to the previous studies, the aim of this project is to further investigate how the molecular topology influences some of the most important physicochemical properties of molecules as well as the compound potency efficiency indices in general. Our results show that among reported molecules in the literature, compounds with only one ring system are smaller in size, less lipophilic and therefore has a higher probability to be less toxic. Interestingly compounds which have a simple topology also show advantage in terms of potency efficiency such as ligand efficiency (LE), ligand lipophilic efficiency (LLE) and ligand-efficiency-dependent lipophilicity index (LELP) compared with compounds which have a more complex topology. Thus a novel hypothesis why compounds with only one ring system are abundant among drugs has been proposed. On average molecules with only one ring system seems to bind more strongly to its protein target; this might reduce the necessary size of the molecule to reach a certain potency level. The reduction in size and lipophilicity reduces the risk of failure in clinical trials. Popular science summary: Molecular topology, A novel descriptor for compound quality assessment Drug discovery and development is a time consuming process which typically takes 15 to 20 years from the target identification until a drug makes it to the market. During this lengthy process, numerous compounds are tested, synthesized and validated in order to achieve the optimal efficacy and safety profile. Historically drug discovery was an iterative process of compound synthesis and in vivo screening. This paradigm has changed by the advancement of in vitro high-throughput screening technology and in silico techniques. The paradigm shift has largely improved hit identification efficiency; however the pharmaceutical industry still faces a high attrition rate. It is critical to identify compounds which are unlikely to succeed (low quality compounds) and to terminate the development of these compounds as early as possible. Recently molecular topology class was proposed; basically, it classifies compounds according to the number of ring systems. It has been reported that the fraction of compounds with only one ring system (1TR) is higher in drugs compared to clinical candidates and general bioactive compounds. This thesis aims to better understand the earlier observation of 1TR compounds’ enrichment in drugs. In this project, how the molecular topology influences some of the most important physicochemical properties of molecules as well as the compound potency efficiency indices was investigated. It showed that among reported molecules in the literature, compounds with only one ring system are smaller in size, less lipophilic and therefore has a higher probability to be less toxic. On average, molecules with only one ring system bind more strongly to its protein target; this reduces the necessary size of the molecule to reach a certain potency level. The reduction in size and lipophilicity reduces the risk of failure in clinical trials. By understanding what properties determine the quality of a compound, it will be possible to deliver drugs to the market more efficiently. Advisor: Ola Engkvist, Hongming Chen (Computational Chemistry, AstraZeneca R&D Mölndal) Master´s Degree Project 60 credits in Bioinformatics, 2011-2012 Department of Biology., Lund University