Molecular modelling.

The enormous pressure that the pharmaceutical and biotech companies are facing, has created the need to apply all available techniques to decrease attrition rates, costs and the time to market. Currently, one of the most widely applied techniques in drug discovery is computational chemistry and molecular modelling. This branch of science is centred on applying the fundamental laws of physics and chemistry to the study of molecules. In the case of drug discovery, the molecules under study are those directly or indirectly involved in human disease. The ultimate aim is to create models and simulations, which can help in the different stages of a discovery pipeline by predicting, rationalizing and estimating the properties of molecules and their interactions, thereby allowing a more rational approach to drug development. This whole trend is now seen both as an alternative and a complement to the more ‘‘brute-force’’ approach exemplified by the application of combinatorial chemistry and high-throughput screening (HTS). The fundamental factor allowing the widespread use of molecular modelling is the central paradigm of today’s drug discovery, the one-disease one-target concept and its implementation. Within this paradigm, a certain human condition is associated with the role played by a particular macromolecule, whose action can be modulated with a small organic molecule in order to achieve a therapeutic effect. With this perspective, drugs are developed in a sequential way. First, a macromolecular target to treat the pathology under study must be found, a process termed as target finding. Then, the search for small molecule binders (hits) for that particular target begins, the so-called hit finding stage. Once found, these binders must be optimized in order to achieve better in vitro activity, selectivity, pharmacodynamic and pharmacokinetic properties, the stage termed as hit to lead. Then, the lead must be optimized in a series of in vivo studies, the stage of lead optimization. Only when the lead has been optimized and tested in several animal models can the project then progress to human clinical trials. Computational chemistry and molecular modelling methods have become central features of all these pre-clinical research stages of the drug-discovery process. When applied to the study of drugs and their receptors, molecular modelling techniques are generally divided into two broad categories. Ligand-based modelling consists of a series of techniques used for creating models and predictions based solely on the structure of the small organic compounds. In contrast, structure-based drug design (SBDD) exploits the knowledge of the 3D structure of one or more biological receptors (targets, the ones sought to modulate and antitargets, the ones sought not to interfere with) and/or their macromolecular ligands. These two broad categories are very often applied in a myriad of different combinations, so the frontier separating them is not clear-cut. Molecular modelling as applied to SBDD has undergone a dramatic change over the last two decades. At first, the simulation of biochemical systems and their interactions was a nearly unfeasible task. The targeted macromolecules were treated in a very simplified way because of the great amounts of computation required. Often only a portion of the whole system could be dealt with, solvent effects were rarely taken into account, and the simulation of complex formation could only be carried out for a small number of molecules. This picture has changed dramatically in the last decade mainly due to two factors. First, as Moore’s law stated in 1965, the number of transistors on a given chip has been doubling approximately every 2 years, with the subsequent impact on computer power. This has allowed an increase in the size of system that can be studied, the degree of accuracy of the models and the number of interactions feasible to calculate on a reasonable time scale. Second, there has been incredible progress in the experimental techniques that the different modelling tools rely on. X-ray crystallography and nuclear magnetic resonance (NMR) have been developed to a level where they are now applied routinely, which has had a tremendous impact on the number of experimentally determined molecular structures available. The number of both small molecules and macromolecules deposited in the Cambridge structural database and the protein data bank (PDB), respectively have increased dramatically. This wealth of experimental information has fuelled the refinement and application of many modelling tools, from force-field and scoring function development to homology modelling. The progress in the reliability of prediction, applicability of the different techniques and higher throughput capacity has enabled the application of structure-based molecular modelling in many phases of drug discovery, as will be reviewed below. First, a brief outline of all the methods is presented, with references to the standard publications in each field. Then, a number of different applications are discussed and structured according to the usual progress of a drug-discovery project. Senior Scientist, Vernalis (R&D), Granta Park, Abington, Cambridge, UK CB1 6GB Molecular Modelling, Grup Uriach, Polı́gon Industrial Riera de Caldes, Av. Camı́ Reial 51-57, 08184, Palau-solità i Plegamans (Barcelona), Spain { This is Chapter 3 taken from the book Structure Based Drug Discovery (Edited by Roderick E. Hubbard) which is part of the RSC Biomolecular Sciences series. { Current address: ICREA and Department de Fisicoquimica, Facultat de Farmacia, Universitat de Barcelona, Av. Joan XXIII, Barcelona s/n 08028, Spain BOOK CHAPTER www.rsc.org/biomolecularsciences | RSC Biomolecular Sciences series