The pharmaceutical industry is “taking too long, spending too much and producing far too little.” This harsh criticism came from someone with intimate knowledge of the industry, John Lechleiter, CEO of Eli Lilly until his retirement at the end of 2016. The analysis was taken up in The Future of Drug Discovery, a publication in which the former executive spared no irony in his assessment of the problems and challenges facing biomedical research. “The crisis in our innovation model comes at a time when we have vastly more scientific knowledge and data than ever, but unless we change the way we do research, we won’t translate this knowledge into advances,” said Lechleiter.
There can be no doubt that there exists an urgent need to develop technologies that substantially improve the efficacy and efficiency of the drug discovery process. This pressure becomes all the more accentuated taking into account the serious problem of antibiotics resistance, a struggle in which the core facilities can play a key role. In this context, researchers at the University of Barcelona have developed a computational method to identify new drugs. Their idea, christened “dynamic undocking” (DUck), was published in the Nature Chemistry journal, and may complement conventional tools used in the discovery of new molecules with biologic activity.
A tool to design drugs more efficiently
Today, most of the approaches that predict whether a molecule will bind to a protein are based on the analysis of the complex’s thermodynamic stability. In other words, its binding affinity is analyzed. Instead of focusing on the situation of equilibrium, in which the two molecules form the best possible interactions, the University of Barcelona team implemented its computational method with the goal of evaluating the complex breakage process, to analyze the possible breaking points and this way determine how the molecules could be improved to make them more resistant to separation. Specifically, DUck is a computational analysis method that calculates the work necessary to reach the quasi-bond state (WQB), at which the ligand has just broken the most important native contact with the receptor.
As an initial proof of concept, scientists used DUck to study a catalog of 41 fragments similar to cyclin-dependent kinase 2 (CDK2) ligands for which the values for the form of binding and half maximal inhibitory concentration (IC50) are known. This process uncovered the surprising relationship that exists between affinity and the quasi-bond state, when they observed that the ligands that bonded to the receptor with the greatest affinity presented a higher WQB value. Once this correlation was proven, the University of Barcelona researchers focused on testing whether the analysis of the breaking process for a complex could be useful in the virtual screening of new molecules, and consequently, drug discovery. To conduct the experiment, DUck was used to distinguish between true ligands and a catalog of decoys designed by the scientists that bound to receptors, like CDK2, serine proteases and the adenosine A2A receptor, as representatives of protein families that are highly relevant in pharmacological research.
From left to right, Francisco Javier Luque, Sergio Ruiz Carmona and Xavier Barril. Source: University of Barcelona.
“This non-equilibrium property is surprisingly effective in virtual screening because true ligands form more-resilient interactions than decoys. Notably, DUck is orthogonal to docking and other ‘thermodynamic’ methods”, says the team led by Dr. Xavier Barril, of the Faculty of Pharmacy and Food Sciences and the Institute of Biomedicine of the University of Barcelona (IBUB). By focusing on a parameter that is different from affinity, the DUck computational method has proven to offer results complementary to those obtained by other techniques. Specifically, the scientists examined the calculations that resulted from other virtual screening systems, such as Glide docking, molecular mechanics Poisson–Boltzmann surface area (MMPBSA) and generalized Born surface area (MMGBSA) rescoring, and concluded that in some cases DUck achieved better analyses. “These results support the idea that structural stability of the binding mode, just like good chemical complementarity, is a necessary—but not sufficient—condition for binding. By imposing both conditions simultaneously, we can multiply the effectiveness of structure-based virtual screening,” the researchers state in an article published in Nature Chemistry.
Lastly, the team assessed the potential of the DUck computational method to identify small molecules that bind to the Hsp90 chaperone. This oncological target, of which hundreds of ligands are known, is a protein of interest and major relevance in the fight against cancer. In the evaluation of the various molecules through the quasi-bond state, scientists could confirm that their computational method multiplies the efficiency of virtual screening by approximately one order of magnitude, achieving hit rates of nearly 40%. Their results show that the DUck tool generates few false negatives, which would correspond to active molecules with a low WQB. Their work indicates that structural stability can be a fundamental parameter in the design of new molecules of pharmacological interest, which complements the results obtained studying the affinity of complexes. “What we demonstrate is that molecules must also form structurally stable (rigid) complexes, and that it is possible to distinguish between active and inactive simply by checking how difficult it is to break certain specific interactions,” concludes Barril. His team is already evaluating the method’s potential in the research and development of new medications against cancer or infectious diseases.
Source: Tirkfl (Wikimedia)
It is not the first time that computational methods are applied to advance in studies related with pharmacological research. Modeling and computation have also been used to reduce the usage of experimental animals in laboratories, as Dr. Jordi Quintana, director of the Drug Discovery Platform of the Scientific Park of Barcelona, told Biocores. In the private sector, examples such as Nostrum BioDiscovery, a spin-off backed by the Institute for Research in Biomedicine of Barcelona (IRB) and the Barcelona Supercomputing Center-Centro Nacional de Supercomputación (BSC-CNS), show that supercomputing can be applied to accelerate the research and development of new drugs.