In silico screening for novel inhibitors of human neutrophil elastase using an integrated approach of quantitative structure activity relationship modeling and protein-ligand docking

The discovery of drug like compounds that inhibit Human Neutrophil Elastase (HNE) could lead to therapeutic products for treating Alpha-1-Antitrypsin Deficiency (A1AD) and acquired forms of Chronic Obstructive Pulmonary Disease (COPD). Efforts have been made towards the development of therapeutic low molecular weight inhibitors for HNE, but as of now, no such compound has been approved by regulatory agencies for treating A1AD. Towards this end, a collaborative, inter-institutional team from the Bio-molecular Medicine Laboratory at Tennessee Tech University and the MeilerLab at Vanderbilt University has pursued a therapeutic drug discovery methodology consisting of a novel integrated computational approach that incorporates virtual high-throughput screening (vHTS), clustering, and protein-ligand docking. Machine learning methods, such as artificial neural networks and decision trees, have been applied to develop Quantitative Structure Activity Relationship (QSAR) models that predict the biological affinity of compounds for HNE based on their IC50 values. The QSAR models were used for vHTS, and the top 0.05% of compounds (125 in total) predicted with the highest binding affinity were selected for a clustering analysis. The clustering analysis compared the atom type and molecular structure of the predicted compounds with a known set of active inhibitors. From vHTS and clustering analysis, 20 compounds were selected from a chemical library of 213,323 such that the greatest difference in structure and atom type was identified to insure the highest probability of finding new chemotypes. These compounds were virtually docked into the HNE protein structure to prioritize them based on the predicted binding affinity and docked conformation. Experimental studies with these 20 compounds are being conducted to validate model predictions and document the potential of these compounds to inhibit HNE.