Integrated machine learning-based virtual screening and biological evaluation for identification of potential inhibitors against cathepsin K.

Cathepsin K is a type of cysteine proteinase that is primarily expressed in osteoclasts and has a key role in the breakdown of bone matrix protein during bone resorption. Many studies suggest that the deficiency of cathepsin K is concomitant with a suppression of osteoclast functioning, therefore rendering the resorptive properties of cathepsin K the most prominent target for osteoporosis. This innovative work has identified a novel anti-osteoporotic agent against Cathepsin K by using a comparison of machine learning and deep learning-based virtual screening followed by their biological evaluation. Out of ten shortlisted compounds, five of the compounds (JFD02945, JFD02944, RJC01981, KM08968 and SB01934) exhibit more than 50% inhibition of the Cathepsin K activity at 0.1 µM concentration and are considered to have a promising inhibitory effect against Cathepsin K. The comprehensive docking, MD simulation, and MM/PBSA investigations affirm the stable and effective interaction of these compounds with Cathepsin K to inhibit its function. Furthermore, the compounds RJC01981, KM08968 and SB01934 are represented to have promising anti-osteoporotic properties for the management of osteoporosis owing to their significantly well predicted ADMET properties.

Identification of novel TACE inhibitors using DNN based- virtual screening, molecular dynamics and biological evaluation.

Rheumatoid Arthritis (RA) is a well-known autoimmune inflammatory disease, distressing roughly 1% of the adult population throughout the globe. Many studies have suggested that overexpression of TNF-α, a pro-inflammatory cytokine, is responsible for the progression of RA. Furthermore, inhibition of the shedding rate of TNF-α is regulated by the TACE (TNF-α converting enzyme) protein and, hence is considered as an important therapeutic target for the prevention of progressive synovial joint destruction in rheumatoid arthritis. In the present study, we have proposed a deep neural network (DNN)-based workflow for the virtual screening of compounds towards the identification of potential inhibitors against the TACE proteins. Subsequently, a set of compounds were shortlisted, based on the molecular docking and subjected to the biological evaluation to validate the inhibitory activities of the screened compounds, determine the practical applicability of the DNN-based model, and strengthen the hypothesis. Out of seven, three compounds (BTB10246, BTB10247, and BTB10245) showed significant inhibition at 10 µM and 0.1 µM concentration. These three compounds also showed a stable and significant interaction potential against the TACE protein as compared with the re-docked complex system and can serve as a novel scaffold for further design of new molecules with improved inhibitory activities against TACE.Communicated by Ramaswamy H. Sarma.