Targeting Cathepsin L in Cancer Management: Leveraging Machine Learning, Structure-Based Virtual Screening, and Molecular Dynamics Studies.

Cathepsin L (CTSL) expression is dysregulated in a variety of cancers. Extensive empirical evidence indicates their direct participation in cancer growth, angiogenic processes, metastatic dissemination, and the development of treatment resistance. Currently, no natural CTSL inhibitors are approved for clinical use. Consequently, the development of novel CTSL inhibition strategies is an urgent necessity. In this study, a combined machine learning (ML) and structure-based virtual screening strategy was employed to identify potential natural CTSL inhibitors. The random forest ML model was trained on IC50 values. The accuracy of the trained model was over 90%. Furthermore, we used this ML model to screen the Biopurify and Targetmol natural compound libraries, yielding 149 hits with prediction scores >0.6. These hits were subsequently selected for virtual screening using a structure-based approach, yielding 13 hits with higher binding affinity compared to the positive control (AZ12878478). Two of these hits, ZINC4097985 and ZINC4098355, have been shown to strongly bind CTSL proteins. In addition to drug-like properties, both compounds demonstrated high affinity, ligand efficiency, and specificity for the CTSL binding pocket. Furthermore, in molecular dynamics simulations spanning 200 ns, these compounds formed stable protein-ligand complexes. ZINC4097985 and ZINC4098355 can be considered promising candidates for CTSL inhibition after experimental validation, with the potential to provide therapeutic benefits in cancer management.

Identifying promising GSK3ß inhibitors for cancer management: a computational pipeline combining virtual screening and molecular dynamics simulations.

Glycogen synthase kinase-3 (GSK3ß), a serine/threonine protein kinase, has been discovered as a novel target for anticancer drugs. Although GSK3ß is involved in multiple pathways linked to the etiology of various cancers, no specific GSK3ß inhibitor has been authorized for cancer therapy. Most of its inhibitors have toxicity effects therefore, there is a need to develop safe and more potent inhibitors. In this study, a library of 4,222 anti-cancer compounds underwent rigorous computational screening to identify potential candidates for targeting the binding pocket of GSK3ß. The screening process involved various stages, including docking-based virtual screening, physicochemical and ADMET analysis, and molecular dynamics simulations. Ultimately, two hit compounds, BMS-754807 and GSK429286A, were identified as having high binding affinities to GSK3ß. BMS-754807 and GSK429286A exhibited binding affinities of -11.9, and -9.8 kcal/mol, respectively, which were greater than that of the positive control (-7.6 kcal/mol). Further, molecular dynamics simulations for 100 ns were employed to optimize the interaction between the compounds and GSK3ß, and the simulations demonstrated that the interaction was stable and consistent throughout the study. These hits were also anticipated to have good drug-like properties. Finally, this study suggests that BMS-754807 and GSK429286A may undergo experimental validation to evaluate their potential as cancer treatments in clinical settings.