Leveraging Scaffold Hopping and Machine Learning for the Innovative Design and Assessment of Pyridine Derivatives as c-Met Inhibitors Using Amide-based Bioisosteres

Background: c-Met is a critical target for cancer therapy in different cancers. Selective inhibitors of c-Met present promising strategies for the suppression of associated signaling pathways, including the HGF/c-MET. Targeted c-Met inhibitors pose greater specificity and decreased contact with off-target biomolecules. Methods: Scaffold hopping was employed to design new c-Met inhibitors. Two pyridine derivatives ([1,2,4]triazolo[4,3-b]pyridazine and 3H-pyrrolo[2,3-b]pyridine) were used as the template and amide-based bioisosteres were utilized as the cores. The c-Met binding site served as the basis for screening the designed compounds, and Machine Learning (ML) models were subsequently created to assess the molecules that were screened. The obtained molecules were then evaluated using ADME/T analysis, Molecular Dynamics (MD) simulations and binding free energy calculations. Results: A total of approximately 4000 newly developed inhibitors were produced, from which six compounds were identified as hit compounds due to their binding affinity to c-Met. The activity of the compounds was assessed against c-Met using ML, leading to the selection of C1 and C3 as the most promising candidates for subsequent evaluation through ADME/T analysis and MD simulation. The results suggested that C1 is a drug-like compound and can target crucial residues such as Met1160, Asp1164, Ile1084, Met1211 and Asp1230. Conclusion: According to the findings, C1 can be considered a favorable drug candidate as a c-Met inhibitor by precisely targeting the essential residues in the binding site of c-Met. This compound warrants further investigation through both in vitro and in vivo research methodologies.