Synergistic Computational Techniques For Discovery of Novel Benzylidene-2,4-thiazolidinedione Derivatives as PTP1B Inhibitors

Protein tyrosine phosphatase 1B (PTP1B) acts as a negative regulator of insulin and leptin signaling pathways by dephosphorylating insulin receptor substrates. In recent years, it has emerged as a promising therapeutic target for the treatment of type 2 diabetes. A total of 48 benzylidene-2,4-thiazolidinedione derivatives were selected for in-silico study from the literature. Following the design of novel compounds, the dataset was utilized to develop 3D-QSAR and pharmacophore models; novel compounds were subsequently generated through R-group enumeration based on QSAR data and further evaluated using molecular docking and virtual screening. Among the generated pharmacophore and 3D-QSAR models, further validation identified the most predictive and statistically significant ones. The model AAANR_1 has been chosen as the best pharmacophore model. The 3D-QSAR study yielded the most successful statistical models, with the atom-based model’s correlation coefficient ([Formula: see text] [Formula: see text] 0.7741) and cross-validated correlation ([Formula: see text] [Formula: see text] 0.8229), and the field-based model’s correlation coefficient ([Formula: see text] [Formula: see text] 0.7652) and [Formula: see text] [Formula: see text] 0.6212. A total of 11,000 compounds were generated using the R-group enumeration study for the production of new compounds. The newly designed compounds were subjected to virtual screening, among which compounds a1, a2 and a3 demonstrated binding interactions with key amino acid residues ALA217 and ARG221 of the target protein (PDB ID: 1C83), similar to those of the co-crystallized ligand, with a docking score of −10.95 kcal/mol. In the molecular dynamics study, compound a3 indicates a much more rigid structure compared to a1 and a2. These molecules may also be utilized in the development of possible PTP1B inhibitors.