Pharmacophore-guided machine learning and molecular simulations of pyrazole-imine ligands targeting estrogen receptor alpha

Estrogen receptor alpha (ERα) plays a crucial role in breast cancer progression, making it a key target for selective estrogen receptor modulators (SERMs). While Raloxifene has demonstrated therapeutic efficacy, resistance and limited bioactivity in some cases necessitate the development of novel ERα inhibitors with improved pharmacological profiles. A computational drug discovery approach integrating molecular docking, molecular dynamics (MD) simulations, and quantitative structure-activity relationship (QSAR) modeling was employed to design and evaluate new ERα inhibitors. Molecular docking was performed using Glide (XP mode) to predict ligand binding affinity and interaction patterns, while MD simulations over 100 ns assessed the stability and conformational dynamics of the protein-ligand complexes. A QSAR model was developed using a dataset of 1,231 compounds from ChEMBL, incorporating XGBoost regression with optimized hyperparameters for robust predictive performance. Compounds 3b, 3a, and 4a showed notable binding affinities (−9.319, −9.121, and −8.867 kcal/mol, respectively) that are comparable to Raloxifene (−9.791 kcal/mol), suggesting their potential as effective ERα ligands primarily through pi-pi stacking with PHE-404 and hydrogen bonding with Glu 353. MD simulations demonstrated that 3a, 4a, and 4b maintained stable receptor interactions (RMSD 0.8), identifying NO₂ (3d), OMe (3c), and Cl (3b) substitutions as key structural features enhancing receptor binding. This study identifies 3b, 3d, and 4b as promising lead compounds with strong binding affinity, stability, and predicted estrogen receptor activity. These findings provide a basis for further experimental validation and structural refinement to develop next-generation SERMs for ERα-positive breast cancer therapy.