Abstract
<title>Abstract</title> <p>Pancreatic cancer remains a highly lethal malignancy with limited targeted therapeutic options, necessitating innovative drug discovery strategies. The epidermal growth factor receptor (EGFR) is a key driver of tumor progression and a validated therapeutic target. Here, we report an integrated generative artificial intelligence (AI)-driven framework for the discovery of novel EGFR inhibitors, leading to the identification of EGFPAN_EMBS as a chemically distinct and potent candidate. A cannabigerol-inspired scaffold was used as the initial seed, and deep generative models, including variational autoencoders, generative adversarial networks, and reinforcement learning, were employed to explore a chemical space exceeding 1.7 million compounds. Lead prioritization based on drug-likeness, pharmacokinetic properties, and target specificity identified EGFPAN_EMBS as a top candidate. Docking studies revealed improved binding affinity compared to cannabigerol, supported by a well-defined interaction network involving key residues within the ATP-binding pocket. Molecular dynamics simulations demonstrated enhanced structural stability, reduced conformational fluctuations, and persistent intermolecular interactions. Free energy landscape and dynamical cross-correlation analyses indicated confinement to a low-energy conformational basin and reduced long-range correlated motions. Binding free energy calculations further confirmed stronger interaction stability for EGFPAN_EMBS. Experimental validation showed potent nanomolar inhibition of EGFR kinase activity, favorable binding kinetics, and strong antiproliferative effects across pancreatic cancer cell lines with high selectivity. Mechanistically, EGFPAN_EMBS suppressed EGFR autophosphorylation and downstream PI3K–AKT–MAPK signaling, inducing apoptosis and G₁-phase cell-cycle arrest.Collectively, EGFPAN_EMBS represents a promising ATP-competitive EGFR inhibitor and highlights the potential of AI-driven discovery pipelines for next-generation anticancer therapeutics. The compound EGFPAN_EMBS has been deposited in the NCBI PubChem database under Compound ID 10972272 and Substance ID 528216732.</p>