Machine-Learning-Driven Discovery of N-Phenylbenzenesulfonamides as a Novel Chemotype for Lactate Dehydrogenase A Inhibition with Anti-Pancreatic Cancer Activity

Lactate dehydrogenase A (LDHA) is a promising target for cancer therapy due to its crucial role in aerobic glycolysis. Despite extensive efforts, the structural diversity of LDHA inhibitors remains limited. Here, we utilized machine learning techniques, particularly reinforcement learning models, along with molecular dynamics simulations and biological validation, to identify N-phenylbenzenesulfonamides as a novel chemotype for LDHA inhibition. Compound 9 was generated and identified as a hit (IC₅₀ = 720 nM), with its sulfonamide moiety forming crucial hydrogen-bonding interactions with LDHA. Structural optimization led to derivatives with enhanced LDHA inhibitory activity, exemplified by compound 28 (IC₅₀ = 156 nM). Compound 28 significantly reduced lactate production and induced apoptosis in pancreatic Mia PaCa-2 cells and demonstrated robust antitumor effects in vivo following oral administration at 100 mg/kg, with no apparent toxicity observed. These findings position 28 as a promising LDHA inhibitor with a novel chemotype for pancreatic cancer treatment.