Jul 3, 2025

Credit: Insilico Medicine Aberrations in fibroblast growth factor receptors FGFR2/3 are major drivers of many cancers—including intrahepatic cholangiocarcinoma, endometrial, breast, gastric, and bladder cancers. However, the development of selective inhibitors targeted FGFR2/3 has long been challenged by the high sequence and structural similarity among FGFR family members, particularly the binding pocket homology among FGFR1, FGFR2, and FGFR3 exceeds 95%. Currently approved pan-FGFR inhibitors lack selectivity and often inhibit FGFR1/4 as well, leading to dose-limiting toxicities such as hyperphosphatemia and diarrhea, reduced treatment efficacy. Compounding these challenges, the emergence of resistance-conferring mutations in FGFR2/3 frequently diminishes the effectiveness of current drugs. To address this challenge, the Insilico team employed advanced molecular modeling techniques in combination with the company’s proprietary generative AI-based chemistry engine, Chemistry42, to efficiently design highly selective FGFR2/3 inhibitors featuring a novel core structure. This research, recently published in the Journal of Medicinal Chemistry , provides an innovative approach for the discovery and development of next-generation cancer therapies targeting FGFR2/3. The process began with structural analysis and molecular modeling, which identified the hinge-binding region (core A) as a key contributor to FGFR2/3 binding. Based on these insights, the researchers defined a pharmacophore model that retained essential hydrogen bond donors and acceptors, as well as relevant features for selectivity and potency. The Chemistry42 platform was then leveraged to generate a library of approximately 10,000 molecules with diverse core structures and linkers, prioritizing those predicted to exhibit strong protein-ligand interactions (as measured by PLI score) and favorable drug-like properties. Based on the generated amide-based scaffold, we found that flexible binding to the kinase hinge region can better accommodate various types of resistance mutations. By filtering for optimal molecular characteristics and high PLI scores, the team identified core3 (C3) as a structural motif likely to deliver high potency and selectivity for FGFR2/3. Further ADMET prediction and Alchemistry—the free energy calculations module of Chemistry42—were subsequently used to rank and optimize the candidate molecules. This process ultimately led to the identification of ISM7594, a covalent dual FGFR2/3 inhibitor distinguished by its unique hinge-binding motif and novel core structure. In the validation studies, ISM7594 exhibits nanomolar inhibitory activity against FGFR2 and FGFR3, with over 100-fold selectivity relative to FGFR1/4. It maintained strong efficacy against clinically relevant FGFR2/3 mutants associated with therapeutic resistance. In cancer cell lines harboring FGFR2/3 alterations, ISM7594 showed robust antiproliferative effects, while exhibiting minimal impact on cells without FGFR aberrations. In preclinical animal models, ISM7594 displayed favorable pharmacokinetic properties, significant tumor growth inhibition, and a reduced toxicity profile compared to less selective FGFR inhibitors. “Our study not only demonstrates the speed and precision of AI-enabled drug design but also the importance of rigorous experimental validation to translate in silico discoveries into real, clinically relevant therapies,” said Xiao Ding, PhD, Head of Chemistry & DMPK and Senior Vice President of Medicinal Chemistry at Insilico Medicine. In February 2025, Insilico published a research paper in the Journal of Medicinal Chemistry entitled “Discovery of Pyrrolopyrazine Carboxamide Derivatives as Potent and Selective FGFR2/3 Inhibitors that Overcome Mutant Resistance,” further demonstrating the process of core structure optimization for novel, highly selective FGFR2/3 inhibitors. Since its founding in 2014, Insilico Medicine has published more than 200 peer-reviewed papers. Leveraging sustained scientific breakthroughs at the intersection of biotechnology, artificial intelligence, and automation, Insilico Medicine secured a position among the Top 100 global corporate institutions for biological sciences and natural sciences publications, and ranked 43rd among global corporate institutions in the USA across all-subject output. Reference [1] Wang, Y. et al. (2025) 'Rational design and identification of ISM7594 as a Tissue-Agnostic FGFR2/3 inhibitor,' Journal of Medicinal Chemistry [Preprint]. https://doi.org/10.1021/acs.jmedchem.5c00928 . About Insilico Medicine Insilico Medicine, a global clinical stage biotechnology company powered by generative AI, is connecting biology, chemistry, medicine and science research using next-generation AI systems. The company has developed AI platforms that utilize deep generative models, reinforcement learning, transformers, and other modern machine learning techniques for novel target discovery and the generation of novel molecular structures with desired properties. Insilico Medicine is developing breakthrough solutions to discover and develop innovative drugs for cancer, fibrosis, central nervous system diseases, infectious diseases, autoimmune diseases, and aging-related diseases. For more information, please visit www.insilico.com . Journal