Bioisostere replacement is a powerful and popular tool used to optimize the potency and selectivity of candidate molecules in drug discovery. Selecting the right bioisosteres to invest resources in for synthesis and subsequent optimization is key to an efficient drug discovery project. In this retrospective study, we used human aldose reductase inhibitors to demonstrate an active learning workflow that prioritizes molecules from a large pool of bioisostere replacements generated by Spark™. This workflow combines two rigorous computational approaches: 3D-quantitative structure activity relationships (3D-QSAR) with shape and electrostatic descriptors¹, and free energy perturbation (FEP) for binding free energy calculations in Flare™. This workflow can rapidly locate the strongest-binding bioisosteric replacements with a relatively modest computational cost (a total of only 16% of the candidate pool was processed with FEP requiring 20% or even less GPU hours than if FEP were to include all candidates). The ROC-AUC for selection of known actives in 80 top-ranked candidates improved to 0.88 from 0.64, and the top picks were enriched with highly potent ALR2 inhibitors, including the well-known clinical candidate Zopolrestat developed by Pfizer.^2,3

Read our open access Active learning FEP paper in ACS Medicinal Chemistry Letters.

References

• Cheeseright, T., et al., Molecular field extrema as descriptors of biological activity: definition and validation. J Chem Inf Model., 2006, 46, 2, 665-6 DOI: 10.1021/ci050357s.
• Mylari, B.L., et al., Novel, potent aldose reductase inhibitors: 3,4-dihydro-4-oxo-3-[[5-(trifluoromethyl)-2-benzothiazolyl] methyl]-1-phthalazineacetic acid (zopolrestat) and congeners. J Med Chem., 1991, 34, 1, 108-1 DOI: 10.1021/jm00105a018.
• Mylari, B.L., et al., Potent, orally active aldose reductase inhibitors related to zopolrestat: surrogates for benzothiazole side chain. J Med Chem., 1992, 35, 3, 457-4 DOI: 10.1021/jm00081a006.