Structure-based and property-compliant library design of 11β-HSD1 adamantyl amide inhibitors

Multiproperty lead optimization that satisfies multiple biological endpoints remains a challenge in the pursuit of viable drug candidates. Optimization of a given lead compound to one having a desired set of molecular attributes often involves a lengthy iterative process that utilizes existing information, tests hypotheses, and incorporates new data. Within the context of a data-rich corporate setting, computational tools and predictive models have provided the chemists a means for facilitating and streamlining this iterative design process. This chapter discloses an actual library design scenario for following up a lead compound that inhibits 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) enzyme. The application of computational tools and predictive models in the targeted library design of adamantyl amide 11β-HSD1 inhibitors is described. Specifically, the multiproperty profiling using our proprietary PGVL (Pfizer Global Virtual Library) Hub is discussed in conjunction with the structure-based component of the library design using our in-house docking tool AGDOCK. The docking simulations were based on a piecewise linear potential energy function in combination with an efficient evolutionary programming search engine. The library production protocols and results are also presented.