A new approach that tackles a key issue in medicinal chemistry — how to achieve maximum diversity in chemical libraries used for screening, while minimizing the size of the libraries — has been recently reported in the Journal of Medicinal Chemistry. The developers of the strategy, Didier Rognan and colleagues, describe how it can be used to increase the potency of a known phosphodiesterase 4 (PDE4) inhibitor by a factor of ∼103 by computationally evaluating just 320 compounds, and synthesizing and testing only 9.

The authors' approach, dubbed the SLF approach, is based on the combinatorial assembly of three types of molecular 'building blocks': a conserved scaffold (S), a variable linker (L) and a variable functional group (F). Importantly, and in contrast to some other general strategies for combinatorial scaffold-based library design, all the building blocks are user-selected. In particular, carefully selecting a limited number of functional groups that do not overlap in 'pharmacophoric space' helps to create maximal chemical diversity while keeping the overall library size down. Furthermore, the combinations were also selected so as to ensure synthetic accessibility.

To demonstrate their strategy, Rognan et al. used the scaffold of the known PDE4 inhibitor zardaverine to build a virtual library by combining 4 closely related such scaffolds, 5 linkers and 16 functional groups. The 320 resultant molecules were then docked into a crystal structure of PDE4, and ranked by predicted binding strength. Inspection of high-ranking molecules suggested that two additional binding pockets in PDE4 not occupied by zardaverine could be targeted by a number of such compounds in the library, and nine of these compounds were selected and synthesized for testing.

Of the nine compounds, five were stronger inhibitors of PDE4 in vitro than zardaverine, and one had an IC50 value of 0.9 nM — ∼900-fold greater than zardaverine. Also, as one of the additional binding pockets identified by the authors shows some degree of variation between different PDEs, these compounds could represent good starting points for attempts to overcome the selectivity problems that have so far limited the clinical development of PDE4 inhibitors. And more generally, the ability of the SLF approach to facilitate rapid and efficient lead optimization could lead to its application for many other targets.