Aggravating Aggregation

Anyone interested in the field of high-throughput screening shouldn't miss this article which appeared online in the ASAP section of J. Med. Chem last week. Generally medicinal chemists can avoid false positives in screens by utilizing the well known Lipinski's Rule of Five or other computational methods that identify potential problematic molecules. Unfortunately, compounds that form colloidal aggregates are particularly troublesome; through sequestration of an enzyme from its substrate, these molecules usually appear to be good inhibitors (with IC50 values as low as 1 micromolar) with rather steep dose response curves. As aggregate-based inhibition is abrogated through the use of moderate concentrations of non-ionic detergents such as Triton X-100 (0.01 to 0.1%), Feng and coworkers developed an assay to test 70,563 compounds for detergent-sensitive inhibition. This screen has really opened my eyes to the prevalence of aggregators among screening hits. Astonishingly, of 1274 beta-lactamase inhibitors identified, 1204 were detergent sensitive, indicating an aggregation based mechanism of inhibition for 1.7% of the library! Anyone that has sorted through thousands or hundreds of initial hits will see the advantage of being able to identify or eliminate these artifacts from screens.

Discovering that a molecule is an aggregator is not a death sentence for its future use; as aggregation is concentration and condition dependent, molecules known to aggregate in one screen might not in a different setting. Additionally, several known drugs are aggregators at concentrations below 100 micromolar, including clotrimazole, nicardipine, delavirdine, and benzyl benzoate as pointed out by this 2003 article in J. Med. Chem.