Chris Lipinski, Massechusetts Uni
Last Updated on Tuesday, 08 May 2012 11:55 Written by Jackie Howard Tuesday, 08 May 2012 10:48
Biologically active chemistry space and the limits of druggability
There is no evidence that biologically active compounds are uniformly distributed throughout chemistry space and there is a lot of evidence to the contrary. As a result, screening truly diverse chemical libraries is the worst way to discover a drug. Poly-pharmacology in biologically active compounds is the rule rather than the exception. The predominant mode of drug discovery for at least two decades has been to discover a single compound that is selective against a mechanistically defined single biology target. That this approach works at all is often a matter of good luck. Complete signaling pathway blockages elicit no phenotypic change 85-90% of the time. This is a consequence of the robust nature of biological signaling networks. By contrast two (or more) modest interventions in a signaling pathway are much more likely to achieve an observable response. The current drug discovery success rate of only 1 in 64 projects is leading to a re-examination of the entire early drug discovery process. The deficits of reductionist thinking are gradually being understood and there is increased appreciation of the merits of phenotypic screening, poly-pharmacology and drug repurposing. Natural products provide the best lessons on druggability possibilities in chemical space. However, discerning general principles is difficult. In contrast to synthetic medicinal chemistry compounds, natural products are not easily categorized by three dimensionally defined scaffold structures. The ability of some natural product drugs to defy druggability rules like the “rule of 5” may in part be explained by their chameleon like physicochemical properties as exemplified by recent studies with Cyclosporin A.