TORONTO—For the last decade, Lipinski's Rule of Five has set the standards for scientists looking to screen small-molecule libraries for druglike compounds. Recently, however, researchers at Advanced Chemistry Development (ACD/Labs) wondered if people aren't throwing out too many babies with the bathwater by using the partition coefficient logP as a measure of lipophilicity. They argue that because most small-molecule drugs include ionizable moieties, the pH-dependent version of logP—logD—might be more appropriate.
As they described in Molecular Pharmaceutics, the researchers examined several drug libraries—including those from ChemDiv, Enamine, Specs, and ComGenex—calculating logP, logD at pH5.5 (the acidity of the small intestine), and the other Rule of Five parameters. They then looked to see how many compounds would pass as being druglike.
They found that using logD in place of logP, anywhere from 0.6% to 2.3% more compounds were classified as druglike. Furthermore, for two libraries, 10% of known drugs would have been excluded for excessive logP values, but more than half of these compounds would have been rescued from the scrap heap using logD.
"While these numbers may initially seem insignificant, when taken in the context of library screening, the 0.6% to 2.3% increase in druglike permeability represents 600-2300 additional leads for every 100,000 compounds screened," the authors write. "Even a small chance of excluding a potential blockbuster drug as a false-negative hit from virtual screening due to unrealistic values is unacceptable."
