Box 1. Rules of engagement.


Chris Lipinski and colleagues at Pfizer Inc. defined the 'rule of five' after analyzing the physicochemical properties of over 2,000 drugs.1 They proposed that compounds would be more likely to be orally available if they had: (1) molecular weight lower than 500 Da, (2) hydrophobicity measure log P less than 5,
(3) fewer than 5 hydrogen bond donors and (4) fewer than 10 hydrogen bond acceptors. All the numbers in the four listed criteria are multiples of five-hence the name 'rule of five'.

The rule of five remains a topic of debate among medicinal chemists who employ it broadly but continue to discuss its pros and cons.2 SciBX addressed this in the conversation with Lipinski.


SciBX: What are the main limitations of using the rule of five in drug discovery?


Chris Lipinski: The problem with the rule of five and with other rules is that they can be misused. The rule of five is a rule with hard cutoffs: log P greater than 5 is bad; log P less than 5 is good. To say a compound with log P of 4.99 is acceptable and with log P [of] 5.01 isn't makes no sense whatsoever, so you need to use common sense.

If your strategy is to screen out anything with a log P greater than 3 and a molecular weight greater than 400 you'll discover many GPCR ligands, phosphodiesterase targets and some kinases, but you will miss just about every possible protein-protein interaction target and you'll have a lot of trouble with peptidergic GPCRs and protease targets.

So by being too limiting on the rules you eliminate great swaths of target space that you'd really like to interrogate.