I came across a nifty Pfizer review published in JMC a couple of years ago about property-based optimization of drug candidates. There are many words of wisdom in it about various pharmacokinetic parameters and processes and related property optimization. One of the most interesting analyses I found was of gut absorption. Apparently:
1. Gut absorption is of two types; passive diffusion through lipid bilayers and passage through tight gap junctions. For drugs that undertake the latter route, size is obviously important. It turns out that dog GI gap junction pores are larger than in humans, so that a drug that passes readily through dogs won't pass readily through humans, highlighting the drawbacks of using dog models for studying human absorption. Also, Caco-2 cells seems to be bad models for drugs that undergo passage through this route. Fortunately, most drugs seem to pass via the first route, that is diffusion through lipid bilayers.
2. P-glycoprotein (PGP) and CYP3A4 (cytochrome P450 isozyme) are important in the gut, even if CYP3A4 expression is not as much as in the liver. The most interesting fact I read was that in case of PGP, the expression increases from top to bottom in the human digestive tract, with the least in the duodenum and the most in the colon. But for CYP the trend is opposite. Clearly this will influence structural considerations based on where the drug is going to be absorbed as well as drug delivery tactics.
3. CYP and PGP often work in concert in the gut, with PGP recycling substrates and offering them multiple times to CYP.
SAR for PGP and CYP definitely seems to be scarce, not surprisingly given that their substrate scope is quite large. Certain molecules such as the well-known immunosuppresants (eg. cyclosporin) have been identified to be rapidly metabolised by these enzymes. On the other hand, hyperforin from St. John's Wort can induce CYP expression and women on birth control pills might suddenly get a nasty surprise and find themselves pregnant; the hyperforin causes excessive metabolism of birth control molecules and reduces their blood levels to non-efficacious doses. Lipophilic drugs with electron rich groups can be choice fodder for CYP. Given the recent review highlighted by KinasePro which concluded that reducing lipophilicty should be a goal for avoiding poor absorption and promiscuous binding, this seems to be another good reason for losing the grease.