Remote-Url: https://www.science.org/content/blog-post/discovering-paxlovid
Retrieved-at: 2022-01-20 08:52:17.001427+00:00

I'd like to recommendthis articleatC&E Newsby Bethany Halford on the development of Pfizer's coronavirus protease inhibitor, Paxlovid (nirmatrelvir). You'll see how the company had a terrific leg up on the problem via its work on the earlier SARS virus, which led to a protease inhibitor candidate that was never used, since the virus had disappeared from the human population by the time it was developed. The graphic at left (viaC&E News) shows Paxlovid at top, then the earlier compound, and at bottom an advanced candidate that dropped out in favor of the final structure.It's a good look at the sort of optimization that goes on in any med-chem project. The SARS compound was not orally available, for example, and the team knew that they had to fix that in order to make a real impact on the pandemic. Oral availability and gut permeability is a problem with several levels, but in this case a clear place to start was getting rid of the (many) hydrogen bonds in that earlier candidate. Pfizer's Dafydd Owen is also quoted in the article about that fused cyclopropyl ring, saying (and he's absolutely right, of course) that "You either win big or you lose big" when you tie things back into a ring in a med-chem optimization. But in this case, it looks like the potency didn't change much (in fact, it went down just a bit), but that was still a win, because it got rid of the leucine from the earlier compound and tying things back onto the next nitrogen got rid of another hydrogen bond donor.Losing that one, though, meant losing a critical contact in the binding site, which is where that trifluoroacetamide came in. The team tried a number of different groups up there to pick up a new interaction where that methoxyindole used to be, and the TFA amide was the best in oral bioavailbility and potency. The article quotes a former colleague of mine, Jeremy Green, to the effect that this wouldn't be most chemist's first choice, and he's sure right about that - I'm sure many medicinal chemists' eyebrows went up when they first saw the Paxlovid structure. Trifluoroacetamides are just about the easiest amides to cleave off a nitrogen, actually; they're quite base-sensitive, but in the stomach and small intestine they're not going to hit the pH that will do the job (and human blood is very carefully buffered so it doesn't get that basic, either).In fact, Pfizer hasjust today sharedmore preclinical information about the compound. Its affinity to the protease is good (Ki about 1 nanomolar), and it maintains potency against the Omicron variant just fine. Pfizer says that its data suggest the the compound should be able to maintain blood levels many times higher than needed to stop Omicron replication in cells, so the pharmacokinetics are in the right range, trifluoroacetamide notwithstanding.Finally, there was the "warhead" group. It turns out that the chiral center next to the benzothiazole compound was more likely to epimerize than the one next to the nitrile, and the latter looked far easier to scale up. That last detail is, of course, a tremendous tiebreaker in a situation like this. Pfizer's chemists spent the whole optimization of the compound constantly looking over their shoulders at the manufacturing issues that each change would introduce, and I'm told that there were a number of ideas that they just didn't even bother with because they would have introduced too many scale-up problems compared to the available alternatives. Paxlovid itself, the article tells us, was first synthesized on July 22nd of last year, with 1.4 kilos prepared by earlier November for tox studies. The fact that it's already being shipped out to pharmacies is mind-boggling. Production is of course still ramping up (seethis blog postfor some thoughts on that process), but this is the hands-down all-time champion example of accelerated drug development, and I know of nothing that even comes close.But remember, it was only possible because of that earlier work on the closely related SARS protease - and remember, at the time (and for some years thereafter!) that work looked like wasted effort for the most part. After all, there was no SARS to treat, and there were no other related coronaviruses out there causing trouble. But that's science: once you discover something, it stays discovered (or it had better, anyway), and that knowledge is out there waiting to be put to a new use. And also remember that there was some luck involved here, too, because there always is. You can bet that the Pfizer team was practically holding their breath as this compound went into tox studies, because (as I like to say) if you're not holding your breath then you haven't been doing drug discovery for very long.Anything can happenas you go into animal studies and into human trials, and we are very fortunate that this compound passed both of them cleanly. That's another reason this will surely remain the record holder for fast development: Paxlovid had as big a head start as you could possibly have, it is in the therapeutic area that has the fastet and most relevant preclinical models of anything in drug discovery, and it dodged every out-of-the-blue failure mode once it went into development. It was primed for quick emergency authorization by the FDA. And it had the backing of a huge company with huge resources, with thousands of very experienced and highly competent people going flat-out to make this drug a reality.So (unfortunately) Paxlovid does not herald some new era where it's just going to take us a few months to crank out the wonder drugs. No, we have the same issues that we always do. But everything came together for this one!