Let’s catch up with what’s going on with Paxlovid, the Pfizer drug that inhibits the Main Protease (MPro) of the coronavirus. Since Merck’s molnupiravir has shown disappointing clinical trial results, Paxlovid is also the main small molecule drug that’s being used to treat the infection.
A brief review, for folks who don’t sit around thinking about small-molecule antivirals all the time. I mean, what are you doing with your life if you’re not, but anyway. Viruses tend to arrive very compactly packaged, and SARS-CoV-2 is a typical example. Remember, when a viral particle infects a cell, its goal is to hijack the cell’s gene-expression and protein-production machinery to make more viral particles. In general, the only things that a virus is ever doing is infecting host species, entering their cells, and forcing those cells to make more viral particles – which then go out and infect more hosts, and so on and so on. All the other details are in service of those goals.
Coronaviruses are RNA viruses, so a viral particle unspools a length of RNA for the ribosomes to take up and produce a protein according to those instructions. The “business end” of that long protein, once it emerges, is a functional protease, a protein-cleaving enzyme that targets the long viral protein itself, breaking it up into the pieces that will need to be assembled into a new viral particle. This sort of bootstrapping approach is a constant feature of viral replication. One of those pieces that gets produced this way is itself another enzyme, a viral RNA polymerase that goes out and assembles more coronavirus RNA to pack into those particles, so you can see that those two enzymes are absolutely crucial to viral replication. Paxlovid is a small-molecule inhibitor of the protease, and molnupiravir is a small-molecule inhibitor of the RNA polymerase. Over and over in antiviral research you will see people targeting such polymerases and “main proteases”; it’s one of the first things you try. There are other targets – things that keep the virus from entering the cell, or which gum up the fitting of the protein pieces as they assemble into new viral particles, for example – but for a new viral pathogen going after the protease and the polymerase are generally at the top of the list.
Now, in its initial clinical trials, Paxlovid showed efficacy in preventing hospitalization or death in unvaccinated patients, both high-risk ones and standard-risk ones. A big question has been whether it can actual prevent disease, rather than treat it after it’s already there – and unfortunately, it looks like we have the answer to that one. Pfizer has announced results of a 3,000 patient trial in people who had someone in their household diagnosed with coronavirus – these people were randomized to get a five-day course of Paxlovid, a ten-day one, or placebo. And there was no difference between the three groups; people got infected at basically the same rate. So prophylactic Paxlovid is not a thing.
The first-order reaction to that is that it’s too bad, because it would be useful indeed to have something that people could take to lower their chances of infection. The second-order take is that any drug that attained wide use in that fashion would just lead to viral resistance developing even faster than before, so this news might not (in the long run) be as sad as it sounds. Remember, all these squintillions of coronavirus particles are constantly throwing off mutations in their RNA sequences as they replicate, and these mutations are getting immediate real-world testing as they get a chance to prove their fitness in reaching those goals mentioned above: infecting new hosts, entering their cells, and producing more viral particles. Every step towards those – such as abilities to evade the human immune system in all its forms (antibodies, T cells, what have you) or spread more easily due to (perhaps) the area of the airway that’s being infected – are also under that same relentless selection pressure. That’s why we keep seeing new varieties of SARS-CoV-2. It’s a new pathogen in humans and none of these things had been optimized yet, so we get to watch it happening in real time, damn it all.
That selection pressure applies (very directly indeed) to small-molecule antiviral therapy. You are challenging the virus directly by interfering with the function of one region of one key protein, so if there are any possible random mutations that allow said protein to function while not being susceptible to the antiviral drug, well, that’s what you’re going to get. You are directly creating the conditions for that to happen, and every drug that targets some protein in a virus (or a bacterium) is asking for it in the same way. The game starts immediately. Do such drug-evading mutations exist? Do they impair the function of the protein or not? Do they affect the other fitness goals of the organism or not? All of these are on a sliding scale, and every patient treated is another opportunity for the pathogen to try the combinations. The good news is that so far, the recent mutational variants of SARS-CoV-2 are not showing changes in the main protease sequence that would affect Paxlovid’s efficacy, but has to be watched closely.
This is why I was really unhappy with the molnupiravir results, as I said at the time. Really successful antiviral treatment regimes (of which there are very few) invariably are cocktails of drugs that hit the virus in several different places at once. The chances of a particular viral particle hitting the combination of mutations that would get it out of more than one jam simultaneously are clearly much lower, so a protease/polymerase inhibitor combination is an obvious thing to try. But we really don’t have that. Paxlovid is getting most of the heat, so I think that viral resistance is likely just a matter of time. Having huge numbers of people taking it Just In Case would only have shortened that timeline even more. If Paxlovid had shown strong efficacy as a preventative, then that could well have been worth the risk (because breaking up further viral transmission is a good result, too), but it didn’t.
And to be honest, I’m not even sure how well the drug performs in the vaccinated population. The clinical trials that have reported so far have all (to the best of my knowledge) been in unvaccinated patients, so we don’t have any hard data on this question. I’m sure that physicians are prescribing it in such cases (and I don’t blame them), but we don’t know how much benefit accrues in these cases nor how that balances against the risk of developing resistant strains of the virus. There’s also a real phenomenon of viral relapse in some people after Paxlovid treatment – that link is a very good look at what’s going on, but as you’ll see, there are a whole lot of questions and very few answers. I’m not even going to speculate until we have more data. . .
