Thread by @michaelzlin

If you haven't heard, FDA approved molnupiravir for COVID19. I've been concerned it could create highly mutated SARSCoV2 and make new enhanced viruses more likely. Today a new study supports the idea that letting viruses sample multiple mutations is risky.

https://twitter.com/michaelzlin/status/1475518939413184518

It's already known that molnupiravir doesn't kill off all mutant viruses after 5 days, but does introduce mutations into the viral genome. That is after all its only mechanism of action. Some of the mutated viable viruses may then hop to other people.

The pro-MOV argument is that MOV drops levels of viable virus more than it increases mutations in the remaining viruses, which might mean fewer later mutational templates. For several reasons, these arguments are not convincing, which means we're making an unusually dangerous bet

The reasons are (1) we don't have data that this best-case scenario is actually true. Amazingly, Merck has never measured the # of mutations in the viruses that do remain viable. FDA didn't demand it before granting EUA either.

(2) MOV increases the rate of both CU and GA transitions, while naturally CU predominates. So MOV is going to lead to a different mix of mutations (singly and in combination) than seen before. The virus can now evolve via sequence paths that were highly improbable before.

It is thus possible that, by reducing previously high probability barriers to reach new genotypes, a burst of MOV-induced mutagenesis could more than make up for any MOV-induced reduction in viral transmission, in terms of long-term evolutionary benefits for the virus.

(3) MOV will lead to more viable viruses with multiple mutations on the same genome. At least at the end of treatment mutations are found at 5x more spots than placebo, so it seems to increase mutagenesis rates from 2 bases at 5 days to 10 bases at 5 days.

And 1 person with virus containing 10 mutations may be worse than 10 people with virus containing 2 mutations, if variants arise from combinations of mutations together confering fitness (even if individually have no benefit). I made this point before:

https://twitter.com/michaelzlin/status/1475272820355592194

Today a new study shows that this kind of phenomenon may occur in Omicron. A study found that 13 mutations in Omicron spike are individually rare in SARSCoV2, so likely not beneficial alone. But together they apparently work to change how spike functions.
nytimes.com/2022/01/24/sci…

Protein engineers know this is very common; often times an individual mutation has no effect but 2 or 3 together in one spot can improve or change a protein function in some useful way. We've seen this repeatedly in our protein work.

Omicron spike shows clusters of 5, 4, and 4 mutations. If each mutation is disadvantageous on its own, then it's harder to accumulate all the mutations needed to reach the benefit of having the combination of them, because the intermediate steps are being selected against.

Thus the way to accelerate the evolution of beneficial multiply muated variants, if individual mutations are disadvantaged but combinations are not, is to induce a burst of heavier mutagenesis, so that each genome has multiple (say 10) mutations.

Of course most of these 10-site mutants are defective, but at least you've created a pool in which a beneficial combination of several mutations can be found in a single genome much more easily, compared to a pool with an average of 2 mutations per genome.

Thus if accumulation of multiple mutations, each slightly deleterious, in the right combination is a rate-limiting step for generating new variants of concern, then heavy mutagenesis in a small pool may be, on net, more dangerous than natural low mutagenesis on many templates.

This is just support for concern #3 above. Concern #2 hasn't yet been addressed experimentally, but I don't think we need to wait to find out. In any case either of these concerns, validated or theoretical, seems more than sufficient to counteract the theoretical assurances.

At least, it seems unwise to let MOV do the mutagenesis in people and hope nothing bad happens, especially when MOV is of such questionable efficacy. Better for the patients, and safer for everyone, to increase the supply of vaccines, use drugs rarely, and use better/safer drugs.

That is, vaccines are cheaper, prevent illness to begin with, and prevent hospitalization at 90% efficacy, not the 30% of molnupiravir (and note MOV is really only for unvaccinated and not previously immune, so if you're vaccinated you don't need it). They are the real answer.

I know @chasewnelson @hjelle_brian @CT_Bergstrom @sarperotto @WmHaseltine @JamesEKHildreth @RickABright were discussing this. Would appreciate your thoughts.