March 2022

Many of us breathed a sigh of relief when COVID-19 vaccines rolled out.  For those with the good luck of living in places with easy access, getting a shot or two could mean a fast track back to normal life — no more masks or social distancing.  As vaccination rates went up, new cases fell, and social restrictions followed the same trend. It made sense to relax the rules we’d been living with once new virus cases slowed. However, recent research reveals potential evolutionary pitfalls in this approach.

Where's the evolution?

During the pandemic, we’ve used two approaches to prevent new infections. First, changes in our behavior can physically prevent a person from encountering the virus.  These changes include masking, physical distancing, limits on how many people can be inside buildings, and the like.  Vaccination, on the other hand, prepares a person’s body to fight the virus in case they do encounter it. Vaccinated people are less likely to get very sick with COVID-19 and are less likely to pass the virus on to someone else.

A team of researchers from Israel, Germany, and the U.S. wondered how these two different approaches (distancing and vaccination) interact with one another and whether they might lead to the evolution of a strain of SARS-CoV-2 that can infect fully vaccinated people. To explore this question, they built a mathematical model including factors like how many people are vaccinated each day, how often the virus spreads from one person to another, and how likely it is that, as the virus is copied, a vaccine-resistant mutant is produced. Then they tested how tinkering with vaccination rates and distancing affected the ability of a vaccine-resistant viral strain to arise and spread.  What they found highlighted a potentially dangerous problem with how vaccination campaigns and distancing measures are often put into practice.

During the pandemic, many communities were monitoring viral spread and imposing strict distancing rules when cases were on the rise but relaxing them once it seemed like the virus was under control (e.g., the number of new infections was going down each day). As vaccination campaigns were rolled out, more people were protected from infection, new cases naturally fell, and so did distancing: gatherings were allowed and mask rules loosened, for example. The strategy seemed sensible, but it fails to account for how these actions might shape viral evolution.

The new research shows that loosening rules on distancing too soon in a vaccination campaign can encourage the evolution of a vaccine-resistant viral strain.  And when you think about this in evolutionary terms, it makes sense. Natural selection relies on four factors: variation, inheritance, selection, and time.  Each time a virus is copied (and they are copied a lot when the virus is spreading in a community) is a chance for a mutation — that is, a variation — to occur.  So a spreading virus means a highly variable viral population, possibly including one with a mutation that makes it vaccine resistant. At the same time, the presence of many vaccinated people means that a vaccine-resistant virus has big selective advantage. It is likely to leave behind a lot of descendants because it can take advantage of a “resource” that other viral variants cannot: vaccinated people.  When these two things (high variation, strong selective advantage) happen at the same time because relaxed distancing measures allow the virus to spread while we are only partway through a vaccination campaign, natural selection is likely to produce a vaccine-resistant viral strain.

So what’s the solution?  The research showed that keeping some distancing rules in place, even after case rates drop because of vaccination, could drastically reduce the odds of a vaccine-resistant strain arising and spreading. Rolling out vaccinations more quickly and convincing more people to get the vaccine helped as well.

The team also simulated a well-known strategy for slowing the evolution of resistance: a cocktail. The idea here is to formulate a vaccine that requires more than one mutation for a virus to be resistant. This might be achieved by mixing vaccines targeting different viral proteins (hence the cocktail reference). Since mutations occur randomly (i.e., without respect to what might be advantageous in a particular situation), a mutant virus might have one mutation that lets a key protein evade detection by a primed immune system, but is much less likely to carry two mutations like this.  We’ve seen that drug cocktails slow the evolution of drug-resistant HIV and antibiotic-resistant bacteria.  Perhaps, the new research suggests, a vaccine “cocktail” could do the same for SARS-CoV-2.

Of course, models are never exactly like the real world and can leave us with plenty of questions. How would the outcome change if vaccinated people mostly socialize with other vaccinated people, or if resistance evolves in small steps instead of with a single mutation, or if vaccination doesn’t completely prevent a person from spreading the virus (as is the case with the current COVID vaccines)? Despite these uncertainties, the new research still sends a powerful message that, in this pandemic, our social policies can have unintended evolutionary consequences. Given how many times the new coronavirus has surprised us so far, such outcomes seem worth considering as we develop policies to live with this virus long term.