Scientists have been studying the evolution of SARS-Cov-2 through experiments on cell cultures. In a laboratory setting, the virus quickly evolves resistance to antibodies when challenged by them. A mainstream view among evolutionary scientists is that the coronavirus variants will develop even more vaccine resistance than what currently exists. However, the theoretical work may not necessarily translate to the real world because there are some limitations to the science.
- We should be realistic about what science can and can’t tell us. We don’t have great science that will help us confidently predict the outcome of the pandemic. The optimism about a return to normal life is risky.
- One possible outcome is that more vaccine resistance emerges; humanity will likely be able to fight this with vaccine updates and more frequent vaccination. Between vaccines and improvements in other interventions, the deadliness and transmissibility of the disease should be significantly lower going forward. However, life won’t fully return to normal.
How science works
To advance science, we perform experiments on the world. The best type of evidence that we can collect is from human test subjects. For ethical reasons, we no longer infect human test subjects with deadly diseases. We now limit human experimentation to potentially beneficial interventions like experimental vaccines and treatments.
If we don’t experiment on actual humans, we can experiment on lab animals and cell cultures. Hopefully, that research will translate to humans. Often that happens but sometimes it doesn’t. For example, hydroxychloroquine and many other drugs demonstrated effectiveness in the lab but not in the real world. Vaccines demonstrated effectiveness in lab animals as well as in humans.
To study the evolution of the SARS-CoV-2 virus, scientists have taken the virus and used it to infect cells growing in cell cultures. They can introduce antibodies to put pressure on the evolution of the coronavirus. In this new environment, the variants that are resistant to antibodies will reproduce faster than their peers. Scientists then sequence the genetics of the fittest strains to try to figure out things like how many beneficial mutations are possible.
Why some viruses develop vaccine resistance and others don’t
Historically, we know that some viruses develop resistance against vaccines while others don’t. The measles virus is one example that didn’t develop resistance against human vaccines.
One theory is that the human immune system produces several different antibodies that are great at attacking measles; for other viruses, it only produces one or two antibodies that are great at attacking the virus. The mutation sets that allow a virus to gain resistance against a specific antibody will have a small fitness penalty. This means that a vaccine-resistant measles variant would need to carry the weight of several sets of antibody-evading mutations, making it very uncompetitive against other strains. Scientists also theorize that it is highly unlikely that such a mutated strain would naturally evolve because it needs to stumble across several sets of beneficial mutations in a very short period of time. This theory is backed up by lab experiments where scientists bred antibody-resistant strains of polio in a lab. They found that the polio virus needs to develop resistance against at least 5 different antibodies to gain theoretical vaccine resistance.
What we know about SARS-CoV-2 is that it easily develops antibody resistance in the lab. In the real world, the virus has already evolved a degree of vaccine resistance. So, it is looking highly likely that SARS-CoV-2 will develop more resistance against vaccines and lower their effectiveness. Historically, vaccines have still been useful even when vaccine-resistant strains exist in the wild. The agriculture industry uses a lot of vaccines for diseases where vaccine-resistant strains are plentiful.
The current state of vaccine resistance
In South Africa, the AstraZeneca vaccine was found to be 10.4% against the South African B.1.351 variant. The scientific paper on the results stated: “A two-dose regimen of the ChAdOx1 nCoV-19 vaccine did not show protection against mild-to-moderate Covid-19 due to the B.1.351 variant.” Unfortunately, we’ve already seen vaccine resistance naturally emerge and dramatically destroy the effectiveness of one vaccine. The mRNA vaccines are more relevant to the future because they are more effective and haven’t yet seen their efficacy destroyed by a variant.
The big picture is that vaccine effectiveness is still high at the moment. Vaccines seem to greatly reduce the chance of dying or requiring hospitalization, even if they lose effectiveness at preventing infection.
- Data from the country of Qatar shows that full vaccination with the Pfizer/Biontech mRNA vaccine is estimated to be 89.5% effective at preventing infection from the B.1.1.7 UK variant and 75% effective at preventing infection from the South African B.1.351 variant. The Qatar data suggests that the Pfizer vaccine is close to 100% effective at preventing ‘severe, critical, or fatal disease’ from the variants.
- The UK PHE has published its data on vaccine effectiveness against the UK variant B.1.1.7 and the Indian variant B.1.617.2 (see page 56 in this PDF). After both doses, UK vaccines are roughly 88.4% effective at preventing infection from the UK variant and 80.8% at preventing infection from the Indian variant. The UK stopped breaking out its data based on vaccine type- this is probably because it wants to portray the AstraZeneca vaccine to be just as safe and effective as the mRNA vaccines.
- An analysis of data from Ontario Canada estimates that the mRNA vaccines are 91% effective at preventing infection and 98% effective at preventing hospitalization/death. It did not present an analysis of the AstraZeneca vaccine that was also used in Ontario.
While the issue of hospitalization has huge economic consequences for society (because every society has imposed social restrictions to protect hospital capacity), there isn’t great science that would help us predict whether or not vaccine protection will hold. A somewhat easier question to answer is whether or not the virus will continue to have pandemic potential in a mRNA-vaccinated world.
A deep dive into the science and its limitations
Evolution experiments indicate that the coronavirus will evolve antibody resistance. If the theory is correct, that antibody resistance will allow the coronavirus to regain most of the transmissibility that it will lose in a mRNA-vaccinated world. Let’s take a look at the theory and the scientific evidence supporting it.
One phenomenon observed in evolution studies is convergent evolution. Multiple lineages of a species will stumble across the same mutation over and over, usually because that mutation provides some type of fitness advantage. When the same mutation shows up over and over again, it is usually because it provides some benefit.
Scientists can try to confirm this fitness advantage by measuring the properties of the protein encoded by that particular mutated gene. In the case of the SARS-CoV-2 spike protein, scientists can measure how well antibodies bind to that protein (and therefore measure antibody resistance). They can also measure how quickly a variant virus can infect cells in a cell culture. What scientists have found is that there are many mutations that provide resistance against specific antibodies without a big impact on the virus’ fitness in a cell culture. Paul Bieniasz and Jesse Bloom have presentations on their research, which are available on Youtube.
Scientists have also studied viral evolution in human beings with weak immune systems, who turn out to be excellent breeding grounds for mutants. They produce enough antibodies to put evolutionary pressure on the virus but not enough antibodies to clear out the infection, giving the virus inside of them plenty of time to evolve resistance.
There are three particular mutations that have become prominent with coronavirus variants: E484K, N501Y, K417N/K417T. These mutations show up frequently in evolution experiments, show up in people with weak immune systems (immunocompromised patients), demonstrate resistance against antibodies in the lab, and are found in some of the most concerning variants circulating in the real world. Both the Brazilian P.1 and South African B.1.135 variants have mutations at E484, N501, and K417.
However, many of the most prominent variants do not have mutations at all three of the locations E484, N501 or K417. The UK variant B.1.1.7 is currently the dominant strain and accounts for roughly three quarters of all infections worldwide. It has a mutation at N501Y but not the other two locations.
The Indian variant B.1.617.2 looks like it will soon become the world’s dominant variant. It does not have mutations at any of those three positions.
The cousin variant B.1.617.1 has a mutation at E484Q. That strain failed to break past 1% of worldwide infections, which you can track over at Outbreak.Info.
There are also versions of B.1.617.2 with a mutation at E484. Those sub-variants also have not grown as fast as the main lineage.
What this data shows is that lab work so far has demonstrated mild predictive power. It is likely that antibody resistance has only been a small factor in determining the fitness of a virus in the real world. That could change when vaccination rates go up. However, there isn’t convincing data that vaccination rates are having that big of an effect on viral evolution.
In the graphic below (from Outbreak.Info), the coloured lines show the prevalence share of different strains in a country (USA) with 40% of its population fully vaccinated (mostly with mRNA vaccines). In a country that is leading the world with mRNA vaccination, you would expect that vaccine-resistant variants would rise the hardest. But the data doesn’t fully support (or reject) that narrative.
The green line shows all variants with the E484K mutation. The yellow and turquoise lines show the Brazilian P.1 and South African B.1.135 variants, which are generally considered to have more vaccine resistance than the other variants shown on the chart below. The blue line shows the Indian B.1.617.2 variant, which is rapidly growing in a disturbing fashion. It seems that the Indian variant is growing faster than variants with more vaccine resistance. (The most recent data is very noisy so don’t put too much faith in it.)
Overall, it seems like variant with the least vaccine resistance (the UK B.1.1.7 variant) is losing share while the variant with medium vaccine resistance (the Indian B.1.617.2 variant) is gaining share the fastest in the US. Vaccine resistance only seems to be a small part of the puzzle at the moment. Something about the Indian B.1.617.2 variant seems to matter a lot more than vaccine resistance.
Scientists do not have a good understanding as to why the Indian variant spreads so much faster than the UK variant. While there are only 18 mutations compared to the original Wuhan strain, scientists don’t really understand what all of the mutations do and why they provide the Indian variant with a huge fitness advantage.
Other problems with antibody theories
The immune system has mechanisms other than antibodies for fighting off viruses. Unfortunately, those mechanisms are difficult to study. We know more about those mechanisms in non-human animals than humans because many experiments are unethical when performed on humans.
A lot of scientists believe that antibody activity correlates to immunity against a disease and therefore vaccine success (they refer to antibodies as a “correlate of protection“). Based on our historical experience with vaccines, we know that antibody activity correlates to immune protection most of the time but not all of the time.
There are also some scientists who will claim that antibody activity has strong predictive power because they have a social/political/economic agenda that would benefit from having certain viewpoints (e.g. they should receive funding, vaccines will save the world, scientists should be viewed as omniscient authorities, etc.).
Coronavirus predictions are risky
I wouldn’t want to predict whether or not life will return to normal. Science is not going to do a good job predicting that because there’s a lot that we don’t know about evolution, viruses, and vaccines. It would also require an accurate prediction of what humans will do. I’ve certainly failed on that front- while I was right about people underestimating the coronavirus, it turns out that the pandemic made stocks go up rather than down. This was especially true for many coronavirus recovery plays like restaurants, movie theatres, etc.
Once again, I think that people are far too optimistic about a return to normal. The politicization of the coronavirus has gotten out of hand and is making it difficult to get good information on the topic. One camp of people is obsessed with pushing the idea that “science”- in the form of vaccines- will save the world (and turn corona into a mild flu). Another camp of people is obsessed with pushing the idea that the coronavirus has always been a mild flu. Because politics-driven narratives tend to dominate news and social media, many people are being led into developing misconceptions about what’s going on in the world.
I do realize that I’m saying that life probably won’t fully return to normal. And that will be unpleasant. But to keep things in perspective, the coronavirus is nowhere as bad as smallpox or other terrible diseases that humanity has lived with. The coronavirus situation will likely also improve because vaccines and other interventions will lead to lower mortality and fewer social restrictions.
I am very bullish on vaccine profits. Moderna’s forward P/E is less than 10 and the company expects to ship 800-1000M doses this year. There is a very good chance that the world moves to annual or biannual vaccination because we know from historical experience that frequent boosting makes vaccines more effective. Biannual vaccination would likely lead to Moderna growing in size to 3B doses / year (e.g. 1.5 B people getting two shots per year). The company has excellent growth prospects and should trade at a growth stock multiple (e.g. 20-30X instead of 10X forward earnings). Analysts do not currently anticipate Moderna turning into a growth stock as their estimates do not anticipate revenue growth.
*Disclosure: I own some shares of MRNA. No position in BNTX (Biontech) or PFE (Pfizer). I received my first shot of the Pfizer.
Presentations by researchers
Robert Seder – Unlike his peers, Seder downplays concerns over variants and the evolution of the coronavirus. He argues in his presentation that antibody activity suggests that the current Moderna vaccine (mRNA-1273) will be highly effective against the South African B.1.135 variant. At the 9:45 mark, he provides his opinion that vaccines will still work against variants “but may be reduced”. Moderna is currently working on a newer vaccine specifically designed against the South African variant; Penny Moore’s second presentation linked below suggests that it could provide broad immunity against both legacy variants, the world’s current dominant variant (the UK variant), and the South African variant.
Jesse Bloom – Bloom gives an excellent starting point on understanding the lab work that is being done to understand the evolution of SARS-CoV-2. In the presentation, he talks about how well/poorly lab results map to real world variants.
Paul Bieniasz – His presentation here goes over his lab’s work using a genetically-modified VSV virus to understand the SARS-CoV-2 spike protein and how it evolves to resist antibodies.
Penny Moore – At the beginning of the presentation “Will SARS Cov-2 Become Resistant to Current Vaccines – Implications”, she provides her opinion that SARS-CoV-2 will become resistant to the current vaccines in light of AstraZeneca results that were released in February. She has another presentation on SARS-CoV-2 infection and vaccination in South Africa. At the 40:44 mark, she talks about how it’s risky to assume vaccine performance based on antibody neutralization titers.
Ralph Baric – Baric is a highly distinguished professor who provides a highly technical presentation on the evolution of the coronavirus (and other topics). At the 38:35 mark, he talks about how he’s “100%” confident that he can make an escape mutant (a virus variant that escapes a specific antibody) if given enough time and energy. At the 50:50 mark, he gives his opinion that SARS-CoV-2 will not disappear.
A scientific paper by Eguia and colleagues (from Jesse Bloom’s lab) talks about evolution in a human coronavirus that causes the common cold. They show that there is significant mutation in the antigenic spike that affects the immunity for many but not all individuals. “If these results extrapolate to other coronaviruses, then it may be advisable to periodically update SARS-CoV-2 vaccines”
Andrew Read studies the evolution of pathogens and provides a presentation on how pathogens evolve resistance against vaccines. One of his more interesting ideas is that vaccines can cause a pathogen to evolve towards more virulence if a deadlier pathogen has greater evolutionary fitness in a vaccinated world (e.g. Marek’s disease).
- Read also has a paper on “Why the evolution of vaccine resistance is less of a concern than the evolution of drug resistance”. The paper has many examples of pathogens that did and did not evolve resistance to vaccines or antibiotics.
- Read has a paper on monitoring COVID-19 for vaccine resistance. Now that vaccine resistance has appeared, his paper is less relevant.
Nature has an article titled: “Could new COVID variants undermine vaccines? Labs scramble to find out”. Here’s an interesting segment of the article which goes over the opinions of some scientists:
But Bloom and other scientists are hopeful that the mutations in the variants won’t substantially weaken the performance of vaccines. The shots tend to elicit whopping levels of neutralizing antibodies, so a small drop in their potency against the variants might not matter. Other arms of the immune response — T-cells, for example — that are triggered by vaccines might not be affected. “If I had to bet right now, I would say the vaccines are going to remain effective for the things that really count — keeping people from getting deathly ill,” says Luban.