Welcome to Science with Shrike! Happy new year! In today’s post, we will explore a roadmap for SARS-CoV2 mutations and spread. Stripping aside the FUD from all sides, the SARS-CoV2 pandemic has played out as expected from the underlying biology. Let’s see how that has played out, and what that means for 2022 and beyond. We’ll talk about some general principles, using flu as an example, and then see how it may apply to coronaviruses.
Antigenic Shift
Before talking about coronaviruses, we’re going to start with a less controversial respiratory virus: influenza. There are 3 different influenza viruses, flu A, B and C. Flu C is mild compared to the others, so it doesn’t get much attention. Flu B is common in kids, has no animal reservoir, and can be lethal. Flu A is usually considered seasonal flu and is the one everyone worries about most. Its original hosts are waterfowl, but it infects a lot of other animals. Flu A is further subdivided by its two spike proteins—hemagluttinin and neuraminidase (HA and NA, or more simply, H and N). HA helps flu bind and get into cells, while NA helps flu detach and leave cells. While there are many different types of H and N (each type gets a different number), the only ones that spread regularly in humans are H1-H3 and N1/N2. For example, the 1918 flu epidemic was H1N1, while the common flu in the early 2000s was H3N2.
So why are these numbers relevant? If your immune system targets H1, it may not be very helpful against H3. The change between numbers (H1 to H2) is an example of antigenic shift. Recall antigens are the targets that your immune system recognizes. Antigenic shift is when a pathogen changes up that target so that your immune system can no longer recognize it. Flu changing between strains is one example of this.
The strongest example, however, comes from the causative agent of African sleeping sickness (Trypanosoma brucei). T. brucei takes ~3 years to kill you. It carries ~900 different versions of its primary antigenic protein (Variable Surface Glycoprotein, VSG), which it can change as needed. Since it takes your immune system about a week to develop adaptive immunity to each version of VSG, T. brucei barely exhausts ¼ of its repertoire evading the immune system before it kills you.
In contrast to T. brucei, antigenic shift in the flu is rarer. Flu doesn’t carry multiple copies of H or N, so multiple viruses need to be in the same host and happen to recombine. This can drive new pandemic strains. One example of this is the change from H3N2 as the predominant flu strain to H1N1 as the predominant flu strain after 2009. Antigenic shift can lead to pandemics because the population does not have widespread immunity to the virus.
Antigenic drift
Another phenomenon related to antigenic shift is called ‘antigenic drift’. They are similar, except for scale. In antigenic Drift, the changes are less pronounced than occurring from a shift. Whereas antigenic shift is often getting a mostly/entirely new protein sequence (H1 to H3), antigenic drift occurs through the accumulation of mutations in the viral proteins (H1 stays H1). These mutations arise because viral polymerases often make mistakes when creating new viruses. Most of these mistakes lead to defective viruses, but viruses play the numbers game. For example, if it takes 1 virion to infect a cell, and the infected cell produces 1000 virions, destroying 99.9% of the new virions leaves that 1 virion left to infect a new cell. Thus, the virus can afford to screw up new virions 1% of the time no problem.
These mutations help the virus survive in two key ways. First, if the mutation helps the virus replicate faster, or infect better, it will confer a survival advantage to the mutant virus. This virus will outcompete all of the other viruses over time and the mutation will become dominant in the population. Second, the mutation can help the virus escape immune control. Once you develop protective immunity, subsequent viral infections will be more limited and/or the virus will fail to establish a productive infection. Mutations that help the virus evade this immunity confer a survival advantage because the mutant virus can better infect people who have immunity. Acquisition of these mutations helps promote immune escape, and enables the virus to survive. This is also why we lack a universal flu vaccine… the flu virus keeps escaping. Notably, seasonal coronaviruses (the ones that give you a cold) are also very good at immune escape, which is one reason we never cured the common cold.
Both kinds of mutations can occur at the same time. However. If the virus has not been in humans for a long time, it still has room to optimize infection. If it has been in humans for a long time, it’s already optimized for infecting humans. At this point it becomes more unlikely that it will be able to improve on the current optimization. In fact, many mutations will make it less able to infect humans. These mutations survive when that trade-off in survival and infectivity (generally lumped as viral “fitness”) is balanced by immune evasion. Thus, on a long time line, viruses have to sacrifice fitness for immune escape. On a short time line, they can improve both fitness and immune evasion. Let’s see how this played out with influenza in the last 100 years or so.
Pandemic flu strains
In the 20th century, there have been several flu pandemics. Starting with the 1918 flu pandemic, H1N1 circulated for ~40 years before being replaced by H2N2 in 1957. H2N2 circulated for 11 years, when it was replaced by H3N2. In 1977, the 1950 H1N1 strain leaked from a lab and circulated along with (but with a lower prevalence than) H3N2 until 2009. In 2009, the swine flu (H1N1) replaced the pre-existing H1N1, and became the predominant flu strain. So how did these viruses supplant each other?
This comes down to immune evasion. Over the course of ~40 years, H1N1 had to progressively trade-off fitness in order to survive in the immune population. When H2N2 arrived on the scene in 1957, it had yet to make any fitness trade-offs. Since both viruses target the same receptors, etc, H2N2 outcompeted H1N1. Then H2N2 itself was outcompeted by H3N2. Notably the same N provided some immunity to H3N2, so H3N2 was less severe than if it had been a different strain. When the 1950 H1N1 was released, it had already traded fitness for survival, so it wasn’t way more competitive than H3N2. Hence it couldn’t displace it. However, there was low enough immunity to it that it could establish in the population (everyone <20 had no immunity to H1N1). The swine flu was then able to displace both strains because it was more fit than either previous strain. It took out the old H1N1 faster than H3N2. Shared immunity between swine flu and the old H1N1 also prevented severe disease in some individuals.
SARS-CoV2
With this framework for flu, let’s now take a look at the SARS-CoV2 pandemic. After it escaped from a lab in Wuhan, China (link), the alpha strain spread across the globe over a period of 4-6 months. As a coronavirus, it is prone to mutation, and various mutations were characterized. The mutations everyone worries most about are in the Spike protein, which is the target for the vaccines. However, other mutations that enhance fitness are also of concern. Notably, these other mutations are usually identified empirically (ie from observing that they’re becoming the dominant viral type in an area). At the international level, variants with enhanced infectious potential may be designated “Variant of concern” by the WHO.
The biggest questions with the variants of concern have been: do the vaccines/natural immunity still work against them, can we still detect them, and how has mortality/infectiousness changed? In practice, delta was the first variant of concern that seemed to escape immunity, with omicron escaping to an even larger degree. However, this escape was not ‘all-or-none’—vaccines still showed efficacy against delta, though the precise extent of efficacy is up for debate. In the case of omicron, it seems to have escaped immune control even better. Both fully vaccinated and previously infected individuals have caught omicron. Additionally, it seems that omicron spreads faster and has lower overall mortality than the earlier strains.
What happened?
It seems to Shrike that omicron has adapted better to humans than the previous strains. The gain in infectivity indicates that SARS-CoV2 is still adapting to humans. Since humans have organs that tissue culture cells lack, it does not help with the viral origins issue beyond suggesting that SARS-CoV2 has not been in humans for very long. The reduction in mortality is also a sign that omicron is adapting to humans. While it is unlikely to give up migrating to the lower respiratory tract entirely, it would be expected to focus more on spread, and remaining an upper respiratory pathogen. Most of the mortality comes from ‘dead-end’ infections, which are not advantageous to the virus.
What we’ve seen so far fits with standard antigenic drift. In the face of the selective pressure to adapt to humans, it is improving its virulence. As an added benefit, it is also escaping immune control. One outstanding question is the role the vaccines played in all of this. Sorting out a definitive effect will be challenging. However, widespread immunity (either via natural immunity or from vaccination) enhances the selective pressure on the virus. Shrike believes that the vaccine accelerated the selection process because it drastically increased the immune population in a short period of time. Insofar as this helps the virus move to a less lethal variant, this is a good development. It also did this while cutting the death rate for the at-risk population 10x.
Where does Omicron go from here?
Now that we have a clear escape mutant, it indicates a few things. First, that while the vaccine was effective at controlling earlier strains, the coronavirus can escape vaccine control. This is not surprising, since it also escapes natural immunity. With the exception of a pig coronavirus, this is how other coronaviruses have gone… they escape vaccines and immunity. If we’re very lucky, SARS-CoV2 will go the route of the pig coronavirus and actually develop into an attenuated vaccine strain…. however, omicron is not it. This means SARS-CoV2 will be endemic, like the flu and the seasonal coronaviruses. It’s not yet clear if a seasonality will establish for SARS-CoV2 or not. Unless a move to an annual vaccine is made, the current vaccines are not expected to be very helpful. However, please note that data on omicron mortality is not well published yet, and does not compare naïve individuals vs natural immunity/ vaccinated individuals—vaccines could still reduce severe disease. The cost/benefit of an annual vaccine would require better information on the incidence of severe disease with omicron and the rate of viral escape. Flu is estimated to kill ~60,000 Americans/year and there is an annual push for a flu vaccine. If SARS-CoV2 is estimated to kill that many or more, it would be consistent with pre-COVID approaches to implement an annual vaccine.
Looking ahead to the next escape mutants (pi? rho?), here are Shrike’s predictions:
1) Change in transmissability. Shrike expects the next major variant or two to become more transmissable, but the variants after that are expected to finally start trading fitness for survival, and become less transmissable. Barring another widespread vaccination campaign, when the virus starts trading fitness for survival, it will look like two strains co-existing, one in the “naïve” population and the other in the previously infected and naïve population. The ratio will change with the proportion of people with immunity. Right now, omicron is displacing the other viral strains. If the next variant is less transmissable, then omicron was fully human adapted.
2) Continued reduction in mortality from the virus. Shrike expects that the virus will cause fewer deaths as it further adapts to its host. Severe disease should be monitored at this point instead of cases. While severe disease will never go to zero, Shrike expects it to be a comparable rate to flu or hopefully seasonal coronaviruses over the next escape mutant or two.
There is one huge caveat to this prediction, which is immunopathology. There is an ongoing debate about what causes COVID19 disease severity. Is it caused by ‘too much immune activation’ or is it caused by ‘not enough immune activation’? It may be that as the virus escapes the immune system, it still triggers immune activation, leading to damage to the host. Hence, severe disease should be monitored, and especially broken down by immune and naïve populations along with age/comorbidities.
3) New escape mutants every 12 months. Based on the alpha (Dec 2019)-> delta (~Nov 2020) -> omicron (Oct-Nov 2021), it looks like it takes ~1 year for an escape mutant to emerge, and then a few months for it to spread everywhere (delta was an issue for India Feb 2021, and then worldwide dominant strain in June 2021). This means that there will likely be a new escape mutant coming out Oct-Nov, which is right around the US 2022 midterm elections. When the time it takes for a new escape mutant to become the new dominant strain increases, that will be a sign that the virus is starting to lose fitness.
How these ideas and predictions could inform state policy is a completely different topic. That also depends on what policy goals are. For example, minimizing death, hospital overload, infections, small businesses, and freedoms all require different approaches for optimal efficacy.
Great write up. Whats your opinion on:
1. The role of BCG vaccines at birth against Cov2.
2. HIV pts with high VL's contributing to mutations of CoV2.
Regarding the immunopathology, what I've seen with my own eyes(Not research papers or stats) is its somewhat skewed towards people with hyper immune response/obeseity.
Immunocompromised indivduals i.e HIV, patients with autoimmune diseases(SLE, Sarcoidosis etc) on high dose corticosteroid tend to fair Covid just as a normal healthy adult. Most people with S.Covid during the first wave (2020) and Delta wave were all obese and ending in ICU with CK storms.