So now we have the Omicron variant to think about. I’m just as glad that I wasn’t around to blog on Monday and Tuesday of this week while the news started hitting, because the main useful thing to say was that there wasn’t much useful to say yet. That’s still largely the case, but we can at least try to get our bearings and get ready to interpret the data that will be coming our way over the next couple of weeks. So here’s my “Intro to Omicron”, for what it’s worth.
What’s This Variant Look Like?
It’s immediately obvious that it has plenty of mutations in it compared to other variants, even Delta (and more on that below). But (as mentioned in yesterday’s post) these mutations are far from evenly distributed. It has a bit of a “greatest hits collection” character to it, because many advantageous mutations keep on being advantageous. There are only two in the region coding for the N protein (the nucleocapsid), although these have been seen before and may be associated with higher viral loads. There’s a three-residue deletion in the ORF1a region (sometimes annotated as NSP6), which has also been seen before, and might have something to do with immune evasion (although that’s not clear). But there are a great many changes in the Spike protein – in fact, a really surprising number to be showing up all at once. For instance, there’s a three-amino-acid sequence that’s been inserted at position 214, and I don’t think that’s ever been seen before. There are several changes around the furin cleavage site (S1-S2) and a whole list of them around the receptor-binding domain (RBD) at the tip of the Spike itself.
Some of these (especially a deletion at 69 and 70) have provided a proxy that helped detect this variant early. A commonly used PCR assay for testing patient samples (TaqPath) targets three different regions of the coronavirus sequence, and one of those is in this part of the Spike. That 69-70 deletion, though, causes that leg of the assay to fail – the other two sequences amplify just fine, but the S-gene part is enough of a mismatch that it doesn’t. So this “S-gene dropout” behavior has been used to monitor variants that get far enough away from the designed sequence – you don’t know what variants they are at that point, but you know that you’re dealing with something that has an odd S protein sequence. You’ll want to go in and do real sequencing after you start seeing this behavior, because several variants have this deletion or others that might do the same thing, but it’s a good early warning. S-gene dropout was in fact a major sign from the South African labs late last month that something new was up.
Where Did It Come From?
Good question! If you look at a tree diagram, you can see that Omicron really jumps off by itself. It is not a direct descendant of the Delta strains; it’s coming in from a bit of a new direction. We don’t see a lot of intermediate sequences, either, steps along the way to Omicron, even though these had to exist at some point. We just didn’t see them to sequence them, which is of course a clue. An immediate hypothesis was that Omicron may have developed in a single immunocompromised human patient over a period of weeks or months. Recall how you go about developing resistance to a new drug in infectious organisms in the lab – you expose the pathogen to sublethal concentrations of the drug and gradually increase it over time. That way, you don’t just immediately kill everything off – you give the virus or bacterium a chance to overcome a lesser challenge before turning up the pressure a bit more, letting mutations accumulate and try their worth over and over as the challenge slowly increases.
That’s just what’s going on inside a coronavirus patient if they can’t mount a full immune response. The virus and the immune system engage in a prolonged battle where neither one can land a decisive blow, and things just keep on evolving. This is why they tell you to take your full course of antibiotics when you have a bacterial infection, and why effective vaccines actually suppress variant formation: if you kill off the pathogens as quickly as possible, they don’t have time to explore their mutational landscape. You need to hit them hard and fast and keep them off-balance. Because if you take it slow, you will give a viral or bacterial infection time to experiment, and you will regret it.
So that’s one possibility, but some virologists thing that Omicron has perhaps too many changes for even that process to have been operating. Another hypothesis (see Helen Branswell’s article today) is that we humans may have first spilled some of our pandemic into an animal host, and then later spilled back over into humans again. That human-to-animal part happens, just as surely as viral infections move from animals to humans – for example, after the pandemic got going, we started finding SARS-Cov-2 in white-tailed deer, and older samples don’t show it. Late last year there was an outbreak in European mink farms from human spillover as well, and many other species have been infected as well. Perhaps Omicron’s precursor (something off the 20B clade) spent the last few months evolving in another species entirely before jumping back into humans. These are both plausible, and neither can be ruled out yet.
Is It More Infectious? More Likely to Escape Our Immune Protections? Or What?
Here’s what we’re going to be trying to figure out in the next days and weeks. Right now it’s just not clear, despite what you might see in some headlines. Omicron cases are popping up in a number of countries, and there will definitely be more. The UK is starting to see a rise in S-gene dropout test results, suggesting that Omicron could be on the way up there. I haven’t seen similar data for the US, partly because our testing and surveillance during the pandemic has too often been a very unfunny joke. But the first US case showed up in California the other day, and you can expect to see “First Omicron Case in [Insert State Name]” headlines over the next couple of weeks, for sure. Israel, Germany, the Netherlands, Japan, Brazil and other countries are all starting to report small numbers of confirmed sequenced cases.
But right now, honestly, it could go either way. Several ways. The sudden apparent rise in South Africa certainly argues for higher transmissibility, but we’ll have to see what happens in other countries as well. Remember, the Lambda variant came up strongly in Peru, Chile, Argentina, and other parts of South American back in the summer, but didn’t really take off in the rest of the world. You can find plenty of stories from back in July and August warning everyone to brace for it, but that just didn’t quite happen (this is not a complaint). Instead, it was Delta that roared around the world. Omicron might kick Delta aside and become the dominant strain. Or the map could turn into a patchwork, with some areas much more affected than others (which, to be sure, has been the dominant mode throughout the whole pandemic). Or, less likely but not impossibly, Omicron could turn out to be less of a problem than it now appears.
Here’s a good thread on Twitter from Trevor Bedford trying to work with the data we have now. He makes several plausible assumptions and shows how these lead to a spectrum of transmissibility and immune escape. Right now, the data support Omicron lying along a curve in that space, with higher transmissibility implying lower propensity for immune escape, and vice versa – you can make the epidemiological data fit with several combinations of those two, but what it doesn’t fit is a Doomsday Coronavirus scenario where both of those are maxed out. If that were true, things would already be worse, believe it. Looking at the large number of Spike region mutations, Bedford believes that it’s likely that Omicron is going to show greater immune evasion at the expense of transmissibility, and in fact it’s possible that it might in the end be less transmissible than Delta is. If this story from Israel pans out, that might be what we’re seeing.
But the greater immune evasion is, of course, still not good news. Again, that does not mean that vaccine protection (or protection via prior infection) is suddenly useless, just that it may have been eroded to some degree that isn’t clear yet. Just as before, you really, really need to get vaccinated, and you really, really need to get a booster if you can do that as well. That’s a far better option than facing either Delta or Omicron with a naive immune system – and if the latter really is a bit better at evading the human immune response, it’s even more desirable to tune that up as much as possible before you get exposed at all. I get a lot of email from people who are worried about vaccine safety, and although I don’t share most of their concerns, I can at least understand them. But the people who refuse to get vaccinated because they apparently think they can fight off the virus better without it, those are the folks that I’m completely baffled by.
What Now?
We gather as much reliable data as we can, as quickly as we can. We carefully monitor the spread of this variant and we try to figure out how many people each new patient has infected. We sequence cases as much as we can to make sure that Omicron isn’t drifting into something even more concerning. Once people have recovered from it, we take blood samples from them and characterize what their immune response looks like and compare it to post-Delta cases and others. And (in experiments that are going on right now) we take blood samples from people who’ve had Delta, others who’ve had two shots of vaccine, and others who’ve had those plus a booster and we see how well their antibodies can neutralize Omicron in vitro. Those experiments will give us the earliest read on what we might expect in all these populations if this variant does really take off in the world, and we’ll be watching the real-world data to make sure that things follow in the way we expect. We don’t know nearly enough today – but we’re going to know a lot more very soon.
