By Derek Lowe
1. Novavax Clinical Data
Word came yesterday that Novavax had very good safety and efficacy in the trial of their recombinant protein vaccine. This is good news. By this point, the vaccine is much less needed here in the US, but it could be a very important part of getting many other countries vaccinated, due to its less demanding storage requirements and (relatively) straightforward production process. The company does intend to file for FDA approval, and is in the last stages of getting all of its manufacturing and quality control procedures ready for that. I hope that this opens up to worldwide usage of this one, and that the company really is ready for large-scale production.
As many readers are well aware, this is a recombinant protein vaccine, not a viral vector (like J&J or the Oxford/AZ vaccines), and not an mRNA one like Moderna or Pfizer/BioNTech. I also hope that this allays some of the worries that many people still have about those two platforms: recombinant protein vaccines have been around for longer, so this one would (presumably) be less of a concern for some potential users.
2. Circulating Spike and S1 Proteins After Vaccination
Let’s talk about one of those concerns in particular. I am still getting question after question about the Spike protein being produced by the mRNA and viral-vector vaccines. Not a day goes by without someone contacting me about this issue, and recently these messages have taken on a slightly different tone. I last wrote about this issue here, and since then there have been other publications that bear on the topic. Here’s a key one: this paper describes the detection of the Spike protein in the bloodstream of patients who received the Moderna vaccine. Now, in that earlier blog post I spent several paragraphs going on about how the Spike protein (because of its transmembrane anchor) would not be floating around in the bloodstream, and you can be sure that many of my correspondents have reminded me of that. But see below for more details!
To be sure, the new paper uses a very sensitive assay to find the protein – Simoa, which is a single-molecule detection method. I’ll explain that one briefly for those who are interested; if you’re not, you can skip ahead and just stipulate that this is pretty much the gold standard for detecting very small amounts of protein in biological samples.
2a: The Assay Technology
So, you take antibodies against your protein of interest (you do need those, just as with the older ELISA assays!) and chemically conjugate them to tiny magnetic beads. These beads have a couple of hundred thousand attachment sites on them, so you’ll be carpeting their surfaces with your antibodies. That said, you then use a large excess of the beads (maybe tenfold) when it comes time to analyze a blood or tissue sample. You’ll be using this technique to detect very small amounts of protein (up to 1000x times smaller than you can work with using ELISA) and ideally, you want the majority of the beads to not capture any sample at all, and the ones that do capture a target protein to only capture one.
A big excess of beads with a lot of capture antibodies on them means that your target protein is going to be vacuumed up out of solution really quickly, and as this post explains, that’s a real advantage in these methods. The more time proteins spend rubbing up against and rolling across surfaces, the more degraded and denatured they get, so speed is definitely a virtue here. Once you’ve done this, then (as with an ELISA) you set up a “sandwich” assay format, where your protein of interest ends up captured by one antibody, while its remained exposed surface then gets bound to another detection antibody. This detection antibody has biotin tags already attached to it, and every time you see biotin stuck onto something, you know that it’s going to be used to grab onto some other reagent that has biotin’s lifelong love, streptavidin, attached to it. Indeed, once you’ve hit the sample-incubated beads with the detection antibody, you add in a streptavidin-beta-galactosidase enzyme conjugate, so that in the end, every bead that captured a target protein now has a detection antibody stuck to that in turn, which is then stuck to a streptavidin-enzyme species. A sandwich indeed!
All these beads are then distributed into a microwell system that the company that own Simoa (Quanterix) will be happy to sell you. These microwells are ridiculously small (femtoliter-sized) and can only hold one bead at a time. You use a layer of oil to seal those into the wells, along with a supply of resorufin beta-galactopyranoside. That one is a fluorescent hand grenade of a molecule: the enzyme that you conjugated to those beads that picked up a target protein is waiting to cleave that galactose sugar off, which liberates the wildly fluorescent free resorufin. And since you’ve got a lot of the stuff sealed into those tiny micron-sized wells, you build up a really bright fluorescent signal very quickly, far more bright and concentrated than you can with an ELISA assay. This confinement-and-fluorescent-buildup is a big part of how you can do single-molecule detection with this technique. You then read off the wells that are glowing compared to the ones that aren’t, and if there are plenty of the latter you’re in good shape to analyze for the real concentration of your target.
2b: S1 and Spike Levels
This paper marks the first time anyone has detected Spike or S1 protein in the bloodstream of vaccinated patients, and that can be put down to the use of this very sensitive assay. The team looked at full-length Spike, the S1 subunit (after the furin cleavage site has done its thing), nucleocapsid (N) viral protein, and antibodies against all of these as well. The assays were baselined against plasma from Covid-19-positive plasma samples from infected patients, healthy subject samples from before the pandemic, pre-pandemic samples from people with other respiratory infections, Covid-19-negative patients, and so on.
Remember, the mRNA and adenovirus vector vaccines will only produce Spike protein, so measuring the coronavirus N protein is a good control (there should be none of that unless you’re infected by the actual coronavirus, since vaccination by these agents cannot make your cells produce it). And indeed, in 13 patients the authors could find S1 protein in 11 of them after the first dose of Moderna vaccine, and Spike in all of them, with no N protein anywhere. The S1 protein started showing up as early as the first day after vaccination, peaked at around day 5, and was undetectable by day 14. The mean peak value was about 68 picograms/mL, although from the figure you can see that this number is influenced by a few high outliers – most of the numbers are 50 or under. That 14-day disappearance is surely because by that time you have raised a strong antibody response to the S1 protein, and it’s being cleared – the antibody levels found in this paper bear that out exactly as you’d expect. As for the full-length Spike itself, it was seen in three of the 13 patients, but later – an average of 15 days after vaccination. After the second vaccination, no S1 protein could be detected at all.
The reason for the differences between the S1 and Spike levels is unclear, but the authors suggest a plausible hypothesis. The mRNA from the vaccine starts being picked up and translated into protein almost immediately, as is clear from the quick detection of S1 protein. That’s there because it’s been cleaved off the full Spike protein, but the reason that the Spike itself isn’t found (at least at the limits of detection in the assay, and it’s a really good assay) is because it’s bound to the cells where it’s produced, by the transmembrane anchor region (discussed in that earlier post I referenced above). The reason that no S1 protein is found after the second vaccination is clear – by then, a robust antibody response to it has had a chance to develop, and the protein gets rapidly cleared from the blood, just like it’s supposed to.
What about those instances where some full-length Spike protein shows up later? Well, consider what happens after the vaccination, when the cells that have taken up the mRNA make and present Spike protein on their surfaces. That (and the free-swimming S1 subunits that are cleaved off of these) set off a response in the adaptive immune system, which after all is the whole point of administering a vaccine. And that response (especially the now-activated T cells) leads to the infected cells being killed off (as is observed with other types of cell-infecting vaccines). At that point some intact Spike protein can get released, now that the cell membrane is being destroyed. The 15-day lag fits that timeline very well.
Mention of that process leads me to address a concern that I’ve also been asked about: if the cells affected by the vaccine are killed off by your body’s own immune system, isn’t that bad? How much muscle tissue, etc., are you losing? I have not seen an estimate – that number would be very hard to pin down! But a decent-sized muscle like the deltoid has on the order of a hundred thousand muscle fibers (each an individual large muscle cell) in it – there are around 250,000 in the bicep, apparently. And each of these is surely getting infected by a great many mRNA particles (they are large, as mentioned, with a lot of surface area). But how many of these are attacked in the end by T cells, I don’t know. Keep in mind, though, that infection by the real coronavirus is surely far more destructive, and you will be losing plenty cells of all description if that happens to you. The vaccine does its thing without replicating, whereas a real infection can flood your body with infectious viral particles.
3. Circulating Spike and S1 Proteins During Coronavirus Infection
That leads us to another interesting question: what are the levels of that S1 protein seen after vaccination as compared to a real infection? This same team of authors has used the same assay technology to study infected patients, which makes comparison very clear-cut. The levels of circulating S1 protein are in fact quite similar, and the profile of decreasing S1 over the course of infection (as the antibody response kicks in) is seen in the same manner.
But let’s get down to what my correspondents are worried about: those animal studies where Spike protein was produced by a pseudovirus, or the S1 subunit was administered directly. Both of these caused pathology in the animals all by itself, without coronavirus itself being present. Why, I get asked constantly, would I allow my own body’s cells to make this stuff, if it’s the cause of all the trouble?
The thing is, it doesn’t appear that it is the main cause. If you go to that first link in this section, the one where they quantified these proteins in sick patients, it looks like S1 levels can be useful in judging severity, but only up to a point. Patients with the highest levels of S1 on admission to the hospital were more likely to end up in to the ICU more quickly, but that could also correlate with total viral load, with lack of a robust immune response, and other factors. And there was no statistically significant difference in death rate between patients with low or high S1 levels. As the paper itself notes, their patients were all admitted to the hospital, so they were already presenting with severe disease – but some of them had single-digit pg/mL concentrations, while others were at 50, 60, or higher. So you can have hospitalization-worthy severe coronavirus but still have quite low S1 concentrations.
Now, you may look at those numbers and say “Hold it – those sound pretty much like the S1 levels you get in the first few days after you’re vaccinated“. They are indeed. But that brings us to another line of argument. I have no patience with the anti-vaccine commentators who are talking about the vaccines mowing people down like wheat, because that is obviously not happening. If hitting an S1 plasma level of 68 picograms/mL was sufficient to destroy your lungs, we would have destroyed tens, hundreds of millions by now. We have not. Nothing of the sort was seen in any of the animal studies, in any of the initial human studies, nor in the crucial human efficacy trials. And nothing of the sort has been seen since these vaccines began being used more broadly in the population.
What about those animal studies, though? I think that this is a question of secondary importance, since it is obvious that the animal and human results diverge. One thing to remember is that both of the animal studies treated the animals (with either the pseudovirus or the S1 protein) by aspirating these directly into their tracheas, which is obviously different than the situation after vaccination. But note that the authors of the Circulation Research paper (the pseudovirus study) conclude by saying that “. . .vaccination-generated antibody and/or exogenous antibody against S protein not only protects the host from SARS-CoV-2 infectivity but also inhibits S protein-imposed endothelial injury“. They are not sounding the alarm about vaccination; they are recommending it.
4. Back to the Novavax Results
There is another line of argument to make. Remember, the Novavax vaccine does not cause cells to produce the Spike protein. It is the Spike protein, injected directly into a person’s body, along with an adjuvant to make the immune reaction that much more vigorous. It does not get cleaved to make S1 protein, because that cleavage site has been mutated to keep that from happening, but it does bind to human ACE2 receptors just like the wild-type protein. If such an injection were causing harm to patients, that would have set off a strong safety signal in the human trials, but no such problems have been seen. Not in the animal studies, not in first dose-finding studies in humans, not in the first efficacy trial, and not in the one whose results were just announced.
It seems clear from all these human trials and the clinical experience to date that the circulating levels of the S1 protein (or the Spike itself) that are sufficient to induce a protective immune response are not in themselves toxic. The animal studies demonstrate that the Spike or S1 can indeed have bad effects on living cells and tissues all by themselves, but the conditions under which this was demonstrated are not those that obtain after vaccination.
And this latest paper showing circulating S1 protein after vaccination? Coupled with the previous paper from the same group, it in fact provides strong evidence that such blood levels are not by themselves the cause of coronavirus symptoms and tissue damage. No, if you want to try for severe, permanent damage, you will need to get infected by real SARS-CoV-2 itself and take your chances. Try your luck, if you like, with the short-term symptoms and with “long Covid” symptoms as well. See if you stay out of the ICU, or if you retain your sense of smell – try them all. If you would rather not spin that wheel, and you shouldn’t, then my strong, heartfelt advice is to get vaccinated. Because then you will be protected.
https://blogs.sciencemag.org/pipeline/archives/2021/06/15/the-novavax-vaccine-data-and-spike-proteins-in-general
