What are the facts about Novavax? I heard that the results from a large trial of Novavax’s COVID-19 vaccine are expected soon. But I hadn’t heard that much about this vaccine itself, so I decided to take a dive into the science in the shot. This vaccine is different from all the other coronavirus/COVID vaccines we’ve seen in that it’s what’s referred to as a “subunit vaccine” or “protein vaccine” or “virus-like particle” (VLP). Basically the vaccine contains lots of lab-grown and purified copies of the coronavirus’ Spike protein (aka S, this is the protein that juts out from the viral membrane, binds to cellular ACE2 receptors, and then shape-changes to dump the virus’ genetic info into the cells). Or at least that’s what it does when it’s on the viral membrane. In the vaccine, there’s no genetic info to dump in – instead the shot just contains little clusters of S protein anchored at their base in a central lipid “bubble” (picture jacks from marbles games). Looks like the virus but, without the genetic info, it’s more like a simplified shell of a virus or a viral mannequin. It can’t infect you and it can’t make more of itself (replicate). What it can do, however, is teach your immune system what the virus looks like so that the immune system can build up specific antibody and T-cell mediated responses to attack and neutralize the virus if it tries to attack. 

How does this differ from the other vaccines we’ve seen? Except for the inactivated viral particles, most of the vaccines all focus on using different ways of introducing your body to the Spike protein. There are a couple reasons for this focus on Spike. One is that it’s the “first thing your immune system sees” since it’s on the virus’ surface, so it gives your body the earliest alert system possible. Also, importantly, since it’s the virus’ entry into the cell, antibodies (little proteins made by the immune system) that can bind to the Spike protein can block the virus from docking on cells. We call these neutralizing antibodies. There are also other non-neutralizing antibodies, as well as a whole ‘nother branch of adaptive immunity called “cell-mediated” immunity. That pathway uses immune cells called T-cells which have receptors which can recognize almost any part of any viral protein because they bind to little chopped up viral proteins that are displayed from the surface of cells. I’m not going to try to go too far into all that because it’s outside of my field and I don’t want to give you any false information. The important take-away is that both of those arms of immunity involve a lot of trial and error to find immune cells with appropriate antibodies or T-cell receptors. And then to make lots of those to “stock up.” So they take time. And it’s a lot better for all that time-consuming stuff to happen before your body encounters the virus rather than while your body is actively trying to fight the virus. 

Therefore, vaccines give your body a heads-start by showing your immune system parts of the virus in a harmless fashion. All vaccines work this way, they just go about the introduction in different ways. We can classify these into 2 main categories which I like to think of as 

  1. “look what I made for you!” and
  2. “make it yourself!” 

The “look what I made for you!” vaccines introduce pre-made viral parts. This can be in the form of inactivated virus (such as SinoVac’s), or viral subunit vaccines, which just give premade, lab-purified, viral pieces, such as Novavax). 

The “make it yourself!” vaccines introduce the genetic instructions for making the Spike protein. These genetic instructions get into the cells (but NOT into your DNA) and direct your cells to make the protein. This class includes mRNA vaccines (including Moderna & Pfizer/BioNTech) and viral vector vaccines (which have the Spike genetic instructions delivered by a harmless virus and include Oxford/AstraZeneca, CanSino, and Sputnik V). More on these other vaccine types: http://bit.ly/covidvaccinebiochemistry 

Before I get too far into the Novavax vaccine, I want to address a couple of questions/concerns I’ve heard about vaccines that contain genetic information. The mRNA is only temporarily in your body before it gets degraded. The mRNA does not get into your DNA or anything like that, so there’s not really any biological reason why a side-effect would pop up in the future. The mRNA you’re getting is the “same” mRNA you would get with a more traditional vaccine – or with the infection, except you don’t have the actual virus, just the single mRNA so it can’t make you sick like the virus does. 

Even in the case of the viral vector vaccines, which do introduce DNA, this DNA can’t get into (integrate with) your DNA. Some viruses, like HIV, can integrate, but they can only do that because they encode and travel with a molecule called integrase which, along with other specific features of the virus, lets them get into DNA. The viruses used for the vaccines do not have those features and they don’t have that integration machinery, so they cannot integrate into your DNA. Furthermore, the virus has also been made unable to replicate, so it can’t spread throughout your body. Hope this article helps: https://bit.ly/3dYLrsJ  

So I don’t want you to get the impression I’m anti-genetic-info-containing vaccines. In fact, I am super grateful because I got my second dose of the Pfizer/BioNTech mRNA vaccine yesterday! But today I want to focus on Novavax because I hadn’t known much about it. Here goes…

You know how there are often “trendy” but vague words that drive you crazy? Well, for me, “nanoparticle” is one of those words… It can mean sooooooo many different things, but usually if you hear nanoparticle, think of some little “package” of molecules. So, for example, the Lipid NanoParticles (LNP’s) that Moderna’s and Pfizer’s shots contain are little spheres of mRNA encapsulated in a variety of lipids (fatty things) which help the particle get into cells and the mRNA get released once inside. 

Novavax also refers to their vaccine as a nanoparticle, but it’s much different. It doesn’t have any genetic info, just proteins anchored at their base in a sort of bubble of detergent called a micelle. I’ve heard these referred to as “rosettes,” kinda like heads of broccoli. Sometimes, proteins are artificially tied together into nanoparticles using some sort of adapter (for example, each protein has a tag on the base which binds to a central bead or core protein). This is NOT the case with the Novavax vaccine. Instead, these particles form naturally as a consequence of the biochemistry. So let’s look a bit at that…

We often think of Spike as “jutting out” from the membrane, but it also goes *through* the membrane. And that membrane is made of lipids, so it’s all greasy and “hydrophobic.” If we call something “hydrophobic,” that means that water doesn’t really want to hang out with it, so it gets excluded from watery networks and instead globs together with other hydrophobic molecules to minimize their overall water-interacting surface area. The liquid inside of a virus and the inside of a cell, as well as the liquid in your body outside of the cell or virus (extracellular fluid) is all water-based (aqueous). So parts of proteins that are not membrane-bound tend to be hydrophilic (water-loved) or “hidden” in the center of proteins where they don’t have to interact with water. But, the “outer world” seen by membrane-spanning parts of proteins is a hydrophobic lipid sea. So these membrane-spanning parts are hydrophobic. So, for example, if you have a full-length membrane-spanning protein like Spike, you have a sort of situation like…

outside environment

hydrophilic ectodomain (ecto- means outside)

hydrophobic transmembrane domain

hydrophilic endodomain (endo- means inside)

inside environment

note: Spike functions as a homotrimer, meaning it’s made up of 3 copies of the same protein chain. much more on Spike here: https://bit.ly/coronavirusspike 

Now say you want to purify that protein to use it as a vaccine. You need to isolate the protein, but imagine what would happen if you removed all the membrane – you’d expose all those hydrophobic parts, water would exclude them, and they’d sort of huddle together in a useless clump called an aggregate. This wouldn’t be good. So, instead, you provide an alternative environment – that of a detergent-based micelle. 

A detergent is an artificial soap. It has a hydrophilic head and a long hydrophobic tail. A bit like a molecular lollipop. 



If you stick it in water, it will form “bubble-like” things called micelles – spheres where the tails are in the center and the heads are on the surface. 




It’s a lot like the phospholipids which make up membranes. Those too have a hydrophilic head and hydrophobic tails, but they have multiple tails 



Since they’re bulkier, they’re less flexible and can’t contort themselves easily into single-layer micelles, so they tend to form phospholipid bilayers, which are like phospholipid sandwiches where the heads are the bread and the tails are the peanut butter. 





Since detergents are so similar to phospholipids, and more flexible, they’re able to kinda slither in between phospholipids, breaking up membranes. But they still have those hydrophobic tails which are able to surround the hydrophobic membrane-spanning portions of the proteins which were embedded in that membrane. This hides them from the water and prevents them from aggregating. As a result, you get micelles with proteins embedded in them! And, when you do this with Spike, you get particles that are similar in size to a real viral particle.  

So it can trick your body into kinda thinking it’s a virus. But it needs help to do this con-artistry, so, in addition to the protein nanoparticles, the Novavax vaccine also includes an “adjuvent” – added chemical(s) which help trigger an immune response. The adjuvant they use is called “Matrix M-1” and it’s based on saponins which are a kind of plant toxins. These specific saponins come from chilean soap bark tree called Quillaja saponaria. Matrix M-1 is a mix – a proprietary one of course, so I don’t know all the details, but I know that in addition to the saponins, it also contains phospholipids and cholesterol and it forms these cool honeycomb-ish cases (which are separate from the the protein nanoparticles so don’t get confused in the micrographs – the jack-like things are the Spike nanoparticles and the honeycomb-like things are the Matrix M-1 cages.) As to how it works, Matrix M-1 is thought to cause a local inflammatory response that calls in immune cells called  “antigen presenting cells” that gobble up the nanoparticles, chop up the Spike protein, and display pieces of it from their surface. https://bit.ly/327OqJA 

Different types of responding cells can lead to different types of immune responses, and that’s out of my area of expertise so I’m not gonna try to explain in too much depth… But hopefully this can help https://nyti.ms/3a4NJFv 

And here’s my basic understanding – if those displayed pieces happen to match the receptors on immune cells called helper T cells, those helper T cells will get activated, make more of themselves, and call in other parts of the immune system to help. The concept of T cell receptors is a bit similar to the concept of antibodies, except that T cell receptors bind to short pieces of chopped-up proteins, rather than “full proteins” from the external environment like antibodies can respond to (although antibodies can also bind to protein pieces). Speaking of antibodies, if B cells happen to have antibodies on their surface that match the Spike protein, and they run into one of those nanoparticles, they can, with the help of activated helper T cells, start making more of that antibody, allowing the body to build up a stock of them to keep guard. This is all a gross oversimplification, and I apologize if I got anything wrong, but I think that this is the general idea. 

Bottom line is your immune system learns to recognize that protein so that if it comes back (in the form of the virus), your immune system will have antibodies at the ready to block it from entering cells and, if the virus does get in, it will have T-cells ready to kill the infected cells before the virus can spread in your body and/or call the immune system into high alert.  The virally-infected cells will chop up the protein and display pieces of it on their surface, similarly to what we saw with the engulfed nanoparticles. Those matching T-cells which you prepared before will now do their thing, preventing the virus from doing it’s thing!

Now, let’s get out of this immunology realm, where I’m less comfortable, to the biochemistry realm, where I’m more comfortable! I want to tell you more about the vaccine production process because it’s really cool and it involves the same insect-cell-based “expression system” I use in my research (although it does so at a much larger scale!). 

“Expression” here refers to making the protein, and the baculovirus expression vector system (BEVS) is a way to get insect cells (usually grown in solution (suspended in insect-cell-food) in flasks or, probably in their case, giant vats). It involves sticking the genetic instructions for a protein of interest into a circular piece of DNA called a bacmid. 

tech note: Since we’re “recombining” DNA, we call this recombinant DNA and the protein that gets made will be called recombinant protein. 

Back to bacmids – you might be more familiar with the term plasmid (and if not, no worries!). They’re similar – they’re both circular extrachromosomal pieces of DNA that bacteria can host for you. The difference with bacmids is that you can stick them into bacteria OR insect cells. This lets you use bacteria to make lots of copies of it, and then stick those copies into insect cells. The insect cells will make protein from them – and – even cooler, they’ll make an insect-infecting virus called a baculovirus which will infect other insect cells, giving them the protein-making instructions too. So you end up with a lot of protein being made. Much more on this here: http://bit.ly/bevsinsect 

When I use BEVS, I use it to express non-membrane, “soluble” proteins. And much of the work on Spike protein that has been done is using a shorter, truncated, version of Spike that contains just the ectodomain (the part that sticks out). This part is soluble and easier to purify and work with. But the Novavax vaccine uses full-length Spike, which is crucial to the nanoparticle formation, because you need those membrane-spanning parts. If you just express and purify the ectodomain, you’ll get individual trimers rather than clusters of trimers. So they use the full-length. Another added bonus to using the full-length is that it’s longer so when antigen-presenting cells chop it up, there are more more potential peptides for T-cells to recognize. 

So they use full-length Spike, but it isn’t fully “normal” (aka wild-type). Instead they introduce a few strategic mutations to make it more stable. Spike is what’s referred to as a type 1 viral fusion protein – it helps the viral particle fuse with cellular membranes (either at the surface of the cell or after being swallowed by the cell into a little membrane-bound pouch called an edosome). This fusion allows the viral genetic material to get into the cell. And it involves some cool molecular gymnastics. After binding to the ACE2 receptor through its receptor binding domains (RBDs) (located in the ectodomain), it gets snipped by a protease, which primes it to undergo a dramatic shape-shift (conformational change) that kinda pries open the cellular membrane so it can fuse with it. For a vaccine, you want Spike to be in the pre-fusion conformation, so they stabilized it in that conformation (shape) by

  • mutating one of the protease cleavage sites (the so-called “furin cleavage site” aka the “polybasic cleavage site) so it can’t get cleaved (they changed 682-RRAR-685 to QQAQ)and 
  • introducing pre-fusion-stabilizing proline mutations. Proteins are made up of building-blocks or “letters” called amino acids. Different amino acids have different properties, and different proteins have different combinations of amino acids, so different proteins different properties. Proline is one of the “weirdo” amino acids because it has this weird ring thing which folds back to contact the protein backbone, preventing flexibility at and around it. Based on prior work by Jason McClellan, who really pioneered Spike protein structural biology, they knew where they could insert change a couple amino acids to prolines to make it harder for Spike to shape-shift. So they did (K986P and V987P mutations)

In a paper introducing their nanoparticles they showed that they were able to express this pre-fusion stabilized Spike in insect cells (they used the same cells I do, Sf9 cells, which come from a fall armyworm moth called Spodoptera frugiperda. After expressing the protein, they extracted it from the insect cell membranes using a detergent called PS-80, which stands for polysorbate 80 and is the same as Tween-80 if you’re familiar with that. Then they took “pictures” of the nanoparticles using a technique called electron microscopy. More on a variation of this here: http://bit.ly/cryoemxray 

The “pictures” (more technically referred to as micrographs) showed that the protein looked normal and that it formed virus-like particles around the size of a viral particle, with up to 14 trimers per particle. In the paper they also present results of experiments showing that it bound to ACE2 receptors as would be expected. https://bit.ly/3a26I3u  

Delivering the protein in these nanoparticles as opposed to just free-floating protein has benefits including protection of the protein until it gets to the antigen-presenting cells and improved immune-system-activating, an effect seen with other nanoparticle-based vaccines https://bit.ly/3dP63U2 

They went into pre-clinical & clinical trials of the nanoparticle/Matrix-M1 mix, which has the name NVX-CoV2373, given as 2 shots spaced 3 weeks apart and stable for up to 3 months in a fridge.

In January they published a paper showing the vaccine proved effective and able to elicit both antibody & T-cell responses in animal studies (the paper was posted as a preprint in June 2020) https://go.nature.com/2QjX0Ta 

In March they announced results of a Phase 2 study showing the vaccine was safe, well-tolerated, and elicited the production of neutralizing anti-Spike antibodies: https://bit.ly/3ak3wR1 

But how well does it work at preventing disease? This is where Phase 3 studies come in. Results of a large Phage 3 studies done in the UK came out in March “found that the vaccine had an efficacy of 96.4% against mild, moderate and severe disease caused by the original coronavirus strain and 86.3% against the B.1.1.7 variant first identified in the United Kingdom.” A smaller study in South Africa found it was less effective (55.4%) against the B.1.351 variant which was first observed there. Even though it was less effective at preventing mild to moderate disease, it “offered 100% protection against severe Covid-19 resulting in hospitalization or death.” https://cnn.it/3a11mWf https://bit.ly/2Qd7nYS 

The results of another large study (~30,000 people), this one done in the US & Mexico, is expected in April, so I expect we’ll be hearing a lot more about it soon. 

As to “why it’s taken so long” – it might be late compared to some other vaccines, but it’s still really fast compared to past vaccine development. And Novavax is a smaller company (based in the US out of Gaithersburg, Maryland) that really had to ramp up their staffing & their production capacity. https://cnn.it/3a11mWf 

They’ve also had some material shortages problems delaying things. But They recently got a boost from the big big company GlaxoSmithKline (GSK) which is going to help with production in the UK. https://bit.ly/3a1zCRo 

Key papers:

Sandhya Bangaru et. al, Structural analysis of full-length SARS-CoV-2 spike protein from an advanced vaccine candidate, SCIENCE27 NOV 2020 : 1089-1094 https://science.sciencemag.org/content/370/6520/1089 

Cheryl Keech, et al., Phase 1–2 Trial of a SARS-CoV-2 Recombinant Spike Protein Nanoparticle Vaccine, December 10, 2020, N Engl J Med 2020; 383:2320-2332

DOI: 10.1056/NEJMoa2026920 https://www.nejm.org/doi/10.1056/NEJMoa2026920 

Tian, JH., Patel, N., Haupt, R. et al. SARS-CoV-2 spike glycoprotein vaccine candidate NVX-CoV2373 immunogenicity in baboons and protection in mice. Nat Commun 12, 372 (2021). https://doi.org/10.1038/s41467-020-20653-8 

In March, This Week in Virology interviewed Matt Frieman, one of the scientists involved in the vaccine production & he does a great job explaining the nanoparticle particulars  https://www.microbe.tv/twiv/twiv-729/ 

more on lipids: https://bit.ly/lipidlove 

more on the hydrophobic effect: http://bit.ly/hydrophobesarenotafraid 

more Covid-19 resources: https://bit.ly/covid19bbresources ⠀

more on topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0  ⠀

Leave a Reply

Your email address will not be published.