Breaking the Vaccine Production Bottleneck: How to make a Messenger RNA vaccine

November 9, 2020

GreenLight Biosciences is working on a Messenger RNA vaccine. In this piece, the team explains why a Messenger RNA vaccine is so vital and how to make it.

The world needs a vaccine. Covid-19 is still sweeping the globe, and killing thousands. All eyes have been on the research – how scientists are establishing which vaccine works, and whether immunity will last. But there’s another, arguably greater problem. In order for the vaccine to establish herd immunity, we will need billions – not millions – of doses. The sheer scale of the problem is hard to imagine.

GreenLight Biosciences has an ambition to match that scale: it is aiming to develop a new vaccine – a messenger RNA vaccine – and to manufacture billions of doses of it from the very start of the process, it is designing the vaccine with that intimidating target in mind. Everything, from the methods used to develop it to the materials used to create it – even down to the part of the virus it targets – has been developed to make that goal possible. 

But first, to understand how they’re doing it and what the challenges are, we need to look at the flu.

The problem with the flu vaccine
A new flu vaccine is created every year. Some years, it works fine. Other years, it does not. The problem is not the science. We understand the influenza virus and we understand our immune system.

The problem is more prosaic: Making enough doses of the flu vaccine. The problem is that each new version of the flu vaccine requires its own, specialised and lengthy manufacturing process. This means we often spend precious months making a vaccine for the wrong virus. 

The standard annual flu vaccine is made from the real flu virus, which is either weakened – “attenuated” – or killed. This sounds straightforward enough, but the flu is not one thing. Each year, new strains of flu emerge from animal populations – often birds or pigs – and mutate to infect humans. Each new flu strain requires research and development efforts to both discover and manufacture the vaccine that creates the right level of immune response.

Different types of vaccines compared with Messenger RNA vaccine
The current record for making a new vaccine entirely from scratch is the mumps vaccine, which took four years. Many others have taken decades.

Fortunately, we now have new kinds of vaccines available, with production processes significantly faster than traditional methods.

Our immune systems are clever. Exposed once to a pathogen, such as a virus, they will remember it. The next time the virus appears, our immune system’s memory allows for a quick response to fight it off.

Vaccines aim to pre-empt that process: to present our bodies with something that looks like the virus – a small part of it, or a weakened version of it – so that when the actual virus arrives, the immune response is ready to go.

While all vaccines share that basic concept, the details differ widely. There are five main kinds of vaccine: attenuated virus, killed virus, protein antigen, viral vector, and messenger RNA.

The first two, attenuated virus and killed virus, we have discussed above. They both work by exposing the immune system to the entire virus, either in a weakened form or in an entirely lifeless one. They have been extremely successful for many decades, but they are slow to produce, with several bottlenecks, notably the need for millions of hens’ eggs, and the very different manufacturing processes between vaccines.

To achieve this, scientists each year make their best guess as to which strains of flu will be prevalent, and they focus on creating vaccines for those strains. The challenge is reaching industrial scale manufacturing capacity capable of inoculating large parts of the world. 

This whole process takes time due to numerous bottlenecks.

Traditional vaccines, such as the attenuated virus form, must be grown in biological substrates. For the better part of a century now, the biological substrate of choice has usually been hens’ eggs. But the precise methods used to grow the virus differ widely from disease to disease, or even from strain to strain, as with flu.

“It’s unique every single time, especially with a novel virus,” says Dr Elisha Fielding, a senior scientist at GreenLight Biosciences. “A new strain may require a whole new process, or at least massive tweaks, every time.”

These tweaks take time. For instance, it can take up to three months to develop the reagents, chemicals which allow manufacturers to test a new vaccine and ensure it is standardised. Also, each batch must be grown in the eggs, which takes around two weeks, creating a severe bottleneck imposed by the number of eggs available.

The reason, therefore, that the flu vaccine has been so hit-and-miss is that the scientists are required to predict one year in advance which of the flu strains will be prevalent. These are the challenges  just for making a new vaccine for flu. When creating a vaccine for an entirely new pathogen, such as Covid-19, the challenge is even greater.

“You’re going to have to figure out how you attenuate, to a level that works,” says Fielding. “And then you have to take that and make it at large scale. If nothing else, you just need an insane amount of chicken eggs.”

Protein antigen vaccines work rather differently. Instead of growing the entire virus, they find a small section, a protein molecule on the outside of the virus. In the case of Covid-19, they use the “spike” protein, the distinctive lumps around the virus’s round body. This protein itself is introduced into the patient, priming the immune system to recognise it. Those, says Fielding, could be produced more quickly than the traditional whole-virus models. But even then, she says “no two proteins are alike. You have to worry about how the protein is decorated, the sugars coating it; if it even expresses recombinantly in cells; whether it is soluble; there are a lot of things to consider which ultimately take time to work through.”

Understanding how proteins behave is notoriously complicated, and even if we manage to work it out for a Covid-19 vaccine, “it doesn’t mean you can find a way to do it for other vaccines.”

Viral vector vaccines work along a different method. Instead of taking the protein, scientists find the genetic code that creates the protein, separate it from the virus itself, and attach it to another virus – often an adenovirus, which causes colds. That virus, usually genetically modified so that it can’t reproduce, is introduced to the body; it hijacks the body’s cells, and gets them to produce the protein, again preparing the immune system for the arrival of the real virus.

The Messenger RNA Vaccine

GreenLight, meanwhile, works with RNA technology, which it hopes will solve many of the world’s most pressing problems, in medicine, agriculture and the environment. But right now they’re hoping to use one type of RNA, called messenger RNA, to develop and manufacture a vaccine for Covid-19.

RNA is the molecule that translates your genetic code – your DNA – into proteins. Your DNA is a long string of small molecules, called nucleotides, linked together to form the two chains you know as the DNA double helix. RNA is a copy of the sequence of nucleotides, which then “reads” that sequence and translates it into a sequence of amino acids, the building blocks of proteins. In short, RNA, and messenger RNA in particular, tells your body how to make proteins.

From the point of view of someone trying to make a vaccine, this has great benefits. We can relatively easily, now, find, isolate and reproduce sections of the virus’ DNA or RNA to provoke an immune response. GreenLight and other groups trying to make messenger RNA vaccines seek to find the RNA sequence that most effectively codes for the Covid-19 spike protein.

By injecting that RNA code into our cells, the patient’s body itself takes over the role of producing the spike proteins that will stimulate an immune response. The patient’s body undertakes the factory-like role of the chicken egg. By using the body as a factory, however, there is no need for the complex, time intensive and expense purification of the final product required when it was grown in a chicken’s egg.

Another key advantage is the flexibility of substituting new RNA sequences.

“A piece of RNA is a piece of RNA. It’s always made up of the same four building blocks.” says Fielding. “So the process of making it is the same, even if the protein it codes for is different.”

Instead of having to make a new, complicated process every time you need to make a different vaccine, you can use the same process each time – you just need to plug the new RNA sequence in, and away you go. Fielding thinks that, once a suitable protein has been identified, it would be possible to create a new vaccine in “weeks, rather than months” – she imagines that a flu vaccine, for instance, could be designed in 12 weeks, instead of 12 months. “It could be done four or five times faster.”

Breaking the vaccine production bottleneck

There are other advantages, too. The lower cost of manufacturing – the sheer ease with which RNA vaccines can be produced – makes it easier to create candidate vaccines for trials. If it costs a quarter of the money and time to research the vaccines, you can research four times as many. If RNA vaccines become standard, the whole pipeline could become faster. It could lead to a world in which new vaccines can be produced within a few weeks of a virus being identified. In contrast to the slow production cycles that result in a hit or miss for traditional vaccines, RNA vaccines hold out the promise of a rapid response, which could become a cornerstone for future pandemic preparedness.

This is still in the future. But in the nearer term, it could be a realistic answer to the basic difficulty of a Covid-19 vaccine: the scale of the problem. If 70% of the population needs to be vaccinated in order to achieve herd immunity, as most scientists believe, that means about five billion doses. If a booster dose is required, that number doubles. Most vaccine manufacturers are talking about making millions of doses, a challenge in its own right at short notice – but GreenLight plans to make billions, either of their own vaccine or of a successful RNA-based vaccine developed by others.

And, because – as Fielding says – a piece of RNA is a piece of RNA, the GreenLight manufacturing process can be used to build any messenger RNA vaccine, including those developed by  other teams.

So from the outset, it has aimed to use low-cost, widely available materials, and a manufacturing process that can scale up easily from laboratory to industry. Hopefully, this technology will help win the war against Covid – and perhaps, one day, it will lead to a flu vaccine that actually works, every year.

How to make an RNA based vaccine

The process of making an RNA vaccine can be essentially divided into three parts: building the molecule; purifying the solution; and formulating it for delivery.

Step One: Building the molecule

“The process of building the molecules is, at its core, the way nature makes it,” says Fielding. In nature, your cells build RNA molecules from a DNA template, by taking the free nucleotides that are floating around in the cell. The GreenLight procedure does exactly that, except in the lab. They take a length of DNA that codes for the protein they want – the template. They put that into a vat full of free nucleotide bases, and use enzymes that will – just as in nature – start grabbing those nucleotides out of solution, and make RNA copies of the DNA template.

There’s one difficulty, which is that nucleotides do not react easily if left to themselves. Imagine a ball sitting in a shallow dip at the top of a hill. It would roll all the way down the hill, but it can’t without a shove to get it out of the dip. In just the same way, some reactions need a little bit of help to get them started. Building RNA out of nucleotides is one such.

Most RNA engineering involves “charged” nucleotides, which have had energy put into them in advance – essentially giving them the shove that they need. This approach has disadvantages. For one thing, charged nucleotides are less stable because they’re more reactive. For this reason, they must be kept frozen. And for another, they’re more expensive to get hold of: they have to be bought from specialist companies.

The GreenLight system uses a very simple form of nucleotides, which are readily available and often used in baby formula production. They can often be cheaply sourced in bulk, are stable and do not need special storage. 

To encourage them to react, GreenLight introduces a catalyst, an enzyme, similar to how cells do it naturally. This reduces the amount of energy required to start the reaction – it removes the walls around the “dip” the ball is sitting in, so that it can roll downhill more easily. “My metaphor is that you can buy wallpaper that you peel and stick, but it’s expensive,” says Fielding. “Or you can get the kind where you paint the paste on yourself. We’re doing the latter.”

Step Two: Purifying the solution

The process of building the molecule rapidly produces lots of the RNA sections that you want, but they are still swimming in a soup of shorter lengths and free nucleotides. “You need to inject it into a person,” says Fielding, “so you have to purify away all the things you used to make the RNA. There are lots of impurities you don’t want.”

This has been a difficult part for messenger RNA engineers, with many using “reverse-phase high-performance liquid chromatography”. In essence, filtering your material through a tube filled with solvent, which separates out the molecules by things like mass and charge. “It’s time-consuming, not scalable, and expensive plus it involves toxic solvents,” says Fielding.

GreenLight is developing a water-based model, which involves no toxic waste, and which – while not yet cheap – is at least less expensive, with systems that can be scaled up to industrial levels. “We wanted to make this messenger RNA available at a large scale and inexpensively.” It has to be, in order to make the billions of doses that GreenLight is planning.

Step Three: Formulating it for delivery

So you have your purified solution of the RNA molecule, and now, you need to “formulate” it. You can’t simply inject a bunch of loose RNA molecules into your bloodstream and hope it works – it would degrade and never get into the cells where it needs to do its work. 

So it is delivered via lipid nanoparticles. Some lipid (fat) molecules have an interesting property: one end of each molecule is attracted to water molecules (hydrophilic), and the other end is repelled by them (hydrophobic). When mixed with water and other molecules, such as RNA, the hydrophobic ends will huddle around the RNA, forming little spheres with the hydrophilic ends on the outside. 

These tiny particles, less than a thousandth of a millimeter across, can make it through the outer membranes of our bodies’ cells, and deliver the RNA into the mechanisms within, where they can start making proteins.

All of these steps can be carried out with commercially available equipment, with the slight exception of the formulation, which involves some trade secrets. “It’s mainly just tanks, some heat, some water,” says Fielding. “Nothing that’s super expensive or difficult to obtain. The beauty of this system is that it’s simple: no solvents, no toxic waste.”
From this process comes the final outcome: Vaccine ready for bottling and injection.