We've updated our Privacy Policy to make it clearer how we use your personal data. We use cookies to provide you with a better experience. You can read our Cookie Policy here.


Turning On the Vaccine Tap – The Challenges of Scaling Up Manufacturing

Turning On the Vaccine Tap – The Challenges of Scaling Up Manufacturing content piece image
Listen with
Register for free to listen to this article
Thank you. Listen to this article using the player above.

Want to listen to this article for FREE?

Complete the form below to unlock access to ALL audio articles.

Read time: 10 minutes

Vaccine manufacturing

According to the World Health Organization’s
COVID-19 vaccine tracker, 102 are in clinical development. Eight vaccines are approved for full use for COVID-19, while seven are in early or limited use. There are four vaccine classes that are particularly prominent among these.

About 30% of vaccines in development are protein subunit vaccines. Most of the remainder are either non-replicating viral vector (16%), DNA vaccines (10%), inactivated virus (16%) or RNA vaccines (16%). Vaccines must be manufactured at huge scale to make inroads into global populations, and each of the
four most popular vaccine types faces its own challenges.

Inactivated vaccines are the most traditional platform, while subunit protein vaccines have boomed over the last few decades. Nucleic acid and viral vector vaccines are important new kids on the block, and their novelty poses difficulties in scaling up production. The leaders in production volume have been Pfizer–BioNTech and Moderna (RNA vaccines), AstraZeneca, Sputnik V and Johnson & Johnson (viral vector) and Sinopharm and Sinovac (inactivated).

Inactivated vaccines

The inactivated vaccine is a traditional production platform used for decades to make, for example, inactivated poliovirus vaccine. By the early 1980s it was possible to
cultivate polioviruses in monkey kidney cells in fermenters of 100 to 1,000 L, disable them and fill them into vials. Today, SARS-CoV-2 is being grown up in vats of monkey kidney cells before being inactivated using a chemical called beta-propiolactone. Most vaccines produced in this way are from two Chinese companies: Sinovac Biotech in Beijing and state-owned Sinopharm.

This older technology was largely ignored by Western pharma companies, but the Chinese companies have surprised industry watchers by ramping up production of these vaccines. In June, Nature 
highlighted how this has allowed China to vaccinate about 20 million people per day. This accounted for 60% of all doses given globally. “This is impressive. It is about twice what the UK is administering each day,” says Zoltán Kis, a chemical engineer at the Future Vaccine Manufacturing Hub at Imperial College London in the UK, though of course China’s population is much bigger.

This is all the more impressive given that inactivated virus requires far more time and effort to make than some of the newer vaccine technologies. Manufacturers must first grow the pandemic virus in living cells, which can take months, and then disable and purify out the virus. Since the virus is inactivated, larger doses may be needed than for live vaccines or for mRNA type vaccines, and some contain adjuvants to boost the immune response. Also, large amounts of infectious virus need to be handled, and the virus needs to retain integrity, which must be confirmed, as noted by immunologists
Fatima Amanat and Florian Krammer in the journal Immunity in April 2020. Still, “inactivated vaccines can be made by almost anybody,” says Krammer, which could allow this form of vaccine in many countries. Though there remains challenges in obtaining equipment and expertise.

The tricky part, according to drug discovery chemist Derek Lowe
in his Science blog, is "inactivating the virus enough so that it can’t infect cells and replicate, but not so much that it presents totally different proteins to the immune system and raises a response that won’t help against the real virus". It takes time to inactivate a virus, says Kis, usually around two weeks (though this varies with the virus), and quality control is crucial to ensure you do not inject live vaccine into people’s arms. “All things being equal, this is the slowest way to make vaccines for sure,” adds Kis.

Manufacturing this vaccine type certainly seemed to take time to scale up, but is now going well. “China’s vaccination campaign got off to a slow start, but has rapidly picked up pace,”
Rongjun Chen at the Future Vaccine Manufacturing Research Hub at Imperial College London told Nature. China probably leveraged its existing capacity for making vaccines based on inactivated viruses such as for influenza and hepatitis A, according to Jin Dong-Yan at the University of Hong Kong in the same article. Another inactivated vaccine is Covaxin, designed by Indian company Bharat Biotech, which announced in May 2021 plans to produce 200 million doses per annum. The science analytics company Airfinity told Technology Networks that it projects that Sinopharm will make just under 1 billion, Sinovac 1.7 billion and Bharat 410 million doses in 2021.

Messenger RNA vaccines

No mRNA vaccine had been licensed prior to the pandemic. Now, two leading vaccines for COVID-19 in Western countries are mRNA-based, including BNT162b2 from Pfizer–BioNTech and mRNA-1273 from Moderna. Manufacturing mRNA vaccines involves a polymerase enzyme creating a string of mRNA by linking nucleotides together, working from a DNA template. The nucleotides themselves are made chemically, immunologist
Drew Weissman at the University of Pennsylvania told Chemistry World, and making the mRNA itself is uncomplicated.

mRNA synthesis takes a few hours, while the entire manufacturing process takes just a few days, notes Kis. The DNA template is broken down using enzymes, and the RNA strand left behind is separated out using tangential flow filtration or various chromatography methods.

The mRNA codes for the spike protein once it gets inside human cells, but the mRNA sequence for the Pfizer–BioNTech and Moderna vaccines must first be mixed with lipids in a precisely controlled process. This slows the manufacturing process down, says Kis, because encapsulating RNA into lipid particles requires exotic mixing technologies: microfluidics brings together miniscule amounts of fluids, to mix the two streams in a precise way. “This is new technology and some of the materials that go into making these vaccines are also new and they don’t have an established supply chain,” says Kis.

Some of the lipids are made by smaller companies, which would have had limited production capabilities at the start of the pandemic, yet owned important intellectual property. mRNA vaccine manufacture has been scaled up during the pandemic, with Moderna projected to make 750,000 doses this year, and Pfizer–BioNTech 2.4 billion, according to Airfinity. Moderna
predicts it will make 3 billion next year. Kis says it is not straightforward to move production to new sites, however, “the experts who know how to do it are busying doing” rather than being involved in training others," he explains.

Stéphane Bancel, the chief executive officer of Moderna, said that it takes six to nine months at a minimum to add significant capacity. Even then, equipment has to be purchased and there is the challenge of scaling up all the consumables for mRNA vaccine from almost nothing at the start of 2020 to enough for billions of doses in 2021. Moreover, with vaccine production being scaled up like never before, even fill and finish will become a bottleneck, says Kis.

Nonetheless, mRNA vaccine manufacture has some fundamental upsides. “These are much simpler processes compared to other vaccines, and they don’t rely on growing cells, which adds to complexity,” says Kis. He points to BioNTech’s purchase of a Novartis facility in Germany that made cancer drugs and which, six months later, was already producing mRNA vaccines; a very quick timeline. The same number of vaccine doses can be produced in 30 to 50 L bioreactors at an mRNA production facility as can be produced in cell-based production facilities that work with around 2,000 L bioreactor volumes. “This can have advantages when setting up new mRNA production sites. The capital costs and facility footprint/area requirements would be lower, compared to cell-based vaccine production,” Kis adds.

A third mRNA vaccine manufacturer, CureVac, was expected to begin producing vaccines, but interim analysis of a Phase III clinical trial, HERALD,
did not meet its prespecified statistical success criteria. Kis had noted that the CureVac vaccine contains lower amounts of RNA per shot, which would have meant one gram of RNA could make more vaccines.

Subunit, recombinant and conjugate vaccines

Instead of provoking an immune reaction against an entire virus, protein subunit vaccines show the body one or two critical components. In the case of SARS-CoV-2, protein subunit vaccines consist of purified spike protein or parts thereof. Since these fragments are incapable of causing disease, subunit vaccines are considered very safe and suitable for people with compromised immune systems, notes
Gavi, the Vaccine Alliance organization. They are also relatively stable, but sometimes require adjuvant chemicals to prod the immune system into a stronger reaction.

All subunit vaccines require living organisms for manufacture. For hepatitis B, yeast are genetically engineered to make a viral protein and then grown in large fermentation tanks. This approach has become extremely popular, and most vaccines on the childhood schedule such as for whooping cough, tetanus and diphtheria are subunit vaccines. This is established technology that has been around for decades,
Peter Hotez at Baylor College of Medicine says, and companies know how to manufacture recombinant proteins at huge scale. The downside is that protein vaccines take longer to develop than most of the other classes, adds Nikolai Petrovsky, vaccine scientist at Flinders University in Adelaide, Australia.

The lead protein subunit vaccine for COVID-19 is from US company Novavax. This vaccine is made up of two parts. First, a recombinant protein nanoparticle that is made by encoding genetic sequences of the spike protein into a baculovirus. These viruses then infect special moth cells, which are cultured to allow the virus make the spike protein. This has all the associated downsides of using cells – it take time to grow the insect cells, infect them with virus and harvest the protein, all while avoiding contamination. The second part of the vaccine is a proprietary adjuvant, Matrix-M, and this relies on a natural ingredient from the bark of a South American tree, Quillaja saponaria. Supply of this compound could pose problems, since it is used in some other vaccines, including a blockbuster shingles vaccine from GSK.

So far, Novavax has yet to gain authorization for its vaccine, but said that it intends to file for authorization with European, UK and US regulators in
the third quarter 2021. In March, Reuters reported that the company was “struggling to source some raw materials." In May, the Washington Post said that Novavax would not seek emergency use authorization in the US until at least July “because of a regulatory manufacturing issue related to an assay.” The report also noted that raw material shortages were delaying the company’s ability to reach manufacturing targets. The company had reached an agreement with the Serum Institute of India that could allow them to produce 2 billion doses per year, “but supply chain issues have forced Novavax to push back its predictions,” according to The New York Times, with Reuters reporting that the firm’s target of 150 million doses per month will take until the third quarter of 2021. Analytics predicts total production for 2021 will be 348 million doses.

Viral vector vaccines

There are
four approved vaccines that rely on viruses to deliver the recipe to make spike protein into our cells. These are the AstraZeneca vaccine developed by Oxford University, the Russian Sputnik V vaccine developed by Gamaleya, the one-shot J&J vaccine and a one-shot vaccine developed by CanSino Biologics. All rely on live adenoviruses. This vaccine type is far more novel than often assumed, being first authorized (for Ebola, by the European Medicines Agency) only in 2020.

Manufacturing viral vector vaccines is not so straightforward, since it involves growing up mammalian cells in tanks at a scale of 2,000 L. “Growing living cells adds complexity and variability to the process,” says Kis. “You have to feed these cells, and they have their own lives and are expressing genes, signaling to each other and secreting.” Growing the cells needed for the AstraZeneca vaccine takes two months using
human embryonic kidney cells, and it is often only at the end that problems might be detected and an entire batch discarded, adds Kis.

The result is that some facilities may miss production targets, and industry is on a learning curve in growing the necessary production cells and then introducing and later purifying the live virus required for the vaccine. AstraZeneca has announced a number of production setbacks,
in Belgium, in Latin America and in the Netherlands. J&J told the EU in March that it had supply issues around ingredients and equipment. “This is normal because these are processes that were developed fast,” says Kis. Lowe told Chemistry World about a J&J site in Europe that he heard “had stubbornly low yields of adenovirus” that they were struggling to troubleshoot. Errors have been made too. US contract manufacturer Emergent Solutions made mistakes which led to 60 million doses of the J&J vaccine being binned, Fierce Pharma reported. The manufacturer also had to discard a number of AstraZeneca batches in 2020. 

There are also indications that Russia is struggling to produce as much Sputnik V as it had hoped, with a Reuters tally indicating that the country had produced 33 million vaccines as of May 12 and exported fewer than 12 million, as
noted in Business Today. The Russian vaccine is more complicated than J&J’s and AstraZeneca’s because it involves producing two different human adenoviruses. “They are trying to outsource production or fine partners to produce it,” says Kis.

One solution to the production challenges of viral-vector vaccines is to build more plants, but this takes time, and many of the specialists in the area are already busy with vaccine production. “We can buy equipment, we can build plants. But in biotechnology, competent people are the most important thing. And there are not very many of them,”
Vikgram Punia, CEO of one private producer, Pharmasyntez, said. Science analytics company Airfinity projects that AstraZeneca will make 2.2 bn doses and J&J will make half a billion doses of their main COVID-19 vaccines, with 291 million doses predicted for Sputnik V and 207 million doses for the CanSino adenovirus vaccine.

The emergence of the delta variant has added urgency to the need to manufacture and roll out vaccines globally. Eventually, companies are likely to generate enough supply for everyone, but for now production challenges remain for all classes of vaccine, in terms of new processes, accessing enough raw materials, as well as obtaining or building manufacturing sites and the expertise needed to make large quantities of COVID-19 vaccines of reliable quality.

About the author

Anthony King is a freelance journalist based in Dublin, Ireland.  He has written on SARS-CoV-2 and/or vaccines for publications such as The Scientist, Chemistry World, National Geographic, Science Magazine, New Scientist and the Irish Times.  He write on a variety of topics, mainly in biology and chemistry, and sometimes science policy and the pharmaceutical industry.  He tweets at: