Volunteers in Wuhan receive an experimental coronavirus vaccine which is in a phase II human clinical trial Credit: TPG / Getty


April 28, 2020   9 mins

There are, I think, three possible futures for the coronavirus pandemic. Either nearly everyone gets it, and the survivors stay immune, and we get herd immunity that way; or it becomes endemic in the population, turning up each year like flu; or we get a vaccine.

In the shorter term, there are other options: decisions to be made about testing and tracing, about when to open schools and bars and football stadiums, about masks. They’re all important. But we can’t keep people in their homes forever, and unless we eradicate every single case of the disease, it will eventually come back. So, as people say, unless everyone gets it and the disease burns itself out, there’s no going back to normal until we get a vaccine.

So: when will that happen? There are bullish reports from Oxford that a new vaccine is showing early promise; that it has protected inoculated macaques from the virus when unvaccinated ones consistently get sick. They think that they could get the first few million doses off the production line and ready for human consumption by September. Other people talk about being able to return to full economic activity in 18 months.

Not everyone is so hopeful, though. A report by a biopharma analyst for the investment bank SVB Leerink, titled “Sober Up! 25 Reasons Not to Count on COVID Vaccine for Herd Immunity in 1-2 years”, suggests that it could be years, rather than months – attesting gloomily that “There is absolutely no evidence that a vaccine can protect against human infection with SARS-nCoV2” and that new kinds of RNA- or DNA-based vaccines are so far completely untested.

As I said last week, predictions are hard, especially about the future. I don’t know which of the two extremes is more likely. But I thought I could go through some of the challenges.

All coronaviruses have a distinctive shape – a ball with spikes coming off it, like an old-fashioned naval mine. The spikes are made of a protein called an S-glycoprotein. Most research into Covid-19 vaccines, says Robin Shattock, a professor of mucosal infection and immunity at Imperial College London, targets those spikes. That is, they try to mimic either the whole spike or a part of it, so that — when the vaccine is injected into our bloodstream — our immune systems will latch onto that S-glycoprotein and learn its shape. Then, if the real coronavirus turns up, not just a disembodied spike but the whole naval mine, our immune systems will remember the protein and attack it.

So your first job is to make a stable version of that protein, separated from the rest of the virus. Then once you have one that looks promising, you test them in animal models.

Assuming it’s safe and effective in ferrets or rhesus macaques (the two most useful animal models for this research, since they’re both susceptible to Covid-19), the next stage is to produce a higher-quality version of the vaccine, test that again in animals to make sure it’s safe, then try it on humans.

The first human trial will be small, and in young, healthy volunteers. It’s just looking to make sure that it’s safe and gives the sort of immune response you want. Then you do more trials, in larger numbers and in older or less healthy volunteers, and eventually you see if the vaccine actually works — if people who are given it are less likely to become infected.

After that, if the vaccine is safe and effective, it will be licensed – in Europe by the European Medicines Agency (EMEA), in the US by the Food and Drug Administration (FDA); once it is licensed, it will be manufactured and then given to the population. And then – fingers crossed – the disease will die out.

So far, so straightforward? Not really. 

One problem that recurs at several stages is that vaccines are different to treatments. If you have some drug that you think will cure a particular cancer, you find people with that cancer, you give some of them the drug and some of them a placebo (or, rather, the usual treatment for that cancer); if they do better on the new drug, it works.

But with vaccines, by definition, you’re giving it to people who don’t have the disease, and then you want to see if they get the disease.

In animal models, that’s not such a huge problem, because you can give them the disease yourself — or “challenge” them, in the terminology. It’s still not ideal when — as now — you’re in a huge rush; it takes some time after exposure for your body to develop enough antibodies to be properly immune. “To see if the vaccine works, the minimal time to get to a level where you want to challenge them is about four weeks,” says Shattock. “So those studies typically take a couple of months.”

But in humans, usually, you don’t want to deliberately give subjects a potentially deadly disease. “I can’t rule out that it might happen,” says Shattock, “but it’s ethically difficult.” Covid-19 is much less dangerous for young healthy people, but it’s not completely without risk, and there’s the the chance “that if you took 10 people in their 20s and gave them the virus, you might just be unlucky”. One or more could get really ill for reasons you don’t understand; they could die. Kirsty Gelsthorpe, a spokesperson for the Association of the British Pharmaceutical Industry, says that the Oxford group certainly isn’t doing it.

Besides, says Shattock, “it’s quite artificial; shoving virus up their nose mimics natural exposure, but it’s not the same”.

So usually, you give the vaccine to a bunch of people in the wild, as it were, and you see if they get the disease less than a control group over some period of time. But, Gelsthorpe points out, at the moment, we’re all locked down. Fewer people are getting the disease. So that causes a problem: it’s good news in the sense that not as many people are getting ill, but it slows down research. You don’t know if your vaccine works if no one gets sick in your control group.

“That’s why many groups,” says Shattock, “including ourselves, are moving very fast. One, because we want to get it made quickly, but two, because we want to be able to test it while infections are still going on.”

There’s another problem. Vaccines are already dealing with a trust problem — anti-vaccination movements around the world have led to significant drops in the takeup of the MMR jab in the West and polio and TB vaccines elsewhere. 

In the phase I trials — the small-scale ones — you’re looking at safety. Usually, that doesn’t mean whether or not it kills people; serious reactions are vanishingly rare, despite the awful and high-profile Northwick Park incident 14 years ago, which left six previously healthy men fighting for their lives. Normally,you’re just looking to see if “their arms are so painful that they can’t go to work, or if they have flu-symptoms that make them feel horrible”, says Shattock.

Of course, if you have some really nasty disease, you can shift your threshold somewhat. “If you’re facing an Ebola epidemic and 50% of people are dying,” says Shattock, “you’re not going to be worried if your arm is sore for a few days.” And since you only need to give it to the most at-risk people, you can be forgiven if you’re not as vigilant as you might be about long-term side effects: “Everyone in the [Ebola-]affected areas thought ‘This is going to kill me. If I take the vaccine I don’t care if I have side-effects in years to come; I won’t die.”

But that’s not true of coronavirus. It’s dangerous, but it’s nowhere near that dangerous, and you have to convince millions of healthy people in their 20s and 30s who aren’t at significant risk to test the vaccine. “The level of what’s tolerable, vs the perceived risk, is very different,” says Shattock. “Safety will be particularly important. You don’t want people turning up five years later with some weird arthritis and asking if it’s caused by the vaccine. 

“Vaccines struggle anyway because of the anti-vaccine movement — the difficulty is that they’re so successful no one sees the disease, they only see the side effects. You don’t want anything that will make that worse.” If the coronavirus vaccine kills some healthy people, it could easily reduce trust in other vaccines and lead to major loss of life.

But let’s imagine we get our vaccine through the testing, and get it licensed. Then we just need to produce it and ship it out. That should be easy, right?

Jeffery Almond, a visiting professor of microbiology at Oxford and for several years the head of research at the French pharmaceutical giant Sanofi, says that the production of a new vaccine usually takes a long time — “10, 12, 15 years” — and that, traditionally, “we always had to engage with manufacturing technology”. What that means is that about half-way through the process of researching the vaccine, they’d have to start designing and building a factory to produce it.

“Typically that process took four or five years,” he says. “Designing the building, getting the equipment ordered, installed, validated, operational. Getting it licensed as a production facility. Minimum three years, then you’d have a facility with a capacity of a few tens of millions.”

We’ve been unlucky with coronavirus, he says, in that everyone expected the next pandemic to be a flu virus, “like 1918, 1957, 1968, 1976, 2009”. There’s industrial capacity for manufacturing flu vaccines — we make millions of doses every year. But those factories won’t work for coronavirus vaccines, and we need billions of doses. “To get a billion doses by this time next year or Christmas or whatever, when you don’t have industrial capacity, is a hell of a challenge,” says Almond.

I can sound a note of optimism here, though. Almond is talking about old-school vaccines, which are made with attenuated viruses grown (in many cases) in chicken eggs or other cultures. There are more modern versions. Some groups, he says, are working on vaccines which target stretches of viral DNA or RNA. “For them the manufacturing is a chemical synthesis of the RNA or DNA,” he says. “It’s intrinsically easier to scale up.”

But so far none of these vaccines exists. “We have plenty under development,” Almond says, “but the manufacturing capacity isn’t there, and we don’t really know if it’s going to work. It’s a bit more risky in that the technology is new.”

His own former employers, Sanofi, are working on vaccines that you can culture in insect cells, which he says is vastly more scalable than the chicken-egg method, and could easily be applied to coronavirus. When used in conjunction with an adjuvant produced by GSK which would, he says, allow you to “reduce the dose per person to a fifth or a 10th of what you’d need without it”, the Sanofi model “is the most straightforwardly scalable version and the manufacturing is already there”.

The Oxford team are using something else entirely — a vaccine that uses another virus, an adenovirus,  as its delivery system. Philip Ridley Smith, the marketing director for the firm Cobra Biologics and Pharmaceutical Services, which works with the Oxford researchers, says that they’re building a 200-litre bioreactor to start producing the vaccine. It would be enough to make a million doses a batch, and, says Ridley Smith, the consortium of three manufacturers that make up the Oxford consortium hope to make six million doses a month.

For that reason, Shattock is quite hopeful that when the vaccine arrives, it will be quite easy to produce enough to vaccinate everyone in Britain pretty quickly. “Solving it on a national basis is relatively straightforward,” he says. Building enough for a global scale will be the tricky bit, although Ridley Smith points out that many developing nations, such as India and China have excellent manufacturing, if not the same refrigerated supply chains.

One trouble will be funding it. Pharma companies are throwing money at this at the moment, but they are in the end commercial enterprises, and vaccines are often not very commercial things: they are given once, and need to be given to people who can’t easily pay for it. Plus, they’ve been burned before — in his book Deadliest Enemy, the epidemiologist Mike Osterholm says that they rushed to get a vaccine for Sars in 2003, urged on by governments and philanthropic bodies; then, when Sars burnt itself out, the governments and philanthropic bodies lost interest. 

“At Sanofi we rushed to make a Sars vaccine,” says Almond. “It got to the point we could test it in primates after 11 months, and by then it had gone away, and we’d spent tens of millions of euros on it never to get it back. The industry does it, but it knows it’s losing money, and it can’t do that too much.”

“I’m not sure how scarred pharma was by that,” says Shattock, “but they’ve certainly been bruised by other projects: Ebola, Sanofi with the dengue programme, GSK with malarial vaccines. They’re not seeing big returns.” It may be different with Covid-19 if it’s around for years, especially if it needs yearly doses, but it may not. 

Osterholm in his book suggests that public-private partnerships — like the US defence contractor model, in which governments put out a specification for a jet fighter, and firms compete to build one in the knowledge that the most suitable will be richly rewarded — may be the best system. That’s probably for future outbreaks, though; this one will be driven by funding from the Gates Foundation, Wellcome, the World Bank, and by goodwill from the pharma companies. “Yes there’s a risk the companies will make a loss,” says Almond, “and on smaller things they might need incentivisation, but on this one they’ll dive in and do the best they can.”

So will we see a vaccine soon?

Ridley Smith is confident: “We would hope that we’d be able to start getting the vaccines out by September” – although he warns that it will depend on the success of the trial, and on enough people in the control branch getting infected to give them working statistics; ironically, too much success at limiting the disease might mean we can’t get a vaccine. The Oxford researchers have encouraged at-risk workers to volunteer for the trials to maximise the chances of getting good information.

Shattock is less so. “If everything went really well, and one of the two UK vaccines work, that could start to be rolled out early next year,” he says. That could happen: “If we were working with no knowledge, it would be a very low percentage chance,” he says. “But because we know a lot about coronaviruses, that increases the chances of success a lot, although many things could still go wrong.” But getting it rolled out globally will take longer; he thinks even 24 months is “very optimistic”. 

And it’s possible, of course, that in 24 months, we might not even need a vaccine any more. The virus might have run its course. “How important will it be by 2022? I suspect we’ll still need it for vulnerable groups, but we may see that by that stage most of the world has had the virus and we’ll have herd immunity.” 

The scientific community has got a lot faster at researching and developing viruses, but it may not yet be fast enough to help us with the coronavirus. Let’s hope this outbreak at least means we’re better prepared for the next one. 


Tom Chivers is a science writer. His second book, How to Read Numbers, is out now.

TomChivers