Novel vaccines have performed remarkably quickly and well

ON NOVEMBER 30TH 1803 the María Pita, a 160-tonne corvette, set sail from Spain for the New World. King Charles IV was keen that his subjects over the ocean benefit from the new technology of vaccination, which used inoculation with pus from blisters due to cowpox, a comparatively minor ailment, to engender immunity against smallpox, a scourge which killed millions—including, a decade earlier, the king’s beloved daughter.

The initial challenge faced by the man in charge of this expedition, Francisco Balmis, was how to get viable cowpox across the ocean. His solution was to take on board 22 boys from an orphanage, two of whom he had infected with cowpox beforehand. During the voyage West the infection was kept viable by being carefully spread from boy to boy, and over the four years which followed, the virus transported and nurtured within them was used to inoculate hundreds of thousands of people from Peru to the Philippines. (Balmis relied on a fresh supply of orphans obtained in Mexico to get the virus safely across the Pacific.)

The Balmis expedition provides a striking illustration of a basic truth. Creating vaccines is a job for biology, and doing biology requires living systems. From Edward Jenner’s discovery in 1796 that cowpox could be used as a vaccine—the word vaccination, coined shortly thereafter, is derived from the Latin vacca, for cow—all the vaccines humankind has used to protect itself were produced in living cells. Until last year.

Over the course of the pandemic a fundamentally new vaccine technology has come into its own. When the first, spectacularly positive results from phase 3 trials of the Pfizer/BioNTech mRNA vaccine were released on November 9th they did not just offer a pandemic exit strategy. They also showed, as did the results of the similar vaccine made by Moderna, that the long process through which science has abstracted biological mechanisms from their fleshy and fibrous substrates has reached a new level. In a small but vital way, medicine has begun to look like programming.

The growth media are not the message

For a long time vaccine production was an often messy and sometimes disgusting business. The polio vaccine developed by Jonas Salk in the 1950s was made in vats of minced kidneys, a process which required thousands of rhesus monkeys to be farmed and killed. In the 1968 flu pandemic Maurice Hilleman, and American virologist, got through chicken eggs by the tonne as he grew 9m vaccine doses in their yolks in just four months.

Though the chicken-egg method is still used for flu vaccines today, later vaccine-makers worked out how to grow soups of individual vaccine producing cells—cell cultures—in vats large enough to supply what was needed to control all sorts of other diseases. Doing so made things safer as well as easier. The polio vaccines made from viruses grown in minced monkey kidneys also contained a somewhat dodgy looking virus, now known as SV-40, that was not detected until it had been injected into tens of millions of people. Had it been carcinogenic—there is no evidence that it was, though some debate on the matter continues—it would have been a disaster to make a nuclear meltdown look small-bore.

But the manufacturing systems that are being used to make the vaccines against covid-19 currently being administered at a rate of millions per day are something else again. They do not produce weakened versions of the virus being vaccinated against, like those used in flu jabs and later versions of the polio vaccine. They do not even produce specific antigens. Instead they just send a message: the genetic sequence which describes the SARS-CoV-2 spike protein. When presented as messenger RNA, or mRNA, this message gets cells to produce the protein just as they would if they were infected by the virus. That gives the immune system a risk-free preview of what infection would look like, allowing it to develop its response ahead of time.

The new vaccines deliver their message in two different ways. In the Oxford/AstraZeneca, Johnson&Johnson and Sputnik V vaccines it is contained in a protein shell derived from an adenovirus. Cells are engineered to make adenovirus particles which have a DNA version of the sequence that describes the spike protein nestled inside them. The particles they produce are then harvested to make vaccine. The adenovirus gets the DNA into the vaccinee’s cells; the cells transcribe the DNA into mRNA which they then use to make spike proteins.

Harvesting virus from a cell culture is hardly a new form of vaccine-making. But the adenovirus, as used in this system, is not really the vaccine. It is not there to produce an immune response to itself. It is the platform which allows the DNA within to get into cells and set about producing antigen there.

An important part of this is that the message in the DNA is independent of the particles into which it is packed and of the cultured cells that do the packing. The same manufacturing infrastructure—the same recipe for growth medium, easily modified versions of the cells, identical protocols—could be used to load a different DNA sequence into particles which would look just the same. The system is not an interconnected whole, as biological systems are in nature. It has been rendered modular.

This modularity is helping vaccine companies produce second-generation jabs that protect against variants of SARS-CoV-2 with which the first generation may cope poorly (see next chapter). Insert the DNA for a version of the spike protein which engenders a better response, reboot the system and off you go. In principle the companies might use the same approach to deliver the DNA for a completely different antigen, and thus a vaccine against a completely different disease. “It’s very different from making vaccines in the old days,” says Greg Lemke, a biologist at the Salk Institute (yes, same Salk). “Then every vaccine was a new project with a whole new infrastructure. With these technologies, every new vaccine is the same project.”

The second of the new technologies, mRNA, works on similarly modular lines, but is yet more radical. A DNA version of the spike gene is used to make huge amounts of mRNA in a cell-free system—a solution that contains an enzyme called RNA polymerase, the chemicals that power its work, and the raw components from which RNA is made, all of which can be bought off the shelf. The mRNA thus produced is then packaged into tiny particles of lipid—the inert material from which the membranes around cells are made. After the original DNA has been harvested no cells are involved. It is all a matter of clean, scalable industrial chemistry.

This simplicity has allowed the mRNA vaccines to be designed and produced on a massive scale in an incredibly short time. Vaccine companies expect to make 2.6bn doses of mRNA vaccine in 2021 using manufacturing techniques proved at scale only last year.

Established drug companies and eager startups expect the systems which are producing the mRNA vaccines to be turned to new purposes. “We’re poised to enter a new age of using this technology,” says Mark Stevenson, chief operating officer of ThermoFisher, which makes polymerases, the components of RNA and the purification resins used in the mRNA-vaccine process. The scale of covid-19 vaccine production has spurred a huge expansion of demand for its wares. Mr Stevenson says the firm will increase its capital investment by up to $1bn in 2021, with a “meaningful proportion” of the new money beefing up the RNA supply chain.

Upgrading an mRNA vaccine is even easier than doing so for an adenoviral one. Change the DNA template and you get a different mRNA—but one which will fit into just the same sort of lipid coating. Ugur Sahin, the boss of BioNTech, says that new versions of his company’s vaccine can be turned around in six weeks. “About three or four weeks of that is testing,” he says. “The vaccine is produced after two weeks.” Nothing besides the sequence being delivered, Mr Sahin believes, will change as the system moves from one product to the next.

When it comes to products destined for the human body, regulators will need convincing that one output of such a modular set up is as safe as another. If it is, then the emergency expansion of new manufacturing technologies is likely to change the world. Easily made, precisely programmed mRNA vaccines are being looked at for a number of infectious diseases—including malaria—as well as in cancer immunotherapy.

Abstraction, repeatability and modularity open new pathways to innovation. For a sense of how this can prove transformative look to the world of computers. Hardware designers work with specified components which do what they are expected to; they have set rules for putting these components together—rules that allow them to build systems far more complex than would be possible if every detail of how every component worked had to be specified from scratch. And users write software which does not need to depend on the quirks of the hardware which embodies it.

The gurus of “synthetic biology”, a school of thought and practice which seeks to re-engineer living systems in a way that makes them easier to engineer further, have been talking about such approaches for decades. They are also seeing them used in an increasing number of industrial settings. The sudden unleashing of RNA as a tool for making vaccines and more looks like a kindred phenomenon. “Basically all these designers are doing, whether they’re packaging mRNA in liposomes or packaging genes in an adenovirus, is they are typing on their computer,” says Dr Lemke. By changing the lines of code they type into their computers they can change the proteins which will, in a few weeks or months, be expressed in bodies from Spain to Peru to the Philippines.

Andrey Zarur, the boss of GreenLight, a biotech company based in Boston, is one of those seized by the possibilities. GreenLight intends to use production processes similar to those used by the vaccine-makers to produce a different sort of RNA molecule which can be used as a pesticide targeted against the Colorado Beetle. The project will make sense only if it can bring costs down dramatically. If it works, though, it will significantly increase the ability to send purposeful genetic messages straight to distant cells.

But GreenLight is also exploring the idea of an mRNA vaccine-production network that serves poorer regions directly, as are others. “It is plausible to consider building factories on different continents,” says Mr Sahin of BioNTech, “and thereby enable independence of the region from global supply.” If need be, such plants could be used to respond to regional changes in the virus’s make-up, making the vaccine most needed right there, right then. In the event that does not prove necessary, making mRNA vaccines closer to the point of use could still have other advantages, such as security of supply.

At the moment it is the adenovirus vaccines, not those based on mRNA, that are being produced at the greatest scale and lower cost, handily outcompeting the older techniques using inactivated SARS-CoV-2 which have been tried in China. With state-of-the-art cell cultures and well established supply chains, AstraZeneca expects 3bn doses of its vaccine to be produced this year. India’s Serum Institute alone plans to make 1bn doses. But cell cultures have economies of scale and require a lot of care and attention, while RNA manufacturing is in its infancy. Smallish plants requiring relatively little capital may prove possible as manufacturing technologies mature, allowing cheap, plentiful and local supply. Whether this will make mRNA vaccines the dominant ones within the next few years remains an open question. But whether it does or not, efforts towards that end will give the world an entirely new bioindustrial infrastructure, one capable of putting ideas into cells, both human and otherwise, more easily than ever before. ■