In the second grade, I learned how to play a game called Word Ladders. Our teacher told her thirty bug-eyed charges that we were going to turn a hen into a fox. In the befuddled silence that followed, she drew a ladder on the chalkboard. On the top rung, she wrote “HEN”; on the bottommost, she wrote “FOX.” On the intervening rungs, to a chorus of ahhhhs from the class, she wrote PEN, PET, SET, SIT, SIX, FIX. Each word differed from its predecessor by a single letter, which could be an addition, a deletion, or a substitution. In this way, it was possible to turn heat into cold, water into wine, a hen into a fox.

When the poliovirus mutates, its genome changes through a similar accretion of tweaks to its base pairs, though the alphabet in this case is limited to four letters – A, T, G, C – representing the four nucleotide bases. Vaccine designers have learned to take advantage of the virus’s changeability. In the 1950s, as thousands of children were succumbing each year to polio’s summertime scourge, Albert Sabin developed a weakened-virus vaccine by passaging poliovirus through cultures of monkey kidney cells. As the virus became inured to the monkey cells, it became maladapted to human ones, and harmless to humans: foxes into hens. Decades later, after a worldwide vaccination campaign brought the poliovirus to the brink of eradication, the WHO announced the emergence of vaccine-derived poliovirus. In one out of two million cases, the weakened vaccine virus could revert to an infectious strain capable of paralyzing the unvaccinated and immunocompromised: hens back into foxes.


Last April, Raul Andino, a virologist at the University of California in San Francisco, wanted to talk about cats. “If you see a picture of a cat in the air, you don’t know whether the cat is jumping or falling,” he told me, over Zoom. He froze with his arms akimbo. “Only a movie will tell you what’s going on.” In order to understand the evolution of the poliovirus, he explained, it is necessary to understand the minute adaptations the virus makes to survive in different environments – in the gut versus the blood, for instance – and also to track its change over the course of decades. The goal of a new polio vaccine, the first in over fifty years, would be to build an “evolutionary cage” around the vaccine virus, Andino told me. Only then would it be possible to tame the menagerie. “We do not want the virus to mutate.” He rapped his right wrist with his left hand, in mock reprimand. “No.”

Andino has devoted most of his forty-year career to studying poliovirus. He is sixty-three, with an impish mien and a penchant for analogy. He knew survivors of paralytic polio during his youth in Buenos Aires, and saw firsthand the evidence of its ravages: down-flexed feet, limp and withered legs. He remembers receiving the polio vaccine in the form of a sugarcube infused with a few pinkish droplets of liquid. (“It was the only vaccine that children wanted more of.”) His father, an oncologist who used to dye tissue samples in the bathroom sink at home, encouraged an easy and early rapport with science. A college biochemistry course kindled an enduring fascination with molecular biology, and Andino left Argentina for MIT in 1986 to work as a postdoctoral researcher with the Nobelist David Baltimore.

Polio seemed like a solved problem by the time Andino arrived in Boston. There were only eight cases of the disease in the U.S. that year, down from some fifty thousand at the epidemic’s peak in 1952. But developments in sequencing technology in the late 70s and 80s had lent certain old problems a sheen of novelty. Biologists began inspecting previously inscrutable genetic material at the base-pair level, and projects from those decades hint at the sweep and ambition of the genomics revolution to come: five years before Andino’s arrival, Baltimore published the full genome of the poliovirus; the Human Genome Project would launch in 1990, with the goal of sequencing the entire human genome by 2005.

Andino and other researchers were keenly interested in the evolutionary tendencies of the finicky poliovirus, which can spawn thousands of mutated copies of itself in a matter of hours. Around the time that Andino began working with David Baltimore, Cara Burns, the current head of the CDC’s polio surveillance division, was starting a PhD at the University of Utah, with the virologist Ellie Ehrenfeld. Burns, who is fifty-eight, wears her shoulder-length brown hair in a neat center part and radiates a quiet resolve. She considered a career in medicine before throwing herself headlong into academic virology. (She finally decided that she wanted to address infectious diseases at their source, she told me, before they could sicken humans.) Like Andino, Burns studied the poliovirus’s replication cycle, which she thought might yield further insight into the ways other viruses reproduce.

Though she worked mostly in the laboratory, Burns found herself drawn to the immediacy and urgency of public health research. Her first exposure to field work was a polio surveillance trip to India, in 1990, where she’d been sent to train WHO health workers in sample collection and the proper management of field labs. During a few hours in a poor neighborhood of Delhi, Burns observed that raw sewage – a well-known vector for polio – flowed in open drains, near street vendors hawking cooked food and produce. A thousand feet from city limits, farmers tended to rice paddies on the banks of the heavily polluted Yamuna River. Burns climbed to the top floor of an apartment complex and found herself among a small group of women and children who’d come to the roof for a respite from the building’s cramped interior. “You could see, across the city, the other rooftop families,” Burns told me. The entire scene – its frenetic energy and fragile ecology – stayed with her for a long time.

By the time Burns joined the Centers for Disease Control in 1998, as a microbiologist in the Center’s division monitoring polio outbreaks, the Pan American Health Organization had declared polio eradicated from the Americas. The virus was still circulating in about fifty countries, including India, Nigeria, Afghanistan, and Pakistan. Outbreaks were typically the result of poor vaccine coverage in areas that were difficult to reach for reasons of geography or political unrest. (The Kosi River Basin, in India, floods during the rainy season, cutting off villagers’ access to roads; parts of Nigeria controlled by Boko Haram remain significantly under-vaccinated.) Eradication seemed possible, though its goalposts had moved. The World Health Assembly had approved a resolution in 1988 committing the World Health Organization to eradicate poliovirus by 2000; in 2013, the Global Polio Eradication Initiative announced a new plan to eradicate polio by 2018. In a departure from the WHO’s usual alphabet soup of acronyms and initialisms, the convergence of resources and strategic planning required for the final push was called, simply, the “Endgame.”

The decades-long delay was unsurprising, even inevitable, to polio researchers – a disease that is asymptomatic in almost 95% of cases tends to go out with a whimper, not a bang. Andrew Macadam, a principal scientist at the National Institute for Biological Standards and Controls in Herts, England, had long suspected that viral reversion might have consequences for the eradication effort. Macadam is fifty-eight, round-faced, and reserved. He studied plant viruses as a graduate student at Imperial College London before joining the NIBSC as a postdoc with the polio researcher Philip Minor, in 1986. On a video call last May, he recalled his early interest in Albert Sabin’s vaccine. New sequencing tools had allowed researchers to “work out how vaccines were attenuated, the molecular basis of attenuation, why they were good vaccines,” Macadam told me. “But there was also this issue in the background of vaccine associated disease, which was very low frequency, but it was still a thing.” Some of the earliest cases of vaccine-derived poliovirus emerged soon after Macadam joined Minor’s group: thirty cases in Egypt, in 1988; twenty-three in Haiti and the Dominican Republic, between 2000 and 2001; twenty-nine in Nigeria, in 2004. These blips soon swelled into a current of vaccine-derived polio outbreaks, most of them in Africa and Asia. There were about three hundred fifty cases in 2019, and more than one thousand in 2020 – a pause in routine polio vaccinations, as a result of COVID, contributed to last year’s spike.

Andino views vaccine-derived polio as a cautionary tale for COVID vaccine coverage. In both cases, infectious variants have arisen as a significant percentage of the population has remained unvaccinated. “The virus is going to navigate this interface between immunized people and non-immunized people,” Andino told me. “It goes to the immunized people, and it gets challenged. Some mutation gets selected that allows the virus to replicate, but then it can jump back to the people that are not immunized.” While there is no risk of the COVID vaccine reverting to virulence – none of the three COVID vaccines approved for emergency use in the US uses whole coronavirus, in either a weakened or killed state – the coronavirus itself is likely to mutate under selective pressure. “The connections are very obvious to me,” Andino told me. “We have already seen that the coronavirus can improve the ability to jump from person to person in the Brazilian, South African, and UK strains.” Ultimately, a new vaccine is a “high-tech solution to a low-tech problem – not vaccinating,” Macadam told me. The best counter, in the case of both polio and COVID, is to vaccinate as many people as possible, and fast.


In the way of people who have been in the periphery of each other’s lives for a long time, the three researchers’ memories of their first meetings are murky. Their early encounters were the result of a kind of narrow serendipity: all three had, on at least several occasions, attended EUROPIC, a conference devoted to research on the family of viruses, called picornaviridae, to which poliovirus belongs. EUROPIC draws a small but dedicated group of attendees who convene, every two years, in a different European city to discuss picornaviruses and bar-hop between scheduled talks. Burns feels certain that she saw Andino at a bar during the 1989 EUROPIC meeting in Bruges. She might have met Macadam there, too, or at one of the other early meetings she attended, in Inari, Finland; Sitges, Spain; or Cape Cod, the single exception to EUROPIC’s European city rule. On a guest list from the 1994 edition of the meeting, held in Korpilampi, Finland, both Andino and Macadam’s names appear. (At some point, the two men discovered that they’d been born on the same day, in the same city – Buenos Aires – exactly five years apart.)

A rising awareness of vaccine-derived polio finally brought the three scientists together as collaborators, in 2010, with the Gates Foundation’s announcement of an initiative to fund the polio eradication effort. The Foundation convened a loose consortium of scientists to oversee the vaccine redesign. Burns’s group had already received a $100,000 Grand Challenges Explorations grant from the Foundation, in 2008, to pursue polio vaccine research. Macadam, acting in his capacity as a coordinator for the consortium (“It was a role that brings people together without being in charge as such,” Macadam told me. “It was a very egalitarian consortium”), reached out to Andino. There was a kickoff meeting the following year in New York City, at a Manhattan townhouse that had once belonged to FDR. “We kind of thrashed things out, talked about what we might do, and that’s really where the collaboration kicked off,” Macadam said.

The Foundation didn’t shy from the fact that its decision had been informed by economics. A press release from 2010, posted on the Gates Foundation’s website, summarized a study that found that polio eradication might save the global economy some fifty billion dollars in healthcare costs and “gains in productivity.” There is no explicit mention of the cost of the eradication program, which has been calculated, in separate reports, to be north of five billion dollars for the next five years. This cost is not evenly distributed across time or geography. Circulating cases are restricted to a few areas of the world, and the last days of the effort – the endgame – are some of the costliest, as countries commit to stringent surveillance and vaccination programs to root out lingering pockets of disease. Donald Henderson, the epidemiologist who led the smallpox eradication effort, was also one of polio eradication’s most vocal opponents. His argument against it evinced a kind of epidemiological trolley problem – funding for polio eradication could be usefully applied to other, more dire public health issues. Polio “doesn’t kill as many as measles,” he said in a 2011 Times article on the Gates Foundation initiative. “It’s not in the top twenty.” Instead of eradication, Henderson said, the WHO should focus on controlling low levels of the disease.

But even stalwarts may be swayed. Henderson had a change of heart after a meeting with Ciro de Quadros, the Brazilian epidemiologist in charge of the polio eradication effort in the Americas. In an article published just two weeks after the Gates Foundation story, he announced his reversal. “It’s not my wont to turn on a dime like this,” Henderson said. “I don’t think I’ve done anything like this before.” De Quadros had hinted at the possibility of an update to the vaccine itself – a further weakening of the vaccine strain that might prevent vaccine-derived polio cases. Henderson was optimistic: “Now the thinking and the muscle have changed.” When he died, five years later, his obituary in the Times made mention of his about-face. The “irascible but often prescient critic” of eradication campaigns had finally come around.

The vaccine update that had so enthralled Henderson was about to become a reality, with caveats. The new vaccine strain would be an engineered version of the old, but it couldn’t be much weaker than its predecessor. “I was coming from the perspective of, how does the Sabin vaccine work?” Macadam told me. “Why is it genetically unstable? And what can we do to make it more stable?” The existing vaccine was a good one, Macadam explained, so he sought to preserve its best aspects as much as possible. Weakening the vaccine strain further might decrease its immunogenicity – its ability to get a rise out of the body’s immune system. “The idea I came up with was to engineer the virus such that, if you had a particular kind of recombination that you didn’t want, it would kill the virus,” Macadam said. He called the mechanism a “kill switch.” Otherwise, the rationally-designed vaccine virus would be much the same as Sabin’s original.

The other researchers took a similar tack. Andino pursued a redesign that would attempt to freeze the virus’s evolution in time: an evolutionary cage. The poliovirus is an unreliable copycat of its own genetic data, Andino explained, and a “sloppy enzyme,” polymerase, is to blame. At each replication cycle, the poliovirus polymerase generates a “cloud” of copies bearing potentially beneficial mutations. Andino taught the polymerase to behave – to stop producing so many mutated versions of the vaccine virus. “It’s like educating the virus: Don’t do that,” Andino told me.

By 2015, Burns had hired a younger scientist, Jennifer Konopka-Anstadt, to contribute to the vaccine redesign effort. Fifteen years Burns’s junior, Anstadt belonged to a new generation of polio researchers, for whom the disease seemed as historical as its most famous patients: FDR, Frida Kahlo, Dinah Shore. She knew Andino and Macadam, elder statesmen in the field, from their Zelig-like appearances on the EUROPIC circuit. Friends and family were befuddled by Anstadt’s interest in the disease – and surprised that there were still circulating cases of it.

Burns and Anstadt applied a technique called codon deoptimization to hamper the virus’s replication. In a genome, a sequence of three consecutive nucleotide bases, called a codon, encodes a single amino acid; in some cases, an amino acid can be encoded by several different codons. The proteins that result from synonymous codons are identical, on the surface, but might take a longer or shorter time to produce. The virus, like a picky eater at dinner, will prefer certain codons and encode others less willingly. “We used to talk about the codon changes being kind of like spinach in your diet,” Burns told me. Burns and Anstadt engineered the vaccine strain with carefully selected synonymous codons. The result was a vaccine virus that looked identical to Sabin’s original strain, but was more difficult to replicate – this meant less virus in the bloodstream, and less virus shed into the environment.

Most new vaccines, including the COVID subunit vaccine, have been engineered with a similarly surgical precision. “In current articles that describe novel technologies, it is often said that they will enable “rational” development of vaccines,” writes the physician Stanley Plotkin, in his History of Vaccination. He continues, a bit tetchily: “The opposite of rational is irrational, but presumably the writers mean to contrast rational with ‘empiric.’” Plotkin, as a co-developer of vaccines for rabies and rubella in the 1960s, wrote from experience. A mortar and pestle, as much as a microscope, were tools of vaccine design in those earlier decades. Vaccine developers concocted strange brews of tissues and organs, in the hope that viruses might become less dangerous to humans as they adapted to the cellular ecologies of other animals: rhesus monkeys, macaques, mice, chickens, rabbits. During the 1940s, Hilary Koprowski, a Polish-born virologist at Lederle Laboratories in Pearl River, New York, vied with Albert Sabin to design the first weakened-virus polio vaccine. Using the serial passage method, Koprowski injected the poliovirus into the brain of an animal that was susceptible to infection: a small, dark-furred rodent called a cotton rat. He blended the animals’ brains into a gray sludge, which he and a willing colleague drank to test for toxicity. The two men suffered no ill effects. (Koprowski later remarked that the macerated brains tasted like cod liver oil.)

Vaccine development, like prospecting for gold, required sound reason and good luck in equal measure. Koprowski, Sabin, and Plotkin knew a good vaccine once they’d administered it to a test subject – or tried it themselves. They could experimentally improve the odds of coming by such a vaccine, but they couldn’t know, a priori, what adjustments to make to the viral genome to weaken it or keep it from going rogue. Today, it is possible to identify such adjustments with dizzying precision: in the eight-thousand-letter sequence of nucleotides that form the poliovirus genome, the error that Macadam calls the “gatekeeper mutation” occurs at the 481st position. Like a rung in Word Ladders, it is a single-letter substitution – an A in place of a G – that seems to precede all other mutations.

Milestones in the eradication effort came and went as the polio researchers worked. In 2014, outbreaks of polio in Africa and the Middle East were designated a Public Health Emergency of International Concern by the WHO. A PHEIC had been declared only once before, in 2009, for the swine flu. (Since the polio declaration, there have been four other PHEICs: two for Ebola, one for the Zika virus, and the most recent for SARS-CoV-2.) In 2016, the Americas celebrated 25 years without a case. 2017 was the first year that vaccine-derived cases outstripped non-vaccine-derived cases, which seemed more a testament to the efficacy of the Sabin vaccine than to the virulence of the vaccine-derived strain. Africa was declared free of non-vaccine-derived polio in August of 2020, leaving circulating cases in only Afghanistan and Pakistan. Africa’s all-clear might have been marquee public health news in any other year. As it was, it was quickly overridden by reports of burgeoning COVID case numbers.

The PHEIC was a bit of bureaucratic rubber-stamping for a reality that the polio researchers had long ago recognized. “The actual designation didn’t change how hard and how fast we were working,” Anstadt told me. “When you are marrying surveillance with vaccine design and research, there’s almost no period where anyone’s resting. If things are going well with surveillance, that gives you more time to do the research work.” Burns’s group could be called on at a moment’s notice to identify a new outbreak of polio. On a few occasions, scientists worked late into the night and tag-teamed with another researcher, who’d arrive in the early morning, to report on the origin of an outbreak within eighteen hours of receiving a sample.

The three groups of researchers had agreed to independently test their approaches to the vaccine redesign – to see if they “had wings,” Macadam told me. Even so, they remained in close communication. Information about edits to the viral genome were sent electronically and replicated in the laboratory; on occasion, samples were packed in dry ice and airmailed. All three groups successfully tested their redesigns on lab mice. The vaccine design consortium decided on two candidate vaccines for clinical trials in humans. The first combined Andino’s engineered polymerase and Macadam’s kill switch; the second combined the kill switch with Burns and Anstadt’s codon-deoptimized design.


For a long time, vaccine trials were a relatively unregulated process. In the 1950s, it wasn’t uncommon for vaccinologists to test on themselves as an ultimate proof of concept, and something like a vote of confidence. Koprowski downed a dose of his own rat-brain smoothie. Both Salk and Sabin tested their polio vaccines on themselves and their families – Salk injected himself, his wife, and their three sons with needles he’d sterilized in a pot of boiling water on the kitchen stove.

In 2017, though, there was Poliopolis. The temporary testing facility was constructed over the course of three days in a parking lot at the University of Antwerp, in Belgium, for the new polio vaccine’s Phase I trial. Sixty-six modular containers formed two rows of small, sterile-looking bedrooms, as well as bathrooms, a fitness room, a kitchen and dining area, showers, a rec room, and laboratory spaces. In the alley between both rows was a parody of a suburban backyard: hammocks, lawn chairs, artificial turf, a tiny barbecue grill. For up to four weeks between May and August, two groups of fifteen adult volunteers lived in Poliopolis, one after the other. Each group received one of the two candidate polio vaccines. Scientists from the university observed the volunteers for polio symptoms and checked their stool for viral shedding. The volunteers were not permitted to set foot outside Poliopolis for the duration of the trial, and news reports of the “container village” took on a lightly dystopian tenor. (‘’The biggest challenge? Keeping everyone in,” read a headline in De Standaard.)

The first human trials at Poliopolis yielded promising results, as did Phase II and III trials conducted in Bangladesh and Panama. Last November, as COVID cases were cresting their second peak, the European Union approved one of the two redesigned strains – with Andino’s engineered polymerase and Macadam’s kill switch – for an emergency use listing. (The codon deoptimization that had made Burns and Anstadt’s redesign so effective also made it harder to produce at scale.) It was the first vaccine to undergo the EUL process, and was followed, a month later, by Pfizer’s COVID vaccine. Since then, 120 million doses of the new polio vaccine have been manufactured by an Indonesian company called Bio Farma, and more than forty million doses of it have been administered in Niger, Nigeria, Liberia, the Congo, Tajikistan, Benin, and Sierra Leone. So far, there have been no signs of viral reversion.

Last June, the GPEI announced a new five-year plan for polio eradication, with the new polio vaccine as its cornerstone. The budget for the new plan would allot about one billion dollars a year to eradication, with some three hundred million dollars dedicated to the vaccine-derived polio response in the first year. By 2026, according to the GPEI’s plan, preparations should begin for a polio-free world: reduction of laboratory stocks of the virus, withdrawal of the weakened-virus vaccine from use, continued surveillance for possible outbreaks.

Since last March, CDC employees at its Atlanta headquarters have been working remotely. Because disease surveillance is essential work, some members of Burns’s group have been visiting the lab on a rotating schedule. Those scientists go to the lab once or twice a week to work alone, with only the pneumatic hum of fume hoods for company. Both Anstadt and Burns set up workstations at their dining room tables. Viral phylogenies and case statistics blended with their home lives; on occasion, Anstadt would leave her computer and return, moments later, to find scribbles from her three-year-old daughter in her notebooks. When we spoke, Burns had recently finished reading An Hour Before Daylight, Jimmy Carter’s account of his Depression-era childhood in rural Georgia. The former president’s foundation, the Carter Center, is leading the other of two eradication programs approved by the World Health Assembly, against the guinea worm. After polio, after guinea worm, Burns thinks that measles will be the next to go.

At the NIBSC, there are plans to make new vaccines, employing new techniques, including something called a virus-like-particle, which mimics the appearance of a virus but is not itself infectious. It seems unlikely that there will be another polio vaccine. “If we would start again, now, there would be little things we might have done differently. But the fact of the matter is, this was our one chance,” Macadam told me. “So even if you can think, well, if I change that nucleotide, it would make it even better – it’s too late to do that.” Andino, waxing expansive at the end of our conversation, told me that an eradication program should seek to eliminate a disease, not its causative agent; on the whole, he told me, viruses are good. “Maybe, in the same way that you eat your yogurt every morning to restore your microbiome, you might have to drink a little bit of virus,” Andino said. A sly smile crept across his face. “Like a viral vitamin.”