Here, for your viewing pleasure, is a very important part of a very special flu virus. It may look like an ordinary protein, but in fact it’s been at the center of a blazing debate about whether our increasing power to experiment on life could lead to a disaster. Not that long ago, in fact, a national security advisory board didn’t even want you to see this. So feast your eyes.
For those who are new to this story let me start back at the beginning, in 1997.
In that year, a child in Hong Kong died of the flu. Doctors shipped a sample of his blood to virus experts in Europe, but they didn’t bother taking a look at it for months. When they did, they were startled to discover that it was unlike any flu they’d seen in a human being before.
Each year, several different flu strains circulate from person to person around the world. They’re known by the initials of the proteins that cover their surface–H3N2, for example, is one common strain. The H stands for haemagglutinin, a protein that latches to a host cell so that the virus can invade. The N stands for neuraminidase, which newly produced viruses then use to hack their way out of the cell.
Birds are the source of all our flu strains. Our feathered friends are hosts to a huge variety of H and N type viruses, which typically infect their guts and cause a mild infection. From time to time, bird flu viruses have crossed the species barrier and adapted to human hosts, infecting our airways and then spreading in air droplets. Since flu spreads so fast around the world, a fair amount of the planet’s population has had some exposure–and thus some immunity–to the flu strains in circulation today. But if a new bird flu should manage to make the leap, we could face a very grim situation–a situation that some scientists worry could rival the 1918 pandemic, which killed some 50 million people.
That’s why the scientists in 1997 were so flustered. The Hong Kong boy had died of a strain of bird flu that hadn’t been found in people before. It came to be known as H5N1.
It turned out that around Hong Kong, chickens were rife with H5N1, including the chickens for sale in live open-air markets. Public health workers slaughtered huge numbers of chickens to stop the outbreak, and, for a time, it seemed like they had beaten the virus. In fact, H5N1 had simply gone into hiding. A few years later it was back–and spreading. Birds carried it across Asia, into Africa and Europe. The New World and Australia have been spared so far, but there’s no reason to think that the virus can’t colonize those continents as well. It will just take the right bird.
Doctors found that the majority of patients hospitalized with H5N1 died. The only comforting thing about H5N1 was that it remained a bird flu. Once inside a human being, the virus couldn’t churn out lots of new viruses capable of spreading to another human. But many bird flu experts consider that a cold comfort. Like all flu viruses, H5N1 has been continually evolving. When the viruses replicate they pick up new mutations–some of which help them replicate faster. Sometimes, two H5N1 viruses co-infect a single cell at once and swap some of their genes, producing hybrids. If this high-speed evolution leads to human-adapted H5N1, we could be dealing with a global cataclysm.
Yet some flu experts doubted this grim prospect. It’s been some 15 years since H5N1 was first discovered, and despite all those years of evolution, the virus has yet to nose its way into our species. Perhaps, some scientists suggested, there was something about the biology of the strain that prevents natural selection from transforming it into a human virus. Skeptics have more recently raised another question about the risk of H5N1: is its mortality rate really all that high? In many studies, scientists have estimated the mortality rate of H5N1 based only on sick people who come to hospitals. It’s possible that a lot of people recover from bird flu infections on their own, and go missing from the statistics. (It’s worth bearing in mind, though, that the 1918 flu only had a mortality rate of 2%. If a virus can infect billions of people, even a low rate like that can lead to terrifying numbers of deaths.)
A few years ago, some bird flu experts decided to test the proposition that H5N1 was a potential human scourge. They would tinker with it to see if it could be transmitted from mammal to mammal, instead of bird to mammal. They might be able to see some warning signs for how this transition could happen in nature. The scientists applied for money from the National Institutes of Health, which considered their idea important enough to sink millions of dollars into it.
Two teams of scientists–one in the Netherlands and the other at the University of Wisconsin–got good results. They could infect ferrets with modified H5N1, and the ferrets could cough up droplets that could infect healthy ferrets. They submitted their findings to the world’s biggest scientific journals, Science and Nature, to let the world know of their discovery. One of the Dutch researchers, Ron Fouchier, referred cryptically to the research at a flu meeting in Malta in September.
It occurred to persons unknown that publishing these results might not be such a great idea. What if individuals bent on destruction got the idea of unleashing a pandemic. Maybe they could use the scientific papers–in particular the description of the experimental methods–to replicate the results. Just knowing the mutations might be enough information for a talented virologist to tweak a bird flu virus into a biological weapon.
The difficult question made its way to the National Science Advisory Board for Biosecurity, a group of scientists, lawyers, and other individuals who meet to discuss how scientists can reduce the risk of such nightmare scenarios. In October, they were asked by the Department of Health and Human Services to consider whether the results of the new studies should be published in full.
Meanwhile, the details from Fouchier’s talk dribbled out into newspapers and magazines.According to many of the articles, Fouchier’s virus didn’t just spread from ferret to ferret: it was also quite deadly. Journalists got wind of the Wisconsin paper, but the Wisconsin researchers, by contrast, were pretty much mum throughout the past few months.
As the controversy heated up, the members of the NSABB remained almost totally silent until they issued a decision in December: the two papers should not include the complete methods and mutations. If scientists needed that information to help them with flu surveillance, they could ask the journals for access to the information.
A battle then broke out–one rarely witnessed in science–between scientists who supported the decision, and those who favored publication. Some scientists wondered why on Earth NIH every approved the research in the first place. Others argued that worrying about evil-doers was a weak justification for stifling scientific freedom.
The tide turned in February–for reasons that have not yet been clearly accounted for. It came to light that Fouchier’s ferrets only keeled over if the viruses were rammed down their trachea. When the ferrets came down with the flu from airborne droplets, they got sick but didn’t die. The NSABB met again in March to reconsider the papers, and voted unanimously to recommend the Wisconsin paper be published in full. The Dutch paper also won approval, although only by a vote of 12 to 6.
Which brings us to today, and the publication of the Wisconsin team’s paper in Nature. Here’s what they did:
The scientists wondered if one reason that H5N1 does a lousy job of spreading in humans is that it struggles to invade human cells. So they introduced mutations into the gene for haemagglutinin, producing a vast collection of mutant H5N1. They then offered the viruses a chance to infect some cells. For their experiment, the Wisconsin scientists used turkey blood cells, which they had studded with the human surface protein that flu viruses attach to. They washed away all the viruses except for the ones that were anchored to the cells.
Through this process, the scientists discovered one particularly human-loving form of H5N1 haemagglutinin. The gene for the protein contained four new mutations. (The figure above illustrates the protein, with the four mutations labeled. The yellow ball is where the protein grabs onto a human cell.) Three of the mutations altered the shape of the protein, while the fourth had a subtler effect. It changed the pH level at which the protein fuses to the cell and allows the genetic material inside the virus to enter the cell. The fact that no one could predict any of these details in advance demonstrates how little we still know about the flu.
Now that the scientists had a gene for human-loving H5N1 haemagglutinin, they inserted it into a human virus. They chose the so-called swine flu strain that burst on the scene in 2009 and then, mercifully, proved to be relatively mild as flu viruses go. (This strain is known as H1N1.) With a little engineering, they produced a swine flu virus studded with the mutant H5N1 haemagglutinin proteins.
The scientists injected this hybrid virus into the noses of ferrets, and it replicated nicely inside them. In later experiments, the scientists found that the virus could spread from one ferret to another. None of the ferrets died, however.
The Wisconsin experiment was artificial, of course, but that doesn’t mean it doesn’t resemble what happens in nature. In an accompanying commentary in Nature, Hui-Ling Yen and Joseph Sriyal Malik Peiris of the University of Hong Kong point out that when scientists mix H5N1 and H1N1 viruses together in a flask, they will readily produce hybrids on their own. What’s more, pigs can get infected with H5N1 and H1N1 at the same time, so they could well be acting as mixing vessels for the viruses.
A hybrid H5N1/H1N1 virus–like any new human flu virus–is a worrisome prospect. But the fact remains that, at least in ferrets, it’s not all that scary. The scarier prospect is that H5N1 bird flu might be able to evolve the same four mutations engineered by the Wisconsin team in their experiment. It’s possible that such a virus would be able to slip easily into humans, but have the same lethality as today’s bird-adapted H5N1. No one can say whether such a virus would behave this way, however. Perhaps the four mutations engineered by the Wisconsin team only works successfully in the company of genes from the 2009 H1N1 virus. There’s only one way to find out: directly engineer those four mutations into bird flu and see what happens to the ferrets.
That appears to be what Fouchier and his Dutch colleagues essentially did. Based on presentations he’s given and on news articles, it seems that the Dutch scientists introduced mutations into H5N1 bird flu and then carried out an evolutionary experiment. They injected the viruses into ferrets, let them replicate, and then drew out newly replicated viruses to administer to another ferret. The viruses mutated inside the ferrets. Some of the mutations sped up their replication, and those mutant viruses came to dominate the population. They were thus more likely to be picked out by the scientists. Eventually, the scientists had evolved viruses that could float from one ferret to another. They then took a look at the genes of the evolved viruses to see which mutations had turned them into mammal flu.
On Tuesday, I participated in a workshop at the National Academies of Science where scientists and public policy experts (and a couple journalists) reflected on the flu controversy. One of the organizers was David Relman, a Stanford microbiologist who serves on the NSABB. During a break, Relman explained to me that he had voted in favor of publishing the Wisconsin paper and against the Dutch paper. For him, it was a matter of whether the risks incurred by making the research public were just too high. The Wisconsin paper came close to the line but didn’t cross it, because the virus contained only one H5N1 protein. If people want to do some evil with the information in the paper, they’ll have to engineer an H5N1 virus with the four-mutation haemagglutinin. And they’d have no guarantee that it would cause a human pandemic.
Relman felt that the Dutch experiment, on the other hand, crossed the line. The Dutch scientists used bird flu from start to finish, and so the final genome they ended up with may be an attractive starting point for malefactors. The fact that the Dutch flu turns out not to be deadly to ferrets doesn’t make Relman any less worried. What matters is that it is bird flu that can now move on its own between mammals. If the Dutch virus were to start infecting humans, it could spread quickly. Along the way, it might be able to evolve into a deadlier form .
Despite Relman’s misgivings, it looks like the Dutch paper is going to be published before too long. After the NSABB reversed its decision in March the Dutch government threw one last obstacle in its path to publication. It decided that the paper required an export license–just like a computer being shipped to China might. Fouchier was granted the licence on April 27, so now the paper is with Science, and soon we will all be able to see its gory details.
That publication won’t mark the end of this story, I think. At the NAS workshop, the signs were clear that this episode is just the start of something much bigger. Roger Brent, a biologist at the Fred Hutchinson Cancer Research Center, put it into history of modern biology. In the 1970s, biologists discovered how to move individual genes from one organism to another. The power to rewrite the book of life caused a lot of consternation, and led to a gathering called Asilomar in 1975, where scientists tried to work out a system for ensuring that no monstrous new creatures would escape a lab and wreak havoc on the world. At the time, just about everyone on Earth who had the wherewithal to perform genetic engineering could get together at Asilomar. Over the past 37 years, these manipulations have become democratized. A far broader group of researchers now have far more power than anyone did in 1975.
Brent, like a number of scientists at the meeting, thought the experiments were a mistake. Brent objected to them becaused they were designed to bring something dangerous into the world that did not exist before. As long as those viruses exist, he argued, there will have to be armed guards making sure that no one steals them. Brent regretted that before the experiments were ready to be published, no scientists spoke out to condemn the risks they carried, or to question whether we might be better off today if scientists had focused more on surveillance or vaccines. It was, Brent declared, “an ethical failure.”
Some of the other scientists who spoke after Brent disagreed with that assessment. “I believe these studies were necessary to alert the world to this possibility,” said Robert Webster, a bird flu expert at St. Jude’s Children’s Research Hospital. Later in the day, Anthony Fauci, the director of the National Institute of Allergy and Infectious Diseases, said that the NSABB’s decision to let the papers be published was a good one. “I personally agree with the decision.”
After both Webster and Fauci spoke, a snowy-haired Englishman walked up to a microphone set up in an aisle for questions. This was Richard Roberts, the chief scientific officer of New England Labs and the winner of the 1993 Nobel Prize in medicine for his discovery that genes are sometimes split into separate segments of DNA. Roberts went after both of the scientists with an understated fury.
He asked how it could be that scientists could tamper with the flu this way, when they could be doing unquestionably valuable research such as developing a universal flu vaccine. “I find it indefensible,” he declared to Webster.
To Fauci, Roberts suggested that many scientists might be afraid to express their objections to the flu research because the leaders of NIH–such as Fauci and the NIH director Francis Collins–have come out so forcefully in support of it. Roberts suggested that some researchers might think twice about questioning “the official NIH line” for fear “their grants might be threatened.”
Fauci replied by calling that idea “preposterous.”
Fauci went on to explain how the NIH has now put in place a formal policy to consider the risks of this sort of research before funding it. In talking with a number of scientists at the meeting, I got all sorts of possible experiments that might fall under this category. Could HIV, for example, evolve to become airborne like flu is? Scientists might find compelling reasons to manipulate smallpox or Ebola. These are all possible, but no one can say how likely it will be that we face a similar controversy over them.
On the other hand, there’s little doubt that scientists will be wanting to do more research on the flu of this sort. Just because the Wisconsin and Dutch researchers found that a few mutations can turn bird flu into mammal flu doesn’t mean that those mutations alone hold the key to the next pandemic. There may be may different ways to cross the species barrier. It will take many replications and variations on these experiments to get to a broad understanding of H5N1′s capacity to make that jump. Whether we choose to gain that knowledge remains to be seen. Fauci promised that the decision-making about these experiments will be transparent and that we in the press will be able to follow it more closely. That’s the kind of knowledge that we journalists definitely like to have out in the open. You can bet we’ll be holding him to his word.
For more information:
[Update 5/2: fixed some of the NSABB chronology--thanks to Brendan Maher of Nature for helping me with the timeline]