Two papers published this week, and one last month, reveal the pandemic potential of H5N1 “bird flu”. One identifies four, another identifies five, genetic changes the virus would have to undergo before it could spread easily in humans, and the third paper suggests some of these changes are already evident in circulating strains.

The papers were written last year, but were held back because of international concerns that making such data public would make it easier for terrorists to make bioweapons.

Update, June 7th, 2013 – Researchers from MIT reported in the journal Cell (June 2013 issue) that the H5N1 and H7N9 bird flu virus strains need just one or a few genetic mutations to become easily human transmissible, which would raise the likelihood of there being a flu pandemic.

H5N1, or “bird flu“, is a subtype of the influenza A virus that can cause illness in humans and many other animals. It has killed tens of millions of birds and spurred the culling of hundreds of millions of others to stop it spreading.

Since 2003, more than half of the 606 cases of human infection of H5N1 reported to the World Health Organization worldwide, have resulted in death. Most of the cases in humans appear to have occurred as a result of contact with infected birds.

So far, H5N1 has not triggered a pandemic in humans because it does not spread easily among mammals, and some scientists believe it never will. To spread easily from one person to another, the virus would have to become airborne, that is develop the ability to spread via tiny droplets that people spray out of their mouths and noses when they cough and sneeze. That is how other flu viruses, like the H1N1, the “swine flu”, that caused a mild pandemic in 2009, spread.

Viruses like H5N1 and H1N1 are mutating all the time. If H5N1 were by chance to acquire some of the properties of H1N1 then it would spread more easily in mammals. One way it could do this is by accumulating chance mutations, another way is by swapping genes with other viruses, for instance while co-infecting an intermediate host (genetic “reassortment”).

Much of the focus of H5N1 research, amid concerns about the bioterrorism risk, has been to investigate how easy it might be for H5N1 to mutate into a readily transmissible form, and if so, which genes would be involved. This information is useful for surveillance, so researchers know what changes to look out for in emerging strains when assessing pandemic risk.

The papers published this week and last month show that it would take about four or five genetic changes for H5N1 to mutate into a form that might trigger a pandemic in humans, and that some of these mutations have already occurred in nature.

In Science this week, one paper describes how the virologist Ron Fouchier of Erasmus MC in Rotterdam, the Netherlands, and colleagues, compared the genetic structure of H5N1 with strains of flu that have caused human pandemics and identified some candidate genes.

Fouchier’s team inserted the candidate genes one by one into an actual strain of H5N1 that was isolated from an Indonesian patient, and tested the artificially mutated strains in ferrets to see how easily they would spread (ferrets are a popular model for human flu research). They conclude that H5N1 would need just five genetic changes to transform it into a form that could spread easily and start a human pandemic.

(Actually, the “conclusion” of five mutations is not strictly accurate, it is a media shorthand that gives a reasonable gist of a more complicated story. Fouchier’s team tested a number of strains and found up to nine mutations in each transmissible strain were relevant, but they all had five mutations in common, so that is how the number five came about. But there is still the possibility that some of the other mutations could play a role.)

Fouchier and colleagues also conclude that the virus could acquire the capacity for airborne transmission between mammals without having to swap genes with another flu virus while co-infecting an intermediate host (that is by accumulation within the strain, without the help of genetic reassortment).

In Nature last month, another paper, from a group led by Yoshihiro Kawaoka of the University of Wisconsin, Madison in the US, and the University of Tokyo in Japan, using a different approach to Fouchier’s team, describes how it would take only four genetic changes, to transform the bird flu virus into an airborne strain.

Instead of taking an actual strain of H5N1 and then inserting genes one by one as the Dutch team did, Kawaoka’s team created a hybrid strain of the bird flu virus and the already airborne-capable H1N1 “swine flu” virus. (In theory, such a strain could come about via “reassortment” or gene-swapping between viruses).

Then they showed, also using ferrets, that the hybrid, thanks to four delicately balanced mutations, would be able to bind more strongly to cells in mammals and replicate in sufficient quantity to be able to spread via respiratory droplets.

In the other Science paper published this week, researchers (including Kawaoka and Fouchier) led by mathematician Derek Smith of the University of Cambridge in the UK, explain how they investigated the likelihood of the emergence of a pandemic strain of H5N1 and how using surveillance data, they looked for evidence of whether some of the mutations described in the Kawaoka and Fouchier papers have already occurred in nature.

They found that several wild strains of H5N1 are three mutations away from the four described in Kawaoka’s paper, and the five in Fouchier’s paper. And they also found a few rare cases where the strain is only two mutations away.

Using some clever maths that they developed themselves, Smith and colleagues then created a model to simulate “within-host viral evolution” to study the factors that would increase or decrease the probability that a virus with some of the mutations could evolve the remaining ones after it had infected a mammalian host.

First author Colin Russell of the University of Cambridge, said in a comment reported in a Science News and Analysis article that a virus that is only three mutations away from acquiring the ability to be airborne is “likely” to do so, but there are so many unknowns they couldn’t put a figure on it. The paper’s main value is that it points others in the right direction:

“These factors, combined with the presence of some of these substitutions in circulating strains, make a virus evolving in nature a potentially serious threat. These results highlight critical areas in which more data are needed for assessing, and potentially averting, this threat,” write the authors.

These papers have appeared after months of international debate about whether this information should be made public, and whether the researchers should have done the experiments in the first place.

In December last year, the US National Science Advisory Board for Biosecurity (NSABB) recommended that the Kawaoka and Fouchier studies, because they reveal so much detail about how the mutations could be introduced, should not be published in full. The concern was that the information could be used by terrorists to create a bioweapon, and this risk outweighed the benefits to public health.

Then, a few months later, after examining revised versions of the papers, the NSABB changed its position, and recommended to the US government that the papers be published in full.

In a statement issued at the end of March, the board said although the papers still contain information that could be useful to those who would use it to do harm, the new information they received caused them to change their risk-benefit calculation such that the benefits now outweight the risks.

An expert panel advising the World Health Organization (WHO) had reached a similar conclusion a month earlier.

This cleared the way for Nature to publish the Kawaoka paper last month, and for Science to publish the Fouchier study this month.

In the meantime, while all this was going on, the influenza scientists responded by imposing a moratorium on themselves, effectively halting some types of H5N1 research, and the US government brought in new legislation to control government-funded studies using dangerous pathogens.

After the initial NSABB decision, the Dutch government also sought to restrict publication of Fouchier’s paper by insisting he apply for an export licence, which he eventually did and received.

Fouchier said this week, that publishing the work in full gives scientists the best possible chance of fighting future flu pandemics.

“We hope to learn which viruses can cause pandemics and by knowing that we might be able to prevent them by enforcing strict eradication programmes,” he told BBC News.

The activities surrounding the publication of these papers have also served to focus attention on what needs to be done to deal more effectively with research that could be misused, the so-called “dual use research of concern” or DURC.

Dr Bruce Alberts, Science’s Editor in Chief, said in a statement reported by the BBC that it looks like we need to develop a “comprehensive, international system for assessing DURC”. He said such a system should allow people with a need to know to have “selected access” to information that is omitted from a published scientific report.

But Fouchier said he doubts such a system is even workable, never mind appropriate.

“You can’t share information with so many people in the field and keep it confidential,” he told the BBC.

Fouchier said the general rule should be to make research information freely available, so scientists can build on it.

Written by Catharine Paddock