Jun 21, 2012 (CIDRAP News) – A study showing that it takes as few as five mutations to turn the H5N1 avian influenza virus into an airborne spreader in mammals—and that launched a historic debate on scientific accountability and transparency—was released today in Science, spilling the full experimental details that many experts had sought to suppress out of concern that publishing them could lead to the unleashing of a dangerous virus.
In the lengthy report, Ron Fouchier, PhD, of Erasmus Medical Center in the Netherlands and colleagues describe how they used a combination of genetic engineering and serial infection of ferrets to create a mutant H5N1 virus that can spread among ferrets without direct contact.
They say their findings show that H5N1 viruses have the potential to evolve in mammals to gain airborne transmissibility, without having to mix with other flu viruses in intermediate hosts such as pigs, and thus pose a risk of launching a pandemic.
The Fouchier study is paired with another in which an international team used mathematical modeling to reach the conclusion that, given how few mutations it took to yield an airborne transmissible virus, such a virus probably could evolve in nature, though it's not possible to quantify the risk. They noted that two of the five mutations in the airborne virus already are common in natural H5N1viruses.
"We now know that we're living on a fault line," said Derek J. Smith, PhD, of the University of Cambridge, senior author of the modeling study, at a Science press conference yesterday. "It could do something. What we need to know is how likely is that."
Second of two controversial studies
Fouchier's study is one of two H5N1 transmissibility experiments that sparked a fierce controversy beginning late last fall, and is the second to see print. The first study, led by Yoshihiro Kawaoka, DVM, PhD, of the University of Wisconsin and the University of Tokyo, was published in Nature in early May.
Kawaoka's team used a somewhat different approach to achieve the same goal as Fouchier. Kawaoka and colleagues introduced mutations in the hemagglutinin (HA) gene from an H5N1 virus and then combined it with seven genes from a 2009 H1N1 virus, creating a hybrid virus that was able to spread by air between ferrets. They said the study demonstrated that a flu virus with H5 HA can spread among mammals—something that natural H5N1 viruses have done extremely rarely, if at all.
When word of the Fouchier and Kawaoka studies got out last fall, it sparked concern about the potential for creation and intentional or unintentional release of a pandemic virus. This led to a review by the US National Science Advisory Board for Biosecurity (NSABB), which advises the Department of Health and Human Services.
After long deliberations, the board recommended in December that both studies be stripped of key details before publication, but that the details be shared with selected scientists and officials on a need-to-know basis.
But after the authors—especially Fouchier—provided some additional information and clarifications, the board recommended on Mar 30 that the full versions of both studies be published, and federal health officials endorsed the recommendation. The NSABB voted unanimously in favor of publication for Kawaoka's paper, but it split 12-6 on Fouchier's.
The heated controversy over publication of the studies prompted a group of 39 leading flu researchers, including Fouchier and Kawaoka, to impose a voluntary moratorium on research designed to increase the transmissibility of H5N1 viruses in mammals. The moratorium remains in effect, awaiting the further development of government policies on biosafety and biosecurity for such research (see related CIDRAP News story).
The controversy also prompted the US government in March to outline a new general policy on the handling of "dual use research of concern," or DURC, meaning research that could be used for either good or ill. US officials said this week that further details on that policy will be released soon in the form of draft guidance for institutional biosafety committees.
In a question-and-answer sheet released with Fouchier's study, Erasmus officials said the paper was revised to better explain the research goals and the benefits to public health, and "display items" were added to explain the virulence and transmission properties of the airborne virus, but the methods and results were not changed. Also, the manuscript was revised to improve clarity, since it was expected to draw a broader readership than such reports usually do.
As for safety, the experiments were conducted in an Animal Biosafety Level 3 (ABSL3+) lab at Erasmus, the report says. Inspectors from the Dutch government and the US Centers for Disease Control and Prevention approved the facilities and procedures before and during the research. Researchers wore protective equipment and were offered seasonal and H5N1 flu vaccines.
Indonesian H5N1 strain used
Fouchier's team started with an H5N1 virus collected in Indonesia and used reverse genetics to introduce mutations that have been shown in previous research to make H5N1 viruses more human-like in how they bind to airway cells or in other ways. Avian flu viruses prefer to bind to alpha2,3-linked sialic acid receptors on cells, whereas human flu viruses prefer alpha2,6-linked receptors. In both humans and ferrets, alpha2,6 receptors are predominant in the upper respiratory tract, while alpha 2,6 receptors are found mainly in the lower respiratory tract.
The amino acid changes the team chose included N182K, Q222L, and G224S, the numbers referring to positions in the virus's HA protein, the viral surface molecule that attaches to host cells. Q222L and G224S together change the binding preference of H2 and H3 subtype flu viruses, changes that contributed to the 1957 and 1968 flu pandemics, according to the report. And N182K was found in a human H5N1 case.
The scientists created three mutant H5N1 virus strains to launch their experiment: one containing N182K, one with Q222L and G2242, and one with all three changes, the report explains. They then launched their lengthy series of ferret experiments by inoculating groups of six ferrets with one of these three mutants or the wild-type H5N1 virus. Analysis of samples during the 7-day experiment showed that ferrets infected with the wild-type virus shed far more virus than those infected with the mutants.
In a second step, the team used a mutation in a different viral gene, PB2, the polymerase complex protein. The mutation E627K in PB2 is linked to the acquisition by avian flu viruses of the ability to grow in the human respiratory tract, which is cooler than the intestinal tract of birds, where the viruses usually reside, according to the report.
The researchers found that this mutation, when added to two of the HA mutations (Q224L and G224S), did not produce a virus that grew more vigorously in ferrets, and the virus did not spread through the air from infected ferrets to uninfected ones.
The passaging step
Seeing that the this mutant failed to achieve airborne transmission, the researchers decided to "passage" this strain through a series of ferrets in an effort to force it to adapt to the mammalian respiratory tract—the move that Fouchier called "really, really stupid," according to a report of his initial description of the research at a European meeting last September.
They inoculated one ferret with the three-mutation strain and another with the wild-type virus and took daily samples until they euthanized the animals on day 4 and took tissue samples (nasal turbinates and lungs). Material from the tissue samples was then used to inoculate another pair of ferrets, and this step was carried out six times. For the last four passages, the scientists used nasal-wash samples instead of tissue samples, in an effort to harvest viruses that were secreted from the upper respiratory tract.
The amount of mutant virus found in the nasal turbinate and nose swab samples increased with the number of passages, signaling that the virus was increasing its capacity to grow in the ferret upper airway. In contrast, viral titers in the samples from ferrets infected with the wild-type virus stayed the same.
The next step was to test whether the viruses produced through passaging could achieve airborne transmission. Four ferrets were inoculated with samples of the "passage-10" mutant virus, and two ferrets were inoculated with the passage-10 wild strain. Uninfected ferrets were placed in cages next to the infected ones but not close enough for direct contact.
The ferrets exposed to those with the wild virus remained uninfected, but three of the four ferrets placed near those harboring the mutant virus did get infected, the researchers found. Further, they took a sample from one of the "recipient" ferrets and used it to inoculate another ferret, which then transmitted the virus to two more ferrets that were placed near it.
Thus, a total of six ferrets became infected with the mutant virus via airborne transmission. However, the level of viral shedding indicated the airborne virus didn't transmit as efficiently as the 2009 H1N1 virus does.
In the course of the airborne transmission experiments, the ferrets showed signs of illness, including lethargy, loss of appetite, and ruffled fur. One of the directly inoculated ferrets died, but all those infected via airborne viruses survived.
When the scientists sequenced the genomes of the viruses that spread through the air, they found only two amino acid switches, both in HA, that occurred in all six viruses: H103Y and T156A. They noted several other mutations, but none that occurred in all six airborne viruses.
"Together, these results suggest that as few as five amino acid substitutions (four in HA and one in PB2) may be sufficient to confer airborne transmission of [highly pathogenic avian flu] H5N1 virus," the researchers wrote.
In further steps, the researchers inoculated six ferrets with high doses of the airborne-transmissible virus; after 3 days, the ferrets were either dead or "moribund." "Intratracheal inoculations at such high doses do not represent the natural route of infection and are generally used only to test the ability of viruses to cause pneumonia," the report notes.
Mutant virus sensitive to antiviral
The team also determined that the airborne-transmissible virus was sensitive to the antiviral drug oseltamivir (Tamiflu), just as natural H5N1 viruses usually are.
In addition, they assessed whether an existing H5N1 vaccine would be likely to offer any protection against the virus. They generated a hybrid virus consisting of the mutant HA and PB2 genes combined with six genes from a mouse-adapted virus called PR8. This hybrid reacted well with antibodies generated by several existing H5N1 vaccines, suggesting that the vaccines could provide some protection.
Finally, the team tested whether antibodies in elderly people (over 70) could recognize the PR8/H5 hybrid virus. The answer was no, suggesting they would have little immunity to it or to the airborne-transmissible mutant.
The scientists note that substitutions Q222L and G224S, which are linked to changes in binding preference, have been found in avian H2 and H3 viruses in nature, though not in H5N1 viruses.
The other three mutations that were found consistently in the airborne-transmissible mutant have all been found in natural H5N1 viruses, which suggests that they don't impair viral "fitness." H103Y has been identified only once, but E627K in PB2 has been found in about 27% of avian H5N1 sequences and about 29% of human H5N1 sequences, the report says. Also, Q222L is one of the four key mutations identified in the Kawaoka study.
"Given the large numbers of HPAI A/H5N1 virus-infected hosts globally, the high viral mutation rate, and the apparent lack of detrimental effects on fitness of the mutations that confer airborne transmission, it may simply be a matter of chance and time before a human-to-human transmissible A/H5N1 virus emerges," the scientists state.
Modeling study suggests possible threat
In the accompanying modeling study, Cambridge researchers sought to estimate the risk that an airborne-transmissible virus like Fouchier's could evolve naturally. Analyzing surveillance data for the past 15 years, they noted that two of the mutations involved (T156A and E627K) are common in natural H5N1 isolates, and a number of isolates had both changes.
The researchers developed a mathematical model of how viruses replicate and evolve in a mammal and assessed the chances that the three additional needed mutations could evolve in a single host or a short chain of transmission. Among the factors they included were random mutation, prolonged infections, transmission, immune response, and deleterious mutations.
"With the information we have, it is impossible to say what the exact risk is of the virus becoming airborne transmissible among humans," senior author Smith said in a Cambridge press release. "However, the results suggest that the remaining three mutations could evolve in a single human host, making a virus evolving in nature a potentially serious threat."
The authors recommend several strategies for further research, starting with increased surveillance in regions where airborne-transmission-enabling mutations have been identified.
Early indications are that the release of Fouchier's study will not end the controversy over its publication and the broader problem of dealing with dual-use research.
At yesterday's press conference, Anthony Fauci, MD, director of the US National institute of Allergy and Infectious Diseases, defended publishing the full details. "I think the benefits that will come out of the Fouchier paper in stimulating thought and pursuing ways to understand better the transmissibility, adaptation, pathogenicity [of H5N1] in my mind far outweigh the risk of nefarious use of this information," he said in response to questions.
"When you get something out in the general literature, you stimulate thought and input from people who at first glance you wouldn't think would have an interest in it," he added. He said publication can trigger involvement by people in a wide range of disciplines, which can pay benefits.
Some biosecurity experts, though, voiced misgivings today.
Eric S. Toner, MD, senior associate in the Center for Biosecurity at the University of Pittsburgh Medical Center, commented by e-mail, "I think that the publication does increase the probability that others will want to replicate the experiments and pursue this line of research further. The more labs that are doing this research, the more likely an accidental escape becomes.
"Clearly if all the lab workers were effectively vaccinated against the virus, the risk of accidental escape would decrease. The paper says the lab workers were offered H5N1 vaccination. It would be nice to know if it was required and that the workers had good antibody titers as well as that the specific H5N1 vaccine had demonstrated efficacy against the research strain."
Toner also said the findings strongly support the concern that a naturally occurring H5N1 pandemic is possible. But he called the findings "only modestly useful" for flu surveillance.
"Many of the mutations they reported were previously known to be ones of concern," he said. "For the other mutations that had not previously been suspected to be associated with mammalian transmission, the paper suggest that it may not be individual mutations that are important but rather various combinations of mutations that produce a similar result. Knowing the significance of all the possible combinations of a nearly infinite number of possible mutations is beyond our current capabilities."
He added that the study underlines "the pressing need for a much more robust global surveillance effort, including much more sequencing and much more timely reporting of sequencing data. It also underscores the need to have a global action plan based on the surveillance. In other words, what do we do that we are not doing now if we find increasing evidence of worrisome mutations?"
An NSABB member who opposed publishing the full version of Fouchier's paper maintained that position today and expressed concern about the bosafety risks if other labs around the world use the findings to launch similar studies.
Michael T. Osterholm, PhD, MPH, director of the University of Minnesota's Center for Infectious Disease Research and Policy, publisher of CIDRAP News, said he supports the type of research done by Fouchier and Kawaoka and thinks it should continue, but not in labs all over the world.
He said the Fouchier and Kawaoka studies mark the first time humans have ever broken the barrier between nontransmissibility and transmissibility of a pathogen in animals. The heart of the issue is transmissibility, not the virulence of the lab-derived viruses, he asserted.
What a ferret-transmissible H5N1 virus would do in humans is unknown, he commented. "But if it started to circulate, such as in swine, all bets are off. Once you break that transmissibility barrier, this virus is open for serious mischief."
The details of Fouchier's study "should've been disseminated on a need-to-know basis," Osterholm said. "I think we punted on this, we didn't exhaust every possibility for disseminating it." He was referring to the conclusion of most officials and scientists involved that it was not possible to quickly devise a way to share the details only with a select group.
"We're making it much easier for everyone to do this," he said. "What if some vaccine manufacturing company in a developing country decides they want to work with this and it gets out?"
He commented that the release of a dangerous flu virus through a lab accident has already happened once, with the re-emergence of the H1N1 virus in 1977: "That was a clear case of a virus that leaked out of work that the Russians were doing."
Osterholm also warned that the issues raised by the Fouchier paper won't go away, saying research papers now in the pipeline will raise even greater concerns about possible misuse and biosafety.
"I think the federal government and the NSABB are very poorly position to deal with the future manuscripts that are already in the pipeline," he said.
Another NSABB member, David A. Relman, MD, of Stanford University, expressed the view that the Fouchier study shouldn't have been done in the first place, let alone published.
"Basically, I was and still am opposed to this particular work, because the risks outweighed the benefits, from the start, and still do," he commented by e-mail. "I don't believe that the benefits, even if they (eg, early detection in birds and animals) could be realized in the near term, are worth the significant risks. Accidental release is a major concern, as well as malignant disregard for public safety.
"We already know enough to push full out for a broadly protective H5 vaccine. What more do we need? Do we really want to wait for a mammalian respiratory transmissible H5 to show up in nature? The general public appears to be largely opposed or at least highly concerned. A real moratorium should continue for the foreseeable future."
A more positive view of the research was offered by Angus Nicoll, CBE, head of the Influenza Program at the European Centre for Disease Prevention and Control.
He said it's significant that the two different experiments by Fouchier and Kawaoka produced viruses with some degree of airborne transmissibility, which "suggests a somewhat greater risk of this happening in nature than if you found there was only one way and one virus that got you to this stage. However it must be remembered that these viruses are only just about transmissible in ferrets (a comment made by Fouchier in Washington was that these viruses really did not spread very well naturally, and their pathogenicity in the ferrets seems to be lower than wild A(H5N1)."
Nicoll added that much more study of the viruses is needed: "Can the results be replicated by others? For example are they still pathogenic in poultry? . . . That is why the moratorium on research is unfortunate though it's quite understandable given the strong feelings and concerns. I hope there will be some agreement in the autumn following a wider meeting organized by WHO [World Health Organization] that allow it to restart under the necessary level of containment."
He also commented that the findings are valuable for identifying which mutations in nature are most worrisome and for guiding the development of H5N1 diagnostic tests and vaccines.
Herfst S, Schrauwen EJA, Linster M, et al. Airborne transmission of influzna A/H5N1 virus between ferrets. Science 2012 Jun 22;336 (early online publication) [Full text]
Russell CA, Fonville JM, Brown AEX, et al. The potential for respiratory droplet-transmissible A/H5N1 infuenza virus to evolve in a mammalian host. Science 2012 Jun 22;336 (early online publication) [Full text]
May 3 CIDRAP News story on Kawaoka study
Nov 17, 2011, CIDRAP News story on the NSABB decision to evaluate the Fouchier study
Jan 20 CIDRAP News story on research moratorium