In a controversial study published 2 years ago Dutch scientists described a lab-modified strain of H5N1 influenza virus that was capable of airborne transmission among ferrets. Now the same researchers say they have identified five specific mutations that gave the virus this ability, a claim that is renewing debate about the risks of conducting and publishing such experiments.
Writing in Cell, the scientists said they identified two combinations of five mutations that affected specific characteristics of the virus and collectively enabled it to spread by air. They assert that the findings will help in the effort to detect early warning signs of flu strains that could cause a pandemic.
But other experts question the scientific value of the findings and argue that they are not worth the risks involved in conducting such experiments and publishing the full details. They assert that the research poses a risk of either accidental or intentional release of dangerous viruses.
Building on 2012 study
The new study builds on a US-government funded study that was published in June 2012 in Science. In that case, Ron Fouchier, PhD, of Erasmus Medical Center in the Netherlands and colleagues described how they used a combination of genetic engineering and serial infection of ferrets to create a mutant H5N1 virus that could spread among the animals without direct contact.
That raised eyebrows because the avian H5N1 virus, while it infects humans relatively rarely, has been deadly in about 60% of the more than 650 known cases. The virus does not spread readily from person to person, but if it gained airborne transmissibility in humans, it would raise the specter of a severe pandemic.
The 2012 findings were published after a long and heated debate about the risks of disclosing the full details. When early word of the findings spread in the fall of 2011, the US National Science Advisory Board for Biosecurity (NSABB) reviewed the study and recommended against publication of the full details. The board recommended the same for a similar study by Yoshihiro Kawaoka, DVM, PhD, and colleagues.
But after the authors of the studies briefed the NSABB and made some clarifications in their manuscripts, in March 2012 the board reversed itself and recommended in favor of full publication, though the vote on the Fouchier study was not unanimous.
The controversy temporarily stalled H5N1 "gain of function" (GOF) research, as it is called, and it also prompted the development of US guidelines for oversight of "dual use research of concern" (DURC), meaning work yielding findings that could be used to do harm as well as good.
In the new study, the researchers started with the mutant virus from their previous experiments that had the lowest number of mutations—nine—consistent with airborne transmission in ferrets, according to the report. Then they systematically deleted one or two mutations at a time, infected ferrets with the resulting strain, and tested whether other ferrets housed nearby (but not in direct contact) became infected. In each round, either two or four ferrets were exposed to the infected ferret.
This process led to the identification of two slightly different combinations of five mutations that the authors say are sufficient to permit airborne transmission of the virus in ferrets. The mutations are:
- E627K in the PB2 gene
- H99Y in the PB1 gene
- H103Y and T156A in the hemagglutinin (HA) gene
- Either Q222L or G224S in the HA gene
Further, the scientists said that the PB2 and PB1 mutations increased viral replication, two of the HA mutations improved the binding of the virus to mammalian respiratory cell receptors, and one HA change increased stability at high temperatures and in more-acidic conditions.
The authors acknowledge that their findings are limited by the number of ferrets used, saying the study "was purposely designed to define a minimal set of substitutions rather than the definitive minimal set of substitutions required for airborne transmission in ferrets." Better defining the minimal set of mutations would require "substantial numbers of ferret pairs," given statistical considerations.
Helping to prevent pandemics?
"By gaining fundamental knowledge about how the influenza virus adapts to mammals and becomes airborne, we may ultimately be able to identify viruses that pose a public health risk among the large number of influenza viruses that are circulating in animals," Fouchier commented in a Cell Press news release. "If we can do this, we might be able to prevent some pandemics in the future."
The report notes that some of the five mutations, particularly PB2 E627K and HA T156A, have been found in wild-type H5N1 viruses, but the full combination of five has not emerged in nature. But because mutations that are "functionally equivalent" to the identified ones exist in nature, the natural emergence of more transmissible H5N1 viruses can't be excluded.
The authors conclude that the findings do not imply that an H5N1 pandemic is imminent, but they suggest a need for a more vigorous search for emerging flu viruses that could pose a pandemic threat.
In a companion commentary in Cell, Matthew S. Miller, PhD, and Peter Palese, PhD, of the Icahn School of Medicine at Mount Sinai, New York City, welcome the findings and call for further GOF studies.
Palese and Miller comment that in the earlier study by Fouchier and colleagues, the virus that became more transmissible also became less virulent. "Assessment of how adaptation in ferrets affects viral fitness, virulence, and transmission . . . is sorely needed to gain a truly holistic perspective of the likelihood that these viruses might cause a pandemic and what characteristics such a pandemic might exhibit," they write.
They go on to assert that the intense reaction to the H5N1 transmission studies in 2012 stemmed "largely from scientific ignorance with regard to how transmission, virulence, and fitness interrelate. Ironically, the only way to address this uncertainty is to move forward with GOF studies that will serve to contextualize how adaptations that mediate mammalian transmissibility affect other properties of the virus."
Critics raise multiple questions
But Marc Lipsitch, DPhil, professor of epidemiology and director of the Center for Communicable Disease Dynamics at the Harvard School of Public Health, questioned the scientific value of the findings and said there are safer ways to study the viruses' transmission characteristics.
In an interview, he said that identifying the mutations in a non-human species in a laboratory setting, using as few as two ferrets per experiment, "to me adds virtually no new biology."
For a hypothetical comparison, he said, "If you show that there are five different ways to prop up a bridge and you set up a lab with no wind and no vibrations and show that if you take three of them away, you can still keep the bridge up, that's kind of interesting, but it doesn't add to our fundamental understanding of how to keep a bridge up."
Further, Lipsitch said a study published by MIT researchers last summer showed that if some of the mutations identified by Fouchier and Kawaoka were joined to a slightly different H5N1 virus's genes, they didn't change the receptor binding to human specificity, as they had in the viruses used by Fouchier and Kawaoka. "What that tells us is that when you know something about isolate A, you know something about that isolate, but not necessarily about another isolate."
He also asserted that, given how few ferrets were used in the study, "the sample size is so small as to be nearly meaningless." Each virus strain that was studied must have some probability of transmission, he said, but "when the sample size is as small as theirs, you can't make conclusions about one probability being higher than another."
Lipsitch also contended that the experiments shouldn't have been done in the first place, because they pose a "measurable risk" of creating a pandemic strain and because there are safer ways to research transmissibility.
He referred to a recent estimate based on data from the Centers for Disease Control and Prevention that the risk of accidental infection in a high-containment laboratory is about 2 per 1,000 lab-years of research. "That's actually a high rate when you consider the potential impact," he added.
He said there are various ways to study transmissibility that are safer, such as using viruses that can't replicate or studying the course of adaptation in existing seasonal flu viruses that originated as avian viruses and evolved into pandemic strains.
Another expert, Michael T. Osterholm, PhD, MPH, who is a member of the NSABB, questioned the decision to publish the full findings, given the risk of laboratory accidents and leaks if others duplicate the research.
"It just further illustrates the inability of our current DURC activities to deal with gain-of-function work and publication. We just keep making it easier and easier for labs around the world to do this work," he said in an interview. He is director of the University of Minnesota's Center for Infectious Disease Research and Policy, which publishes CIDRAP News.
Osterholm said there are many labs capable of doing GOF research that are not under US government oversight because they don't receive US funding.
Contending that the risk of flu outbreaks resulting from lab leaks is real, he observed that an H1N1 virus that re-emerged in 1977-78 resulted from an accidental release from a Russian research program on live attenuated flu vaccines.
Osterholm also reiterated an argument he and others made during the publication controversy 2 years ago: that there should be a way, short of publication, to share critical details of potentially dangerous studies with those who have a legitimate need to know.
"There's got to be some kind of information category that's between open source and classified, so those who have a need to know can have access," he said.
"As long as we allow open-source publication, what we've done is put all the 'dangerous work' in a government-supported lab, but then we've allowed everyone else to do it," he said. He likened that approach to supposing that "if we fix three of five screen doors on a submarine, it's going to be OK."
Linster M, van Boheemen S, de Graaf M, et al. Identification, characterization, and natural selection of mutations driving airborne transmission of A/H5N1 virus. Cell 2014 Apr 10;157(2):329-39 [Abstract]
Miller MS, Palese P. Peering into the crystal ball: influenza pandemics and vaccine efficacy. (Commentary) Cell 2014 Apr 10;157(2): 294-9 [Abstract]
Apr 10 Cell Press news release
Apr 4 CIDRAP News story "Experiments render H7N1 virus more contagious in ferrets"
Jun 21, 2012, CIDRAP News story on earlier Fouchier H5N1 study
Jun 7, 2013, CIDRAP News story covering MIT studies