After more than 6 months of heated discussion, the second group that succeeded in making the H5N1 avian flu transmissible between ferrets, considered a good model for human transmission, has published its results. The paper, which came out today (June 21) in Science, demonstrates that only five mutations are needed to confer this aerosol transmissibility among mammals, and that re-assortment between different types of viruses—a technique used by the other group, which published its results last month in Nature—is not necessary.
“Five mutations are not very many—fewer than previously estimated,” said Eddie Holmes, who studies virus evolution at Pennsylvania State University but was not involved with the research. The two papers together are the first “really good experimental data” addressing the question of how little mutation H5N1needs to undergo before it becomes transmissible between mammals via the air.
Though avian H5N1 has shown itself quite lethal in the few human cases identified, which raises the specter of a virulent pandemic, it’s been circulating in birds since 1997 without efficiently jumping to mammals. This 15-year lag has led some in the influenza community to argue that the virus will not and cannot switch to mammalian hosts, explained Paul Thomas, who studies immune responses to influenza at St. Jude Children’s Research Hospital in Memphis, but did not participate in the study. This argument makes the Science and Nature studies important proofs of principle, he said.
The previous Nature study, led by Yoshihiro Kawaoka at the University of Wisconsin, Madison, at aimed to identify mutations that could make H5 viruses, which have never initiated a human pandemic, transmissible in mammals. They randomly mutated the hemagluttinin (HA) gene—the H5 part of the virus—until they found two that caused it to switch from bird to human target receptors, then swapped it for H1 in an H1N1 strain (the strain responsible for 2009’s swine flu pandemic). The resulting H5N1 contained the mutated avian H5, with the N1 and internal proteins from H1N1, mimicking the natural process of viral re-assortment, which might occur if avian flu jumped from a domestic duck to a pig already infected with swine flu, for example.
In the Science study, researchers examined a different possibility: that H5N1 might acquire the ability to transmit among mammals via the air though mutation alone, no re-assortment necessary. To make the leap, H5N1 would need to change in specific ways, such as changing its target receptor from 2,3-linked sialic acids found in birds to the 2,6-linked sialic acids in human and ferret upper respiratory tract, and replicating efficiently at mammals’ lower body temperatures.
To examine what mutations could make this possible, the researchers, led by Ron Fouchier at Erasmus Medical Center in the Netherlands, used a strain of H5N1 taken from a human patient in Indonesia in 2005. They gave the virus an initial mutational “boost” by adding three mutations identified from the 1918, 1957, and 1968 H2 and H3 pandemics. Two of these mutations helped HA switch from binding bird receptors to human ones, and another mutation, in the polymerase protein PB2, helped the virus replicate at 33 degrees Celsius, rather than 41 degrees, the average body temperature of birds.
The scientists then began passaging the manipulated H5N1 and wild-type H5N1 between ferrets by hand. After 10 passages, both the mutated and wild-type strains had acquired new mutations, but only the manipulated H5N1 could be transmitted from an infected ferret to an uninfected ferret through the air.
Examination of the manipulated H5N1’s gene sequences identified two additional mutations present in all air-transmissible strains, both modifying the HA sequence and enhancing binding to both human and bird sialic acid receptors.
This result is similar to the mutations identified in Kawaoka’s study, said Fouchier in a press conference. “We both find … loss of glycosylation at the tip of the HA molecule, and this loss of glycosylation seems to increase the receptor binding specificity of the HA,” Fouchier explained. And though not all the mutations identified in the two studies match, “the mutations that are not identical still have a similar phenotypic trait,” he added.
Fouchier also noted that his group’s strategy allowed them to examine which genes in the internal proteins might mutate to facilitate H5N1’s mammalian airborne transmission. Mutations occurring frequently, though not universally, generally targeted polymerase genes, suggesting that enhancing viral replication will likely enhance transmission, he noted.
What these results mean for the risk of a human pandemic is still unclear, said Holmes. Because the virus needed “help” in the form of three initial mutations, scientists can’t say what chances there are of H5N1 acquiring these in the wild. Furthermore, the mutated virus lost virulence, becoming non-lethal, as it gained the ability to transmit via the air, suggesting there may be a tradeoff between virulence and transmission. But although ferrets are generally considered the best model system available, they aren’t human, Holmes noted, so “we don’t know what the virus would be like if it evolved human-to-human transmission.”
Still, the separate studies highlight that while H5N1 can acquire airborne transmission between mammals, “there’s no single evolutionary trajectory,” said Andrew Mehle, a virologist at the University of Wisconsin, Madison, who was not involved in the research. But scientists are moving toward a more mechanistic understanding of virus transmission and how flu viruses might jump between species, Mehle said. “If [researchers] know the virus needs to accomplish X, Y and Z [to transmit between humans], can we think of other ways it can do that?"
S. Herfst et al., “Airborne transmission of Influenza A/H5N1 virus between ferrets,” Science, 336:1534-41, 2012.