H5N1 Transmission in Ferrets
By Amesh A. Adalja, MD, FACP, May 11, 2012
Fears of an impending H5N1 influenza pandemic have existed since the first cases of human infection appeared 15 years ago. A pandemic could occur only if the virus acquired mutations that made it more adept at transmission among humans. Two recent and controversial studies have raised the spectre of increased mammalian droplet transmissibility. Now, after prolonged international debate about the biosecurity concerns posed by these studies of genetically modified H5N1 influenza viruses, the first paper, by Yoshihiro Kawaoka and team, has been published by Nature (online, May 2, 2012).
Focus on HA Binding Domain
One of the factors that restricts the host range of H5N1 is the predilection of the viral hemagglutinin globular head to bind host sialic acid linked to galactose by an α2,3 linkage. The α2,3 linkage is found in avian species; the α2,6 linkage is found in mammals. Using the GeneMorph II kit and working in a BSL-2 containment setting, Kawaoka and colleagues introduced random mutations into the binding region of the HA gene in an H5N1 virus in which the polybasic HA cleavage sequence was removed. They recovered 370 viruses with a preference for the α2,6 linkages needed for optimal human adaptation and 9 isolates that were used for more detailed experimentation and characterization.
When the viruses where characterized, the induced mutations clustered around the binding pocket region of the HA molecule. Mutations known to confer human binding properties, including N186K, S227N, and Q226L, were present, along with new mutations not previously known to play a role in human binding. The researchers then subjected 8 of the 9 viruses to solid-phase binding assays in which one virus with strong α2,6 binding properties remained (1 of the 9 viruses reverted to wild type, leaving 8 for study). This remaining virus contained the following HA mutations: E119G, V152I, N224K, and Q226L. These mutations were then tested singly and in combination, revealing that the N224K/Q226L combination was responsible for the shift to α2,6 binding. This binding preference was confirmed in cell culture with human tracheal cells. The specific mutation N224K may alter the orientation of the 220-loop of the HA-receptor binding domain allowing a better interaction with the mutation at position 226 and α2,6 linked receptors.
Combination with 2009 H1N1
One way that an H5N1 pandemic may occur would be through reassortment with a seasonal influenza strain in an animal, such as a pig. When the 2009 H1N1 pandemic occurred, the fear of that occurrence became real. To test the possibility of a 2009 H1N1-H5N1 reassortment in a pig, the HA gene from the mutant H5N1 virus was combined with the remaining 7 genes of a 2009 H1N1 virus in a BSL-3+ setting approved by the CDC and USDA. The researchers generated 3 combination viruses that had the virulence-enhancing polybasic cleavage site intact. The viruses they created included one that had all 4 mutations, one that had only the N224K/Q226L combination mutation, and one with a wild type H5 HA.
Three ferrets were then infected with the virus via nasal inoculation. At day 3, both of the mutated recombinant viruses of interest showed diminished replicative ability in ferret tracheas. Nasal replication levels were similarly low with the quadruple mutant virus, but not with the dual mutant. By day 6, the virus with the N224K/Q226L mutation still showed high levels of nasal replication. When the dual mutant virus was isolated from 2 of the ferrets, the researchers observed a new mutation at site 158; that mutation caused the loss of glycosylation (a factor in tissue tropism) at the site.
Ferret Transmission Studies
To establish whether a reassorted H5/H1N1 virus could be transmitted between ferrets by respiratory droplets, inoculated caged ferrets were placed next to naïve caged ferrets. Of the viruses tested, the triple mutant (N158D/N224K/Q226L) virus was recovered from 2 of 6 contact ferrets. Seroconversion occurred in 5 of 6 ferrets. No ferrets suffered fatal infection. New mutations were identified in the transmitted viruses (K193N, A242S, T318I). Virus from the contact ferret that was shedding the highest titer contained the N158D, N224K, Q226L, and T138I mutations and was selected for further study. In transmission studies, this virus was isolated from 4 of 6 contact ferrets and caused all contact ferrets to seroconvert. No fatal infections occurred. In other experiments, the T318I mutation alone was not found to be sufficient for transmissibility or α2,6 binding, but it was found to stabilize the HA molecule with respect to pH-dependent fusion with host cells and heat stability.
Virulence and Vaccine Neutralization
To assess the virulence of the mutant virus (N158D/N224K/Q226L/T318I), the researchers compared it with 2009 H1N1 and found that the 2 groups of ferrets had no statistically significant difference in weight loss, and they had similar pathological changes in nasal passages. Each of the 3 reassorted viruses caused lung pathology.
To determine whether existing H5 influenza vaccines would protect against the mutant virus, sera from those vaccinated with the U.S.-licensed vaccine was tested against the virus and found to be neutralizing in HAI assays. The virus was also susceptible to oseltamivir.
Limited Applicability of Results
The discovery of mutations that confer respiratory transmissibility of an H5/H1 chimera virus in ferrets is important in that it expands understanding of basic influenza biology. However, it may be difficult, for a number of reasons, to use this data to make inferences about virus behavior in a natural setting. The researchers studied a combination virus that included 2009 H1N1 virus, which could have exerted an influence on mammalian transmissibility as strong as that of the 7 genes that came from H1N1, and especially PB2. Further, the pathway to an H5N1 pandemic may not require reassortment with H1N1.
Of the 4 mutations identified in this study, N158D, which caused loss of glycosylation, has already been observed in some areas, some of which also harbor the human-adapting mutation at site 627in PB2.Theoretically, identifying these mutations could aid influenza surveillance and allow the rate and locations at which these mutations are occurring in the wild to be assessed; however, widespread and timely sequence surveillance is not being conducted in most of the affected countries.
This study underscores the fact that in order to harness the full benefit from the knowledge gained by this study, an augmented surveillance capacity is needed.
Reference: Imami M, Watanabe T, Hatta M, et al. Experimental adaptation of an influenza H5 HA confers respiratory droplet transmission to a reassortant H5 HA/H1N1 virus in ferrets. Nature 2012. http://www.nature.com/nature/journal/vaop/ncurrent/full/nature10831.html. Accessed May 7, 2012.