Kawaoka paper published on aerosol transmission of H5 influenza virus in ferrets

h5 ha changesOne of two papers on avian influenza H5N1 virus that caused such a furor in the past six months was published today in the journal Nature. I have read it, and I can assure you that the results do not enable the construction of a deadly biological weapon. Instead, they illuminate important requirements for the airborne transmission of influenza viruses among ferrets. Failure to publish this work would have compromised our understanding of influenza viral transmission.

The paper from Kawaoka’s group focuses on the viral hemagglutinin (HA) protein, an important determinant of whether influenza viruses can infect birds or mammals. In the image, the HA is shown as blue ‘spikes’ on the virion surface; a single HA molecule is shown at right. Avian influenza viruses prefer to attach to cells via a specific form of sialic acid that differs from the form bound by mammalian influenza viruses. This difference in receptor preference is one reason why avian influenza viruses do not transmit among mammals.

Kawaoka’s group used a random mutagenesis and selection approach to identify amino acid changes in the avian H5 HA protein that allow it to bind human receptors. These changes are located around the sialic acid binding pocket in the HA head (figure). Some of the amino acid changes were previously known, but there are also some new ones reported, expanding our understanding of how the HA binds sialic acids. Some of the HA amino acid changes allow virus binding to ciliated epithelial cells of the human respiratory tract (wild type H5 HA cannot). All of this is important new information.

The H5 HA genes with these amino acid changes were then substituted for the HA gene in a 2009 H1N1 pandemic virus, and this reassortant virus was inoculated intranasally into ferrets. The viruses did not replicate well in the ferret trachea, but viruses recovered from the animals contained a new change in the HA protein that improves replication. This change (asparagine to aspartic acid at amino acid 158) is known to prevent attachment of a sugar group to the HA and enhance binding to human receptors. Viruses with this change probably have a replicative advantage in ferrets.

A reassortant virus with HA amino acid changes N158D/N224K/Q226L transmitted through the air to 2 of 6 ferrets. Viruses recovered from one of the animals contained a new change in the HA protein, T318I. A virus with four amino acid changes in the H5 HA (N158D/N224K/Q226L/T318I) replicates well in ferrets and transmits efficiently, although the infection is not lethal.

Even more interesting are the results of experiments to understand how these HA amino acid changes affect viral transmission. The N224K/Q226L amino acid changes that shift the HA from avian to human receptor specificity reduce the stability of the HA protein. The N158D and T318I changes, which were selected in ferrets, restore stability of the HA.

There are three key questions concerning this work that must be answered.

Would an H5N1 virus with the changes N158D/N224K/Q226L/T318I transmit among humans? Probably not. The virus tested by the authors derived 7 of 8 RNA segments from a human H1N1 strain, which is well adapted for human transmission. It is likely that changes in other avian influenza viral proteins would be needed for human transmission. It might also be that entirely different changes in the H5 HA are required for transmission in humans compared with ferrets.

Is this information useful for the surveillance of circulating H5N1 strains; specifically, would the emergence of these HA changes signify a virus with pandemic potential? I don’t believe so. These are mutations that enhance the transmission of H5 viruses in ferrets, and their effect in humans is unknown. Ferret transmission experiments are not meant to be predictive of what might occur in humans.

If these results are not predictive of what might happen in humans, why were these experiments done? (to paraphrase Laurie Garret at the New York Academy of Sciences Meeting on Dual Use Research). A substantial portion of this work goes far beyond surveillance of H5N1 strains: it provides a mechanistic framework for understanding what regulates airborne transmission of avian H5 influenza viruses. In the Kawaoka study, amino acid changes that improve the stability of the HA protein were selected for during replication and transmission of the H5 viruses in ferrets. In other words, stability of the HA protein is an important property that allows efficient airborne transmission among ferrets. Additional experiments can now be designed to extend this idea. If such stabilizing changes can be shown to be important for transmission of human strains, then they might be a valuable marker of influenza transmission.

The Kawaoka paper is a significant piece of work that substantially advances our understanding of what viral properties control airborne transmission of influenza viruses. To view it as enabling construction of a bioweapon is highly speculative and fundamentally incorrect.

M. Imai, T. Watanabe, M. Hatta, S.C. Das, M. Ozawa, K. Shinya, G. Zhone, A. Hanson, H. Katsura, S. Watanabe, C. Li, E. Kawakami, S. Yamada, M. Kiso, Y. Suzuki, E.A. Maher, G. Neumann, Y. Kawaoka. 2012. Experimental adaptation of an influenza H5 HA confers respiratory droplet transmission to a reassortant H5 HA/H1N1 virus in ferrets.   doi: 10.1038/nature10831.

The NSABB speaks on influenza H5N1

The National Science Advisory Board for Biosecurity (NSABB) has published “Adaptations of Avian Flu Virus Are a Cause for Concern”, an explanation of their recommendations with respect to influenza H5N1 research (versions at Science and Nature). It starts with the statement that advances in technology now allow manipulation of microbial genomes in ways that could be misused, leading to global harm. They define dual-use research as “research that could be used for good or bad purposes”.

The authors begin their discussion of influenza H5N1 with the usual incorrect statement about the lethality of the virus:

Highly pathogenic avian influenza A/H5N1 infection of humans has been a serious public health concern since its identification in 1997 in Asia. This virus rarely infects humans, but when it does, it causes severe disease with case fatality rates of 59%.

The reference for this information is a WHO summary of confirmed human cases of H5N1. Both WHO and NSABB ignore the serological evidence for many mild or inapparent H5N1 infections. Omitting these data leads to an overestimation of the virulence of the virus, which has apparently played a large role in the NSABB’s decision.

Next, they engage in extensive speculation:

If influenza A/H5N1 virus acquired the capacity for human-to-human spread and retained its current virulence, we could face an epidemic of substantial proportions.

The virus has been circulating since the 1990s and has not acquired the capacity for human to human spread. This doesn’t mean it never will, but the possibility seems remote. The statement ‘retaining its current virulence’ of course refers to the erroneous 59% case fatality rate. What if the fatality rate is 0.1%, like seasonal influenza?

In discussing influenza H5N1 transmission in ferrets, the NSABB notes the value of the research:

The research teams that performed this work did so in a well-intended effort to discover evolutionary routes by which avian influenza A/H5N1 viruses might adapt to humans. Such knowledge may be valuable for improving the public health response to a looming natural threat.

Many have written that the research should never have been done, and that there are no benefits for human health (New York Times, Tom Inglesby, DA Henderson). Clearly the NSABB believes otherwise.

Next the NSABB describes their consideration of risk assessment of the H5N1 ferret studies. Their conclusion:

We found the potential risk of public harm to be of unusually high magnitude. Because the NSABB found that there was significant potential for harm in fully publishing these results and that the harm exceeded the benefits of publication, we therefore recommended that the work not be fully communicated in an open forum.

But there is no description of how they reached this conclusion. What data did they consider when making this decision? What were the benefits and the potential harms, and how did they weigh them? Apparently we must take the word of the panel that they reached the right decision, even though we cannot know what information they used. To convey their decision in this manner is unacceptable and sends the message that the committee did not consider specific data during their deliberations.

They conclude:

The life sciences have reached a crossroads. The direction we choose and the process by which we arrive at this decision must be undertaken as a community and not relegated to small segments of government, the scientific community, or society.

This is precisely why the decision to redact publication should not have been made by the NSABB or any small group of individuals. I agree that this is an ‘Asilomar moment’, a time when scientists must meet to decide what types of microbial research should be regulated. This should be a discussion among a large group of scientists, not bioterrorism policy analysts.

I understand the need to regulate certain types of experiments on microbes. But when I balance the benefits and risks of the H5N1 ferret transmission experiments, it does not make sense to stamp them as dual use and restrict publication of the results. Let publication proceed and then decide how to decide on how to move forward.