At last, one of the papers investigating H5N1 influenza transmission in ferrets has been published in the journal Nature yesterday. To recap the controversy briefly: news of experimental studies investigating transmissibility of avian H5N1 influenza hit the news this fall, igniting a fierce debate about biosecurity, “dual use” research, and the damage that censorship can have on scientific advancement. An advisory group, the NSABB, recommended partial censorship of the data, perhaps believing that redacting specific data from the publications would prevent information on how to generate a highly pathogenic mammalian transmissible virus from getting into the hands of bioterrorists or others incapable of handling such viruses safely. However, after some new data or clarification of data presented in a revised version of the submitted manuscript, the NSABB recommended publication in full.
Avian H5N1 influenza virus has caused sporadic infections in humans who have close contact with infected animals. Human-to-human transmission has not been observed. But could an H5 virus mutate or reassort, allowing human-to-human transmission?
One thing that has been lost in this whole controversy is that this study is actually a great demonstration and application of evolution. How does a virus switch to a new host and transmit efficiently between individuals? Start with a diverse population that varies in a particular trait (in this case, ability to bind the human receptor). Put it through selective pressure. This occurred in several steps: first, select for viruses that can bind the human receptor in vitro. From those that bind, select ones that can efficiently replicate in the respiratory tract of the animal. Finally, take the efficient replicators and allow for transmission. At each step of selection, mutations that naturally occur during viral replication further diversify the population resulting in variants that possess the desired property. Those variants get selected for the next experiment.
This study focused on one particular influenza virus protein, hemagglutinin (HA). HA is on the surface of the virion and is what the virus uses to attach to the host cell, the first step in viral infection. HA of human influenza viruses bind a sugar on the surface of human cells, which is slightly different from that found in the avian respiratory tract. Avian viruses, of course, bind to the form found in birds. H5 shows a strong preference for binding the avian receptor, so Karaoka et al were interested in finding out if H5 could change to recognize the human receptor, allowing more efficient transmission between mammals.
To address this, they began with a mutagenesis technique to introduce random mutations in the globular head (the receptor binding part) of HA, then used an in vitro approach to select for mutants that bound the human receptor (selection step 1). Through this process they identified three H5 variants that gained the ability to bind the human receptor while maintaining the ability to bind the avian receptor, and one that switched specificity completely to the human receptor. Several of the mutations identified in this study had already been shown in previous studies to be important in receptor specificity.
To test if the variant H5s conferred binding to human receptors in vivo, sections of human tracheal tissue were exposed to the viruses and only two were able to infect (selection step 2). This suggests the virus can infect human epithelium of the upper respiratory tract.
Next, the two remaining variants were used to infect ferrets. Both replicated in ferret respiratory tracts, but one replicated to higher levels. When they sequenced the virus that they isolated from that ferret, it was different from what they had put in: a new mutation had appeared. This new mutation presumably confers the property of better replication in the ferret respiratory tract, so it outgrew the original input virus (selection step 3).
Using this new virus (now with a total of 3 mutations in H5), transmission was tested in ferrets. Compared to the original H5, which did not transmit via aerosol, the 3-mutation variant did transmit, although between only 2 of 6 animal pairs. Again, they sequenced the virus that was present in the contact animals and found that it was different than what had gone in to the inoculated animals. Yet another mutation had appeared (selection step 4). This additional mutation appears to enhance transmission: the new virus, now with 4 mutations, transmitted more quickly and between more pairs of animals than the 3-mutation virus. Although the virus can transmit, none of the infected animals died, but they did show pathology at the site of infection.
So what does this all mean? The best model available for influenza transmission studies is ferrets. Ferrets aren’t humans, so its important to keep in mind that this is a model that helps us understand what viral or host factors are involved in aerosol transmission in these mammals, and maybe, but not necessarily, in humans. Since we don’t know what is necessary for human-to-human transmission, it is valuable to have an animal model to give us some ideas of what to look for. It can provide some good hypotheses on what mediates transmission in humans, which would then have to be further tested. Obviously, specificity for the human receptor is necessary, but the mutations identified tell us more than that. Mutations that change specificity are not sufficient for transmission. It turns out that those mutations also decrease the stability of the HA protein. The additional mutations acquired through the selection steps compensate for that, and enhance stability. So now we know that HA stability is important in influenza transmission. Between ferrets. That’s probably true for humans too, but it would need to be tested.
If you are still reading, you are obviously procrastinating, and are probably avoiding studying for your final exam. But here are some more thoughts on the controversy overall. This is a really interesting paper, with nothing particularly frightening or worrisome about it. Certainly not any more so than other papers doing similar work that were published without so much controversy. If “dual use” research needs to be regulated, it needs to be done before the work is done, not after. If the NSABB was concerned about this kind of research, why only express concern once the experiment succeeds? In my intro biology class, we read another paper, published in 2005, which was addressing the exact same question, in an almost identical way. The difference is that they failed to make a transmissible virus. If there is a concern about this kind of research, a concern that it is too risky to do these kinds of experiments, shouldn’t the alarm have been raised regardless of the outcome? It just doesn’t make sense to me why it suddenly became so concerning. If anything good has come of this controversy, it is the widespread discussion that this has stimulated on the importance of open communication of scientific data, the importance of not censoring in science. Ironically, had we all been given access to the data, like through a journal publication, it would have been apparent that there wasn’t anything to be concerned about.
I agree that there shouldnt be any censorship. What I meant by my statement is essentially what you said: if there are experiments of concern for which we need to consider whether the experiment should be done at all, that needs to be considered before actually doing the work. The NSABB seemed to think that H5 ferret transmission experiments like these fall under the category of serious concern, but they only expressed concern once a paper was submitted showing some success. Perhaps I dont fully understand the role of the NSABB, but it seems to me that if they are only called in at the stage of publication, the only thing they can be is a censorship body, and that obviously is not OK.
In response to the statement “If there is a concern about this kind of research, a concern that it is too risky to do these kinds of experiments, shouldn’t the alarm have been raised regardless of the outcome?”:
I don’t believe that regulation of information of this type should be censored at all, and thus, I think any public safety concerns regarding such work should be sorted at the “Should we even do this experiment?” phase of the process. That said, I see no reason to avoid doing these experiments and perhaps even more concerning ones. We will, regardless of the outcome of these experiments, be gaining knowledge; why conduct research at all if we can only go so far before gaining more knowledge becomes a bad thing, for whatever reason?
I don’t think we should let there be a threshold that kicks in to prevent publishing research whenever people fear public safety because, although the remote possibility exists, I think it highly unlikely that someone with malicious intent will successfully develop a “super-virus.” Experts have been studying such topics for a while; what is the likelihood that a terrorist will figure out how to create such a virus before experts in the field? And, if we fear terrorists like this, why avoid publishing important data on such a matter that may help us understand such viruses before those with malicious intent do? Wouldn’t it be better to understand what we may one day deal with (as a result of evolution or human intervention) before times become dire?
I’m thrilled to hear that this paper was finally published, and it is unclear to me why there was such anxious hysteria surrounding it. Karaoka et al succeeded in making a transmissible virus and also discovered factors that possibly mediate transmission in humans through a ferret model. This information needed to be released in order to gain more knowledge about a potential pandemic.
It was especially interesting how Karaoka et al took an evolutionary approach to virus transmissibility. The design of this experiment really echoed the excerpt we read from Paul Ewald’s “Plague Time.” Looking at transmissibility in terms of selection pressures (H5 in vitro, H5 in vivo, then see if the virus can transmit from ferret to ferret) reinforced Ewald’s idea that naturally occurring mutations diversify a population during viral replication. This diversity results in variants that are more “fit” for survival. It is these mutants that are more capable of transmission.
It seems as though mutations enhance transmissibility, but not necessarily the effectiveness or strength of the virus. For example, when H5 possessed four mutations, the virus was transmitted between more ferret pairs at a faster rate. However, none of these ferrets actually died from the virus. Therefore, it can be concluded that an increased number of mutations make H5 less stable.
The fact that Karaoka et al could even produce a transmissible virus is a feat; the scientific community has now gained a greater understanding of how H5N1 adapts to selective pressures in the environment.