Variants of an avian influenza (H5N1) virus strain that are transmitted through the air among ferrets were recently isolated in two laboratories, one at the University of Wisconsin at Madison and the other at the Erasmus Medical Center in Rotterdam, the Netherlands. The National Science Advisory Board for Biosecurity has asked that critical details of that research be withheld from publication to prevent their use by bioterrorists. This decision rests on unconvincing scientific grounds, and is a weak test case on which to set a precedent for restricting research that can benefit humanity.
The experiments were meant to determine what makes this influenza virus transmissible. It is lethal for birds, humans, and ferrets, but is not readily transmitted among humans or ferrets. After 10 serial ferret-to-ferret passages, a virus was obtained that could spread through the air from an infected ferret to an uninfected one in a nearby cage. The board advised that methods for how the virus was isolated and the amino-acid changes leading to transmissibility should not be published.
The advisory board wants to restrict access to this information because the avian influenza virus is widely believed to be highly pathogenic; more than half of the 600 seriously ill people infected with it who have been admitted to hospitals have died. Unfortunately, that statistic does not provide information on the case fatality rate, because we do not know the total number of human infections that have occurred.
Results from a recent study of rural Thai villagers indicate that 73 of 800 individuals examined (9.1 percent) have antibodies against one of two different H5N1 strains. If we were to find that a substantial portion of rural Asians has been infected with the virus—with either mild or no disease—it would drastically change our view of the lethal power of the virus.
The advisory-board decision is based on the assumption that experimental results obtained in ferrets would predict what occurs in humans. This is a faulty argument. Ferrets are a good model for influenza experiments because they display flu-like symptoms, immune responses, and pathological alterations—including elevated temperature, weight loss, and changes in tissue structure—similar to those of humans, but ferret and human influenza are not the same. Influenza virus strains behave differently in humans and ferrets. For example, the 2009 pandemic H1N1 virus, which caused severe infections in some ferret studies, had relatively mild effects in humans.
Central to the decision of the advisory board is the assumption that the H5N1 virus that can be transmitted among ferrets can cause disease in humans. In this context it is important to remember that passage of viruses in a different host is a strategy used for reducing viral virulence in humans. The infectious, attenuated yellow-fever-virus and poliovirus vaccines have been produced in this way. So the possibility exists that passage of the H5N1 virus in ferrets will attenuate its virulence in humans, a possibility not considered by the advisory board.
It also seems unlikely that the ferret-adapted H5N1 would be used for bioterrorism, as its potential for transmission and lethality in humans is unknown. Bioterrorists are not scientists wishing to answer questions; their goal is to instill terror. There are many other viruses and bacteria that are known to be lethal for humans that would be suitable for their purpose.
A broader issue is whether laboratory-modified viruses can cause disease in humans. Genetic manipulation of viruses is generally done without understanding of the requirements for efficient replication, disease production, and transmission in humans. Viruses engineered in this way are not subject to the strong selection pressures that sift through the enormous and diverse collection of viruses to produce a human pathogen. To think that we can duplicate this environment in the laboratory indicates a severe case of scientific hubris.
The purpose of a scientific paper is not only to report experimental findings that answer a question, but also to provide sufficient methodological detail to enable others to repeat and extend the findings. What the board has suggested—to publish the H5N1 findings without these details—is contrary to this purpose and would make a mockery of the journals in which the papers appear. A more sensible path would be to publish the findings, recognizing that a dialogue must take place to settle the issues revealed by these experiments.
The National Science Advisory Board on Biosecurity is not a suitable body to undertake such a dialogue—it is too small and cannot realistically consider all aspects of infectious diseases. A broad collection of microbiologists, similar to the one assembled in Asilomar, Calif., in 1975 to debate the safety of recombinant-DNA experiments, should be convened to discuss what types of experiments merit regulatory attention and restrictions on publication.
No one can guarantee that the H5N1 virus passed in ferrets would not be lethal and transmissible in humans. However, the scientific arguments I have presented suggest that the likelihood of this scenario is minimal. We must balance the potential dangers of disseminating this information with the dangers of setting a precedent for restricting scientific publication—an action that would surely lead to highly restrictive regulations on scientific research and publication.