Two groups of scientists — one based in the Netherlands [1] and the other in the United States [2] — have carried out detailed studies on the experimental evolution of the A/H5N1 virus, more commonly known as the “bird flu”. Public health officials have documented at least 600 cases of humans infected by A/H5N1 since it was first reported in Hong Kong in 1997. Half of those infected were killed, a strikingly high mortality rate compared to most strains of the flu. While this mortality rate has garnered widespread media attention, the impact of A/H5N1 on humans has remained relatively small due to its inability to transmit efficiently from human to human. However, the virus is highly transmissible in some non-human animals, especially birds, and has caused large-scale epidemics in these animals.

People who have direct and extended contact with birds infected with A/H5N1 (such as those who work on chicken farms) are at the highest risk for infection, because the virus has many more chances to make the jump into these humans. Currently, the virus is only known to pass readily from infected birds to humans, and not between humans. But viruses are always changing and evolving, developing random mutations in their genetic material that can change the way they behave (this is why we need new flu vaccines every year). What if the A/H5N1 virus evolves the ability to pass from human to human? A virus that is both easily transmissible and as deadly as A/H5N1 could wreak havoc on the scale of the 1918 Spanish Influenza, which infected about a third of all people and killed at least 3% of the world’s population. In other words, how worried should we be about a repeat of the 1918 flu pandemic?

With this concern in mind, the two groups of scientists from the Netherlands and the US set out independently to determine how difficult it would be for the A/H5N1 virus to acquire mutations that confer the ability to jump from person to person. Their results were submitted to the scientific journals Science and Nature last year, but it took ten months for both papers to finally be published and be freely available online [1,2]. The publication of these papers was fraught with controversy, as scientists and the government attempted to find a balance between publishing research that could potentially benefit public health, and unveiling information that could be misused to the detriment of society.

Making a virus more dangerous

To determine whether A/H5N1 has the capacity to become more transmissible among humans, the two groups of scientists took similar but distinct approaches.

The scientists from the University of Wisconsin-Madison, who published the first paper in Nature, created a hybrid virus containing genes from the H5N1 bird flu and the 2009 H1N1 swine flu, which is known to “reassort” or shuffle its genetic materials easily with other flu viruses. Then, by adding random mutations to the genetic materials of the resulting virus, they were able to search for new mutations that would help the virus bind to human cells (the first prerequisite for infection). They used these mutated viruses to infect ferrets (a model for studying flu transmission in humans) and allowed the virus to evolve further in the ferrets.  Ferrets are a good model for studying human flu infection because they are susceptible to the same flu strains as humans and exhibit many similar symptoms, including fever, runny nose, and sneezing. As in humans, the original H5N1 virus could not readily jump between two ferrets. However, the scientists found that their hybrid virus only needed to acquire four mutations to enable transmission from ferret to ferret [2].

Instead of creating a hybrid virus and making random mutations, the group from the Netherlands allowed the original virus to evolve on its own as it transmitted from ferret to ferret. They first infected ferrets with either wild-type A/H5N1(the virus normally found in the “wild”), or A/H5N1 that they had engineered to contain three mutations believed to help with ferret-to-ferret transmission. These mutations turned out to be insufficient on their own to increase transmission, but the scientists predicted that with a few additional mutations, the virus might acquire this trait (Figure 1). The researchers allowed the virus to grow and divide in the ferret for four days before using nasal samples from the infected ferret, presumably containing the virus, to pass the infection on to a second ferret. This process was repeated ten times before the virus from the last ferret was harvested and analyzed to identify its mutations. After passaging the virus through just ten ferrets and allowing random mutations to develop naturally, the virus had evolved the ability to infect a healthy ferret living in the cage next to an infected ferret. This transmissible A/H5N1 only has five additional genetic mutations compared to the naturally occurring A/H5N1 [1].

These findings were startling — many people expected that mammal-to-mammal transmission would require more than just a few mutations. That said, it is impossible to estimate the chances that these specific mutations might naturally occur outside of a controlled lab environment [3]. Also surprising was the fact that these transmissible viruses were much less dangerous — they caused nowhere near the 50% human mortality seen with A/H5N1, though the mortality figures between humans and ferrets might not be entirely comparable. In other words, compared to wild-type A/H5N1, the variants observed in the lab could more readily jump between mammals, but the infection may be less serious.

Figure 1. Schematic of A/H5N1 evolution in ferrets, the experiment performed by the group from the Netherlands. The scientists started with “P0 wt” wild-type (normal) virus, or “P0 mut” virus that they had engineered to contain three mutations believed to help with transmission. The mutations are represented by yellow marks on the blue-colored lines, which represent the genetic material of the virus. The first ferret (P1, for “passage 1”) was infected with one of these two viruses, and the virus was allowed to grow in the ferret. After about four days, the animals were humanely sacrificed and nasal turbinates (NT, part of the nasal cavity) were homogenized and used to infect the next ferret. After the 6th ferret (P6), solution rinsed through the ferret’s nose (NW or nasal wash) was used to infect the next ferret. Finally, nasal washes from the 10th ferret were collected, and the resulting virus was analyzed to identify the genetic mutations it had acquired along the way (black marks on the yellow-colored genetic materials). (Image from [1])

The long road to publication

Nowadays, a ten-month interval between submission and publication is certainly not unheard of for scientific papers. Once a paper is submitted to a scientific journal, the journal editors send it to other scientists in the field for the process of peer review. These reviewers take a few weeks to read and respond with comments and suggestions on the manuscript. They also provide recommendations to the journal editors to help them decide whether they should publish or reject the paper, or ask the authors to revise and resubmit the article. The authors must then perform additional experiments to address reviewer concerns, and may go through multiple rounds of review and revision before their paper is ultimately accepted and then scheduled for publication. Not surprisingly, this process often takes months.

The reason for the publication delay for these papers was different, however.  For months, journal editors and governmental agencies debated whether the papers should be published at all — not because the reviewers disliked a particular experiment or wanted more data, but because publishing the details of these experiments might pose an international biosecurity threat.

The National Science Advisory Board for Biosecurity (NSABB), a government panel based at the National Institutes of Health, asked Science and Nature in December 2011 to consider not publishing these papers in their entirety. The panel was concerned that publishing the exact experimental details, including the specific mutations that allowed the wild-type A/H5N1 strains to travel between ferrets, would allow others to replicate these experiments.

Generally, authors of peer-reviewed scientific publications are required to reveal the details that would allow their experiments to be repeated in other labs — this level of detail keeps scientists honest and encourages the sharing of knowledge. The government panel wasn’t concerned about highly trained scientists replicating the experiments in their own, tightly regulated laboratories. Rather, they were concerned that terrorists might take advantage of the information to generate a deadly bioweapon — bird flu virulent and contagious enough to cause a pandemic as disastrous as the 1918 Spanish Influenza [4]. Perhaps, the panel recommended, the details could be released only to scientists with special security clearance to avoid this issue.

Proponents of the work argued that it was important to know which mutations allowed transmission so that future strains of the virus can be monitored for their pandemic-producing potential. For example, if public health officials were to identify a rash of birds infected with a strain of the virus that was only one mutation away from being able to jump between humans, that would serve as an alert to stockpile vaccines and flu medication in case the final mutation arose (however, whether officials, especially in developing countries, have the resources to actually carry out this kind of monitoring has been questioned). Furthermore, the research has been presented in sufficient detail at enough scientific conferences that even without publication in Science and Nature, the information would still be readily available to anyone determined to find them.

Ultimately, the World Health Organization, after convening its own expert panel meeting, issued a statement supporting uncensored publication of both papers. And after considering the fact that the mutant viruses may be less dangerous than natural ones, the NSABB reversed its decision on the publications [5]. For better or for worse, the complete experimental details from the two flu studies are now freely available for anyone to view [1,2]. While the controversy may be over for now, similar controversies are bound to arise again where the benefits and risks posed by particular scientific studies must be carefully assessed. Ultimately, all research has the potential to be misused. Society at large must continue to promote open discussion and meaningful deliberation on how we should balance academic freedom with other societal values and concerns.

Laura Strittmatter is a graduate student in Chemical Biology at Harvard University.


[1] Sander Herfst et. al., “Airborne Transmission of Influenza A/H5Na Virus Between Ferrets,” Science June 22, 2012 <>

[2] Masaki Imai et. al., “Experimental adaptation of an influenza H5 HA confers respiratory droplet transmission to a reassortant H5 HA/H1H1 virus in ferrets,” Nature May 2, 2012 <>

[3] Carl Zimmer, “The Evolution of Bird Flu, and the Race to Keep Up,” The New York Times June 25, 2012 <>

[4] Denise Grady and William J. Broad, “Seeing Terror Risk, U.S. Asks Journals to Cut Flu Study Facts,” The New York Times December 20, 2011 <>

[5] Denise Grady, “Panel Says Flu Research Is Safe to Publish,” The New York Times March 30, 2012 <>

Leave a Reply

Your email address will not be published. Required fields are marked *