Figure 16.1 Opening Pandora’s box …
Cartoon prepared by Elizabeth Stevens, Department of Biomedical Communications, St Jude Children’s Research Hospital, Memphis
The work of resurrecting the 1918 Spanish influenza virus demonstrated that scientists now have hitherto-unimagined power. We can change a virus’s RNA to make the virus multiply, or turn a non-pathogenic influenza virus (such as a virus from migratory birds) into one that can spread and kill. But should we? Is this not akin to opening Pandora’s box and letting the evil out?
The 1918 Spanish flu virus is considered to be one of the deadliest infectious agents in the history of mankind. It infected at least 30 per cent of humans globally and killed perhaps 5 per cent of those people. At the peak of the pandemic, cities throughout the world struggled to function and to pick up the dead. So, think what might occur if either of the two bird flus described in previous chapters developed the ability to spread from human to human. H5N1 kills about 60 per cent of humans infected, and H7N9 kills about 30 per cent of those infected.
Twenty years after it was first detected in Hong Kong, H5N1, the first bird flu virus, is still present in domestic poultry in China, Vietnam, Indonesia, Cambodia, Bangladesh and Egypt and periodically spreads to humans, mainly through live bird markets, causing sporadic outbreaks. The second bird flu virus, H7N9, first detected in Shanghai in 2013, is still confined to China. At the time of writing, H5N1 has infected 859 people, with 453 deaths; H7N9 has infected 1532 people, with 581 deaths. The latter virus possesses properties that might make it more likely than H5N1 to acquire the ability to transmit from human to human. To date, thankfully, it has not done so.
These high death rates in humans are difficult to interpret because so many people work in the poultry industry without apparently being infected. The key question here, then, is whether some people are especially susceptible to bird flu viruses. Is there a genetic susceptibility of a section of the human population? The question is not yet answered, but with more and more human genome analysis under way, it may soon be. I believe we will find that there are people who are genetically more susceptible to bird flu, but that the eternally mutating and reassorting influenza virus will potentially circumvent everyone’s defence mechanisms and, as in 1918, most people will become susceptible.
However, complacency about bird flu viruses and their supposed inability to spread from human to human is on the rise. ‘Since it hasn’t happened in 20 years, it is unlikely to occur’ is the general attitude. Some scientists also point out that in the known history of influenza, only three subtypes of influenza viruses – H1, H2 and H3 – have caused pandemics in humans, and that the problematic influenza waves seem to cycle through these three groups, so we don’t need to worry about any others.
Concerned by this complacency about the possible dangers of bird flu to humans, in 2006 the US National Institutes of Health Blue Ribbon Panel on Influenza Research and the WHO Influenza Research Agenda embarked on an investigation to find out whether the H5N1 virus could ever acquire the ability to spread from human to human. Their research would mean deliberately modifying the virus so that it would gain the ability to spread from animal to animal. Such experiments are referred to as ‘gain of function’ experiments. The specific goal was to create an H5N1 influenza virus that could spread between separated cages of ferrets.
Two groups of scientists set to work. One group, at the Erasmus Medical Centre in Rotterdam in the Netherlands, was led by Ron Fouchier; the other, at the University of Wisconsin, was led by Yoshihiro Kawaoka. The Fouchier group modified an avian H5N1 influenza virus isolated from a human infected in Indonesia in 2005 to make it transmissible between ferrets. They first caused the RNA of the virus to change by a process known as site-directed mutagenesis: they changed the virus’s genetic code so that it could multiply in mammals. Then they infected the ferrets through the nose with the virus. On the fourth day after infection, the virus from the first ferret was used to directly infect a second ferret, and this process was repeated 10 times (referred to as passages). In the end, the resulting virus spread by aerosol between ferrets in different cages.
Meanwhile, the Kawaoka group used a hybrid virus that contained the haemagglutinin protein of an H5N1 influenza virus isolated from a human infected with H5N1 bird flu in Vietnam in 2004, combined with seven gene segments from the H1N1 virus that caused the 2009 human pandemic. They first introduced random mutations into the genetic segments coding for the H5 component, then created the hybrid virus and finally infected the ferrets, again via the nose. This H5N1 hybrid virus was also shown to be transmissible by aerosol between ferrets in separate cages.
Each of these studies was carried out in high-security facilities under the same stringent guidelines and safeguards required for the work on resurrecting the 1918 influenza virus. The scientists involved were likewise vaccinated against the H5N1 virus and wore full protective gear.
These two studies clearly demonstrated that the H5N1 influenza viruses can acquire the ability to spread between ferrets. It is highly probable that such viruses would also be transmissible from human to human. The studies also showed that a small number of changes, perhaps as few as five, would permit the viruses to spread. It was apparent that transmissibility could come about either by changing an avian H5N1 influenza virus or by mixing genes from different viruses and producing hybrid H5N1 viruses.
This information sheds light on how pandemic influenza viruses might be generated in nature. The highly lethal 1918 Spanish influenza viruses may have been generated by exposure to the chemical agents used during World War I. Perhaps deadly mustard gas caused the key mutational changes in the original virus. High troop concentrations in the trenches would have set the scene for human-to-human transmission of the mutated virus. Repeated transmission from human to human is what the Fouchier group modelled with its 10 passages of their mutated virus in ferrets.
While the 1918 pandemic virus probably involved the transmission of a purely avian influenza virus to humans, subsequent pandemics – including the H2N2 Asian pandemic of 1957, the H3N2 Hong Kong pandemic of 1968 and the H1N1 pandemic of 2009 – all involved hybrid viruses that possessed novel haemagglutinin and neuraminidase components from an avian source but retained some gene segments from 1918 influenza viruses. The experiments by the Kawaoka group with the H5N1 virus adapted to ferrets mimics the genesis of these particular pandemic viruses.
Figure 16.1 Opening Pandora’s box …
Cartoon prepared by Elizabeth Stevens, Department of Biomedical Communications, St Jude Children’s Research Hospital, Memphis
When knowledge of the generation of a ferret-transmissible H5N1 became known to the scientific community, a firestorm erupted. On Monday, 12 September 2011, I had breakfast with Fouchier at the fourth European influenza conference for scientists in Malta. During the meal, he told me of their extremely interesting results and said he would present them in his keynote address that morning. I realised the huge impact the findings would have, and I remember sitting with Peter Doherty, the Nobel laureate immunologist, in the session and saying to him that the public would accuse scientists of having opened Pandora’s box (Figure 16.1).
Unfortunately some people got the idea from Fouchier’s presentation that the H5N1 virus not only was spread from ferret to ferret but that it had killed the ferrets. This was incorrect. The ferrets were killed in a separate study, when the H5N1 virus was injected directly into their windpipe (trachea). But the excitable science journalists in the audience associated transmissibility with death, considerably raising the level of public concern about generating such deadly viruses, and leading to articles in major newspapers about scientists engineering bioterrorism agents. The public firestorm was fuelling itself.
The results from studies by the Kawaoka group on ferret-transmissible H5N1 influenza viruses were released after they had been sent to scientific journals for publication. The studies confirmed the work of the Fouchier group: with only a small number of changes in the haemagglutinin component, H5N1 influenza viruses could be produced that spread from ferret to ferret. The Kawaoka group also pointed out that the changes required to make H5N1 transmissible were already present in the various H5N1 viruses that were causing disease in poultry and humans, but not all changes were present in the same virus. The question was, and is, how long will it take Mother Nature to put these five changes needed into the correct orientation in the same virus? While these studies raise legitimate concerns about bioterrorism and the accidental escape of killer viruses, they also alert the world that as long as these bird flu viruses are circulating, sooner or later nature will deal a deadly hand.
These two studies raised enormously important issues for the scientific community to consider. Is it appropriate to conduct this research, let alone publish the findings? In response to the public outcry, scientists working on influenza viruses called a voluntary moratorium on all gain-of-function studies on emerging viruses in 2011. The studies were considered ‘dual use research of concern’ (DURC). This meant that, on one hand, the studies on the production of ferret-transmissible H5N1 viruses had clear benefits to knowledge, such as improving our understanding of the potential for the virus to spread in humans and cause a pandemic, and they were important to the development of vaccines and antiviral medicines. But on the other hand, they raised serious concerns about the possibility of accidental release of the H5N1 viruses and of their nefarious use as a bioterrorism agent. The argument was that the blueprint for producing such a terrifying agent should not be published.
It was a difficult time for the NSABB for deciding whether to publish the scientific papers by the Fouchier and Kawaoka groups. They were strongly lobbied from both sides. The committee first erred on the side of caution and opted to publish them only after the removal of key information about the methods used, but after numerous meetings and consultations the papers were published in their entirety in leading journals – the Fouchier group’s report in Science,99 and the Kawaoka group’s report in Nature.100 The papers emphasised the high levels of biosafety and biosecurity enforced and the fact that the H5N1 influenza viruses containing the dangerous building blocks were already circulating in various parts in the world; the greatest threat was from Mother Nature.
After the publication of these papers, additional biosafety precautions were mandated, and the voluntary moratorium on gain-of-function studies was lifted in 2013. Several more scientific papers were published describing experiments increasing the biological activity of influenza viruses (i.e. gain of function), giving rise to more discussion of whether experiments with viruses like H5N1 should be permanently banned. Then two serious violations of biosecurity at the CDC in Atlanta – one involving the release of anthrax spores and a second involving the sending of an influenza virus culture contaminated with live H5N1 virus to another laboratory – set off alarm bells in the press and among public health scientists. Although neither of these events resulted in human infection or spread of the agents, they served as a wake-up call and pointed to a need for additional levels of control.
These incidents underlined that even the regulators of biosecurity and biosafety can make mistakes, and that many levels of security are necessary. A group of 18 leading scientists – the Cambridge Working Group – called for an immediate cessation of all gain-of-function studies and a thorough examination of all aspects of the control and regulation of dangerous infections like H5N1 influenza virus agents, known as select agents. The National Institutes of Health in the US immediately reimposed the moratorium on gain-of-function research, and the White House Office of Science and Technology Policy requested that the National Academy of Sciences, Engineering and Medicine and the NSABB consider all aspects of the problem.
Those bodies were tasked with organising two open discussions between scientists and the general public on the risks and benefits of gain-of-function research. The issue was to be fully debated, and participants were to make recommendations on the way forward. In recognition of the global significance of the issue, the leading scientific organisations from around the world were involved. I was one of many participants at both symposia. The first took place in December 2014, the second in March 2016. The process has been extremely thorough and detailed, involving participants from many countries in which the biological research under way could result in a disease agent’s gaining function.
It is worth considering the term ‘gain of function’, for it is often thought to refer only to functions that are of high risk to society, such as person-to-person transmissibility in the case of a virus. But the term also refers to highly desirable functions needed for the production of influenza vaccine. For example, when first isolated from humans, influenza viruses often grow poorly in chicken embryos. To make a virus multiply to the levels we need for large-scale vaccine production, we first make a hybrid virus that includes gene segments from known and safe, high-growing vaccine strains. We then ‘passage’ the hybrid viruses in the chicken embryo, creating far more virus. These studies clearly involve gain of function, but they do not pose a risk. The problem is with DURC gains of function, those that carry both scientific benefits and public health risks.
On 19 December 2017 the US National Institutes of Health lifted the funding pause on gain-of-function experiments involving influenza, SARS and Middle East Respiratory Syndrome that was implemented in October 2014. They also provided recommendations for the evaluation and oversight of the proposed gain-of-function research of concern, recommended by the NSABB. Before any research begins, the project must be assessed to determine whether it is in the DURC category. This involves a rigorous three-stage review and assessment. If the research is found to be of concern, it goes to a final panel that considers whether the proposed research meets all the guidelines of risk mitigation; only if it does is the study approved, and even then it must undergo national inspections. While the proposed recommendations apply to all research funded by the United States government they are not enforceable for privately funded research or in other countries. The hope is that involving the general public as well as international organisations will lead to similar guidelines being adopted.
These multiple levels of review minimise risk levels. Regardless, nothing in life is totally risk free. Scientists dealing with such issues must police themselves closely to make sure they follow the guidelines, cut no corners and institute buddy systems by which scientists supervise one another’s work. It is still unresolved who sent anthrax spores by mail to United States news offices and two United States senators, killing five people and injuring 17 others beginning on 18 September 2001. A scientist who worked at the biodefense laboratories at Fort Detrick was suspected and he committed suicide in July 2008. The motive was never resolved but this event amply illustrates the need for a buddy system to reduce such risks.
The only way to reduce the risk to near zero would be to recommend indefinitely retaining the gain-of-function moratorium on such studies. The reader may agree wholeheartedly with this solution to the problem, but in fact even this would not completely eliminate the risk. Pandora’s box is already open, and not every scientist in the world has to abide by the recommendations of United States scientific bodies.
Additionally, Mother Nature is doing her best to keep us very concerned about the spread of the second bird flu (H7N9) in China. There have been over 1000 reported human infections since the first one in 2013, with approximately 30 per cent of infected people now dead. The use of vaccine to H7N9 influenza in poultry in 2018 has dramatically reduced the number of human cases and the incidence of disease in poultry. To date this virus has not spread out of China, which may be explained by its absence in domestic and wild ducks. However, as long as the Asian H5Nx and H7Nx viruses continue to circulate they are a threat to animal and human health. Additional research is urgently needed to prepare new medicines and better vaccines. We cannot achieve this without continued gain-of-function studies, conducted in full compliance with all the guidelines.