8.
Gain of function
‘Nuclear weapons need large facilities, but genetic engineering can be done in a small lab. You can’t regulate every lab in the world. The danger is that either by accident or design, we create a virus that destroys us.’
STEPHEN HAWKING
Bird flu is a disease of wild birds that can transmit into poultry. One of the most lethal strains, known as influenza A H5N1, has periodically devastated poultry farms in Asia. In 1997, in Hong Kong, it began occasionally infecting human beings with an alarming mortality rate of 60 per cent. Fortunately, the H5N1 virus is usually caught directly from birds during the slaughter and handling of diseased birds or their carcasses. Sustained chains of person-to-person infection are uncommon, occurring only through very close contact. If that were to change, and the virus gained the ability to jump easily between people, it could cause a frightening pandemic.
How easy would it be for the virus to change so that it could spread efficiently between people? In May 2012, Dr Yoshihiro Kawaoka’s research group at the University of Wisconsin and the University of Tokyo published a paper reporting experiments in which they had enabled H5N1 viruses to spread through the air between mammals. A few weeks later, another team of scientists led by Dr Ron Fouchier at Erasmus University in Rotterdam published a similar paper with similar results. The news sent shockwaves through the scientific community because the research seemed to involve deliberately making viruses more dangerous to mammals, including human beings.
These were not rogue teams of researchers working in secret. Both studies had received funding from the US National Institutes of Health (NIH). The papers were published in prestigious scientific journals: one in Nature, the other in Science. Each manuscript had been submitted for peer review almost a year earlier in August 2011. The motive behind the work was good: scientists like Dr Kawaoka and Dr Fouchier wanted to know if and how H5N1 could mutate into a more contagious form among mammals and which emerging variants or mutations should raise the alarm. Given the number of high-density chicken farms in the world, there is a concern that the virus could cause widespread outbreaks and mutate to become more transmissible between people – similar to the seasonal flu – causing a devastating pandemic. At the time, there was no flu shot for H5N1 readily available for large populations.
In December 2011, Dr Fouchier told the New York Times, ‘There are highly respected virologists who thought until a few years ago that H5N1 could never become airborne between mammals . . . I wasn’t convinced. To prove these guys wrong, we needed to make a virus that is transmissible.’ So Dr Fouchier’s team deliberately encouraged the virus to become transmissible through the air in a secure laboratory (BSL-3+), by first generating mutations in a key gene (haemagglutinin protein, or HA, the influenza equivalent of the coronavirus spike protein) and then ‘passaging’ the virus through animals. Passaging is a common practice and one that, as we will describe, also played a role in various experiments with coronaviruses. The word simply means to infect a first batch of animals, then take the virus from that batch to infect a second batch of animals, repeating this process as many times as desired. The effect is to encourage the virus to evolve, by not-so-natural selection, to suit the new host species or to spread between hosts in a specific way. Both Dr Fouchier’s and Dr Kawaoka’s teams decided to use ferrets, which are easier to handle in the laboratory than monkeys and have an immune response to flu more similar to that of humans than mice do. In order to become more contagious in mammals, the virus would have to switch its attack from the virus entry receptor found in birds to the version of the same receptor found in mammals. Could it do so, and how easily could this happen?
Dr Fouchier’s team wrote: ‘We designed an experiment to force the virus to adapt to replication in the mammalian respiratory tract and to select virus variants by repeated passage.’ The final virus was sufficiently infectious that ferrets housed in cages near but physically separate from an infected animal started to fall ill. Dr Fouchier’s findings suggested that as few as five mutations could enable airborne transmission of the virus between mammals. H5N1 thus had the potential ‘to evolve directly to transmit by aerosol or respiratory droplets between mammals, without reassortment in any intermediate host, and thus pose a risk of becoming pandemic in humans’.
Ferrets were used in influenza virus gain-of-function research.
ITAR-TASS News Agency/Alamy Stock Photo
In contrast, Dr Kawaoka’s group did not largely rely on serial passaging to evolve a more transmissible variant of H5N1. Instead, they showed that they could introduce random mutations in the H5N1 HA gene; create chimeric viruses with genes from H1N1 swine flu and H5N1 bird flu; identify the best performing variants in the ‘library’ of mutants in host cells (at BSL-2; the scientists reasoned that they had replaced a HA cleavage site with a non-virulent sequence to attenuate the viruses so the work could be performed at this lower biosafety level); and then confirm that one of these variants was indeed efficiently transmissible by air in ferrets (at BSL-3+). Dr Kawaoka’s group also found that only three or four mutations were required to make the H5N1 HA capable of supporting efficient transmission in ferrets. In both approaches, scientists had harnessed techniques in the lab simulating and accelerating the process of evolution by artificial selection.
Before doing the experiments in negative-pressure safety cabinets in BSL-3+ laboratories, Dr Fouchier’s team took every possible precaution, acquired every requisite permit and welcomed inspectors not just from the Dutch government but from the US CDC. They even offered H5N1 influenza vaccines to the researchers and prepared quarantine hospital rooms. Dr Kawaoka’s lab also performed their work in BSL-3+ spaces approved by the US CDC and Department of Agriculture. Their facility had ongoing biosecurity monitoring and all personnel had undergone an assessment by the US Criminal Justice Information Services Division and completed rigorous biosafety and Select Agent training. Nothing went wrong. Rather, the reason for the shockwaves was that the experiments succeeded too well. It was too easy for scientists to create dangerous viruses that could spread through the air between animal models of human disease.
In late 2011, before publication, the media picked up on these radical pieces of work and amplified the concern of some scientists that such ‘gain of function’ experiments were too risky to allow. The phrase ‘gain of function’ typically refers to any experiment that results in a pathogen acquiring some new or improved capability, such as the ability to infect a new type of cell or a new host species. Why deliberately change a bird-transmissible virus so that it could transmit through the air among mammals? On 20 December 2011, the US National Science Advisory Board for Biosecurity, which advises the Department of Health and Human Services, recommended that both scientific teams and the editors of both journals should omit the methods and details of their work that could allow nefarious actors to develop dangerous pathogens. However, as a professor of pathology, microbiology and immunology at the University of California at Davis, Dr Nicole Baumgarth, pointed out: ‘If two research labs have done this already, nobody is going to stop a third and fourth lab from doing the same. These are routine procedures done in many labs around the world.’
Scientists debating the issue crystallised into two camps, one asserting that such experiments must continue to help the world prepare for pandemics, the other arguing that the experiments might themselves trigger pandemics and should cease. Citing the rapid emergence of the 2009 swine flu, Drs Francis Collins, Anthony Fauci and Gary Nabel from the NIH posited that ‘new data provide valuable insights that can inform influenza preparedness’ and that ‘the engineered viruses developed in the ferret experiments are maintained in high-security laboratories’. Their opinion was published in the Washington Post in December 2011, under the headline ‘A Flu Virus Risk Worth Taking’.
This optimism was not shared by some other scientists and journalists. ‘We cannot say there would be no benefits at all from studying the virus. We respect the researchers’ desire to protect public health,’ said a New York Times editorial on 7 January 2012, ‘but the consequences, should the virus escape, are too devastating to risk.’ The chairman of preventative medicine at Vanderbilt University School of Medicine in Nashville, Dr William Schaffner, said, ‘People in that lab need to have a careful discussion on how to keep that virus in the lab secure. Viral escape is quite real.’
By 20 January 2012, still before the results had been published, Dr Fouchier and Dr Kawaoka joined other scientists around the world in declaring a voluntary and temporary sixty-day moratorium on further experiments to generate highly pathogenic viruses that are more transmissible between mammals. ‘Despite the positive public health benefits these studies sought to provide, a perceived fear that the ferret-transmissible H5 HA viruses may escape from the laboratories has generated intense public debate in the media on the benefits and potential harm of this type of research,’ they wrote. ‘We realize that organizations and governments around the world need time to find the best solutions for opportunities and challenges that stem from the work.’
However, it was too late. This gain-of-function work was too tantalising, too novel, and had been published in high-profile scientific journals. Dr Baumgarth was right on the money. The very next year, a laboratory at the Harbin Veterinary Research Institute in China successfully mixed a duck isolate of H5N1 with the 2009 H1N1 flu virus, creating 127 recombinant viruses to show that the avian virus could become transmissible among mammals in hypothetical agricultural scenarios. The work was once again published in Science in June 2013. The ease with which laboratories around the world could design and engage in their own unique gain-of-function studies to create highly transmissible viruses was becoming clear. Some scientists claimed that these were possible natural trajectories of virus recombination or evolution that could occur on farms or in the wild without laboratory intervention. If true, how could one distinguish a lab-derived recombinant strain from a natural recombinant strain of virus?
An ineffective moratorium
Two years later, a run of three mistakes involving smallpox, anthrax and the bird flu virus in US laboratories poured oil on the fire of the gain-of-function debate. Live anthrax was accidentally shipped under conditions that exposed up to seventy-five workers to the bacteria. Several vials of forgotten 1950s-era smallpox were discovered in a cold-storage room of a Food and Drug Administration lab; two were tested and found to still contain live virus despite the predictions of smallpox experts. And the US CDC accidentally sent a highly virulent H5N1 influenza strain that can sicken humans (initially mistaken as a milder H9N2 strain) to another lab. In the last case, the CDC lab had been notified of the error on 23 May, but the leadership only got wind of it on 7 July, a delay the CDC chief, Dr Thomas Frieden, called very troubling: ‘I’m disappointed by what happened and frankly I’m angry about it.’
In response to these shocking incidents at a ‘superb laboratory’ (Dr Frieden’s words), the FDA and NIH searched their labs to check for more forgotten smallpox samples, and the CDC temporarily closed the anthrax and influenza labs that had shipped those dangerous pathogens, alongside halting all pathogen shipments from its BSL-3 and BSL-4 laboratories. A CDC deputy director was appointed to review and supervise biosafety at the CDC, and those who had violated protocols or reporting rules were disciplined.
Despite these measures, the accidents reignited fears about the safety at other labs working with dangerous or even enhanced human pathogens, such as those of Dr Kawaoka and Dr Fouchier. In July 2014, Dr Marc Lipsitch had organised the Cambridge Working Group, based in Cambridge, Massachusetts, to lobby for a moratorium on gain-of-function research. They described themselves as a ‘group of concerned scientists and experts in legal, ethical and other dimensions of Potential Pandemic Pathogen research’ formed to ‘enhance public understanding of the biosafety risks of such research’. The group pointed out that laboratory incidents had been growing in frequency, ‘occurring on average over twice a week with regulated pathogens in academic and government labs across the country’. If similar incidents were to occur with newly created, highly transmissible pathogens, the risks would be substantially increased, leading to outbreaks that could be impossible to control. Dr Lipsitch argued that gain-of-function experiments ‘unjustly require unconsenting populations to bear pandemic risk while promising them no realistic prospect of benefit’. Dr Steven Salzberg, an influenza researcher at Johns Hopkins University, agreed, adding that ‘the benefits of gain-of-function research are minimal at best’ and ‘could easily and far more safely be obtained through other avenues of research’.
By October 2014, the White House Office of Science and Technology Policy had announced a funding moratorium on gain-of-function research. Some scientists were surprised to hear that the pause on new federal funding applied not just to influenza but also to two other viruses: SARS and MERS. Dr Ralph Baric of the University of North Carolina at Chapel Hill, a legend in coronavirus research, was attending a daughter’s wedding that weekend and only heard about the moratorium when he returned to his laboratory on the Monday and opened his emails. It was devastating news. ‘It took me 10 seconds to realize that most of them were going to be affected,’ he said of his research projects. He wrote to the NIH arguing that ‘this decision will significantly inhibit our capacity to respond quickly and effectively to future outbreaks of SARS-like or MERS-like coronaviruses . . . Emerging coronaviruses in nature do not observe a mandated pause.’
Yet Dr Baric had himself already raised concerns about the kind of work in which he had made ground-breaking advances. In a lengthy 2006 review of reverse genetics (a method to understand the function of a new gene by altering it or inserting it into a different genetic context) experiments on viruses, entitled ‘Synthetic Viral Genomics: Risks and Benefits for Science and Society’, he detailed how a clever bioterrorist could create infectious pathogens undetected by any existing surveillance system: ‘It is conceivable that a bioterrorist could order genome portions from various synthesis facilities distributed in different countries throughout the world and then assemble an infectious genome without ever having access to the virus. To our knowledge, no international regulatory group reviews the body of synthetic DNAs ordered globally to determine if a highly pathogenic recombinant virus genome is being constructed.’
The coronavirus experiments
The coronavirus experiments that raised concern deserve close scrutiny if we are to understand what two laboratories in particular, Dr Baric’s in Chapel Hill, North Carolina, and Dr Shi’s in Wuhan, Hubei, were capable of doing with SARS-like viruses. Dr Baric had joined the University of North Carolina, Chapel Hill, in the 1980s to work on coronaviruses, then somewhat low-profile creatures that were responsible for some common colds but had never apparently caused serious outbreaks of human disease. In 1995, he published a book chapter about mouse hepatitis virus (MHV), arguing, in its title, that ‘coronaviruses may be potentially important emerging viruses’. He described experiments in which MHV gained the ability to infect the cells of hamsters, by using serial passaging, as described earlier. The virus could ‘rapidly alter its species specificity and infect rats and primates’.
After the epidemics of SARS, then MERS, coronavirus research was no longer backwater. Baric now held professorships in epidemiology, microbiology and immunology, and was an acknowledged world expert on the genomes of coronaviruses. His lab had a reputation for continually breaking new ground in manipulating coronavirus genomes. This included inserting new sequences into them using a ‘no-see’m’ method, patented in 2006, that Dr Baric named after a small biting insect sometimes encountered on beaches in North Carolina. The beauty of this technique, compared with traditional genetic engineering, was that it allowed a seamless insertion with no tell-tale ‘restriction site’ sequence attached because ‘the sites are removed during reassembly, leaving only the desired mutation in the final DNA product’. This is a vital point, given the argument often still made against laboratory-leak hypotheses for Covid-19, that restriction site ‘scars’ in its genome would be expected if genetic manipulation had occurred. Dr Baric’s no-see’m method was used from at least 2016 by the WIV scientists in some of their experiments and has since been enhanced by other seamless methods developed since 2006. In fact, once the SARS-CoV-2 genome was made public in January 2020, it took coronavirus experts in Europe and America just one to two months to separately devise and demonstrate reverse genetic systems to create synthetic SARS-CoV-2 genomes in the lab. The synthetic genome could have been used to create an infectious virus that would have left no signs of having been constructed in a laboratory.
In 2007, Dr Shi published a paper announcing her arrival at the forefront of this field of reverse genetics in coronaviruses. She and her colleagues used a ‘pseudovirus’ made from HIV, the virus that causes AIDS. Their aim was to test the ability of various spike receptor-binding domains found in SARS-like bat viruses to use the ACE2 receptors from a human, a civet or the horseshoe bat R. pearsonii. A pseudovirus has some of the attributes of a virus but is not capable of making more copies of itself even after entering a host cell; this makes the research safe because even if the virus were to leak, it cannot transmit from host to host.
The reason that Dr Shi’s group had to test the spike of a novel bat SARS-like virus in a pseudovirus was that they could not isolate the SARS-like virus itself. Often specimens from the wild will show evidence of the presence of a virus genome, but in such a degraded state that an infectious virus itself cannot be recovered. Using the pseudovirus system, they found that a bat SARS-like virus’s spike was ‘unable to use ACE2 proteins of different species for cell entry’. However, when they replaced just the receptor-binding domain of the bat virus spike with that of the 2003 SARS virus that had infected human beings, the hybrid protein had the ability of the 2003 SARS virus to use human ACE2 as a receptor. This indicated that this relatively short segment of the spike was both necessary and sufficient to convert a virus from bats into one that could infect human cells via the ACE2 receptor.
Remarkable though this experiment was, it was not altogether surprising. Seven years before, in 1999, researchers at Utrecht University in the Netherlands had done a similar reverse genetics experiment using cat and mouse viruses. They replaced the part of the spike that sticks out from the virus membrane in a mouse coronavirus, MHV, with the equivalent section from a cat coronavirus, called infectious peritonitis virus. The chimeric virus they thus created could infect cat cells but not mouse cells. The scientists inferred that the quality of the fit between the coronavirus spike and its entry receptor on host cells was what determined which host species or cell types a coronavirus can infect.
Scientists are sometimes collaborators, but they can also be competitors. The Baric group published an even more ground-breaking experiment in the same year, 2007, as Dr Shi’s team, though not involving bats. They manufactured five coronavirus strains with their genome backbones derived from the genome of a SARS virus. The group used one of the first SARS virus variants to have been sequenced in 2003, named after Carlo Urbani, a courageous WHO epidemiologist who died after treating some of the first patients and raising the alarm on the new infectious disease. But Dr Baric’s team replaced the spike with the equivalent from viruses extracted from infected palm civets, raccoon dogs, and the early, middle and late phases of the human SARS epidemic. They found that the human variants grew well in cultures of human ‘airway’ cell cultures, but the animal versions did not. This suggested that the animal versions are less pathogenic in humans but can evolve, while spreading among humans, to become more effective human pathogens. The reference to human airway cell cultures is a significant one. One of Dr Baric’s collaborators, Dr Raymond Pickles, had by now ingeniously managed to culture the cells that line the human respiratory tract, so that they formed a two-layered floating tissue complete with tiny waving hairs called cilia, to simulate the lining of the human lung. This was an ideal set-up in which to test the virulence of viruses in a model of a human being’s respiratory tract or lungs.
Another important new technique developed around this time was the creation of ‘humanised’ mice to accurately and rapidly study diseases such as SARS. In 2007, three separate teams of scientists succeeded in developing lines of healthy mice that expressed human ACE2 (hACE2) receptors to study the effect of SARS viruses. The reason was that SARS viruses do not infect mice easily because of an incompatibility with the mouse ACE2 receptor. The word ‘express’ in this context has a specific meaning: that a gene has been inserted and is actively being decoded to produce a protein. In most cases the mouse version of the ACE2 gene remains functional, but in some cases the human version replaces it. When infected with SARS virus, the hACE2 mice at the University of Iowa lost weight, became lethargic with laboured breathing and died within a week. By contrast, mice that did not express hACE2 displayed no clinical disease or mortality. The mice developed at the University of Texas Medical Branch in Galveston, Texas, exhibited ‘ruffled fur, lethargy, rapid and shallow breathing, and persistent weight loss’ when similarly infected. The hACE2 mice developed at the Institute of Laboratory Animal Science in Beijing suffered from severe lesions in the lungs alongside degeneration and necrosis in other organs including the brain. Almost a decade later, in 2016, Dr Baric’s team also developed a line of humanised mice and injected SARS into their noses, whereupon they ‘exhibited rapid weight loss and death between days 4 and 5’. These mice were shared with Dr Shi’s group at the WIV. After SARS-CoV-2 had emerged, it was revealed that a team in Beijing had been developing hACE2 mice in which the human ACE2 receptor replaced the mouse version of ACE2; in other words, this mouse was biologically as close to humans with regards to ACE2 as scientists could get. The scientists in Beijing inoculated their young and old hACE2 mice with the SARS-CoV-2 virus and found that the aged mice developed pneumonia.
Armed with these new techniques, the Baric lab could anticipate significant advances against coronaviruses. A seminal 2008 paper by Dr Baric and Dr Mark Denison of Vanderbilt University, and their colleagues achieved the feat of effectively creating a synthetic virus. At the time, no bat coronavirus had yet been successfully cultured in the laboratory, making it challenging to understand how coronaviruses were jumping species. Dr Denison’s team embarked upon a daring experiment. Using the sequences from the 2003 SARS virus and the four known bat SARS-like coronaviruses at the time, the scientists designed a hypothetical ‘consensus’ sequence, that is, a sequence that uses the most frequent letter at each position. Surprisingly, this worked. Chimeric viruses with spikes that were part SARS and part bat SARS-like virus were able to infect cells. This was a spectacular advance, and the team proudly announced ‘the design, synthesis, and recovery of the largest synthetic replicating life form’. They had laid the foundations for future generations of scientists to create infectious pathogens not found in nature by weaving natural virus sequences together, even without access to samples of actual, physical virus particles. The scientists were optimistic, claiming that their work represented ‘an approach for rapid recovery and testing of newly identified pathogens, and which may improve public health preparedness and intervention strategies against natural or intentional zoonotic-human epidemics’. Adopting their methods, scientists could more quickly and precisely understand what makes a new virus able to infect human cells, how it causes disease, and ultimately how to develop therapeutics or vaccines against these viruses.
In 2013, Dr Baric met Dr Shi at a scientific meeting and asked if she would share one of the new, unpublished virus sequences closely related to the original SARS virus that her group had discovered. He wanted to create and study a chimeric virus using the spike of the new sarbecovirus in the backbone of his laboratory mouse-adapted strain of the 2003 SARS virus. Dr Shi graciously shared the sequence with Dr Baric. The result was that during the moratorium on gain-of-function research, in 2015, Dr Baric published a landmark paper with Dr Shi showing that a SARS-like virus from Chinese horseshoe bats could potentially lead to an outbreak in humans. The reason that the two groups came to work together is clear. Baric had the greater expertise in synthesising coronavirus backbones and creating chimeras, but Shi had a growing library of bat coronavirus genomes collected during her collaborations with Dr Daszak’s EcoHealth Alliance. Unlike in 2007, this library now included some viruses that were closely related to SARS – R. sinicus bat viruses that had been collected in Yunnan in the years running up to 2015. In the words of Dr Daszak, describing this collaboration to another virologist in December 2019 before news of the Covid-19 pandemic had broken, ‘You can manipulate [coronaviruses] in the lab pretty easily . . . you can get the sequence [of the spike gene], you can build the protein, and we work with Ralph Baric at UNC to do this, insert into backbone of another virus.’
Dr Vineet Menachery, then a postdoctoral researcher in Baric’s lab, took the backbone of a SARS virus that had been slightly adapted to grow in mice and replaced its spike protein with that of a bat SARS-like virus sampled by Dr Shi’s group. The resulting chimeric virus could use human ACE2 to enter and infect human airway cells that had been cultured in the laboratory. When tested in mice, the virus caused severe damage in the lung, and was resistant to treatment by antibodies. It was also resistant to an early design of a vaccine against the SARS virus, in the form of an inactivated whole SARS virus – not a vaccine that was yet considered safe for use in the human population. Worryingly, the scientists found that, compared with having the spike of the 2003 SARS virus in this mouse-adapted SARS backbone, the bat spike in the same backbone showed a gain in virulence. The main conclusion they drew from the experiments was that bat viruses on their own were potentially dangerous. They did not need to be refracted through an intermediary species to start an epidemic. As the paper put it, ‘our results suggest that the starting materials required for SARS-like emergent strains are currently circulating in animal reservoirs.’ Dr Daszak said the study raised the profile of a wild virus ‘from a candidate emerging pathogen to a clear and present danger’.
Yet if the experiment underlined the risk of natural zoonosis, it also emphasised the risk of laboratory experiments on such viruses. The paper itself warned that ‘scientific review panels may deem similar studies building chimeric viruses based on circulating strains too risky to pursue, as increased pathogenicity [virulence] in mammalian models cannot be excluded’. It added: ‘The potential to prepare for and mitigate future outbreaks must be weighed against the risk of creating more dangerous pathogens.’ The work had been initiated before the funding moratorium and reviewed by the NIH, which funded and approved its continuation. It was nonetheless contentious. An article in Nature quoted virologists questioning whether the results of the experiment could justify the risks. Dr Simon Wain-Hobson, of the Pasteur Institute in Paris, said: ‘If the virus escaped, nobody could predict the trajectory.’ Dr Richard Ebright of Rutgers University said: ‘The only impact of this work is the creation, in a lab, of a new, non-natural risk.’
The work shifts to China
Several months before the gain-of-function moratorium began, the EcoHealth Alliance won a five-year, multi-million-dollar grant from the US National Institute of Allergy and Infectious Diseases (NIAID), with Dr Daszak as the principal investigator. Its purpose was ‘understanding the risk of bat coronavirus emergence’. The research would be carried out by the EcoHealth Alliance’s overseas partners, including the WIV: ‘Predictive models of host range (i.e. emergence potential) will be tested experimentally using reverse genetics, pseudovirus and receptor binding assays, and virus infection experiments across a range of cell cultures from different species and humanized mice.’ In all, according to the EcoHealth Alliance, the Wuhan institute received about $600,000 from this grant.
One project funded by this grant resulted in a 2017 paper in the journal PLoS Pathogens by Dr Ben Hu, one of Dr Shi’s doctoral students who had stayed on in the group as a researcher at the WIV. It described the creation of recombinant viruses that carried the spike genes of eight novel SARS-like coronaviruses sampled from bats, spliced into the genetic backbone of another SARS-like virus that the group had successfully isolated from bats and named WIV1. Two of these recombinant viruses (as did the natural WIV1 virus) proved capable of infecting and replicating in cells expressing human ACE2.
The paper summed up the achievements of the scientists’ visits to the Shitou cave near Kunming, Yunnan, yielding fifteen full-length genomes of SARS-like coronaviruses found in bats, as well as a total of three viable viruses isolated from the samples: WIV1, WIV16 and Rs4874. The genomes of WIV16 and Rs4874 were 99.9 per cent identical, and both came from samples collected on 21 July 2013, though WIV16 was published in 2016 and Rs4874 in 2017. The spike proteins of the two viruses were the most closely related yet to that of the 2002–3 SARS virus, with more than 97 per cent protein-sequence identity, and could use human ACE2 as an entry receptor. This discovery suggested to the scientists that some of the diverse SARS-like viruses found in that cave could be capable of transmitting directly from bats to humans. Based on the SARS-like viruses that had been found in the cave over the course of five years, 2011 to 2015, the authors hypothesised that the direct progenitor of the 2002–3 SARS virus may have arisen from recombination among the viruses found in that region in Yunnan. The scientists pointed out that the closest village to the cave was just over a kilometre away and, in 2003, there was a civet farm in Kunming that sold civets to Guangdong for consumption. The scientists did not know whether bats had transmitted the SARS virus to civets in Yunnan that were transported to Guangdong, or whether it was bats in Guangdong that had carried the virus and passed it to civets in the same province. However, they proposed continued monitoring of the SARS-like viruses to probe for and prevent spillover in these areas.
It is not clear if Dr Hu’s work, funded by the NIH, fell under the type of gain-of-function research for which new US federal funding had been paused. Dr Anthony Fauci, head of the NIAID, had helpfully defined gain-of-function research back in 2012 in a speech: ‘What historically investigators have done is to actually create gain-of-function by making mutations, passage/adaptation or other newer genetic techniques such as reverse genetics and genetic re-assortment.’ This did seem to describe Dr Hu’s work. Yet in 2021, in response to a question from US Senator Rand Paul, Dr Fauci categorically denied that NIH had funded gain-of-function research in Wuhan. Certainly, there is room for disagreement here depending on the definition of gain of function in use. In particular, the funding pause did not apply to the study of natural viruses unless there was a ‘reasonable expectation that these tests would increase transmissibility or pathogenicity’. In experiments in which only natural bat SARS-like viruses, not known to infect humans, were being recombined, it could be reasoned that the resulting viruses would not exhibit increased transmissibility or virulence in humans – after all, the scientists were only mixing and matching parts of bat viruses and not of human viruses. For instance, Dr Baric had argued that the mouse-adapted human SARS virus (from the 2003 epidemic) ‘actually makes the germs less able to infect human cells’ and are safer to work with in the laboratory. Bats are more distantly related to humans than mice are. The US government said that the point at which gain-of-function research must stop would be determined for individual grants via discussions between the scientists and their funding officers.
Aside from this confusion over what was considered gain-of-function research, the moratorium only targeted new funding. On funding that had already been granted, it merely called for a pause, meaning that there was no enforcement. This loophole was further enlarged by a footnote that read: ‘An exception from the research pause may be obtained if the head of the funding agency determines that the research is urgently necessary to protect the public health or national security.’ The NIH, responding in early 2021 to questions from Fox News, argued that the government-funded EcoHealth Alliance-WIV research did not breach the moratorium because the work ‘characterized the function of newly discovered bat spike proteins and naturally occurring pathogens and did not involve the enhancement of the pathogenicity or transmissibility of the viruses studied’.
In 2017, the moratorium ended under the Trump administration, and the NIH announced that funding of gain-of-function experiments involving influenza, MERS and SARS would resume under a new framework. Each research proposal would be assessed on a case-by-case basis by the Department of Health and Human Services. ‘If we are pursuing this research in an active way, we will be much better positioned to develop protection and countermeasures should something bad happen in another country,’ argued Dr Carrie Wolinetz of the NIH. Dr Lipsitch of the Cambridge Working Group disagreed: ‘I still do not believe a compelling argument has been made for why these studies are necessary from a public health point-of-view.’
Some scientists expressed concern that the wording of the new framework was too narrow, affecting only those research projects that were ‘reasonably anticipated to create, transfer, or use enhanced PPPs [potential pandemic pathogens]’. To count as an enhanced PPP, a pathogen had to be highly transmissible between humans, highly virulent and have resulted from an enhancement of either transmissibility or virulence. Notably, the new framework said that ‘Enhanced PPPs do not include naturally occurring pathogens that are circulating in or have been recovered from nature, regardless of their pandemic potential.’ The scientists were experimenting with viruses collected from bats, with some, albeit unknown, potential for transmitting to humans. Even if pandemic potential was suspected – and indeed repeatedly emphasised in numerous EcoHealth Alliance-WIV publications – the experts appointed by the NIH determined that this type of research fell outside the scope of the funding pause and the 2017 framework. They stressed that the EcoHealth Alliance application had been rigorously reviewed and ‘judged to be very high priority’ because of the emergence of the original SARS virus. In any case, this research would have almost certainly continued with Chinese government funding even if the NIAID grant had been paused. In today’s world of expensive research, $600,000 of funding over five years is small beer. A top laboratory in China such as Dr Shi’s group at the WIV, holding at any time numerous such grants and sharing funding with other top labs, would not have folded if a measly $120,000 did not make its way from the US into its coffers each year.
The vaccine hope
What was the point of all this virus hunting and genetic manipulation? By the end of 2017, the Wuhan scientists led by Dr Shi had achieved most of their original goals on the trail of SARS-like viruses. They had tracked down the likely bat origin of the 2003 SARS virus in Yunnan, although it was still a mystery how the virus had made its way to Guangdong. They had isolated three viruses from bat samples collected up to 2013. They had made chimeric viruses to test the infection abilities of diverse spikes in SARS-like viruses found in the wild. They had challenged these viruses against human cells and humanised mice. They had gathered huge amounts of data on the pandemic threats that these SARS-like viruses presented, indicating a probable spillover zone in southern China where people who handled wild animals or lived near bat caves were found to carry antibodies against SARS viruses. In addition to assessing the risk from viruses, the obvious priority now was surely to use this knowledge to prevent a pandemic occurring.
In this area in recent years a new word has begun to crop up with growing frequency: vaccine. The dream of designing a vaccine against all SARS-like viruses was very much within the ambitions of those studying SARS viruses. In this way, people living in areas prone to the spillover of SARS viruses could be vaccinated to pre-empt an outbreak of novel coronaviruses. As early as 2008 Dr Baric and Dr Denison had declared that ‘to protect against future emerging zoonotic pathogens, it is crucial to develop cell culture and animal models to test vaccines and therapeutics, ideally against entire families of organisms, such as [coronaviruses]’. In 2019, one of the reasons Dr Daszak cited for the laboratory work was to help develop a broad-spectrum vaccine against new coronaviruses: ‘The logical progression for vaccines is, if you’re going to develop a vaccine for SARS, people are going to use pandemic SARS [that is, the virus that caused the 2003 epidemic] . . . but let’s try and insert some of these other related [viruses] and get a better vaccine.’
So was a vaccine against SARS viruses on the horizon? It is not mentioned in any of the key grant programmes. The National Science Foundation of China announced a new grant to support continuing work on the ‘Evolution mechanism of the adaptation of bat SARS-related coronaviruses to host receptor molecules and the risk of interspecies infection’. It was to run for three years from the start of 2018 with Dr Shi as the principal investigator. That same year, Dr Hu in Dr Shi’s lab received a further 250,000 yuan grant under the WIV’s Youth Science Fund beginning in January 2019 to study the ‘pathogenicity of two new bat SARS-related coronaviruses to transgenic mice expressed ACE2’. This probably refers to WIV1 and either one of WIV16 or Rs4874 (near identical viruses), which had been isolated and characterised in cell cultures, but not yet in laboratory animals. Incidentally, Francisco de Ribera (a tenacious data sleuth who we will meet later in the book) has been able to identify and account for all the published viruses isolated at the WIV, from WIV1 to WIV19, with one exception. He could find no published account of WIV15. When details of WIV6 were published in June 2021, Ribera commented: ‘We found WIV6. A little bit strange, but unrelated with SARS-like viruses . . . This leaves WIV15 as the only missing live isolate. It was always the most suspicious due to the range of dates in which it should have been isolated.’ Based on the date range, Ribera speculated that WIV15 might have possibly been isolated from a Mojiang sample.
In 2018, a further large five-year grant was announced by the Chinese Academy of Sciences under a ‘Pathogen host adaptation and immune intervention’ special project. This project focused on discovering new viruses in animals, as well as humans, afflicted with unknown pathogens, understanding how these viruses could jump between species and cause disease, and how they were able to evolve to defeat both the immune system and treatments. Using this knowledge, the scientists aimed to develop and test novel antibody therapy and vaccines against emerging infectious diseases. They would enlist cutting-edge technologies, including human organs-on-a-chip and humanised mice, as well as in-laboratory directed evolution. There were five sub-projects in total. Dr Shi co-led the first alongside a different Dr Shi. And it seems that Dr Shi’s own group was also to receive funding from this strategic priority research programme of the Chinese Academy of Sciences (grant number XDB29010101) valued at about $1.35 million, from July 2018 to June 2023.
Thus, in the year preceding the Covid-19 outbreak, research was ramping up in China, and was clearly aimed at comprehensively cataloguing natural viruses with pandemic potential, including SARS-like viruses, and devising vaccines against them. With the impressive track record of Dr Shi’s group, it seemed inevitable that vaccines and therapeutics against SARS viruses would soon be developed, giving scientists the upper hand against emerging coronaviruses.
In America, in October 2018, Dr Baric’s group published a paper suggesting how to design a live vaccine that would work against SARS viruses. Briefly, vaccines work by raising immunity against real infections and to do this they are either inert, dead versions of (usually parts of) a virus or they are ‘live’ but attenuated viruses: still capable of causing an infection, but a very mild one. Live vaccines have been used successfully against measles, mumps, rubella, yellow fever and chickenpox. The power of live vaccines is that, although the virus is weakened, they cause an infection similar to the natural virus and can stimulate strong immunity against it that can last a lifetime. The problem with live vaccines is that sometimes they revert to the untamed form, usually via recombination with a natural, closely related virus that happens to infect the vaccinated person or animal. For instance, sporadic cases of vaccine-derived poliovirus have occurred when the weakened poliovirus used in the oral vaccine continues to circulate in a particular population for a long period of time, allowing it, through mutation, to regain its ability to cause severe disease. For this reason, the US only uses the inactivated polio vaccine and not the live attenuated virus vaccine.
Live attenuated African swine fever vaccines sometimes mutate back into a pathogenic form. In early 2021, reports began to surface that some of the recent African swine fever outbreaks in China had been caused by unlicensed vaccines being administered to large numbers of pigs. However, due to the issue’s political sensitivity, ‘reporting of the recent African swine fever outbreaks was extensively covered up’, according to Channel News Asia. These attenuated viruses are difficult to detect and can cause long-term and widespread infections across pig farms, with the potential to devastate the pork industry in China – particularly because some farms feed their pigs with kitchen waste that may still carry live viruses.
Dr Baric’s team was trying to get around this limitation of live vaccines by making sure that their live attenuated SARS virus would not recombine with other SARS-like viruses and revert to virulence. They were going to do it by rewriting a small section of special text in the genome of the virus. This text, known as a transcription regulatory sequence (TRS), triggers the expression of genes. The team showed that within each of the nine regulatory sequences of the SARS genome there was a six-letter text that was the same – ACGAAC – and if they replaced this in each case with a slightly longer and markedly different text – UGGUCGC – they had effectively changed the password on the whole network. This attenuated the virus, weakening its virulence by messing up the efficiency of its gene transcription. Encouragingly, Dr Baric and his colleagues found that inoculating aged mice with this TRS mutant could protect against a lethal SARS virus challenge. This change in TRSs also meant that an attenuated version of the virus would not revert to virulence if it found itself in the same cell as a virulent version of the virus since its ability to start swapping sequences and repairing faulty parts had been compromised. Thus, they had made a weakened and recombination-resistant virus as ‘a candidate strategy for a broadly applicable, rapidly implementable CoV vaccine platform’. There is a note of hope, even triumph, in the final words of the paper’s discussion: ‘This attenuation strategy . . . could bring live-attenuated CoV vaccines within the reach of realization in the face of the ever-growing threat of new human and animal CoV-based epidemics.’ Sadly, this technology was not developed in time to contribute to ending the Covid-19 pandemic. It would be other kinds of novel vaccines, messenger-RNA (mRNA) vaccines and recombinant vector vaccines, rather than live attenuated vaccines, that would come to the rescue of humanity during 2021.
Alongside raising hopes for a vaccine, the research in both Dr Baric’s and Dr Shi’s laboratories held a darker promise, as a demonstration of how to synthesise novel viruses. Once other labs saw that it was possible to stitch together sequences from different SARS-like viruses from bats and bring the final chimera to life, they soon began to copy these techniques or invent their own unique approaches. The public sees only the research that scientists decide to publish, not the projects that fail, that remain unfinished, or that they choose to keep secret. This is particularly troubling in recent years when it has become possible for labs to install their own gene-synthesis machines that print out any sequence. Enforcing worldwide surveillance of the gene synthesis performed by each of these machines is impossible.
Yet it is important to note that it remains impossible even today to magically conjure up a virus based on no sequences. Access to virus samples is not needed, but scientists invariably have to use the sequences of natural virus genomes as a starting point for their genetic designs. For this reason, it can be difficult to distinguish a natural from a synthetic virus without knowing what natural viruses could have been used as a template and what laboratory techniques had been deployed. In the absence of both pieces of knowledge, it is just about impossible to know from the sequence alone whether a given genome is a product of natural recombination or of laboratory engineering.
A whistleblower emerges
The idea that these techniques could have been used to make bioweapons, and that Covid-19 might have been caused by such a weapon, was especially inflamed when, in April 2020, a young Chinese scientist-turned-whistleblower fled to the United States in fear of being ‘disappeared’ in China, and began to give interviews about the early events of the pandemic. Dr Limeng Yan had previously worked on influenza vaccines and is a co-first author on a Nature paper about how SARS-CoV-2 causes disease in hamsters. She blew the whistle on two of the senior authors of the paper, one of whom was her supervisor, Dr Leo Poon, claiming that they had known about the human-to-human transmissibility of SARS-CoV-2 but failed to relay this crucial information to the WHO and the world in a timely manner. According to stories on Fox News (July 2020) and in the New York Times (November 2020), the saga unfolded as follows. The Hong Kong laboratory in which Dr Yan worked was a WHO reference laboratory, one required to inform the WHO of developments concerning influenza viruses and pandemics. In December 2019, Dr Poon asked Dr Yan to look into the SARS-like cases that were being reported in Wuhan. Dr Yan, who had trained as a doctor in mainland China, tapped into her network in various medical facilities across the country. On 31 December, a friend at the Chinese CDC told her that the virus was transmitting from person to person, weeks before China or the WHO confirmed human-to-human transmission, and she reported this to Dr Poon. However, by early January, her contacts were going silent. The doctors said: ‘We can’t talk about it, but we need to wear masks.’ Dr Yan again reported this to Dr Poon on 16 January, but was told ‘to keep silent, and be careful’. By 19 January, Dr Yan had leaked this information to Lu De, a YouTuber and China critic, who told his hundred thousand followers that the novel coronavirus was transmissible between people. The very next day, the Chinese National Health Commission (and the WHO in turn) confirmed human-to-human transmission of the virus. In his video, Lu De added that the virus had been deliberately engineered and released by the Chinese government, pointing to research on bat SARS-like viruses in Zhoushan, in Zhejiang province, published in September 2018 by scientists at the Third Military Medical University, Chongqing, and the Research Institute for Medicine of Nanjing Command.
Dr Yan’s escape story revealed a problem that one of us (Alina) pointed out on Twitter: ‘If there is one thing that this entire saga has made clear – it is that whistleblowers (as it pertains to SARS-CoV-2) have no obvious safe route of sharing their information.’ When she fled to America, Dr Yan had to rely on the contacts of Lu De, an anti-Chinese Communist Party billionaire named Miles Guo and his friend Steve Bannon, ex-advisor to Donald Trump. (Steve Bannon was arrested on Guo’s yacht in August 2020 for allegedly defrauding donors in a crowdfunding campaign to build the US border wall with Mexico; he pleaded not guilty.) According to the New York Times, Dr Yan was packaged by Guo and Bannon as ‘a whistleblower they could sell to the American public’, and Guo recounted on his show telling Dr Yan: ‘Don’t link yourself to Bannon, don’t link yourself to [Miles] Guo Wengui . . . Once you mention us, those American extreme leftists will attack and say you have a political agenda.’ Dr Yan gave a series of interviews to right-wing media in the United States, deviating far from her original whistleblower report about human-to-human transmission of the virus to talk about topics such as hydroxychloroquine treatment, the controversial anti-malarial drug championed by Donald Trump as a treatment for Covid-19. As Alina pointed out on Twitter, the Yan episode showed how easily a whistleblower can be tainted by their dependency on their host or saviour, someone who they now have to rely on for security for the rest of their life.
In early September, a meeting was arranged between Dr Yan and Dr Daniel Lucey, an infectious diseases expert from Georgetown University. In a blog post at the end of June 2020, Dr Lucey had himself published questions about whether the virus might have emerged from a laboratory. According to the New York Times, Dr Lucey ‘said that Dr. Yan appeared to genuinely believe that the virus had been weaponized but struggled to explain why’. Undeterred, Dr Yan published a lengthy preprint that month, setting out her claims that SARS-CoV-2 had been created in a laboratory. The day after publication, on 15 September, she went on Tucker Carlson Tonight, one of the most popular Fox News shows. Asked if she believed that the Chinese government had released the virus on purpose, Dr Yan replied: ‘Yes, of course, it’s intentionally.’ Inevitably, Dr Yan’s report and interview were harshly criticised and debunked by scientists.
Alina, as well as other scientists, soon found numerous problems with the scientific claims in the Yan preprint. Most crucially, the Zhoushan viruses found by the Third Military Medical University, Chongqing, on which the entire report hinged were more than three thousand mutations away from SARS-CoV-2 – therefore they were highly unlikely to have been the template virus genome for SARS-CoV-2. This destroyed the report’s credibility, and had diverted attention away from RaTG13, the Mojiang miners and the missing WIV virus database. Within a month, Dr Yan had released another preprint, this time titled ‘SARS-CoV-2 Is an Unrestricted Bioweapon: A Truth Revealed through Uncovering a Large-Scale, Organized Scientific Fraud’. This one did not mince words: ‘Records indicate that the unleashing of this weaponized pathogen should have been intentional rather than accidental. We therefore define SARS-CoV-2 as an Unrestricted Bioweapon and the current pandemic a result of Unrestricted Biowarfare.’
Despite the drama surrounding these preprints, the possibility cannot be dismissed that laboratories in America, China and elsewhere have been developing cutting-edge genetic-manipulation technologies, partly to be ready to fend off their use by terrorists, but also to use them against their enemies if necessary. ‘The technology to synthetically reconstruct genomes is fairly straightforward and will be used, if not by the United States, then by other nations throughout the world,’ Ralph Baric had warned in 2006 – and technology has advanced considerably since then. In 2021, the US State Department published a statement that included this claim: ‘The United States has determined that the WIV has collaborated on publications and secret projects with China’s military. The WIV has engaged in classified research, including laboratory animal experiments, on behalf of the Chinese military since at least 2017.’
However, at this stage, in the absence of harder evidence, we think allegations that SARS-CoV-2 is a bioweapon or a vaccine trial that went wrong are a distraction. If the virus came from a laboratory, it is much more likely that it was a leak from experiments designed to understand viruses that pose potential pandemic threats. The SARS-CoV-2 virus does not appear to have the features of an attenuated vaccine. If anything, it seems to have features that promote increased transmissibility, virulence and ability to evade the immune system, and it would have made no sense to release a bioweapon in the very city in which a world-renowned lab was performing such similar research, if the plan was to do so without raising suspicions.