Magazine

You Should Be Afraid of the Next ‘Lab Leak’

Listen to This Article

Audio Recording by Audm

To hear more audio stories from publications like The New York Times, download Audm for iPhone or Android.

The National Emerging Infectious Diseases Laboratories is a seven-story concrete fortress in Boston’s South End, hemmed in by Interstate 93 on one side and Boston University’s medical school on another. Unmarked by overt signs or logos, NEIDL — known as “needle” — is surrounded by a vast manicured lawn that would make for an excellent picnic spot, but for the high-spired steel fence and the constant surveillance by a police force stationed at nearby checkpoints. The safeguards ensure that passers-by will keep out. Yet almost everything else about the lab — its layout, its operations, its work protocols — is in service to an opposing imperative: keeping dangerous things, and especially very deadly things, in.

The building has one of the larger collections of Biosafety Level 4 and Biosafety Level 3 labs in the world. These kinds of facilities are where research on the planet’s most threatening pathogens takes place. Ebola, Lassa, Marburg viruses: All are designated for Level 4 work by the National Institutes of Health, meaning they are both transmissible and highly pathogenic, with few (or no) treatments for those who become infected with them. The pathogens studied in a Level 3 lab, like Mycobacterium tuberculosis or SARS-CoV-1, the predecessor to the novel coronavirus responsible for Covid-19, are slightly less lethal but still dangerous; the risks in this level of lab remain significant, but the safety level is a notch lower.

A secure biolab building is architecture for containment: boxes within boxes, each providing a hermetic boundary to prevent the escape of something risky inside. Within one of NEIDL’s Level 4 labs, for instance, researchers work on viruses in a glass safety cabinet, which they access through gloves. They also wear a hooded pressurized suit, fully zipped, that is connected at the small of their back to a hose that pumps in HEPA-filtered air drawn from outside the building. A negative-pressure air system prevents the escape of anything airborne; the room itself is situated off a corridor, entered by way of locked doors, accessible only to credentialed personnel who pass through a security barrier that scans their irises.

Ronald Corley, a microbiologist at Boston University and NEIDL’s director for the past seven years, showed me around the facility in mid-October. About 150 people work there. Studies were underway on Ebola, Lassa and the virus that causes Crimean Congo hemorrhagic fever. The lab has its own insectary, Corley remarked, which hatches mosquitoes to study the transmission of arboviruses, including Zika, dengue and West Nile. (The bugs are dewinged after infection, so they can’t fly away.) But the pandemic had impelled many researchers to shift their work to SARS-CoV-2, which for the past 18 months had become the facility’s main focus.

Outside one of the Level 4 suites, we watched through a window of ballistic glass as the researchers inside worked slowly and methodically in their puffed suits, like astronauts in low gravity. The concrete walls, Corley pointed out, were a foot thick. “And we’re actually looking into a different building,” he added. You wouldn’t know it from outside, but the Level 4 labs were in a building separate from but tucked within the larger NEIDL mother ship. In the event of an earthquake, the buildings could move independently; should a breach of containment happen, the entire Level 4 suite of labs could be evacuated, sealed and decontaminated with a noxious gas to kill an errant pathogen. “BSL-4s are all about multiple layers of protection, right?” Robert Davey, a microbiologist at Boston University, told me later. “It’s layer upon layer upon layer, so if one of them fails, which it should never do, but if it happens, you’re still OK.”

There is little doubt that buildings like NEIDL are among the most sophisticated and secure in the world. But whether the biolabs are safe enough — or safe as they could be — is more difficult to say. Even if a lab leak in Wuhan can’t explain the origins of the SARS-CoV-2 virus (a claim so far unproved and, according to many virologists, unlikely), the idea is nevertheless made plausible by the fact that breaches have occurred at other high-containment labs, including in China. In 2004, for instance, two lab researchers in Beijing became infected with SARS-CoV-1, which in turn spread to at least seven other people outside the lab.

In 2015, the N.I.H. commissioned a consulting firm called Gryphon Scientific to do a risk assessment of certain types of research at U.S. labs. The result, released in April 2016, was a thousand-page report that concluded, among other things, that experiments to improve the transmissibility of coronaviruses in a lab could “significantly” increase the chance of a pandemic “due to a laboratory accident.” Nonetheless, the report noted, “loss of containment” incidents for viruses are rare. And a breach would not necessarily mean an outbreak. Any accident, whether involving a natural or a lab-concocted “P.P.P.” — the term stands for a “pathogen with pandemic potential,” among the scariest possible lab escapees — would have a low chance of leading to a global crisis.

The Gryphon conclusions had to rely on estimates to fill in the gaps in the available evidence, however. Some analyses were based on meager data and others on analogous scenarios at nuclear power plants or chemical factories — facilities where the risks of bad outcomes might be small but the possible consequences catastrophic. When I asked Rocco Casagrande, the study’s lead author and a former United Nations weapons inspector, what we don’t know about the global risks of high-containment labs, he ticked off a long list. “Well, it’s still almost everything,” he said. “We don’t know how often incidents happen, or how often those incidents lead to exposures. We don’t know what factors are driving those exposures and incidents. We don’t know what features, like training or extra containment equipment or engineering controls, are effective at mitigating those incidents. So we have no idea if we’re spending way too much on biosafety or way too little or just the right amount. And if we should spend more, what should we spend it on?”

American biolabs, Casagrande stressed, are not necessarily unsafe. But because transparency is not required of all of them, they remain “a big black box,” where the related debate about risks and improvements exceeds our actual knowledge. To prevent the next pandemic, we might want to shine some light inside them. And for reasons that have as much to do with the labs, where the manipulation of pathogens sometimes enhances them, as with nature, where viruses increasingly pass from animals to humans, we might want to do so soon.

Credit…Illustration by Rachel Levit Ruiz

The world’s microbiologists and virologists now experiment with more types of dangerous viruses and bacteria than ever before, and they do so with far greater frequency. In a comparative historical sense, the work is being done in a more regulated — and probably more considered — manner. “A lot of people equate the presence of a laboratory with danger, instead of a way to safely work with organisms,” Gigi Gronvall, a biosecurity expert at Johns Hopkins University, told me. “But Louis Pasteur didn’t have a biosafety cabinet, and a lot of people died because they didn’t have these protections.”

In recent years, however, work that was once confined to a tiny number of facilities in the U.S. and Europe has expanded across the globe. This past summer, an academic study of Biosafety Level 4 labs around the world — whether run by governments, militaries, academic institutions or private companies — calculated that there are at least 59 in operation, under construction or in the planning stages, including about two dozen set up in just the past decade. Filippa Lentzos, an associate professor at King’s College London who helped conduct the study, told me that while most countries declare their Level 4 facilities to the U.N., there is no official international database keeping track of the labs and no requirement for governments to acknowledge their existence — either publicly or to the World Health Organization. “These labs are there so we can work with these pathogens in safe, secure ways,” Lentzos said. But a fundamental problem is that biocontainment facilities have discrete approaches to safety and risk assessment — and to transparency. “Right now,” Lentzos said, “these labs are spreading into other parts of the world, where you have different kinds of regimes, more authoritarian regimes, for instance, where the concept of openness is particularly challenging.”

In fact, the most concerning aspect about high-containment biolabs is that, considered as a collective, they may only be as safe as the worst lab among them: A breach or a breakdown at one could imperil us all. That risk is usually broken down into two categories. The first is biosafety: the effort to ensure through training and technology that workers stay free of infection and illness — not only for their own sake but also for that of surrounding communities. Then there is biosecurity, which focuses on the potential theft or misuse of dangerous biological agents. A related, more recent concern is cyberbiosecurity; as experimentation becomes more about manipulating data than samples in a wet lab, critical information — virus genomes, say — risks becoming subject to remote tampering. It’s one reason the newest high-containment labs have “air gapped” networks that are cut off from public internet traffic.

The international scientific community has worried about biolabs since they began proliferating two decades ago. But it has generally had difficulty moving beyond meetings or discussions of case-by-case incidents. So these labs remain separate and unequal. Some facilities, like NEIDL, have state-of-the-art technology and rigorous training regimens for workers; others have lower standards. Not all labs are in localities where strict policies govern the research (and handling) of emerging pathogens. And all labs may not be able to count on the ample, stable funding needed for maintenance and security. “We cannot quantify risk,” Lentzos said when I asked her to assess the world’s high-containment research. “We can only make qualitative judgments.”

Casagrande, for one, thinks the systematic collection of data would help. By his estimates, an agency like the National Institute for Occupational Safety and Health could spend as little as $10 million a year to better understand the risks in the U.S. He also suggests trying to answer basic questions that have so far gone unaddressed: What really occurs when, say, the contents of a flask are dropped and an infectious agent is released? What are the fluid dynamics of a spill?

Rebecca Moritz, who oversees biosafety operations at Colorado State University, told me that rigorous and standardized rules for training people who work in high containment, which are currently nonexistent, would be a large step forward, too. “We need to bring institutions to a minimal equal level,” she said. Like others I spoke with, Moritz also said it would be useful to create an anonymized database for “near misses” and other episodes, like one used for the airline industry.

By regulation, NEIDL discloses lab incidents to Boston’s health department as they occur; a quarterly report from B.U. typically lists two or three events — mouse bites, needle sticks, small tears in gloves or biosuits. “People hear the words ‘lab incident,’ and they automatically assume catastrophe, whereas 99.9 percent of the time, the risks of that incident were completely mitigated or actually were nothing at all,” Moritz said. “But then there are kind of these freak one-offs, sometimes that you couldn’t have anticipated.” These can be quite serious. A 2015 report in USA Today on lab safety unearthed documents showing that a lab at Texas A&M University repeatedly lost its negative-pressure air system in 2013. Another facility, at the Centers for Disease Control and Prevention, endured software failures. A number of labs were shown to have poor inventory records or had shipped toxins, like anthrax, that workers had mistakenly believed were inactivated.

The most unpredictable dangers presented by biolabs involve human error. While watching “hot work” on Ebola through a window at NEIDL, I was cautioned not to draw the researchers’ attention, lest they be startled and drop a flask or rip a glove. In the 2016 Gryphon report, Casagrande posited that an accident is 100 times as likely to result from human error as from mechanical failure. For this reason, the standardization of working practices could be one of the most effective improvements. At NEIDL, at least 100 hours of training are required before you can work in a Level 4 lab. This training begins in rooms built specifically for simulations; a full-time instructor there, described to me as “sadistic,” tests an employee’s panic levels when confronted with problems like a power failure or a torn suit. To get the chance to work with human pathogens at NEIDL’s labs, moreover, researchers need to be aware of how their lives intersect with safety considerations. “We try to have basically a no-fault policy for someone who has stressors at home, to ensure that they can opt out of working in containment,” Corley told me. Everyone at the laboratory, he said, receives an annual psychological assessment. They are also sent fake phishing emails to test their wariness of hackers.

When Robert Davey described a typical day at a Level 4 lab to me, he said he first must find someone to go with him — a “two-person rule” at NEIDL means he cannot be in the lab alone, in case he needs to be rescued. Once he and a partner reach the dressing room on the Level 4 floor, they check the gloves attached to their biosuits; if more than seven days have passed since the gloves were last used, they will have to change them, a task that can take 15 minutes. Next, he inspects his suit for holes. “Then I need to take all my clothes off, everything, and put on scrubs and socks,” Davey said. He carefully dons a pair of gloves, puts on the hooded suit with gloves attached to it and then adds yet another pair of gloves. After a hose from a ceiling pipe is connected to a valve on his suit — “you inflate like the Michelin Man,” he said — he passes through an air lock and into his lab. Being obsessive-compulsive in the Level 4 lab can be helpful, Davey said. He lines up his tasks in “nice little rows and columns within the safety cabinet” and constantly reminds himself not to cross one arm over the other and risk a spill. Lately, his work testing therapeutic drug candidates has been defined by a plate with 384 holes, each containing a live virus sample, usually SARS-CoV-2.

Researchers generally stay in a Level 4 lab no more than a few hours. The low humidity, the constant whoosh of air circulating around your head — the conditions can be fatiguing. When Davey completes his work, he and his partner must agree to leave together. Because talking is difficult, they motion to each other.

First they clean up, because there is no custodial staff in a Level 4 lab. Then, in an air lock and still suited, each will take a 10-minute shower that sprays them from all sides with a chemical detergent. When they’re finished, they move out of the sealed space, carefully remove their suits, gloves, scrubs and socks and take conventional showers. Then they get dressed. “Being a BSL-4 person, you never smell,” Davey said. “You’re always freshly washed.” But the safety protocols mean that completing an hour’s worth of work takes three or four hours.

I asked Davey, as well as Elke Mühlberger, another researcher at NEIDL, if they were ever fearful. Once they became comfortable with the pressurized suits, they said, they experienced a kind of joy in the “privileges” of the work, as well as confidence in containment measures. To Mühlberger, in fact, working in a Level 2 or Level 3 facility feels riskier than being in a Level 4 lab, where the safety protocol is so stringent; the day before she gave birth to her second son, she told me, she spent the morning working with the Ebola virus in a Level 4 lab. Once inside, there are no cellphones, no email, no small talk — only the pathogens and the white noise of air swirling around her ears. “It’s really very relaxing,” she said. Her work is focused on the planet’s most formidable threats, she acknowledged. But it is in many ways an escape from the world itself.

Is that world better off with or without high-containment biolabs? It’s a question not easily resolved. The work that goes on inside them involves a nontrivial degree of risk, which is why NEIDL, with its vaults and barricades and bulwarks — including its operational protocols — resembles a modern-day citadel. Yet no amount of engineering, infrastructural or human, can reduce to zero the chance of bad things coming out of biolabs. On the other hand, without them, we would lack all sorts of treatments for diseases like Covid-19 and Ebola. For now, the world seems to agree that we need these facilities.

Next summer, the C.D.C. will break ground on a new high-containment laboratory complex on its campus in Atlanta. One ambition is to supplement an aging biolab with a five-story, state-of-the-art facility that includes two Level 3 suites and six Level 4 suites. These will be largely dedicated to studying viruses with more fearsome fatality rates: Ebola, Nipah, Marburg, Chapare. Construction will take about three years, followed by a two-year commissioning process to ensure safety expectations are met. The cost has been reported to be at least $350 million — a significant jump from the $280 million (adjusted for inflation) that built the NEIDL facilities. Melissa Pearce, who will oversee the new lab, told me that she and her C.D.C. colleagues have toured North American facilities in recent years to survey current best practices and design ideas.

Ideas that are too new won’t necessarily be adopted. “When you’re designing a Biosafety Level 4, the thought of using new technology tends to give you pause,” Pearce told me. “It’s like the first year of a brand-new model of a car — you tend to not want to buy that, because there are probably some bugs that need to get worked out.” So, many of the improvements in Atlanta are likely to be incremental. Some of the researchers on the planning team believe that the spaces in current Level 4 labs are too narrow, for example, so there will be more room within new suites for workers to move around freely. A new chemical shower off the hallway will allow the staff to sanitize equipment more efficiently.

To talk to people at the C.D.C. is to be struck by how close to the next pandemic they think we might be — and how important, should a little-known infectious agent again explode in the general population, the research done on exotic viruses in containment there and elsewhere will be in directing us toward therapies or a cure. That’s the expectation at NEIDL, too, where Mühlberger has recently been working with the Lloviu virus, a relative of Ebola, which was first identified in bats in Eastern Europe 10 years ago. A group in rural Hungary extracts small amounts of blood from local bat colonies, searching for Lloviu. If the virus is present, the group sequences and sends the genetic information to her. She then compares its viral properties with other pathogens to better understand potential dangers. “We don’t know yet whether it causes disease in humans or not,” she said. “But if it causes disease, about 200 million people live in the area where these bats roam.”

When I asked Joel Montgomery, the head of the viral special pathogens branch at the C.D.C., whether our awareness of new pathogens is a result of improved surveillance or of more viruses having increased opportunities to jump into humans, he seemed to think both factors were responsible. The ability to test new viruses, thanks to nucleic-acid-sequencing capabilities, is far better than it was 10 or 20 years ago. “But I think we are interacting with our environment much more now than we have before, and just the sheer number of people on the planet has increased,” he said, which also affects population densities. “And so we’re going to see outbreaks — epidemics, pandemics — happening more frequently. It most certainly will happen.”

Our high-containment facilities, moreover, may have to deal with threats hatched in labs as well as what comes from nature. Take, for example, pox diseases. The C.D.C.’s campus in Atlanta is home to one of two Level 4 labs left in the world that harbors the live variola virus, which causes smallpox and was declared eradicated globally in 1980. (The other cache is in Russia.) Victoria Olson, a deputy director of lab science and safety at the C.D.C., told me that the lab keeps samples because studies using a live virus could help scientists develop diagnostics, treatments and vaccines should smallpox re-emerge, or should a similar poxvirus appear. Monkey pox, which has caused recent outbreaks in Africa, where it has a fatality rate of 10 percent, is already a serious concern; Alaska pox was just identified in 2015. More alarming, perhaps, is the potential that someone outside the world of known biolabs might cook up a version of a poxvirus, using the tools of genetic engineering. Smallpox had an average case-fatality rate of about 30 percent; Americans have not been immunized against it since 1972. A synthetic smallpox — or even a synthetic super smallpox, which could be deadlier than the original — is not much of an intellectual leap.

It’s a frightening notion, of course. But one premise behind biolabs is to be ready — ready to test new vaccines and therapeutics, ready to apply insights from old pathogens to new ones. And even in an age of vast computing power, there are no expectations — by either Corley at NEIDL or those I spoke with at the C.D.C. — that scientists will be able to make computational models as effective as the painstaking studies being done in Level 4 labs. That seems reason enough to keep striving to quantify the risks and improve the safety of the work being done there: If our containment research is not replaceable by digital simulations, and if our pathogenic enemies are real and growing in number, it may be best to keep them close when we can — to keep them in, that is, rather than out.


Jon Gertner has been writing about science and technology for the magazine since 2003. His most recent article examined how CO2 can be incorporated into products to make an impact on climate change.

Related Articles

Back to top button