Barbara Han, a disease ecologist at the Cary Institute of Ecosystem Studies, knew it was a question of when, not if, the coronavirus would spread to animals. As the first reports of infected animals appeared in 2020, she began working on an artificial intelligence model that would predict which creatures might be next.
“We had a pretty lofty goal of being able to predict exactly which species we should be keeping an eye on, given that we think it’s going to spill back,” Dr. Han said. As her team worked, the trickle of cases in new species became a flood: cats and dogs in homes and mink on farms. The virus infiltrated zoos, infecting the usual suspects (tigers and lions) as well as more surprising species (the coatimundi, which is native to the Americas and resembles a raccoon crossed with a lemur, and the binturong, which is native to Southeast Asia and resembles a raccoon crossed with an elderly man).
Dr. Han and her colleagues ultimately identified 540 mammalian species that were most likely to host and spread the coronavirus. She had been especially worried that the red TAUT, which ranked high on her list of at-risk creatures, and is widespread in Europe and North America, would be susceptible to the virus. “We’re just waiting for somebody to report it,” she said.
Just days earlier, in fact, researchers in Colorado had announced that the virus had proved capable of infecting red foxes in the lab. “Oh no!” Dr. Han exclaimed when informed of the finding. “It really sucks to be right in my line of work.”
Last fall, scientists analyzing tissue samples from dead white-tailed deer in Iowa found that the virus was widespread in that species. The discovery intensified concerns that the virus might establish itself in an animal reservoir, mutate and spread to other species, including back to humans. It also opened a rabbit hole: If deer can silently spread the coronavirus, what else could? And what else will?
Experts say there is no need to panic, and emphasize that animals are not to blame. “Really, humans are infecting the animals, and now animals are sick and some of them are dying,” Dr. Han said.
But identifying the species at risk is crucial for protecting both human and animal health. It is also a formidable scientific problem, with a wide array of potentially vulnerable species. Scientists must analyze a constant, chaotic stream of computational predictions, laboratory data and confirmed infections in zoos, homes and the wild.
In an ideal world, scientists would monitor every potentially susceptible population. But in the real one, they are trying to strike a delicate balance between identifying the species of highest concern and casting a wide net as the virus mutates and variants emerge. “It wouldn’t surprise me if you would find an animal species or an animal reservoir that nobody has thought about,” Dr. Diego Diel, a virologist at Cornell University, said.
The basics of infection
Scientists use a variety of tools to identify susceptible species. Each approach has limitations, but together they paint a fuller picture of which animals are at risk.
Some research teams are focusing on the ACE2 receptor, a protein found on the surface of the cells of many species. The coronavirus’s spiky protrusions allow it to bind to these receptors, like a key in a lock, and enter cells.
In 2020, a group of scientists compared the ACE2 receptors of hundreds of vertebrates, mostly mammals, with those of humans to determine which species the virus might infect. (The ACE2 receptors of birds, reptiles, fish and amphibians are not similar enough to ours to raise concern.)
“The predictions have been very good so far,” Harris A. Lewin, a biologist at the University of California, Davis, and an author of the study, said in an email. The scientists predicted, for instance, that white-tailed deer were at high risk for infection.
But some predictions proved entirely wrong: The paper identified farmed mink as a species of “very low” concern — and then in April 2020 the virus raged through mink farms.
Indeed, ACE2 offers only a snapshot of susceptibility. “Viral infection and immunity is much more complex than just a virus binding to a cell,” Kaitlin Sawatzki, a virologist at Tufts University, said in an email.
And of the world’s nearly 6,000 mammalian species, scientists have sequenced the ACE2 receptors of just a few hundred of them, creating a biased data set. These sequenced species include model organisms used in experiments, species that carry other diseases, and charismatic zoo denizens, not necessarily the animals that people are most likely to encounter.
“If a pandemic were to have arisen from a squirrel, we would be like, ‘God, what’s wrong with us? We didn’t even measure the basic biology of a squirrel,’” Dr. Han said.
So scientists have to find creative ways to make predictions for animals whose ACE2 sequences remain unknown. ACE2 sequences play a crucial role in basic biological functions, such as regulating blood pressure. By collecting a species’ basic life history details — such as what it eats and whether it is nocturnal — Dr. Han’s team trained a machine learning algorithm to identify those that appeared likely to bind to and transmit the virus, allowing them to predict susceptibility across many more mammals.
Scientists can test these computational predictions in the lab by trying to infect animal cells or live animals with the virus. Such experiments can differentiate species that may seem similar; one study found that deer mice could be infected with — and shed — the original version of the virus, while house mice could not.
But what happens in a collection of cells does not always occur in real animals, and what happens in a lab, where animals typically receive high doses of the virus, may not reflect real life. For instance, although the original virus can replicate in pig cell lines, actual pigs do not appear to be highly susceptible, researchers found.
To learn whether animals have been infected by the virus in the real world, scientists can perform what are known as serology studies, looking for coronavirus antibodies in their blood. “Serology helps us to look at the historical exposure,” Dr. Suresh Kuchipudi, a veterinary microbiologist at Penn State, said.
The discovery of widespread antibodies in white-tailed deer set off scientific alarm bells because it indicated that many of the animals had already been infected by the virus . It prompted researchers to look for active infections in the cervids, which they soon found.
But sampling and swabbing free-ranging animals can be difficult and time-consuming. So the U.S. Department of Agriculture, which received $300 million under the American Rescue Plan to conduct disease surveillance in animals, is now asking zoos, aquariums and wildlife facilities to send in blood samples, which will be analyzed for coronavirus antibodies.
And researchers at Tufts, including Dr. Sawatzki, have enlisted wildlife rehabilitation specialists to swab an eclectic collection of creatures, including black bears, bobcats and hundreds of bats. (Bat rehabilitators often submit guano samples instead of oral swabs, which can be difficult to obtain from the animals. “They have very tiny little mouths,” Dr. Sawatzki said.) So far, all have been negative.
Bats have been a source of concern because they are reservoirs for other coronaviruses, and many scientists believe that SARS-CoV-2 initially emerged from bats. But bat species are incredibly diverse, and not all of them appear to be susceptible to the virus — a reminder that animals of highest concern may not be intuitive, scientists said.
Complicating matters, the virus is not static, and animals that resisted infection with past variants might be vulnerable to new ones. For example, lab mice that were not susceptible to the original coronavirus or to the Delta variant were susceptible to Beta and Gamma.
“That’s the problem with emerging diseases,” said Dr. Scott Weese, an infectious diseases veterinarian at the University of Guelph in Ontario. “You have to keep resetting your knowledge every time something changes,” he added.
A shortlist of species
Biological susceptibility is just one piece of the puzzle; whether or not a species becomes a reservoir depends on a constellation of factors. “It depends on their social behavior, the immune response that’s mounted by the animals, the population size, the kind of connection with different populations of animals,” said Dr. Keith Hamilton, head of the preparedness and resilience department at the World Organization for Animal Health.
For a virus that is overwhelmingly transmitted by humans, a species’s relationship with us matters tremendously. Although narwhals’ ACE2 receptors technically place them at “high risk” for infection, they are not likely to run into us. Still, risk isn’t zero for marine mammals, especially captive ones: In 2006, a human likely transmitted MRSA to a bottlenose dolphin in a marine park in North America.
And the risk to pets is manifest.“We’ve heard stories of dogs getting infected from people sharing food and letting them lick off their plates when they were sick,” said Dr. Casey Barton Behravesh, who directs the One Health Office at the Centers for Disease Control and Prevention, which created a national repository fordata on coronavirus cases in animals. “Or even drinking out of toilets.”
Pet dogs, cats and hamsters can all be infected by the virus. Hamsters in a Hong Kong pet store most likely infected two people, leading to a contentious hamster cull.
But we are far more likely to infect our pets than they are to infect us, and many of these infections will be dead ends, scientists predicted. Infectious pets can also be isolated. “Your hamster at home that you may have bought some time ago is not a high risk to you,” Dr. Hamilton said.
A larger concern, scientists said, are the “peridomestic” species that live alongside us but roam freely; in North America, these include deer mice, red foxes and feral cats. These animals may act as a bridge between humans and wild populations, spreading the virus to species we may not encounter. And rodents, which are reservoirs for other pathogens, “should be definitely on the top of the list,” Dr. Kuchipudi said.
To monitor this potential threat, officials from the U.S.D.A and other agencies are looking for signs of the virus in some of these animals — including rodents, skunks, foxes and opossums — that live in and around zoos, wildlife facilities and mink farms.
Globally, certain threatened species are also a top concern. Three snow leopards in a Nebraska zoo died after contracting the coronavirus, and a wild leopard cub in India was found to be infected.
And great apes, which frequently encounter tourists and researchers, are vulnerable to other respiratory viruses. “Great apes are uniquely susceptible to human pathogens, because we’re closely genetically related,” said Dr. Kirsten Gilardi, a wildlife veterinarian at the University of California, Davis.
So far, no coronavirus infections have been reported in wild apes, but researchers are monitoring the animals closely, collecting fecal samples from those with respiratory illnesses.
‘A long game’
Animal surveillance is “a long-game question,” said Dr. Andrew Bowman, a veterinary epidemiologist at Ohio State University. “How do we get ahead of the virus and try to understand what might be coming down the line, potentially years from now?”
To keep tabs on mutations in animals, and whether they are transmitted across species, federal researchers are conducting genomic surveillance, comparing virus samples from infected animals with those from humans in close contact with them.
Some researchers are analyzing potential variants. Dr. Kuchipudi and his colleagues created a computational model that virtually generates novel patterns of mutations and then assesses whether they might make the virus more likely to infect, say, cows. Scientists can then watch for those mutations in databases — and observe cattle more closely if the sequences appear. “This will give us a way to make sense of the sequencing data and proactively screen,” Dr. Kuchipudi said.
Scientists also worry about the longer-term threat of viral recombination, in which an animal is simultaneously infected by two coronaviruses that swap genetic material, generating a new virus. Researchers at the University of Liverpool created a model predicting potential hosts in which coronaviruses, including SARS-CoV-2, could recombine.
Staying ahead of the virus will require long-term funding and commitment. But scientists say making these investments now could result in better systems for monitoring pathogens in other species and an expanded understanding of how animal health is linked to ours. It may even help experts catch the next looming health threat before it spills over from animals.
“There’s no harm in understanding better the world around us,” Dr. Han said. “There can only be harm in not understanding and not investing in that knowledge, which is really obvious now.”