Beloved as pets, Syrian hamsters are winning another kind of attention from scientists trying to understand and defeat COVID-19. Fifteen years ago, scientists found the hamsters could readily be infected with the coronavirus that causes severe acute respiratory syndrome (SARS). Their symptoms were subtle, so the animals didn't get much traction as a model for the disease. But with COVID-19, caused by a related virus, SARS-CoV-2, the model's prospects appear brighter.
When physician scientist Jasper Fuk-Woo Chan of the University of Hong Kong (HKU) and co-workers recently infected eight hamsters, the animals lost weight, became lethargic, and developed ruffled fur, a hunched posture, and rapid breathing. High levels of SARS-CoV-2 were found in the hamsters' lungs and intestines, tissues studded with the virus' target, a protein receptor called angiotensin-converting enzyme 2 (ACE2). These findings "closely resemble the manifestations of upper and lower respiratory tract infection in humans," Chan and co-authors wrote in a 26 March paper in Clinical Infectious Diseases.
That team is but one of dozens of groups that are racing to develop animal models that can help find effective COVID-19 vaccines and treatments and clarify precisely how SARS-CoV-2 causes disease. The teams are often shorthanded because of the pandemic's shelter-in-place restrictions, but they are collaborating intensively. Each Thursday, the World Health Organization arranges a video conference of nearly 100 scientists, regulators, and funders who are collectively working with a menagerie of lab animals, including mice, ferrets, and several species of monkeys. "A lot of the traditional silos of information are really coming down," says the group's co-chair, William Dowling, who works on vaccine development at the Coalition for Epidemic Preparedness Innovations.
The group swaps the latest data—and tips, such as different viral transmission strategies and the most likely places to find the pathogen in animals. "Everybody has been thrown into a rush to get an animal model that's faithful to the human condition and reproducible," says Chad Roy, a monkey researcher at the Tulane National Primate Research Center. "I don't want to say it's enjoyable because it's a tough time right now, but it's a refreshing way to approach this problem."
The wide range of species may be an asset. "You need the right model for the right question," says Vincent Munster, whose team at the Rocky Mountain Laboratories branch of the U.S. National Institute of Allergy and Infectious Diseases is focusing on monkeys. He cautions against dismissing an animal model simply because SARS-CoV-2 produces an effect, such as death from a brain infection, that doesn't reflect typical disease in humans. "That's a big misunderstanding," he says, noting that "humans don't have a tail, either."
A top priority is to test experimental vaccines by immunizing animals and then "challenging" them with the virus—experiments that must be done in biosafety level 3 labs. Animal models could also warn of dangers of COVID-19 vaccines and drugs; challenges of some experimental vaccines against the related SARS virus, for example, triggered antibodies that enhanced disease severity. Furthermore, experiments with animals may explain why children rarely develop symptoms, how readily SARS-CoV-2 transmits through aerosolized particles versus larger droplets, and whether host genetic factors make some people more susceptible to severe disease. One monkey study has already shown that animals that clear a SARS-CoV-2 infection can resist reinfection for at least 1 month.
Mice—easy to handle and breed—have long been the mainstay of biomedicine, and a good mouse model would be a boon for COVID-19 research. But they shrug off infection with SARS-CoV-2 because the mouse ACE2 receptor has so many key differences from the human one. "It's funny how the virus can have such devastation in humans, and then you can give a million particles to a mouse and it's inert," says Timothy Sheahan, who is developing mouse COVID-19 models at the University of North Carolina (UNC), Chapel Hill.
Chan, working with his HKU colleague, microbiologist Kwok-Yung Yuen, and others, pinpointed the problem by doing a cross-species comparison of the segment of ACE2 to which SARS-CoV-2 first attaches. In the mouse, 11 of 29 amino acids of this critical domain differed from the human version. (Rats had 13 differences, but hamsters only had four.)
One way around the roadblock is to engineer mice that express both the mouse and the human versions of the receptor's gene, ACE2. In 2007, Stanley Perlman of the University of Iowa did just that to study SARS. Although the SARS coronavirus can infect mice through their ACE2, they only develop mild symptoms. Equipped with the human ACE2, mice succumb to a lethal brain disease. This model helped evaluate potential SARS vaccines and treatments, and also teased out the impact of different immune responses.
But demand for the modified animals dwindled after the SARS outbreak subsided in 2003, and Perlman gave them to the mammoth nonprofit mouse supplier Jackson Laboratory (JAX). JAX froze the animals' sperm, and since SARS-CoV-2 surfaced, has raced to breed the mouse again. "We've had over 1000 requests at this point," says Nadia Rosenthal, JAX's scientific director.
A Chinese team that also years ago engineered mice to express the human ACE2 protein to study SARS still had some of the transgenic animals and has already infected them infected with SARS-CoV-2. These transgenic animals lost weight and showed signs of pneumonia but little else, Qin Chuan of Peking Medical Union College and colleagues reported in a preprint published on bioRxiv 28 February. "That's really very, very, very mild disease," Perlman says.
While Perlman has been waiting for JAX to send him newly bred versions of his own mouse, he created a quick and dirty model by stitching the human gene for ACE2 into an adenovirus, which he then used to infect mice so that some of their lung cells made the receptor. When infected with SARS-CoV-2, these mice lost 20% of their weight—more than twice what Qin's team saw—and had ruffled fur, another sign of illness. But none died. "I consider the adenovirus model a stopgap," Perlman says.
The ideal mouse model, he says, will have its own ACE2 gene disabled and only express the human version. Perlman says he would then passage SARS-CoV-2 through the mouse strain repeatedly until the virus adapts and causes severe disease in the rodent. If this works, researchers can dial up or down the impact of a mouse infection by the dose of virus they give.
To create a more faithful mouse model, researchers at JAX are using the genome editor CRISPR to change the sequence of the native mouse ACE2 so that the encoded protein is recognized by the virus. This approach should direct only mouse cells that naturally make the receptor to produce the humanized version. "Everyone I talk to is crying out for authentic human ACE2 expression in mice," Rosenthal says.
Sheahan, in collaboration with UNC's Ralph Baric, is instead tailoring the virus to the mouse. His group is genetically tweaking the surface protein on the virus so that it can infect unaltered mice.
Other SARS-CoV-2 researchers are turning to rats. They are no more susceptible to COVID-19 than mice, but their larger size is an advantage. "You often want to do repetitive bleeding in an experiment, and you can't do that with mice," says Prem Premsrirut of Mirimus, whose company is collaborating with an academic group that's using CRISPR to create a rat model with a human ACE2 receptor. Vaccine studies, for example, often assess how different doses affect antibody responses over several days. Premsrirut notes that "most toxicology studies" of drugs also start in rat. "If you can study a drug directly in rats, you're a step ahead."
Chan's group, which in January contributed to one of the earliest studies to document human-to-human transmission of SARS-CoV-2 and asymptomatic infection, was first to publish data on the new hamster model, but close behind is another team also at HKU. Led by Hui-Ling Yen, the researchers reported similar results with the rodents on Springer Nature's preprint server, In Review, on 1 April. "Some of the hamster data actually look really good," Munster says.
These hamster studies may help further illuminate how the virus spreads. Both groups' experiments put one infected animal in a cage with an uninfected hamster and found transmission of SARS-CoV-2 occurred every time. Although Chan and colleagues suspect the transmission occurred through respiratory droplets, they noted that hamsters eat feces and the scientists could not rule out fecal-oral spread.
Ferrets are a mainstay of research on another respiratory disease, influenza, because the flu virus not only infects them, but produces symptoms that mimic the human disease. Infected ferrets even sneeze, readily spreading flu though the air. The animals may not prove as faithful a model for COVID-19, however. The virus does infect them and causes increases in body temperature, Young Ki Choi of Chungbuk National University and colleagues reported online on 6 April in Cell Host & Microbe. But it did not replicate to high levels and the ferrets didn't develop other symptoms.
The team did find evidence that ferrets might mimic one aspect of COVID-19: respiratory transmission. The animals they infected not only spread SARS-CoV-2 to cage mates, but to two of six ferrets in adjoining cages. Although researchers suspect SARS-CoV-2 primarily transmits through relatively large respiratory droplets that quickly fall to surfaces, this finding suggests finer particles, able to drift in the air for longer periods and over longer distances, can also carry infectious virus. "Aerosol infection is not as highly efficient as direct contact, but it's possible," concludes co-author Jae Jung of the University of Southern California.
Jung also suggests that working with ferrets older than the young ones used in the initial experiment could improve his group's animal model. In humans, for reasons that remain unclear, SARS-CoV-2 strikes the elderly much harder. Jung has seen the same happen in ferrets with a virus that causes loss of white blood cells called platelets. Young ferrets infected with that virus had no symptoms, but 93% of older ones died, his team reported last year in Nature Microbiology. "If we factor in the age of the ferrets, we may be able see more severe disease," after SARS-CoV-2 infection, Jung says. "It's probably similar to the situation in humans."
The animals likely to carry the most weight in assessing potential drugs and vaccines are monkeys. Although they are expensive and difficult to handle, their close genetic relationship to humans often makes monkeys the gate keeper to clinical trials of drugs and vaccines. "This is going to be our near clinical model that we're going to rest heavily on," Roy says.
Intense efforts to infect four different monkey species with SARS-CoV-2 began shortly after the isolation of the virus from people. "There's not been an emergent species that leads me to say, 'Oh wow, this is it,'" says Roy, who is testing African greens and rhesus macaques, and has looked closely at infection data from cynomolgus monkeys. (Marmosets are also being examined.)
In a Dutch study of eight cynomolgus monkeys inoculated with SARS-CoV-2, the four oldest monkeys developed higher levels of the virus in nose and throat swabs compared with younger animals. None of the monkeys developed symptomatic disease, but autopsies found some lung damage in two of four animals. "This looks like what you see in mild cases of humans," says Bart Haagmans from Erasmus University Medical Center, whose team published its data on 17 March on bioRxiv. The model, he suggests, might work better than one that causes severe disease for evaluating vaccine safety because health problems would be easier to detect.
Monkey studies have also begun to explore questions about immune protection. Two rhesus monkeys that recovered from being infected with SARS-CoV-2 at Peking Union Medical College were resistant to reinfection 4 weeks later. The finding provides a hint of good news, as it suggests both natural infections and vaccine-triggered immunity will provide at least some subsequent protection.
Like ferrets, monkeys are being used to address the controversial issue of how much risk people face from aerosol spread of SARS-CoV-2, which will inform debates about the value of face masks and the risks of transmission at, say, a supermarket or in a classroom. Roy and, separately, Douglas Reed at the University of Pittsburgh are staging experiments in air chambers that attempt to infect monkeys through this route, which both might increase pathogenicity and offer clues about transmission risks. "We're trying to get enough virus into them to get some kind of disease," says Reed, who studies African greens and for 7 years shared an office with Roy at a U.S. Army infectious disease lab. "If this virus remains replication competent after many minutes or hours in a small particle aerosol that can be subsequently inhaled, wow, that's a big deal," Roy adds.
Humans who suffer from severe COVID-19 often have underlying diseases, such as hypertension or diabetes, and Roy says researchers may have to find or create monkeys with these comorbidities to develop the most meaningful model. "Monkeys, in general are pretty resilient animals and they handle viral diseases pretty well," he notes.
The list of animal models may soon grow. A recent study published online on 8 April by Science, for example, reported that the virus can infect cats. Autopsies showed the infection led to "massive" lesions in their nasal passages, trachea, and lungs.
Dave O'Connor of the University of Wisconsin, Madison, who is studying SARS-CoV-2 in cynomolgus monkeys, says the field will ultimately winnow down models. "It might turn out that some models are not really worth pursuing after we do this sort of foundational work, but I just don't think we're there yet. We need to let the data guide us."