For the past few years, a handful of Maryland horse farms have gotten a unique kind of pest control. They don’t use poison or set out traps. Instead, a local scientist comes in to rid them of their mice. “I was just an additional big cat running around,” says Stephan Rosshart, an immunologist formerly at the National Institutes of Health in Bethesda, Maryland.
Rosshart’s hunts were part of a long search to solve a plague of immunology studies: Research mice—raised in the sterile, unnatural setting of a laboratory—don’t have the same bacteria, viruses, and pathogens living on them as humans do. And that can make them poor models to study human infection and disease.
Rosshart started to collect wild mice in 2014. His goal wasn’t to use these animals for research models (although some scientists have proposed this). Rather, he wanted them to be the surrogate parents of lab mice to give them a more natural microbial profile.
He turned to organic horse farms to find wild mice that had not been exposed to any pesticides or antibiotics. Rosshart set traps baited with creamy peanut butter—“They don’t like the crunchy one,” he says—and checked them twice daily so the rodents didn’t get stressed out from being in the trap for too long. “It stressed me, because I got up at 4 a.m. and sometimes checked the traps at midnight, but that’s OK,” he says.
Rosshart caught more than 200 wild mice from the farms, took them back to the lab, and implanted them with lab mouse embryos. When the lab embryos were born to the wild mouse surrogates, the species and abundances of microbes on the skin, guts, and genitalia of these “wildling” mice were similar to those found in wild mice, not lab mice, he and colleagues report today in Science.
To assess how well the wildlings modeled human immune responses, the researchers tested two therapies that had seemed promising in traditional lab mice, but failed in humans. One, an antibody treatment that had an anti-inflammatory effect in mice, nearly killed healthy human volunteers when researchers tried it in 2006. When Rosshart tried this, the lab mice survived—as expected—whereas the wildlings died of sepsis.
The other treatment, a potential therapy for sepsis, had increased the survival rates of mice with sepsis, but had the opposite effect in humans. When Rosshart used the treatment on wildlings with a sepsislike condition, they died at a higher rate than untreated controls, just like in the human trials. These experiments suggest the wildling mice could be a more accurate model of human immune responses in preclinical trials.
Rosshart and his colleagues also wanted to see whether wildlings would retain their wild mouse–like microbes in lab settings. When they treated the wildlings with antibiotics and changed their diets, they found that the amounts and proportions of bacteria in and on the wildlings’ bodies changed less drastically than those of lab mice that underwent the same process, and soon bounced back. The wildlings also retained their wild microbiomes for several generations, and even passed on their microbes to cage mates that had been raised in the lab. (This happened through typical mouse behaviors such as eating each other’s poop.)
Integrating the wildlings into laboratory settings, Rosshart says, could be relatively easy. Researchers who wanted to use wildlings for specific experiments could establish colonies of the mice themselves, and refresh their microbiota once in a while by cohousing the wildlings with wild mice, he says. Every few generations, scientists could create more wildlings by implanting the lab mouse embryos into wild mice again.
David Masopust, an immunologist at the University of Minnesota in Minneapolis, says Rosshart’s study was well-designed. But he cautions that more research is needed to determine whether the new method is the best way to create “dirty” mice. Masopust, for example, keeps his lab mice in cages with mice from pet shops to achieve a more humanlike immune response. “I don’t know that … their way is really any better than other approaches,” he says.