Can mice model human behaviors? Maybe better than you think

neurobehavioral mouse assays

A mouse surrounded by computer screens turns its head when it notices lines moving across one of them, as a camera captures this evidence of visual acuity. A chamber similarly equipped with video cameras tests social interaction between mice. A small swimming pool, with shapes on its walls as navigational cues, lets scientists gauge a mouse’s spatial memory. A pint-sized treadmill, with a tiny camera to watch foot placement, measures gait.

Here in the Neurobehavioral Developmental Core at Boston Children’s Hospital, managed by Nick Andrews, PhD, the well-tended mice also have opportunities to play: “If you have a happy mouse,” says Andrews, “researchers get better, more consistent results.”

Those results help researchers better understand conditions such as autism, stroke, traumatic brain injury and inherited deafness, and gauge how well a new treatment is working or not working.

The Core’s newest addition, the “Smart Chamber,” is essentially a tiny sound recording studio. It picks up sounds that mice use to communicate and that humans are incapable of hearing—great for behavioral testing on mouse pups that normally call out for their mother. “In autistic models,” explains Andrews, “the mice take longer to communicate, or they don’t make any noise at all.”

neurobehavioral assays mice
A social interaction test: The blue tracks in these images show the path of a mouse sharing a cage with a new mouse (yellow star). The mouse at left is actively checking out its cagemate; the mouse at right appears disinterested, spending most of its time in the corner.

But can these models really capture the complexity of neurobehavioral disorders? Is there such a thing as an autistic mouse? Mouse models have their limitations, but when done right, they reveal more about human disorders than you might think.

Replicating clinical disorders

Once researchers discover a human gene they believe is responsible for a disease, they can try to turn it on or off in mice, using pharmaceuticals or gene therapy, and observe how the animals’ behavior is affected. Mouse genes and human genes are “extraordinarily similar,” says Peter Tsai, MD, PhD, a researcher and child neurologist who has used the Core since its inception.

Neurobehavioral assay mice
A spatial memory test: At left, the path of a mouse swimming a water maze, seeking a platform with food. At right, the mouse, cued by the geometric shapes, navigates directly to the platform.

Tsai works on a mouse model of tuberous sclerosis, a genetic disease that causes benign tumors in the brain, heart and elsewhere in the body. Ninety percent of patients with tuberous sclerosis develop epilepsy, and half of them exhibit autism symptoms at some point in their lives.

The mouse model lets Tsai evaluate social behaviors as compared with control mice: how much time a mouse spends with another mouse, whether it shows more interest in interacting socially with another mouse or in examining an object, and whether it exhibits classic autistic symptoms like repetitive behavior. But since he can’t communicate with a mouse, he will never know for sure whether these behaviors are analogous to what a person with autism experiences.

“Without a doubt, mouse behaviors are not completely representative of human behavior,” he says. “However, they provide a critical tool to evaluate gene and brain function while also providing a platform to evaluate potential therapies.”

muscular dystrophy mouse model
This plot charts the movements of mice with Duchenne muscular dystrophy (DMD) before and after exercise. Treated mice (far right) were almost as active as normal mice.

Michela Fagiolini, PhD, director of the Core, researches new treatments for Rett syndrome. This genetic disorder mostly affects young girls who develop normally into their toddler years before suddenly regressing. They have autism-like features such as hand wringing and appear “closed off” from the world around them. They have intellectual disabilities, often lose their language skills, and have motor and respiratory problems.

Rett syndrome is caused by a mutation in the MECP2 gene on the X chromosome; the corresponding mouse gene was discovered nearly 15 years ago. Using this mouse model, Fagiolini is testing new therapies to see if she can reverse or at least slow the regression in Rett syndrome.

Like their human counterparts, the mutant mice develop normally at first but then begin to display nearly identical symptoms to the girls’; they even wring their tiny paws. But Fagiolini and her team noticed a symptom that hadn’t been described in the clinical literature: Their mice were also losing their sight. They asked their clinical colleagues whether their patients with Rett syndrome could see normally and found that no one had ever tested this.

“So we went back to the Rett clinic at Boston Children’s and started to measure vision in the patients,” Fagiolini says.  What they found was astonishing: The children were, in fact, processing visual stimuli very differently than typically developing children.

Evaluating treatments and gene interactions

Beyond just describing disorders, the Core can document the efficacy of treatments—sometimes equally dramatically. PhD student Charles Askew, from the lab of Jeffrey Holt, PhD, and Gwenaelle Géléoc, PhD, recently tested a possible cure for genetic deafness. After treating deaf mice, he placed them in the Core’s startle boxes—which let out a quick beep and capture the animals’ startle or “jump” reflex. When Georgia Gunner, a lab assistant in the Core, increased the volume, the mice suddenly started jumping. “I couldn’t help but shout, ‘It’s working!’” Gunner says. “It was really cool to see the rescue of the hearing.”

Though far from perfect, mouse models are proving their value in studying complex behavioral disorders. In the future, rather than study one gene or one disease at a time, researchers would like to do cross-comparisons of different genes that result in similar symptoms. This could yield a better understanding of how genes have co-evolved and how they interact—like grouping similarly colored puzzle pieces together rather than constructing the image piece by piece.

Neurologist Mustafa Sahin, MD, PhD, for for example, is comparing autistic features produced by mutations in genes like PTEN, Shank3, CntNap2 and the fragile X gene Fmr1. Such methods could be applied to other complex disorders like schizophrenia or addiction. “It’s one of the most exciting things that animal models will allow us to do,” he says.

Join us at the Global Pediatric Innovation Summit + Awards 2014 on October 30-31 at Boston’s Seaport World Trade Center. Seats are limited, so register today at Please use the code VECTOR at check-out for a 10% discount.