Drug companies have spent billions of dollars searching for therapies to reverse or significantly slow Alzheimer’s disease, to no avail. Some researchers argue that the best way to make progress is to create better animal models for research, and several teams are now developing mice that more closely simulate how the disease devastates people’s brains.

The US National Institutes of Health (NIH), the UK Dementia Research Institute and Jackson Laboratory (JAX) — one of the world’s biggest suppliers of lab mice — are among the groups trying to genetically engineer more sophisticated rodents. Scientists are also probing the complex web of mutations that influences neurological decline in mice and people.

“We appreciate that the models we had were insufficient,” says Bruce Lamb, a neuroscientist at Indiana University in Indianapolis who directs the NIH-funded programme. “I think it’s sort of at a critical juncture right now.”

Alzheimer’s is marked by cognitive impairment and the build-up of amyloid-protein plaques in the brains of people, but the disease does not occur naturally in mice. Scientists get around this by studying mice that have been genetically modified to produce high levels of human amyloid protein. These mice develop plaques in their brains, but they still do not display the memory problems seen in people.

Many experimental drugs that have successfully removed plaques from mouse brains have not lessened the symptoms of Alzheimer’s disease in people. One high-profile stumble came last month, when three companies reported that their Alzheimer’s drugs — from a class called BACE inhibitors — had failed in large, late-stage clinical trials. Although the drugs successfully blocked the accumulation of amyloid protein in mice, they seemed to worsen cognitive decline and brain shrinkage in people.

A complex genetic stew

The drive for better mouse models comes as genomics studies are linking the most common form of Alzheimer’s — late onset — to dozens of different genes ¹ . This diversity suggests that each case of the disease is caused by a different combination of genetic and environmental factors. “There is no single Alzheimer’s disease,” says Gareth Howell, a neuroscientist at Jackson Laboratory (JAX) in Bar Harbor, Maine.

Howell argues that scientists’ reliance on inbred lab mice with only a few genetically engineered mutations might have limited research. His own work suggests that in mice, just as in people, genetic diversity plays a part in determining how neurodegeneration progresses.

When Howell’s team modified two genes associated with early-onset Alzheimer’s in both lab mice and their wild cousins, all of the animals developed amyloid plaques. But although the more-inbred lab mice did not display any outward signs of Alzheimer’s, a portion of the genetically diverse wild mice suffered memory problems. The researchers think that a combination of plaques and unknown genetic factors caused these symptoms. They presented the results this month at a meeting of the Society for Neuroscience in San Diego, California.

Another study, by neuroscientist Catherine Kaczorowski at JAX, suggests that animals’ genetic make-up affects how they respond to environmental triggers. Her group bred genetically diverse wild mice to lab mice with mutations that cause amyloid plaques to form. Some of the resulting offspring were more likely to develop cognitive problems if they ate a high-fat diet, yet other mice with the diet had a lower risk of these symptoms, Kaczorowski reported at the San Diego meeting.

Remaking rodents

Understanding how this expanded universe of genetic factors affects Alzheimer’s risk will require a host of new animal models with different combinations of mutations. Several efforts to engineer these next-generation mice are already under way.

In 2016, the NIH started the MODEL-AD consortium to develop more Alzheimer’s mice and make them available to researchers. Project scientists engineer mice with different genetic mutations associated with early- or late-onset Alzheimer’s, and test the animals to see whether they display signs of the disease. They then post descriptions of each mouse type in an online database. Lamb says that the team has released about 30 mouse varieties, and received more than 500 orders for the animals from academic scientists and biotechnology companies.

And in January, the UK Dementia Research Institute in London launched a similar programme. Scientists there are developing model mice whose brains show the amyloid plaques and tangles of another protein, called tau, that appear in people with Alzheimer’s. To mimic the brain inflammation that the disease causes, the group is implanting neural immune cells grown from human stem cells into the brains of mice.

Ultimately, researchers hope that the models will reveal ways to predict whether a person will respond to a particular Alzheimer’s therapy. And having a better understanding of how inflammation and genes drive the disease could help to identify it in people before plaques and tangles have formed, says Rudolph Tanzi, a neurologist at Harvard University in Cambridge, Massachusetts. “That’s why it’s so important to have those animal models available and really start working on all these genes.”

But Bart de Strooper, a molecular biologist at the Catholic University of Leuven (KU Leuven) in Belgium, urges caution. De Strooper, who directs the UK programme, says that none of the next-generation animals is likely to be a perfect analogue for people. “The biggest mistake you can make,” he says, “is to think you can ever have a mouse with Alzheimer’s disease.”