Rethink mouse experiment protocols

Using mutant mice to find treatments for devastating diseases such as Alzheimer’s has failed, but that could change by harnessing diversity in the animals, rather than stamping it out, according to authors of a review in the journal Science Translational Medicine.

Neuroscientists Elizabeth Fisher, from University College London, and David Bannerman, from the University of Oxford, both in the UK, pull no punches in calling out the status quo.

“Despite the hundreds of mouse models of human neurodegenerative disease, we still have no cure for any major form of neurodegeneration,” they say.

“Neurodegenerative diseases are common, largely untreatable, and certainly incurable and create a huge health and social burden worldwide.”

Dementia, 70% of which is caused by Alzheimer’s, affected 50 million people around the globe in 2018 at a cost of $1 trillion. In the same year, 10 million people had Parkinson’s disease, which manifests in shaking, stiff limbs and slow movements.

Spearheading the push to understand these illnesses is the humble mouse, with genes altered to give it a version of the disease that works as a template for how it might progress, and be stopped, in humans.

Rather like lemmings, however, a lot of mice go over what has been called a “clinical trial cliff”: a drug works in mice but fails spectacularly when tested in people, or has toxic side effects.

One cause, argue Fisher and Bannerman, lies in a conflict buried deep within the scientific method.

“As scientists, our cultural viewpoint is to minimise variation within our studies to maximise our chances of identifying notable differences between experimental and control conditions,” they write.

In short, it is science bedrock that if you want to know that a gene mutation really causes a disease, or a drug really cures it, you try to keep everything else in your study the same. That means using mice that are near genetically identical, housed under the same conditions.

While that might up the statistical power of studies and make them easier to replicate, there is a big downside.

“When attempts are made to translate findings in animal models to the much more heterogeneous human clinical population (which varies in terms of age, gender, genetic background, environment, and life history), these attempts have failed consistently,” write the authors.

The real world is messy, and illness is often caused by multiple gene variants in a dynamic environment. Stress is a case in point.

Social isolation triggers the stress hormones cortisol and adrenalin in animals and people, and that affects disease. Mice engineered to have Alzheimer’s put in solitary confinement make more beta amyloid, the protein laid down as brain plaques in that disease. On the flipside, humans with Motor Neurone Disease (MND) who are married, which is (mostly) a stress busting factor, can live eight months longer.

When an intangible such as stress holds so much sway, the enormity of creating a faithful animal model, and of the results translating seamlessly to people, hit home.

Our best response, argue the authors, is very much from the “if you can’t beat ’em, join ’em” playbook.

“Embracing and understanding variation may be of great benefit for translation,” they write.

It is a project, they say, that will involve far more extensive documentation of all aspects of an experiment, from the precise composition of rat chow (an excess of phytoestrogen-containing soy may be protective in MND) to whether behavioural tests were done during the day (affecting results in nocturnal animals).

Initiatives in the UK such as the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines attempt to do just that, specifying everything from the number of cage companions an animal has to their bedding material.

Moreover, the authors write, the new environment of machine-learning and big data informatics is well suited to crunching all those extra numbers.

It is a frame shift in scientific thought, upon which many lives may hang.

They conclude: “Embracing rather than rejecting variability in our mouse studies, and then understanding both its sources and its underlying mechanisms, could be of great benefit for successful translation to clinical patient subgroups.”