From mice to medicine - Six questions with Nadia Rosenthal

How working with genetically diverse could revolutionize our understanding of human disease and treatment.

You've just completed a clinical trial for a promising new medication, but when it hits the market, it only works for about 10-20% of patients. Why? The inconvenient truth is that our genetic uniqueness means that a one-size-fits-all approach rarely works. This challenge isn't just limited to humans-our tiny, furry research counterparts, mice, face the same issue.

In a recent Nature Biotechnologycommentary, spearheaded by JAX Mammalian Genetics Scientific Director Nadia Rosenthal, PhD, FMedSci, a team of genetics researchers at The Jackson Laboratory (JAX) has proposed a new approach to developing treatments. This comes on the heels of the U.S. FDA's adoption of the Modernization Act 2.0, which opens the door for alternatives to traditional animal testing. JAX's message is clear: to truly advance medicine, we must use genetically diverse mice that better reflect the genetic tapestry of the human population.

In this Q&A, we'll dive into this new approach and explore how it could reshape our understanding and treatment of human diseases.

What is the main issue with the current use of traditional mouse models in drug discovery?

Let's start with the reasons mouse models have been so critical for medical research.

Scientists have been studying animals to help model human diseases for over a hundred years. They developed inbred laboratory mouse strains because they have many of the same biological features as humans, but have compressed lifespans, breeding at only 6 weeks of age and living only 2 to 3 years. This helps us rapidly follow the progression of rare diseases that start after birth or common diseases that come with age, such as heart failure or diabetes.

Most importantly, mice and humans share virtually the same set of genes. 80 million years ago, humans and mice shared a common ancestor, making our genomes comparably similar. Almost every gene found in one species has been found in a closely related form in the other.

Accordingly, it's not surprising that mice and humans share many features and suffer from many of the same diseases. In mice, however, we can experimentally test the function of the genes we have in common. Scientists can mimic the effect of DNA alterations that occur in human diseases and carefully study the consequences of these DNA misspellings. Mouse models also afford the opportunity to test possible therapeutic agents and evaluate their effects.

At JAX, we keep over 13,000 different strains and each one has a unique genetic makeup. We have over 100 years of careful breeding records to rely on, so we know exactly which genetic variations each of these strains carry.

If someone needs a mouse that tends towards extreme obesity, we suggest the NZO mouse strain, which gains weight unusually fast when fed a 'Western diet' of high fat and high sugar. If someone is studying Type 1 Diabetes, we suggest the NOD mouse strain that has a high likelihood of developing autoimmune disease.

Why do inbred mice often fail to accurately replicate human conditions, such as cancer and diabetes?

There's a popular notion that drugs successfully trialed in mice don't pan out in humans because mice are wired differently. As a result, pharma companies are moving away from mice as an experimental model and placing their bets on human cells.

One might assume that if a drug is approved for use in people, all patients will respond as advertised -- but an inconvenient truth, which most doctors know, is that most medications that pass clinical trials work in only a relatively low fraction of patients (10-20%). This is because we are genetically unique and most drugs only work on individuals with a specific genetic makeup. This diverse response is highly likely to play out in cells isolated from different patients, each responding to a given drug in a different way.

The same is true in mice. If you test a medication in a single mouse or the cells from that mouse, the chances of it working are about the same as if you tested it in one person or cells isolated from that person.

To complicate matters, unlike patients, lab mice are inbred, like purebred dogs - each with their own set of characteristics and disease susceptibilities. If you want mice to respond to medications like humans, you need to make them more diverse, like humans.

So, when someone claims a drug doesn't work the same in mice and humans, I ask them, "Which mouse and which human? And how many different mice did you test?"

What new approach to drug discovery are you proposing?

At JAX, we are breeding mouse mutts -- think labrodoodles or cockapoos! We randomly cross different mouse strains to give us mice with genetic diversity that is more closely related to human genetic diversity. This allows us to follow inherited traits that underlie more complex diseases like heart disease, dementia or diabetes. These 'designer mice' are healthier, unique and highly variable in their response to medications, just like people.

We are also developing new panels of mice that are derived directly from different locations in the wild. These mice vary even more widely than the limited genetic repertoire we have in the lab, as they have been cordoned off from their wild ancestors for over a hundred years and bred for specific traits that may affect the way they respond to disease.

Why is this important to the pre-clinical research community, or the U.S. Food and Drug Administration?

Until recently, for safety and efficacy, the FDA mandated testing new drugs in an animal model before starting human trials. Recently they revised their mandate to allow drugs to be tested only on human cells, bypassing the need for animal testing. Animal experiments are time consuming, expensive and carry ethical issues, so the thinking is that reducing or eliminating the use of animals in drug testing would benefit everyone -- including the animals.

The problem with this reasoning is that we still don't know how much cells in a dish tell us about human disease. Cells may develop cancer in a dish, but it's not clear how we can study the complex interactions of tissues, immune responses, or psychiatric conditions. Safety is still a big concern: a drug that doesn't affect cell viability in a dish may have negative effects on tissues or organ systems in the whole body.

How could the proposed approach revolutionize our understanding of treatment responses in patients?

Cell models carry big advantages - you can test thousands of cells at a time with thousands of compounds. But cells carry just as much genetic diversity as the people they come from. So, unless you know a lot about the genetic makeup of each cell, it's hard to distinguish features cause by a drug response from those that caused by underlying genetics.

We still have only a rudimentary understanding of how our genetic background affects disease, and there's a limit to how much we can deduce from cell behavior. What if a drug gives the desired effect in a cell from your body but not in one from mine? Does that mean the drug is no good? Or that it only works in people who have an unknown genetic feature that you are lucky enough to carry?

Imagine testing that same drug on cells from two mice with different known genetics. Let's say mouse A responds but B doesn't. We can go back to the lucky mouse strain A and use our knowledge of mouse pedigrees to map the specific genetic feature that allows the drug to work.

And it gets better. We can now test our hypothesis by introducing that genetic feature from strain A into cells from strain B to see if the drug now works. If it does, we can now screen people for genetic feature A, and that subset of people should all respond to the drug.

Why is this important? Remember that only 10-20% of humans respond to any drug. By prescreening human patients for features that either make people responsive or not, you could be much more precise in drug delivery. We call that precision medicine.

Can you provide an example of how diverse mice have provided valuable insights into a specific disease?

Take the COVID-19 pandemic. When it hit, scientists were unable to test vaccines or antivirals because mice don't get COVID-19. This is because the protein on the surface of human cells that admits the SARS-CoV-2 virus is slightly different in the mouse.

The solution was obvious - genetically engineer the human version of the gene into the mouse, and voilà, the mouse comes down with COVID-19 when exposed to the virus.

The only problem was that the effect was really severe. We know now that there is a broad spectrum of responses to COVID-19 in people - some get really sick and others are asymptomatic. There is currently no way of predicting the outcome. How do we model that spectrum of response in mice?

By now you may have come up with the answer: try it on mouse mutts! We crossed the mouse bearing the human gene to 10 different diverse strains of mice in our collection and tested the virus on a generation of mouse mutts.

The results were spectacular: One mutt got very sick, another got a respiratory infection but recovered and another got infected but showed no sign of illness. All of these results were reproducible, because the underlying genetics is reproducible.

Knowing the genetics of these mice now allows us to trace the inherited basis of this difference in symptoms and use that knowledge to understand how to better treat diseases like long COVID.

So -- don't underestimate what is going on in your kitchen cupboard at night! All that genetic diversity could lead to the next medical breakthrough.

Public link

Please use the above public link if you want to share this noodl on another website.