Choose the wrong model, and years vanish testing drugs that never help patients
In the long effort to heal what multiple sclerosis quietly destroys, a team of American researchers has offered a corrective to a foundational assumption: that two widely used laboratory models of the disease are interchangeable. By mapping the genetic activity of more than 400,000 brain cells across mice and humans, scientists at the University of Notre Dame have shown that the choice of model shapes the very nature of what is being studied — and that choosing poorly may explain why so many promising treatments dissolve before they reach patients. It is a reminder that in science, as in life, the frame through which we look determines what we are able to see.
- Nearly 3 million people live with MS and no existing treatment can restore the myelin once it is lost — the urgency to find one has driven decades of animal research that may have been partially misdirected.
- A new analysis reveals that two mouse models used almost interchangeably produce fundamentally different cellular responses, meaning years of drug testing may have been built on a mismatched foundation.
- The cuprizone model mirrors the stress MS places on myelin-producing cells, while the LPC model better replicates the immune system's inflammatory reaction — conflating the two risks chasing cures that work in the wrong context.
- Researchers are now calling for model-specific research strategies and experiments in aging and repeatedly injured animals to close the gap between laboratory results and human disease.
- Experts caution that neither model fully captures human MS complexity, and unresolved questions — whether drugs reach the brain, whether remyelination can be measured — mean the road from mouse to patient remains longer than this breakthrough alone can shorten.
On a morning set aside to acknowledge multiple sclerosis, American researchers released findings that could redirect how scientists search for treatments that actually heal. For years, two mouse models — cuprizone and lysophosphatidylcholine — have been used almost interchangeably to study the disease. A sweeping analysis of more than 400,000 brain cells suggests that assumption was wrong, and that the difference between the two models matters enormously.
Multiple sclerosis destroys myelin, the protective sheath around nerve fibers, leaving nearly 3 million people worldwide to navigate fatigue, vision loss, numbness, and unpredictable balance failures. Current therapies can slow progression but cannot restore myelin once it is gone. Because living brain tissue from MS patients is nearly impossible to obtain, animal models remain the only practical stand-in — making the choice of model a decision with real consequences.
The cuprizone model introduces a toxic compound through food, gradually eroding myelin across the brain. The LPC model injects a substance directly, producing a localized lesion within days. Both cause demyelination, but through different mechanisms. Led by Katrina Adams at the University of Notre Dame, researchers used single-cell RNA sequencing to map gene activity across 112,000 mouse brain cells and cross-reference it with 321,565 samples of human MS tissue. The result was a clear distinction: cuprizone pushes myelin-producing oligodendrocytes into a stress state that closely mirrors human lesions, while LPC provokes a more aggressive and prolonged immune response from the brain's microglia.
The practical stakes are direct. A researcher studying how myelin-producing cells struggle or attempt repair should use cuprizone; one studying immune response should use LPC. Choosing the wrong model risks years of testing drugs that succeed in the wrong biological context and fail in human trials.
The team acknowledged that neither model fully captures the cellular diversity of actual human lesions, framing their work as a roadmap rather than a solution. Buenos Aires neurologist Dr. Jorge Correale praised the study's rigor while noting that some findings conflict with earlier research suggesting meaningful differences between mouse and human oligodendrocytes — which may help explain why myelin-regenerating drugs that work in animals so often disappoint in patients. Questions about whether candidate drugs reach the central nervous system, and how remyelination can be reliably measured, remain open. The path from mouse to human, the findings suggest, is clearer now — but still long.
On a morning when the world pauses to acknowledge multiple sclerosis, a team of American researchers released findings that could reshape how scientists hunt for treatments that actually work. For years, neurologists have relied on two mouse models to study the disease—cuprizone and lisofosfatidilcolina—using them almost interchangeably, as if they were equivalent tools. A new analysis of more than 400,000 brain cells suggests they are not equivalent at all, and the choice between them matters more than anyone realized.
Multiple sclerosis destroys myelin, the insulating sheath that wraps around nerve fibers in the brain and spinal cord. Nearly 3 million people worldwide live with the disease, experiencing sudden waves of fatigue, blurred vision, tingling in the limbs, and balance problems that come and go without warning. Current treatments can slow the disease's progression and reduce flare-ups, but none can regenerate the myelin once it's gone. That gap—between managing symptoms and actually healing the damage—is where the mouse models come in. Because obtaining active brain tissue from living MS patients is nearly impossible, researchers have no choice but to work with animal stand-ins.
The cuprizone model works by mixing a toxic compound into a mouse's food, gradually destroying myelin across the entire brain. The lisofosfatidilcolina model is more direct: scientists inject the substance straight into the brain, creating a small lesion within days. Both produce demyelination, the loss of that protective coating. But they do it in fundamentally different ways, and until now, no one had rigorously compared what those differences meant.
Scientists at the University of Notre Dame, led by Katrina Adams and her team, analyzed 112,000 brain cells from mice exposed to both models and cross-referenced that data with 321,565 samples of human brain tissue from MS patients. They used single-cell RNA sequencing, a technique that reads which genes are active in each individual cell, building a genetic map one cell at a time. What emerged was a clear distinction. Cuprizone pushes oligodendrocytes—the cells that manufacture myelin—into a state of severe stress that closely mirrors what happens in human MS lesions. Lisofosfatidilcolina, by contrast, triggers a much more aggressive and prolonged inflammatory response from the brain's immune cells, the microglia.
This distinction has immediate practical consequences. If a researcher wants to understand what happens to myelin-producing cells under stress—whether they're struggling, dying, or attempting to repair themselves—cuprizone is the better choice because the damage unfolds gradually. If the goal is to study how immune cells respond to myelin loss, lisofosfatidilcolina offers a more intense and realistic inflammatory environment. Choose wrong, and a drug that shows promise in the lab might fail completely in human trials, wasting years and resources on a dead end.
The Notre Dame team also discovered that neither model fully captures the cellular diversity present in actual human lesions. Adams acknowledged this limitation while framing the work as a roadmap: evidence-based guidance for researchers deciding which model suits their specific question. The team proposes extending experiments to aging models and repeated injuries as the next step, moving closer to the messy reality of the disease as it actually unfolds in patients.
Dr. Jorge Correale, a neurologist and neuroimmunologist at FLENI in Buenos Aires, praised the study's methodological rigor while noting its constraints. The work integrates genetic data from mice and humans in a way that identifies disease-associated states in oligodendrocytes that persist across species. But Correale pointed out that some findings contradict earlier research from McGill University suggesting human and mouse oligodendrocytes differ in important ways—which might explain why drugs that regenerate myelin beautifully in animals often fail in patients. He also raised questions that remain unanswered: Do all candidate drugs actually reach the central nervous system? How can scientists reliably measure whether remyelination is actually happening? These gaps suggest the path from mouse to human remains longer than the headlines might suggest.
Meanwhile, clinical trials are underway testing potential MS interventions, with some focused on halting disease progression and others on early detection through biological markers. The Notre Dame findings don't answer these questions, but they do provide a clearer lens for asking them. By matching the right model to the right research question, scientists can avoid years of chasing false leads and move faster toward therapies that might actually restore what the disease has taken away.
Citas Notables
Our analysis of these two models provides a roadmap based on solid scientific evidence that we hope will advance the study of MS and related diseases— Dr. Katrina Adams, University of Notre Dame
This could explain why some drugs with excellent results in remyelination in animals are not effective in patients with multiple sclerosis— Dr. Jorge Correale, FLENI neurologist
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter which mouse model researchers choose? Aren't they both studying the same disease?
They're studying the same disease, but through different windows. One model shows you what happens to the cells that make myelin when they're under stress. The other shows you what the immune system does in response. If you're trying to develop a drug that protects those myelin-making cells, you need the first model. If you're trying to calm down the immune attack, you need the second. Pick wrong and you spend years testing something that works beautifully in mice but means nothing for actual patients.
How did researchers figure this out?
They sequenced the genes active in hundreds of thousands of individual brain cells from both mouse models and compared them to human brain tissue from MS patients. It's like reading the genetic signature of each cell to see what it's actually doing. When they lined up the data, the patterns didn't match equally. One model looked much more like human disease in certain ways.
Does this mean we're close to a cure?
It means we're closer to asking the right questions. The study doesn't give us a cure, but it gives us a map for how to search more efficiently. That matters because drug development is expensive and slow. Every year wasted on a dead-end therapy is a year patients aren't getting better.
What's still missing?
Everything about how the disease actually works in a living human body. These models are still simplified versions of reality. They don't capture all the different types of cells involved in real lesions. They don't tell us whether drugs can actually cross the blood-brain barrier to reach the tissue. And some research suggests human and mouse cells behave differently in ways we don't fully understand yet.
So the mouse models are useful but imperfect?
Exactly. They're the best tool we have right now, but they're still a proxy. The real test always comes in human trials, and that's where many promising mouse discoveries fail. This study just makes the proxy a little more honest about what it can and can't tell us.