Scientists have created mice with functioning human brain cells inside their skulls. Why?
Because mice are not humans.
Researchers want better ways to study psychiatric and neurodevelopmental conditions that do not naturally occur in mice. Drugs can look promising in animal experiments and then fail spectacularly when tested in actual humans. So rather than abandoning the mouse model, scientists have tried to make the mice more human. The animals were genetically engineered to develop almost none of their own cerebral cortex, the part of the brain involved in memory, senses and higher-level processing. Human skin cells were then reprogrammed into brain-like organoids and implanted into their brains. The human cells grew, organised themselves and connected with the mice’s existing brains and spinal cords. It is an extraordinary experiment. But it also exposes one of the biggest problems with animal research. The animals we experiment on are not miniature humans.
A review of animal-free research methods found that around 92% of drugs fail during human clinical trials, primarily because problems with safety and effectiveness were not predicted by animal data. Around 40% to 50% of trials fail because a drug does not work, while another 20% to 30% fail because of toxicity. So science has arrived at a remarkable workaround. If a mouse is not human enough to model human disease, make the mouse more human.
Around 6 months after surgery, behavioural testing reportedly found that they behaved largely like other mice. But that should not be particularly reassuring. The ethical problem was never dependent on a mouse suddenly developing human thoughts. Mice already think. They already experience the world. They already have interests of their own. They did not need human brain cells to become worthy of moral consideration.
The mouse is not the problem
Species differences are exactly why animal experiments can be so unreliable.
Humans and mice may share large amounts of genetic material, but differences in how genes are regulated and expressed can result in very different biological responses. Even closely related primates have failed to predict human outcomes in areas including HIV vaccine development.
Then there is the laboratory itself.
Stress caused by confinement, unnatural environments, artificial lighting, noise, handling and witnessing procedures performed on other animals can affect experimental results. In one mouse pain study cited in the review, environmental stressors accounted for 42% of experimental variability.
We take an animal from an artificial environment, alter their biology, subject them to invasive procedures and then ask whether the resulting data accurately predicts what will happen inside a human being.
There are better options.
Human cells can be grown into organoids. Organs-on-chips can recreate aspects of human organs and even connect several organ systems together. Computer models can predict biological responses. AI systems can analyse existing data before anyone manufactures or administers an experimental compound.
Researchers have even developed an AI-powered “virtual mouse” for nanomedicine, using existing study data to model interventions without conducting new experiments on living mice.
The UK government itself now accepts that these technologies need far greater investment. Its Replacing Animals in Science strategy is backed by £75 million to accelerate the development, validation and adoption of non-animal methods. Part of that effort includes investment in human tissue models, organoids, AI, lab-grown heart tissue and other methods designed to replace experiments on animals.
But there is still no overall timetable for ending animal experimentation. Instead, the government is pursuing targeted replacement of particular tests and models where alternatives can be validated. That leaves an obvious danger. New technologies can become additional tools rather than replacements. Researchers could use organoids, AI, computer modelling and organs-on-chips and still move into animal experiments afterwards.
The review into animal-free research specifically warns that alternatives risk being incorporated alongside animal testing instead of actually replacing it. And now we have perhaps the strangest illustration of where the old model can lead. Mice are poor models for some human conditions precisely because mice are not humans. The response has been to remove much of a mouse’s cerebral cortex, implant human brain tissue and see whether that makes them more useful to us.
The scientific question is whether doing so creates a better model. The ethical question should come much earlier.
Why should a mouse have to become more human before we start questioning what we are doing to them?