Engineered brain cells clear Alzheimer's plaques with a single injection in mice

Science, March 5, 2026

Ten months. That is roughly how much additional independent living the current generation of Alzheimer's drugs, monoclonal antibodies that clear amyloid plaques from the brain, can provide. These therapies require high-dose infusions once or twice a month, indefinitely. They work, but they are demanding.

A study published March 5 in Science introduces a fundamentally different approach. Researchers at Washington University School of Medicine in St. Louis engineered brain cells called astrocytes to carry chimeric antigen receptors (CARs), turning the brain's most abundant cell type into targeted plaque-clearing machines. In mice, a single injection either prevented plaques from forming or cut existing plaque burden in half.

Borrowing from cancer immunotherapy

The concept borrows directly from CAR-T cell therapy, which has transformed treatment for certain blood cancers. In CAR-T therapy, a patient's T cells are genetically modified to recognize and attack cancer cells. The WashU team applied the same logic to a different cell type and a different target.

Astrocytes are the most abundant cells in the brain. They already play a housekeeping role, tidying up the cellular environment. But they do not normally target amyloid beta specifically. The researchers, led by senior author Marco Colonna and first author Yun Chen, designed a gene encoding a chimeric antigen receptor tuned to grab amyloid beta proteins. They delivered this gene to astrocytes using a harmless virus injected into the mice.

The CAR homing device on the astrocyte surface enabled the cells to capture and engulf amyloid beta. With this newly acquired specificity, the astrocytes concentrated their cleaning efforts exclusively on amyloid plaques.

Two groups of mice, two different outcomes

The researchers used mice carrying genetic mutations that cause amyloid beta plaques to saturate the brain by six months of age. They injected two groups with the CAR-expressing virus: young mice before plaques had formed, and older mice whose brains were already saturated with plaques.

Three months after treatment, the young mice remained plaque-free at an age when untreated animals normally have brains full of harmful plaques. The older mice, treated after plaques had already accumulated, showed a 50% reduction in amyloid beta plaques compared to control mice that received a virus without the CAR gene.

David Holtzman, co-author and the Barbara Burton and Reuben M. Morriss III Distinguished Professor of Neurology at WashU, noted the parallel with existing antibody treatments: this CAR-astrocyte approach is also more effective when given earlier. But unlike monthly infusions, it required only a single injection.

Why astrocytes instead of microglia

The brain already has dedicated immune cells, microglia, responsible for clearing waste. But in neurodegenerative disease, microglia become overwhelmed and dysfunctional. Rather than trying to repair the brain's existing cleanup crew, the WashU team recruited astrocytes as reinforcements.

The choice has practical advantages. Astrocytes are abundant and already distributed throughout the brain. They can be targeted with viral delivery systems that are well established in neuroscience research. And because they are not immune cells, they may avoid some of the inflammatory side effects associated with microglial activation.

From mice to humans: the distance remaining

This is a mouse study, and the gap between clearing plaques in a mouse model and treating human Alzheimer's disease is substantial. Mouse models of Alzheimer's, while useful, do not fully recapitulate the human disease. Mice that accumulate amyloid plaques do not develop the full spectrum of tau pathology, neurodegeneration, and cognitive decline seen in patients.

Safety is a major concern. Delivering a permanent genetic modification to brain cells via viral injection raises questions about long-term effects. The CAR-astrocytes are designed to target amyloid beta, but off-target effects on normal brain cell functions need to be carefully evaluated. The researchers acknowledge that more work is needed to optimize the approach and address potential side effects.

The team has filed a patent related to the CAR-astrocyte engineering approach. Future work will focus on fine-tuning the CAR design and exploring whether the same strategy could be adapted to target brain tumors by switching the CAR homing device to recognize tumor markers instead of amyloid.