For decades, Alzheimer’s research has focused on two targets: amyloid plaques and tau tangles — the protein deposits that accumulate in the brains of patients. Two drugs, lecanemab and donanemab, can now clear amyloid and slow cognitive decline.
But they share a stubborn limitation: neither stops brain cells from dying. And neither works against the broader family of diseases — called tauopathies — where tau accumulates without amyloid ever appearing.
A new study from Washington University School of Medicine in St. Louis points toward a different culprit, and a different solution. The real killer in Alzheimer’s, the research suggests, may not be the tau protein itself — but the immune system’s response to it.
The Immune System Gone Wrong
When tau accumulates in the brain, something unexpected happens: T cells flood in. These are the immune system’s specialist soldiers, typically dispatched to fight infections and cancer. In Alzheimer’s, they appear in high numbers precisely where tau is most abundant. And growing evidence suggests they are driving much of the neurodegeneration — the actual death of brain cells — that eventually robs patients of their memory and independence.
This is a genuinely important shift in how the disease is understood. Tau tangles don’t directly kill neurons in the way previously assumed. They trigger an immune assault that does the damage instead.
The question the Washington University team set out to answer was: how do the T cells get in? And once that was known — could they be stopped?
Following the Chemical Trail
T cells navigate through the body by following chemical signals called chemokines. The researchers had already found that one particular chemokine — CXCL10 — was elevated in mice with tau accumulation, and is elevated in humans with Alzheimer’s. T cells carry a receptor protein on their surface called CXCR3, which is essentially a sensor tuned to follow this exact signal like a homing beacon.
The team confirmed that when either the CXCL10 signal or the CXCR3 receptor was absent, T cells simply did not enter the brain — even when the researchers deliberately provoked inflammation to drive an immune response. The CXCR3 pathway was the gate.
Blocking the Gate
Having identified the mechanism, the team tested whether blocking it could protect the brain. They injected young mice — old enough to have tau accumulation but not yet showing significant brain cell loss — with an antibody designed to block CXCR3. Treatment was given every five days for three and a half months.
The results were striking. Treated mice had roughly half as many T cells in their brains as untreated animals. More importantly, they retained about 40% more brain tissue in memory centres, showed significantly less nerve cell damage, and performed better on memory tests.
The tau tangles themselves were unchanged — the antibody didn’t clear the protein, and wasn’t designed to. But neurodegeneration was meaningfully reduced anyway. The damage was coming from the immune response to tau, not from tau directly — and interrupting that response protected the brain regardless of what the tau was doing.
A Crucial Practical Advantage
One finding is particularly significant for anyone thinking about how this could eventually become a treatment: the antibody worked without crossing the blood-brain barrier. It remained at the brain’s border, blocking T cells from entering, rather than penetrating brain tissue itself.
This matters enormously. Getting drugs into the brain is one of the most formidable challenges in neurology — the blood-brain barrier is designed to keep foreign molecules out, and the vast majority of candidate drugs fail to cross it in useful quantities. A treatment that works from the outside sidesteps that problem entirely.
“You wouldn’t have to design the drug to get into the brain — which is a big deal,” said senior author David Holtzman. “And you don’t have to get rid of the tau to get this therapeutic effect.”
Existing Drugs Already Target CXCR3
The potential path to human treatment is closer than it might first appear. Drugs that suppress specific T cell activity already exist — some are already used to treat multiple sclerosis and other autoimmune conditions. These weren’t designed with Alzheimer’s in mind, because tauopathies have never been thought of as autoimmune disorders. But this research suggests they might respond to the same therapeutic logic.
“This study shows for the first time in an animal model that these diseases respond to a specific T-cell therapy,” Holtzman said. If the safety profile is acceptable, repurposing existing T cell-targeting drugs for clinical testing in Alzheimer’s patients is a plausible next step — one that could move faster than developing an entirely new compound from scratch.
What This Doesn’t Change — Yet
The honest caveats are important. This is mouse data, and Alzheimer’s has a long and humbling history of results that looked transformative in rodents and failed in humans. Longer-term studies are needed to confirm that protecting brain tissue in the short term translates into sustained cognitive benefit. And the CXCR3 pathway is involved in more than just T cell brain homing — blocking it could have effects elsewhere in the immune system that haven’t yet been characterised.
But as a conceptual advance, the research is meaningful. It identifies a specific, druggable mechanism responsible for the brain cell death that existing treatments cannot stop. It works independently of tau clearance, meaning it could in principle be combined with amyloid-targeting drugs for a more comprehensive attack on the disease. And it applies not just to Alzheimer’s but to the entire family of tauopathies — including progressive supranuclear palsy and frontotemporal dementia — for which there are currently no disease-modifying treatments at all.
Alzheimer’s research has been reoriented before by a single key insight about mechanism. The discovery that amyloid triggers tau, and that tau triggers the immune assault that kills neurons, is still being fully unpacked. This study takes another step in that unpacking — and points, for the first time, toward stopping the killing without having to first clear the protein that started it.