Editorial NexoScience
For years, the fight against Alzheimer’s has been waged almost entirely inside the skull. Researchers have poured effort into clearing the toxic protein clumps that build up in the brain, amyloid and tau, with limited success. New experiments in mice point somewhere else. Some of the damage may be ordered from lymph nodes in the neck.
The work comes from a team that includes neurologist David Holtzman of Washington University School of Medicine in St. Louis, and it was published Sept. 3 in the journal Nature Neuroscience. The researchers studied mice bred to develop tau tangles and the brain loss that follows them. When the mice lacked one small group of immune cells that lives mostly outside the brain, their brains shrank far less, more neurons survived and the animals behaved more normally. The tau was still there. So the question of whether Alzheimer’s starts outside the brain has a partial, mouse-sized answer: at least part of the attack seems to be set in motion there.
T cells were already the prime suspects. In brains with tau disease, immune cells called T cells pile up, especially CD8+ T cells, the cytotoxic kind that kill other cells. In earlier work on the same mouse model, the group found that wiping out T cells eased the neurodegeneration. In those brains, CD8+ T cells made up most of the invading T cells, about 60% to 70%, and many were copies of the same few clones. That pattern hints that they were reacting to something specific. What nobody knew was how they got their orders to go in.
Every soldier needs a briefing. A naive T cell doesn’t attack anything on its own. It first has to be “primed,” and the cells that do the priming are dendritic cells. They pick up fragments of material from damaged or foreign cells and show them to T cells, a bit like an officer handing out a photo of the target. One subtype, conventional type 1 dendritic cells (cDC1s), specializes in “cross-presentation,” showing scraps collected from outside the cell to CD8+ T cells. They’re rare. In these mice, cDC1s made up only about 0.4% to 0.6% of all white blood cells, roughly one cell in every 200.
The team took the briefers out of the picture. The mice carried a human tau gene with a disease mutation plus a human version of the APOE gene called APOE4. The researchers crossed them with mice missing a short stretch of DNA that switches on a gene called Irf8 only in cDC1s. Without that switch, cDC1s never mature, and the cells vanished from the spleen, the brain’s draining lymph nodes, the membranes around the brain and the brain itself. The team checked that microglia, the brain’s resident immune cells, still made the Irf8 protein as usual, so the effect wasn’t coming from them.
The brains held up. By 9.5 months of age, ordinary tau mice showed clear shrinkage of the hippocampus, a memory hub, and of nearby cortex, along with swollen fluid-filled ventricles. Tau mice without cDC1s kept much more of that tissue. They had more neurons in the dentate gyrus, a part of the hippocampus, and they scored higher at nest building, a natural behavior that tau damage erodes. Blood levels of neurofilament light chain, a protein that leaks out of dying neurons, were lower in males without cDC1s, but not in females.
Tau didn’t go away, and that’s the interesting part. Levels of tau and of its chemically tagged, disease-linked form barely changed. There were small drops in some fractions, nothing like the gap in brain loss. Here’s where it gets strange: the mice were protected while still carrying the protein everyone blames. “A lot of the damage that tau is causing is due to the inflammatory response that it’s eliciting,” Holtzman told Science News. In males, microglia and star-shaped support cells called astrocytes were also calmer without cDC1s. In females the microglial markers didn’t differ, which the authors put down to the milder brain damage females show in this model.
Far fewer killer T cells got in. Tau mice had about four to five times more immune cells in the brain than mice without tau. Removing cDC1s cut that crowd by roughly 40%. In the hippocampus, T cells dropped by about half. The cut was selective: CD8+ T cells fell sharply, while CD4+ T cells, a helper type, stayed level. The CD8+ cells that did make it in were less activated and showed less clonal expansion, which is what you’d expect if they had never been properly briefed. And T cell counts in the blood and spleen didn’t change, so the mice weren’t simply short of T cells overall.
The orders are given outside the brain. Even in brains with obvious shrinkage, cDC1s were barely detectable. To find where priming happens, the team injected into the brain cells loaded with ovalbumin, an egg-white protein, and gave the mice T cells built to recognize it. Those T cells never showed up in the brain or its membranes. They multiplied in the deep cervical lymph nodes, which sit in the neck and drain fluid from the brain, and to some degree in the spleen. In a closer copy of the real thing, using neurons that displayed the protein, multiplication happened only in those neck lymph nodes. It also happened only when cDC1s were present. The authors say brain material likely reaches the nodes through the meningeal lymphatic system, a network of drainage vessels in the membranes that wrap the brain.
A second knockout told the same story. To check that the briefing itself mattered, and not just the presence of the cells, the researchers bred tau mice lacking Wdfy4, a gene cDC1s need for cross-presentation. These mice still had cDC1s, but the cells couldn’t show their scraps to CD8+ T cells. The result looked familiar: better-preserved hippocampus, calmer glia and a sharp drop in CD8+ T cells in the brain.
The T cells themselves stir up trouble. In another test, the team took T cells from sick 9.5-month-old tau mice and injected them into the hippocampus of 6.5-month-old tau mice, an age with mild tau disease and no obvious shrinkage. Eight weeks later, these brains showed much stronger glial activation than brains given T cells from healthy mice. Volume and tau levels hadn’t changed yet. The T cells seemed to fire up the brain’s own immune cells before neurons started dying.
There are hints in people, but only hints. In brain tissue from eight people with primary tauopathies, such as progressive supranuclear palsy and Pick’s disease, CD8+ T cells were clearly more common than in five control brains. Data from human spinal fluid and brain membranes also suggested that dendritic cells stay stable with age and in Alzheimer’s, ready to present whatever they find. Whether cDC1s actually do this job in human brains remains unverified, the authors noted.
The caveats are real. This is a mouse study, and the mice were engineered to make human tau with a disease mutation, so they model the tau side of Alzheimer’s rather than the whole disease. The cDC1s were removed from birth, something that isn’t practical in people at risk. Some benefits were clear only in males. And the brain fragments that actually set off the T cells are still unknown. The team’s analysis turned up candidates, including pieces of tau and of neurofilament light chain, but none has been confirmed. The results would need to be replicated in human brains, Holtzman said.
The prize would be a drug that never has to enter the brain. Most dementia drugs have to get past the blood-brain barrier, the tight lining that keeps most substances out of brain tissue. If the trouble starts in the neck’s lymph nodes, a treatment could work there instead. It won’t be simple. Cutting off the brain’s lymphatic drainage has made things worse in other mouse models, so blocking the whole route is a bad idea. The authors suggest something narrower: stopping T cell priming before neurodegeneration sets in while leaving the drainage intact. Holtzman says his colleagues are already looking for targets that could dial these cells down in adulthood.
Source: Hu, H., Lin, P. B.-C., Zeng, C., Li, Y., Bosch, M. E., Emmerson, J. T., et al. “Priming of CD8+ T cells by peripheral dendritic cells exacerbates tau-mediated neurodegeneration.” Nature Neuroscience (2026). DOI: 10.1038/s41593-026-02427-5
This article summarizes published scientific research for general information. It is not medical advice and does not replace consultation with a qualified health professional.
