Trans-Synaptic Spread of Tau Pathology In Vivo

Alzheimer's disease doesn't strike the brain all at once. It starts in one small region and then traces the brain's own wiring outward, like a fire following a fuse. That's not a metaphor. Decades of post-mortem mapping have shown that tau pathology, the neurofibrillary tangle disease at the heart of Alzheimer's, begins in the entorhinal cortex and spreads from there in a predictable anatomical sequence. The question is why. Does it spread because neurons are passing something to each other across synapses, or does it just appear wherever the brain happens to be vulnerable? Liu and colleagues set out to answer that. Tau is a protein that normally stabilizes the internal scaffolding of neurons — the microtubules that give axons their structure. In Alzheimer's, tau becomes hyperphosphorylated, misfolds, and clumps into neurofibrillary tangles inside cells. The Braak staging framework, developed from systematic post-mortem mapping, shows that these tangles appear first in the entorhinal cortex, then in the hippocampus, and then in neocortical regions. It's one of the most reproducible patterns in neuroscience. But correlation with anatomy doesn't tell you about the mechanism. The regions that get sick in sequence are also anatomically connected in sequence, so you can't distinguish "the disease travels along the wiring" from "those regions are independently vulnerable" just by looking at human tissue. To solve this, Liu and colleagues built a transgenic mouse with a key restriction in the design. They drove expression of pathological human tau, specifically the four-repeat P301L mutation version, almost exclusively in the entorhinal cortex using a promoter for the neuropsin gene that produces robust activity in the superficial layers of the medial entorhinal cortex and little else. The logic is clean: if human tau later shows up in downstream brain regions where the transgene is not being expressed, it must have gotten there from the entorhinal cortex. They sampled animals at two ages: young mice around 10 to 11 months and old mice at 22 months or older, using three different tau antibodies to map the pathology. MC1 detects an abnormal tau conformation. CP27 is human-tau specific. AT8 marks tau phosphorylated at sites associated with disease. In the young mice, the picture was contained. Human tau immunoreactivity was concentrated in the superficial layers of the medial and lateral entorhinal cortex, primarily in axons and neurites rather than cell bodies. Crucially, it was also visible at the terminal zones of the perforant pathway — the axonal projection that connects entorhinal cortex layer two and three neurons to the hippocampus, delivering terminals to defined molecular layers of the dentate gyrus. This is tau sitting in its own axons, at the far ends of its own projections. The postsynaptic neurons, the dentate gyrus granule cells on the receiving end, showed no human tau accumulation at this stage. When the team looked for mature tangles using thioflavin-S staining, they found essentially nothing. Silver staining for argyrophilic material was likewise clean. This is the before picture: tau in the source region and its own axons, pathology not yet mature and not yet mobile. Then came the old mice. At 22 months and beyond, the distribution changed dramatically. Intense human tau immunoreactivity, detected by MC1, CP27, and AT8, appeared in the subiculum, in hippocampal pyramidal neurons especially in the CA1 region, and in dentate gyrus granule cells. Scattered immunoreactive neurons appeared in the perirhinal cortex and secondary somatosensory cortex. This is a disease that has moved. And crucially, these downstream regions do not express the transgene. The neuropsin promoter doesn't drive tau there. So the protein had to arrive from somewhere. There's a directional clue in where the tau went and where it went missing. In old mice, axonal tau at the perforant pathway terminal zones was significantly reduced, while tau accumulated in the cell bodies and dendrites of the postsynaptic neurons downstream. Tau relocated from axons to somatodendritic compartments in the source neurons and appeared in the somatodendritic compartments of the target neurons. That pattern is consistent with transfer across the synapse. Add to that the thioflavin-S result: mature, beta-sheet-rich tangles were found in a small number of medial entorhinal cortex neurons in old mice, and their formation in the source region coincided with the appearance of tau in downstream regions. The accumulation of tangles in the entorhinal cortex and the spread of pathology outward appear to be linked events. Now, there's an obvious objection. Maybe the neuropsin promoter is leaking. Maybe it's driving low-level tau expression in dentate gyrus granule cells directly, and what looks like spread is just ectopic expression. Liu and colleagues tested this rigorously. They used laser-capture microdissection to physically isolate approximately one thousand dentate gyrus granule cells per sample, distinguishing tau-protein-positive from tau-protein-negative neurons before capture. They then extracted RNA from those pooled cells and ran quantitative reverse transcription polymerase chain reaction with human-specific tau primers. The result was unambiguous on the specificity side: a non-transgenic control sample produced no human tau amplification, with a cycle threshold above 40, meaning essentially no signal. Putative genomic DNA contamination ran at least twenty-five-fold below the messenger RNA-derived signal in proper controls. What they found in the transgenic animals was low but detectable human tau messenger RNA in dentate gyrus granule cells. In one old mouse, tau-protein-positive granule cells had twenty-seven percent higher human tau messenger RNA than tau-protein-negative granule cells from the same animal. The team also crossed the neuropsin-tTA activator line to a nuclear lacZ reporter and to an amyloid precursor protein responder, and in both cases found minimal or no activity in dentate gyrus granule cells. The key argument: the sparse, low-level messenger RNA signal in granule cells cannot plausibly account for the widespread, intense human tau protein accumulation seen in old mice throughout the hippocampus. Ectopic expression is minimal. Transfer is the more likely explanation. What emerges from this work is a mouse model that recapitulates the earliest stages of Alzheimer's tauopathy — Braak stages one through three — in anatomical sequence, and the data support trans-synaptic propagation as the mechanism. Pathology follows the wiring. The entorhinal cortex projects to the subiculum, CA1, and dentate gyrus via the perforant pathway, and those are exactly the regions where tau appears next, in the same order that human Braak staging predicts. The model isn't just consistent with the human disease pattern; it actively tests and supports the connectivity hypothesis over the vulnerability hypothesis. The implication worth considering is this: if tau moves between neurons synaptically, then synapses are intervention points. A disease that propagates through the brain's own communication infrastructure could, in principle, be intercepted there. Liu and colleagues are explicit that this mouse model provides a platform for testing mechanisms and functional outcomes associated with disease progression. And the deeper significance of the finding is unsettling in a precise way — the brain's architecture, the very circuitry that enables memory and cognition, may be what Alzheimer's disease exploits to spread. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.