In Alzheimer disease and related disorders, the microtubule-associated protein tau aggregates and forms cytoplasmic lesions that impair neuronal physiology at many levels. In addition to affecting the host neuron, tau aggregates also spread to neighboring, recipient cells where the misfolded tau aggregates, in a manner similar to prions, actively corrupt the proper folding of soluble tau, and thereby impair cellular functions. One vehicle for the transmission of tau aggregates are secretory nanovesicles known as exosomes. Here, we established a simple model of a neuronal circuit using a microfluidics culture system in which hippocampal neurons A and B were seeded into chambers 1 and 2, respectively, extending axons via microgrooves in both directions and thereby interconnecting. This system served to establish two models to track exosome spreading. In the first model, we labeled the exosomal membrane by coupling tetraspanin CD9 with either a green or red fluorescent tag. This allowed us to reveal that interconnected neurons exchange exosomes only when their axons extend in close proximity. In the second model, we added exosomes isolated from the brains of tau transgenic rTg4510 mice (i.e. exogenous, neuron A-derived) to neurons in chamber 1 (neuron B) interconnected with neuron C in chamber 2. This allowed us to demonstrate that a substantial fraction of the exogenous exosomes were internalized by neuron B and passed then on to neuron C. This transportation from neuron B to C was achieved by a mechanism that is consistent with the hijacking of secretory endosomes by the exogenous exosomes, as revealed by confocal, super-resolution and electron microscopy. Together, these findings suggest that fusion events involving the endogenous endosomal secretory machinery increase the pathogenic potential and the radius of action of pathogenic cargoes carried by exogenous exosomes.

The cause of protein accumulation in neurodegenerative disease is incompletely understood. In Alzheimer9s disease (AD), the axonally enriched protein Tau forms hyperphosphorylated aggregates in the somatodendritic domain. Consequently, a process of subcellular relocalization driven by Tau phosphorylation and detachment from microtubules has been proposed. Here, we reveal an alternative mechanism of de novo protein synthesis of Tau and its hyperphosphorylation in the somatodendritic domain, induced by oligomeric amyloid‐β (Aβ) and mediated by the kinase Fyn that activates the ERK/S6 signaling pathway. Activation of this pathway is demonstrated in a range of cellular systems, and in vivo in brains from Aβ‐depositing, Aβ‐injected, and Fyn‐overexpressing mice with Tau accumulation. Both pharmacological inhibition and genetic deletion of Fyn abolish the Aβ‐induced Tau overexpression via ERK/S6 suppression. Together, these findings present a more cogent mechanism of Tau aggregation in disease. They identify a prominent role for neuronal Fyn in integrating signal transduction pathways that lead to the somatodendritic accumulation of Tau in AD. Synopsis Amyloid‐β (Aβ) promotes de novo protein synthesis of Tau and its hyperphosphorylation in the somatodendritic domain, in a process mediated by Fyn kinase‐dependent activation of ERK/S6 signaling. This finding presents a novel mechanism for somatodendritic Tau accumulation and aggregation in disease. Fyn kinase boosts Tau translation and phosphorylation. Aβ promotes translation of axonally enriched Tau in the somatodendritic domain. Aβ activates the ERK/S6 pathway via Fyn to induce Tau translation. Tau accumulates in Aβ‐depositing, Aβ‐injected and Fyn‐overexpressing brains. Fyn activity is required for Aβ‐induced Tau synthesis.

Alzheimer's disease is characterized by the accumulation of amyloid-β (Aβ) and Tau in the brain. In mature neurons, Tau is concentrated in the axon and found at lower levels in the dendrite where it is required for targeting Fyn to the spines. Here Fyn mediates Aβ toxicity, which is vastly abrogated when Tau is either deleted or a truncated form of Tau (Tau(1-255)) is co-expressed. Interestingly, MAP2, a microtubule-binding protein with mainly dendritic localization that shares Fyn-binding motifs with Tau, does not mediate Aβ's synaptic toxicity in the absence of Tau. Here we show in hippocampal neurons that endogenous Tau enters the entire spine, albeit at low levels, whereas MAP2 only enters its neck or is restricted to the dendritic shaft. Based on an extensive mutagenesis study, we also reveal that the spine localization of Tau is facilitated by deletion of the microtubule-binding repeat domain. When distinct phosphorylation sites (AT180-T231/S235, 12E8-S262/S356, PHF1-S396/S404) were pseudophosphorylated (with glutamic acid, using alanine replacements as controls), Tau targeting to spines was markedly increased, whereas the pseudophosphorylation of the late phospho-epitope S422 had no effect. In determining the role physiological Fyn has in the spine localization of Tau, we found that neither were endogenous Tau levels reduced in Fyn knockout compared with wild-type synaptosomal brain fractions nor was the spine localization of over-expressed pseudophosphorylated or P301L Tau. This demonstrates that although Fyn targeting to the spine is Tau dependent, elevated levels of phosphorylated Tau or P301L Tau can enter the spine in a Fyn-independent manner.