Extracellular PHF-tau modulates astrocyte mitochondrial dynamics and mediates neuronal connectivity

Abstract

Background: Tau is an intracellular protein that plays a crucial role in stabilizing microtubules. However, it can aggregate into various forms under pathological conditions and be secreted into the brain parenchyma. While the consequences of tau aggregation within neurons have been extensively studied, the effects of extracellular paired helical filaments of tau (ePHF-tau) on neurons and astrocytes are still poorly understood.

Methods: This study examined the effect of human ePHF-tau (2N4R) on primary cultures of rat neuroglia, focusing on changes in neurites or synapses by microscopy and analysis of synaptosome and mitochondria proteomic profiles after treatment. In addition, we monitored the behavior of mitochondria in neurons and astrocytes separately over three days using high-speed imaging and high-throughput acquisition and analysis.

Results: ePHF-tau was efficiently cleared by astrocytes within two days in a 3D neuron-astrocyte co-culture model. Treatment with ePHF-tau led to a rapid increase in synaptic vesicle production and active zones, suggesting a potential excitotoxic response. Proteomic analyses of synaptosomal and mitochondrial fractions revealed distinct mitochondrial stress adaptations: astrocytes exhibited elevated mitochondrial biogenesis and turnover, whereas neuronal mitochondria displayed only minor oxidative modifications. In a mixed culture model, overexpression of tau 1N4R specifically in astrocytes triggered a marked increase in mitochondrial biogenesis, coinciding with enhanced synaptic vesicle formation in dendrites. Similarly, astrocyte-specific overexpression of PGC1alpha produced a comparable pattern of synaptic vesicle production, indicating that astrocytic mitochondrial adaptation to ePHF-tau may significantly influence synaptic function.

Conclusions: These findings suggest that the accumulation of PHF-tau within astrocytes drives changes in mitochondrial biogenesis, which may influence synaptic regulation. This astrocyte-mediated adaptation to tauopathy highlights the potential role of astrocytes in modulating synaptic dynamics in response to tau stress, opening avenues for therapeutic strategies aimed at astrocytic mechanisms in the context of neurodegenerative diseases.

Keywords: Astrocytes; Live imaging microscopy; Mitochondria; Synapse; Tau.

Fig. 1

ePHF-tau is internalized by astrocytes in a neuron-glia 3D cell culture model. a Experimental setup: primary neuron-glia 3D co-cultures were treated with ePHF-tau, followed by live imaging for 72 h and post-fixation immunofluorescence (IF) for astrocytic (GFAP) and neuronal (NeuN, NFL) markers. b Overview of the neuron-glia 3D cultures showing neurospheroids (NS) and their surrounding microenvironment, including the peri-NS region and neurite sheaths. Scale bar: 100 µm. Examples at higher magnification highlight astrocytes (top right) and axonal/spine structures (bottom right). Scale bar: 10 µm. c Time-lapse series showing the gradual disappearance of ePHF-tau aggregates over 60 h. Red arrows point to the aggregates, with an inset providing a higher magnification of an aggregate being cleared at a specific location. Scale bar: 50 µm. d Time series quantifying ePHF-tau clearance per region of interest (ROI) in treated (red) versus control (DMSO, black) conditions, segmented using an AI-based model. Shaded areas represent SEM. e Time-lapse images of ePHF-tau uptake by astrocytes at different time points after aggregate addition (10, 20, 30, and 40 min). Red and yellow arrowheads indicate ePHF-tau aggregates, and green arrows point to astrocytes. Scale bar: 50 µm. f A confocal photograph of the neurospheroid positive for DAPI and the histogram represents the number of neurospheroids in the cultures. g A confocal photograph of astrocytic GFAP labelling in neuropsheroid and histogram presenting the GFAP surface by neuroscpheroide. h A confocal photograph of neuronal NeuN labelling in neuropsheroids and histogram showing NeuN surface by neuroscpheroid. i A confocal photograph of neuronal NFL labelling in neuropsheroids and histogram showing NFL surface by neuroscpheroide. (n = 3 cultures, 25 neurospheroids, 5 ROIs per condition;)

Fig. 2

ePHF-tau induces modifications in synaptic markers in neuronal cultures. a Experimental design: treatment with ePHF-tau, live imaging over 72 h, followed by immunostaining for pre- (VGLUT1) and postsynaptic (PSD-95) markers. Bottom: brightfield and fluorescence images showing synaptic vesicle detection (light blue dots) in neurite sheaths. Scale bar: 50 µm. b Time series of synaptic vesicle density in control (black) and ePHF-tau-treated (red) conditions, relative to the time of treatment. Statistical comparisons for each time points (multiple t-tests) are shown as with gray bars on the right Y-axis as -log(P-value) and comparison over complete days 1, 2, and 3 are shown with asterisks (*) above the graph. c Fluorescence images of synaptic dye at 24, 48, and 72 h, with blue circles marking segmented synaptic vesicles in control and treated conditions. d Combined illustration and methodological overview of neurospheroid immunostaining analysis. The inset shows an example of a 3D volume of interest (VOI) analyzed for synaptic markers. The schematic on the right illustrates the approach used for segmenting active zones and postsynaptic compartments. Scale bar: 100 µm. e 3D visualization of the active zone (gray) along with VGLUT1 (magenta) and PSD-95 (green) in segmented regions of interest (ROI). Scale bar: 5 µm. f–h Quantification of total VGLUT1 volume (f), total PSD-95 volume (g), and active zone volume normalized to PSD-95 surface area (h) in PERI-NS sheath (n = 3 cultures, 25 neurospheroids, 5 ROIs per condition; ***P < 0.001)

Fig. 3

Proteomic analysis of synaptosome and mitochondrial fractions reveals differential protein regulation under ePHF-tau treatment. a Experimental workflow to investigate the interactions of ePHF-tau in a rat primary mixed culture with 72-h ePHF-tau treatment, followed by isolation of synaptosome-enriched fractions (SEF; n = 3 for DMSO, n = 3 for ePHF-tau) and mitochondria-enriched fractions (MEF; n = 4 for DMSO, n = 4 for ePHF-tau) for proteomic analysis. b Quality check of SEF fractions: Synaptic marker enrichment (presynaptic and postsynaptic proteins) in DMSO and tau-treated conditions, confirming the expected protein content in SEF fractions. c Quality check of SEF fractions based on cell-type enrichment markers (astrocyte, microglia, neuron, oligodendrocyte, and endothelial) showing representation of different cell types within the synaptosome fractions under DMSO and tau-treated conditions. d Differentially expressed functional annotations (2D enrichment) in SEF. Enriched terms are shown in red region, and depleted terms are shown in blue region. Dots colors indicate the annotation types. e–g Box plots comparing the relative abundance of MAPT (e), GFAP (f), and TUBB3 (g) between DMSO and ePHF-tau conditions, showing differential regulation of these proteins in the MEF. h Protein–Protein interaction network of the modified proteins in SEF/MEF

Fig. 4

ePHF-tau negatively affects mitochondrial turnover in neurites. a Experimental design: primary mixed cultures transduced with mitochondrial sensors (MitoTimer) were imaged over 72 h in the presence of ePHF-tau. b Diagram of the MitoTimer biosensor measuring mitochondrial morphology and turnover/redox state. c Data normalization process: mitochondrial features were normalized to baseline (BL) and then compared to controls. d Mitochondrial segmentation illustrating the focus on individual mitochondria (purple) versus clusters (green) in neurons (left) and astrocytes (right). Scale bar: 10 µm. e Representative Mitotimer fluorescence images in neuronal neurites (left) and astrocytic processes (right) at baseline and > 60 h after ePHF-tau treatment. Red arrows indicate swollen, red mitochondria. Scale bar: 10 µm. f–t Comparative analysis of mitochondrial features in and between neurons and astrocytes. Each row represents a specific feature: fluorescence intensity at 555 nm (f, g, h), 488 nm (i, j, k), redox state (l, m, n), mitochondrial surface area (o, p, q), and mitochondrial count (r, s, t). The left columns (f, i, l, o, r) show the effects of ePHF-tau treatment (red lines) versus control (black lines) in neurons, the middle columns (g, j, m, p, s) depict the same comparison in astrocytes, and the right columns (h, k, n, q, t) compare neurons (blue lines) with astrocytes (green lines) relative to their respective controls. Gray bars in each graph represent -log(P-value) for comparisons at each time point on the right y-axis. Asterisks above the graphs indicate P-values for comparisons over entire days (means and SEM, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

Fig. 5

Astrocytic modifications alone enhance synaptic vesicle density in neurons. a Experimental design: primary mixed cultures with astrocytes infected to overexpress WT human Tau-1N4R-V5 or PGC-1α were stained with a synaptic dye and imaged over 72 h in the presence of ePHF-tau. b Immunofluorescence images of astrocytes (GFAP) stained for Tau-V5 tag after 72 h treatment. Quantification of Tau-V5 intensity in peri-nuclear space of GFAP-positive cells is shown on the right. Scale bar, 50 μm; dashed blue line: peri-nuclear space. c Microscopic images of cultures infected with G1B3-Tau-1N4R-V5 and stained for V5 (green) and TFAM (red). Cells can be positive or negative for V5 expression (as displayed by text on merge image). Right, a bar plot of the quantification of the mean TFAM intensities in nucleus. d Time series of synaptic vesicle density in neurons co-cultured with G1B3-Tau-1N4R-V5 astrocytes (orange) versus DMSO-treated controls (black). Statistical comparisons for each time point (multiple t-tests) are shown as gray bar on the right Y-axis as -log(P-value) and statistical comparison for data subsets aggregated by day is shown with asterisks (*) above the graph. e Fluorescence images of synaptic dye at baseline and 72 h in control, 1N4R, and PGC-1α conditions, with blue circles marking segmented synaptic vesicle. Scale bar, 10 μm. f Same as (c) but with G1B3-PGC-1α infection of astrocytes. g Same as (d), with G1B3-PGC-1α infection of astrocytes. *P < 0.05, **P < 0.01, ***P < 0.001