Selective vulnerability of human synapses to soluble tau oligomers

Abstract

Background

Soluble tau oligomers (tauO) are early, synaptotoxic drivers of dysfunction in tauopathies. While selective vulnerability is well documented at the cellular level, emerging evidence suggests that synaptic subtypes may differ in their susceptibility to tau pathology. Still, the key factors that shape synaptic vulnerability to toxic tauO, particularly in humans, remain poorly understood.

Objective

To define the synaptic compartments and subtypes most vulnerable to tauO and identify molecular correlates underlying this susceptibility.

Methods

Synaptosomes were isolated from cognitively normal human autopsy specimens and acutely challenged with preformed recombinant tauO. Flow cytometry with multiplexed immunophenotyping resolved tauO engagement across intact pre- and postsynaptic compartments and excitatory versus inhibitory subtypes. Functional effects were assessed by microtransplanting synaptic membranes into Xenopus laevis oocytes and recording ligand-gated GABAergic and glutamatergic responses. Complementary LC-MS/MS proteomics of brain-derived tau oligomers (BDTO) from PBS-soluble hippocampal lysates of primary age-related tauopathy (PART) cases were analyzed using SynGO enrichment to identify molecular correlates.

Results

TauO preferentially engaged presynaptic compartments and showed elevated association with GABAergic synapses. Functionally, acute tauO exposure selectively enhanced GABA_AR-mediated responses, with no effect on AMPAR-mediated currents. The PART BDTO interactome was enriched for presynaptic vesicle-associated proteins involved in vesicle cycling and neurotransmitter release, consistent with a presynaptic axis of vulnerability.

Conclusions

This integrative analysis identifies a compartment- and subtype-specific vulnerability of human synapses to tauO, highlighting a presynaptic inhibitory bias as a potential driver of synaptic dysfunction and tau propagation in early-stage tauopathies.

Keywords

Alzheimer's disease primary age-related tauopathy (PART)synaptic vulnerability tau oligomers

Introduction

Tau is a microtubule-associated protein canonically implicated in axonal transport and cytoskeletal stability, but is now recognized also as a key modulator of synaptic architecture, signaling, and plasticity.^1–5 In tauopathies—neurodegenerative disorders characterized by aberrant tau accumulation—tau adopts misfolded conformations that assemble into oligomers, paired helical filaments, and ultimately neurofibrillary tangles.⁶ Among these, soluble tau oligomers (tauO) have garnered particular attention as early, synaptotoxic species that correlate strongly with cognitive decline.^7–11 TauO have also been implicated in the trans-synaptic propagation of pathology via prion-like propagation, further implicating the synapse as both a vulnerable target and a conduit for early tau-mediated injury.^12–16

Within this framework, emerging evidence suggests that the accumulation and functional consequences of tauO may not be uniformly exerted across all synapses.¹⁷ While selective vulnerability has long been appreciated at the cellular level, where specific neuronal populations exhibit disproportionate susceptibility to tau pathology, recent studies indicate that distinct synaptic subtypes, distinguished by features such as cytoarchitecture, neurotransmitter signaling, or anatomic localization, may differ also in their capacity to engage and respond to pathological tau species.^17–20 Yet, the structural, functional, and molecular determinants of this selective synaptic vulnerability, especially as they pertain to tauO, remain largely ill defined.

To begin addressing this gap, we recently examined autopsy brain specimens from individuals with primary age-related tauopathy (PART), leveraging PART as an amyloid-free model to isolate synapse-specific vulnerability to tau pathology in a companion study.²¹ Using complementary biochemical and electrophysiological approaches, we quantified native tau species and assessed ligand-gated synaptic responses from the same tissue; we found that elevated tauO burden in PART was correlated with a reduction in the synaptic excitation-to-inhibition (sE/I) ratio. However, whether this shift reflects diminished excitation, enhanced inhibition, or a more complex interplay remains unresolved. In particular, it is unclear whether tauO preferentially engage and modulate specific functional synaptic subtypes (excitatory or inhibitory) or synaptic compartments (presynaptic or postsynaptic).

To further elucidate the structural, functional, and molecular determinants of synaptic vulnerability to tauO, here we implemented a translationally relevant platform using human synaptosomes derived from autopsy brain specimens of aged, cognitively normal individuals. Synaptosomes were acutely challenged with preformed recombinant tauO, and synapse-specific tauO engagement was quantified via flow cytometry using a multiplexed synaptic immunophenotyping panel. To evaluate changes in synaptic function, synaptic membranes from tauO-treated synaptosomes were microtransplanted into Xenopus laevis oocytes, enabling measurement of ligand-gated GABAergic and glutamatergic responses. Complementing these structural and functional analyses, we profiled the synaptic interactome of brain-derived tau oligomers (BDTO) isolated from PBS-soluble hippocampal homogenates of PART autopsy specimens, aiming to identify candidate proteins and pathways associated with synaptic vulnerability to soluble tauO.

Our findings suggest that tauO preferentially engage presynaptic terminals, with additional preference for inhibitory synapses. Functionally, tauO exposure selectively potentiated GABAergic currents, while glutamatergic (i.e., kainate-sensitive AMPAR) responses remained unaffected. Proteomic analysis of PART BDTO revealed a vesicle-associated interactome enriched in presynaptic release machinery. Together, these findings delineate a synapse-type- and compartment-specific pattern of oligomeric tau engagement and implicate presynaptic inhibitory terminals as selectively vulnerable targets in early-stage tauopathy (Figure 1).

Figure 1.

Schematic overview of the experimental workflow. This study investigates selective synaptic vulnerability to tauO using a multi-faceted approach integrating cellular, molecular, structural, and functional techniques. (Left panel) Synaptosomes isolated from Control human SMTG or WT mouse cortex were treated ± rtauO and analyzed via FC, EM, and TEVC electrophysiology. (Right panel) BDTO were isolated from PART hippocampal tissue and analyzed via LC-MS/MS proteomics and SynGO enrichment to identify synapse-specific pathways and interactomes. This integrative strategy enables the characterization of selective synaptic vulnerability to tauO across pre- versus postsynaptic and excitatory versus inhibitory compartments/subtypes.

Methods

Clinical and pathological profiling of human postmortem samples

Human postmortem brain tissue was obtained through the Alzheimer's Disease Research Center (ADRC) at the Sanders-Brown Center on Aging, University of Kentucky, under established material transfer agreements. All participants were enrolled in longitudinal studies of aging and dementia, with informed consent obtained prior to enrollment. All procedures adhered to institutional IRB guidelines.

Annual clinical assessments were conducted by experienced clinicians trained in cognitive and neurological evaluation, including administration of the Mini-Mental State Examination (MMSE) and the Clinical Dementia Rating (CDR) scale. Standardized neuropathological evaluations were performed by board-certified neuropathologists using established criteria from the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) and Braak staging to assess AD pathology. Control cases exhibited no cognitive impairment (CDR = 0, MMSE ≥ 26) and were free of AD-related neuropathological changes. Cases classified as PART demonstrated mild to moderate tau pathology (Braak stages I–II), predominantly confined to the hippocampus, in the absence of Aβ plaques, in accordance with established diagnostic criteria.^22–24 Notably, previous work from our group confirmed the absence of detectable Aβ pathology in synaptosomal preparations derived from this PART cohort.¹³ Individuals carrying the APOE ε4 allele were excluded from the present study, as APOE ε4 has been shown to modulate tau's protein interactome,²⁵ influence the extent and topology of network-level tau spreading,²⁶ and disrupt NMDAR function.²⁷

All specimens were de-identified prior to transfer to the University of Texas Medical Branch (UTMB), rendering their use exempt from additional IRB oversight under CFR §46.101(a)(1). Comprehensive demographic, clinical, and neuropathological characteristics of all subjects are provided in Supplemental Table 1.

Isolation and preparation of synaptosomes from human brain tissue

Standardized protocols for synaptosome isolation, quality control, and downstream applications have been optimized and validated in our vast prior work.^13,28–31 In the present study, synaptosome integrity and synaptic enrichment were confirmed by transmission electron microscopy and LC-MS/MS proteomic profiling with subsequent SynGO analysis (Supplemental Figure 1).

For FC experiments, synaptosomes were isolated from frozen autopsy samples of the superior middle temporal gyrus (SMTG) obtained from cognitively intact Control subjects (n = 8; 4 male, 4 female). To ensure preservation of synaptic ultrastructure and minimize degradation, only samples with a postmortem interval of 5 h or less were used. In brief, 20 mg of snap-frozen tissue were homogenized in ice-cold Syn-PER reagent (ThermoFisher Scientific; cat# 87793) at a ratio of 10 μL per mg of tissue, supplemented with 1% protease and phosphatase inhibitor cocktails. Homogenates were centrifuged at 1200 × g for 10 min at 4°C to clear myelin, nuclei, and other cellular debris. Resulting supernatants were extracted and further centrifuged at 15,000 × g for 20 min at 4°C to isolate the crude synaptosomal (P2) fraction. Synaptosome pellets were gently resuspended in 100 µL of freshly prepared HEPES-buffered Krebs-like solution (HBK; 143.3 mM NaCl, 4.75 mM KCl, 1.2 mM MgSO₄·7H₂O, 1.2 mM CaCl₂, 20.1 mM HEPES, 0.1 mM NaH₂PO₄, and 10.3 mM D-glucose; pH adjusted to 7.4). HBK was used within one week of preparation to maintain buffer stability and reduce risk of microbial contamination. To prevent synaptosome clumping, 0.5% Pluronic F-68 non-ionic surfactant (ThermoFisher Scientific; cat# 24040-032) was added to the suspension. Synaptosome concentrations were determined using a Guava EasyCyte 8 flow cytometer (MilliporeSigma, Burlington, MA, USA), based on event counts within a predefined size gate corresponding to synaptosomal dimensions, as described in Kadamangudi et al., 2024.¹³ All samples were diluted or concentrated to achieve a final working concentration of 1 × 10⁶ synaptosomes/μL for downstream applications.

For MSM (microtransplantation of synaptic membranes) experiments, synaptosomes were prepared as described above. Following the final centrifugation step, P2 pellets were resuspended in Syn-PER containing protease and phosphatase inhibitors (2 µL per mg of starting tissue). Total protein concentrations were quantified fluorometrically using a QFX fluorometer (DeNovix, New Castle, DE, USA), and volumes were adjusted to a final concentration of 2 mg/mL. Aliquots were immediately stored at −80°C to prevent degradation from repeated freeze–thaw cycles.

Preparation and fluorescent labeling of preformed recombinant tau oligomers

Preformed recombinant tau oligomers (rtauO), generated from monomers derived from the full-length (2N4R) tau construct, were kindly provided by the laboratory of Dr Rakez Kayed and prepared in accordance with previously established protocols.^32–34 All procedures involving rtauO utilized low-retention plasticware to minimize protein loss.

Fluorescent labeling of rtauO was performed using the Microscale Protein Labeling Kit with Alexa Fluor™ 488 (Thermo Fisher Scientific, cat# A30006), following the manufacturer's instructions. Labeling conditions were optimized for full-length human tau (45.9 kDa) to ensure efficient conjugation and reproducibility across preparations. Briefly, tauO were diluted to 1 mg/mL in 1× Dulbecco's phosphate-buffered saline (PBS; Corning, cat# 46-013-CM). Alexa Fluor 488 dye was added at a dye-to-protein molar ratio of 50:1 to achieve an optimal degree of labeling (DOL). The reaction mixture was gently mixed and incubated for 30 min at room temperature. Following conjugation, excess unbound dye was removed via spin filtration using Bio-Gel P-6 fine resin columns, as recommended in the kit protocol. The DOL and post-labeling protein concentration were determined by measuring absorbance at 280 nm (A280), adjusted for fluorophore-specific absorbance, using a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). Protein concentrations were further validated using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, cat# 23227). Efficient fluorophore conjugation and successful removal of unreacted dye were confirmed by UV-Vis spectroscopy, which showed a strong absorbance peak at 495 nm in the purified rtauO fraction and negligible signal in the flow-through (Supplemental Figure 2a). To confirm structural integrity, western blot analysis was routinely performed to verify maintenance of the oligomeric conformation following labeling (Supplemental Figure 2b).

Flow cytometry

rtauO challenge and immunophenotyping of human synaptosomes for flow cytometry. Both the synaptosome rtauO challenge assay and flow cytometry (FC) immunophenotyping protocols utilized in this study were adapted from our previously published methods.^13,29 Here, we briefly outline each protocol separately and then describe their integration to assess selective rtauO engagement across distinct synaptic compartments and functional subtypes. FC experiments employed pooled synaptosomes (n = 8 subjects per pool) isolated from autopsy SMTG specimens from healthy Control individuals, minimizing potential confounding effects of pre-existing tau pathology. We previously demonstrated consistent rtauO engagement across individual preparations²⁹; thus, sample pooling enhanced sample yield and experimental feasibility. Synaptosome aliquots were rapidly thawed at 37°C for no more than two minutes to minimize aggregation.^35,36 To reduce sample loss, low-retention pipette tips and microcentrifuge tubes were consistently utilized. Samples were maintained on ice unless otherwise specified, with centrifugation and wash steps performed at 10,000 × g for 5 min at 4°C.

For synaptic rtauO challenge, synaptosomes were incubated with Alexa Fluor 488-conjugated rtauO for 1 h at room temperature (RT) in a final volume of 20 µL HBK. Samples were then washed three times with HBK and either analyzed immediately or stored at 4°C for no more than 24 h before analysis. Prior to reading, samples were diluted with 200 µL DPBS and gently vortexed. To determine optimal conditions for rtauO challenge, titration experiments were performed by fixing either the synaptosome concentration (100,000 synaptosomes/µL) while varying rtauO concentrations (2.5 nM to 2.0 µM), or fixing the rtauO concentration (25 nM) while varying synaptosome concentrations (3700 to 300,000 synaptosomes/µL). These experiments revealed that 25 nM rtauO produced a near half-maximal engagement response (∼60%), and that synaptosome concentrations below 33,000 synaptosomes/µL yielded maximal signal intensity; accordingly, these parameters were adopted for all subsequent experiments (Supplemental Figure 3).

Prior to initiating immunophenotyping experiments, the following solutions and buffers were prepared: HBK buffer (recipe outlined in Isolation and preparation of synaptosomes from human brain tissue section); Sucrose/EDTA/Tris buffer (SET buffer; 320 mM sucrose, 1 mM EDTA, and 5 mM Tris base [pH 7.4]); CellTrace neutralization buffer (1% bovine serum albumin [BSA] in HBK); fixation solution (4% paraformaldehyde [PFA] in PBS); blocking-only buffer (5% BSA in SET buffer); blocking + permeabilization buffer (5% BSA and 0.1% Tween-20 in SET buffer); and antibody diluent (0.5% BSA and 0.1% Tween-20 in SET buffer). For synaptosome immunophenotyping, samples underwent four key steps: (1) CellTrace labeling, (2) fixation, (3) blocking and permeabilization, and (4) antibody staining. Synaptosomes (∼1 × 10⁶ per reaction) were labeled with 5.0 µM CellTrace dye in 100 µL HBK buffer at 37°C for 20 min, protected from light. The exact starting quantity of synaptosomes and the reaction volume are not critical, provided that the specified working concentrations are maintained. Excess dye was neutralized with CellTrace neutralization buffer (5× volume), followed by centrifugation, resuspension, and incubation in 100 µL SET buffer for 10 min at RT to allow acetate hydrolysis. Synaptosomes were then fixed by adding an equal volume of fixation buffer to achieve a final concentration of 2% PFA and incubated for 15 min at RT. After fixation, synaptosomes were washed with blocking-only buffer to neutralize residual PFA, followed by a simultaneous blocking and permeabilization step via incubation in blocking + permeabilization buffer for 30 min at RT. Primary and secondary antibodies were diluted in antibody diluent and sequentially incubated (100 µL) for 30 min each at RT, with two intermediate washes using SET buffer. Working concentrations and item specifications of antibodies are provided in Supplemental Table 2. Following both antibody incubation steps, samples were washed twice, resuspended in SET buffer, and analyzed immediately or stored at 4°C for up to 24 h. Before reading, samples were diluted with 200 µL DPBS and gently vortexed.

The antibody panel used for synaptosome immunophenotyping was selected based on four key criteria: (1) compartment specificity of the target protein, (2) commercial validation of the antibody for FC applications, (3) availability of pre-conjugated fluorophore formats to eliminate the need for secondary staining, and (4) minimal spectral overlap among selected fluorophores. All antibodies used in multiplexed panels, except for STX1A and VAMP2, were pre-conjugated. For STX1A and VAMP2, secondary antibodies were included where applicable. Antibodies were titrated against isotype-matched controls to confirm specificity and optimize signal-to-noise ratios (Supplemental Figure 4). To ensure reliable colocalization and distinguishability of synaptic subpopulations, fluorophores were chosen to allow resolution of both positive and negative populations across at least two, and preferably three markers per panel (Supplemental Figure 5). Co-staining experiments confirmed expected colocalization patterns for pre- and postsynaptic markers (Supplemental Figure 6), and final panel design was iteratively optimized to balance resolution, signal intensity, and fluorophore compatibility for multiplexed FC.

To assess rtauO engagement across specific synaptic compartments and functional subtypes, the rtauO challenge was integrated within the immunophenotyping workflow. Following CellTrace labeling and neutralization, synaptosomes were centrifuged and resuspended in 100 µL HBK (assuming a starting amount of ∼1 × 10⁶ synaptosomes) in lieu of SET buffer, then divided equally (50 µL each) into rtauO-positive (rtauO⁺) and rtauO-negative (rtauO^–; HBK-only) groups for incubation at RT for 1 h. Wash steps were performed in HBK as described above, followed by resuspension in SET buffer. Subsequent fixation, blocking, permeabilization, and antibody incubation steps adhered strictly to the standard immunophenotyping protocol and were performed in parallel for rtauO⁺ and rtauO^– groups. When multiple samples were processed simultaneously, procedural steps through blocking and permeabilization were scaled proportionally. For example, to immunophenotype five samples, 5 × 10⁶ synaptosomes were labeled in 500 µL HBK, then evenly divided into rtauO⁺ and rtauO^– groups (250 µL each). Following blocking and permeabilization, each group was further split into five 50 µL aliquots for multiplexed antibody staining. All wash steps and reagent volumes were proportionally scaled to maintain consistent working concentrations.

Instrumentation and acquisition parameters for flow cytometry small particle measurement

FC experiments were performed using either the Guava EasyCyte 8 (EMD Millipore, Burlington, MA) or the BD FACSymphony A5 SE (BD Biosciences, San Jose, CA), with instrumentation selected based on experimental needs. With the exception of updated optical configurations for additional probes and fluorophores, all instrument settings and acquisition parameters were adopted directly from our recently published work.¹³ These updates, implemented on the BD FACSymphony A5 SE, enabled expanded multiplexed detection of synaptic markers with minimal spectral interference.

The Guava EasyCyte 8 was used exclusively for synaptosome enumeration and rtauO titration experiments due to its built-in absolute particle counting capability, which obviates the need for reference beads. Synaptosomes were acquired at a flow rate of 0.5 µL/s (“Low” setting) using SSC triggering. A total of 10,000 events within the designated size gate were acquired per sample.

Synaptosome immunophenotyping was conducted on the BD FACSymphony A5 SE flow cytometer. This instrument was selected for its multispectral detection capabilities, supported by five spatially separated lasers and a dedicated array of detectors. These features enabled efficient multiplexing and precise resolution of diverse synaptic sub-populations. Detailed optical configurations are provided in Supplemental Table 3. Synaptosomes were acquired using dual-threshold triggering on side scatter (SSC) and CellTrace Blue fluorescence to ensure reliable detection of structurally intact particles, as fluorescence-based thresholding has been shown to significantly outperform standard FSC-based detection for small particle analysis.³⁷ Data were collected using BD FACSDiva Software v9.0, with fluorescence compensation applied at the start of each run using single-color compensation beads (ThermoFisher, cat# 01-2222-42) and matrix calculations performed according to standard cytometry practices.³⁸ Samples were acquired for 3 min per condition.

To ensure accurate and reliable detection of small particles such as synaptosomes, several methodological considerations were implemented across both platforms. Instrument sensitivity and size calibration were assessed on both instruments prior to each run using standardized ApogeeMix “Micro” (cat# 1493, Apogee Flow Systems) and Spherotech polystyrene (Spherotech Inc., Lake Forest, IL, USA) bead standards. These calibrations supported consistent definition of the synaptosome size gate (0.5–1.3 µm) (Supplemental Figure 7). Optimization of linear detection range and event rate fidelity for both instruments has been described previously.¹³ Lastly, to minimize coincidence detection and particle aggregation, we implemented several additional quality control measures, including shortened centrifugation steps and spin time, optimal sample dilution, and the use of non-ionic buffers when feasible.^13,37

Flow cytometry data analysis

FC data analysis, including gating, quantification, and the generation of histograms and density plots, was performed using FlowJo v10 (BD Life Sciences, Ashland, OR). Synaptic compartments and functional subtypes were differentiated using Boolean (conditional) gating strategies, defined by the co-expression or exclusion of established synaptic markers. Gating thresholds were determined using antibody-negative (antibody^–) controls, ensuring identical implementation across rtauO⁺ and untreated rtauO⁻ samples. Importantly, rtauO⁻ samples served as internal gating references, facilitating accurate delineation of rtauO⁺ events within rtauO treated samples. A comprehensive list of the conditional gating strategies used to define individual synaptic populations is provided in Supplemental Figure 5, and a representative gating workflow is illustrated in Supplemental Figure 7.

While the identification of rtauO⁺ synaptosomes provided qualitative confirmation of oligomer engagement, quantitative analysis was based on the median fluorescence intensity (MFI; detector: B537; fluorophore: Alexa-488) of the entire synaptic subpopulation. This approach is particularly advantageous for analyzing small particles such as synaptosomes, where inherent signal heterogeneity and optical limitations frequently hinder accurate discrimination between dim and truly negative events—a potential limitation of percentage-based metrics in this context.³⁹ Furthermore, MFI was chosen as the primary readout due to its sensitivity in capturing the full range of fluorescence intensity, thus enabling robust detection of shifts in oligomer engagement.

To account for technical variability and background signal, MFI values for each synaptic subpopulation were normalized in two stages. First, MFI values were divided by the MFI of the total rtauO⁺ synaptosome population to control for inter-sample variation in rtauO delivery. Second, this value was further normalized to the MFI of the same synaptic subpopulation in rtauO⁻ samples, thereby correcting for background fluorescence, autofluorescence, or other sources of optical noise. The resulting normalized MFI values were used for all downstream analyses. Samples were analyzed in technical duplicate within the same experimental day, and each experimental paradigm was repeated on separate days to confirm reproducibility. All MFI metrics and event counts were exported using FlowJo's Batch Export function, with subsequent statistical analyses conducted in GraphPad Prism v10.0 (GraphPad Software, La Jolla, CA).

Assessment of synaptic function via microtransplantation of synaptic membranes

Protocols for isolating synaptosomes, including detailed characterization and stringent quality control procedures, have been optimized and validated in our vast prior studies.^13,28–31 Similarly, our group has rigorously established and validated methodologies for assessing synaptic functionality through microtransplantation of synaptic membranes (MSM) into Xenopus laevis oocytes.^31,40–43 Below, we summarize key steps of this procedure.

For MSM experiments, pooled synaptosomes (n = 4 per pool) derived from autopsy specimens of the SMTG from healthy Control individuals were utilized. Pool size was limited to four samples based on a preliminary functional screen assessing baseline γ-aminobutyric acid (GABA) and kainate receptor responses (Supplemental Table 4). Specifically, synaptosome preparations from all available Control cases were screened, and the samples with the highest average receptor responses were selected for pooling. As shown previously, measurable electrophysiological responses to GABA and kainate were observed only in oocytes microtransplanted with synaptic membranes, with GABA selectively activating GABA_A receptors and kainate selectively activating AMPA receptors.⁴⁰

rtauO challenge experiments were conducted in a manner analogous to our FC experiments, with minor modifications. Briefly, synaptosomes were incubated with either rtauO or HBK-only control for 1 h at room temperature in a final volume of 20 µL HBK, followed by three sequential washes in HBK (10,000 × g for 5 min at 4°C). To achieve optimal transplantation of synaptic receptors on to Xenopus oocyte membranes, the final protein concentration of rtauO-treated synaptosomes was carefully maintained at a minimum of 2 µg/µL.^31,40,42 Accordingly, 10 µg of total synaptosomal protein was used per condition, allowing the final pellet to be resuspended in 5 µL HBK. This preparation yielded sufficient material for multiple injection rounds, as each oocyte required only ∼50.6 nL (detailed below), typically supporting the injection of 5–10 oocytes per sample. Protein concentrations were confirmed using a DeNovix QFX fluorometer (DeNovix, New Castle, DE, USA). Treated synaptosomes were either immediately injected or stored at −80°C until use.

Xenopus laevis oocytes were harvested in compliance with guidelines stipulated by the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee (IACUC protocol #1803024) at UTMB. Briefly, frogs were anesthetized in a 0.17% MS-222 bath, humanely euthanized, and ovaries surgically excised. Oocytes were subsequently isolated and defolliculated via gentle agitation in 1× Barth's solution (SBS; Ecocyte BioScience Inc, Dortmund, Germany) supplemented with 2 mg/mL collagenase (Sigma-Aldrich, St Louis, MO, USA) for 4 h at 30°C. Healthy stage V and VI oocytes were selected under a light microscope and stored in fresh SBS at 16°C until microinjection.

Microinjection needles were generated from 3.5-inch borosilicate glass tubes (Drummond Scientific, PA, USA) using a micropipette puller (Model P-1000; Sutter Instruments, Novato, CA, USA), and tips were trimmed uniformly to a length of approximately 4 mm to facilitate consistent injection volumes. Prior to injection, oocytes were transferred individually to SBS. Each microinjection needle, freshly prepared for each sample, was prefilled with 0.5 mL/cc mineral oil and mounted onto a Nanoject II microinjector (Drummond Scientific, PA, USA). Synaptosome preparations were adjusted to a concentration of 2 ng/nL, homogenized by ultrasonication (3 cycles of 5 s each) in an ice-cold FS20D ultrasonic cleaner (Thermo Fisher Scientific, Waltham, MA, USA), and maintained on ice between cycles. Approximately 50.6 nL of the synaptosome preparation was injected per oocyte, with injections carried out under the oocyte's surface within a nylon mesh-lined glass Petri dish. Injected oocytes were incubated at 16°C in SBS within a 96-well conical-bottom plate (Thermo Fisher Scientific).

Stock solutions of GABA (Sigma-Aldrich, St Louis, MO, USA) and kainate (Tocris Bioscience, Bristol, UK) were prepared in deionized water and stored at −20°C. Prior to electrophysiological assessment, working solutions (final concentrations: 1 mM and 100 µM) were freshly diluted in a magnesium-free custom formulation of Normal Frog Ringer (NFR) solution (Ecocyte BioScience Inc.). Two-electrode voltage clamp (TEVC) recordings were performed using a Roboocyte2 recording system, consumable measuring heads, and glass electrodes pre-filled with 3 M KCl (Thermo Fisher Scientific) and embedded with Ag/AgCl-treated silver wires (Multi-Channel Systems, MCS GmbH, Reutlingen, Germany). Electrode resistances were maintained between 200 and 1000 kΩ, and a minimum resting membrane potential of −15 mV was confirmed prior to recordings. Oocytes were voltage-clamped at a holding potential of −80 mV, with a drug application rate set at 5000 µL/min controlled by a peristaltic pump. Experimental sequences for application and recording were programmed via JavaScript, with detailed scripts accessible at https://github.com/LimonLab-UTMB. Subsequent analysis, including drift correction, baseline adjustment, and peak amplitude measurements, was conducted using Roboocyte2 + software (Multi Channel Systems, MCS GmbH). Outliers from peak amplitude measurements were identified using the interquartile range (IQR) method with tail quantiles set at 0.25 and a multiplier (Q) of 1.5. Data points outside this quantile range were excluded from subsequent analyses.

Post hoc analysis of synaptic pathways within the PART BDTO interactome

Building on previously published proteomic data from individuals with Primary Age-Related Tauopathy (PART), we re-analyzed the brain-derived tau oligomer (BDTO) interactome to evaluate enrichment of synaptic components and processes.⁴⁴ This dataset was originally generated from hippocampal autopsy tissue (n = 4), in which BDTOs were isolated from PBS-soluble fractions via co-immunoprecipitation using the oligomer-selective T18 antibody, followed by liquid chromatography–tandem mass spectrometry (LC-MS/MS). Non-specific interactors were rigorously excluded using Significance Analysis of INTeractome (SAINTq), yielding a high-confidence protein network for each pathological condition. All experimental procedures, including tissue processing, BDTO isolation, LC-MS/MS sample preparation, and DIA-based proteomic acquisition and quantification, are detailed in the original publication, in which raw data were made publicly available.⁴⁴

For the present study, we performed post hoc functional annotation of the PART BDTO interactome using the Synaptic Gene Ontologies (SynGO) knowledgebase (https://www.syngoportal.org/), a curated resource for synaptic protein function. High-confidence PART BDTO interactors (as defined by SAINTq FDR < 0.05) were uploaded to the SynGO portal and analyzed under ‘maximum stringency’ settings, with the total human brain proteome used as the reference background. Enrichment analyses were conducted independently for Cellular Component (CC) and Biological Process (BP) ontologies. SynGO-generated sunburst plots were used to visualize significant enrichment across hierarchical levels, from top-level synaptic compartments and processes (inner rings) to finer subcategories (outer rings). Significantly enriched second-level terms were further annotated to highlight synaptic vesicle-related proteins and presynaptic functional pathways.

Electron microscopy (EM)

Transmission EM for morphology quality control. The structural integrity and cytoarchitecture of synaptosome preparations were assessed using transmission electron microscopy (TEM), following a previously established protocol in our laboratory.¹³ Pooled synaptosomes were pelleted by centrifugation (10,000 × g for 10 min at 4°C), yielding a visible ∼50 μL pellet. Pellets were fixed overnight at 4°C in a solution containing 2.5% formaldehyde and 0.1% glutaraldehyde in 0.05 M cacodylate buffer. Post-fixation, samples were rinsed three times with 0.1 M cacodylate buffer and further fixed in 1% osmium tetroxide (OsO₄) in 0.1 M cacodylate buffer (pH 7.4). Following en bloc staining with 2% aqueous uranyl acetate in 0.1 M maleate buffer, synaptosomes were dehydrated through a graded ethanol series and embedded in Poly/Bed 812 epoxy resin (Polysciences, Warrington, PA, USA) via progressive infiltration with propylene oxide. Polymerization was completed at 60°C overnight. Ultrathin sections were prepared using a Leica EM UC7 ultramicrotome (Leica Microsystems, Buffalo Grove, IL, USA), post-stained with 0.4% lead citrate, and imaged using a JEM-1400 transmission electron microscope (JEOL USA) operating at 80 kV. Digital micrographs were captured using a bottom-mounted Orius SC-200 camera (Gatan, Pleasanton, CA).

EM Immunogold staining of ultrathin sections for qualitative validation of synaptic rtauO engagement. Immunogold electron microscopy was employed to qualitatively confirm synaptic engagement of rtauO. Synaptosomes were isolated from frozen temporal cortex of adult C57BL/6 wild-type (WT) mice (n = 4; 2 M, 2F; 8–9 months old; Mus musculus, JAX #000664), pooled, and incubated with 2.5 μm rtauO (derived from full-length human tau, 2N4R), then washed thoroughly following the same protocol described for FC and MSM rtauO challenge experiments. WT mouse tissue was selected to minimize cross-reactivity with endogenous human tau and leverage prior findings from our group showing comparable synaptic tau engagement profiles between human and murine synaptosomes.²⁹ The presence of rtauO was detected using the oligomer-specific T22 antibody, developed in-house by Dr Rakez Kayed's laboratory. T22 has previously been validated for immunoprecipitation and ultrastructural (EM) labeling of pathogenic tau aggregates in human brain specimens.^{12,34,45–48} Animal housing conditions complied with USDA guidelines, including a 12:12 light/dark cycle and unrestricted access to food and water. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at UTMB and were designed to minimize animal distress.

Following rtauO treatment, mouse synaptosomes were pelleted (10,000 × g for 10 min at 4°C) and fixed in 2% glutaraldehyde overnight at 4°C. Samples were stained en bloc with 2% aqueous uranyl acetate, dehydrated in 50% and then 75% ethanol, and embedded in LR White resin (medium grade; Electron Microscopy Sciences, Cat#14381). Ultrathin sections (70–80 nm) were prepared on a Leica EM UC7 ultramicrotome and mounted on Formvar-carbon-coated nickel grids (EMS, Hatfield, PA). Grids were incubated in a humid chamber on drops of blocking buffer (0.1% BSA and 0.01 M glycine in 0.01 M PBS) for 15 min, followed by incubation with T22 (1:50 in 1% BSA/PBS) for 1 h at room temperature and overnight at 4°C. After washing, grids were incubated for 1 h at room temperature in the dark with 6 nm gold-conjugated anti-mouse IgG secondary antibody (Electron Microscopy Sciences, Cat#25123) diluted 1:20 in 1% BSA. Following additional washes in blocking buffer, PBS, and distilled water, grids were fixed in 2% aqueous glutaraldehyde for 5 min at room temperature. Sections were stained with 2% uranyl acetate for 5 min, washed, air dried, and counterstained with lead citrate for 30 s. Final imaging was conducted using a JEM-1400 TEM (JEOL USA) at 80 kV, with images acquired on a bottom-mounted Orius SC-200 camera (Gatan).

Systematic literature review of tauO–synapse studies in human tissue

To contextualize our findings within the broader literature on tauO and synaptic pathology in human systems, we conducted a systematic PubMed search using defined Boolean operators. Of note, our search was focused on three main criteria: tau oligomers, synapses, and human/patient specimens. The search string was:

(“tau oligomer"[Title/Abstract] OR “tau oligomers"[Title/Abstract] OR “oligomeric tau"[Title/Abstract]) AND (“synapse"[Title/Abstract] OR “synapses"[Title/Abstract] OR “synaptic"[Title/Abstract] OR “synaptic vulnerability"[Title/Abstract] OR “synaptic dysfunction"[Title/Abstract])

This search yielded 67 articles in total (see Supplemental File 2 for the full list), including 47 primary research articles and 20 reviews. After filtering for studies involving human tissue, 29 primary articles remained. Following individual review, key studies specifically addressing tauO-associated synaptic vulnerability in human specimens were curated and summarized in Table 1 of the Discussion.

Table 1.

Summary of prior studies examining tau oligomer interactions with human synapses.

Study	Tissue/Diagnosis/Region	Experimental approach	Pre- versus Postsynaptic findings	Excitatory versus Inhibitory findings
Sánchez-Aced et al., 2025⁴⁹	AD–frontal & temporal cortex	AT + dSTORM	AT: T22⁺ tauO more frequently found in presynaptic terminals. dSTORM: tauO found closer to SYPH versus PSD95.	Not addressed.
Kadamangudi et al, 2024¹³	AD & PART–hippocampus	Co-IP + LC-MS/MS of BDTO interactome	PART BDTO interactome enriched in presynaptic vesicle proteins. AD BDTOs showed balanced pre/post enrichment.	No separable excitatory or inhibitory synaptic signature found in PART or AD BDTO interactome.
Taddei et al., 2023⁵⁰	AD–visual cortex	Expansion microscopy	TOC1⁺ tauO colocalized more with Bassoon⁺ presynaptic (49%) versus PSD95⁺ postsynaptic (33%) puncta; greater loss of Synapsin-1⁺ presynaptic terminals.	Not addressed.
Colom-Cadena et al., 2023¹²	AD–temporal & occipital cortex	AT + EM immunogold	AT: T22⁺ tauO showed a presynaptic > postsynaptic gradient in excitatory pairs. EM: tauO localized to presynaptic vesicles.	AT: 1.25x loss of excitatory synaptic pairs, particularly around dystrophic neurites; inhibitory not addressed.
Singh et al., 2020⁵¹	AD & NDAN–frontal cortex & hippocampus	MSM + WB	Not addressed.	AD synapses show tauO accumulation and ↓ sE/I. NDAN synapses lack tauO and maintain normal sE/I.
Usenovic et al., 2015⁵²	hiPSC neurons	IHC + neurotransmitter release assay	Synapsin-1 and SYPH⁺ puncta were decreased in tauO treated hiPSC neurons.	hiPSC neurons seeded with tauO showed increased basal GABA release; glutamate release was unaffected.
Tai et al., 2014⁵³	AD–temporal cortex	IF	Alz-50⁺ tauO detected symmetrically across pre- and postsynaptic terminals in bipartite (15% versus16%) and hemi-synapses (23% versus 33%).	Not addressed.

Survey of PubMed-indexed studies assessing tauO localization and synaptic impact in human tissue. Entries are ranked (row-wise) by publication date and organized (column-wise) by diagnostic classification, methodological approach, and key findings relevant to selective synaptic vulnerability. AD: Alzheimer's disease; AT: array tomography; Alz50: antibody against misfolded and/or aggregated tau species; BDTO: brain-derived tau oligomers; co-IP: co-immunoprecipitation; dSTORM: direct stochastic optical reconstruction microscopy; sE/I: synaptic excitatory-to-inhibitory ratio; EM: electron microscopy; hiPSC: human induced pluripotent stem cells; IF: immunofluorescence; IHC: immunohistochemistry; LC-MS/MS: liquid chromatography–tandem mass spectrometry; MSM: microtransplantation of synaptic membranes; NDAN: non-demented with Alzheimer's neuropathology; PART: primary age-related tauopathy; PHF: paired helical filament; PSD95: postsynaptic density protein 95; SYPH: synaptophysin; TOC1 and T22: anti-tau antibody specific to oligomeric conformation; VGLUT1: vesicular glutamate transporter 1; WB : western blot.

Data analysis and visualization software

All statistical analyses and graphical representations were performed using GraphPad Prism 10.0 (GraphPad Software, La Jolla, CA, USA). Figures were prepared with BioRender (BioRender.com) and Adobe Illustrator (Adobe Inc., San Jose, CA, USA).

Results

To assess whether tauO selectively engage specific synaptic compartments and functional subtypes, SMTG synaptosomes isolated from autopsy specimens of Control individuals were treated with rtauO (25 nM) and analyzed via FC immunophenotyping. Experimental conditions were rigorously optimized for accurate detection of small particles and synaptic resolution (see Methods, Flow cytometry). The working concentration of rtauO (25 nM) was selected based on titration studies demonstrating near half-maximal engagement (∼60%) and a plateau at higher concentrations, thereby avoiding signal oversaturation (Supplemental Figure 7). Synaptic compartments and functional subtypes were delineated using carefully constructed Boolean (conditional) gating strategies based on the co-expression or exclusion of established synaptic markers. A full list of these strategies and a representative gating workflow are provided in Supplemental Figures 4 and 5.

Synapses were delineated by compartment (presynaptic versus postsynaptic) and functional subtype (excitatory versus inhibitory). rtauO engagement was significantly greater (∼1.73-fold) in the presynaptic compared to the postsynaptic compartment (Figure 2(a); t(14) = 4.09, p = 0.001, 95% CI −0.79, −0.25]). When comparing functional subtypes, inhibitory synapses exhibited significantly higher rtauO association (∼1.83-fold) than excitatory synapses (Figure 2(b); t(6) = 3.68, p = 0.01, 95% CI [0.18, 0.90]). Further stratification into four synaptic sub-categories—(1) presynaptic excitatory, (2) presynaptic inhibitory, (3) postsynaptic excitatory, and (4) postsynaptic inhibitory—revealed the highest magnitude of rtauO engagement in presynaptic inhibitory synapses. Among these, the only statistically significant difference was observed between presynaptic and postsynaptic excitatory synapses, with the former showing ∼1.72-fold greater rtauO binding (Figure 2(c); t[18] = 3.59, p_adj = 0.008). Representative FC histograms illustrate the distribution of rtauO engagement within unique synaptic populations (Figure 2(d)–(f)). Collectively, these findings reveal a compartment- and function -specific pattern of synaptic rtauO engagement, with affinity for presynaptic and inhibitory synapses.

Figure 2.

rtauO preferentially engage presynaptic compartments and inhibitory synapses. FC immunophenotyping was performed on pooled Control SMTG synaptosomes (n = 8) following treatment with 25 nM rtauO. rtauO engagement was compared a. between presynaptic and postsynaptic compartments, b. between excitatory and inhibitory synapses, and c. as a substratified analysis across excitatory and inhibitory synapses within pre- and postsynaptic compartments. Below each analysis, corresponding FC histograms illustrate fluorescence intensity distributions for rtauO across indicated gating strategies. MFI values for rtauO were extracted from each sub-gate and normalized to total rtauO⁺ signal (controlling for rtauO delivery) and corresponding rtauO⁻ signal (correcting for background fluorescence and optical noise). Data are presented as mean ± SD measured across two independent experiments. Normality was assessed via the Shapiro–Wilk test. Panels a and b were analyzed using unpaired t-tests. In panel c, a one-way ANOVA was performed to assess overall group differences, followed by post hoc pairwise t-tests, with Bonferroni correction, comparing across synaptic subtypes (excitatory versus inhibitory) and compartments (presynaptic versus postsynaptic) (α = 0.05). **p < 0.01, *p < 0.05.

To determine whether tauO modulate synaptic receptor function, we performed TEVC recordings in Xenopus laevis oocytes microtransplanted with synaptosomes isolated from the SMTG of Control individuals. Prior to MSM, synaptosomes were pre-incubated with increasing concentrations of rtauO (0, 10, 25, or 50 nM), and responses were elicited using 100 μM kainate or 1 mM GABA to selectively activate AMPA-type glutamate receptors (AMPARs) and GABA_A receptors (GABA_ARs), respectively. Kainate-evoked AMPAR currents remained largely unchanged across conditions (Figure 3(a)), with no significant differences in peak amplitude observed among treatment groups (Kruskal–Wallis p > 0.05). In contrast, GABA_AR responses displayed a marked enhancement following rtauO pre-treatment (Figure 3(b)). Peak amplitudes were significantly elevated in oocytes receiving synaptosomes exposed to 25 nM and 50 nM rtauO, relative to 0 nM controls (Dunn's post hoc test: p < 0.05 and p < 0.01, respectively), suggesting a selective potentiation of GABAergic transmission.

Figure 3.

rtauO selectively potentiates GABAergic synaptic currents in Xenopus laevis oocytes. Human synaptosomes (pooled from n = 4 donors, SMTG region, Control cases) were pre-treated with increasing concentrations of rtauO (0, 10, 25, or 50 nM) and microtransplanted into Xenopus laevis oocytes. TEVC recordings were used to measure synaptic responses mediated by a. AMPARs following application of 100 μM kainate and b. GABA_ARs following application of 1 mM GABA. Representative current traces for each condition are shown above their corresponding bar plots. Peak current amplitudes are presented as mean ± SD from technical replicates across three independent experiments. Data distribution was assessed via the Shapiro–Wilk test. Statistical differences were evaluated using the Kruskal–Wallis test with Dunn's post hoc correction. *p < 0.05, **p < 0.01 denote significant differences compared to the 0 nM rTauO condition.

To examine selective synaptic vulnerability to tauO from a molecular perspective, BDTOs and their associated proteins were isolated from PBS-soluble fractions of PART hippocampal autopsy tissue via co-immunoprecipitation using the tau oligomer-specific antibody T18. This approach enabled the identification of synaptic proteins associated with endogenous BDTOs in their native structural conformation. Associated proteins were characterized by LC-MS/MS proteomic analysis, and the resulting interactome was analyzed using SynGO to assess enrichment across structural (Cellular Component, CC) and functional (Biological Process, BP) ontologies. In addition, isolating BDTO from the PBS-soluble total homogenate, rather than a pre-enriched synaptosomal fraction, offered two methodological advantages. First, this approach selectively targets “soluble” tauO, which are widely considered to be the most neurotoxic species.^33,54–56 Second, it supports unbiased profiling of protein interactors by not limiting detection to synaptic compartments.

BDTO-associated proteins demonstrated broad enrichment across synaptic pathways in both the CC and BP ontologies (–log₁₀ Q = 5.7 and 4.2, respectively; Figure 4(a)–(d)), as indicated by the red central node in the CC map and the purple central node in the BP map. Within the CC ontology, proteins were enriched across both pre- and postsynaptic compartments, with a bias towards presynaptic components (–log₁₀ Q = 6.4; Figure 4(a)). Notably, substructural enrichment was observed for synaptic vesicles (SVs) and related machinery (Figure 4(b)). Seven proteins were identified within this vesicle-enriched category: amphiphysin (AMPH), ATPase H⁺ transporting V1 subunit A (ATP6V1A), clathrin heavy chain (CLTC), Parkinsonism-associated deglycase (PARK7), Ras-related proteins (RAB14 and RAB2A), and synapsin-1 (SYN1). Of these, six (AMPH, ATP6V1A, CLTC, RAB14, RAB2A, SYN1) localized to the synaptic vesicle membrane, and four (AMPH, ATP6V1A, CLTC, SYN1) were specifically annotated to the extrinsic component of that membrane.

Figure 4.

Proteins within the PART BDTO interactome are enriched in presynaptic components and synaptic vesicle-associated processes. BDTO were co-immunoprecipitated (antibody: T18) from PBS-soluble fractions of PART hippocampal autopsy tissue (n = 4) and subjected to LC-MS/MS proteomic analysis. Resulting datasets were interpreted in SynGO. a, b. Cellular Component and c, d. Biological Process categories. Each sunburst plot represents hierarchical enrichment, with top-level terms in the innermost ring and more specific subcategories radiating outward. Color-coded segments indicate significantly enriched synaptic compartments and pathways associated with BDTO. Individual proteins comprising respective enriched pathways are listed within dashed boxes. SV: synaptic vesicle. Stringency evidence filter was set to: ‘Remove annotations with only proteomics evidence’.

BP ontology analysis revealed significant enrichment in processes related to presynaptic vesicle cycling and synaptic terminal organization, with vesicle cycling exhibiting the strongest enrichment (Figure 4c,d). Eleven proteins were associated with this pathway, including previously noted AMPH, ATP6V1A, CLTC, PARK7, and SYN1, along with adaptor-related protein complex 2 subunit sigma 1 (AP2S1), Rho GDP dissociation inhibitor alpha (ARHGDIA), complexin-2 (CPLX2), cathepsin D (CTSD), heat shock protein 8 (HSPA8), and SH3GL interacting endocytic adaptor 1 (SGIP1). Among these, six (AMPH, AP2S1, CLTC, HSPA8, PARK7, SGIP1) were specifically associated with synaptic vesicle endocytosis. Together, these findings define the molecular landscape of the PART BDTO interactome and highlight preferential association with presynaptic compartments and synaptic vesicle-associated processes, particularly those involved in vesicle cycling and endocytosis.

To corroborate these proteomic findings at the ultrastructural level, we performed EM immunogold labeling on murine synaptosomes acutely challenged with rtauO. Synaptosomes were derived from the temporal cortex of WT mice to minimize off-target detection of endogenous human tau and to leverage prior demonstrations of conserved rtauO engagement profiles across rodent and human synapses.²⁹ Samples were labeled with the oligomer-selective T22 antibody to determine the subcellular localization of rtauO. In vehicle-treated synaptosomes, minimal immunogold labeling was detected, consistent with background signal (Supplemental Figure 8a–(d)). In contrast, synaptosomes exposed to rtauO exhibited robust labeling, with immunogold particles distributed along vesicle-rich zones and presynaptic membranes (Supplemental Figure 8e–(h)).

Discussion

Synapses are early and selectively vulnerable targets in tauopathies, prone to both tau accumulation and functional impairment; yet the molecular mechanisms underlying this vulnerability, particularly in relation to soluble tauO, remain poorly understood. Notably, whether tauO preferentially associate with specific synaptic compartments (e.g., presynaptic versus postsynaptic) or functional subtypes (e.g., excitatory versus inhibitory), and how such targeting might influence synaptic function is especially unclear. In this study, we implemented a novel flow cytometry–based immunophenotyping panel for human synaptosomes, which, in combination with electrophysiological assessment of ligand-gated synaptic responses, enabled us to delineate the vulnerability of specific synaptic compartments and functional subtypes to tauO. To anchor our findings in clinical biology, we profiled the synaptic interactome of BDTO isolated from PBS-soluble hippocampal lysates of PART autopsy specimens as a surrogate for early-stage AD, thus capturing native binding partners of soluble tauO in the absence of Aβ pathology. Our results reveal a convergent pattern in which tauO a) preferentially engage presynaptic and inhibitory synapses, b) selectively potentiate synaptic GABAergic responses, and c) interact with a molecular network enriched in presynaptic proteins involved in synaptic vesicle dynamics. Collectively, these findings provide key insights into how tauO selectively engage and modulate human synapses, advancing our understanding of synapse-specific vulnerability in early tauopathy.

The present study introduces a novel experimental strategy enabling direct, real-time assessment of tauO engagement and its functional impact at native human synapses—offering a new framework for linking these interactions with synaptic outcomes in human tissue. To accurately contextualize our findings within the existing literature, we performed a systematic PubMed search using defined Boolean operators (see Methods, Systematic literature review of tauO–synapse studies in human tissue) to identify studies examining tauO-associated synaptic vulnerability in human tissue; the curated list of studies and relevant findings are summarized in Table 1. Most studies were limited to AD—a secondary tauopathy co-occurring with Aβ pathology, liming our ability to isolate tau-specific effects—and were observational in nature, relying on postmortem tissue to infer synaptic involvement. Although mechanistic investigations of tauO have been conducted in non-human systems, such models do not fully recapitulate the human synaptic milieu and often involve overexpression of mutant tau or non-physiological tau species. Nonetheless, we interpret our findings with reference to this body of literature, giving precedence to prior studies conducted in human tissue (summarized in Table 1).

Pre- versus postsynaptic vulnerability to tauO

Using multiplexed FC immunophenotyping, we found that tauO engage both pre- and postsynaptic compartments in human synapses, yet with a notable preference for presynaptic association. This asymmetry was further reflected in our proteomic analysis of the PART BDTO interactome, which showed greater enrichment of presynaptic proteins, particularly those related to synaptic vesicle cycling (e.g., components of the vesicle docking and fusion machinery and neurotransmitter release), as compared to postsynaptic elements. EM immunogold labeling further confirmed rtauO localization to presynaptic, vesicle-containing structures.

Our findings are consistent with numerous prior studies utilizing postmortem human brain tissue, which collectively point to the preferential localization of tauO within presynaptic compartments. As summarized in Table 1, these studies span a range of methodologies—including array tomography, dSTORM, electron microscopy, immunohistochemistry, and immunofluorescence—and employ diverse tauO conformation-specific probes (e.g., T22, TOC1, Alz50, PHF1) as well as markers for both presynaptic (e.g., Bassoon, Synaptophysin, Synapsin-1) and postsynaptic (e.g., PSD95, MAP2) compartments. Importantly, we were able to achieve comparable resolution in our own study using multiplexed immunophenotyping, which enabled the differential discrimination of synaptic compartments and functional subtypes within a single human-derived sample. Notably, five of the seven studies report stronger presynaptic than postsynaptic tauO association. For example, Colom-Cadena et al. (2023) identified abundant tauO⁺ puncta within presynaptic terminals of the temporal cortex, spanning areas devoid of neurofibrillary tangles, with EM immunogold confirming tauO colocalization with synaptic vesicles¹²—an observation which mirrors our findings. Similarly, presynaptic degeneration has emerged as an early feature not only in AD but also in primary tauopathies such as frontotemporal dementia.⁵⁷ Biochemical studies reinforce these imaging observations wherein affinity-purification mass spectrometry has consistently shown that tau species isolated from human brain—namely, monomeric,⁵ oligomeric,¹³ hyperphosphorylated,^25,58 and fibrillar tau⁵⁹—are enriched for synaptic vesicle–associated proteins. Supporting this, a meta-analysis of synaptic pathology, analyzing over 400 publications performed specifically in human AD postmortem specimens, identified presynaptic vesicular machinery as disproportionately downregulated or lost.⁶⁰

Mechanistic studies in tauopathy models have illuminated how tau might perturb presynaptic function. In Drosophila and rodent models, tau accumulates in presynaptic terminals and binds synaptic vesicles via its N-terminal domain, impairing vesicle mobility and neurotransmission.⁶¹ This interaction disrupts vesicle mobility and impairs neurotransmitter release, mirroring the role of Synapsins, which also tether vesicles to the actin cytoskeleton to maintain reserve pools.⁶² Consistent with this, Synapsin-1 emerged as a prominent interactor in our PART BDTO dataset, potentially reflecting a shared mechanistic pathway. Collectively, converging evidence from human postmortem analyses and mechanistic model systems supports a consensus view that tau oligomers preferentially target presynaptic compartments, where they likely disrupt vesicle-associated processes critical for neurotransmission, though the precise mechanisms remain to be determined.

Excitatory versus inhibitory synaptic vulnerability to tauO

FC immunophenotyping further revealed that rtauO preferentially engage GABAergic more than glutamatergic synapses. Complementary electrophysiological recordings from MSM experiments showed a selective enhancement of GABA_AR currents following acute rtauO exposure, with no corresponding change in kainate-mediated AMPAR responses. Such preferential engagement of inhibitory synapses contrasts intriguingly with the predominant narrative in AD pathology, which has frequently emphasized heightened vulnerability of excitatory neurons to NFT accumulation and overt degeneration.^{20,59,63–65} However, prior investigations into selective neuronal vulnerability have predominantly focused on AD, inherently confounded by co-existing Aβ pathology (Table 1). It is well established that Aβ pathology independently promotes neuronal hyperexcitability,^66–68 complicating interpretations regarding tau-specific vulnerability.

Given conflicting evidence regarding tau's role in neural excitability,^68,69 our findings underscore the need to disentangle tau pathology from Aβ effects at human synapses. Recent evidence suggests that tau, particularly in its soluble form, contributes to neural network hypoexcitability; several studies in tau transgenic models observe reduced hippocampal excitability^70–72 as well as cortical hypoactivity.^73–75 For instance, Busche et al. (2019) illustrated neuronal hypoactivity in rTg21221 mice, which express non-fibrillar wild-type human tau at levels comparable to NFT-forming models (rTg4510), reinforcing the notion that soluble tau alone can drive significant neuronal dysfunction independent of overt NFT formation.⁷³

Focusing on the role of tauO, we recently found that aggregation-prone tau species—particularly tauO and PHF-tau—were associated with a reduced synaptic E/I ratio in PART autopsy brain specimens.²¹ These observations are further supported by studies performed in human-derived model systems, wherein iPSC-derived neuronal cultures seeded with rtauO exhibited an increase in basal GABA release, while glutamate release remained unaffected.⁵² Likewise, treatment of organotypic slice cultures with oligomeric tau generated from the pro-aggregant Tau^RDΔ construct led to increased inhibitory filipodia density, which selectively innervate interneurons.⁷⁶

Suggestive mechanistic implications of selective synaptic vulnerability to tauO

Our findings, in conjunction with prior literature, offer novel insights into potential mechanisms underlying selective vulnerability of human synapses to tauO. We speculate two non-mutually exclusive mechanistic frameworks: (1) tauO promote presynaptic GABA release, leading to increased postsynaptic GABA_AR activity, or (2) tauO engage presynaptic terminals—both GABAergic and glutamatergic—without necessarily promoting vesicle release, and instead modulate GABA_AR function, perhaps directly (e.g., conformational or allosteric mechanisms) or indirectly (e.g., receptor clustering via gephyrin stabilization). While these models remain conjectural and necessitate further empirical validation, our data favor the latter scenario.

In our MSM electrophysiology paradigm, synaptosomes were pretreated with rtauO, thoroughly washed, and microtransplanted into Xenopus oocytes. If rtauO induced transient neurotransmitter release during pretreatment, released GABA would likely be eliminated during washing and would not persist through membrane transplantation. Our flow cytometry data also favor the plausibility of postsynaptic GABA_AR modulation over presynaptic GABA release. Within excitatory synapses, rtauO exhibit a presynaptic bias; however, this appears to lack functional significance, as no modulation of kainate-mediated AMPAR responses was observed. In contrast, GABAergic synapses demonstrate more balanced engagement across both pre- and postsynaptic compartments, accompanied by increased GABA_AR receptor-mediated responses in the MSM assay. Having ruled out presynaptic GABA release as a likely cause, this effect most plausibly reflects direct postsynaptic modulation, perhaps induced by tauO binding as discussed below. Together, these observations suggest an affinity for postsynaptic GABAergic sites—where GABA_ARs are enriched.⁷⁷

Independent studies reinforce this hypothesis. Hill et al. (2019) elegantly demonstrated that rtauO, derived from the same full-length (2N4R) tau construct utilized in the current study, injected selectively into the presynaptic compartment led to a marked reduction in synaptic strength—an effect not observed with postsynaptic delivery.⁹ Unlike classical inhibition of neurotransmitter release (e.g., mediated by receptors such as adenosine A1 or GABA_B), which typically reduces release probability and thereby diminishes synaptic depression across a train of excitatory postsynaptic potentials (EPSPs), presynaptic rtauO paradoxically increased synaptic depression.⁹ One interpretation is that tau, following activity-dependent presynaptic release,^15,78,79 may bind to postsynaptic GABA_AR in a sustained agonistic or allosteric fashion, leading to prolonged receptor activation. Alternatively, given tau's well established association with microtubules,⁸⁰ tauO may modulate GABA_AR clustering or trafficking by influencing gephyrin–microtubule dynamics, thereby stabilizing receptor localization at inhibitory synapses.^81,82 Of note, these postulated frameworks remain speculative and warrant further mechanistic investigation.

Limitations and future directions

While our study provides novel evidence of the selective vulnerability of human synapses to tauO, several limitations should be noted. First, we cannot definitively determine whether the observed tauO–synapse interactions are inherently pathogenic or reflect physiological roles of tau at the synapse. However, multiple independent investigations support a synaptotoxic role for tauO,^7,8 and their absence from synapses in cognitively intact individuals with high NFT burden further reinforces a pathogenic interpretation.⁵¹ Second, although the use of human synaptosomes enhances the translational relevance of our findings, the experimental system remains reductionist. Isolated exposure of synaptosomes to rtauO does not fully recapitulate the complex environment of the human brain, where endogenous tauO may exist in diverse structural conformations and carry disease-relevant posttranslational modifications. Finally, our analyses were limited to synaptosomes from the temporal cortex of Control cases and BDTO from the hippocampus of PART cases; however, susceptibility to tauO may vary across brain regions or disease stages.^18,20

Future studies should aim to identify the specific synaptic targets of tauO, including potential interactions with GABA_AR subunits and scaffolding proteins such as gephyrin. Proteomic pull-downs and super-resolution imaging may help map these interactions and guide functional validation. It will also be important to examine regional and cell-type specificity, as tauO may display distinct patterns of synaptic engagement across brain areas and neuronal subtypes. Investigating synapses from the locus coeruleus and hippocampus—regions known to be highly vulnerable—as well as from more resilient areas such as the cerebellum, could reveal intrinsic factors underlying selective vulnerability. Additionally, the interplay between tauO and other amyloidogenic proteins—including Aβ, α-synuclein, and TDP-43—remains poorly understood. In AD, for instance, it is unclear whether AβO modulate tauO's synaptic binding profile or vice versa. Co-incubation experiments, combined with the synaptic immunophenotyping platform developed here, could help determine whether and how these oligomers interact to drive synaptic dysfunction and contribute to the transition from early Aβ-driven hyperexcitability to later tau-associated hypoactivity.

Conclusion

In summary, our study provides novel insights into the selective synaptic vulnerability to tau oligomers in human synapses, revealing a pronounced presynaptic and inhibitory bias. Through integrative structural, functional, and proteomic analyses, we delineate specific synaptic compartments and pathways preferentially targeted by oligomeric tau, with particular emphasis on presynaptic vesicle machinery and GABAergic signaling. These findings refine our understanding of tau-driven synaptic dysfunction, help reconcile prior inconsistencies in the literature, and lay the groundwork for future therapeutic strategies aimed at promoting synaptic resilience across the tauopathy spectrum.

Supplemental Material

sj-txt-1-alz-10.1177_13872877261416539 - Supplemental material for Selective vulnerability of human synapses to soluble tau oligomers

Supplemental material, sj-txt-1-alz-10.1177_13872877261416539 for Selective vulnerability of human synapses to soluble tau oligomers by Shrinath Kadamangudi, Laura Sanchez-Sanchez, Rakez Kayed, Agenor Limon and Giulio Taglialatela in Journal of Alzheimer's Disease

Supplemental Material

sj-docx-2-alz-10.1177_13872877261416539 - Supplemental material for Selective vulnerability of human synapses to soluble tau oligomers

Supplemental material, sj-docx-2-alz-10.1177_13872877261416539 for Selective vulnerability of human synapses to soluble tau oligomers by Shrinath Kadamangudi, Laura Sanchez-Sanchez, Rakez Kayed, Agenor Limon and Giulio Taglialatela in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

This work was supported by the NIH/NIA grants F30AG085974 to SK, R21AG089708 to GT, R01AG070255 to AL, R01AG054025 to RK, and R01AG073133 to GT and AL. We extend our sincere gratitude to the donors and families that make these studies possible. We thank Dr Peter Nelson from the University of Kentucky for dissection and neuropathological evaluation of the postmortem autopsy specimens utilized in this study, facilitated in part by the University of Kentucky Alzheimer's Disease Research Center supported by P30AG072946. We also thank Drs. Nemil Bhatt and Md Anzarul Haque (members of RK lab group) for providing preformed and tested recombinant tauO; Dr Vsevolod Popov and the UTMB Electron Microscopy for their training and assistance with acquisition of electron microscopy data; Dr Lee Palmer, Dr William Russell, and the UTMB Mass Spectrometry Core Facility, for their assistance with acquisition and interpretation of mass spectrometry data; Ara XIII of the Moody Brain Health Institute for assistance with manuscript revision.

ORCID iDs

Shrinath Kadamangudi

Laura Sanchez-Sanchez

Ethical considerations

Postmortem frozen human brain tissues used in this study were obtained through established material transfer agreements with the ADRC at the Sanders-Brown Center on Aging, University of Kentucky. All procedures involving human specimens were conducted in accordance with IRB guidelines at the originating institutions. Donor samples were fully de-identified prior to transfer to the University of Texas Medical Branch (UTMB); therefore, the use of these specimens did not constitute human subjects research and did not require UTMB IRB approval, in accordance with 45 CFR §46.101(a).

Consent to participate

All donors were enrolled in longitudinal brain aging studies at the contributing institutions, and written informed consent for participation and postmortem tissue donation was obtained prior to death.

Consent for publication

Not applicable

Author contribution(s)

Shrinath Kadamangudi: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Validation; Visualization; Writing – original draft; Writing – review & editing.

Laura Sanchez-Sanchez: Investigation; Methodology; Writing – review & editing.

Rakez Kayed: Resources.

Agenor Limon: Funding acquisition; Methodology; Resources; Supervision; Writing – review & editing.

Giulio Taglialatela: Conceptualization; Project administration; Resources; Supervision; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the NIH/NIA grants F30AG085974 (SK), R21AG089708 (GT), R01AG070255 (AL), R01AG054025 (RK), R01AG073133 (GT and AL), and P30AG072946 (University of Kentucky).

National Institute on Aging (grant number F30AG085974, R01AG070255, R01AG073133, R21AG089708).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

Raw proteomic data used for the post hoc PART BDTO synaptic interactome analysis was made available in Jamison et al. (2025), .⁴⁴ All other forms of raw data will be provided by the authors upon reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Guo

Noble

Hanger

. Roles of tau protein in health and disease. Acta Neuropathol 2017; 133: 665–704.

Regan

Whitcomb

Cho

. Physiological and pathophysiological implications of synaptic tau. Neuroscientist 2017; 23: 137–151.

Robbins

Clayton

Kaminski Schierle

. Synaptic tau: a pathological or physiological phenomenon? Acta Neuropathol Commun 2021; 9: 149.

Siano

Falcicchia

Origlia

, et al. Non-canonical roles of tau and their contribution to synaptic dysfunction. Int J Mol Sci 2021; 22: 10145.

Tracy

Madero-Pérez

Swaney

, et al. Tau interactome maps synaptic and mitochondrial processes associated with neurodegeneration. Cell 2022; 185: 712–728.e14.

Hernández

Ferrer

Pérez

, et al. Tau aggregation. Neuroscience 2023; 518: 64–69.

Guerrero-Muñoz

Gerson

Castillo-Carranza

. Tau oligomers: the toxic player at synapses in Alzheimer’s disease. Front Cell Neurosci 2015; 9: 464.

Gutierrez

Limon

. Synaptic disruption by soluble oligomers in patients with Alzheimer’s and Parkinson’s disease. Biomedicines 2022; 10: 1743.

Hill

Karikari

Moffat

, et al. Introduction of tau oligomers into cortical neurons alters action potential dynamics and disrupts synaptic transmission and plasticity. eNeuro 2019; 6: ENEURO.0166–19.2019.

10.

Kaniyappan

Chandupatla

Mandelkow

E-M

, et al. Extracellular low-n oligomers of tau cause selective synaptotoxicity without affecting cell viability. Alzheimers Dement 2017; 13: 1270–1291.

11.

Lasagna-Reeves

Castillo-Carranza

Sengupta

, et al. Tau oligomers impair memory and induce synaptic and mitochondrial dysfunction in wild-type mice. Mol Neurodegeneration 2011; 6: 39.

12.

Colom-Cadena

Davies

Sirisi

, et al. Synaptic oligomeric tau in Alzheimer’s disease — A potential culprit in the spread of tau pathology through the brain. Neuron 2023; 111: 2170–2183.e6.

13.

Kadamangudi

Marcatti

Zhang

W-R

, et al. Amyloid-β oligomers increase the binding and internalization of tau oligomers in human synapses. Acta Neuropathol 2024; 149: 2.

14.

Leake

. Oligomeric tau might spread trans-synaptically in Alzheimer disease. Nat Rev Neurosci 2023; 24: 393–393.

15.

Sokolow

Henkins

Bilousova

, et al. Pre-synaptic C-terminal truncated tau is released from cortical synapses in Alzheimer’s disease. J Neurochem 2015; 133: 368–379.

16.

Stancu

I-C

Vasconcelos

Ris

, et al. Templated misfolding of Tau by prion-like seeding along neuronal connections impairs neuronal network function and associated behavioral outcomes in tau transgenic mice. Acta Neuropathol 2015; 129: 875–894.

17.

Taddei

Duff

. Synapse vulnerability and resilience underlying Alzheimer’s disease. eBioMedicine 2025; 112: 105557.

18.

Hardy

Duff

. Selective vulnerability in neurodegenerative diseases. Nat Neurosci 2018; 21: 1350–1358.

19.

Martínez-Serra

Alonso-Nanclares

Cho

, et al. Emerging insights into synapse dysregulation in Alzheimer’s disease. Brain Commun 2022; 4: fcac083.

20.

Mrdjen

Fox

Bukhari

, et al. The basis of cellular and regional vulnerability in Alzheimer’s disease. Acta Neuropathol 2019; 138: 729–749.

21.

Kadamangudi

Sanchez-Sanchez

Limon

, et al. Regional correlates of tau pathology and synaptic function in primary age-related tauopathy. J Alzheimers Dis 2026; 110, 649–665.

22.

Costoya-Sánchez

Moscoso

Silva-Rodríguez

, et al. Increased medial temporal tau positron emission tomography uptake in the absence of amyloid-β positivity. JAMA Neurol 2023; 80: E1–E11.

23.

Crary

Trojanowski

Schneider

, et al. Primary age-related tauopathy (PART): a common pathology associated with human aging. Acta Neuropathol 2014; 128: 755–766.

24.

Jellinger

Alafuzoff

Attems

, et al. PART, a distinct tauopathy, different from classical sporadic Alzheimer disease. Acta Neuropathol 2015; 129: 757–762.

25.

Thierry

Ponce

Martà-Ariza

, et al. The influence of APOEε4 on the pTau interactome in sporadic Alzheimer’s disease. Acta Neuropathol 2024; 147: 91.

26.

Montal

Diez

Kim

C-M

, et al. Network tau spreading is vulnerable to the expression gradients of APOE and glutamatergic-related genes. Sci Transl Med 2022; 14: eabn7273.

27.

Ramakrishna

Jhaveri

Konings

, et al. APOE4 Affects basal and NMDAR-mediated protein synthesis in neurons by perturbing calcium homeostasis. J Neurosci 2021; 41: 8686–8709.

28.

Franklin

Taglialatela

. A method to determine insulin responsiveness in synaptosomes isolated from frozen brain tissue. J Neurosci Methods 2016; 261: 128–134.

29.

Marcatti

Fracassi

Montalbano

, et al. Aβ/tau oligomer interplay at human synapses supports shifting therapeutic targets for Alzheimer’s disease. Cell Mol Life Sci 2022; 79: 222.

30.

Marcatti

Jamison

Fracassi

, et al. A method to study human synaptic protein-protein interactions by using flow cytometry coupled to proximity ligation assay (syn-FlowPLA). J Neurosci Methods 2023; 396: 109920.

31.

Scaduto

Lauterborn

Cox

, et al. Functional excitatory to inhibitory synaptic imbalance and loss of cognitive performance in people with Alzheimer’s disease neuropathologic change. Acta Neuropathol 2023; 145: 303–324.

32.

Lasagna-Reeves

Castillo-Carranza

Guerrero-Muñoz

, et al. Preparation and characterization of neurotoxic tau oligomers. Biochemistry 2010; 49: 10039–10041.

33.

Puangmalai

Bhatt

Montalbano

, et al. Internalization mechanisms of brain-derived tau oligomers from patients with Alzheimer’s disease, progressive supranuclear palsy and dementia with Lewy bodies. Cell Death Dis 2020; 11: 1–16.

34.

Sengupta

Kayed

. Tau oligomers as pathogenic seeds: preparation, characterization, and propagation in vitro and in vivo. In: Smet-Nocca

(ed.) Tau protein: methods and protocols. New York: NY: Springer US, 2024, pp.147–183.

35.

Daniel

Malladi

Kettle

, et al. Analysis of synaptic vesicle endocytosis in synaptosomes by high-content screening. Nat Protoc 2012; 7: 1439–1455.

36.

Gleitz

Beile

Wilffert

, et al. Cryopreservation of freshly isolated synaptosomes prepared from the cerebral cortex of rats. J Neurosci Methods 1993; 47: 191–197.

37.

Hobson

Sims

. Critical analysis of particle detection artifacts in synaptosome flow cytometry. eNeuro 2019; 6: ENEURO.0009–19.

38.

Roederer

. Compensation in flow cytometry. Curr Protoc Cytom 2002; Chapter 1: Unit 1.14.

39.

Welsh

Van Der Pol

Arkesteijn

GJA

, et al. MIFlowCyt-EV: a framework for standardized reporting of extracellular vesicle flow cytometry experiments. J Extracell Vesicles 2020; 9: 1713526.

40.

Lauterborn

Scaduto

Cox

, et al. Increased excitatory to inhibitory synaptic ratio in parietal cortex samples from individuals with Alzheimer’s disease. Nat Commun 2021; 12: 2603.

41.

Limon

Reyes-Ruiz

Miledi

. Microtransplantation of neurotransmitter receptors from postmortem autistic brains to Xenopus oocytes. Proc Natl Acad Sci U S A 2008; 105: 10973–10977.

42.

Miller

Moreno

Gutierrez

, et al. Microtransplantation of postmortem native synaptic mGluRs receptors into xenopus oocytes for their functional analysis. Membranes (Basel) 2022; 12: 931.

43.

Zeppillo

Schulmann

Macciardi

, et al. Functional impairment of cortical AMPA receptors in schizophrenia. Schizophr Res 2022; 249: 25–37.

44.

Jamison

Kadamangudi

Tumurbaatar

, et al. Comparative analysis of brain-derived tau oligomer interactomes in Alzheimer’s disease, non-demented with Alzheimer’s neuropathology, and primary age-related tauopathy: implications for neurodegeneration and cognitive resilience. J Alzheimers Dis 2025; 106: 1486–1508.

45.

Al-Hilaly

Foster

Biasetti

, et al. Tau (297-391) forms filaments that structurally mimic the core of paired helical filaments in Alzheimer’s disease brain. FEBS Lett 2020; 594: 944–950.

46.

Al-Hilaly

Pollack

Vadukul

, et al. Alzheimer’s disease-like paired helical filament assembly from truncated tau protein is independent of disulfide crosslinking. J Mol Biol 2017; 429: 3650–3665.

47.

Lasagna-Reeves

Castillo-Carranza

Sengupta

, et al. Identification of oligomers at early stages of tau aggregation in Alzheimer’s disease. FASEB J 2012; 26: 1946–1959.

48.

Lo Cascio

Park

Sengupta

, et al. Brain-derived tau oligomer polymorphs: distinct aggregations, stability profiles, and biological activities. Commun Biol 2025; 8: 1–18.

49.

Sánchez-Aced

Moya-Llamas

Aumatell Escabias

, et al. Enhancing lateral resolution using two-colour direct stochastic optical reconstruction microscopy to unravel synaptic tau pathology in Alzheimer’s disease. Neuropathol Appl Neurobiol 2025; 51: e70010.

50.

Taddei

Perbet

Mate De Gerando

, et al. Tau oligomer–containing synapse elimination by microglia and astrocytes in Alzheimer disease. JAMA Neurol 2023; 80: 1209.

51.

Singh

Allen

Fracassi

, et al. Functional integrity of synapses in the central nervous system of cognitively intact individuals with high Alzheimer’s disease neuropathology is associated with absence of synaptic tau oligomers. J Alzheimers Dis 2020; 78: 1661–1678.

52.

Usenovic

Niroomand

Drolet

, et al. Internalized tau oligomers cause neurodegeneration by inducing accumulation of pathogenic tau in human neurons derived from induced pluripotent stem cells. J Neurosci 2015; 35: 14234–14250.

53.

Tai

H-C

Wang

Serrano-Pozo

, et al. Frequent and symmetric deposition of misfolded tau oligomers within presynaptic and postsynaptic terminals in Alzheimer’s disease. Acta Neuropathol Commun 2014; 2: 146.

54.

Lasagna-Reeves

Castillo-Carranza

Sengupta

, et al. Alzheimer brain-derived tau oligomers propagate pathology from endogenous tau. Sci Rep 2012; 2: 700.

55.

Gerson

Kayed

. Formation and propagation of tau oligomeric seeds. Front Neurol 2013; 4: 93.

56.

McBride

, et al. Transmission of tauopathy strains is independent of their isoform composition. Nat Commun 2020; 11: 7.

57.

Taddei

Duff

. Synapse vulnerability and resilience across the clinical spectrum of dementias. Nat Rev Neurol 2025; 21: 353–369.

58.

Drummond

Pires

MacMurray

, et al. Phosphorylated tau interactome in the human Alzheimer’s disease brain. Brain 2020; 143: 2803–2817.

59.

Otero-Garcia

Mahajani

Wakhloo

, et al. Molecular signatures underlying neurofibrillary tangle susceptibility in Alzheimer’s disease. Neuron 2022; 110: 2929–2948.e8.

60.

De Wilde

Overk

Sijben

, et al. Meta-analysis of synaptic pathology in Alzheimer’s disease reveals selective molecular vesicular machinery vulnerability. Alzheimers Dement 2016; 12: 633–644.

61.

Zhou

McInnes

Wierda

, et al. Tau association with synaptic vesicles causes presynaptic dysfunction. Nat Commun 2017; 8: 15295.

62.

Shupliakov

Haucke

Pechstein

. How synapsin I may cluster synaptic vesicles. Semin Cell Dev Biol 2011; 22: 393–399.

63.

Broekaart

DWM

Sharma

Ramakrishnan

, et al. Molecular signatures of regional vulnerability to tauopathy in excitatory cortical neurons. Acta Neuropathol 2025; 149: 60.

64.

Possenti

Freer

, et al. A tau homeostasis signature is linked with the cellular and regional vulnerability of excitatory neurons to tau pathology. Nat Neurosci 2019; 22: 47–56.

65.

Leng

Eser

, et al. Molecular characterization of selectively vulnerable neurons in Alzheimer’s disease. Nat Neurosci 2021; 24: 276–287.

66.

Busche

Hyman

. Synergy between amyloid-β and tau in Alzheimer’s disease. Nat Neurosci 2020; 23: 1183–1193.

67.

Palop

Mucke

. Synaptic depression and aberrant excitatory network activity in Alzheimer’s disease: two faces of the same coin? Neuromolecular Med 2010; 12: 48–55.

68.

Targa Dias Anastacio

Matosin

Ooi

. Neuronal hyperexcitability in Alzheimer’s disease: what are the drivers behind this aberrant phenotype? Transl Psychiatry 2022; 12: 1–14.

69.

Kazim

Seo

Bianchi

, et al. Neuronal network excitability in Alzheimer’s disease: the puzzle of similar versus divergent roles of amyloid β and tau. eNeuro 2021; 8: ENEURO.0418-20.2020.

70.

Hatch

Wei

Xia

, et al. Hyperphosphorylated tau causes reduced hippocampal CA1 excitability by relocating the axon initial segment. Acta Neuropathol 2017; 133: 717–730.

71.

Hoover

Reed

, et al. Tau mislocalization to dendritic spines mediates synaptic dysfunction independently of neurodegeneration. Neuron 2010; 68: 1067–1081.

72.

Mondragón-Rodríguez

Salas-Gallardo

González-Pereyra

, et al. Phosphorylation of Tau protein correlates with changes in hippocampal theta oscillations and reduces hippocampal excitability in Alzheimer’s model. J Biol Chem 2018; 293: 8462–8472.

73.

Busche

Wegmann

Dujardin

, et al. Tau impairs neural circuits, dominating amyloid-β effects, in Alzheimer models in vivo. Nat Neurosci 2019; 22: 57–64.

74.

Marinković

Blumenstock

Goltstein

, et al. In vivo imaging reveals reduced activity of neuronal circuits in a mouse tauopathy model. Brain 2019; 142: 1051–1062.

75.

Menkes-Caspi

Yamin

Kellner

, et al. Pathological tau disrupts ongoing network activity. Neuron 2015; 85: 959–966.

76.

Decker

Krüger

Sydow

, et al. Pro-aggregant tau impairs mossy fiber plasticity due to structural changes and Ca++ dysregulation. Acta Neuropathol Commun 2015; 3: 23.

77.

Sigel

Steinmann

. Structure, function, and modulation of GABAA receptors. J Biol Chem 2012; 287: 40224–40231.

78.

Brunello

Merezhko

Uronen

R-L

, et al. Mechanisms of secretion and spreading of pathological tau protein. Cell Mol Life Sci 2020; 77: 1721–1744.

79.

Pooler

Phillips

Lau

DHW

, et al. Physiological release of endogenous tau is stimulated by neuronal activity. EMBO Rep 2013; 14: 389–394.

80.

Avila

Lucas

Pérez

, et al. Role of tau protein in both physiological and pathological conditions. Physiol Rev 2004; 84: 361–384.

81.

Essrich

Lorez

Benson

, et al. Postsynaptic clustering of major GABAA receptor subtypes requires the γ2 subunit and gephyrin. Nat Neurosci 1998; 1: 563–571.

82.

Nykänen

N-P

Kysenius

Sakha

, et al. γ-aminobutyric acid type A (GABAA) receptor activation modulates tau phosphorylation. J Biol Chem 2012; 287: 6743–6752.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.02 MB

4.24 MB

0.00 MB