Abstract
In sporadic neurodegenerative diseases, the endogenous proteins α-synuclein in Parkinson's disease and tau in Alzheimer's disease undergo pathogenic prion-like propagation over many years, accumulating in both soluble and insoluble forms in neurons including synapses, where they impair synaptic transmission and potentially cause various neuronal symptoms. To investigate the functional outcome of such synaptic accumulation, we induced accumulation of endogenous proteins in murine and human synapses by incubating mouse (of either sex) neuronal cultures with pathogenic preformed fibrils (pffs). Two weeks after treatment with human α-synuclein or tau pff, the respective endogenous proteins accumulated in neurons including presynaptic terminals, where we also observed tubulin accumulation, suggesting microtubule over-assembly. These were not associated with mRNA upregulation and were prevented by pharmacological stimulation of autophagy. Both pffs caused accumulation of p62 in cell bodies, suggesting compromised protein degradation. pHluorin imaging in synapses indicated a marked prolongation of vesicular endocytic time, which was rescued by pharmacological depolymerization of microtubules or by the overexpression of full-length dynamin 1. Since dynamin is a high-affinity binding partner of microtubules as well as an endocytic key molecule, over-assembled microtubules can sequester dynamin, thereby inhibiting endocytosis. We conclude that pff-induced accumulation of α-synuclein or tau in presynaptic terminals can disrupt vesicle endocytosis through a common mechanism. Since endocytosis-dependent vesicle recycling is critical for maintaining neurotransmitter release, its disruption can affect the neurocircuitry operations involved, thereby causing diverse symptoms associated with neurodegenerative diseases. Thus, our data suggest a common molecular mechanism underlying synaptic dysfunctions associated with Parkinson's and Alzheimer's diseases.
- dynamin sequestering
- neurodegenerative diseases and synapse
- preformed fibrils
- synaptic dysfunction
- synuclein pff
- tau pff
Significance Statement
The accumulation of the pathogenic proteins α-synuclein and tau drives prion-like trans-neuronal propagation and underlies distinct neurodegenerative diseases, such as Parkinson's and Alzheimer's disease. Using a synaptic culture model of protein propagation, we identified a shared mechanism of synaptic dysfunction caused by these otherwise distinct proteins. In our models, propagated α-synuclein or tau disrupt protein degradation pathways, leading to their accumulation. These accumulated proteins promote excessive microtubule assembly and sequester the key endocytic protein dynamin, eventually impairing synaptic vesicle endocytosis. This cascade results in synaptic dysfunction that could compromise neurocircuit operations critical for brain functions. Our clarification of these mechanistic steps will improve our understanding of the synaptic pathophysiology underlying diverse symptoms of distinct neurodegenerative diseases.
Introduction
In Parkinson's disease (PD) and Alzheimer's disease (AD), α-synuclein and tau play pathogenic roles, with their intracellular aggregates propagating trans-neuronally in a prion-like manner into neuronal networks and eventually causing synaptic loss (Holland et al., 2023; Holmes et al., 2024) and neurodegeneration (Goedert, 2015; Jucker and Walker, 2018). In postmortem brain tissues from PD and AD patients, α-synuclein and tau are significantly accumulated in the synaptic compartment (Kramer and Schulz-Schaeffer, 2007; Tai et al., 2012; Colom-Cadena et al., 2023), but it is unclear whether such protein accumulations may affect synaptic function. Previous studies using genetic overexpression of α-synuclein (Nemani et al., 2010; Lundblad et al., 2012), or tau (Yin et al., 2016), or direct presynaptic infusion of soluble α-synuclein (Eguchi et al., 2017) or tau (Hori et al., 2022) show that elevated synaptic concentrations of tau or α-synuclein can affect synaptic vesicle turnover and neurotransmission. However, it still remains unclear whether accumulations of endogenous wild-type proteins can cause synaptic dysfunctions as in the case of sporadic AD and PD.
α-Synuclein is a small protein localized in presynaptic terminals (Nakajo et al., 1990; Jakes et al., 1994; Withers et al., 1997) in weak associations with synaptic vesicles (SVs; Fortin et al., 2005; Taoufiq et al., 2020). α-Synuclein normally exists in monomeric/tetrameric (Bartels et al., 2011; Burre et al., 2013) forms, but upon accumulation, it oligomerizes into fibrils and aggregates in neurites and cell bodies, producing the so-called Lewy pathology (Spillantini et al., 1998), which is a hallmark of synucleinopathies including PD. Like α-synuclein in PD, tau plays a central role in AD progression. Tau normally binds to axonal microtubules (MTs), but upon hyperphosphorylation detaches from MTs and elevated soluble tau aggregates into neurofibrillary tangles (NFTs), a pathological hallmark of tauopathies (Grundke-Iqbal et al., 1986; Kosik et al., 1986; Greenberg and Davies, 1990). Elevation of Lewy pathology in NFTs and soluble form of tau are closely linked to disease progression and cognitive decline (Gotz et al., 2008; Koss et al., 2016).
For both α-synuclein and tau, propagation is initiated by oligomers released via exosomal exocytosis (Emmanouilidou et al., 2010; Saman et al., 2012) followed by internalization into nearby neurons through endocytosis (Konno et al., 2012; Wu et al., 2013). The internalized oligomers are then transported via the endosomal-lysosomal pathway, where they seed endogenous proteins, thereby promoting their further accumulation and aggregation (Lee et al., 2008; Clavaguera et al., 2009; Masuda-Suzukake et al., 2013; Colom-Cadena et al., 2023). Impairments in protein degradation pathways exacerbate this process (Lee et al., 2008). In the PD and AD seeding models in culture, application of preformed fibrils (pffs) of recombinant proteins (Polinski et al., 2018) reproduce Lewy pathology (Volpicelli-Daley et al., 2011; Wu et al., 2019) or NFTs (Guo and Lee, 2011, 2013). Injections of α-synuclein pffs (Luk et al., 2012; Masuda-Suzukake et al., 2013) or tau (Iba et al., 2013; Peeraer et al., 2015) pffs in animal models also lead to PD- and AD-like pathology. These demonstrate that synaptic propagation and seeding of toxic oligomeric proteins are central and pivotal events in the early pathogenesis of both sporadic PD and AD and furthermore, can be modeled in vitro by application of pffs.
We utilized this model to address the mechanisms of synaptic dysfunction associated with sporadic PD and AD; we applied pff of human α-synuclein or tau to murine cochlear neurons forming giant synapses in culture (Dimitrov et al., 2016) or human iPSC-derived neurons in culture to investigate molecular and functional changes induced by this treatment. Our results indicate a prominent accumulation of endogenous pathogenic proteins associated with MT over-assembly in presynaptic terminals. Functionally, synaptic vesicle endocytosis measured using pHluorin assay from giant presynaptic terminals in murine neuronal culture showed a marked slowing, which could be rescued by an MT-depolymerizer agent, or overexpression of full-length dynamin 1. Furthermore, the tubulin and pathogenic protein accumulation could be prevented by an autophagy-stimulating reagent. These results suggest that a common cellular/molecular synaptic mechanism underlies synaptic dysfunctions in sporadic PD and AD.
Materials and Methods
All experiments were performed in accordance with guidelines of the Physiological Society of Japan and experiment regulations at Okinawa Institute of Science and Technology Graduate University.
Cell culture
Giant synapses were reconstructed in culture as previously described (Dimitrov et al., 2016). In short, ventral cochlear nucleus (VCN) and medial nucleus of the trapezoid body (MNTB) in the brainstem region of neonatal pups (P0, both sexes, ICR strain) were excised and enzymatically dissociated using a papain-based solution. Cells were then plated on 35 mm dishes (Ibidi) precoated overnight with 0.2 mg/ml poly-d-lysine (Millipore) in 100 mM borate buffer (Thermo Fisher Scientific), then washed three times with distilled water and dried. Cells were plated as a mixture of VCN (2 × 105 cells/dish) and MNTB-derived cells (1 × 105 cells/dish). Serum-free hippocampal astrocyte-conditioned culture medium was supplemented with the serum replacement B27 Plus (Thermo Fisher Scientific) and 20 mM KCl, NGF, BDNF, NT3 (50 ng/ml each), and FGF2 (10 ng/ml), half of which was exchanged every 4 d. VCN neurons for Western blot and RT-PCR were cultured without MNTB cells in Neurobasal Plus supplemented with 20 mM KCl, B27 plus, NGF, BDNF, NT3 (50 ng/ml), and FGF2 (10 ng/ml) and exchanged by half every 4 d.
Cultures of PSC normal cell line (HPS0331, Riken) were maintained in StemFit (Ajinomoto) medium and exchanged every day. Cells were passaged using Versene (Thermo Fisher Scientific) and plated in a 35 mm laminin-precoated dish (3.5 × 105 per dish, iMatrix; Nippi). Cells were differentiated using lentiviral transduction with doxycycline-induced expression of Ngn2 (Zhang et al., 2013) in DMEM/F12 (Thermo Fisher Scientific) supplemented with NEAA, N2 (Thermo Fisher Scientific) and GDNF, BDNF, CNTF (50 ng/ml each; PeproTech) for 1 d (DIV-3; Fig. 7a) and after 24 h (DIV-2) doxycycline was added. The next day (DIV-1), medium was changed to include additionally 1 µg/ml puromycin, and after 24 h (DIV0), the cells were dissociated with Accutase and replated on 35 mm dishes for imaging (Ibidi). After replating, cells were maintained in Neurobasal Plus medium (Thermo Fisher Scientific) supplemented with B27 Plus, GDNF, BDNF, CNTF (50 ng/ml each; PeproTech), and doxycycline 0.5 µg/ml (until DIV10) and half-changed every 4 d. During replating of differentiated neurons, primary mouse astrocytes (1 × 105 per dish) were added to each dish, and the medium was supplemented with CultureOne or AraC (Thermo Fisher Scientific) to restrict glial proliferation. Primary mouse astrocytes were cultured in DMEM (Thermo Fisher Scientific; supplemented with 5% fetal bovine serum) and passaged at least twice before usage. Human recombinant pffs of α-synuclein (αS pff) and tau were obtained from StressMarq.
DNA and transfections
For expression of pHluorin, cells were electroporated (Neon Transfection System, Thermo Fisher Scientific) with a synaptophysin fusion with two pHluorin copies (Zhu et al., 2009) under the CAG promoter (Dimitrov et al., 2016). For expression of Ngn2 in iPSC, pFUGW lentiviral plasmids (pFUGW-Ngn2 and pFUGW-tTA) carrying Ngn2 were a kind gift from Thomas Südhof laboratory. For Dynamin 1, human Dynamin 1 (BC063850) mRNA coding sequence was obtained from DNA form (Japan; Clone ID 6184100), amplified by PCR (Q5 PCR master mix, NEB) and ligated to a lentiviral vector (VectorBuilder) under the control of EF1a promoter, with a second expression of mCherry. Lentiviral particles were produced by transfecting LentiX-293 cells (Takara) using polyethylenimine (PEI) with pFUGW-Ngn2 or pFUGW-tTA and three helper plasmids (pVSVg, RRE, REV; Zhang et al., 2013). After transfection and washing, the medium was changed to harvesting media (DMEM supplemented with 5% fetal bovine serum). After 36 h, total media was collected and filtered through a 0.45 µm PES filter (Millipore), mixed with Lenti-X concentrator (3:1, Takara), left for 24 h at 4°C, and centrifuged at 1,500 × g for 1 h. After centrifugation, the pellet containing the lentiviral particles was resuspended in 1 ml PBS and stored in 0.1 ml aliquots at −80°C.
Live imaging
pHluorin imaging was performed as previously described (Dimitrov et al., 2016). Briefly, VCN neurons were transfected with pCAG-SypH2x during culture plating, before coculture with the MNTB neurons using the Neon Electroporation System (Thermo Fisher Scientific). On the day of experiments, culture dishes were set on a confocal microscope stage (LSM780, Zeiss) and perfused through an in-line heater set at 37°C with bath solution containing the following (in mM): 140 NaCl, 5 KCl, 10 HEPES, 10 glucose (Nacalai Tesque) and adjusted to pH 7.4 and 290–310 mOsm/L. To block spontaneous synaptic activity, the bath solution contained the following (in µM): 25 CNQX, 25 APV, 1 strychnine, 25 picrotoxin (Tocris). For pHluorin imaging, calyx terminals were identified from resting pHluorin fluorescence. Synaptic vesicular exo-endocytosis was triggered by stimulation (20 Hz, 1 s using a bipolar electrode placed close to the terminal), and imaging ROI was set for the time-series acquisition at 0.6–1 Hz.
Immunofluorescence
Cultured cells were fixed with 4% PFA, 4% sucrose in PBS, and permeabilized with 0.5% saponin in PBS for 10 min. Blocking was done in wash buffer (PBS, 0.05% saponin, 0.02% Tween 20) with 10% Normal Goat Serum (NGS) for 1 h at room temperature (RT). The primary antibodies were diluted in wash buffer with 5% NGS and incubated overnight (∼18 h) at 4°C and then washed three times. The secondary antibodies were diluted in wash buffer with 5% NGS and incubated for 1 h at RT and thereafter washed three times with wash buffer and twice with PBS and then imaged. Confocal imaging was performed on a confocal laser scanning LSM780 microscope equipped with Plan-Apochromat 63×, 1.4 NA oil immersion lens (Carl Zeiss). For quantifying fluorescence intensity levels, the region of interest was delimited around calyceal terminals and background fluorescence was subtracted using ImageJ software.
Western blot
Cultures were washed two times with PBS (1 mM MgCl2, 2 mM CaCl2) and total protein extracted with M-PER mammalian protein extraction reagent supplemented with Halt protease inhibitor cocktail (Thermo Fisher Scientific). Protein concentrations were measured using Bradford protein assay (Thermo Fisher Scientific). SDS-PAGE of samples (10 µg total protein per lane) were performed in the Mini Gel Tank in a Bolt 4-12% Bis-Tris Mini gel according to the manufacturer's instructions (Thermo Fisher Scientific). Proteins were transferred onto a PVDF membrane (Immunblot, Bio-Rad) using the Mini Blot Module, with Bolt Transfer Buffer according to the manufacturer's instructions (Thermo Fisher Scientific). After the transfer, the membranes were washed briefly in distilled water and incubated in 0.04% PFA in Tris-buffered saline (TBS) for 30 min then washed in TBS three times for 5 min. This PFA fixation step was taken only for α-synuclein immunoblotting, whereas other membranes were washed briefly and added directly to the blocking buffer. After washing, blocking was done overnight in TBS with 10% Normal Goat Serum (NGS) or 30 min in StartingBlock (Thermo Fisher Scientific). The primary antibodies were diluted in Signal Booster solution A (Beacle) and incubated with membranes overnight at 4°C on a shaker at 50 rpm. The secondary antibodies were diluted in Signal Booster solution B (Beacle) and incubated with membranes for 1 h at RT on a shaker at 50 rpm. After each primary and secondary antibody incubations, membranes were washed three times in TBS with 0.05% Tween 20 (TBS-T) for 20 min each. For development, chemiluminescent substrate (Clarity Western ECL, Bio-Rad) was used and imaged.
End-point RT-PCR
Total RNA was extracted using RNeasy Micro kit (Qiagen), and RNA concentration and integrity (RIN) was determined using the Bioanalyzer 2100 (Agilent). Only samples with RIN >7 were used for further analyses. Reverse transcriptase reaction was run with 10 ng of RNA with a ProtoScript II cDNA synthesis kit (NEB), and then 2.5 µl of the reaction products was run with Taq polymerase (OneTaq HS, NEB) at the annealing temperature (51°C) for 27 cycles. PCR product was run on a 2% Agarose Tris-Acetate-EDTA gel and imaged. Primer sequences were TTCATGGAGTGACAACAGTGGC (mouse SNCA forward), CTTCCTGTGGGTACCCCTCC (mouse SNCA reverse), GTGCAGATAATTAATAAGAAGCTG (mouse Tau forward), CCTGATCACAAACCCTGCTTGGCCAG (mouse Tau reverse), CCTGGATACCGCAGCTAGGA (18S forward), GCGGCGCAATACGAATGCCC (18S reverse).
Antibodies
The primary antibodies used were mouse anti-alpha-synuclein [MJFR1, 1:200 dilution for immunofluorescence IF, 1:1,000 for Western blot (WB), ab138501, Abcam], mouse anti-alpha-synuclein (Syn-1, 1:500 IF, 1:1,000 WB, BD610787, BD Transduction Laboratories), guinea pig anti-VGLUT1 (1:4,000 IF, AB5905, Merck), rabbit anti-Tau (1:1,000 IF, 1:3,000 WB, ab75714, Abcam), rabbit anti-beta3-tubulin (1:1,000 IF, #2200, Sigma), mouse anti-alpha-tubulin (1:1,000 IF, DM1A, Abcam), guinea pig anti-syp1 (1:1,000, SySy#101004, Synaptic Systems), rabbit anti-LC3 (MBL#PM036MS), guinea pig anti-p62 (MBL#PM06MS). Secondary antibodies used for IF were Goat IgG conjugated with Alexa Fluor Plus 488, 568 or 647 (1:400, Thermo Fisher Scientific). The secondary antibodies for WB were goat anti-mouse IgG, HRP-conjugate (1:10,000, ab205719, Abcam), and goat anti-rabbit IgG, HRP-conjugate (1:20,000, ab7090, Abcam).
Electron microscopy
The pffs were diluted in distilled water, stained with 2.5% uranyl acetate for 5 min, washed with water, and imaged with a transmission electron microscope (JEM-1230R, JEOL).
Analyses, graphing, and statistics
Images were analyzed and prepared with Zen (Zeiss) and ImageJ (NIH) software packages. For all images and blots, the only modification was a linear enhancement to brightness applied over the whole image. Fluorescence and chemiluminescence intensity values were extracted using ImageJ. Colocalization was evaluated using the “coloc2” plugin in ImageJ. Figures were arranged with Inkscape vector graphics editor (https://inkscape.org). Graphing was done with seaborn and matplotlib Python libraries. Statistical analyses were done with pandas, scypi, and statistics Python libraries. Statistical significance level was set at p < 0.05. Code is reposited on GitHub at https://github.com/DiDiToP/DD2023seeding-stats-and-graph.git.
Results
Human α-synuclein or tau pffs induced the accumulation of endogenous α-synuclein or tau in cultured murine neurons
We first examined whether endogenous proteins might accumulate in murine brainstem neurons in culture by applying recombinant human α-synuclein or tau protein pff (αS pff or T pff, respectively; both at 2 µg/ml) to cultures on day 5 in vitro (DIV 5) and cultured thereafter for 13–16 d (until DIV 18–21; Fig. 1a). Electron micrographs of human αS pff (Fig. 1b) and human T pff (Fig. 1c) showed rod-shaped structures of pffs with a median length of 29 nm or 42 nm, respectively, for αS and T pff. Western blots of sonicated αS and T pffs confirmed their purity (Fig. 1b,c).
Human α-synuclein or tau pffs induced accumulation of endogenous α-synuclein or tau in cultured murine neurons. a, Pff-induced seeding culture protocol. At DIV5, α-synuclein (aS) or tau (T) pffs (2 µg/ml in final concentration) were added to primary culture of murine cochlear neurons. At DIV 18–21, the cultures forming giant presynaptic terminals were used for imaging, immunochemistry, or end-point RT-PCR. b, c, Left panels, Electron micrographs of human αS pff (b) and human T pff (c). Scale bars, 50 nm. Right histograms indicate length distribution of αS pff (b, 29 nm in median) and T pff (c, 42 nm in median). Western blots in insets derived from sonicated pff (10 ng) of αS (b) or tau (c). d, Western blot for αS in whole cell lysate of cochlear neurons treated with human αS pff. Bar graph shows the results of densitometric quantification and statistics. Normalized mean value ± SEM; 1 ± 0.12 in control versus 2.54 ± 0.13 in pff-treated neurons (significant difference in t test, p = 0.016, n = 5 each). e, Western blot for tau in whole cell lysate of cochlear neurons after human tau pff treatment with densitometric quantifications and statistics. Normalized mean ± SEM; 1 ± 0.02 in control versus 1.27 ± 0.06 (significant difference in t test, p = 0.029, n = 4 cultures each). f, End-point RT-PCR quantification of SNCA cDNA as a measure of total SNCA mRNA content in cultures with (αS pff) or without (Control) αS pff treatment. Normalized mean ± SEM; 1 ± 0.22 in controls versus 0.38 ± 0.19 in αS pff-treated neurons (ns with p = 0.056 in t test, n = 5 cultures each). g, End-point RT-PCR quantification of tau cDNA as a measure of total tau mRNA content in cultures with (T pff) or without (Control) tau pff treatment. Normalized mean ± SEM: 1 ± 0.09 in control versus 0.75 ± 0.12 (ns in t test with p = 0.2, n = 4 each). Data was obtained from VCN neurons alone. Bar graph data was normalized to the mean value of Control, which was set to 1. Asterisks in bar graphs, *p < 0.05, ***p < 0.001.
In pff-treated culture, at DIV 18–21, Western blotting of total protein lysates from VCN neuron showed a 2.5-fold increase in full-length (15 kDa) α-synuclein (p = 0.02, n = 5) together with a prominent 12 kDa truncated form (Fig. 1d), the latter of which is found in brain tissues from PD patients (Campbell et al., 2001) as well as in seeding models in culture (Mahul-Mellier et al., 2020). α-Synuclein accumulation was noticeable at 22 h and full-length and truncated α-synuclein accumulated to the maximal level at 70 h (data not shown). Thus, human α-synuclein pffs trigger rapid seeding accumulation of endogenous murine α-synuclein as previously reported (Volpicelli-Daley et al., 2011). Like αS pff, incubation of cultured VCN neurons with human T pff significantly elevated total tau protein (13–16 d treatment; p = 0.03, n = 4; Fig. 1e).
In contrast to elevated α-synuclein protein in VCN neurons, expression of the corresponding mRNA (measured using SNCA cDNA RT-PCR) did not increase but underwent a slight albeit statistically insignificant decrease (p = 0.056, n = 5; Fig. 1f). This apparent decrease might presumably reflect a feedback inhibition of SNCA transcription by overexpressed α-synuclein. Like SNCA transcript, tau (MAPT) mRNA did not increase after T pff treatment (p = 0.2, n = 4; Fig. 1g).
Accumulation of endogenous α-synuclein or tau in calyceal presynaptic terminals associated with MT over-assembly in pff-treated cultures
In cocultures of neurons from the ventral cochlear nucleus (VCN) and medial nucleus of trapezoid body (MNTB), glutamatergic synapses with giant calyx-like axo-somatic presynaptic terminals are frequently formed (illustrated in Fig. 2a; Dimitrov et al., 2016). α-Synuclein and type 1 vesicular glutamate transporter (VGLUT1) were immunolabeled in calyceal terminals and their fluorescence intensity in confocal optics was quantified using densitometry. After treatment of culture with human αS pff (Fig. 1a), presynaptic α-synuclein signal increased by 100% from control (p < 0.001), whereas VGLUT1 signal was unchanged (p = 0.11; Fig. 2b). Like α-synuclein, after T pff treatment, presynaptic tau signal increased by 63% (p < 0.001) with no change in VGLUT1 level (p = 0.17; Fig. 2c). Although our available antibody recognizes both murine and human α-synuclein (Syn-1), another human α-synuclein-specific antibody (MJFR1), which stained human α-synuclein in human iPSC-derived neurons (Fig. 7), did not stain calyceal terminals of murine neurons in culture (data not shown), confirming that accumulation of endogenous murine α-synuclein was induced by human α-synuclein pff treatment.
Pff treatment-induced upregulation of endogenous α-synuclein or tau in calyceal presynaptic terminals in culture associated with enhanced MT assembly. a, Schematic illustration of an axon terminal of a VCN neuron forming a calyceal giant synapse on a soma of MNTB neuron in reconstituted culture. b, Immunofluorescence images of VGLUT1 (green) and α-synuclein (αS; red) in a calyceal presynaptic terminal with (right panels) or without (left, control) αS pff treatment. Scale bar, 10 µm. Box plots indicate fluorescent signal intensities of VGLUT1 (upper panel) and αS (lower panel) with or without αS pff treatment. Normalized median (IQR against control) for VGLUT1; 1 (0.50) in control versus 1.16 (0.74) in αS pff-treated terminals [no significant difference (ns) in Mann–Whitney test, p = 0.11, n = 73 controls and 75 pff-treated terminals]. Normalized median (IQR) for αS; 1 (1.11) in control versus 2.02 (1.78) in αS pff-treated terminals (significant difference in Mann–Whitney test, p = 1.1 × 10−13, n = 186 controls and 149 pff-treated terminals). c, Immunofluorescence images for VGLUT1 (green) and tau (red) in a calyceal presynaptic terminal with (right panel) or without (left) tau pff treatment. Scale bar, 10 µm. Quantification of normalized VGLUT1 and tau content in calyceal terminals with (T pff) and without (Control) tau pff treatment. Normalized median (IQR) for VGLUT1; 1 (0.57) in control versus 0.87 (0.29) in T pff-treated (ns with p = 0.17 in Mann–Whitney test, n = 28 controls and 40 treated terminals). Normalized median (IQR) for tau; 1 (0.78) versus 1.63 (0.40) in T pff-treated (significant difference, p = 0.0005, n = 28 controls and 40 treated terminals). d, e, VGLUT1 immunofluorescence (green, top rows) and βIII-tubulin (white, bottom row in d) or α-tubulin (white, bottom row in e) in calyceal terminals with (right panel) or without (left panel) treatment with αS pff (d) or tau pff (e). Box plots show quantification of tubulin immunofluorescence intensity in the VGLUT1-positive regions. Normalized median (IQR); 1 (1.03) in control versus 2.25 (1.48) in αS pff-treated terminals (significant difference in Mann–Whitney test, p = 0.0008, n = 24 controls and 21 treated terminals) and 1.29 (0.44) in T pff-treated terminals (significant difference in Mann–Whitney test, p = 0.02, n = 28 control and 41 treated terminals). Box plot data was normalized to the median value of Control, which was set to 1. Asterisks in box plots, *p < 0.05, ***p < 0.001.
α-Synuclein can bind to tubulins and assemble them into microtubules (MTs) in cell-free assays and in cultured cells (Alim et al., 2004; Cartelli et al., 2016; Toba et al., 2017). Tau normally stabilizes MTs in axonal compartments, and both phosphorylated and nonphosphorylated soluble tau can assemble tubulins (Shahpasand et al., 2012). In slice experiments, loading of presynaptic terminals with human recombinant α-synuclein (Eguchi et al., 2017) or tau (Hori et al., 2022) induces MT over-assembly. We examined whether endogenous α-synuclein or tau accumulated in culture by pff treatment might likewise assemble MTs. Immunofluorescence staining of tubulin in giant presynaptic terminals primarily reveals assembled MTs, as free tubulin is washed out during permeabilization (Piriya Ananda Babu et al., 2020). This method thus enables quantification of MT content in presynaptic terminals. In cultured calyceal presynaptic terminals, pff-triggered accumulation of endogenous murine α-synuclein or tau was associated with a significant increase in MT assembly, a 125% (p = 0.0008; Fig. 2d) increase in αS pff-treated cultures, and a 29% (p = 0.02; Fig. 2e) increase in T pff-treated cultures.
Accumulation of α-synuclein or tau in calyceal presynaptic terminals inhibits vesicle endocytosis
Using the pH-sensitive fluorescent protein pHluorin (Miesenbock et al., 1998) fused to synaptophysin (Zhu et al., 2009), we next investigated whether synaptic vesicle (SV) endocytosis might be affected by seeded endogenous proteins. In calyceal terminals overexpressing a synaptophysin-pHluorin fusion protein, repetitive stimulations (20 Hz for 1 s) of presynaptic axons induced exocytosis (vesicular pH neutralization) followed by endocytosis (vesicular acidification; Fig. 3a,b) with a median half-decay time of 9 s in controls without pff treatment (at 37°C; Fig. 3c,d). In α-synuclein pff-treated calyceal terminals, the endocytic median half-decay time was prolonged (Fig. 3c) to 30 s (p < 0.001; Fig. 3d). Likewise, in tau pff-treated terminals SV endocytosis was significantly prolonged (p = 0.008; Fig. 4d,e,f).
αS pff treatment impairs vesicular endocytosis. a, Live imaging of pHluorin (SypH2x) signals at calyceal presynaptic terminals (shown in 8 bit pseudocolor, 0–255 shown as blue to red) in cultures with (bottom row) or without (top row) αS pff treatment. Images were taken immediately before stimulation (left column), during stimulation (20 Hz for 1 s; middle column), and 30 s after stimulation (right column). Insets (top right corners) show higher magnification images of a single presynaptic swelling as outlined in the left row. Scale bar is 10 µm and 1 µm, respectively, for the main and inset images. b, PHluorin signal recorded from a single calyceal synapse in cultures with (bottom trace; αS pff) or without (top trace; control) αS pff treatment. c, pHluorin signal decay with (red; αS pff) or without (black; control) αS pff treatment superimposed after normalization at their peaks. Traces indicate mean and SEM from 10 terminals each. d, Quantification of the half-decay time with (right; αS pff) or without (left; control) αS pff treatment. Median (IQR) half-decay times were 9.1 (3.9) s in control versus 31 (28) s (significant difference in Mann–Whitney test, p = 4 × 10−6, n = 20 and 19 terminals, respectively). Asterisks in box plot, ***p < 0.001.
Nocodazole rescued endocytosis impaired by the pff treatment. a, Live pHluorin imaging (SypH2x) of calyceal terminals treated with αS pff. Pseudocolor (8 bit, 0–255 as blue to red) fluorescent images of the same giant synaptic terminal before (top row) and after (bottom row) application of nocodazole (20 µM, 30 min). Left panels show images immediately before stimulations, middle panels show images during stimulation for 1 s at 20 Hz, and right panels show images 30 s after stimulation. b, Single pHluorin signal traces in αS pff-treated (left traces) or untreated (right traces, control) calyceal terminals, before (top traces) and after (bottom traces) application of nocodazole (blue arrow, 20 µM, 30 min). A train of stimulations (20 Hz, 1 s) is given at each arrowhead. c, Mean trace of pHluorin signal decay (top) in control terminals without αS pff treatment (left traces) and αS pff-treated terminals (right) before (black traces) and after nocodazole application (superimposed blue traces). Bottom box plots indicate half-decay time (from left to right) in control terminals without (C) or with nocodazole application (CN), in comparison with αS pff-treated terminals without (Sp) or with nocodazole application (SpN). Median (IQR); 10 (5.5) s in control versus 11 (6.4) s (CN), 29 (8.0) s (Sp) and 14 (6.1) s (SpN), respectively. Statistically significant difference between C and Sp (p = 0.0002, n = 11 terminals each) and between Sp and SpN (p = 0.02, n = 13 terminals each). No significant difference between C and CN (p = 0.79, n = 11) and between C and SpN (p = 0.3, n = 11 and 13 terminals, respectively). Statistical test was done by Kruskal–Wallis test with Dunn post hoc after Holm correction. d, Pseudocolor presentation of pHluorin live imaging of tau pff-treated calyceal terminal treated with T pff before (top panels) and after (bottom panels) application of nocodazole (20 µM, 30 min). e, Single pHluorin signal traces in tau pffs-treated (T pff; left traces) or untreated (right traces, control) calyceal terminals, before (top traces) and after (bottom traces) application of nocodazole (blue arrow, 20 µM, 30 min). A train of stimulations (20 Hz, 1 s) is given at each arrowhead. f, pHluorin signal decay trace (top) in control terminals without (left traces) and with tau pff treatment (right) before (black traces) and after nocodazole application (superimposed blue traces). Bottom box plots indicate half-decay time (from left to right) in control terminals without (C) or after nocodazole (CN), in tau pff-treated terminals without (Tp) or with nocodazole application (TpN). Median (IQR); 9.0 (4.5) s in control (C) versus 8.5 (3.5) s (CN), 24 (27) s (Tp) and 13 (6.5) s (TpN), respectively. Statistically significant difference between C and Tp (p = 0.008, n = 13 and 20, respectively) and between Tp and TpN (p = 0.04, n = 20 and 11, respectively). No significant difference between: C and CN (p = 0.77, n = 13 and 12, respectively) and between C and TpN (p = 0.3, n = 13 and 11, respectively). Statistical test was done by Kruskal–Wallis test with Dunn post hoc after Holm correction. Asterisks in bar graphs, *p < 0.05, **p < 0.01, ***p < 0.001.
Microtubule depolymerization rescues impairment of SV endocytosis by pff-induced accumulation of α-synuclein or tau
The increase in intrasynaptic MTs induced by pff seeding suggests that MT over-assembly might underlie the prolonged SV endocytosis. To test this, we bath-applied nocodazole, a potent MT depolymerizer, to pff-treated cultures. Mild nocodazole treatment (30 µM, 30 min at 37°C), which depolymerizes endogenous MTs by ∼20% in calyceal terminals of cultured neurons (Guillaud et al., 2017), by itself had no effect on endocytosis in control presynaptic terminals (Fig. 4c,f, CN) but reversed the slowed endocytosis by αS pff (29 s in half-decay time; Fig. 4a–c, αSp) to control level (14 s, αSpN; p = 0.02; Fig. 4c). Nocodazole likewise reversed the slowed endocytosis by T pff (from 25 to 13 s in half-time; p = 0.04; Tp vs TpN; Fig. 4d–f). Our above results of pff-induced increase in synaptic tubulin immunoreactivity (Fig. 2), which together with the sensitivity to nocodazole (Fig. 4), indicates enhanced microtubule over-assembly in the synaptic terminal. Thus, endogenous α-synuclein or tau accumulated in presynaptic terminals of cultured neurons over-assembled MTs, thereby causing endocytic slowing, like recombinant human proteins infused into presynaptic calyceal terminals in slice (Eguchi et al., 2017; Hori et al., 2022).
Overexpression of full-length dynamin 1 rescues tau pff-induced impairment of SV endocytosis
The monomeric GTPase dynamin, which is originally discovered as an MT-binding protein (Shpetner and Vallee, 1989), plays a critical role in endocytic SV fission (Hinshaw and Schmid, 1995; Takei et al., 1995). Presynaptic MTs over-assembled by α-synuclein or tau (Fig. 2) can sequester soluble dynamin, thereby preventing SV endocytosis (Hori et al., 2022). To test this hypothesis, we overexpressed full-length dynamin 1 (DNM1) in tau pff-treated cells using a lentiviral dynamin 1/mCherry bicistronic expression system at DIV15 (Fig. 5a). In this method, mCherry signal is expected to be positive where dynamin 1 is overexpressed. Since endogenous dynamin 1 is a synaptosomal protein highly expressed in presynaptic terminals (Taoufiq et al., 2020), we examined whether T pff-induced endocytic slowing is rescued in presynaptic terminals where dynamin 1 is overexpressed. We detected positive mCherry signal in some presynaptic terminals, but not in others or postsynaptic cells (Fig. 5b). The pHluorin endocytosis imaging assay was subsequently performed on both mCherry-positive (Dynamin 1-overexpressing) and mCherry-negative calyceal terminals (Fig. 5b,c). At mCherry-negative terminals (Tp), the mean endocytic half-decay time was 26 s like those in culture without DNM1 overexpression (Fig. 4d–f). In contrast, at mCherry-positive calyceal terminals (TpD), the mean endocytic half-decay time was shortened to 15 s (p = 0.03) toward the control without tau pff treatment (C). mCherry-positive terminals without tau pff treatment showed a similar endocytic time course (C and CD; Fig. 5c,d), indicating that DNM1overexpression by itself had no effect on SV endocytosis. Thus, dynamin 1 overexpression rescued SV endocytosis impaired by endogenous tau accumulation in pff-treated calyceal terminals in culture.
Dynamin 1 overexpression prevents endocytic impairment caused by tau pff treatment. a, Schematic illustration for the protocol of T pff application (at DIV5) and lentiviral infection of dynamin 1-mCherry construct (DIV15). b, Live confocal images of pHluorin (green) and dynamin (DNM)1-mCherry (red) at calyceal terminals in a culture dish. c, Half-decay time of pHluorin signal after stimulation in control terminals (C, black) or in terminals overexpressing DNM1 (CD, gray), and in T pff-treated terminals without (Tp, blue) or tau pff-treated terminals overexpressing DNM1 (TpD, magenta). d, Quantifications of half-decay times. Median (IQR); 9.0 (4.3) s in control (C), 10 (2) s (Cp), 26 (18) s (Tp), 4.5 ± (6.3) s (TpD). Statistically significant difference between C and Tp (p = 0.000003, n = 12 and 20, respectively), between Tp and TpD (p = 0.03, n = 20 each) and between C and TpD (p = 0.01, n = 12 and 20, respectively). No significant difference between C and CD (p = 0.76, n = 12 and 13, respectively). Statistical test was done by Kruskal–Wallis test with Dunn post hoc after Holm correction. Asterisks in bar graphs, *p < 0.05, ***p < 0.001.
Autophagy-stimulating reagent relieves the pff-induced accumulation of α-synuclein and tau in calyceal terminals
In our culture model, pff treatment induced the accumulation of α-synuclein and tau without increasing their mRNAs (Fig. 1f,g), suggesting that their protein degradation system might be compromised. Hence, we tested whether the autophagy-stimulating drug rapamycin (Noda and Ohsumi, 1998) affects the protein accumulation by pff. Incubation of α-synuclein pff-treated culture with 500 nM rapamycin for 5 d (from DIV 14 to DIV19) prevented the presynaptic accumulation of α-synuclein (Fig. 6a,b) and tubulin (Fig. 6a,c). Likewise, in tau pff-treated culture, rapamycin prevented tau (Fig. 6d,e) and tubulin (Fig. 6d,f) accumulation in calyceal terminals. Thus, in both α-synuclein and tau pff-treated cultures, rapamycin prevented the seeded accumulation of α-synuclein and tau and the associated MT over-assemblies. These results suggest that compromised autophagy might underlie the seeded accumulation of α-synuclein and tau after pff treatment in culture.
Pff-seeded synaptic accumulation of αS or tau and associated MT assemblies is rescued by rapamycin and causes consistent increase in cell body p62 immunofluorescence. a, Immunofluorescence for αS (top row; red) or tubulin (bottom row; white) in calyceal terminals from αS pff-treated cultures without (left panel) or with (right panel) rapamycin treatment. b, Statistical analyses of αS content in the calyceal terminals in control cultures (C), αS pff-treated cultures (Sp), and αS pff-treated cultures with rapamycin treatment (SpR). Normalized median (IQR): 1 (1.9; C), 4.56 (3.3; Sp), and 1.9 (2.3; SpR), respectively. Statistically significant difference between C and Sp (p = 0.00009, n = 21 and 20, respectively) and between Sp and SpR (p = 0.032, n = 20 and 18, respectively). No significant difference between C and SpR (p = 0.1, n = 21 and 18, respectively). c, Statistical analyses of tubulin content in the giant synapses in control cultures (C), αS pff-treated cultures (Sp), and αS pff-treated cultures with rapamycin treatment (500 nM) for 5–7 d before imaging. Normalized median (IQR); 1 (0.88; C), 1.71 (0.39; Sp), and 0.94 (1.03; SpR), respectively. Statistically significant difference between C and Sp (p = 0.003, n = 21 and 20, respectively) and between Sp and SpR (p = 0.0002, n = 20 and 18, respectively). No significant difference between C and SpR (p = 0.36, n = 21 and 18, respectively). d, Immunofluorescence images of a calyceal terminals stained for tau (top row, red) or tubulin (bottom row, white), after treatment with T pff, before rapamycin treatment (left panels), or after application of 500 nM rapamycin for 5 d (right panels). e, Statistical analyses of tau content in the giant synapses in control cultures (C), tau pff-treated cultures (Tp), and tau pff-treated cultures with rapamycin treatment (TpR). Normalized median (IQR): 1 (0.58; C), 1.95 (1.00; Tp), and 1.55 (0.70; TpR). Statistically significant difference between C and Tp (p = 8 × 10−8, n = 27 and 25 terminals, respectively), between Tp and TpR (p = 0.045, n = 25 and 24, respectively), and between C and TpR (p = 0.001, n = 27 and 24, respectively). f, Statistical analyses of tubulin content in the giant synapses in control cultures (C), tau pff-treated cultures (Tp), and tau pff-treated cultures with added 500 nM rapamycin for 5–7 d before imaging (TpR). Normalized median (IQR): 1 (0.71; C), 1.57 (0.58; Tp), and 0.95 (0.58; TpR). Statistically significant difference between C and Tp (p = 0.0001, n = 27 and 25, respectively) and between Tp and TpR (p = 0.0004, n = 25 and 24, respectively). No significant difference between C and TpR (p = 0.81, n = 27 and 24, respectively). g, Immunofluorescence images of CN neurons treated with 2 µg/ml αS pff (right column) for 2 weeks showing immunostaining of α-synuclein (top), p62 (middle-top), LC3 (middle-bottom), and DAPI (bottom). Scale bar, 10 µm. h, Quantification of fluorescence intensity in cultures treated with αS pff from cell soma ROIs normalized to median values of control for p62 (top) and LC3 (bottom). Normalized median (IQR) for p62; 1 (0.446) in control versus 1.616 (0.765) in αS pff-treated cultures (significant difference, p = 0.00069, n = 20 control and 24 pff-treated cells). Normalized median (IQR) for LC3; 1 (0.188) in control versus 1.685 (0.964) in αS pff-treated cultures (significant difference, p = 0.00002, n = 20 control and 24 pff-treated cells). i, Immunofluorescence images of CN neurons treated with 2 µg/ml T pff (right column) for two weeks showing immunostaining of tau (top), p62 (middle), and DAPI (bottom). Scale bar, 10 µm. j, Quantification of fluorescence intensity in cultures treated with T pff from cell soma ROIs normalized to median values of control for p62 (left graph). Normalized median (IQR) for LC3; 1 (0.508) in control versus 1.381 (0.604) in tau pff-treated cultures (significant difference, p = 0.004, n = 18 control and 25 pff-treated cells). k, Merged fluorescence images from T pff-treated cells showing immunostaining of tau (red) and p62 (green). Scale bar, 10 µm. l, Merged fluorescence images from αS pff-treated cells (right column; left column is control) showing immunostaining of α-synuclein (red) and p62 (green) and DAPI (blue). In middle panels, α-synuclein (red), LC3(green), and DAPI (blue). In bottom panels, LC3 (red), p62 (green), and DAPI (blue). Scale bar, 10 µm. m, Quantification of fluorescence intensity overlap coefficient (Manders overlap coefficient). Overlap coefficient was calculated for pff-treated cultures overlap between p62 and αS (in αS pff-treated cultures), tau (in T pff-treated cultures), LC3 (as positive control; in αS pff-treated cultures), and DAPI (as negative control; in αS pff-treated cultures). P62 Manders overlap coefficient median (IQR): 0.02 (0.55; αS), 0.55 (0.35; tau), 0.54 (0.37; LC3), and 0.04 (0.11; DAPI); n = 43, 30 cells from αS and tau pff-treated cultures, respectively. Statistically significant difference between αS and tau (p = 4 × 10−7), αS and LC3 (p = 3 × 10−8), tau and DAPI (p = 2 × 10−7), and LC3 and DAPI (p = 1 × 10−8). No statistical difference between αS and DAPI and tau and LC3. Data for rapamycin treatment was obtained from VCN/MNTB neuron cocultures; data for p62 quantifications was obtained from VCN neurons alone. Box plot data was normalized to the median value of Control, which was set to 1. Statistical test was done by Kruskal–Wallis test with Dunn post hoc after Holm correction, or with Mann–Whitney test for two-sample comparisons. Asterisks in bar graphs, *p < 0.05, **p < 0.01, ***p < 0.001.
Pff treatment causes accumulation of autophagic machinery components in cochlear neuronal cell bodies
Given that rapamycin rescued pff-induced protein accumulation and MT over-assembly, using immunostaining of the autophagy adaptor protein p62(SQSTM1), we examined whether the autophagic machinery function might be affected by pff treatment. Since p62 binds and coaggregates with ubiquitinated target proteins for subsequent degradation, autophagy inhibition leads to an increase in p62 protein levels (Bjørkøy et al., 2005). In cochlear neurons in culture treated with αS pff for 2 weeks, p62 showed clear accumulation colocalized with another autophagosomal marker protein LC3 (Fig. 6g,h). Likewise, tau pff-treated neurons showed p62 accumulation (Fig. 6i,j).
Furthermore, we examined whether tau or α-synuclein or tau co-localize with p62 in pff-treated cultures (Fig. 6k,l). In these analyses, we utilized the p62-LC3 colocalization as a positive control and the p62-DAPI (cell nucleus marker) colocalization as a negative control. Clear overlaps were observed between tau and p62 immunofluorescence (Fig. 6k,m), whereas αS and p62 signals showed no overlap (Fig. 6l,m). Thus, tau, but not αS, colocalized with p62. The reason for the lack of αS in the p62-autophagosome is unclear. One possibility would be that their colocalization might be transient within a distinct time window in culture, which we missed for αS pff in our present protocol. More systematic examination would be necessary to identify the timing of colocalization of p62 with pathogenic protein oligomers.
Pff treatment of human iPSC-derived neurons induces the protein accumulation in small presynaptic puncta
To examine whether the effects of pff treatment at murine calyceal synapses can be reproduced in bouton-type human central synapses, we differentiated human-induced pluripotent stem cells (iPSCs) into neurons (Materials and Methods) that formed functional synapses (data not shown). We then treated them with human αS pff or tau pff (at 2 µg/ml) for 14 d (Fig. 7a). After treatment of human iPSC-derived neurons with αS or tau pff, synaptophysin-positive glutamatergic terminals (puncta) showed a significant increase in αS (Fig. 7b–d) or tau (Fig. 7e–g) immunofluorescence both accompanied by tubulin accumulation (Fig. 7d,g). These results predict that trans-neuronal seeding propagation of the pathogenic proteins in humans can slow down synaptic vesicle endocytosis.
Pff treatment in human iPSC-derived neurons causes accumulation of the αS or tau and tubulin in synaptophysin-labeled presynaptic terminals. a, Schematic illustration for the protocol for differentiation of human iPSC cultures into neurons by Neurogenin 2 transcription factor and addition of neurotrophic factors and then replated in coculture with mouse glial cells (DIV 0). b, Immunofluorescence images of iPSC-derived human neurons without treatment (top row, Control) or treated with 2 µg/ml of αS pff for 2 weeks (bottom row, αS pff) showing immunostaining of synaptophysin 1 (syp1, left panels, green), α-synuclein (middle panels, red), and tubulin (right panels, white). Scale bar, 50 µm. c, High magnification immunofluorescence images from control (left column) and αS pff-treated (right panels) cultures immunostained as in b. Scale bar, 1 µm. d, Quantification from synaptophysin-positive puncta in regions of interest (ROIs, 1 µm2). Left panel box plots indicate αS contents in synapses with normalized median (IQR): 1 (1.16; Control) versus 1.59 (1.29; αS pff) with significant difference in Mann–Whitney test (p = 2.9 × 10−22, n = 282 and 310, respectively). Right panel box plots indicate βIII-tubulin content in synapses with normalized median (IQR); 1 (0.91; Control) versus 1.38 (1.11; αS pff) with significant difference in Mann–Whitney test (p = 2.32 × 10−10, n = 282 and 310, respectively). e, Immunofluorescence images of iPSC-derived human neurons without treatment (top row, Control) or treated with tau pff (2 µg/ml) for 2 weeks [bottom row, tau (T) pff] showing immunostaining of synaptophysin 1 (syp1, left panels, green), tau (middle column, red), and tubulin (right column, white). Scale bar, 50 µm. f, High magnification immunofluorescence images from control (left panels) and tau pff-treated (right panels) cultures immunostained as in e. Scale bar, 1 µm. g, Quantification from synaptophysin-immunolabeled synaptic puncta ROIs in tau pff-treated cultures. Left panel box plots indicate tau content in synapses with normalized median (IQR); 1 (0.67; Control) versus 1.50 (1.08; T pff) with significant difference in Mann–Whitney test (p = 0.0001, n = 109 for both). Right panel box plots indicate α-tubulin content in synapses with normalized median (IQR); 1 (0.46; Control) versus 1.50 (0.96; T pff) with significant difference in Mann–Whitney test (p = 6 × 10−9, n = 109 for both). Asterisks in bar graphs, ***p < 0.001.
Discussion
Toxic effects of α-synuclein or tau on synaptic transmission have been well documented in studies using overexpression or overloading in cell culture or acute brain slice (Nemani et al., 2010; Scott and Roy, 2012; Xu et al., 2016; Eguchi et al., 2017; Hori et al., 2022). If the accumulation of α-synuclein or tau can impair synaptic functions prior to neurodegeneration, early behavioral symptoms in PD or AD could result from defective neurocircuitry operations. However, it remains open whether endogenous proteins can cause such toxicity (Sulzer and Edwards, 2019). Hence, to investigate the synaptic toxicity of the endogenous wild-type (WT) proteins, we applied the protein pff-induced propagation model (Volpicelli-Daley et al., 2011; Luk et al., 2012; Masuda-Suzukake et al., 2013; Wu et al., 2019) to the giant synapse reconstituted in culture (Dimitrov et al., 2016). After 2 weeks treatment of cochlear neurons in culture with the human recombinant protein pff, murine endogenous α-synuclein or tau accumulated in neurons, including their giant presynaptic terminals. This protein upregulation was unassociated with an increase in mRNAs and prevented by stimulating autophagy with rapamycin, suggesting that the protein accumulation resulted from compromised degradation rather than de novo synthesis. In the pff-induced protein seeding hypothesis, extracellular protein pffs are retrieved into cells via endosomal-lysosomal pathway (Lee et al., 2008), and upon lysosomal rupture, endogenous proteins are seeded in abundance unless damaged lysosomes are autophagocytosed (Kakuda et al., 2023). Our results of the autophagy stimulant rapamycin, rescuing the pff-induced increase in endogenous α-synuclein or tau, are consistent with this hypothesis. These results are also compatible with the age-dependent accumulation of WT α-synuclein or tau protein in sporadic PD and AD.
In the rodent brainstem slice, recombinant monomeric α-synuclein or tau directly loaded in giant presynaptic terminals induces MT over-assembly (Eguchi et al., 2017; Hori et al., 2022). In the present study, we confirmed that the accumulation of endogenous α-synuclein or tau, seeded by pff treatment, accompanied MT over-assembly in the murine giant synapse as well as in human iPSC-derived neurons in culture. Thus, MT over-assembly in presynaptic terminals can be a common feature in the early stage of sporadic PD and AD. The functional outcome of the protein accumulation and associated MT over-assembly was a slowdown of vesicle endocytosis as shown by the pHluorin assay in culture or by presynaptic membrane capacitance measurements in slice (Eguchi et al., 2017; Hori et al., 2022).
MT over-assembly and endocytic impairments can be mechanistically linked by the monomeric GTP-binding protein dynamin, which is a high-affinity binding partner of MTs (Shpetner and Vallee, 1989) and plays a fission role in vesicle endocytosis (Hinshaw and Schmid, 1995; Takei et al., 1995). Therefore, it is suggested that over-assembled MTs in presynaptic terminals sequestrate cytosolic dynamin, thereby inhibiting vesicle endocytosis (Hori et al., 2022). Consistently, in the present culture model, overexpression of the full-length dynamin 1 rescued the endocytic defect caused by endogenous tau accumulation after pff treatment. Likewise, in slice, the synthetic dodeca peptide PHDP5, which inhibits MT-dynamin binding in vitro, rescues endocytic defect when coloaded with tau in giant presynaptic terminals (Hori et al., 2022). Furthermore, in vivo, intranasal applications of PHDP5 to two types of AD model mice restored their defective learning and memory task performances in Morris's water maze test back to normal levels (Chang et al., 2024).
When vesicle endocytosis is impaired by tau loading, exocytic release of transmitter is consequently impaired after various delay depending upon the frequency of synaptic transmission (Hori et al., 2022). Namely, during high-frequency transmission, releasable vesicles at release sites are quickly depleted unless the sites are refilled in time with new vesicles recycled via endocytosis. Physiologically, high-frequency synaptic transmission plays essential roles in cognition (Sabatini and Regehr, 1999; Buzsaki and Draguhn, 2004) and in motor control (Sugihara et al., 1993). In the electroencephalography, gamma oscillation at 30–100 Hz accompanies various cognitive processes (Herrmann et al., 2004; Lachaux et al., 2012), and it is impaired in AD mouse model (Nakazono et al., 2017). Therefore, preferential impairments of high frequency neurotransmission by accumulated α-synuclein or tau (Eguchi et al., 2017; Hori et al., 2022) can explain symptoms associated with AD or PD.
Our results in culture as well as in slice models (Eguchi et al., 2017; Hori et al., 2022) indicate that presynaptic accumulation of α-synuclein or tau induces common mechanisms leading to synaptic dysfunction. Since α-synuclein and tau propagate to different brain regions during the progression of sporadic PD and AD (Goedert, 2015), synaptic impairments by a common mechanism can result in region-specific distinct symptoms, depending upon the physiological roles of neurocircuitry in the protein-propagated regions. Namely, synaptic dysfunction in the basal ganglia caused by accumulated α-synuclein can impair the high-frequency input-dependent dopamine release (Lundblad et al., 2012; Janezic et al., 2013) and motor controls, whereas weakened synaptic efficacy by accumulated tau in the hippocampal and cerebral cortical regions will impair memory induction and retention.
Our protein seeding culture model revealed molecular and functional mechanisms underlying synaptic dysfunctions associated with neurodegenerative diseases in mouse primary giant synapse cultures culture, as well as in human synapses formed by neurons differentiated from human iPSCs. Thus, the inhibition of MT over-assembly, block of MT-dynamin interaction, and prevention of seeding protein accumulation can all be a common target of therapeutic development against synaptic defects and early symptoms associated with AD and PD. In this regard, our human iPSC-derived neuronal culture model will provide an intermediate platform for the development of therapeutic reagents.
Footnotes
This work was supported by the Okinawa Institute of Science and Technology to T.T.; KAKENHI Grants JP20K07771 to D.D. We thank Yukiko Goda for comments and Patrick Stoney for editing the paper. We also thank Omar Quaret Sorr and Qiyi Qian for their technical assistance and Toshio Sasaki for electron microscopy. We are grateful to Masato Hasegawa and Tomohiro Miyasaka for providing us with human α-synuclein and tau pffs at the early stage of this study. We also thank Thomas Südhof for sending us the lentiviral plasmids for Ngn2 and rtTA.
The authors declare no competing financial interests.
D.D.’s present address: Synapse Biology Unit, Okinawa Institute of Science and Technology Graduate University, Okinawa 904-0495, Japan.
- Correspondence should be addressed to Dimitar Dimitrov at ddimitrov{at}oist.jp or Tomoyuki Takahashi at ttakahas{at}oist.jp.
