Introduction
The brains of individuals with Alzheimer’s Disease (AD) are characterized by two anatomical hallmarks, beta-amyloid (Aβ)-containing senile plaques and neurofibrillary tangles (NFTs), which consist of twisted fibers of the protein tau. Although most people spontaneously develop NFTs as they age, those with AD tend to develop far more [
4]. Importantly, the degree of dementia in AD patients is highly correlated with the frequency of NFTs [
30]. Thus, understanding the mechanisms of the sporadic development of tau aggregates is necessary to produce effective therapeutic strategies for dementia in sporadic AD. To date, these mechanisms have been poorly understood.
Some studies have examined, both in vitro and in vivo, the possible role of tau phosphorylation on its aggregation [
31,
33,
43], suggesting that under some conditions tau phosphorylation may increase its capacity for self-assembly [
2]. The cellular ability to carry out protein degradation also affects tau accumulation and aggregation. For example, proteasomal inhibition increases the total level of tau and facilitates the formation of detergent-protective tau aggregates in rat brains [
25]. In addition, tau phosphorylation is a signal for its degradation by the ubiquitin-proteasome system [
23]. Perturbation of autophagy also enhances tau aggregation in a cell model consisting of overexpressed human tau [
11]. Furthermore, the stimulation of autophagy ameliorates tau pathology in tau-overexpressing mice, also indicating the involvement of autophagy in tau degradation [
39]. NFTs are found in Niemann-Pick disease type C, which affects lysosomal function, suggesting that impairment of lysosomes might be one of the causes of tau aggregation [
52].
Tau is categorized as a microtubule-associated protein and contributes to the stabilization of the cytoskeleton [
12] and to neuronal development [
5,
44]. Classically, tau has been considered to be involved in axonal functions, such as axonal transport of molecular cargo [
9], because of its stable localization in axons [
3]. Recent studies, however, have demonstrated the activity- or transmitter- dependent expression of tau in the somato-dendritic regions of neurons [
22,
49]. Therefore, tau was also expected to be involved in dendritic functions, as was shown by several groups that reported a critical role for tau in a form of synaptic plasticity, long-term depression (LTD) [
20,
27,
37]. These studies showed that phosphorylation of tau is promoted by LTD-inducing stimuli [
20,
27] and that phosphorylated tau is required for the formation of LTD [
37]. Additionally, a study using the overexpression of wild-type human tau in mice showed that age-dependent accumulation of phosphorylated tau in the somato-dendritic region was negatively correlated with spine number and neural activity [
19]. Furthermore, in AD brains, a similar negative correlation between the accumulation of phosphorylated tau and spine density has been reported [
26]. Those studies suggest that the phosphorylation state of tau influences dendritic functions, such as synaptic functions, in adult and aged brains.
In the present study, we examined whether the tau modification processes that accompanies LTD contributes to the formation of detergent-protective tau aggregates, sarkosyl-insoluble (SI) tau [
24]. We found that LFS formed SI tau aggregates in an age-dependent manner in vivo. Furthermore, the mechanism that leads to the age dependency of tau oligomerization was examined using electrophysiological, biochemical, and pharmacological techniques.
Materials & methods
Animals
C57/BL6J mice were used for all experiments except where otherwise noted. Tau-deficient mice (Mapttm1Hnd/J, The Jackson Laboratory) were maintained by backcrossing with C57/BL6J mice. All mice were kept on a 12-h light/12-h dark cycle at 23 °C and had free access to food and water. In the present study, only male animals were used. Mice were divided into two age categories: adult mice, which were 4–10 months old and aged mice, which were 20–24 months old. More detailed age ranges of the animals used in each experiment are described in the Results.
Drugs and antibodies
The drugs used in this study were as follows: picrotoxin (Sigma), a GABA-receptor inhibitor; two different autophagy inhibitors, 3MA (3-methyladenine; Santa Cruz Biotechnology) and Bafilomycin (Bafilomycin A1; AdipoGen); and the proteasome inhibitor MG132 (Chemscene). The antibodies used were as follows: anti-GluR2 (Millipore), anti-LC3 (M152–3; MLB), A0024 (Daco), T22 (Millipore), Tau5 (Millipore), anti-pS396-tau (Invitrogen), anti-NDP52 (GeneTex), Alexa Fluor 488– and Alexa Fluor 568–conjugated secondary antibodies (Invitrogen), gold-conjugated (5 nm) secondary antibodies (British BioCell International), and horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch Laboratories).
In vivo electrical stimulation and fEPSP recording
The methods for the electrical stimulation and fEPSP measurement in in vivo preparations were based on ones described previously [
20]. Briefly, each mouse was anesthetized with a 1.5% isoflurane–air mixture and fixed in a stereotaxic device (model 900, David Kopf Instruments). The skull of each mouse was exposed, and two holes were bored through the skull to reach the brain surface (with a center position of −1.75 mm from the bregma and 1.75 mm from the midline and −0.5 mm from the bregma and −0.5 mm from the midline). After a 1-h recovery period during which anesthesia was maintained, an electrode assembly and a cannula (#38 syringe tip) were inserted toward the stratum radiatum (projection area of Shaffer’s collaterals) of the hippocampus CA1 region and brain ventricle. The location of the electrode assembly was estimated based on the change in the form of the field excitatory postsynaptic potential (fEPSP) triggered by an electrical pulse with a duration of 100 µs.
When LFS-induced LTD was measured in vivo, 0.0333 Hz electrical stimulation (pulse duration = 100 s) was continuously applied from 1 h after the electrode position was fixed to monitor evoked fEPSPs and their stability. After stable fEPSPs (i.e., the ratio of the minimum to maximum slope was <0.8) were confirmed over a 15-min or longer period (fEPSPs were required to be stable for 1–5 h during the experimental condition), the baseline slope of the test stimulation–induced fEPSP was measured for 15 min. Then, after application of LFS (stimulation with 900 pulses at 1 Hz), the temporal change in the test stimulation–induced fEPSP was measured for 1 h. For this measurement, the electrical signal was amplified 100 times (ER-1; Cygnus Technology), digitized (Digidata 1321A; Axon Instruments), and processed on a computer. The amplitude and slope of each recorded fEPSP and fiber volley were measured by a custom application based on MATLAB (version 2013a, Mathworks Inc.). fEPSPs were analyzed only when the maximal amplitude was >1 mV, and the latency of the minimum peak from the stimulus was <7 ms.
When LFS-induced tau oligomerization was assessed in vivo, the schedule for the electrode and cannula penetration (1 h after the operative treatment), injection of chemical or vehicle solution (3.5 or 4.25 h after operative treatment), and LFS application (4 h after operative treatment) was fixed to normalize effects from anesthetization or operation. For the purpose of the experiment, LFS (900 pulses at 1 Hz) was applied twice successively (1800-pulse protocol) to enhance its influence. For intracerebral ventricle injection of solutions, a syringe pump (KTS310, Morumachi) was used. The injection speed was 1 μl/min, with a total injection volume of 3 μl. The solutions used consisted of 10 μM Bafilomycin in artificial cerebrospinal fluid (aCSF: 124 mM NaCl, 3 mM KCl, 26 mM NaHCO3, 1.25 mM NaH2PO4, 2 mM CaCl2, 1 mM MgSO4, and 10 mM d-glucose, bubbled with 95% O2/5% CO2) containing 1% DMSO or 10 mM 3MA in aCSF. Application of these solutions did not change the basic fEPSP amplitude (data not shown).
The electrode assembly used in this study consisted of two pairs of bipolar electrodes (stimulating and recording electrodes), which were made from tungsten or insulated nichrome wire. The tungsten wire, which had a polished end, was used to minimize physical damage to the tissue in which it was inserted. The nichrome wire, which had a gold-coated tip, was used to minimize the influence of metal elution on physiological responses. There were no notable differences in the stimulating effects on tau aggregation between the two wire types (data not shown).
Acute slice preparations
For in vitro electrophysiology experiments, acute hippocampal slices were obtained from aged and adult mice. After decapitation, the brain was rapidly removed and placed in ice-cold aCSF with 1 mM kynurenic acid. Transverse hippocampal slices (350 μm thick) were prepared using a Vibratome (VT1200S, Leica Biosystems). Hippocampal slices were stored in aCSF (20–25 °C) for 1–2 h before being transferred to the recording chamber, in which they were submerged in aCSF containing 20 μM picrotoxin at 32 °C with a flow rate of 2 ml/min. Picrotoxin, a GABA receptor antagonist, was used to reduce the effect of GABA-related effects. Extracellular field potentials were recorded in the CA1 region using glass electrodes containing aCSF. A stimulating electrode in CA2 was used to evoke fEPSPs with a test stimulus of a single pulse (15–20 μA constant current pulse inducing fEPSPs with a 50% amplitude relative to the maximum, 100-s duration, repeated at 30-s intervals). For this measurement, the electrical signal was amplified 100-fold (ER-1; Cygnus Technology), digitized (Digidata 1321A; Axon Instruments), and processed on a computer. The slope of the evoked fEPSP was measured using custom software based on MATLAB (version 2013a). In this experiment, one or two of the slices obtained from an individual mouse were used for drug application, and one or two of the remaining slices were used for vehicle application. Drug solutions and their respective vehicles were as follows: 0.1 μM Bafilomycin and aCSF containing 0.01% DMSO and 20 μM picrotoxin, 1 mM 3MA and aCSF containing 1% water and 20 μM picrotoxin, and 0.1 μM MG132 and aCSF containing 0.01% DMSO and 20 μM picrotoxin.
Sarkosyl-insoluble (SI) fraction
Each isolated hippocampus was weighed and homogenized with 30 vol of cold HEPES-sucrose buffer (HSB: 320 mM sucrose; 4 mM HEPES; 2 mM EDTA, pH 7.4) with protease inhibitors (Sigma, diluted 1:100) and phosphatase inhibitors (Nacalai Tesque, diluted 1:100) and was centrifuged at 1000 g for 15 min at 4 °C to remove nuclear material and cell debris, resulting in the S1 fraction. Then, a high-salt sarkosyl fraction was obtained by adding an equal volume of extra-high-salt HSB (HSB with 1.6 M NaCl) with 2% sarkosyl solution to the S1 fraction. This mixture was incubated at 37 °C for 2 h and then was separated into the SI fraction and the sarkosyl-soluble (SS) fraction by centrifugation (200,000 g, 4 °C, 1 h).
SDS-PAGE and western blotting
The methods of SDS-PAGE and western blotting were previously described [
19]. In brief, each obtained fraction was analyzed by SDS-PAGE and western blotting. For SDS-PAGE, each fraction obtained was suspended in Laemmli sample buffer and subjected to SDS-PAGE using a 5–20% gradient gel (Wako). Separated proteins were blotted onto polyvinylidene difluoride (PVDF) membranes (GE Healthcare). The membranes were incubated with primary antibody (room temperature, 2 h) in Tris-buffered saline (TBS; 50 mM TrisHCl, 500 mM NaCl, pH 7.6), followed by the appropriate species of horseradish peroxidase–conjugated secondary antibody (room temperature, 30 min) in TBS. Chemiluminescent detection (ECL, GE Healthcare) was used for visualization. Quantification and visual analysis of immunoreactivity were performed with a computer-linked LAS-4000 Bio-Imaging Analyzer System (GE Healthcare). Antibody dilutions were as follows: A0024, 1:20,000; anti-ps395 tau, 1:1000; tau5, 1:500; anti-LC3, 1:1000; anti-NDP52, 1:1000; anti-GluA2, 1:1000; all secondary antibodies, 1:10,000.
Blue native (BN)-PAGE and western blotting
For BN-PAGE analysis, each S1 fraction obtained from a hippocampus was subjected to centrifugation (12,500 g, 4 °C, 20 min) and divided into the crude synaptosomal (P2) fraction (i.e., the pellet), in which PSD-95 was largely recovered, and the synaptosome-depleted fraction (S2). The P2 fraction pellet was suspended in 50 μl native lysis buffer (NativePAGE Sample Prep Kit, Invitrogen) with 0.1% Triton X-100 and run on a Tris-Bis gradient gel (3–12% Bis-Tris Protein Gels, Novex). After blotting to a PVDF membrane using a transfer tank (Mini Blot Module, Novex), the tau oligomers labeled with tau oligomer–specific antibody T22 (diluted 1:500 in TBS) or anti-tau A0024 (diluted 1:20,000 in TBS) were visualized by a chemiluminescence method using the LAS-4000 system.
Immunoprecipitation
To analyze the components of the molecular complex detected by T22, a commercial immunoprecipitation kit (direct magnetic IP/Co-IP kit, Pierce) was used. To bind T22 antibody on magnetic beads (NHS-activated magnetic beads; Pierce), 10 μg of magnetic beads was washed with ice-cold 1 mM HCl, and then 500 μl of T22 solution (10 μl T22 diluted in 500 μl TBS) was added, and the mixture was incubated at room temperature for 1 h. The beads were isolated with a magnet and washed with Elution buffer (Pierce), and the reaction was quenched by the addition of 3 M ethanolamine. Then each P2 fraction pellet, which had been diluted in 200 μl HSB with 0.5% Triton X-100, protease inhibitors (diluted 1:200), and phosphatase inhibitors (diluted 1:200), was exposed to the antibody-bound beads overnight at 4 °C. The beads were again isolated with a magnet and washed with HSB and then distilled water before being rinsed in 10 μl Elution buffer to recover tau oligomers. After neutralization by the addition of 1 μl of 2 M Tris solution, the resulting immunoprecipitated proteins were analyzed by SDS-PAGE and western blotting as described above.
Electron microscopy
The morphological features of tau oligomers in the SI fraction were investigated by an immunogold electron microscopy technique [
51]. In brief, the SI pellet obtained from each hippocampus was washed three times with TBS and resuspended in 50 μl TBS. For electron microscopy, the TBS solution was incubated with primary antibody (A0024, diluted 1:200 or T22, diluted 1:50) in TBS for 2 h at 4 °C. After washing, the samples were absorbed onto glow-discharged supporting membranes on 400-mesh grids and incubated with a 5-nm colloidal gold–conjugated secondary antibody (diluted 1:200) for 90 min. After fixing with 2% glutaraldehyde, grids were negatively stained with 2% sodium phosphotungstic acid, dried, and then examined with a transmission electron microscope (Tecnai 12, FEI). In examination using T22 antibody, SI samples suspended in TBS were directly exposed to a supporting membrane on a mesh and were immunostained with the same primary and secondary antibodies on the mesh after blocking with 0.05% bovine serum albumin (BSA) and 1% normal horse serum.
Immunohistochemistry and immunofluorescence staining
Mice were deeply anesthetized with pentobarbital (50 mg/kg) and then transcardially perfused with 10% formalin. Brains were postfixed in the same fixative for 16 h and embedded in paraffin and sectioned (4–6 mm) in the coronal plane. Deparaffinized sections were treated with Target Retrieval Solution (Dako) for 20 min at 80 °C, blocked in 0.1% BSA/TBS, and incubated with primary antibodies (anti-LC3, 1:500; anti-GluA2, 1:1000) in 0.1% BSA/TBS overnight at 4 °C. Immunohistochemical staining was performed with immPRESS Reagent kit (Vector) and immPACT DAB (Vector). For fluorescent staining, slices were incubated with Alexa Fluor 488– and Alexa Fluor 568–conjugated secondary antibodies (1:500) in 0.1% BSA/TBS overnight at 4 °C. A LSM700 laser confocal microscope (Zeiss) was used for fluorescent observations.
Statistical analysis
In the present study, when the normalized levels of the SI tau in ipsilaterally stimulated hippocampi were compared to their internal controls (i.e., contralateral ones obtained from the same animals), the one-sample t-test against a theoretical value of ‘1’ was used because the normalized level of contralateral ones was logically ‘1’. When the normalized levels of the SI tau in stimulated hippocampi were compared to the ones in another group, unpaired t-test was mainly used. In other cases, we used unpaired t-test, paired t-test or two-ways ANOVA. These analyses were performed with Prism 7 (GraphPad Software, Inc.).
Discussion
The present study showed that LFS-induced LTD in aged hippocampus is critically dependent on the throughput level of the 3MA- and Bafilomycin-sensitive pathway but not on the MG132-sensitive pathway. This suggests that a part of the LTD cascade switches with age from a proteasome-sensitive to autophagy-sensitive one. Contributions of the MG132-sensitive pathway on NMDA-induced, AMPA-induced, and electrical stimulus–induced internalization of AMPARs have been described (see [
13]). In contrast, NMDA application triggers autophagy, which contributes to AMPAR internalization and/or trafficking in cultured hippocampal neurons, although the physiological role of NMDA-induced autophagy is unclear [
46]. These findings suggest that neurons are able to use those different protein-degradation pathways to control AMPAR internalization. Age-dependent reduction of proteasome activity in the hippocampus, cortex, and spinal cord in rats has been described [
18]. Thus, it is possible that the alteration influences neuronal selection of a protein-degradation pathway relating to LFS-induced LTD. At a minimum, the age-dependent alteration in the response to the drugs used in this study represents an age-dependent change in the dynamic properties of the protein degradation system, which is required for LTD.
Meanwhile, the present study demonstrates the importance of the switching of the protein-degradation pathway in an age-dependent manner relative to the risk of tauopathy formation, because, as shown here, LFS induces not only LTD but also tau aggregation in the aged hippocampus. Similar to LTD formation in the aged mice, tau aggregation was also sensitive to the ALP, especially during the process of autophagosome formation. It was reported that the autophagic marker p62/SQSTM1 accumulates and colocalizes with hyperphosphorylated tau in human tauopathies [
35]. In addition, studies that analyzed the aggregation mechanism of pro-aggregate mutant tau have also suggested the importance of chaperone-mediated autophagy for the formation of tau aggregates [
8,
53]. As there are notable differences between those studies and our study with respect to the tau species used and the experimental system, the results cannot be directly compared. However, those independent studies lead to a similar conclusion that the early stage of tau aggregation in vivo requires a certain protein accumulation mechanism.
Extracellular tau detected in human cerebrospinal fluid increases in an age-dependent and pathogenesis-dependent manner [
15,
48]. Analysis of interstitial fluid in a tauopathy mouse model has shown an extracellular secretion of tau aggregates [
50]. Moreover, extracellular tau oligomers are taken up by neurons via endocytosis mechanisms and contribute to tau pathology [
10]. If tau oligomers form within autophagosomes or lysosomes as a result of LFS, they could be candidates for subsequently becoming extracellular tau aggregates, because tau aggregates within intracellular vesicles are secreted extracellularly and can propagate to other cells [
15]. In contrast, as in the case of tau aggregation during chaperone-mediated autophagy [
8,
53], there is the possibility that tau aggregation occurs on the lysosomal/phagosomal surface. In this case, the LTD-ALP cascade may be involved more directly in cytoplasmic pathogenesis during tauopathy. Therefore, the early phase of the LTD-ALP cascade potentially acts as a supplier of tau aggregates in aged brains, which contributes to the pathogenesis in direct and/or indirect ways.
The analysis of AD model mice has demonstrated that over-expressing mutant Aβ alters electrical activity patterns in the brain [
1,
28] and shifts the characteristics of synaptic plasticity (i.e., long-term potentiation is down-regulated and LTD is up-regulated) [
14,
38]. Thus, the toxic form of Aβ potentially enhances the tau-aggregation cascade that accompanies LTD. In AD model mice, neurons that accumulate Aβ oligomers show impaired proteasome activity [
45] and enhanced activation of autophagy [
6]. Therefore, it is possible that those AD-related factors may influence tau aggregation via the LTD-ALP cascade.