Background
Tauopathies are a heterogeneous group comprising dementias and movement disorders, neuropathologically characterized by prominent intracellular accumulations of neurofibrillary tangles formed of tau in neurons and glia. Accumulating evidence suggests that the conversion of physiological tau to pathological tau plays a central role in the development of tauopathy. In particular, abnormal phosphorylation and fragmentation of tau have been proposed as important post-translational modifications that lead to pathogenic forms of tau [
19]. In addition, a range of inter-related cellular processes, including microtubule disorganization [
18,
55,
56], activation of the unfolded protein response (UPR) [
22,
34,
52], activation of the nutrient sensor mammalian target of rapamycin complex 1 (mTORC1) [
9,
49], and deficiencies in insulin signaling [
41,
45], also promote cell dysfunction in tau-mediated neurodegeneration. However, the cellular events linking pathological changes in tau to cell dysfunction and the pathogenesis of tauopathies are largely unknown.
We previously described a 35 kDa C-terminal tau fragment (Tau35), lacking the N-terminus of tau but containing all four microtubule-binding repeats (4R), that is present in 4R tauopathies [
53]. When expressed in transgenic mice, Tau35 induces several key features of tauopathy, including accumulation of abnormally phosphorylated tau, dysregulation of glycogen synthase kinase-3β (GSK3β) activity, progressive cognitive and motor deficits, and loss of synaptic proteins [
7]. Here we have used a cell model to investigate the molecular mechanisms that are affected by Tau35 expression. Our findings suggest that aberrant tau cleavage may have a key role in disrupting physiological signaling pathways involved in the development of tauopathy.
Materials and methods
Plasmids
Plasmids encoding full-length 2N4R human tau (FL-tau) or Tau35 were generated in pcDNA 3.1D/V5-His-TOPO vector (Invitrogen), which carries a neomycin resistance gene, a V5 epitope, a 6 × His tag and the promoter from cytomegalovirus. The original plasmid encoding FL-tau in bacterial expression vector pRK172 was a kind gift from Professor Michel Goedert (MRC Laboratory of Molecular Biology, Cambridge, UK). cDNA sequences corresponding to FL-tau and Tau35 were each inserted into the multiple cloning site of the pcDNA 3.1D/V5-His-TOPO vector at BamHI-XbaI [
12]. Alpha-tubulin N-acetyltransferase 1 (αTAT1) plasmid [
2] was a kind gift from Professor Jacek Gaertig (University of Georgia, USA).
Cell maintenance and transfection
Mycoplasma negative Chinese hamster ovary (CHO) cells, acquired from the European Collection of Authenticated Cell Cultures, were grown at 37 °C with 5% CO2 in Ham’s F12 medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 units/mL penicillin and 100 μg/mL streptomycin (Thermo Fisher Scientific). 24 h before transfection, cells were plated at a density of 3.7 × 104 cells/cm2 in 6-well or 12-well plates. CHO cells were transiently transfected with plasmids (2 μg plasmid/well for 6-well plate or 1 μg plasmid/well for 12-well plate) using jetPEI™ (Polyplus Transfection) according to the manufacturer’s instructions. 24 or 48 h after transfection, cells were processed for biochemical assays, or fixed for immunocytochemistry.
Generation of stable CHO cell lines
CHO cells were transiently transfected with plasmids encoding either FL-tau or Tau35, as described above. Non-transfected CHO cells were included as controls. 48 h after transfection the medium was replaced by Ham’s F-12 medium as above, with the addition of 800 μg/mL G418 (Santa Cruz). After selection, G418-resistant cells were transferred to 145 mm diameter dishes for clonal isolation. Cell clusters were isolated using cloning cylinders (Sigma) and transferred to 6-well plates for clonal expansion. Further characterization of the G418-resistant cells was undertaken using western blots to examine the stable expression of tau protein and immunocytochemistry to assess the homogeneity of the cell lines. Clonal cells homogenously expressing 2N4R tau or Tau35, termed CHO-FL and CHO-Tau35, respectively, were selected and maintained in CHO cell growth medium without G418 at 37 °C in 5% CO2.
Cell treatments
24 h before treatment, CHO-FL, CHO-Tau35 and untransfected CHO cells were seeded at a density of 3.7 × 104 cells/cm2. For insulin treatment, cells were treated with 100 nM insulin (Sigma) for 30 min at 37 °C before washing in phosphate-buffered saline (PBS, 137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4, 2 mM KH2PO4, pH 7.4). For LiCl treatment, cells were treated with 5 mM LiCl, or 5 mM NaCl (control) at 37 °C for 24 h, then washed with PBS. For thapsigargin treatment, cells were treated with 800 nM thapsigargin for 5 h at 37 °C, then washed with PBS. After treatment, cells were either scraped into ice-cold Tris-HCl buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM Na3VO4, Complete protease inhibitor and Complete protease inhibitor cocktail [Roche]), lysed in 2× Laemmli sample buffer and heated at 95 °C for 10 min for analysis on western blots, or fixed for immunocytochemistry, as described below.
In situ microtubule binding assay
In situ microtubule binding was assayed as described previously [
40]. Briefly, 24 h before the experiment, CHO-FL and CHO-Tau35 cells were plated (3.7 × 10
4 cells/cm
2). Cells were rinsed with warm PBS and scraped into warm PIPES buffer (80 mM piperazine-N,N′bis-2-ethanesulfonic acid, pH 6.8, 1 mM guanosine-5′-triphosphate, 1 mM MgCl
2, 1 mM ethylene glycol-bis(2-aminoethyl)-N,N,N′,N′-tetraacetic acid, 0.5% (
w/
v) Triton X-100 and 30% (
v/v) glycerol) containing Complete protease inhibitor (Roche), 20 mM NaF, 0.5 μM okadaic acid (Merck), and 10 μM taxol (Sigma). Cell lysates were centrifuged at 5000 g for 10 min at ambient temperature, and an aliquot of the supernatant was retained (total, T). The remaining post-nuclear lysate was centrifuged at 100,000 g for 1 h at 37 °C. The supernatant (unbound fraction, U) was collected, and the pellet (bound fraction, B) was rinsed twice in PIPES buffer, pelleted at 100,000 g, and then resuspended in PIPES buffer. All fractions were suspended in 2× Laemmli sample buffer and heated at 95 °C for 10 min prior to analysis on western blots.
Western blots
Proteins in cell lysates and sub-cellular fractions were separated on sodium dodecyl sulfate polyacrylamide gel electrophoresis. Electrophoresed proteins were transferred onto 0.45 μm nitrocellulose membranes. Membranes were blocked in Odyssey blocking buffer (Li-Cor Biosciences), 3% (
w/
v) dried skimmed milk in Tris-buffered saline/0.2% (
v/
v) Tween 20 (TBST), or 5% (
w/v) bovine serum albumin in TBST for 30 min at ambient temperature, then incubated overnight at 4 °C in primary antibodies (Additional file
1: Table S1). After washing, membranes were incubated for 60 min at ambient temperature with the appropriate fluorophore-conjugated secondary antibody (Alexa Fluor® 680 goat anti-mouse immunoglobulin G (IgG) or IRDye™ 800 goat anti-rabbit IgG, Invitrogen). Antigens were visualized using an Odyssey® infrared imaging system (Li-Cor Biosciences). Images were analyzed using Li-Cor Image Studio Lite software (Li-Cor Biosciences).
Immunocytochemistry
Cells on coverslips were washed 3 times in PBS, fixed for 10 min in 4% (
w/
v) paraformaldehyde at 37 °C or in ice-cold methanol at − 20 °C for 10 min. Paraformaldehyde-fixed cells were permeabilized using 0.25% (
v/
v) Triton X-100 for 10 min at ambient temperature and washed in PBS. Following incubation in blocking buffer (10% (
v/v) fetal bovine serum in PBS, pH 7.4) for 30 min at ambient temperature. Cells were incubated in primary antibody overnight at 4 °C, followed by secondary antibody for 60 min at ambient temperature (Additional file
1: Table S1). Nuclei were stained using Hoechst 33342 (5 μg/mL bisbenzimide in PBS). Fluorescence microscopy was performed using a Leica DM5000B fluorescence microscope equipped with a 63×/1.25 immersion objective, and a digital camera (DFC360 FX, Leica) using Leica Application Suite Advanced Fluorescence Software (Leica). Images were processed and analyzed using ImageJ software [
44].
Quantification and statistical analysis
Quantitative analyses were performed using Microsoft Excel and GraphPad Prism 7. Normality for individual variables was determined by the Shapiro-Wilk test and analyzed using Student’s unpaired t-test or one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc tests. Differences were considered statistically significant when P < 0.05.
Discussion
We previously identified a 35 kDa C-terminal tau fragment termed Tau35 (residues 187–441 of FL-tau), in 4R human tauopathy brain [
53]. Tau35 is generated by cleavage of human tau, resulting in a tau fragment that lacks the N-terminal domain and part of the proline-rich domain, but contains all four microtubule-binding repeats and an intact C-terminus (Fig.
6e). Minimal expression of Tau35 in transgenic mice causes several key features of human tauopathy [
7]. Here we investigated the molecular mechanisms underlying the development of disease-related phenotypes using cells stably expressing Tau35.
Tau phosphorylation plays a key role in regulating tau localization and function, and aberrant phosphorylation of tau reduces its ability to bind to microtubules [
16,
46]. When expressed in CHO cells, Tau35 displayed elevated phosphorylation at several epitopes associated with the development of human tauopathy, in which aggregates of highly phosphorylated and fragmented tau are present [
4,
53], highlighting the relevance of this Tau35 cell model to human tauopathy.
Tau35 has a reduced ability to bind to microtubules, despite the presence of all four microtubule binding repeats (residues 244–401 of human FL-tau) and an intact C-terminus. Moreover, reducing Tau35 phosphorylation using lithium did not increase the interaction between Tau35 and microtubules, suggesting that the reduced binding was not due to increased phosphorylation of Tau35 in CHO cells. Our findings therefore support the view that the reduced microtubule binding ability of Tau35 is due to the absence of amino acid sequences present in the N-terminal half of tau. These results parallel those of others showing that N-terminally cleaved tau species display altered interactions with microtubules [
11,
31,
57] and indicate that the N-terminal region of tau is important for its association with microtubules. It is possible that an extended region of tau, encompassing domains outside the microtubule binding and flanking regions, may be required in order to facilitate microtubule binding and stabilization since truncated forms of tau corresponding to residues 1–255 and 256–441 also exhibit reduced abilities to polymerize microtubules [
57]. Notably, unlike Tau35, a different tau fragment comprising residues 124–441, displays an increased ability to bind and stabilize microtubules compared to FL-tau [
11]. Taken together, these findings indicate that a sequence of amino acids located between residues 124–186 of FL-tau may be critical for its interaction with microtubules.
Tau35 was unable to induce significant microtubule bundling, even after enhancing tubulin acetylation, consistent with the proposal that microtubule bundling is promoted by complementary intermolecular dimerization between the N-terminus and the proline-rich domains of tau [
42]. Such a model provides an explanation for this loss of function of Tau35, since it lacks N-terminal tau residues, which may be required for regulating microtubule organization.
The adverse effects of Tau35 on insulin signaling also distinguish it from FL-tau. In particular, FL-tau interacts with and reduces the activity of phosphatase and tensin homologue on chromosome 10 (PTEN), thereby promoting Akt activation [
30]. Notably, knocking out tau also reduces Akt activity and impairs the hippocampal response to insulin [
30]. Our finding of increased Akt activity in the presence of FL-tau supports this view, whereas cleavage of tau to generate Tau35 could potentially prevent or perturb its interaction with PTEN, attenuating Akt activation and resulting in enhanced GSK3 activity. Defective inactivation of GSK3β in CHO-Tau35 cells was also observed in CHO-Tau35 cells exposed to insulin. Whereas FL-tau facilitates insulin signaling, attenuated Akt phosphorylation in response to insulin in CHO-Tau35 cells suggests the induction of insulin resistance.
We also identified activation of mTORC1/SK61 signaling as a potential mediator of the adverse impact of Tau35 on insulin signaling in CHO-Tau35 cells, resulting in IRS1 phosphorylation. Altered IRS1 phosphorylation and insulin resistance have been reported in several tauopathies, including Alzheimer’s disease, progressive supranuclear palsy and corticobasal degeneration [
54]. However, the mechanisms leading from pathological changes in tau, to insulin resistance in human tauopathy are not well understood. Notably, upregulation of mTORC1 activity increases both tau phosphorylation and tau pathology [
9]. Our data therefore support the view that Tau35 may trigger inhibitory phosphorylation of IRS1 through activation of mTORC1/S6K1 signaling, which in turn exacerbates phosphorylation of Tau35.
Intriguingly, we identified chronic activation of the PERK and ATF6 branches of the UPR induced by Tau35, which has also been found in human tauopathy [
8,
39,
47]. Activation of the PERK and ATF6α branches of the UPR lead to expression of pro-apoptotic factor CHOP, which is also elevated by Tau35 [
21,
25]. Interestingly, the IRE1α branch of the UPR does not appear to be affected, indicating selectivity in Tau35-induced UPR activation. Prolonged PERK signaling impairs cell proliferation and promotes apoptosis, whereas IRE1α signaling enhances cell proliferation [
27]. Such divergence in the activation of PERK and IRE1α may be indicative of persistent ER stress [
26,
27] in CHO-Tau35 cells, which ultimately results in an imbalance between detrimental and protective effects of UPR activation. It has been suggested that PERK can facilitate the translation and activation of ATF6α [
48], thus, it will be of interest to examine whether Tau35 directly triggers ATF6α activation, or whether this is the result of prior PERK activation. Tau35 also renders CHO cells more susceptible to thapsigargin-mediated activation of the UPR, which has been linked to accumulation of abnormally phosphorylated tau [
1,
29]. The mechanisms that contribute to UPR activation in tauopathy are unclear. It has been proposed that soluble tau oligomers are the driving force behind tau-induced ER stress [
1] and these could impair ER-associated degradation, resulting in UPR activation [
14,
38].
Activation of both mTORC1/S6K1 and UPR signaling are associated with neurodegenerative disease [
3] and importantly, crosstalk between these two pathways is increasingly recognised. Such interactions include ATF6α-mediated upregulation of mTORC1 [
43] and UPR activation contributing to insulin resistance [
6,
33]. Given the inter-dependence of these two pathways, suppressed insulin signaling may be the synergistic consequence of activation of both the UPR and mTORC1/S6K1 pathways in CHO-Tau35 cells and potentially also in human tauopathy brain.
Acknowledgements
We thank Professor Peter Davies (The Feinstein Institute for Medical Research) for providing PHF1 antibody. We thank Professor Michel Goedert (MRC Laboratory of Molecular Biology, Cambridge, UK) for the kind gift of the Tau40 plasmid that was used to generate the constructs herein. We also thank Professor Jacek Gaertig (University of Georgia, USA) for generously providing the plasmid expressing αTAT1.