Tauopathy induced by low level expression of a human brain-derived tau fragment in mice is rescued by phenylbutyrate

Abstract

Human neurodegenerative tauopathies exhibit pathological tau aggregates in the brain along with diverse clinical features including cognitive and motor dysfunction. Post-translational modifications including phosphorylation, ubiquitination and truncation, are characteristic features of tau present in the brain in human tauopathy. We have previously reported an N-terminally truncated form of tau in human brain that is associated with the development of tauopathy and is highly phosphorylated. We have generated a new mouse model of tauopathy in which this human brain-derived, 35 kDa tau fragment (Tau35) is expressed in the absence of any mutation and under the control of the human tau promoter. Most existing mouse models of tauopathy overexpress mutant tau at levels that do not occur in human neurodegenerative disease, whereas Tau35 transgene expression is equivalent to less than 10% of that of endogenous mouse tau. Tau35 mice recapitulate key features of human tauopathies, including aggregated and abnormally phosphorylated tau, progressive cognitive and motor deficits, autophagic/lysosomal dysfunction, loss of synaptic protein, and reduced life-span. Importantly, we found that sodium 4-phenylbutyrate (Buphenyl®), a drug used to treat urea cycle disorders and currently in clinical trials for a range of neurodegenerative diseases, reverses the observed abnormalities in tau and autophagy, behavioural deficits, and loss of synapsin 1 in Tau35 mice. Our results show for the first time that, unlike other tau transgenic mouse models, minimal expression of a human disease-associated tau fragment in Tau35 mice causes a profound and progressive tauopathy and cognitive changes, which are rescued by pharmacological intervention using a clinically approved drug. These novel Tau35 mice therefore represent a highly disease-relevant animal model in which to investigate molecular mechanisms and to develop novel treatments for human tauopathies.

Keywords: 4-phenylbutyrate; lysosomal degradation; progressive supranuclear palsy; tau; tauopathies.

© The Author (2016). Published by Oxford University Press on behalf of the Guarantors of Brain.

Figures

Post-translational modification of tau is common in human tauopathies. Bondulich et al. generate transgenic mice expressing low levels of a truncated form of tau (Tau35) that is associated with human tauopathy. Tau35 mice develop progressive tau neuropathology and cognitive impairment, modelling human disease. The approved drug 4-phenylbutyrate rescues these abnormalities.

Figure 1

Tau expression in human tauopathy and Tau35 mouse brain. (A) Construct used to generate Tau35 mice with the human tau promoter (phTau), upstream of the Tau35 sequence with the haemagglutinin tag (HA). The hypoxanthine phosphoribosyltransferase promoter (pHprt) and exons 1–3 enabled targeted integration of the Tau35-HA transgene. The lower panel shows a schematic representation of the expressed Tau35-HA protein in comparison to full-length human tau (441 amino acids). The amino terminal domain of tau contains two inserts (N1, N2), followed by a central proline-rich domain and the microtubule (MT) binding domain, which comprises four repeats (R1-R4). Tau35 retains the majority of the proline-rich domain, four MT binding repeats and an intact C-terminus. The epitopes of the phospho-dependent (orange boxes), conformation-dependent (black boxes) and region-specific (grey boxes) tau antibodies used in this study are indicated above full-length tau. (B) Western blot of insoluble fractions of control and progressive supranuclear palsy (PSP) brain reveals the 35 kDa tau species (asterisk) in the premotor cortex of human PSP brain probed with TP70, the C-terminal tau antibody. (C) Reverse transcription (RT)-PCR to confirm Tau35 expression. Primers recognizing mouse tau (exons 7 and 13) generate a 612 bp band, corresponding to endogenous mouse tau in wild-type (WT) and Tau35 mice. Primers recognizing tau exon 9 and haemagglutinin generate a 558 bp band only in Tau35 mice. (D) RT-PCR of endogenous mouse and total tau in wild-type and Tau35 mice. PCR products amplified using primers for mouse tau (exons 7 and 13, 612 bp) or total tau (exons 9 and 13, 348 bp) in wild-type and Tau35 mice, standardized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Results are shown as mean ± standard error of the mean (SEM). (n = 6 mice for each genotype). (E) Sagittal sections show widespread haemagglutinin labelling in Tau35 mouse brain (upper panels, scale bar = 2 mm). Higher magnifications of the hippocampal CA1 region show strongly haemagglutinin-positive pyramidal neurons in Tau35 mice (lower panels, scale bar = 200 μm). Western blots of frontal region and hippocampus/associated cortex (HC) show haemagglutinin protein expression only in Tau35 mice.

Figure 2

Progressive neuromuscular impairment and reduced survival of Tau35 mice. (A) Limb clasping is apparent in Tau35 mice from 2 months of age (image shows 8 months), with all Tau35 animals affected by 18 months (n = 40 mice). Clasping is not observed in wild-type (WT) mice at any age examined. (B) Spine curvature is apparent in Tau35 but not wild-type mice, at 14 months of age. A progressive reduction in the kyphotic index in Tau35 mice after 4 months of age indicates increasing spine curvature. Values shown are mean ± SEM, n = 8 mice for each genotype. (C) Kaplan-Meier survival plots reveal a median lifespan of 717 days for Tau35 mice, compared to 788 days for wild-type mice (P < 0.05, log-rank test, pairwise multiple comparison and Bonferroni correction), n = 10 mice for each genotype. (D) Tau35 mice have a significantly reduced latency to fall from an accelerating rotarod at all ages tested (4–16 months). Values shown are mean ± SEM, n = 8 mice for each genotype. (E) The grip strength of Tau35 mice declines steadily with age. Values shown are mean ± SEM, n = 8 for each genotype. (F) Haematoxylin and eosin staining of quadriceps and latissimus muscle sections from wild-type and Tau35 mice at 8 and 16 months. Muscle fibres from Tau35 mice show internal nuclei at 8 and 16 months (arrows). Scale bar = 60 μm. Graphs show increased internal nuclei (8 and 16 months) and altered distribution of muscle fibre diameter (minimal Feret’s diameter, 16 months) in Tau35 mice. Values shown are mean ± SEM, n = 3 mice for each genotype. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3

Tau35 mice exhibit reduced hippocampal-dependent spatial learning ability in the Morris water maze. (A) Visible platform (VP) training in the Morris water maze was followed by 4 days of hidden platform training. Escape latencies were tracked for independent cohorts of Tau35 and wild-type (WT) mice at 4, 6, 8, 10 and 12 months of age. At 8, 10 and 12 months of age, Tau35 mice exhibit longer escape latency on the fourth day of testing compared to wild-type mice. The distance swam by Tau35 mice searching for the hidden platform is significantly increased compared to wild-type mice at 8, 10 and 12 months of age. (B) During the probe trial (1 min), the percentage occupancy of the target quadrant is significantly reduced for Tau35 mice compared to wild-type mice at 8 months. Tau35 and wild-type mice exhibit equivalent swim speeds during the probe trial indicating no apparent impairment in swimming ability at the ages tested. Values shown in A and B are mean ± SEM, n = 8 mice for each genotype. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4

Tau pathology in Tau35 mouse brain. CA1 hippocampal sections of Tau35 mice at 2, 8, and 14-16 months of age, and wild-type (WT) mice aged 16 months (right), labelled with tau antibodies, PHF1, TOC1, MC1, AT8 and TP007, and counterstained with haematoxylin. PHF1 revealed tau-positive labelling at 2 months. Tau-positive labelling of inclusions in Tau35 brain, including dystrophic neurites and neuropil threads, was apparent with antibodies PHF1, TOC1, MC1, AT8 and TP007 at 14-16 months (insets). PHF1 revealed increased cytoplasmic labelling at 2 months and tau-containing inclusions from 8 months, which were more abundant at 14–16 months. Wild-type mouse hippocampal sections showed no labelling of inclusions with these antibodies at 14–16 months of age (right). n = 3 mice for each genotype, scale bar = 200 µm.

Figure 5

Tau phosphorylation changes in Tau35 mouse hippocampus. Western blots of (A) PHF1, total tau, and β-actin, (B) TG3 and total tau, and (C) AT270 and total tau, reveal a significant increase in tau phosphorylation in Tau35 mice at each phosphoepitope, compared to wild-type (WT) mice, whereas the total amount of tau relative to actin is equivalent in both genotypes. (D) Western blots of Tau1 and total tau, show a reduction in tau phosphorylation in Tau35 mice. Values shown are mean ± SEM, n = 6 mice for each genotype. *P < 0.05, ***P < 0.001.

Figure 6

Biochemical changes in Tau35 mouse hippocampus. Western blots show (A) phosphorylated (inactive) glycogen synthase kinase-3β (pGSK3β) relative to total (T-) GSK3β, is significantly decreased in Tau35 mice, indicating increased GSK3β activity. The total amounts of GSK3α and GSK3β, and of pGSK3α, are unchanged in Tau35 mice. (B) Western blot shows the amounts of LC3-I and LC3-II, relative to actin, are increased in Tau35 mice. (C) p62 is significantly increased, relative to actin, in Tau35 mice. (D) Mature (active) cathepsin D is reduced in Tau35 mice, whilst the amount of pro-cathepsin D is unchanged. (E) Western blots show a small but significant decrease in acetylated α-tubulin in Tau35 mice. The total amount of tubulin relative to actin is equivalent in both genotypes. (F) The amount of synapsin 1 is reduced relative to tubulin, whereas synaptophysin is unchanged in Tau35 mice (G). Values shown are mean ± SEM, n = 6 mice for each genotype. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7

Phenylbutyrate rescues disease-related changes in Tau35 mice. (A) Morris water maze (10 months) testing of 4-phenylbutyrate (PBA, dotted lines) and vehicle-treated (solid lines) Tau35 (circles) and wild-type (WT, squares) mice. (B) Tau35 mice treated with PBA show improved learning in the probe trial at 10 months of age. (C) Grip strength of Tau35 (10 months) is restored by PBA treatment. Values shown in A–C are mean ± SEM, n = 8 mice for each group. Western blots show that Tau35 mice treated with PBA have (D) reduced phosphorylated tau (PHF1), relative to total tau, whereas the amounts of LC3-I and LC3-II are unchanged, relative to actin (E). PBA significantly reduced p62 in Tau35 mice compared to vehicle-treated animals (F). The amount of mature cathepsin D, but not pro-cathepsin D, is rescued by PBA treatment of Tau35 mice (G). Acetylated α-tubulin (H) and synapsin 1 (I) are both increased in Tau35 mice following PBA treatment, whereas the amount of synaptophysin is unchanged (J). Values shown in D–J are mean ± SEM, n = 5 mice for each group. (K) Haematoxylin and eosin staining of quadriceps muscle sections from Tau35 mice treated with vehicle (left) or PBA (right). Muscle fibres from vehicle-treated Tau35 mice show internal nuclei at 10 months (left, arrows). Scale bar = 60 μm. Graph shows reduced internal nuclei in Tau35 mice treated with PBA. Values shown are mean ± SEM, n = 3 mice for each group. *P < 0.05, **P < 0.01, ***P < 0.001.