Inhibition of O-GlcNAcase leads to elevation of O-GlcNAc tau and reduction of tauopathy and cerebrospinal fluid tau in rTg4510 mice

Abstract

Background: Hyperphosphorylation of microtubule-associated protein tau is a distinct feature of neurofibrillary tangles (NFTs) that are the hallmark of neurodegenerative tauopathies. O-GlcNAcylation is a lesser known post-translational modification of tau that involves the addition of N-acetylglucosamine onto serine and threonine residues. Inhibition of O-GlcNAcase (OGA), the enzyme responsible for the removal of O-GlcNAc modification, has been shown to reduce tau pathology in several transgenic models. Clarifying the underlying mechanism by which OGA inhibition leads to the reduction of pathological tau and identifying translatable measures to guide human dosing and efficacy determination would significantly facilitate the clinical development of OGA inhibitors for the treatment of tauopathies.

Methods: Genetic and pharmacological approaches are used to evaluate the pharmacodynamic response of OGA inhibition. A panel of quantitative biochemical assays is established to assess the effect of OGA inhibition on pathological tau reduction. A "click" chemistry labeling method is developed for the detection of O-GlcNAcylated tau.

Results: Substantial (>80%) OGA inhibition is required to observe a measurable increase in O-GlcNAcylated proteins in the brain. Sustained and substantial OGA inhibition via chronic treatment with Thiamet G leads to a significant reduction of aggregated tau and several phosphorylated tau species in the insoluble fraction of rTg4510 mouse brain and total tau in cerebrospinal fluid (CSF). O-GlcNAcylated tau is elevated by Thiamet G treatment and is found primarily in the soluble 55 kD tau species, but not in the insoluble 64 kD tau species thought as the pathological entity.

Conclusion: The present study demonstrates that chronic inhibition of OGA reduces pathological tau in the brain and total tau in the CSF of rTg4510 mice, most likely by directly increasing O-GlcNAcylation of tau and thereby maintaining tau in the soluble, non-toxic form by reducing tau aggregation and the accompanying panoply of deleterious post-translational modifications. These results clarify some conflicting observations regarding the effects and mechanism of OGA inhibition on tau pathology, provide pharmacodynamic tools to guide human dosing and identify CSF total tau as a potential translational biomarker. Therefore, this study provides additional support to develop OGA inhibitors as a treatment for Alzheimer's disease and other neurodegenerative tauopathies.

Keywords: Alzheimer’s disease; Neurodegeneration; O-GlcNAc; OGA; Tau; Tauopathy.

Fig. 1

Dose-dependent increase of total O-protein in HEK293 cells treated with Thiamet G. A quantitative sandwich immunoassay (see Methods) was used to detect O-protein after treating the cells with various concentrations of Thiamet G for 6 h. The data shown are the mean ± standard error of the mean (SEM) from a single experiment performed in triplicate and are representative of three independent experiments

Fig. 2

Determination of OGA enzyme and O-protein levels in OGA iKD mice. a Binding of [³H]Thiamet G in brain homogenates prepared from WT (n = 5) and OGA iKD (n = 4) mice was determined as described in Methods. Individual saturation binding curves were performed using total brain homogenates prepared from each animal and the data shown are the mean ± standard deviation of values for each group of animals. The K_D of [³H]Thiamet G binding to OGA was similar in WT and OGA iKD mouse brain (K_D = 1.7 nM and 1.9 nM, respectively), while the B_max values were reduced by 80% in the OGA iKD mouse relative to the WT mouse (433 fmol/mg protein and 78 fmol/mg protein, respectively). b Total O-protein in brain homogenates from WT (n = 5) and OGA iKD mice (n = 5) was measured by a quantitative sandwich immunoassay as described in Methods. Brain O-protein levels were 1.4-fold higher in OGA iKD mice relative to the WT mice. The data shown are the mean ± SEM. Each sample was assayed in duplicate. *p = 0.048 (t-test)

Fig. 3

Brain O-protein levels following acute Thiamet G treatment. C57BL6 mice (n = 8 per group) were given a single oral dose of vehicle (water) or Thiamet G at 10 or 500 mg/kg. Brain and plasma were collected 6 h after dosing. a Concentrations of Thiamet G in plasma (right) and brain (left). b Elevation of O-protein in total brain homogenates after Thiamet G treatment. All data shown are the mean ± SEM. Each sample was assayed in duplicate

Fig. 4

Effect of chronic treatment of rTg4510 mice with Thiamet G on body weight, food intake and brain O-protein. rTg4510 mice (n = 35 per group) were treated with Thiamet G (500 mg/kg/day in the diet) either for 8 weeks from 8 to 16 weeks of age or for 4 weeks from 12 to 16 weeks of age. Body weight and food consumption were monitored weekly. Plasma and brains were collected at the end of the study. Baseline groups of naïve (untreated) rTg4510 mice were also sacrificed at 8 or 12 weeks of age (n = 10 per age group). a Body weight and b food intake as measured weekly during the 8 weeks of the study. Animals in all groups showed an increase in body weight and food intake during the treatment period (p < 0.0001), but no significant differences were observed between the treatment groups (body weight, p = 0.6, food intake, p = 0.22). c and d Plasma and brain Thiamet G concentrations (c) and O-protein levels in total brain homogenates (d) measured at the end of the study. The data shown are the mean ± SEM. Each sample was assayed in duplicate

Fig. 5

Chronic Thiamet G treatment reduces tauopathy in rTg4510 mice. Thiamet G (500 mg/kg/day) or vehicle were administered in diet to rTg4510 mice (n = 35 per group) from either 8–16 or 12–16 weeks of age. Baseline animals (n = 10 per group) were sacrificed at the indicated age and did not receive any treatment. Total tau, aggregated tau and various p-tau species were measured in the brain insoluble fraction using AlphaLISA-based immunoassays as described in Methods. The species of tau measured were: a p-tau recognized by PHF6; b global pThr phosphorylated tau; c tau aggregates; and d total tau. The data shown are the mean ± SEM. Each sample was assayed in duplicate. *p < 0.05 compared to vehicle-treated animals (t-test). e and f Statistically significant correlations between levels of pThr tau, PHF6 tau and tau aggregates were observed using data from all vehicle and Thiamet G-treated animals (r² = 0.98 for pThr tau and PHF6 tau, r² = 0.94 for tau aggregates and PHF6 tau, p < 0.0001 in both cases)

Fig. 6

Effects of chronic Thiamet G treatment of rTg4510 mice on CSF tau. Thiamet G (500 mg/kg/day) or vehicle were administered in diet to rTg4510 mice (n = 35 per group) from either 8–16 or 12–16 weeks of age. Baseline animals (n = 10 per group) were sacrificed at the indicated age and did not receive any treatment. Eight weeks of treatment with Thiamet G treatment significantly reduced CSF total tau (a; *p < 0.05 by t-test) and CSF pT181 tau (b), although the latter effect did not reach statistical significance. CSF total tau and pT181 tau were measured as described in Methods. The data shown are the mean ± SEM. Each sample was assayed in duplicate. c Statistically significant correlation of CSF total tau and pT181 tau in vehicle and Thiamet G-treated groups (r² = 0.86, p < 0.0001)

Fig. 7

Validation of the “click” chemistry method for detection of O-tau. a Flow chart describing the procedure for O-tau detection (see Methods for further details). b Detection of a 55 kDa TAMRA-labeled protein in rTg4510 mouse brain homogenate (upper panel) that is recognized by the HT7 anti-tau antibody (lower panel). Note that this putative human O-tau band migrates at the known molecular weight of soluble human tau (55 kDa) and was seen only in brain samples from rTg4510 mice that overexpress human tau (lanes labeled TG), but not in brain samples from wild type C57BL6 mice (lanes labeled WT). c The 55 kDa TAMRA labeled band was insensitive to PNGase F treatment (PNG), required Gal-T1 treatment and was abolished by β-elimination (β-EL), thus further validating this band as O-tau

Fig. 8

Thiamet G treatment increases O-tau levels. rTg4510 mice were treated with Thiamet G (500 mg/kg/day) for 8 weeks from 8 to 16 weeks of age. Brains were harvested at the end of study to prepare total brain homogenate, soluble and insoluble fractions. a Thiamet G treatment led to elevated O-tau in the total brain homogenate and soluble fractions as detected by TAMRA (TAM) labeling, but had no effect on total tau as detected by HT7 antibody on Western blots (HT7). The data shown are from one vehicle and one Thiamet G-treated animal and are representative of 5 animals in each group. b Quantification of O-tau normalized to total tau from the data shown in a. Unpaired t-test was used for comparison of vehicle and Thiamet G treatment p = 0.0013, 0.0008 and 0.014 for brain stock, soluble and insoluble fractions, respectively