All reagents and antibodies were purchased from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise stated. Thiamet G was synthesized at the Department of Medicinal Chemistry, H. Lundbeck A/S, Valby, Denmark according to published methods [26]. Thiamet G at 10 µM was tested in an in vitro diversity profile panel composed of 71 binding and 27 enzyme assays broadly defined with roughly an equal number of selective, central, and peripheral therapeutically relevant targets (Eurofins Cerep, Celle-Lévescault, France).

Murine cortical neurons were isolated from day E14–16 non-transgenic or rTg4510 embryos and cultured in Neurobasal medium supplemented with 2% B-27 supplement with antioxidants, 0.5 mM L-glutamine, 100 U/mL penicillin, and 0.1 mg/mL streptomycin (all solutions from Invitrogen, Waltham, MA, USA) as described [35]. Embryos were genotyped using primer pairs as described [11]. Cortical neurons were treated with 1 µM cytosine arabinoside at days in vitro (DIV) 4 to halt proliferating cells and incubated with Thiamet G dissolved in phosphate buffered saline (PBS) at DIV 7. Culture medium was supplemented with 20 mM glucose before addition of Thiamet G.

For detection of cell viability following Thiamet G addition, the percentage of viable cells in the neuronal cultures was quantified by their capacity to reduce 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) following incubation with 0.5 mg/mL MTT for 60 min.

Cells were lysed in RIPA-buffer (50 mM Tris HCl, 1% NP-40/IGEPAL CA630, 0.25% Na-deoxycholate, 150 mM NaCl and 1 mM egtazic acid (EGTA)) supplemented with a 1% protease and phosphatase inhibitors mix (Roche, Basel, Switzerland) and 200 nM hexokinase inhibitor PUGNAc [O-(2-Acetamido-2-deoxy-D-glucopyranosylidenamino) N-phenylcarbamate] (Tocris, Bristol, UK). Following 30 min of lysis on ice, cell lysis was spun down (15,000 $\times$g, 15 min) and the protein concentration of the resulting supernatant measured. Protein measurements were performed using the bicinchoninic acid (BCA) Protein Assay Kit from Pierce as per manufacturer’s instructions (ThermoFisher Scientific, Waltham, MA, USA). For detection of total tau, phosphorylated tau, and O-GlcNAcylated proteins, 5 µg total protein was dissolved in sodium dodecyl sulfate (SDS)-sample buffer containing dithiothreitol (DTT, 100 mM) and used for western blotting.

The generation of rTg4510 mice was described previously [11]. In our colony, the tau transgene responder was expressed in the FVB/N mouse strain (Taconic, Silkeborg, Denmark), and the tTA activator system was maintained in the 129S6 mouse strain (Taconic, Silkeborg, Denmark). Their F1 progeny carried responder and activator transgenes in the rTg4510 mice along with non-transgenic (FVB/129) littermates. Only F1 mice were used in this study. All mice were bred in specific-pathogen-free conditions at Taconic (Silkeborg, Denmark) and genotyped from tail DNA using the primer pairs as described [11]. Mice were group-housed, where water and food (Brogaarden, Lynge, Denmark) were available ad libitum along with enrichment materials. The light/dark cycle was 12 h; room temperature was 21 $\pm$ 2 °C and a relative humidity of 55% $\pm$ 5%. Animal experiments were performed in accordance with the European Communities Council Directive no. 86/609, the directives of the Danish National Committee on Animal Research Ethics, and Danish legislation on experimental animals (license no. 2014-15-0201-00339).

For single dosing experiments, non-transgenic mice were dosed subcutaneously 75 mg/kg Thiamet G and rTg4510 and non-transgenic littermate mice were dosed per os 200 mg/kg Thiamet G or vehicle (10% HP-beta-CD, pH 4). Mice were euthanized 0.5, 1, 2, 4, 6, or 24 hours (h) following dosing and blood and brain tissue samples were collected. For subchronic and chronic dosing experiments, mice were dosed by drinking water for 1 week or 18 weeks, respectively, where 3.75 mg/mL Thiamet G was included in the drinking water. Water consumption was 5 mL/day per mouse and the average weight of the rTg4510 mice was 30 grams. Accordingly, a daily Thiamet G dose of approximately 600 mg/kg was given. Thiamet G was found to be stable in drinking water at room temperature for 1 week and the water supply was changed weekly. To suppress transgene expression, doxycycline (200 mg/kg) enriched chow (Bio-Serv, Flemington, NJ, USA) was administered to 6 week old rTg4510 mice in the control group ad libitum for 18 weeks during the study.

Mice were euthanized by cervical dislocation. Immediately following decapitation, blood was collected in ethylenediaminetetraacetic acid (EDTA)-coated tubes. Samples were kept on ice before being centrifuged at 20,000 $\times$g for 10 min at 4 °C to obtain plasma samples, which were then stored at –80 °C until use for exposure analysis. Brains were rapidly extracted and divided sagittally down the midline to yield two hemispheres, which were snap-frozen on dry ice and stored at –80 °C until use for biochemical or exposure analysis, respectively.

Forebrain homogenates were separated by differential centrifugation into Tris-buffered saline (TBS)-extractable (S1), S1 precipitate (S1p), and sarkosyl-insoluble pellet (P3) fractions as established previously [13] for tau biochemistry performed as described [36]. Briefly, forebrain homogenates were separated by centrifugation at 27,000 $\times$g for 20 min at 4 °C in S1. The pellet was re-homogenized in 5 volumes of high salt/sucrose buffer and centrifuged as above. The supernatant was collected and incubated with 1% sarkosyl for 1 h at 37 °C, followed by centrifugation at 150,000 $\times$g for 1 h at 4 °C to obtain P3. S1 was separated by further centrifugation at 150,000 $\times$g for 20 min at 4 °C to obtain S1p. All buffers were supplemented with 200 nM PUGNAc (Tocris, Bristol, UK).

Western blotting, analysis and quantitation of immunoreactivity was performed as described previously [36]. Total protein O-GlcNAcylation was detected using mouse monoclonal CTD110.6 [37] or RL2 (ThermoFisher Scientific, Waltham, MA, USA) antibody. Total tau was detected using mouse monoclonal Tau5 (Abcam, Cambridge, UK) and rabbit polyclonal tau E1 antibodies (raised against amino acids 19–33 (GLGDRKDQGGYTMHQ) of the longest isoform of human tau 2N4R) [38]. Tau phosphorylation was detected using rabbit polyclonal antibodies pS199/202 (Invitrogen, Waltham, MA, USA), pS400 (Invitrogen, Waltham, MA, USA), and mouse monoclonal pS396 antibody D1.2 [35]. Tau5 is a pan tau antibody and recognizes human and murine tau at non-phosphorylated and phosphorylated epitopes. The E1 antibody recognizes exclusively human tau at non-phosphorylated and phosphorylated epitopes, while pS199/202, pS396 and pS400 antibodies recognize the phosphorylated epitopes of murine and human tau. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used for normalization and detected with mouse monoclonal (Meridian Life Science, Memphis, TN, USA) or rabbit polyclonal (Abcam, Cambridge, UK) antibody.

Forebrain homogenates were analyzed using a Crosslink immunoprecipitation kit from Pierce (ThermoFisher Scientific, Waltham, MA, USA) in accordance with the manufacturer’s protocols. A set of 2 samples containing homogenates from a vehicle and a Thiamet G treated rTg4510 mouse were processed in parallel. The A/G agarose resin was crosslinked to 10 µg mouse CTD110.6 and 300 µg pre-cleared brain lysate was incubated with the resin-coupled O-GlcNAc antibody. Samples were eluted, collected by centrifugation, dissolved in SDS-sample buffer containing DTT (100 mM), and the immunoprecipitated O-GlcNAcylated proteins were probed with the E1 tau and O-GlcNAc RL2 antibodies by western blot.

Concentrations of Thiamet G were determined in plasma and brain homogenate using ultra performance liquid chromatography (UPLC) followed by tandem mass spectrometry (MS/MS) detection. Brain samples were homogenized by adding four times the volume of dimethyl sulfoxide (DMSO):2-propanol:H20 (20:30:50). The mobile phases A and B, which consisted of water/acentonitrile (50:50) containing 10 nM NH${}_4$Ac and 0.1% NH${}_4$OH, and water/acentonitrile (5:95) containing 10 nM NH${}_4$Ac and 0.1% NH${}_4$OH, were pumped as a gradient through an analytical column (Waters Acquity BEH Amide 1.7 µm, 2.1 $\times$ 50 mm, Waters, Milford, MA, USA). Detection was performed with a Sciex-API 4000 MS (AB Sciex, Framingham, MA, USA) using electro spray with positive ionization mode with a parent $>$ daughter transition of 249.2 $>$ 156.9 Da. The retention time was 0.73 min. The lower limit of quantification was 5.0 ng/mL in plasma and 2.5 ng/g in brain tissue samples (peak S/N $>$6). Data analysis was performed with Analyst 1.4.2 software (AB Sciex, Framingham, MA, USA).

The free fraction of Thiamet G in plasma and brain homogenates was measured by equilibrium dialysis and liquid chromatography (LC), followed by tandem mass spectrometry (MS/MS) analysis. Thiamet G dissolved in DMSO and diluted in 50% acetonitrile (ACN)/H${}_2$O was spiked in a final concentration of 2 µM to blank plasma or brain homogenates. Spiked plasma and brain homogenates were dialyzed against an equal volume of dialysis buffer for 5 h at 37 °C. In the post-dialysis plasma and brain homogenates, the total concentration (bound and unbound) of Thiamet G was measured. In the post-dialysis buffer, the unbound concentration of Thiamet G was measured. Based on these measurements the free fraction of Thiamet G was calculated.

A locomotor activity assay was performed in activity cages equipped with infrared light sources and photocells as described previously [34]. The frequency with which the mouse crossed the photo beams was used as a measure of activity in a 3-hour locomotor activity test. Hyperactivity was defined as total activity counts in the non-transgenic littermate population higher than the mean + 3 $\times$ SD. In this study $>$10,037 total activity counts were defined as hyperactivity. Data was collected using UMOTWin (Ellegaard systems, Aarup, Denmark).

All data were analyzed with GraphPad Prism version 9 (GraphPad Software, San Diego, CA, USA). Paired samples t tests using a 95% confidence interval and a two-tailed p value was used to analyze the immunoprecipitation (IP) data sets. All other data comparing two groups were analyzed using unpaired t tests, with a Welch’s correction factor, 95% confidence interval, and two-tailed p value. The difference in the number of hyperactive mice between the groups was analyzed with a chi-square test, with p values of 0.05 considered significant.

Thiamet G is reported as selective inhibitor of the eukaryotic OGA [26]. We confirmed that 10 µM Thiamet G did not show binding or activity on 98 other central and peripheral therapeutically relevant targets in a Cerep diversity profile screen (data not shown). We investigated consequences of OGA inhibition by Thiamet G in primary cortical neurons isolated from rTg4510 mouse embryos over-expressing human 0N4R P301L tau. Overexpression of the transgene tau did not lead to tau hyperphosphorylation, aggregation, or toxicity in rTg4510 primary neurons. Neuronal functions and viability were unchanged compared to cortical cultures isolated from non-transgenic littermates (data not shown). Neurons were treated with Thiamet G in concentrations ranging from 0.0001 to 100 µM for 6 and 24 h. No cell toxicity was observed in neurons following Thiamet G treatment, measured by a MTT reduction test (data not shown). A substantial increase in global O-GlcNAc levels of numerous proteins at the range from 30 kDa to 220 kDa or greater was detected 6 h after Thiamet G treatment by western blotting using the CTD110.6 antibody (Fig. 1A). We observed approximately 100% increased overall O-GlcNAc signal in neurons treated with 100 nM Thiamet G compared to buffer control (Fig. 1B). By densitometry quantification and normalization to GAPDH signal, the half-maximal effective concentration (EC${}_{50}$) Thiamet G was calculated to be 33 nM in rTg4510 neuronal cultures (Fig. 1B).

Fig. 1.

Concentration-dependent increase in protein O-GlcNAc levels in neurons from rTg4510 mice treated with Thiamet G. Cortical neurons isolated from rTg4510 mouse embryos were incubated at days in vitro (DIV) 7 with Thiamet G at concentrations as indicated. After 6 hours, neurons were lysed and analyzed by western blot for global protein O-GlcNAcylation (O-GlcNAc), human tau (E1), total tau (Tau5), and phosphorylated tau at S199/202 (pS199/202) and S396 (pS396) epitopes. For normalization, the O-GlcNAC blot was re-probed with GAPDH. The E1 blot was re-probed with pS396 and GAPDH, and the pS199/202 blot was re-probed with Tau5 and GAPDH. Phosphorylated and non-phosphorylated human 0N4R transgene tau was displayed at 55 kDa with the E1 antibody. Murine and human 0N4R transgene tau species are displayed at 45 and 55 kDa, respectively, with the Tau5 antibody. (A) Representative western blots from n = 3 experiments. (B) Quantification of western blots is shown as graphs in absolute units O-GlcNAc/GAPDH and in % tau/GAPDH with buffer control (0) set to 100%. Individual samples are presented as filled shapes. Thiamet G EC${}_{50}$ based on densitometric data from total O-GlcNAcylation (O-GlcNAc) western blot presented in A was calculated to 33 nM. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Next, the impact of OGA inhibition by Thiamet G on the phosphorylation of mouse and human transgene tau was assessed in rTg4510 neuronal cultures. Normally phosphorylated human 0N4R P301L tau runs as 55 kDa band on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Hyperphosphorylated 0N4R P301L tau mobility is shifted and runs as 64 kDa band on SDS-PAGE. We detected transgene human 0N4R P301L tau displayed at 55 kDa with the E1 antibody specific for human tau by western blotting (Fig. 1A). Additionally, neurons expressed endogenous murine tau isoforms which were displayed at 45 and 55 kDa on SDS-PAGE and detected together with the human transgene tau with the phosphorylation-independent Tau5 antibody. Human and murine tau were phosphorylated at S199/202 and S396 epitopes as detected with the pS199/202 and pS396 tau antibodies by western blotting. When treated with Thiamet G, we observed a marked increase in global O-GlcNAc levels in the 0.1–100 µM concentration range (Fig. 1A), and no changes in the normal total human and murine tau levels detected with the E1 and Tau5 antibodies (Fig. 1B). No changes of tau phosphorylation at the S199/202 and S396 epitopes of murine and human tau were observed. Using densitometry quantification and normalization to GAPDH signaling, we confirmed that Thiamet G treatment did not lead to significant changes in phosphorylation at S199/202 and S396 epitopes of human or mouse tau in rTg4510 primary neurons (Fig. 1B).

Initially, we tested whether Thiamet G can cross the blood brain barrier after a single subcutaneous dose of 75 mg/kg and calculated that the half-life of Thiamet G was approximately 1.5 and 3.5 h in plasma and brain, respectively (Table 1). Next, we investigated oral administration and dosed 24 week old rTg4510 and non-transgenic littermate mice with a single dose of 200 mg/kg acutely for 6 or 24 h or subchronically for 1 week with 600 mg/kg/day Thiamet G in drinking water. At this age rTg4510 mice exhibited prominent tau pathology including NFTs [34, 36] and hyperphosphorylated tau (Supplementary Fig. 1). We analyzed Thiamet G concentrations in plasma and brain tissue (Table 2), and total O-GlcNAc levels in brain tissue (Fig. 2). Non-transgenic littermate (data not shown) and rTg4510 (Fig. 2) mice displayed similar Thiamet G concentrations in plasma and brain tissue, and a comparable increase in global O-GlcNAc levels. After 6 h, 24 h, and 1 week of dosing we measured 0.5, 0.2, and 1.1 µM Thiamet G, respectively, in brain samples from rTg4510 mice (Table 2). The free fraction of Thiamet G in the brain was determined to be 67%, leading to approximately 360, 120, and 720 nM unbound Thiamet G after acute 6 h, 24 h, and 1 week dosing in the drinking water, respectively (Table 2). These concentrations were approximately 11-, 4-, and 22-fold higher than the measured EC${}_{50}$ in cortical neurons, indicating that the concentration of Thiamet G in the brain was sufficient to inhibit OGA. Accordingly, we confirmed an elevation in O-GlcNAc of numerous proteins at the range from 30 kDa to 220 kDa or greater, as determined by western blotting for the three oral dosing schemes (Fig. 2A). Quantification of O-GlcNAc positive bands indicated a 1.9, 1.6, and 3.5-fold increase in global O-GlcNAc levels 6 h, 24 h, and 1 week after Thiamet G administration, respectively (Fig. 2B). To investigate whether Thiamet G brain exposure changed over age and the progression of tau pathology in the rTg4510 mice, we dosed 6 week old rTg4510 and non-transgenic littermate mice subchronically for 1 week with 600 mg/kg/day Thiamet G in the drinking water. At this age, rTg4510 mice did not express tau pathology [34, 36] and hyperphosphorylated tau species were not detectable (Supplementary Fig. 1). The 6 week old non-transgenic littermates and rTg4510 mice displayed similar Thiamet G concentrations in plasma and brain tissue (data not shown). We also observed no difference in brain exposure in 6 or 24 week old rTg4510 mice treated subchronically for 1 week with Thiamet G.

Fig. 2.

Acute and subchronic Thiamet G dosing increases protein O-GlcNAc levels. Soluble fraction of brain homogenates from 24 week old rTg4510 mice treated either with vehicle (n = 5) or Thiamet G (n = 5) acutely per os (200 mg/kg) for 6 or 24 hours, or subchronically for 1 week (600 mg/kg/day), were analyzed by western blot for global protein O-GlcNAcylation (O-GlcNAc). For normalization, blots were re-probed for GAPDH. (A) Representative western blots from n = 3 experiments. (B) Quantification of western blots is shown as graphs in absolute units O-GlcNAc/GAPDH, individual samples are presented as filled circles and in lines as means $\pm$ SD (Unpaired t test with Welch’s correction, ***p 0.001). SD, Standard deviation.

Next, we investigated whether tau phosphorylation and hyperphosphorylation were modulated by OGA inhibition with Thiamet G under the acute treatment regimen in aged rTg4510 mice. As previously characterized by others [11, 32] and us [34, 36], rTg4510 mice develop an age-dependent neuropathology displaying pre-tangle tau and NFTs corresponding to oligomeric and fibrillar hyperphosphorylated tau [13] at 24 weeks of age (Supplementary Fig. 1).

Tau pathology in rTg4510 mice, which overexpress human 0N4R P301L tau, can be analyzed biochemically by differential centrifugation and sarkosyl fractionation of brain homogenates (Supplementary Fig. 2). On SDS-PAGE, normally phosphorylated human 0N4R P301L tau is displayed as a 55 kDa band. Hyperphosphorylated 0N4R P301L tau’s mobility is shifted and displayed as 64 and 70 kDa bands. The majority of hyperphosphorylated tau species run at 64 kDa (Supplementary Fig. 1). Hyperphosphorylated 64 kDa and 70 kDa tau species isolated in the sarkosyl-insoluble pellet fraction (P3) consist of fibrillary tau aggregates, as characterized by immunoelectron microscopy [13], and are regarded as the biochemical equivalent of NFTs [11, 32]. Hyperphosphorylated 64 kDa tau species in the soluble fraction (S1) precipitate under high speed centrifugation into the precipitate fraction of S1 (S1p). These low speed soluble and high speed precipitable hyperphosphorylated 64 kDa tau species consist of tau oligomers, as characterized by immunoelectron microscopy, and are regarded as the biochemical equivalent of pre-tangle tau [13].

We isolated forebrain homogenates into soluble (S1) and sarkosyl-insoluble pellet (P3) fractions by differential centrifugation (Supplementary Fig. 2) as previously characterized [13]. To investigate the different tau species, we used the human tau specific E1 antibody and phospho tau antibodies that detect phosphorylation at the S199/202 and S396 epitopes. Normal human 0N4R P301L tau was displayed as 55 kDa in S1 detected with the E1 and phospho tau pS199/202 and pS396 antibodies (Fig. 3A,C). Hyperphosphorylated tau species were displayed as mobility shifted tau of 64 kDa in S1, and 64 kDa and 70 kDa in P3 (Fig. 3A–D). Levels of soluble 55 kDa human tau were not changed in brain samples 6 and 24 h after acute treatment with Thiamet G (S1, Fig. 3A,C). Levels of soluble hyperphosphorylated 64 kDa tau in S1 were not altered after acute Thiamet G treatment. Additionally, neither insoluble hyperphosphorylated 64 kDa or 70 kDa tau species in P3 were altered (Fig. 3A–D).

Fig. 3.

Acute and subchronic Thiamet G dosing and effects on hyperphosphorylated tau species. Brain homogenates from 24 week old rTg4510 mice treated either with vehicle (n = 5) or Thiamet G (n = 5) acutely per os (200 mg/kg) for 6 and 24 hours or for 1 week subchronically in the drinking water (600 mg/kg/day) were fractioned into soluble (S1) and sarkosyl-insoluble (P3) fractions and analyzed by western blot for total human tau (E1) and phosphorylated tau at S199/202 (pS199/202) and S396 (pS396) epitopes. For normalization of S1 blots GAPDH was used. The E1 blot was re-probed with pS396 and GAPDH and the pS199/202 blot was re-probed with GAPDH. The P3 E1 blot was re-probed with pS396. Phosphorylated and non-phosphorylated human 0N4R tau transgene is displayed at 55 kDa in S1 fraction, hyperphosphorylated tau is mobility-shifted and displayed at 64 and 70 kDa in S1 and P3 fractions. (A,C,E) Representative western blots from n = 3 experiments. (B,D,F) Quantification of western blots is shown as graphs in S1 % tau/GAPDH and P3 % tau/S1 tau with vehicle set to 100%. Individual samples are presented as filled shapes and in lines as means $\pm$ SD.

Soluble 55 kDa tau was not reduced by subchronic Thiamet G treatment for 1 week and phosphorylation levels remained unchanged (S1, Fig. 3E). On the other hand, subchronic Thiamet G treatment led to a 20–30% decrease in soluble 64 kDa and insoluble 64 and 70 kDa hyperphosphorylated tau species (Fig. 3E,F), although this difference did not reach statistical significance. From these results, we concluded that Thiamet-G in acute and subchronic oral doses reaches sufficient brain concentrations, increases the total protein O-GlcNAc levels in brains, but has no effect on phosphorylation levels of soluble 55 kDa human 0N4R P301L tau. In subchronic 1 week oral dosing, Thiamet G showed a trend to reduce soluble and insoluble hyperphosphorylated tau species in 24 week old rTg4510 mice.

Based on the trend observed in aged rTg4510 mice, we initiated chronic Thiamet G treatment in young rTg4510 mice to investigate the long-term effect of OGA inhibition on O-GlcNAc levels and accumulation of tau pathology over time. 6 week old rTg4510 mice, which did not present with hyperphosphorylated tau (Supplementary Fig. 1) or tau pathology as we reported previously [34, 36], were continually treated for 18 weeks with 600 mg/kg/day Thiamet G in the drinking water. Biochemical analysis was performed on brain homogenates isolated from 24 week old rTg4510 and non-transgenic littermate mice, where we found similar brain concentrations. We measured 1.1 µM Thiamet G in brain samples from rTg4510 mice, which were similar to the brain sample concentrations of Thiamet G found in rTg4510 mice treated subchronically for 1 week (Table 2). The unbound Thiamet G concentration in the brain amounted to 740 nM, which was 22-fold higher than the EC${}_{50}$ measured in cortical neurons (Table 2). These results indicate that brain concentrations were stable during the chronic treatment regimen and sufficiently high for OGA inhibition. Accordingly, a marked increase in global O-GlcNAc levels was observed in brain homogenates from rTg4510 mice treated with Thiamet G compared to vehicle (n = 7 for each treatment group; Fig. 4A). Quantification of O-GlcNAc positive bands indicated a significant 3.5-fold increase in global O-GlcNAc levels after 18 weeks of Thiamet G administration (Fig. 4B). We concluded that OGA was sufficiently inhibited and global O-GlcNAc levels constantly increased in rTg4510 mice treated for 18 weeks with Thiamet G.

Fig. 4.

Chronic Thiamet G treatment increases protein O-GlcNAcylation in aged rTg4510 mice. Soluble fraction of brain homogenates from 24 week old rTg4510 mice treated either with vehicle (V, n = 7) or Thiamet G (T, n = 7) in the drinking water (600 mg/kg/day) from 6 weeks of age were analyzed by western blot for global protein O-GlcNAcylation (O-GlcNAc). For normalization, the blot was re-probed for GAPDH. (A) Representative western blot from n = 3 experiments. (B) Quantification of western blot is shown as a graph in absolute units O-GlcNAc/GAPDH. Individual samples are presented as filled circles and in lines as means $\pm$ SD (Unpaired t test with Welch’s correction, ***p 0.001).

Next, we investigated whether chronic Thiamet G treatment for 18 weeks in rTg4510 mice could affect O-GlcNAcylation of human P301L tau. We performed immunoprecipitation (IP) experiments with forebrain homogenates of rTg4510 mice treated with vehicle or Thiamet G, and immunoprecipitated O-GlcNAc modified proteins with the CTD110.6 antibody, which selectively recognizes O-GlcNAcylated proteins (in pairs, n = 4 for each treatment group; Fig. 5A). As expected, we detected a significant 3-fold increase in protein O-GlcNAcylation in the immunoprecipitates from rTg4510 mice treated with Thiamet G compared to vehicle (Fig. 5B). We then probed the O-GlcNAc immunoprecipitates with the human tau specific E1 antibody and detected mostly 55 kDa tau. 64 kDa tau bands were almost undetectable in the O-GlcNAc immunoprecipitates. We repeatedly observed considerable variations in the 55 kDa tau levels between the individual immunoprecipitates (Fig. 5A). However, we detected a mild but significant 1.5-fold increase of 55 kDa tau species in O-GlcNAc immunoprecipitates from rTg4510 mice treated with Thiamet G (Fig. 5B), suggesting that chronic Thiamet C treatment for 18 weeks could increase tau O-GlcNAcylation.

Fig. 5.

Chronic Thiamet G treatment increases tau O-GlcNAcylation in aged rTg4510 mice. Pairs of brain homogenate from 24 week old rTg4510 mice treated either with vehicle (V) or Thiamet G (T) in the drinking water (600 mg/kg/day) from 6 weeks of age were prepared individually for immunoprecipitation of O-GlcNAcylated proteins with the CTD110.6 antibody. Immunoprecipitates (n = 4 pairs) were analyzed by western blot for O-GlcNAc and human tau with the E1 antibody. (A) Representative western blots from n = 3 experiments. (B) Quantification of western blots is shown as graphs in absolute units O-GlcNAc and human 55 kDa tau. Individual samples are presented as filled circles and in lines as means $\pm$ SD (Paired t test *p 0.02 O-GlcNAc, *p 0.03 tau).

Next, we investigated whether chronic Thiamet G treatment prevented accumulation of soluble and insoluble hyperphosphorylated tau in 24 week old rTg4510 mice. S1 and P3 were isolated from forebrain homogenates and different human tau species were measured using E1 and phospho tau pS199/202, pS396, and pS400 antibodies, respectively. At 24 weeks of age, vehicle treated rTg4510 mice displayed considerable levels of soluble hyperphosphorylated 64 kDa tau in S1 (Fig. 6A,B) as well as insoluble hyperphosphorylated 64 and 70 kDa tau species in P3 (Fig. 6C,D), in agreement with our biochemical characterization (Supplementary Fig. 1) reported previously [34, 36]. In Thiamet G treated rTg4510 mice we observed a significant reduction of soluble hyperphosphorylated 64 kDa tau by 45% in S1 based on detection with the phospho tau antibodies pS199/202, pS396, and pS400 (Fig. 6B). Insoluble hyperphosphorylated 64 and 70 kDa tau in P3 was also significantly reduced by 42% (Fig. 6D). The highest reduction was observed with the pS400 tau antibody showing 52% and 49% reduction of hyperphosphorylated tau species in S1 and P3, respectively. We did not observe any changes in the total levels or phosphorylation of soluble 55 kDa tau in S1 (Fig. 6A). These results suggest that long-term OGA inhibition by Thiamet G reduced accumulation of soluble and insoluble hyperphosphorylated tau when treatment was initiated before rTg4510 mice exhibited tau pathology.

Fig. 6.

Chronic Thiamet G treatment reduces accumulation of hyperphosphorylated tau in aged rTg4510 mice. Brain homogenates from 24 week old rTg4510 mice treated either with vehicle (V, n = 7) or Thiamet G (T, n = 7) in the drinking water (600 mg/kg/day) from 6 weeks of age were fractioned into soluble (S1) and sarkosyl-insoluble (P3) fractions and analyzed by western blot for total human tau (E1) and phosphorylated tau at S199/202, S396, and S400 epitopes. For normalization of S1 blots GAPDH was used. The E1 blot was re-probed with pS396 and GAPDH. pS199/202 and pS400 blots were re-probed with GAPDH. The P3 E1 blot was re-probed with pS396. Phosphorylated and non-phosphorylated human 0N4R tau was displayed at 55 kDa in S1 fractions. Hyperphosphorylated tau was mobility-shifted and displayed at 64 and 70 kDa in S1 and P3 fractions. (A,C) Representative western blots from n = 3 experiments. (B,D) Quantification of western blots is shown in graphs in S1 64 kDa tau/GAPDH in % and P3 64 and 70 kDa tau/S1 tau in %. Individual samples are presented as filled shapes and in lines as means $\pm$ SD (Unpaired t test with Welch’s correction, ***p 0.001, **p 0.005).

rTg4510 mice develop a non-mnemonic behavioral defect, characterized by a hyperactive phenotype that increases with the progression of tau pathology [33, 34]. We reported previously that between 46–59% of the rTg4510 mice in our colony were hyperactive when they reach 24 weeks of age. Suppression of transgene tau expression with doxycycline from 6 weeks of age reduced the number of rTg4510 mice displaying the hyperactive phenotype at 24 weeks of age, suggesting that hyperactivity is dependent on tau pathology [34]. Here, we investigated whether long-term treatment with Thiamet G influence hyperactivity in the rTg4510 mice. 56 non-transgenic littermates (n = 36 treated with vehicle and n = 20 treated with Thiamet G) and 72 rTg4510 mice (n = 36 treated with vehicle and n = 36 treated with Thiamet G) were dosed from 6 weeks of age with 600 mg/kg/day Thiamet G in the drinking water and tested for increased locomotor activity at 24 weeks of age. Non-transgenic littermates displayed normal locomotor activity between 963 and 8124 total activity counts with an average of 3848 $\pm$ 2063 total activity counts (Fig. 7A). Total activity counts higher than the mean + 3 $\times$ SD of the non-transgenic population, which was $>$10,037 total activity counts, were defined as hyperactivity. 50% of the rTg4510 mice in the vehicle treated group (n = 18) displayed an abnormal increase in locomotor activity, with an average of 45,014 $\pm$ 20,157 total activity counts (Fig. 7A). Chronic Thiamet G dosing had beneficial effects on hyperactivity in 24 week old rTg4510 mice. In the Thiamet G treatment group we found that the average total activity count for rTg4510 mice decreased to 7837 $\pm$ 9419 compared to 24,625 $\pm$ 25,025 in the vehicle treated group (Fig. 7A). The number of hyperactive rTg4510 mice decreased significantly to 19% (n = 7, Fig. 7B) with an average of 22,863 $\pm$ 13,100 total activity counts in the remaining hyperactive mice compared to 45,014 $\pm$ 20,157 in the vehicle treated group. As anticipated, we did not observe any effect of Thiamet G treatment on activity levels in the non-transgenic littermates (Fig. 7A). As a control, doxycycline administration from 6 weeks of age reduced the number of rTg4510 mice displaying the hyperactive phenotype to 10% at 24 weeks of age (data not shown) in agreement with our previous findings [34]. These results indicate that OGA inhibition by chronic Thiamet G administration initiated in young rTg4510 mice significantly reduced hyperactivity later in aged rTg4510 mice.

Fig. 7.

Chronic Thiamet G reduces hyperactivity in aged rTg4510 mice. Non-transgenic littermates (non-tg, n = 56) and rTg4510 mice (n = 72) treated with vehicle (non-tg, n = 36; rTg4510, n = 36) or Thiamet G (non-tg, n = 20; rTg4510, n = 36) in the drinking water (600 mg/kg/day) from 6 weeks of age were tested in a 3-hour locomotor activity test in a novel cage environment at 24 weeks of age. (A) Data are presented as total activity counts. Hyperactivity was defined as $>$10,037 total activity counts. Individual mice are presented as filled shapes and in lines as means $\pm$ SD. (B) Data are presented as number of hyperactive and non-hyperactive rTg4510 mice. Chi-square test = 7.001, **p = 0.0081.

Taken together, these in vivo results show that Thiamet G can cross the blood brain barrier in rTg4510 mice and non-transgenic littermates following oral dosing. The free fraction of Thiamet G in the brain of rTg4510 mice after chronic dosing was 740 nM, which is 22-fold above the EC${}_{50}$ measured in rTg4510 primary neurons and considered sufficient for full target engagement. Western blots revealed that chronic Thiamet G treatment over 18 weeks increased overall brain levels of O-GlcNAc and reduced significantly soluble and insoluble hyperphosphorylated tau species in 24 week old rTg4510 mice. Furthermore, a significant reduction in the number of hyperactive rTg4510 mice was observed, demonstrating a positive effect for chronic OGA inhibition in this mouse tauopathy model.