The microtubule (MT)–stabilizing protein tau disengages from MTs and forms intracellular inclusions known as neurofibrillary tangles in Alzheimer’s disease and related tauopathies. Reduced tau binding to MTs in tauopathies may contribute to neuronal dysfunction through decreased MT stabilization and disrupted axonal transport. Thus, the introduction of brain-penetrant MT-stabilizing compounds might normalize MT dynamics and axonal deficits in these disorders. We previously described a number of phenylpyrimidines and triazolopyrimidines (TPDs) that induce tubulin post-translational modifications indicative of MT stabilization. We now further characterize the biologic properties of these small molecules, and our results reveal that these compounds can be divided into two general classes based on the cellular response they evoke. One group composed of the phenylpyrimidines and several TPD examples showed a bell-shaped concentration-response effect on markers of MT stabilization in cellular assays. Moreover, these compounds induced proteasome-dependent degradation of α- and β-tubulin and caused altered MT morphology in both dividing cells and neuron cultures. In contrast, a second group comprising a subset of TPD molecules (TPD+) increased markers of stable MTs in a concentration-dependent manner in dividing cells and in neurons without affecting total tubulin levels or disrupting MT architecture. Moreover, an example TPD+ compound was shown to increase MTs in a neuron culture model with induced tau hyperphosphorylation and associated MT deficits. Several TPD+ compounds were shown to be both brain penetrant and orally bioavailable, and a TPD+ example increased MT stabilization in the mouse brain, making these compounds potential candidate therapeutics for neurodegenerative tauopathies such as Alzheimer’s disease.
Alzheimer’s disease (AD) and related tauopathies are characterized by hyperphosphorylation and aggregation of the microtubule (MT)–associated protein, tau (Lee et al., 2001b; Ballatore et al., 2007). Neuronal tau is normally thought to stabilize MTs and regulate axonal transport, whereas hyperphosphorylated tau shows reduced interaction with MTs (Alonso et al., 1994; Merrick et al., 1997), facilitating the formation of insoluble tau aggregates termed neurofibrillary tangles and neuropil threads within the neuronal soma and dendritic processes, respectively (Wood et al., 1986; Lee et al., 2001b). Transgenic mouse models of tauopathy display fewer MTs, increased MT hyperdynamicity, axonal dystrophy, and reduced fast axonal transport (FAT) (Zhang et al., 2005, 2012; Barten et al., 2012). Similarly, there is evidence of MT deficits in the AD brain (Hempen and Brion, 1996; Cash et al., 2003). Ultimately, MT abnormalities likely contribute to neuronal dysfunction, neuron loss, and cognitive decline, and introduction of MT-stabilizing compounds might attenuate these deficits in AD and related tauopathies (Ballatore et al., 2012; Brunden et al., 2012).
Studies performed in mouse tauopathy models support the hypothesis that increasing MT stability imparts a beneficial effect on neuronal function and disease course. For instance, weekly intraperitoneal injection of the MT-stabilizing compound paclitaxel improved FAT and motor behavior in addition to alleviating spinal cord tau pathology in T44 tau transgenic mice (Zhang et al., 2005). Although paclitaxel is unable to cross the blood–brain barrier (BBB), sufficient levels of drug likely reached the spinal cord through uptake at neuromuscular junctions (Zhang et al., 2005). More recently, our laboratory demonstrated that administration of the brain-penetrant MT-stabilizing compound, epothilone D (EpoD), to PS19 transgenic mice, which develop tau inclusions within the forebrain (Yoshiyama et al., 2007), rescued neuronal loss and attenuated deficits in axonal architecture, FAT, and cognition, while decreasing tau neuropathology (Brunden et al., 2010; Zhang et al., 2012). The salutary effects of EpoD were further confirmed in two additional tau transgenic mouse models (Barten et al., 2012). Although EpoD has since progressed to clinical trials in patients with AD and represents a promising drug candidate, it is the only small molecule MT-stabilizing agent reported to date to show beneficial effects in transgenic mouse models with brain tauopathy. Thus, there is a need to identify alternative brain-penetrant MT-stabilizing agents, ideally those that could be readily synthesized and orally administered.
Cevipabulin [TTI-237 (5-chloro-6-[2,6-difluoro-4-[3-(methylamino)propoxy]phenyl]-N-[(2S)-1,1,1-trifluoropropan-2-yl]-[1,2,4]triazolo[1,5-a]pyrimidin-7-amine) CNDR-51533], a non-naturally occurring triazolopyrimidine (TPD), has been reported to possess MT-stabilizing activity and efficacy in murine tumor xenograft models (Zhang et al., 2007; Beyer et al., 2008). However, cevipabulin does not cross the BBB and thus would not have therapeutic potential in neurodegenerative disease. We previously reported on the activity of a number of brain-penetrant TPD and phenylpyrimidine (PPD) congeners that possess oral bioavailability and, unlike EpoD, do not inhibit the P-glycoprotein (P-gp) transporter that protects the brain from exposure to xenobiotics and many drugs (Lou et al., 2014; Cornec et al., 2015). In many cases, these compounds demonstrated activity in a QBI-293 cellular assay that measures acetylated tubulin, a post-translational modification of the α-tubulin subunit that is thought to occur only on polymerized MTs and is a commonly used marker of stable MTs (Schulze et al., 1987; Brunden et al., 2011; Lou et al., 2014).
Based on potency in the acetyl-tubulin assay and favorable pharmacokinetic properties, the previously disclosed TPD “9” (CNDR-51555; (S)-5-chloro-6-(4-(3-(dimethylamino)propoxy)-2,6-difluorophenyl)-N-(1,1,1-trifluoropropan-2-yl)-[1,2,4]triazolo[1,5-a]pyrimidin-7-amine) and the PPD “20” (CNDR-51549; (S)-6-chloro-2-(pyrazin-2-yl)-5-(2,4,6-trifluorophenyl)-N-(1,1,1-trifluoropropan-2-yl)pyrimidin-4-amine) (Lou et al., 2014) were chosen for further characterization of biologic actions, including evaluation of MT-stabilizing activity in primary neuronal cultures. Notably, these studies revealed several unexpected features in the biologic activities of these compounds. In particular, the lead PPD and TPD prototypes, as well as many related congeners, revealed unusual bell-shaped or inverse U concentration-dependent changes in acetyl-tubulin levels and caused alterations in MT structure. In addition, these compounds induced degradation of α- and β-tubulin, even at doses that increased acetyl-tubulin levels, and proved ineffective in increasing markers of stable MTs in primary neurons. However, select TPD analogs were identified that retained MT-stabilizing activity without negatively affecting tubulin levels or cytoskeletal architecture. This particular subset of TPDs, hereafter referred to as TPD+ compounds, also promote MT stability in primary neurons and in the cortex of wild-type mice. Furthermore, we demonstrate that TPD+ compounds can rescue MT deficits in neurons treated with okadaic acid (OA), an in vitro model of tau hyperphosphorylation with associated loss of MT structure. Based on their MT-stabilizing activity in cortical neurons and their ability to enter the brain, these TPD+ compounds represent potential therapeutic candidates for the treatment of neurodegenerative tauopathies.
Materials and Methods
EpoD was prepared as previously described (Lee et al., 2001a; Rivkin et al., 2004). The spectroscopic properties of the compound were identical to those reported in the literature. Cevipabulin (CNDR-51533) was synthesized as described (Zhang et al., 2007). All other TPD and PPDs tested were synthesized following published procedures (Zhang et al., 2007, 2009; Lou et al., 2014) (Supplemental Methods). In each case, compound purity was > 95% as determined by liquid chromatography (LC)–mass spectrometry (MS) and nuclear magnetic resonance analyses.
Determination of Plasma and Brain Drug Concentrations.
All animal protocols were approved by the University of Pennsylvania Institutional Animal Care and Use Committee. Test compounds were administered to CD-1 or B6SJL mice. Both female and male mice were used but were not mixed within experimental groups. The average group age for each study ranged from 2.0 to 5.7 months. For standard single time point brain and plasma determinations, mice were injected intraperitoneally with a single dose of 5 mg/kg compound dissolved in dimethylsulfoxide (DMSO). For oral bioavailability studies, 10 mg/kg compound dissolved in PEG400/dimethylacetamide/water (1:1:1) with 0.5% methyl cellulose was administered via oral gavage. One hour after compound administration, mice were euthanized and perfused following an Institutional Animal Care and Use Committee–approved protocol. Whole brain hemispheres were homogenized in 10 mM ammonium acetate, pH 5.7 [50% (w/v)], using a hand-held sonic homogenizer. Plasma was obtained from blood collected in 0.5 M EDTA solution and centrifuged for 10 minutes at 4500g at 4°C. Fifty-microliter aliquots of brain homogenate or plasma were mixed with 0.2 ml acetonitrile and centrifuged at 15,000g, and the resultant supernatants were subjected to LC–tandem mass spectrometry (MS/MS) analysis. The LC-MS/MS system includes an Acquity UPLC instrument and a TQ MS instrument controlled using MassLynx software (Waters Corporation, Milford, MA). Compound detection was performed using multiple reaction monitoring of their specific collision-induced ion transitions. Five microliters of each sample was separated on an Acquity BEH C18 column (1.7 µm, 2.1 mm × 50 mm) at 35°C. Operation was in positive electrospray ionization mode, with mobile phase A of 0.1% (v/v) formic acid in water and mobile phase B of acetonitrile with 0.1% (v/v) formic acid at a flow rate of 0.6 ml/min using a gradient from 5% to 95% B over 2 minutes, followed by wash and re-equilibration steps. The MS instrument was operated with a desolvation temperature of 450°C and a source temperature of 150°C. Desolvation and source nitrogen gas flows were 900 and 50 l/h, respectively. Source and MS/MS voltages were optimized for each compound using the MassLynx AutoTune utility. Standard curves were generated for each compound from brain homogenate and plasma samples that had compound added at concentrations ranging from 1 to 1000 nM and extracted as above.
Measurement of Acetyl-Tubulin in Cortical Tissue.
Wild-type CD-1 female mice (aged 2 to 3 months) received two injections of 1 or 5 mg/kg CNDR-51657 ((R)-5-chloro-N-(3-methylbutan-2-yl)-6-(2,4,6-trifluorophenyl)-[1,2,4]triazolo[1,5-a]pyrimidin-7-amine) or CNDR-51555 spaced approximately 18 hours apart. Four hours after the second injection, mice were euthanized by an approved protocol and cortices were dissected from each brain and placed immediately in ice-cold radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris, 150 mM NaCl, 5 mM EDTA, 0.5% sodium deoxycholate, 1% NP-40, and 0.1% SDS, pH 8.0) containing protease inhibitor (PI) cocktail (Sigma-Aldrich, St. Louis, MO), 1 mM phenylmethylsulfonyl fluoride (PMSF) (Sigma-Aldrich), and 3 µM trichostatin A (TSA) (Sigma-Aldrich). Tissue was homogenized with a hand-held battery-operated pestle motor mixer and then sonicated to complete the lysis. Samples were centrifuged at 100,000g for 30 minutes at 4°C and supernatant was transferred to a new Eppendorf tube. Remaining pellets were resuspended in RIPA buffer and homogenized, sonicated, and centrifuged again, as before. Supernatant from the second centrifugation step was pooled with that from first spin. Samples were assessed for protein concentration by the bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, Waltham, MA) and enzyme-linked immunosorbent assay (ELISA) analysis of acetyl- and α-tubulin levels was performed, as described below.
Cell-Free MT Assembly Assay.
Compounds CNDR-51555 or CNDR-51549 were spotted onto a 384-well clear plate at concentrations derived from serial dilutions starting with 60 µM to 0.475 µM. A 0.5-μl aliquot of each compound was added at 100× the desired final concentration. Ice-cold porcine brain tubulin (30 µM) (Cytoskeleton Inc., Denver, CO) in PEM buffer (80 mM 1,4-piperazinediethanesulfonic acid, pH 6.8, 5 mM EGTA, and 1 mM MgCl2) containing 2 mM GTP was quickly added to each well containing compound. The assembly assay was performed in a SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA) at 37°C with absorbance of 350 nm with reads at 1-minute intervals for 45 minutes. Each assembly reaction was performed in triplicate. At the final time point, material was recovered from wells and samples were centrifuged at 100,000g for 30 minutes at 37°C to sediment assembled tubulin. The supernatants were recovered (supernatant fractions) and pellets (pelletable fractions) were dissolved in PEM buffer in an equal volume as the supernatant fraction. For electrophoresis, 5× Laemmli buffer was added to each supernatant fraction and pelletable fraction sample to achieve 1× Laemmli concentration, and samples were boiled at 100°C for 5 minutes. Two microliters of each sample was loaded per lane on 10% polyacrylamide gels. Gels were stained with Coomassie blue for quantification of protein in the supernatant and pelletable fractions.
QBI-293, ADR-RES, and COS-7 Cell Culture.
QBI-293, a derivative of human embryonic kidney 293 cells, and COS-7 cells (American Type Culture Collection, Manassas, VA) were maintained in Dulbecco’s modified Eagle’s medium (Mediatech Inc., Manassas, VA) containing 10% fetal bovine serum (FBS) (Atlanta Biologicals, Lawrenceville, GA), 2 mM l-glutamine (Mediatech), 50 U/ml penicillin, and 50 µg/ml streptomycin (1% penicillin/streptomycin; Thermo Fisher Scientific). Adriamycin-resistant (ADR)-RES cells (National Cancer Institute, Frederick, MD) were maintained in RPMI medium (Mediatech) containing 10% FBS, 2 mM l-glutamine, and 1% penicillin/streptomycin. For compound testing, QBI-293 cells were dissociated with trypsin/EDTA (Thermo Fisher Scientific) and plated at a density of 6 × 105 cells/well in six-well plates. After overnight incubation, the medium was aspirated and fresh medium containing vehicle or test compound was added. For immunocytochemistry experiments, sterilized glass coverslips coated with 0.1 mg/ml poly(d-lysine) (PDL) (Sigma-Aldrich) and 0.1% swine gelatin were placed into 24-well plates. COS-7 cells were plated at a density of 7.5 × 104 cells/well onto the coverslips and incubated overnight prior to compound treatment. Cells were maintained at 37°C in a humidified atmosphere (5% CO2) for all experiments.
Primary Cortical Neuron Cell Culture.
Dissociated primary cortical neurons from embryonic day 18 Sprague-Dawley rats were obtained through the University of Pennsylvania tissue culture service. Neurons were plated onto 0.1 mg/ml PDL-coated six-well plates at a density of 7.5 × 105 cells/well in neurobasal medium (Thermo Fisher Scientific) supplemented with 2% B27 (Thermo Fisher Scientific) and containing 0.5% penicillin/streptomycin. For immunocytochemistry experiments, neurons were plated onto PDL-coated coverslips in 24-well plates at a density of 8.0 × 104 cells/well. Compound treatments were performed at 3, 7, or 10 days in vitro (DIV). For drug treatments, one-half of the medium was removed from each well and replaced with an equal volume of fresh neuronal medium containing compound (at 2× concentration) to avoid complete medium changes. Neurons were maintained at 37°C in a humidified atmosphere (5% CO2) for all experiments.
In Vitro Compound Treatment.
For experiments requiring proteasomal inhibition, cells were pretreated with 10 µM MG-132 (benzyl N-[(2S)-4-methyl-1-[[(2S)-4-methyl-1-[[(2S)-4-methyl-1-oxopentan-2-yl]amino]-1-oxopentan-2-yl]amino]-1-oxopentan-2-yl]carbamate; EMD Millipore, Billerica, MA) in DMSO for 2 hours prior to 4-hour treatment with MT compounds (with or without MG-132). For experiments involving lysosomal inhibition, cells were pretreated with 100 µM chloroquine (CQ) for 30 minutes prior to the addition of MT compounds (with or without CQ). To induce tau hyperphosphorylation in rat primary cortical neurons, 10 DIV cultures were treated for 8 hours with 15 nM OA. Vinblastine, CQ, and OA were purchased from Sigma-Aldrich and paclitaxel from Cytoskeleton Inc., whereas the rest of the MT compounds used in this study were synthesized in house and are described in the chemistry section above or in the Supplemental Methods.
P-gp Inhibition Assay.
ADR-RES cells, which express the human P-gp transporter, were plated at a density of 3.5 × 104 cells/well of 96-well clear plates in 0.1 ml medium. After overnight incubation at 37°C/5% CO2, cells were treated with 50 µl verapamil (a P-gp inhibitor; Sigma-Aldrich), cyclosporin A (a P-gp substrate; Sigma-Aldrich), or test compounds dissolved in DMSO and incubated for 17 minutes (37°C/5% CO2). After incubation, 50 µl of 2 µM Calcein AM (Thermo Fisher Scientific) was added to each well (final concentration of 1 µM) and plates were incubated for an additional 25 minutes (37°C/5% CO2). The medium was removed and cells were washed three times with cold wash buffer [Dulbecco’s modified Eagle’s medium without phenol red (Thermo Fisher Scientific), 10% FBS, and 2 mM l-glutamine]. Calcein fluorescence was analyzed with excitation at 485 nm and emission at 530 nm. Vehicle (0.2% DMSO)–treated cells served as 0% inhibition controls, whereas verapamil (100 µM)–treated cells served as 100% inhibition controls.
Extraction of Whole-Cell or MT-Enriched Lysates.
To obtain whole-cell extract from primary neurons and QBI-293 cells, cells were washed once with 1× phosphate-buffered saline (PBS; pH 7.4) and were then lysed in 200 µl RIPA buffer containing PI cocktail, 1 mM PMSF, and 1 µM TSA. Lysed cells were scraped into 1.5-ml Beckman ultracentrifuge tubes (Beckman Coulter Inc., Brea, CA), briefly sonicated, and centrifuged at 100,000g for 30 minutes at 4°C. After centrifugation, the supernatant from each sample was collected and analyzed for protein content by BCA assay. To separate the free tubulin fraction from the tubulin incorporated into MTs, fractionation under MT-stabilizing conditions was performed. Methods were adapted from a previously reported protocol (Joshi and Cleveland, 1989). Cells were strictly maintained at 37°C and washed with prewarmed microtubule-stabilization buffer (MSB; 0.1 M 4-morpholineethanesulfonic acid, 1 mM MgSO4, 2 mM EGTA, 0.1 mM EDTA, and 10% glycerol, pH 6.8). For the extraction of soluble tubulin, 500 µl MSB containing 1% Triton X-100, 10 µM paclitaxel (Cytoskeleton Inc.), PI, PMSF, and TSA (concentrations same as above) was added to each well and incubated at 37°C for 5 minutes with occasional gentle agitation. The solution was then transferred to a 1.5-ml Eppendorf tube and centrifuged for 2 minutes at 16,000g. The supernatant was collected as the soluble (MSB) fraction. Cellular cytoskeletons remaining bound to the tissue culture plate were dissolved in 250 µl (for QBI-293 cells) or 300 µl (for neurons) SDS buffer (50 mM Tris, 150 mM NaCl, and 2% SDS, pH 7.4) plus PI, PMSF, and TSA. The cytoskeletal (SDS) fraction was combined with the pellet (if any was obtained) from centrifugation of the Triton-soluble fraction and sonicated to complete the solubilization. The neuronal MSB fraction was concentrated using centrifugal protein concentrators (EMD Millipore) and all samples were brought up to 150 µl (one-half the volume of the SDS fraction). The protein content in the MSB fraction from QBI cells was sufficient for immunoblotting without performing a concentration step. The MSB fraction was analyzed for protein concentration by the BCA assay.
Acetyl- and α-Tubulin ELISA.
The ELISA was performed as previously described (Brunden et al., 2011). Briefly, 384-well plates were coated with 12G10 α-tubulin antibody (10 µg/ml; Covance, Princeton, NJ) in 30 µl cold 0.1 M bicarbonate buffer. 12G10 anti–α-tubulin antibody was originally deposited in the Developmental Studies Hybridoma Bank. After overnight incubation at 4°C, the plates were blocked in Block Ace solution (Bio-Rad, Hercules, CA) for a minimum of 24 hours at 4°C. Neuronal or QBI-293 cell homogenates, or cortical homogenates from in vivo studies, were diluted in C buffer (0.02 M sodium phosphate, 2 mM EDTA, 0.4 M NaCl, 1% bovine serum albumin, and 0.005% thimerosal, pH 7.0). Typically, 2-fold dilutions from 66.7 ng/µl to 16.7 ng/µl for neurons, 266 ng/µl to 22.2 ng/µl for QBI cells, or 13.3 ng/µl to 6.67 ng/µl for brain homogenate were performed and 30 µl of the sample was added to wells in duplicate. Plates were sealed, centrifuged, and incubated overnight at 4°C. After incubation with antigen, wells were aspirated and washed with PBS containing 0.05% Tween-20 and 0.005% thimerosal (PBS-Tween buffer). A horseradish peroxidase (HRP)–acetyl-tubulin reporter antibody was prepared by conjugating acetyl-tubulin primary antibody (Sigma-Aldrich; clone 6-11B-1) to HRP using a commercially available peroxidase labeling kit (Roche Applied Science, Indianapolis, IN). The HRP-acetyl-tubulin reporter antibody [1:1000 (v/v)] or preconjugated HRP–α-tubulin [1:5000 (v/v); ProteinTech Group, Chicago, IL] diluted in C buffer were added to appropriate wells (30 µl/well). The plates were sealed and incubated at room temperature for 4 hours on a platform rocker, followed by washing with PBS-Tween buffer. Peroxidase substrate solution (KPL, Gaithersburg, MD) was added to each well and the reaction was quenched after 10 minutes with 10% phosphoric acid. Plates were read on a SpectraMax M5 plate reader at an absorbance of 450 nm. The amount of acetyl- and α-tubulin protein in each sample was extrapolated using standard curves generated from serial dilutions of known protein standard (acetyl- or α-tubulin) concentrations.
Equal amounts of protein from cell extracts in RIPA, MSB, or SDS buffers were dissolved in sample buffer, loaded onto 10% polyacrylamide gels (or 15% gels for LC3 detection), and separated by electrophoresis. Protein was then transferred from gels onto 0.2 µm–pore size nitrocellulose membranes (Bio-Rad), which were blocked in 5% milk/Tris-buffered saline (TBS) for 1 hour at room temperature (23°C). Membranes were then incubated overnight in primary antibody diluted in blocking buffer at 4°C on a rocking platform. Primary antibodies consisted of acetyl-tubulin [clone 6-11B-1, 1:3000 (v/v); Sigma-Aldrich], 12G10 α-tubulin [1:3000 (v/v); Covance], β-tubulin [1:1000 (v/v); Abcam, Cambridge, MA], detyrosinated (glu)-tubulin [1:3000 (v/v); Millipore], ubiquitin [1:1000 (v/v); Millipore], LC3 [1:1000 (v/v); MBL International, Woburn, MA], and AT8 [1:2000 (v/v); Thermo Fisher Scientific]. Primary antibodies to glyceraldehyde-3-phosphate dehydrogenase [1:2000 (v/v); Advanced Immunochemical Inc., Long Beach, CA], β-actin [1:4000 (v/v); ProteinTech Group], and histone H3 [1:3000 (v/v); Cell Signaling Technology, Danvers, MA] were used as loading controls for quantification. In addition, a reversible protein staining kit for nitrocellulose membranes (Thermo Fisher Scientific) was used to confirm equal protein loading. After overnight incubation in primary antibody, membranes were washed three times with TBS containing 0.05% Tween-20 buffer and then incubated in infrared dye-conjugated goat anti-mouse or goat anti-rabbit secondary antibodies [1:20,000 (v/v); Li-Cor Biosciences, Lincoln, NE] for 1 hour at room temperature. After incubation with secondary antibodies, membranes were washed three times in TBS containing 0.05% Tween-20 buffer and then imaged using an Odyssey infrared imaging system (Li-Cor). Relative protein amounts were quantified using ImageStudio software (Li-Cor) and normalized to loading control proteins.
After treatment, primary neurons or COS-7 cells on coverslips were washed rapidly with prewarmed PEM buffer (80 mM 1,4-piperazinediethanesulfonic acid, 5 mM EGTA, and 1 mM MgCl2, pH 6.8) and fixed with 0.3% glutaraldehyde (in PEM buffer) at 37°C for 10 minutes. After fixation, cells were washed and membranes were permeabilized with 0.2% Triton X-100 for 15 minutes. Unreacted aldehydes were removed by three consecutive 10-minute treatments with 1 mg/ml sodium borohydride diluted in PBS. Coverslips were washed with PBS containing 0.1% Triton X-100 (PBS-T) and incubated in blocking buffer (3% FBS/3% bovine serum albumin in PBS) for 1 hour at room temperature. Acetyl-tubulin [1:3000 (v/v); Sigma-Aldrich] and β-tubulin [1:3000 (v/v); Abcam] antibodies were diluted in blocking buffer and applied overnight at 4°C. After primary antibody incubation, coverslips were washed with PBS-T and incubated with Alexa Fluor 594– or Alexa Fluor 488–conjugated secondary antibodies [1:2000 (v/v); Thermo Fisher Scientific] for 1 hour at room temperature. After incubation in secondary antibody, coverslips were washed with PBS-T and mounted on double-frosted microscope slides using 4′,6-diamidino-2-phenylindole Fluoromount-G (Southern Biotech, Birmingham, AL) to label nuclei. Slides were imaged by fluorescence microscopy (Olympus, Center Valley, PA) and representative photographs for each condition were acquired.
Linear mixed-effect models were used to compare the outcomes in cell culture studies. The fixed effects in the linear mixed-effects model were the treatment types and replicate runs, whereas experiment-specific random intercepts were used to account for the correlation between repeated measures within an experiment. Treatment effects were assessed with NCSS10 software (NCSS, Kaysville, UT). Results were plotted graphically and expressed as the average ± S.E.M. using GraphPad Prism software (GraphPad Software Inc., La Jolla, CA). Each study that underwent statistical analysis was performed in a minimum of three independent experiments (n ≥ 2 per study). For in vivo studies, treatment effects were assessed by one-way analysis of variance with Dunnett’s post hoc test for comparison with vehicle-treated control animals using GraphPad Prism software (n = 3 mice per study). P values < 0.05 were considered statistically significant.
Prototypic PPD and TPD Compounds Induce Proteasome-Dependent Degradation of α- and β-Tubulin.
We previously demonstrated that the prototype PPD and TPD examples CNDR-51549 and CNDR-51555 (compounds 20 and 9, respectively, in Lou et al., 2014; refer to Table 2 for chemical structures) increase levels of acetylated tubulin after 4 hours of treatment with doses up to 1 µM in QBI-293 cells (Lou et al., 2014). However, further evaluation of these compounds revealed that increasing concentrations beyond 1 μM no longer increased acetyl-tubulin levels in the QBI-293 cells, as determined by ELISA (Fig. 1A). Thus, rather than observing an additional increase or a plateau effect on acetyl-tubulin levels when tested at 10 µM, a bell-shaped or inverse U concentration-response curve was observed in which both compounds appeared inactive at the highest dose (Fig. 1A). This unusual concentration response was not observed in assays in which CNDR-51549 or CNDR-51555 was incubated with purified tubulin, as a linear concentration-dependent increase was observed both in light scattering that resulted from the formation of higher-order structures and in the amount of assembled tubulin that was found in the pellet fraction after centrifugation (Supplemental Fig. 1). The observed decrease in acetyl-tubulin in the cellular assay could have resulted from an effect that left stable MTs intact but somehow altered tubulin acetylation. However, immunoblot analysis of detyrosinated (glu)-tubulin, an independent marker of stable MTs that depends on the enzymatic removal of the COOH terminus of α-tubulin (Kreis, 1987), revealed that treatment of QBI-293 cells with 1 μM CNDR-51555 resulted in a significant increase of glu-tubulin, whereas 10-µM treatment with CNDR-51555 did not increase this marker of stable MTs (Fig. 1, C and D). Interestingly, the 10 µM treatment with CNDR-51549, which did not increase acetyl-tubulin levels appreciably relative to control cells, still increased glu-tubulin levels above baseline, albeit to a lesser extent than the 1 µM concentration of this compound (Fig. 1, E and F). To further investigate the relative absence of effect on stable MT markers after treatment with 10 μM of the prototypic PPD and TPD compounds, the total amount of cellular α-tubulin available to be incorporated into MTs at each treatment concentration was assessed by ELISA (Fig. 1B). Surprisingly, α-tubulin levels were significantly decreased at concentrations of CNDR-51555 (100 nM and 1 µM) and CNDR-51549 (1 µM) that increased acetyl-tubulin and glu-tubulin (Fig. 1, A and C–F). Cellular α-tubulin levels declined in a concentration-dependent manner, with only 27% and 55% of control levels remaining after 4 hours of treatment with 10 µM CNDR-51555 or CNDR-51549, respectively, as determined by ELISA (Fig. 1B). In accordance with these findings, QBI-293 cells treated with 100 nM of the TPD, cevipabulin (CNDR-51533), also demonstrated a similar decrease in α-tubulin levels (56% ± 2.7% of those from untreated cells; graph not shown). Assessment of modified (acetyl- and glu-tubulin) as well as total α- and β-tubulin by immunoblot analyses verified the ELISA results and confirmed that β-tubulin levels are reduced in a similar manner to α-tubulin after treatment with CNDR-51555 (Fig. 1, C and D) or CNDR-51549 (Fig. 1, E and F). This effect appeared specific to tubulin, because levels of other cytoskeleton-associated proteins, including β-actin, were not altered by treatment with either compound.
The observation of decreased α- and β-tubulin levels upon treatment with the PPD and TPD compounds at concentrations that increased acetyl- and glu-tubulin levels was unexpected, particularly given our prior observation that 100-nM cevipabulin treatment appeared to yield normal MT morphology as judged by acetyl-tubulin immunofluorescence staining of QBI-293 cells (Lou et al., 2014). To evaluate morphologic changes in the MT network after treatment with the MT-targeted compounds, immunocytochemistry was performed in COS-7 cells, because it is easier to visualize MTs in these cells than in QBI-293 cells. Cells were treated with vehicle (DMSO), 100 nM EpoD, a natural product believed to bind to the same site on the β-tubulin subunit of assembled MTs as taxol, or 1 or 10 µM CNDR-51555 or CNDR-51549. After 4 hours of treatment, the cells were fixed and double labeled for acetyl-tubulin and β-tubulin. There was a marked increase in acetyl-tubulin immunoreactivity after treatment with EpoD, 1 µM CNDR-51555 (Fig. 2A), or 1 μM CNDR-51549 (Supplemental Fig. 2A). However, whereas β-tubulin labeling appeared normal in the EpoD-treated cells, the cells exposed to 1 μM CNDR-51555 showed some evidence of altered β-tubulin distribution and intensity, and 10-µM treatment with either CNDR-51555 or CNDR-51549 resulted in barely detectable acetyl-tubulin, with β-tubulin labeling that was diffuse and, in the case of CNDR-51555, greatly diminished compared with both control and EpoD-treated cells (Fig. 2; Supplemental Fig. 2).
There is some precedent for MT-directed agents altering cellular α- and β-tubulin levels, with isothiocyanates having been previously identified as MT-binding compounds capable of inducing rapid tubulin degradation (Mi et al., 2009a). It was suggested that isothiocyanates induce MT-misfolding upon covalent binding to tubulin, with eventual degradation of α- and β-tubulin by an incompletely characterized proteasome-mediated pathway (Mi et al., 2009b). In addition, another study demonstrated loss of α-tubulin after administration of the vinca alkaloid, vincristine (Huff et al., 2010). However, this effect was only observed in neural HCN2 cells and not in a variety of other cell lines (Huff et al., 2010). In this regard, treatment of QBI-293 cells with 1 µM of the vinca alkaloid vinblastine, which promotes MT depolymerization and redistribution of tubulin from cytoskeletal to soluble cellular pools (Supplemental Fig. 3, A and B), does not significantly alter total levels of α- or β-tubulin (Supplemental Fig. 3, C–E). These findings are significant in light of previous studies, which demonstrated that cevipabulin displaced vinblastine in competitive binding assays (Beyer et al., 2008, 2009). Thus, if the cevipabulin analog CNDR-51555 interacts with the vinca site of MTs, it elicits an effect on cellular tubulin levels that differs from that of vinblastine.
To elucidate the mechanism of decreased α- and β-tubulin after TPD or PPD treatment, we assessed the contributions of autophagic-lysosomal and proteasomal protein degradation pathways. Autophagic clearance through lysosomal pathways comprises a major mechanism of cellular protein degradation (Mizushima and Komatsu, 2011); we thus treated QBI-293 cells with the lysosome inhibitor, CQ (Wibo and Poole, 1974), 30 minutes prior to and during administration of 1 µM CNDR-51555 or CNDR-51549 to determine whether autophagic inhibition prevented the reduction in tubulin levels. CQ had no effect on the reduction in α-tubulin levels induced by treatment with 1 µM CNDR-51555 or CNDR-51549, as assessed by ELISA (Fig. 3A). Likewise, immunoblot analysis of lysates from QBI-293 cells treated with 1 µM CNDR-51555 or CNDR-51549 in the absence or presence of CQ demonstrated that the compound-mediated decrease in β-tubulin levels could not be rescued by treatment with CQ (Fig. 3, B and C). Accumulation of LC3-II protein (lower molecular weight band on LC3 blot) confirmed effective inhibition of the autophagic pathway in response to CQ (Fig. 3, B and C). Thus, alterations in tubulin levels in response to the prototype TPD and PPD compounds are unlikely attributable to autophagic protein clearance.
Proteasome-dependent protein degradation represents a highly used pathway for cells to degrade unnecessary or damaged proteins (Lecker et al., 2006), and as noted, binding of certain isothiocyanates to MTs has been demonstrated to induce tubulin degradation by proteasomes (Mi et al., 2009a). In this context, 2-hour pretreatment of QBI-293 cells with the proteasome inhibitor, MG-132, followed by the addition of CNDR-51555 for an additional 4 hours, significantly attenuated the loss of α-tubulin observed after treatment with CNDR-51555 alone, as assessed by ELISA analysis of RIPA-soluble cell lysates (Fig. 4A). However, MG-132 only partially rescued tubulin levels in this fraction after treatment with CNDR-51555 and did not rescue α-tubulin loss from the RIPA-soluble fraction of CNDR-51549–treated cells (Fig. 4A). Western blotting of lysates from MG-132–treated cells revealed the presence of accumulated ubiquitinated proteins, thereby confirming proteasome inhibition (Fig. 4B). Because proteasome inhibition might lead to the accumulation of ubiquitinated and aggregated proteins with decreased RIPA solubility (Demasi and Davies, 2003), as has been previously observed for tubulin after treatment of cells with certain isothiocyanates (Mi et al., 2009a,b), the RIPA-insoluble fraction from treated cells was dissolved in 2% SDS-containing buffer and analyzed for tubulin content by immunoblotting. Significant accumulation of α- and β-tubulin was observed in the SDS-soluble fraction from cells treated with the combination of MG-132 and CNDR-51555 or CNDR-51549, but not under any other condition (Fig. 4, C and D). Because the SDS concentration in the insoluble fraction prevented ELISA analysis, equal proportions of RIPA-soluble and SDS-soluble cellular fractions were combined for each treatment and assessed by immunoblotting, which revealed nearly complete rescue of total α- and β-tubulin levels after dual treatment of MG-132 and CNDR-51555/CNDR-51549 compared with compound treatment alone (Fig. 4, E–H). These findings suggest that both of these compounds induce proteasome-dependent tubulin degradation. However, unlike what was previously reported for isothiocyanate-treated cells (Mi et al., 2009a), we observed no evidence of increased insolubility of tubulin in the absence of MG-132; thus, the mechanism by which the PPD and TPD compounds induce tubulin degradation is presently unknown, although it may relate to alterations in tubulin conformation.
Identification of a TPD Series that Exhibits MT-Stabilizing Activity without Reducing Total Tubulin Levels.
Although our prototypic PPD and TPD compounds seem to possess MT-stabilizing properties at certain concentrations based on increases in acetyl-tubulin and glu-tubulin, their substantial negative effect on total tubulin levels and apparent disruption of MT architecture likely preclude them from providing any therapeutic benefit in AD or related tauopathies. However, further evaluation of additional members of the TPDs and PPDs (Lou et al., 2014) led to the identification of one compound (CNDR-51597 or (S)-5-chloro-6-(2,4,6-trifluorophenyl)-N-(1,1,1-trifluoropropan-2-yl)-[1,2,4]triazolo[1,5-a]pyrimidin-7-amine; compound 8 in Lou et al., 2014), which induced elevated acetyl-tubulin at concentrations up to 10 µM without decreasing α-tubulin levels (Fig. 5A; Table 1). Like CNDR-51555, this compound belongs to the TPD class of heterocycles. However, unlike the former, CNDR-51597 lacks the alkoxy side chain, suggesting the possibility of a structure-activity relationship in which the nature of the substituent in the para position may be ultimately involved in determining the cellular phenotype. Indeed, among all TPDs tested, those containing an alkoxy side chain (e.g., cevipabulin, CNDR-51555, and CNDR-51567 ((S)-3-(4-(5-chloro-7-((1,1,1-trifluoropropan-2-yl)amino)-[1,2,4]triazolo[1,5-a]pyrimidin-6-yl)-3,5-difluorophenoxy)propan-1-ol)) demonstrated a bell-shaped dose-response curve in the acetyl-tubulin assay, as well as a decrease in total α-tubulin levels (Table 2), whereas TPD congeners in which the side chain is replaced by a fluoride (e.g., CNDR-51539 (5-chloro-N-(2,2,2-trifluoroethyl)-6-(2,4,6-trifluorophenyl)-[1,2,4]triazolo[1,5-a]pyrimidin-7-amine), CNDR-51597, CNDR-51647 (5-chloro-7-(4-methylpiperidin-1-yl)-6-(2,4,6-trifluorophenyl)-[1,2,4]triazolo[1,5-a]pyrimidine), CNDR-51655 (5-chloro-N-methyl-N-(2,2,2-trifluoroethyl)-6-(2,4,6-trifluorophenyl)-[1,2,4]triazolo[1,5-a]pyrimidin-7-amine), and CNDR-51657) were found to produce linear dose responses in the cell-based assays, with no evidence of a reduction of total tubulin (Table 1). Hereafter, we refer to the improved examples in Table 1 as TPD+ compounds. Although these results indicate a structure-activity relationship for the TPD series, all of the PPD compounds elicited a similar cellular phenotype as CNDR-51549 (Table 2), regardless of whether the substituent in the para position is a fluoride or an alkoxy side chain.
Although no TPD+ compounds were active at concentrations of 10 or 100 nM, compounds that increased acetyl-tubulin at 1 µM underwent additional in vivo testing for brain exposure. Wild-type mice received a 5-mg/kg intraperitoneal injection of compound and brain and plasma drug levels were determined 1 hour post-injection. Resultant average brain levels and blood/plasma (B/P) ratios are indicated in Table 1 for selected compounds. One of the most potent compounds identified in the acetyl-tubulin assay, CNDR-51657, also demonstrated a favorable in vivo profile, with brain levels averaging 812 ± 61 nM and a high B/P ratio of 2.7 ± 0.7 (Table 1). Although previously reported intraperitoneal dosing with CNDR-51555 and CNDR-51549 resulted in higher overall brain exposure than CNDR-51657 (1300 ± 200 nM for CNDR-51555; 2900 ± 100 nM for CNDR-51549), the B/P ratios were much less favorable (0.27 ± 0.2 for CNDR-51555; 0.58 ± 0.01 for CNDR-51549), thus resulting in much higher peripheral drug exposure (Lou et al., 2014). An average concentration of 670 ± 580 nM of CNDR-51657 was detected in brain tissue of mice 1 hour after a single oral gavage dosing of 10 mg/kg, revealing that the compound has reasonably good oral absorption in addition to excellent brain exposure.
Finally, the effects of newly synthesized compounds were evaluated for activity in a P-gp inhibition assay. Of the compounds described in Tables 1 and 2, the PPD compound CNDR-51554 ((S)-6-chloro-5-(4-(3-(dimethylamino)propoxy)-2,6-difluorophenyl)-2-(pyrazin-2-yl)-N-(1,1,1-trifluoropropan-2-yl)pyrimidin-4-amine) and TPD+ compound CNDR-51647 exhibited the highest P-gp inhibitory activity in the assay, achieving 29% and 37% inhibition, respectively, at the highest concentration tested of 33.3 µM (Supplemental Table 1). CNDR-51657 inhibited P-gp activity by approximately 19% at the 33.3-µM concentration, but activity decreased in a concentration-dependent manner and was negligible at doses that resulted in increased acetyl-tubulin (Supplemental Fig. 4; Supplemental Table 1). Full dose-response curves of CNDR-51657 and previously tested compounds EpoD, CNDR-51555, and CNDR-51549 are provided (Supplemental Fig. 4). These data suggest that the majority of newly synthesized TPD+ compounds would interfere minimally with P-gp function at concentrations that increase acetyl-tubulin.
To characterize the biologic actions of the TPD+ compounds in greater depth, CNDR-51657 was selected as a prototype to represent this series and to undergo additional testing for comparison with the original TPD lead, CNDR-51555. CNDR-51657 significantly increased levels of acetylated tubulin at 1 µM and 10 µM by 2.4- and 3.9-fold, respectively, as determined by ELISA (Fig. 5B; Table 1), whereas α-tubulin levels were not decreased at any dose (Fig. 5B). Immunoblot results confirmed that CNDR-51657 also increased levels of glu-tubulin in a dose-dependent manner and did not negatively affect levels of β-tubulin (Fig. 5, C and D). To rule out the possibility that a bell-shaped concentration response in acetyl-tubulin levels or a loss of α-tubulin was not observed with CNDR-51657 simply due to decreased potency of this compound compared with the PPD and TPD prototypes, QBI-293 cells were treated with higher concentrations (10, 30, or 60 µM) and assessed for levels of acetyl- and α-tubulin by ELISA. A concentration-dependent increase of acetyl-tubulin was observed at these higher compound exposures, and α-tubulin levels remained at those of vehicle-treated cells (Supplemental Fig. 5). For comparison, CNDR-51533 (cevipabulin)–treated cells demonstrated a significant reduction in α-tubulin at 100 nM, even though this concentration induced a significant increase in acetyl-tubulin levels (Supplemental Fig. 5). These results suggest CNDR-51657, and likely other TPD+ compounds, behave in a mechanistically distinct manner than TPD congeners bearing the alkoxy side chain.
CNDR-51657 Increases Levels of Post-Translationally Modified Tubulin in Cytoskeletal Cellular Fractions without Altering MT Morphology.
To assess the effects of treatment with CNDR-51657 on the amount of free versus MT-incorporated tubulin, QBI-293 cells were treated with increasing doses of CNDR-51657 for 4 hours, followed by harvest under MT-stabilizing conditions. Soluble proteins were first extracted from cells adherent to the tissue culture dish in MT-stabilizing buffer containing 1% Triton X-100 (soluble fraction). Under these conditions, the majority of the cytoskeleton, as well as the nuclei, remained bound to the tissue culture dish (Joshi and Cleveland, 1989). After extraction of soluble proteins, the remaining cytoskeleton was solubilized in 2% SDS-containing buffer (cytoskeletal fraction). Analysis of both fractions by immunoblotting revealed that nearly all of the post-translationally modified tubulin was detected within the cytoskeletal/SDS fraction, as expected (Fig. 6A). Under these extraction conditions, α- and β-tubulin present in the soluble fraction represent free tubulin, whereas levels within the cytoskeletal fraction represent tubulin that has been incorporated into MTs. Treatment with 1 µM or 10 µM CNDR-51657, as well as with 0.1 μM EpoD, significantly increased the levels of acetyl-tubulin in the cytoskeletal fraction (Fig. 6, A and C); the 10 µM treatment with CNDR-51657, as well as EpoD, significantly increased levels of glu-tubulin (Fig. 6, A and C). Acetyl- and glu-tubulin were not detected within the soluble fraction at any dose of CNDR-51657, suggesting that MTs remained intact (Fig. 6A). In addition, significant increases in β-tubulin were observed within the cytoskeletal fraction after treatment with 1 or 10 µM CNDR-51657 or 0.1 μM EpoD (Fig. 6, A and C). The amount of α-tubulin within the cytoskeletal fraction was also slightly increased by 1 and 10 μM CNDR-51657, although significance was only reached at the 1 µM dose or with 0.1 μM EpoD (Fig. 6, A and C). The increase in cytoskeletal tubulin was associated with a corresponding decrease in levels of α- and β-tubulin in the soluble fraction (Fig. 6, A and B). Collectively, these data suggest that TPD+ compounds such as CNDR-51657 cause redistribution of soluble tubulin to MTs, thereby increasing MT mass in a manner similar to the MT-stabilizing taxanes and epothilones (Schiff and Horwitz, 1980; Bollag et al., 1995).
To evaluate cytoskeletal architecture after treatment with CNDR-51657, COS-7 cells were treated for 4 hours with 1 μM of the compound and the MT structure was examined by immunocytochemistry. An increase in acetylated tubulin was observed and the levels and distribution of β-tubulin did not appear to change relative to vehicle-treated cells with normal MT structures (Fig. 6D). These findings are in contrast with what was observed at a similar concentration of CNDR-51555 (Fig. 2). At a 10 µM concentration of CNDR-51657, β-tubulin immunolabeling remained robust, although tubulin bundling appeared to occur around the nucleus (Fig. 6E), which was also observed upon treatment with 0.1 μM EpoD (Fig. 2) and could be a result of mitotic block. The distribution of β-tubulin within the cells treated with 10 µM CNDR-51657 is substantially different than that observed within an identical concentration of CNDR-51555 or CNDR-51549 (Fig. 6E), with the latter compounds causing dramatic redistribution of β-tubulin, including the formation of punctate structures after treatment with CNDR-51555. These findings demonstrate that CNDR-51657 increases both markers of stable MTs and MT mass, characteristic of well described MT-stabilizing compounds, and differs markedly in its activity relative to the structurally related compounds CNDR-51555 and CNDR-51549.
CNDR-51657 Increases Markers of Stable MTs in Rat Primary Cortical Neurons without Disrupting Neuronal Morphology.
Because the assessment of MT-stabilizing activity of the TPD and PPD compounds was performed only in dividing cells (QBI-293 and COS-7 cell lines), an evaluation of compound effects on the MTs of primary neurons was important to assess their potential for future testing in transgenic mouse models of tauopathy. Neuronal MTs serve as the major architectural framework for maintaining cellular morphology, and they also serve as the railways for the transport of critical cellular proteins and organelles. Because neuronal MTs are highly stable under normal conditions within the brain and in mature neuronal culture systems, rat cortical neurons were assessed at a young age (3 DIV) when MTs tend to exhibit increased dynamicity compared with older cultures. Although CNDR-51555 and CNDR-51549 were shown to increase acetylated tubulin in dividing cells at moderate concentrations, both compounds failed to increase acetyl-tubulin in primary neurons at the tested concentrations after 24 hours of treatment (Fig. 7A). In fact, acetyl-tubulin levels were significantly decreased by CNDR-51555 at 100 nM and 1 µM, and by CNDR-51549 at 1 µM (Fig. 7A). Similarly, there was a concomitant decrease in α-tubulin levels, as assessed by ELISA (Fig. 7B). Evidence of toxicity was observed after 10 µM treatment with either compound, suggesting that neurons may exhibit heightened sensitivity to these molecules. Because of the observed toxicity in the neuronal cultures, subsequent studies in primary neurons were conducted at concentrations not exceeding 1 µM.
Unlike the results obtained after treatment of neurons with CNDR-51555 or CNDR-51549, ELISA results demonstrated that treatment of rat primary cortical neurons at 3 DIV with 1 µM CNDR-51657 for 24 hours induced a significant increase in acetyl-tubulin without a concomitant decrease in α-tubulin (Fig. 7C). The increase in acetyl-tubulin was similar in magnitude to that observed after treatment with the positive control, 100 nM EpoD (Fig. 7C). These results were confirmed by immunoblotting, which also demonstrated a significant increase in glu-tubulin with no change in total tubulin levels (Fig. 7, D and E). In contrast, both glu-tubulin and total tubulin amounts were decreased by treatment with CNDR-51555 or CNDR-51549 (Supplemental Fig. 6).
Levels of post-translational modification and total tubulin from neurons treated with CNDR-51657 were also assessed in the soluble and cytoskeletal fractions after harvest in MT-stabilizing buffer. Significantly increased acetyl- and glu-tubulin levels were observed in the cytoskeletal fraction from neurons treated at 3 DIV with 1 µM CNDR-51657 or 0.1 µM EpoD for 24 hours (Fig. 8, A and C). Similar to the results obtained in QBI-293 cells, 1 µM treatment with CNDR-51657 increased α- and β-tubulin levels in the cytoskeletal fraction, although only the increase in β-tubulin reached significance (Fig. 8, A and C). A concomitant decrease was observed in α- and β-tubulin levels in the soluble fraction, reaching significance for α-tubulin (Fig. 8, A and B). These findings again suggest that TPD+ compounds, like EpoD, increase both MT stability and MT mass across multiple cell types.
Although CNDR-51657 and EpoD were only effective in increasing markers of stable MTs in young cultures (3 DIV), because of a majority of tubulin already being incorporated into MTs and post-translationally modified in older neuron cultures, we nonetheless examined the effects of compounds on more mature cultures in which processes and interneuronal connections have been more fully established. Rat primary cortical neurons at 7 DIV were examined by immunocytochemical staining with acetyl-tubulin antibody after 24-hour treatment with 1 µM CNDR-51657 or CNDR-51555. In comparison with vehicle-treated neurons, those exposed to CNDR-51555 for 24 hours demonstrated significant alterations in morphology and diminished acetyl-tubulin immunoreactivity (Fig. 8D). However, neurons treated with CNDR-51657 had a similar morphology to vehicle-treated neurons (Fig. 8D). Neurons treated with 100 nM EpoD also appeared similar to vehicle-treated neurons, although processes appeared slightly shorter and less branched (Fig. 8D), which could account for the slight decrease observed in total β-tubulin levels in whole-cell extracts from EpoD-treated neurons (Fig. 7, D and E). Overall, our findings suggest that TPD+ compounds, but not other TPD or PPD analogs, increase MT stabilization in primary neurons without significantly altering neuronal morphology or disrupting established networks.
CNDR-51657 Restores Neuronal MT Stability in Response to Tau Hyperphosphorylation In Vitro.
We have provided evidence that the TPD+ compound, CNDR-51657, can increase markers of MT stability in primary neuronal cultures. To evaluate whether this compound can stabilize neuronal MTs under conditions of tau loss of function, we treated rat primary cortical neurons (10 DIV) with OA in the absence or presence of CNDR-51657 for 8 hours. OA potently inhibits the serine threonine protein phosphatases PP1 and PP2A (Bialojan and Takai, 1988), and treatment of neurons with OA has been shown to induce hyperphosphorylation of tau as well as decrease markers of MT stability, including acetylated tubulin (Merrick et al., 1997; Das and Miller, 2012). A significant increase in phosphorylated tau recognized by the AT8 antibody, which detects tau phosphorylated at serine 202 and threonine 205, confirms the efficacy of OA treatment (Fig. 9A). Furthermore, the reduction in acetyl- and glu-tubulin levels in OA-treated neurons demonstrates a loss of MTs that likely resulted from the disengagement of tau from MTs (Fig. 9, A and B). Although treatment of 10 DIV neurons for 8 hours with 1 µM CNDR-51657 alone had no effect on levels of post-translationally modified tubulin, coadministration of CNDR-51657 significantly attenuated the loss of acetyl- and glu-tubulin observed in OA-treated neurons (Fig. 9, A and B). Moreover, immunolabeling of acetyl-tubulin in these cultures demonstrated the preservation of both acetyl-tubulin levels and MT morphology in neurons treated dually with OA and CNDR-51657 compared with OA treatment alone (Fig. 9C). These findings suggest that whereas TPD+ compounds have little effect on mature neurons in culture, they offer protection against MT loss and/or dysfunction under conditions of tau hyperphosphorylation and disengagement from MTs, as could occur in AD and related tauopathies. Conversely, treating neurons simultaneously with OA and CNDR-51555 or CNDR-51549 did not result in an inhibition of acetyl-tubulin loss induced by OA treatment (Supplemental Fig. 7). Each of these compounds was administered at 100 nM, because higher concentrations of either compound were shown to reduce neuronal acetyl-tubulin (Fig. 7; Supplemental Fig. 6). These studies further differentiate CNDR-51657 and the TPD+ compounds from the other TPD and PPD examples.
CNDR-51657 Increases Acetyl-Tubulin in the Cortex of Wild-Type Mice.
To examine the in vivo effects of CNDR-51657, wild-type CD-1 mice were administered two doses of 1 or 5 mg/kg CNDR-51657, approximately 18 hours apart, by intraperitoneal injection. Four hours after the second dose, mice were euthanized and cortices were dissected for analysis of acetyl- and α-tubulin levels by ELISA. Acetyl-tubulin levels were significantly increased in both treatment groups compared with vehicle-treated animals (Fig. 10A). α-Tubulin levels were not significantly affected by either dose of CNDR-51657 (Fig. 10A). When the acetyl-tubulin/α-tubulin ratio was determined for each treatment group, a dose-dependent increase was observed (Fig. 10A). Past studies demonstrated an increase in the brain acetyl-tubulin/α-tubulin ratio of mice treated with the TPD CNDR-51555 daily for 4 or 6 days (Lou et al., 2014). To directly compare the efficacy of the TPD+ CNDR-51657 with that of the TPD CNDR-51555 in this study, wild-type CD-1 mice were injected with 1 mg/kg CNDR-51657 or 1 or 5 mg/kg CNDR-51555 following the same dosing scheme as described above. Acetyl-tubulin levels were highest in cortices from mice treated with 1 mg/kg CNDR-51657, but an increase was also observed after treatment with 5 mg/kg, but not 1 mg/kg, CNDR-51555 (Fig. 10B). α-Tubulin levels were not significantly changed across groups, although a trend toward a decrease was observed in mice treated with CNDR-51555 (Fig. 10B). Interestingly, we previously observed a decrease in brain α-tubulin levels after 6 days of treatment with CNDR-51549 or CNDR-51555, although at the time it was thought that this represented variability in the brain homogenates, because we were then unaware of the general phenotype of this compound class (unpublished data). The acetyl-tubulin/α-tubulin ratio was significantly increased in cortical tissue from mice treated with 1 mg/kg CNDR-51657 and 5 mg/kg CNDR-51555 (Fig. 10B). These studies suggest that the TPD+ CNDR-51657 can elicit an in vivo pharmacodynamic effect on acetyl-tubulin levels without negatively affecting α-tubulin levels and is also active in vivo at a lower dose than the previously disclosed TPD, CNDR-51555.
AD and related tauopathies are characterized by the formation of pathologic aggregates composed of the MT-associated protein, tau. Hyperphosphorylation and fibrillization of tau are associated with a concurrent decrease in MT-bound tau, and there is evidence that this results in MT deficits in cell culture (Alonso et al., 1994; Merrick et al., 1997) and in transgenic mouse models of tauopathy (Zhang et al., 2005, 2012; Barten et al., 2012). Moreover, there is evidence of MT deficits in the AD brain, including decreased acetylated tubulin (Hempen and Brion, 1996) and a reduction in the total number and length of MTs within neurons (Cash et al., 2003). Recently, postmortem evaluation of the AD brain revealed decreased levels of total α-tubulin, in addition to a decreased acetylated and glu-tubulin, although there was no correlation between the presence of tangles and post-translationally modified tubulin within individual neurons. This is consistent with the idea that it is the initial hyperphosphorylation of tau and its disengagement from MTs that results in MT deficits (Zhang et al., 2015).
Neurons are particularly vulnerable to dysfunction resulting from MT instability, because they rely on MT networks both for architectural support and for the axonal transport of critical cellular proteins and organelles (Alonso et al., 1997; Baird and Bennett, 2013). Thus, the loss of tau function would be predicted to be detrimental to MT function and neuronal health, as supported by the aforementioned observations in the AD brain. It was therefore somewhat surprising that genetic knockout of murine tau was initially reported to result in no overt phenotype (Harada et al., 1994; Dawson et al., 2001). However, mounting evidence suggests that age-dependent MT and motor function deficits develop in tau-depleted mice (Dawson et al., 2010; Lopes et al., 2016). In addition, primary neurons derived from tau knockout mice show impaired neurite outgrowth (Dawson et al., 2001). Finally, upregulation of additional MT-associated proteins, such as MAP1A, has been observed in tau knockout models (Harada et al., 1994), hampering the ability to discern the full effect of tau loss of function. Transgenic overexpression models have provided additional means to explore the consequences of tau hyperphosphorylation and the accumulation of pathologic tau. Indeed, our laboratory has shown that tau transgenic mice exhibit MT deficits, axonal dystrophy, behavioral impairments, neuron loss, and shortened life span (Ballatore et al., 2007; Brunden et al., 2012; Yoshiyama et al., 2013), which recapitulates many features of AD and related tauopathies. Although therapeutic strategies in AD have traditionally focused on lowering amyloid-β extracellular plaque load, tau pathology has received increasing attention over recent years, given that the levels of neurofibrillary tangles correlate more closely with cognitive decline than amyloid plaque burden (Wilcock and Esiri, 1982; Arriagada et al., 1992; Gómez-Isla et al., 1997). In this context, multiple approaches are being investigated that are aimed at reducing tau hyperphosphorylation or misfolded tau species (Brunden et al., 2009; Lee et al., 2011). An alternate approach is to target tau loss of function. In this context, our group (Brunden et al., 2010; Zhang et al., 2012) and others (Barten et al., 2012) have demonstrated in transgenic mouse models of brain tauopathy that the MT-stabilizing compound, EpoD, exerts improvements in MT and axonal function that result in reduced neurodegeneration and enhanced cognitive function. The observation that EpoD-treated tau transgenic mice also had reduced tau pathology (Barten et al., 2012; Zhang et al., 2012) suggests that MT dysfunction contributes to the deposition of insoluble tau or the inability of diseased cells to clear tau aggregates.
Here, we have described the biochemical properties of a series of non-naturally occurring MT-binding compounds that had been originally identified as potential antifungal agents and then later developed as candidate cancer chemotherapeutics (Zhang et al., 2007, 2009). An in-depth analysis of several TPD and PPD congeners (Lou et al., 2014) reveals interesting new elements of structure-activity relationships that led to the identification of two distinct groups of compounds based on the particular cellular phenotype elicited. Whereas active PPDs and TPDs bearing an alkoxy side chain in the para position promoted both an increase in markers of stable MTs and degradation of tubulin in a proteasome-dependent manner, the active TPD congeners lacking the alkoxy side chain did not decrease total tubulin levels or affect MT architecture. These “TPD+” small molecules induced concentration-dependent increases in markers of stable MTs and increased MT mass without obvious morphologic alteration to existing MT networks. Moreover, the prototype TPD+ small molecule, CNDR-51657, is brain penetrant, orally bioavailable, and stabilizes MTs in primary neurons and in the mouse brain after peripheral administration. Given these desirable properties, the TPD+ compounds represent a new class of potential candidate drugs for the treatment of neurodegenerative tauopathies.
Although the changes in MT structure and the reduction of tubulin levels caused by certain TPD compounds, exemplified by CNDR-51555, and by all tested PPDs indicate that these compounds would not be beneficial for the treatment of tauopathies, it is possible that these effects may prove useful in an oncology setting, particularly in glioblastoma, a highly fatal condition in which traditional therapeutics have failed to achieve sufficient brain exposure (Katsetos et al., 2015; van Tellingen et al., 2015). Although we do not fully understand why many of the TPD and PPD compounds induce proteasomal degradation of tubulin, one hypothesis is that these compounds promote the formation of aberrant or unstable MT structures. The promotion of anomalous MT structures that are subject to proteolysis would be consistent with the observed increase of acetyl-tubulin and glu-tubulin at concentrations in which reductions of total tubulin are also observed. The finding that TPD+ compounds increase markers of stable MT in the absence of tubulin degradation suggests that the nature of the substituent in the para position may ultimately determine the nature of the MT structure that forms, with TPD+ compounds lacking an alkoxy moiety inducing normal MTs that are no longer targeted for removal.
The ability of the TPD+ prototype, CNDR-51657, to increase markers of stable MTs in primary neurons without evidence of tubulin degradation is highly significant, especially since other PPD and TPD compounds that increased markers of stable MTs in QBI-293 cells failed to increase these markers in neurons. Importantly, CNDR-51657 was shown to increase MT mass within QBI-293 cells and neurons, and thus mimicked in many ways the effects of EpoD in these cells. Furthermore, treatment of older neuronal cultures (7 DIV) that have more mature neuritic processes with CNDR-51657 did not appear to alter neuronal morphology or disrupt existing MT networks in our culture systems. This contrasts dramatically with the effect of the related TPD, CNDR-51555, which caused significant MT disruption and truncation of neuritic processes. Finally, CNDR-51657 was able to alleviate MT deficits in neurons treated with OA, an in vitro model of tau hyperphosphorylation with associated loss of stable MTs, whereas CNDR-51555 was not.
In summary, we have characterized a number of small molecules containing PPD and TPD core structures and have demonstrated differential effects on cellular MTs that can be related to certain structural features of the compounds. A preferred set of TPD+ small molecules exhibit MT-stabilizing properties in multiple cell types, including primary neurons. Moreover, several examples from this class were found to penetrate the BBB and have excellent brain exposure and, importantly, CNDR-51657 was shown to increase acetylated tubulin in the mouse brain after peripheral administration. Thus, these compounds, as exemplified by CNDR-51657, hold promise as potential candidates for the treatment of tauopathies.
The authors thank Amy Lam and Sarah DeVaro for laboratory assistance. They also acknowledge the following donors to the Center for Neurodegenerative Disease Research: the Karen Cohen Segal and Christopher S. Segal Alzheimer Drug Discovery Initiative Fund, the Paula C. Schmerler Fund for Alzheimer’s Research, the Barrist Neurodegenerative Disease Research Fund, the Eleanor Margaret Kurtz Endowed Fund, the Mary Rasmus Endowed Fund for Alzheimer’s Research, Mrs. Gloria J. Miller, and Arthur Peck.
Participated in research design: Kovalevich, Cornec, Lee, Smith, Ballatore, Brunden.
Conducted experiments: Kovalevich, Cornec, Yao, James, Crowe.
Performed data analysis: Kovalevich, James, Crowe.
Wrote or contributed to the writing of the manuscript: Kovalevich, Cornec, James, Lee, Trojanowski, Smith, Ballatore, Brunden.
- Received December 2, 2015.
- Accepted March 14, 2016.
This research was supported by the National Institutes of Health National Institute on Aging [Grant R01-AG044332] and in part by the Marian S. Ware Alzheimer Foundation.
- Alzheimer’s disease
- blood–brain barrier
- bicinchoninic acid
- days in vitro
- enzyme-linked immunosorbent assay
- epothilone D
- fetal bovine serum
- horseradish peroxidase
- liquid chromatography
- benzyl N-[(2S)-4-methyl-1-[[(2S)-4-methyl-1-[[(2S)-4-methyl-1-oxopentan-2-yl]amino]-1-oxopentan-2-yl]amino]-1-oxopentan-2-yl]carbamate
- mass spectrometry
- tandem mass spectrometry
- microtubule-stabilization buffer
- okadaic acid
- phosphate-buffered saline
- protease inhibitor
- phenylmethylsulfonyl fluoride
- radioimmunoprecipitation assay
- Tris-buffered saline
- trichostatin A
- Copyright © 2016 by The American Society for Pharmacology and Experimental Therapeutics