|Home | About | Journals | Submit | Contact Us | Français|
We have previously shown that the L-type calcium channel (LCC) antagonist nilvadipine reduces brain amyloid-β (Aβ) accumulation by affecting both Aβ production and Aβ clearance across the blood-brain barrier (BBB). Nilvadipine consists of a mixture of two enantiomers, (+)-nilvadipine and (−)-nilvadipine, in equal proportion. (+)-Nilvadipine is the active enantiomer responsible for the inhibition of LCC, whereas (−)-nilvadipine is considered inactive. Both nilvadipine enantiomers inhibit Aβ production and improve the clearance of Aβ across the BBB showing that these effects are not related to LCC inhibition. In addition, treatment of P301S mutant human Tau transgenic mice (transgenic Tau P301S) with (−)-nilvadipine reduces Tau hyperphosphorylation at several Alzheimer disease (AD) pertinent epitopes. A search for the mechanism of action of (−)-nilvadipine revealed that this compound inhibits the spleen tyrosine kinase (Syk). We further validated Syk as a target-regulating Aβ by showing that pharmacological inhibition of Syk or down-regulation of Syk expression reduces Aβ production and increases the clearance of Aβ across the BBB mimicking (−)-nilvadipine effects. Moreover, treatment of transgenic mice overexpressing Aβ and transgenic Tau P301S mice with a selective Syk inhibitor respectively decreased brain Aβ accumulation and Tau hyperphosphorylation at multiple AD relevant epitopes. We show that Syk inhibition induces an increased phosphorylation of the inhibitory Ser-9 residue of glycogen synthase kinase-3β, a primary Tau kinase involved in Tau phosphorylation, by activating protein kinase A, providing a mechanism explaining the reduction of Tau phosphorylation at GSK3β-dependent epitopes following Syk inhibition. Altogether our data highlight Syk as a promising target for preventing both Aβ accumulation and Tau hyperphosphorylation in AD.
Alzheimer disease (AD)2 is the most prevalent form of dementia in the elderly. AD pathology is characterized by the presence of extracellular deposits of amyloid β (Aβ) peptides and by the accumulation of intraneuronal tangles made of hyperphosphorylated Tau. Although the amount of fibrillar Aβ peptides deposited as senile plaques at autopsy does not correlate with AD cognitive impairment (1), the levels of the soluble forms of the peptide appear to correlate better with the severity of the dementia (2). Recent longitudinal positron emission tomography imaging studies aimed at detecting Aβ in the brain of subjects with mild cognitive impairment, AD, and normal controls revealed that elevated brain Aβ levels at baseline are associated with greater cognitive worsening in both AD, mild cognitive impairment, and normal individuals over a period of 18 months (3). In addition, all familial forms of AD identified so far are caused by mutations in the amyloid precursor protein (APP) or in presenilin proteins resulting in increased brain Aβ accumulation and alteration of Aβ production (4, 5), suggesting that Aβ plays an important role in the development of AD. Interestingly, a coding mutation found in the APP gene resulting in a reduction of Aβ production is protective against AD and cognitive decline in the elderly (6, 7) further supporting the view that Aβ is a key culprit in the pathobiology of AD. Aβ is generated by the sequential proteolytic cleavage of APP by β- and γ-secretase. The γ-secretase cut is imprecise and leads to the formation of Aβ peptides of different size, mainly Aβ38, Aβ40, and Aβ42. Soluble Aβ oligomers are now widely recognized as pathogenic because they have been shown to inhibit synaptic function, to cause synaptic degeneration, to induce Tau hyperphosphorylation, and to impair memory (8,–11).
The Tau pathology in AD is also the subject of intensive research, and many aberrantly hyperphosphorylated sites on Tau have been identified in AD. Tau is known to bind to microtubules and to promote their polymerization and stabilization. Tau hyperphosphorylation results in its dissociation from microtubules, which affects microtubule-dependent axonal transport and promotes Tau oligomerization and aggregation, eventually leading to neurodegeneration. Several kinases responsible for Tau hyperphosphorylation have been identified; among them cyclin-dependent kinase 5 (Cdk5) and glycogen synthase kinase 3 (GSK3) are considered important candidates responsible for pathological Tau phosphorylation in AD (12,–14). Cdk5 has been shown to indirectly affect Tau hyperphosphorylation by regulating GSK3β activity establishing GSK3β as a key mediator of Tau phosphorylation at disease-associated sites (12). GSK3β appears to be a pivotal enzyme in AD as it regulates Aβ production and mediates pathological Tau hyperphosphorylation induced by Aβ (15,–17).
We have shown previously that a subset of L-type calcium channel blockers (CCB) belonging to the class of antihypertensive dihydropyridines display some Aβ-lowering properties by affecting both the processing of APP and the clearance of Aβ across the BBB (18, 19), properties that are unrelated to the inhibition of calcium channels or to their anti-hypertensive activity (18). Among the dihydropyridine CCB tested, we identified nilvadipine for its ability to readily cross the BBB, to lower brain Aβ levels, to improve cognition, and to prevent cerebral blood flow deficits in a transgenic mouse model of AD overexpressing Aβ (18, 20). Nilvadipine is a clinically used antihypertensive that we have shown to be well tolerated and to stabilize cognition in AD patients compared with untreated patients (21, 22). Interestingly, long term use of nilvadipine in subjects with mild cognitive impairment has also been shown to prevent cognitive decline and to reduce the incidence of AD conversion (23) suggesting that nilvadipine may have disease-modifying benefits. Nilvadipine consists of a mixture of two enantiomers, (+)-nilvadipine and (−)-nilvadipine, in equal proportion. (+)-Nilvadipine is known to be the active isomer responsible for the CCB/anti-hypertensive properties of nilvadipine, whereas (−)-nilvadipine is a much weaker CCB (24,–28). As nilvadipine is an anti-hypertensive compound, the dosage of nilvadipine in AD patients may be limited as nilvadipine may induce an unwanted drop in blood pressure in these patients possibly resulting in adverse events. For these reasons, we explored the potential Aβ-lowering properties of (−)-nilvadipine and (+)-nilvadipine in vitro using a cell line overexpressing Aβ. We tested the impact of (+)- and (−)-nilvadipine on the clearance of Aβ peptides across the blood-brain barrier (BBB) in vitro and in vivo because we have shown previously that racemic nilvadipine stimulates the BBB clearance of Aβ (18). Additionally, we evaluated the Aβ-lowering properties of (+) and (−)-nilvadipine in vivo following an acute treatment in Tg PS1/APPsw mice. We also tested the impact of (−)-nilvadipine on Tau phosphorylation using a transgenic mouse model of tauopathy. Finally, we searched for a possible mechanism of action of nilvadipine that can explain the impact of nilvadipine on Aβ and Tau phosphorylation.
Racemic nilvadipine, (−)-nilvadipine, and (+)-nilvadipine were synthesized as described previously (26) and were obtained from Archer Pharmaceuticals. BAY61-3606, phorbol 12-myristate 13-acetate (PMA), human recombinant TNF-α, dimethyl sulfoxide (DMSO), 2-mercaptoethanol, imidazole, sodium chloride, and phenylmethylsulfonyl fluoride (PMSF) were purchased from Sigma. KT5270 was obtained from R&D Systems. All antibiotics, fungizone, PBS, culture media, and fetal bovine serum were purchased from Invitrogen. Fluorescein-Aβ(1–42) was purchased from rPeptide (Bogart, GA). The MPER reagent and the protease/phosphatase inhibitors mixture were purchased from Thermo Fisher Scientific. Guanidine hydrochloride was obtained from EMD Millipore.
7W CHO (29) cells stably transfected with human APP751 were maintained in DMEM containing 10% fetal bovine serum, 1× mixture of penicillin/streptomycin/fungizone mixture, and 0.3% geneticin as a selecting agent. Cells were cultured in 96-well culture plates and treated for 24 h with a dose range of BAY61-3606 (0.5, 1, 5, and 10 μm), a dose range of (−)-nilvadipine (1, 5, 10, and 20 μm), (+)-nilvadipine, and a racemic mixture of nilvadipine consisting of an equal amount of (+)- and (−)-nilvadipine. Potential cytotoxicity of the different treatments was routinely evaluated using the cytotoxicity detection kit (Roche Diagnostics), and no significant toxicity was observed for the different treatments (data not shown). Following the treatments with nilvadipine enantiomers, Aβ40 and Aβ42 were analyzed in the culture medium by using commercially available sandwich ELISAs (Invitrogen) according to recommendations of the manufacturer. Following the treatments with a dose range of BAY61-3606, Aβ38, Aβ40, and Aβ42 were quantified by electrochemiluminescence using multiplex Aβ assays according to the manufacturer's recommendations (Meso Scale Discovery, MD). All experiments were performed at least in quadruplicate for each treatment dose. Additionally, sAPPα was detected by Western blot in the culture medium surrounding 7W CHO cells using the antibody 6E10 (Signet Laboratories Inc.), which recognizes amino acids 1–17 of Aβ, and sAPPβ was detected in the culture medium using an anti-human sAPPβ antibody (Immuno-Biological Laboratories Co. Ltd., Gunma, Japan) as we described previously (30).
The in vitro model of the BBB consisting of a polarized human brain microvascular endothelial cell monolayer grown on cell culture inserts that separate into apical (“blood”) and basolateral (“brain”) compartments was established as described previously by our group (18, 19, 38,–41). Aβ exchange dynamics across the BBB model were examined using a fluorometric Aβ42 assay as we described previously (18, 19, 31,–34). Briefly, the apical (receiver) side of the membrane was exposed to various concentrations of racemic nilvadipine, (−)-nilvadipine, (+)-nilvadipine, or BAY61-3606. The donor compartment was sampled at time 0 to establish the initial concentration of fluorescein-labeled Aβ(1–42) (FlAβ(1–42)) in each group. Following exposure of the insert to the well containing FlAβ(1–42), samples were collected from the apical compartment at various time points up to 90 min to assess the movement of FlAβ(1–42) across the human brain microvascular endothelial cell monolayer (basolateral-to-apical). The samples were analyzed (λex = 485 nm and λem = 516 nm) for FlAβ(1–42) using a BioTek Synergy HT multidetection microplate reader (Winooski, VT). The apparent permeability (Papp) of FlAβ(1–42) was determined using the following equation: Papp = 1/AC0·(dQ/dt), where A represents the surface area of the membrane; C0 is the initial concentration of FlAβ(1–42) in the basolateral compartment, and dQ/dt is the amount of FlAβ(1–42) appearing in the apical compartment in the given time period. The apparent permeability of FlAβ(1–42) in the presence of drug was compared with control (i.e. no drug exposure) and expressed as a percentage. We corrected for permeability resistance associated with the blank membrane as reported previously (31).
Mice were maintained under specific pathogen-free conditions in ventilated racks at the Association for Assessment and Accreditation of Laboratory Animal Care International accredited vivarium of the Roskamp Institute. All the experimentations involving mice were reviewed and approved by the Institutional Animal Care and Use Committee of the Roskamp Institute before implementation and were conducted in compliance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.
(−)-Nilvadipine and (+)-nilvadipine were administered intraperitoneally to 26-week-old Tg PS1/APPsw mice (35) for 4 consecutive days at a dosage of 2 and 4 mg/kg of body weight. (−)-Nilvadipine and (+)-nilvadipine were dissolved in 50% DMSO/PBS immediately before the injection. Vehicle-treated Tg PS1/APPsw mice received an intraperitoneal injection of 50% DMSO in PBS for 4 days. Within 1 h after the last injection, animals were humanely euthanized, and their brains (without cerebella) were snap-frozen in liquid nitrogen until analyzed for Aβ40 and Aβ42 levels using commercially available ELISA kits (Invitrogen). Briefly, brains were homogenized at 4 °C in MPER reagent (Thermo Fisher Scientific Inc.) containing 1 mm PMSF and 1× protease/phosphatase inhibitors mixture (Thermo Fisher Scientific Inc.) and centrifuged at 15,000 × g for 30 min at 4 °C. The supernatant-containing detergent-solubilized Aβ was collected and denatured in 5 m guanidine for 1 h before being diluted with the sample diluent provided in the Aβ ELISA kits and assessed for Aβ40 and Aβ42 according to the manufacturer's protocol. The pellet containing detergent-insoluble material was resuspended in 5 m guanidine and diluted with the sample diluent provided in the Aβ ELISA kits before assaying for Aβ. Protein concentrations were measured in the guanidine extracts using the BCA method (Invitrogen), and results of brain Aβ40 and Aβ42 were calculated in picograms/mg of protein.
BAY61-3606 was administered intraperitoneally to 28-week-old Tg PS1/APPsw mice for 4 consecutive days at a dosage of 2 mg/kg of body weight. BAY61-3606 was dissolved in sterile PBS. Vehicle-treated Tg PS1/APPsw mice received an intraperitoneal injection of PBS for 4 days. Mice were humanely euthanized 30 min after the last injection, and their brains (without cerebella) were snap-frozen in liquid nitrogen until analysis for Aβ38, Aβ40, and Aβ42 levels by electrochemiluminescence using multiplex Aβ assays according to the manufacturer's recommendations (Meso Scale Discovery). Brain homogenates for the quantification of Aβ peptides were prepared as described above.
Short hairpin RNAs (shRNAs) were used to stably knockdown SYK gene expression in HEK-NFκB luciferase reporter cells (Panomics, CA). SYK shRNAs cloned into the GIPZ lentiviral vector were purchased from Origene (Rockville, MD). Five different SYK shRNAs and a scrambled control shRNA were used separately for transfection. Approximately 1 million HEK-NFκB luciferase reporter cells detached with TripLE (Invitrogen) were mixed together with 10 μg of shRNA in genepulser cuvettes (0.4 cm) (Bio-Rad) and electroporated using a square wave protocol, 110 V, one pulse length of 25 ms using a Bio-Rad gene pulser, and were seeded in 6-well plates. 48 h later, the cell culture media were replaced by selective media containing 6 mg/ml puromycin for stable selection of transfected cells. Single colonies resistant to puromycin were isolated. Western blots were run with cell lysates from several of these clones, including the control shRNA (control) to confirm silencing of the SYK gene using the anti-Syk antibody from Santa Cruz Biotechnology (1:500 dilution). The clones (83D04, 38G10, and 69F04) with the most profound silencing were chosen to evaluate the effect of Syk deficiency on NFκB luciferase activity. Briefly, NFκB activation was triggered with 10 ng/ml PMA for 24 h, and NFκB luciferase activity was quantified by chemiluminescence using the Luc-Screen Extended Glow from Applied Biosystems with a Synergy HT chemiluminescence reader (Biotek) as we described previously (36, 37). Similarly, a clone of 7W CHO cells overexpressing APP stably transfected with SYK shRNAs to knock down Syk expression was selected. Functional inactivation of Syk activity in 7W CHO cells was characterized by quantifying RAF phosphorylation (Ser-338) in response to PMA in regular 7W CHO cells and in 7W CHO cells transfected with SYK shRNAs by Western blot using a phospho-RAF (56A6) rabbit antibody (Cell Signaling Technology).
The plasmid pDW445 (38) containing the BirA biotin-protein ligase gene was obtained from Addgene and digested with BamHI and HindIII restriction enzymes to remove the BirA gene. The BirA gene was then ligated into pBac-1 (EMD Millipore), which had been digested with the same restriction enzymes to give pBacBirA. The human SYK gene was synthesized and had added to the 5′ end a BamHI restriction site followed by the ATG start codon. A Gly-4–Ser linker followed by an AviTag (Avidity) GLNDIFEAQKIEWHE (which is biotinylated on the Lys by the BirA enzyme (39)) and a His6 tag followed by a stop codon and an EcoRI restriction site (GenScript) were added to the 3′ end of the human SYK gene. The synthesized gene (SykA6H) was blunt-end cloned into pUC57 using an EcoRV site to give pSykAvi6H; the gene sequence was verified by DNA sequencing (GenScript). The pSykAvi6H plasmid was digested with BamHI and EcoRI, and the SykAvi6H gene was ligated into pBac-1 that had been digested with the same restriction enzymes to give pBacSykAvi6H. Both pBacBirA and pBacSykAvi6H were transfected into Sf9 insect cells as described previously (39) with the following modification: the flashBAC plasmid (Oxford Expression Technologies, Inc.) was used to generate the recombinant baculoviruses fBAC-BirA and fBAC-SykAvi6H. The recombinant baculoviruses were amplified and used to infect Sf9 insect cells grown in shaker flasks at 27 °C with GlutaMAX media (Invitrogen) supplemented with 100 μm biotin.
The fBAC-BirA and fBAC-SykAvi6H co-infected insect cells were harvested after 60–70 h and centrifuged at 2,000 × g. The pellet was then resuspended in ice-cold lysis buffer (50 mm sodium phosphate, pH 7.4, 100 mm NaCl, 0.5% Nonidet P-40, 5 mm 2-mercaptoethanol, and complete EDTA-free protease inhibitor mixture (Roche Applied Science and Invitrogen)) and then incubated on ice for 15 min with gentle mixing every 3 min to keep cells suspended. The sample was then centrifuged at 6,000 × g for 10 min at 5 °C, and the supernatant was removed, placed in a new tube, and centrifuged at 15,000 × g for 10 min at 5 °C. The supernatant was collected and placed in a new tube on ice. The crude lysate was batch-bound to Talon Superflow resin (GE Healthcare) with inversion for 1 h at 5 °C, batched washed with Buffer A (50 mm sodium phosphate, pH 7.4, 300 mm sodium chloride), supplemented with 8 mm imidazole, and then poured into an FPLC column. The column was washed with Buffer A and 8 mm imidazole, and then the imidazole concentration was increased to 150 mm to elute SykAvi6H. The column fractions were immediately made with 5 mm EDTA and then 50% glycerol and stored at −20 °C. SykAvi6H was incubated in 50 mm sodium phosphate, pH 7.4, 50 mm sodium chloride at 30 °C for 30 min and found to remain at full length (data not shown). The SykAvi6H was >95% biotinylated as determined using streptavidin-agarose resin (Thermo Fisher Scientific Inc., IL) in a pulldown assay (data not shown).
The Syk enzymatic reactions were performed using human full-length recombinant Syk prepared as described above and the Omnia Y peptide 7 kit following the recommendations of the manufacturer (Invitrogen). Enzymatic reactions were carried out at 30 °C using a Biotek Synergy HT plate reader. Fluorescence intensity readings (λex = 360 nm/λem = 485) were collected every 39 s for 40 min.
A fortéBio Octet Red platform (fortéBio Corp., Menlo Park, CA) was used to study the binding kinetics of (−)-nilvadipine and BAY61-3606 to immobilized human full-length SYK that was biotinylated in situ as described above. All the binding assays were performed at 30 °C with an agitation set at 1,000 rpm using solid black 96-well plates (Grieger Bio-One) in fortéBio kinetic buffer to minimize nonspecific interactions. The final volume for all the solutions was 200 μl/well. Human biotinylated SYK was loaded on the surface of super-streptavidin biosensors (fortéBio Corp.). SYK biosensors were washed several times in kinetic buffer prior to performing the binding experiments. Reference super-streptavidin biosensors loaded with biotinylated PEG4 (Thermo Fisher Scientific Inc.) were used to correct for nonspecific binding of (−)-nilvadipine and BAY61-3606 to the biosensors. For (−)-nilvadipine and BAY61-3606, the association step was performed for 120 s, the dissociation step for 180 s followed by an additional wash step of 30 s in kinetic buffer. SYK-loaded biosensors and reference sensors were gradually subjected to the same dose range of (−)-nilvadipine (2.5, 5, 10, 20, 40, and 80 μm) or BAY61-3606 (0.625, 1.25, 2.5, 5, and 10 μm). Data analysis and curve fitting were done using the Octet software version 6.4 (fortéBio Corp.). Steady-state kinetic analyses were performed for every data set to calculate the KD value using the estimated response at the equilibrium for each concentration of (−)-nilvadipine and BAY61-3606. An average KD value (± S.E.) representative of six experiments with (−)-nilvadipine and BAY61-3606 was calculated.
Tg Tau P301S mice (line PS19) (40) were purchased from the Jackson Laboratory. (−)-Nilvadipine was administered intraperitoneally at a dosage of 2 mg/kg of body weight to 25-week-old Tg Tau P301S mice for 10 consecutive days. Vehicle-treated Tg Tau P301S received an intraperitoneal injection of 50% DMSO in PBS (vehicle used to dissolve (−)-nilvadipine) for 10 consecutive days. Tau phosphorylation was quantified by Western blots using the following Tau antibodies: AT8 (Thermo Fisher Scientific Inc., IL) and PHF-1 (41) as we described previously (42). BAY61-3606 (dissolved in PBS) was administered intraperitoneally at a dosage of 2 mg/kg of body weight to 25-week-old Tg Tau P301S mice for 5 consecutive days. Tau phosphorylation at multiple epitopes was quantified in brain homogenates by dot blots as we described previously (42) using PHF-1, RZ3, CP13 (48), and 9G3 (MédiMabs, Canada) antibodies.
SH-SY5Y cells were purchased from American Type Culture Collection (Manassas, VA). SH-SY5Y cells were grown in DMEM/F-12 medium supplemented with 10% serum and 1% penicillin/streptomycin/fungizone. BACE-1 mRNA quantifications were performed by real time quantitative RT-PCR using the housekeeping gene GAPDH as a reference mRNA, and BACE-1 protein expression was analyzed by Western blots as we described previously (36, 37). SH-SY5Y cells were treated with a dose range of BAY61-3606 for 30 min and 24 h to assess the impact of Syk inhibition on AKT phosphorylation and GSK3β Ser-9 phosphorylation by Western blots using phospho-selective Ser-9 GSK3β and phospho-Ser-473 AKT antibodies (Cell Signaling Technology) at 1:1,000 dilutions. The effects of BAY61-3606 on Tau phosphorylation at multiple AD-relevant epitopes (9G3 (phospho-Tyr-18), PHF-1, RZ3, and CP13) were studied by Western blot as we described previously (42). The involvement of PKA in mediating the stimulation of Ser-9 phosphorylation of GSK3β was tested by using the selective PKA inhibitor KT5270 (Tocris, CA). SH-SY5Y cells were pretreated with KT5270 for 5 min prior to being challenged with 10 μm BAY61-3606 for a period of 30 min. The impact of these treatments on Ser-9 GSK3β phosphorylation and on CREB phosphorylation (Ser-133) was followed by Western blots using phospho-selective antibodies from Cell Signaling Technology at 1:1,000 dilutions.
Results are expressed as mean ± S.E. Statistical analyses were performed using SPSS Version 12.0.1 for Windows. Data were examined for assumption of normality using the Shapiro-Wilk statistic and for homogeneity of variance using the Levene's test. Statistical significance was determined by Student's t test, univariate, or repeated measures analysis of variance (ANOVA) where appropriate followed by post hoc comparisons with the Bonferroni method. For data not satisfying assumptions of normality and homogeneity of variance, a nonparametric Mann-Whitney test was used. p values < 0.05 were considered significant.
The impact of a dose range of (−)-nilvadipine and (+)-nilvadipine and a racemic mixture of nilvadipine containing an equal amount of each enantiomer on Aβ production was tested in vitro using 7W CHO cells overproducing Aβ. Following 24 h of treatment with the pure enantiomers or the racemic mixture of nilvadipine, a dose-dependent inhibition of Aβ production was observed (Fig. 1A). We found overall that the (+)- and (−)-enantiomers or the racemic mixture of nilvadipine have similar effects on Aβ production in vitro. We have previously shown that racemic nilvadipine affects the β-cleavage of APP and reduces sAPPβ secretion (18). Analyses of sAPPβ secretion in the culture media surrounding 7W CHO cells reveal that (−)-nilvadipine inhibits sAPPβ secretion suggesting an inhibition of the β-cleavage of APP (Fig. 1B). We found that (−)-nilvadipine was, however, unable to directly affect BACE-1 activity using a cell-free assay (data not shown). We therefore evaluated the possible impact of racemic nilvadipine and (−)-nilvadipine on BACE-1 expression. Tumor necrosis factor-α (TNFα) has been shown to induce BACE-1 expression and to contribute to brain accumulation of Aβ peptides (43). We therefore tested the possible impact of (−)-nilvadipine and racemic nilvadipine on TNFα-induced BACE-1 transcription in human neuron-like SH-SY5Y cells. We found that both (−)-nilvadipine and racemic nilvadipine reduce BACE-1 mRNA expression (Fig. 1C) induced by TNFα. In addition, a reduction in BACE-1 protein levels was observed following treatment of HEK293 cells with (−)-nilvadipine or racemic nilvadipine (Fig. 1D) further suggesting that the inhibition of Aβ production observed following nilvadipine treatment is mediated in part by a reduction of BACE-1 expression.
Next, we assessed the impact of a dose range of nilvadipine enantiomers on Aβ transcytosis across the BBB using an in vitro model employing human brain microvascular endothelial cells that we have abundantly characterized (18, 19, 31,–34) since we have shown previously that racemic nilvadipine increases the clearance of Aβ across the BBB (18). We found that both the (+)- and (−)-nilvadipine enantiomers enhance Aβ42 clearance from the brain to the peripheral side of the in vitro BBB model (Fig. 2A). Altogether, these data show that (−)-nilvadipine retains similar Aβ-lowering properties compared with (+)-nilvadipine or racemic nilvadipine, and we further demonstrate that the Aβ-lowering properties of nilvadipine are not related to a blockade of L-type calcium channels.
We tested the effect of an acute treatment with (−)-nilvadipine or (+)-nilvadipine on brain Aβ levels using Tg PS1/APPsw mice, and we observed that both (−)-nilvadipine and (+)-nilvadipine acutely reduced brain Aβ levels with similar potency (Fig. 2, C and D). We also evaluated the impact of (−)-nilvadipine on the clearance of human Aβ42 across the blood-brain barrier using wild-type mice as we described previously (18). Wild-type mice were intracranially injected with human Aβ42, and the appearance of human Aβ42 was followed in the blood of the animals. Data show that (−)-nilvadipine stimulated the clearance of human Aβ42 across the BBB as more human Aβ42 was detected in the plasma of (−)-nilvadipine-treated mice than control animals (Fig. 2B).
We explored the possible impact of (−)-nilvadipine on Tau phosphorylation using Tg Tau P301S mice because this model of Tauopathy displays Tau hyperphosphorylation at multiple AD relevant epitopes (40, 42). Tg Tau P301S mice were treated for 10 days with a daily intraperitoneal injection of (−)-nilvadipine (2 mg/kg) or a vehicle. Western blot analyses of brain homogenates show that (−)-nilvadipine significantly reduces Tau phosphorylation in AT8 (phosphorylated Ser-199/Ser-202/Thr-205) and PHF-1 (phosphorylated Ser-396/Ser-404) epitopes (Fig. 3).
Overall, our data show that the Aβ-lowering properties of (−)-nilvadipine are similar to racemic nilvadipine and (+)-nilvadipine revealing that the Aβ inhibition observed following nilvadipine treatment is not due to an inhibition of the L-type calcium channel and suggesting that another target is responsible for the Aβ-lowering properties of nilvadipine. This prompted us to explore in further detail the mechanisms responsible for the Aβ-lowering properties of nilvadipine and to identify a possible molecular target controlling both Aβ production, Aβ clearance across the BBB, and Tau hyperphosphorylation. As (−)-nilvadipine and racemic nilvadipine inhibit BACE-1 transcription, we evaluated whether (−)-nilvadipine was impacting NFκB activation because NFκB has been shown to play an important role in the regulation of BACE-1 transcription and expression (36, 37, 43, 44). We observed that (−)-nilvadipine inhibits NFκB activation in response to TNFα (Fig. 4A) or phorbol 12-myristate 13-acetate (PMA) (Fig. 4B) by using an NFκB-luciferase reporter cell line to monitor NFκB activation, thus suggesting a possible mechanism responsible for the inhibition of BACE-1 transcription following nilvadipine treatment. The signaling pathways responsible for the activation of NFκB by TNFα or PMA have been described in the literature, and we explored whether elements of these pathways were impacted by (−)-nilvadipine. For instance, the small GTPase Rho and its downstream effector Rho-associated coiled-coil containing protein kinase (ROCK) have been shown to contribute to TNFα induction of NFκB activation (45). We found that nilvadipine was ineffective at inhibiting ROCK activity in a cell-free assay excluding ROCK as possible target of nilvadipine (data not shown). In addition, we observed that Rho inhibition with C3 cell-permeable transferase (CT04) and ROCK inhibition with Y27632 were unable to prevent the inhibition of NFκB activation induced by (−)-nilvadipine and did not significantly impact BACE-1 expression ruling out the possible involvement of Rho/ROCK for mediating the Aβ-lowering properties of (−)-nilvadipine (data not shown). PMA is a known agonist of PKC, which leads to the activation of the PKC/RAS/RAF/MEK/MAPK pathway that ultimately induces NFκB activation (46,–48). Therefore, we evaluated whether nilvadipine could inhibit some members of this signaling pathway. In particular, we monitored RAF phosphorylation following treatment with (−)-nilvadipine and observed that (−)-nilvadipine prevents RAF phosphorylation induced by PMA (Fig. 4, C and D) suggesting that (−)-nilvadipine is impacting a target upstream of RAF. We tested the possible effect of (−)-nilvadipine on RAS activity and found that (−)-nilvadipine was unable to affect RAS activity in a cell-free assay ruling out RAS as a possible target of (−)-nilvadipine (data not shown). Tyrosine kinases, including Syk and Bruton's tyrosine kinase (BTK), are activated following PMA treatment (49, 50), act upstream of RAS/RAF (51, 52), and mediate the activation of the NFκB pathway (53). We tested a selective BTK inhibitor (BTK inhibitor III, 1-(3-(4-amino-3-(4-phenyloxyphenyl)-1H-pyrazolo[3,4-d]pyrimidin-1-yl)piperidin-1-yl)prop-2-en-1-one, N-acryloyl-(3-(4-amino-3-(4-phenyloxyphenyl)-1H-pyrazolo[3,4-d]pyrimidin-1-yl)piperidine) on Aβ production, and we found that this compound was unable to significantly inhibit Aβ production (data not shown) suggesting that the Aβ-lowering properties of nilvadipine are not mediated via an inhibition of BTK. We explored whether Syk activity was impacted by (−)-nilvadipine as nilvadipine reduces RAF phosphorylation. Using a cell-free assay using human recombinant Syk, we observed that (−)-nilvadipine dose-dependently inhibits Syk activity (Fig. 5A). The Syk inhibitor BAY61-3606 was used as a positive control in the Syk activity assay, and a dose-dependent inhibition of Syk activity was observed with BAY61-3606 as expected (Fig. 5B). To ensure that the reduction in Syk activity observed was not due to an interaction of the peptide substrate with (−)-nilvadipine, we also verified that (−)-nilvadipine was able to directly bind to Syk. We measured the binding affinity of (−)-nilvadipine for Syk using biolayer interferometry and confirmed that (−)-nilvadipine binds to human recombinant Syk with a binding dissociation constant (KD) of 2.1 μm (Fig. 5C), further suggesting that Syk is the possible target impacted by nilvadipine. Binding kinetics were also performed using BAY61-3606 as a positive control (Fig. 5D), and a KD of 0.4 μm was obtained for BAY61-3606. In addition, we verified that pharmacological inhibition of Syk with BAY61-3606 resulted in a blockade of NFκB activation and that genetic down-regulation of SYK using shRNA also prevented NFκB activation (data not shown) thus highlighting Syk as a key player of NFκB activation. The magnitude of the NFκB inhibition following (−)-nilvadipine treatment was also reduced in clones of HEK293 NFκB luciferase reporter cells in which Syk expression had been silenced (data not shown) further suggesting that Syk is required to mediate the inhibition of NFκB activity induced by (−)-nilvadipine.
As (−)-nilvadipine affects both Aβ production and Aβ clearance across the BBB, we investigated whether pharmacological inhibition of Syk could also affect these two processes. We found that Syk inhibition with the selective Syk inhibitor BAY61-3606 suppresses Aβ production in 7W CHO cells overexpressing APP (Fig. 6A). We also tested other Syk inhibitors with a chemical structure distinct from BAY61-3606 (3-(1-methyl-1H-indol-3-yl-methylene)-2-oxo-2,3-dihydro-1H-indole-5-sulfonamide and 2-(2-aminoethylamino)-4-(3-trifluoromethylanilino)-pyrimidine-5-carboxamide), and we observed that these compounds inhibited Aβ production in vitro (data not shown). We found that, like (−)-nilvadipine, Syk inhibition with the selective Syk inhibitor BAY61-3606 resulted in decreased sAPPβ secretion, BACE-1 mRNA, and BACE-1 protein expression (data not shown). To further demonstrate the involvement of Syk in the regulation of Aβ production, we suppressed syk expression in 7W CHO cells overexpressing Aβ using shRNA. To verify that Syk activity has been functionally suppressed in 7W CHO cells transfected with SYK shRNA, we monitored RAF phosphorylation. As expected, RAF phosphorylation induced by PMA as well as basal RAF phosphorylation were reduced in 7W CHO cells transfected with SYK shRNA confirming a reduction in Syk activity (Fig. 6B). Interestingly, Aβ production in 7W CHO cells transfected with SYK shRNA compared with 7W CHO cells (Fig. 6C) was significantly reduced, further demonstrating the involvement of Syk in the regulation of Aβ production.
Because we have shown that (−)-nilvadipine improves the clearance of Aβ across the BBB, we also explored whether Syk inhibition could affect the transport of Aβ across the BBB. We show that the selective Syk inhibitor BAY61-3606 stimulates the transport of Aβ across the BBB in vitro mimicking the biological activity of (−)-nilvadipine in this model (Fig. 7A). We tested other Syk inhibitors of unrelated chemical structures in our in vitro BBB model and observed that all these Syk inhibitors stimulated the clearance of Aβ from the basolateral to apical side of the BBB (data not shown) further demonstrating the involvement of Syk in the clearance of Aβ across the BBB.
To further validate Syk as a potential target for regulating brain Aβ levels and Aβ clearance across the BBB, we investigated the impact of the selective Syk inhibitor BAY61-3606 on these different processes in vivo. BAY61-3606 was selected over other Syk inhibitors because it has been shown previously to inhibit Syk activity in vivo in rodents (54, 55). We observed that BAY61-3606 significantly reduces brain Aβ38, Aβ40, and Aβ42 levels in Tg PS1/APPsw mice (Fig. 7, C and D). In addition, we found that BAY61-3606 stimulates the clearance of Aβ across the BBB in wild-type mice as demonstrated by increased circulating plasma levels of human Aβ42 in mice treated with the Syk inhibitor compared with vehicle-treated mice following the intracranial injection of human Aβ42 (Fig. 7B).
Syk has been shown to mediate the phosphorylation of Tau at Tyr-18 (56). We evaluated the possible effects of the selective Syk inhibitor BAY61-3606 on Tau phosphorylation in Tg Tau P301S mice using a dot-blot approach that we have validated for monitoring Tau phosphorylation in P301S mice (42). We tested the impact of BAY61-3606 on Tau phosphorylation at Tyr-18 and also investigated whether other Tau phosphoepitopes were affected by the treatment. We observed a reduction in Tau phosphorylation at the Tyr-18 epitope as expected (Fig. 8) in BAY61-3606-treated mice. Interestingly, we also detected a reduction in Tau phosphorylation at PHF-1 (Ser(P)-396/Ser(P)-404) and CP13 (Ser(P)-202) in epitopes following treatment of Tg Tau P301S mice with BAY61-3606, whereas the RZ3 (Thr(P)-231) Tau epitope was not significantly impacted (Fig. 8) suggesting that Syk inhibition may also control the activity of other downstream kinases involved in Tau hyperphosphorylation.
We conducted some mechanistic studies in human neuron-like SH-SY5Y cells to delineate a possible mechanism of action responsible for the reduction of Tau phosphorylation observed at multiple Tau epitopes following Syk inhibition. Because we observed a reduction of Tau phosphorylation at the characteristic GSK3β motif (Ser(P)-396/Ser(P)-404 (PHF-1)) following treatment of Tg Tau P301S mice with BAY61-3606, we investigated whether pharmacological inhibition of Syk could affect GSK3β by monitoring the phosphorylation of GSK3β at Ser-9, which is known to inactivate the enzyme (57). We observed that pharmacological inhibition of Syk with BAY61-3606 stimulates Ser-9 phosphorylation of GSK3β in SH-SY5Y cells (Fig. 9, A and B) suggesting that blocking Syk activity results in GSK3β inhibition. We further confirmed that possibility by showing that Tau phosphorylation at the typical GSK3β sites (PHF-1 and CP13) is reduced following treatment of SH-SY5Y cells with BAY61-3606, whereas Tau phosphorylation at the RZ3 site was not significantly impacted in SH-SY5Y cells (Fig. 9C). GSK3β phosphorylation at Ser-9 has been shown to be mediated by PKA and AKT (protein kinase B) (57, 58). We therefore examined whether Syk inhibition was affecting AKT and PKA. We found that treatment of SH-SY5Y cells with BAY61-3606 inhibits AKT phosphorylation (Fig. 9, A and B), which is consistent with previous studies (59) investigating the impact of Syk inhibition on AKT activation. Our data therefore show that AKT cannot mediate the increased GSK3β Ser-9 phosphorylation induced by Syk inhibition in SH-SY5Y cells. We then explored whether a selective PKA inhibitor (KT5270) was able to mitigate GSK3β Ser-9 phosphorylation induced by treatment of SH-SY5Y cells with BAY61-3606. We found that KT5270 effectively suppressed GSK3β phosphorylation at Ser-9 induced by BAY61-3606 (Fig. 10B) suggesting that this event is mediated by an activation of PKA. PKA is a known substrate of Syk, and it has been shown that Syk inhibits PKA activity by phosphorylating Tyr-330 of the PKA catalytic subunit (60), further supporting our observation. To verify that Syk inhibition with BAY61-3606 resulted in PKA activation, we evaluated the phosphorylation of CREB, a direct PKA substrate, following pharmacological inhibition of Syk. We found that Syk inhibition with BAY61-3606 induced CREB phosphorylation, although that event is inhibited in the presence of a selective PKA inhibitor (Fig. 10, A and B) further showing that Syk inhibition results in PKA activation.
Hypertension in mid-life has been associated with an increased risk of dementia, including AD in the elderly (61). Interestingly, the double-blind, placebo-controlled systolic hypertension study (Syst-Eur) has revealed that the long term use of a dihydropyridine CCB (nitrendipine) reduces the risk of dementia by 55%, including AD (62, 63), suggesting that antihypertensive therapies may be beneficial against dementia. Chronic treatment with nilvadipine has been shown to prevent the conversion of mild cognitive impairment to AD (23) suggesting that nilvadipine may have disease-modifying properties. In addition, a short treatment duration with nilvadipine in AD patients has been shown to reduce cognitive decline (21, 22).
In this study, we investigated the Aβ-lowering properties of nilvadipine enantiomers, (+)-nilvadipine and (−)-nilvadipine. (+)-Nilvadipine is the enantiomer of racemic nilvadipine responsible for the anti-hypertensive activity of the compound (24,–26). We confirmed ex vivo in rat aortae that (−)-nilvadipine was ~1,000 time less potent than (+)-nilvadipine for opposing the vasoconstriction induced by a selective L-type calcium channel agonist (FPL64176) showing that it is a much weaker L-type calcium channel antagonist than (+)-nilvadipine (data not shown). Interestingly, we observed that both enantiomers are equipotent at lowering Aβ production in a cell line overexpressing Aβ thus showing that the Aβ-lowering properties of nilvadipine are not related to L-type calcium channel inhibition. We have shown previously that a subset of CCB dihydropyridines prevent brain Aβ accumulation by opposing Aβ production and by stimulating the clearance of Aβ across the BBB (18). Therefore, we also investigated the impact of nilvadipine enantiomers on the transport of Aβ across the BBB to determine whether the two enantiomers also retain the ability to increase the clearance of Aβ across the BBB. Our data show that (−)-nilvadipine also stimulates the clearance of Aβ across the BBB both in vitro and in vivo duplicating the properties of the racemic mixture and further substantiating the fact that the Aβ-lowering properties of nilvadipine are not related to L-type calcium channel inhibition. As (−)-nilvadipine retains the anti-amyloidogenic properties of the racemic mixture but is deprived of anti-hypertensive activity, (−)-nilvadipine may represent a new lead compound for the treatment of AD. Interestingly, we found that (−)-nilvadipine also reduces Tau phosphorylation at AD-pertinent epitopes (AT8 and PHF-1) in a mouse model of tauopathy (Tg Tau P301S mice) suggesting that (−)-nilvadipine may also be beneficial for modulating Tau phosphorylation in AD.
The fact that (−)-nilvadipine affects pathological Tau phosphorylation and Aβ accumulation prompted us to further explore the mechanism of action of (−)-nilvadipine. Our search for the mechanism responsible for the Aβ-lowering properties of nilvadipine led to the identification of Syk as a possible target responsible for mediating the effects of (−)-nilvadipine on Aβ and Tau phosphorylation. We show that nilvadipine enantiomers inhibit Syk activity in a cell-free assay, whereas analysis of binding kinetics by biolayer interferometry shows that nilvadipine enantiomers bind to full-length human Syk with relatively high affinity (KD 2.1 μm) identifying Syk as a possible nilvadipine target responsible for its Aβ-lowering properties. We further confirmed this possibility by showing that pharmacological inhibition of Syk in vitro and in vivo as well as down-regulation of syk expression result in a reduction of Aβ production. Moreover, Syk inhibition reduces sAPPβ production and BACE-1 transcription mimicking the effect of (−)-nilvadipine. In addition, we show that Syk inhibition increases the clearance of Aβ across the BBB recapitulating to the full extent the Aβ-lowering properties of (−)-nilvadipine. Our data therefore identify Syk as an important regulator of Aβ production both in vitro and in vivo. We show additionally that (−)-nilvadipine, as well as a selective Syk inhibitor, or a reduction of syk expression using shRNA all prevent NFκB activation suggesting a possible mechanism for the reduction of BACE-1 transcription observed following Syk inhibition as NFκB has been shown previously to play an important role in the regulation of BACE-1 transcription and expression (36, 37, 44, 64). Because NFκB is involved in the regulation of neuroinflammation and cytokine production, these data also suggest that Syk may serve as an important target for regulating neuroinflammation. Pharmacological inhibitors of Syk have been developed and represent potential immunomodulatory agents for the treatment of autoimmune and inflammatory conditions. Interestingly, Aβ has been shown to stimulate Syk activity resulting in microglial activation and neurotoxicity (65,–70).
Syk has also been shown to phosphorylate Tau in vitro, primarily at the Tyr-18 residue (56), which is considered to be an early event in the pathophysiology of AD (56, 71). We show that treatment of SH-SY5Y cells with the selective Syk inhibitor BAY61-3606 results in a partial reduction of Tau phosphorylation at Tyr-18 but also an almost complete inhibition of Tau phosphorylation at Ser-202 (CP13) and Ser-396/Ser-404 (PHF-1). The partial inhibition of Tau phosphorylation at Tyr-18 following Syk inhibition may be explained by the fact that other tyrosine kinases are also known to phosphorylate Tau at this particular residue (71). Additionally, we show that treatment of Tg Tau P301S mice with the selective Syk inhibitor BAY61-3606 results in a reduction of Tau Tyr-18 phosphorylation as expected but also in a significant decrease in Tau phosphorylation at other epitopes, including Ser-202 (CP13) and Ser-396/Ser-404 (PHF-1) that are phosphorylated by GSK3β (13). Because Syk inhibition results in decreased Tau phosphorylation at typical GSK3β sites in Tg P301S mice and SH-SY5Y cells, we investigated whether pharmacological inactivation of Syk could affect GSK3β. We found that inhibition of Syk in SH-SY5Y cells stimulates Ser-9 phosphorylation of GSK3β, which is known to inactivate the enzyme (57). We hypothesized that by this mechanism, Syk inhibition results in a reduction of Tau phosphorylation at Ser-202 and Ser-396/Ser-404 epitopes phosphorylated by GSK3β. Previous data have revealed that Syk phosphorylates Tyr-330 in the catalytic subunit of PKA resulting in the inactivation of PKA (60). We therefore tested the possible involvement of PKA for mediating the increased GSK3β Ser-9 phosphorylation induced by Syk inhibition. Our data show that PKA inactivation with KT5270 opposes the increased Ser-9 phosphorylation of GSK3β induced by Syk inhibition suggesting that Syk inhibition favors PKA activation providing a mechanism explaining the inactivation of GSK3β following Syk inactivation. This is also supported by the fact that phosphorylation of CREB (a substrate of PKA) is increased following Syk inhibition, whereas PKA inhibition decreases CREB phosphorylation induced by Syk inhibition. Altogether these data suggest that PKA activation following Syk inhibition plays an important role in the regulation of pathological Tau phosphorylation at GSK3β sites. Recent data have also emphasized this possibility by showing that PKA stimulation using a cAMP phosphodiesterase inhibitor leads to an inhibition of GSK3β activity and attenuates Aβ-induced tauopathy (72). GSK3β has also been shown to regulate Aβ production in vivo by controlling BACE-1 gene expression in a NFκB-dependent manner (15); therefore, the inhibition of GSK3β activity following Syk inhibition may also contribute to the inhibition of Aβ and BACE-1 expression observed following Syk inhibition. Alternatively, Syk inhibition may also result in a suppression of NFκB activation via an inhibition of RAF phosphorylation and AKT phosphorylation, which consequently is expected to lower BACE-1 expression and Aβ production. Interestingly, we show that Syk inhibition results in a stimulation of CREB phosphorylation via PKA activation suggesting that Syk inhibition could also have neuroprotective properties because stimulation of CREB phosphorylation by PKA is known to confer neuroprotection (73,–75). Also, PKA and CREB phosphorylation are down-regulated in AD brains (76) and in transgenic mouse models of AD overexpressing Aβ, whereas stimulation of PKA/CREB phosphorylation with various compounds improves memory in AD mouse models (77,–80) suggesting that Syk inhibition has the potential to reduce memory impairments by stimulating PKA/CREB signaling. We have summarized the Syk signaling events that regulate AD-related processes highlighting the potential therapeutic application of Syk inhibitors in AD (Fig. 11).
Altogether our data show that Syk plays a key role in the regulation of Aβ production, Tau hyperphosphorylation, and NFκB activation, suggesting that Syk inhibition has the potential to oppose the three main pathologies found in AD brain (Aβ deposition, Tau hyperphosphorylation, and neuroinflammation) and may therefore represent a particularly attractive target for the treatment of AD.
We are grateful to Dr. Peter Davies (Albert Einstein College of Medicine, Bronx, NY) for the gift of the PHF-1, CP13, and RZ3 antibodies. We thank Dr. Michael Wolf (Harvard Medical School, Boston) for providing 7W CHO APP-overexpressing cells.
*This work was supported by the Roskamp Foundation.
2The abbreviations used are: