Concurrent Treatment with Taxifolin and Cilostazol on the Lowering of β-Amyloid Accumulation and Neurotoxicity via the Suppression of P-JAK2/P-STAT3/NF-κB/BACE1 Signaling Pathways

Taxifolin is a potent flavonoid that exerts anti-oxidative effect, and cilostazol increases intracellular cAMP levels by inhibiting phosphodiesterase 3 that shows antiinflammatory actions. BACE1 (β-site APP cleaving enzyme 1) is the rate-limiting enzyme responsible for the β-cleavage of amyloid precursor proteins to Aβ peptides. In this study, endogenous Aβ and C99 accumulation was explored in N2a Swe cells exposed to 1% FBS medium. Increased Aβ and C99 levels were significantly attenuated by taxifolin alone and in combination with cilostazol. Increased phosphorylated JAK2 at Tyr1007/1008 (P-JAK), phosphorylated STAT3 at Tyr 705 (P-STAT3) expressions and increased expressions of BACE1 mRNA and protein in the activated N2a Swe cells were significantly attenuated by taxifolin (10~50 μM), cilostazol (10~50 μM) alone and in combination at minimum concentrations. In these cells, decreased cytosol IκBα expression was elevated, and increased nuclear NF-κB p65 level and nuclear NF-κB p65 DNA binding activity were significantly reduced by taxifolin and cilostazol in a similar manner. Following STAT3 gene (~70% reduction) knockdown in N2a cells, Aβ-induced nuclear NF-κB and BACE1 expressions were not observed. Taxifolin, cilostazol, or resveratrol significantly stimulated SIRT1 protein expression. In SIRT1 gene-silenced (~50%) N2a cells, taxifolin, cilostazol, and resveratrol all failed to suppress Aβ1-42-stimulated P-STAT3 and BACE1 expression. Consequently, taxifolin and cilostazol were found to significantly decrease lipopolysaccharide (1–10 μg/ml)-induced iNOS and COX-2 expressions, and nitrite production in cultured BV-2 microglia cells and to increase N2a cell viability. In conclusion, taxifolin and cilostazol strongly inhibited amyloidogenesis in a synergistic manner by suppressing P-JAK2/P-STAT3-coupled NF-κB-linked BACE1 expression via the up-regulation of SIRT1.


Introduction
Alzheimer's disease (AD) is characterized by increased amyloid β (Aβ)-containing extracellular plaque and intracellular neurofibrillary tangles, which are associated with synaptic failure and cognitive deficits [1]. Enhanced amyloidogenic processing of amyloid precursor protein (APP) by β-and γ-secretase increases intracellular level of soluble oligomeric Aβ, which results in pronounced synaptic failure and eventually in memory decline [2,3]. Theoretically, Aβ accumulation can be reduced in AD patients by suppressing Aβ production or enhancing Aβ degradation and clearance. A membrane-associated C-terminal fragment of APP, C99, is liberated by the action of β-secretase, and this is subsequently cleaved by γ-secretase to produce Aβ peptide [4].
BACE1 (β-secretase, a membrane-bound aspartyl protease β-site APP cleaving enzyme 1) is a rate-limiting enzyme for β-amyloid production [5]. The expression of BACE1 protein and its activity have been demonstrated to be elevated in the brains of AD patients [6,7]. Buggia-Prevot et al. [8] proposed Aβ 1-42 acts as a regulator of BACE1, and suggested exacerbated Aβ production modulates BACE1 promoter transactivation and its activity via an NF-κB-dependent pathway. Furthermore, Aβ has been shown to activate nuclear transcription factor NF-κB [9,10], which is activated during the early stages of AD, where RelA/p65 plays a critical role in neurons and astrocytes surrounding amyloid plaques in the brain, and elevates oxidative stress [11].
In addition, constitutive Janus kinase 2 (JAK2)/signal transducer and activator of transcription 1 (STAT1) signaling has been demonstrated to contribute to endogenous BACE1 expression and subsequent Aβ generation in neurons, and inhibition of the JAK2/STAT1 signaling pathway by AG490 (a JAK2 inhibitor) reduced the expression of endogenous BACE1 and Aβ production [12]. Grivennikov and Karin [13] postulated STAT3-mediated nuclear NF-κB activation plays an important role in the pathogenesis of cancer and neurodegenerative disease, despite the fact NF-κB is not the only transcription factor that cooperates with STAT3.
Previous studies have shown [17,18], cilostazol inhibits proinflammatory factor (including remnant lipoprotein particle, TNFα and LPS)-stimulated increases in IκBα degradation and NF-κB p65 activity, and thus, suppresses the productions of inflammatory cytokines. Interestingly, Paris et al. [19] showed NF-κB inhibitors decreased both Aβ 1-40 and Aβ 1-42 production and suggested NF-κB inhibitors might be of therapeutic importance for the treatment of AD.
It is generally said that the currently available hitting-one-target drugs are insufficient for the treatment of AD, and thus, research designed to identify drug therapies that target multiple action sites is attracting more attention. Considering these points of view, combinatorial drug therapy that target multiple pathogenic factors are worthy of investigation for prevention and treatment of AD. For this reason, we investigated the efficacies of taxifolin alone and in combination with cilostazol with respect to the regulation of BACE1, Aβ accumulation, and the inhibition of neuroinflammation with a view towards possible intervention in Alzheimer's disease.
In the present study, we examined the inhibitory effect of taxifolin and taxifolin plus cilostazol on the JAK2/STAT3/NF-κB signaling axis in Aβ production in N2a cells stably expressing human APP Swedish mutation (N2a Swe cells) with focus on the suppressive effects of taxifolin and of taxifolin plus cilostazol on the Aβ-induced mRNA and protein expressions of BACE1. In addition, we elucidated the mechanism(s) by which Aβ 1-42 -induced STAT3 and NF-κB activation are down-regulated by these drugs, and verified that taxifolin and cilostazol both inhibit P-STAT3, nuclear NF-κB p65, and BACE1 expression by upregulating SIRT1.

Results
Effects on Aβ-induced P-JAK2 and P-STAT3 expressions in activated N2a Swe cells Initially, we sought to confirm the endogenous production of Aβ is accompanied by increases in the expressions of P-JAK2 and P-STAT3 in N2a Swe cells [20]. As shown in Fig 1Aa and 1Ab, activated N2a Swe cells upon exposure to 1% FBS revealed a large accumulation of intracellular Aβ (4 kDa) and an elevated C-terminal fragment β (C99) level (11 kDa) in a time (3-48 hr)-dependent manner with maximum values at 12~24 hr (Fig 1Aa and 1Ab). After obtaining this confirmation, we assessed the expressions of P-JAK2 and P-STAT3 in activated N2a Swe cells.
Effects on Aβ-induced BACE1 mRNA and protein expressions in N2a Swe cells Shimmyo et al. [22] reported natural flavonoids, such as, myricetin, quercetin, and kaempferol, directly inhibit BACE1 enzyme activity in a concentration-dependent manner and concomitantly inhibit Aβ 1-42 levels. As many of the flavonoids tested inhibited β-cleavage of APP, we sought to determine whether treatments with taxifolin or taxifolin plus cilostazol directly inhibit β-secretase (BACE1) protein.

Discussion
The present study shows that endogenously accumulated Aβ and C-terminal fragment β (C99) liberated from activated N2a Swe cells markedly increased the expressions of P-JAK, P-STAT3 and enhanced the nuclear translocation of NF-κB p65 and its activity, and that these events subsequently resulted in significant increases in the mRNA and protein expressions of BACE1 and amyloidogenesis. Taxifolin and cilostazol when combined even in lower concentrations Recently, we showed that endogenously accumulated Aβ in N2a Swe cells (expressing human APP Swedish mutation) was significantly and concentration-dependently diminished by cilostazol. SIRT1-linked retinoic acid receptor-β activation by cilostazol underlies its pharmacological actions with respect to increasing ADAM10/α-secretase activity and reducing intracellular accumulations of full-length-APP and Aβ [28]. The present study highlights the synergistic effects of co-treatment with taxifolin plus cilostazol at low concentrations in terms of lowering of the intracellular accumulations of Aβ and C99 by suppressing BACE1 expression and reducing the generations of proinflammatory substances.
It is noteworthy that BACE1 gene promoter contains a NF-κB binding site [8]. Chen et al. [29] have reported both BACE1 and NF-κB p65 levels significantly increased in the brains of AD patients, and that increased NF-κB levels upregulate BACE1 promoter activity and BACE1 transcription. In the present study, when N2a Swe cells were exposed to 1% FBS (to overproduce Aβ and C99), P-JAK2 and P-STAT3 expressions were significantly elevated at 3 hr, and these up-regulations were suppressed by taxifolin, cilostazol, and by AG490 (a JAK2 inhibitor), indicating P-STAT3 expression is increased via the mediation of P-JAK2 activation. The expression of P-STAT3 in cytosol was significantly elevated at 3 hr under 1% FBS and then nuclear P-STAT3 expression increased from 3 to 12 hr, and its elevation was suppressed by taxifolin plus cilostazol, which indicates a decrease in nuclear P-STAT3 levels by these drugs.
Several lines of study have revealed NF-κB inhibitors are therapeutically important for treatment of AD pathology because the inhibitors block inflammatory processes and directly inhibit the production of Aβ peptides [19] and NF-κB inhibition prevents Aβ-induced BACE1 promoter transactivation [8,29]. Herein, we assessed the impact of taxifolin on the Aβ-stimulated degradation of IκBα in cytosol and on the nuclear translocation of NF-κB. Notably, cilostazol has been reported to inhibit NF-κB activation [17,18]. Consistent with these reports, decreased cytosolic IκBα expression and increased nuclear NF-κB p65 levels in activated N2a Swe cells were totally prevented by pretreating taxifolin plus cilostazol. Furthermore, co-treatment with lower concentrations of taxifolin plus cilostazol caused striking increases in IκBα expression and decreased nuclear NF-κB p65 expression, whereas the monotherapy with same concentration could not influence. These findings were further supported by our NF-κB p65 DNA binding assay results, which showed that the increase in NF-κB activity was substantially decreased by co-treatment with lower doses of taxifolin and cilostazol. Overall, these results confirm the synergistic effect of combinatorial taxifolin plus cilostazol therapy.
Earlier, it was reported interactions between STAT3 and p300 or histone acetyltransferase (HAT) are required for STAT3-dependent transcription [30], and that the reversible acetylation of RelA regulates the duration of the nuclear activity of NF-κB [31]. After phosphorylation of its critical tyrosine residue (Tyr 705), STAT3 dimerizes due to a phosphotyrosine-SH2 domain interaction [32]. According to a report issued by Grivennikov and Karin [13], activated STAT3 interacts with RelA/p65 to recruit p300 HAT complex and causes RelA acetylation that prolongs its nuclear retention. Once dimerized, STAT transcription factors enter the nucleus where they interact with RelA/p65, and subsequently recruit p300-HAT. In addition, p300 can acetylate p65 and increase its nuclear retention, and thereby, enhance its transcriptional activity [33]. Given that STAT3 prolongs the nuclear retention of NF-κB, P-STAT3 inactivation by taxifolin plus cilostazol may decrease NF-κB activity.
Taxifolin is a dihydroquercetin flavonoid found in onions and French maritime pine bark [14]. Several studies have reported that natural flavonoids, such as, myricetin and quercetin directly inhibit BACE1 in a concentration-dependent manner and concomitantly inhibit Aβ 1-42 levels [22]. The present study shows taxifolin, cilostazol, and quercetin all significantly diminished the mRNA and elevated protein expressions of BACE1 in activated N2a Swe cells.
NAD + -dependent protein deacetylase, particularly SIRT1, has attracted attention since it was reported it down-regulated the transactivation potential of the RelA/p65 subunit of NF-κB [34]. Lee et al. [20] previously reported cilostazol (3-10 μM) stimulates SIRT1 mRNA and protein expression and increases SIRT1 activity in N2a cells. Of the seven sirtuins (SIRT1-7), SIRT1 has been strongly implicated in AD, that is, loss of SIRT1 was found to be closely associated with Aβ accumulation in the brains of AD patients [35,36]. In the present study, taxifolin, cilostazol, and resveratrol all significantly stimulated SIRT1 protein expression and SIRT1 activity. Interestingly, in N2a cells transfected with SIRT1 siRNA, both taxifolin and cilostazol failed to suppress the Aβ 1-42 -stimulated expressions of P-STAT3 and BACE1, which contrasted to that observed in cells transfected with scrambled siRNA duplex (negative control), indicating both taxifolin and cilostazol inhibit the expressions of P-STAT3 and BACE1 by up-regulating SIRT1. These results suggest BACE1 inhibition could be a valid therapeutic target in AD.
In agreement with the suggestion made by Fusco et al. [37] that CREB directly regulates SIRT1 mRNA and protein expression in neurons, Lee et al. [20] showed a decrease in CREB/ P-CREB Ser133 and of SIRT1 mRNA was induced by endogenous Aβ overproduction in activated N2a Swe cells. Alternatively, when these cells were pretreated with cilostazol, P-CREB Ser133 expression was significantly elevated and the decrease in SIRT1 mRNA expression was prevented. In view of these findings, we believe it likely that the up-regulation of P-CREB Ser133 by cilostazol increases SIRT1 transcription in neurons. However, it remains to be determined how taxifolin increases SIRT1 expression.
NF-κB is a central regulator of microglial responses to proinflammatory stimuli, including LPS and cytokines [38]. A number of reports have shown LPS strongly activates microglia and induces the COX-2 gene, leading to the synthesis of PGE2 [39]. Furthermore, the expressions of iNOS and COX-2 are known to be regulated by the activation of NF-κB. Reportedly, there is one NF-κB DNA consensus sequence within COX-2 promoter, and two NF-κB DNA consensus sequences within iNOS promoter [40][41][42]. In the present study, when BV-2 microglia were exposed to LPS, nitrite levels (a marker of NO production) significantly increased in conjunction with the expressions of iNOS and COX2, and these increased variables were attenuated by taxifolin (3~100 μM) and by cilostazol (3~30 μM). In addition, we observed the reduced viability of N2a cells by Aβ 1-42 was significantly prevented by 10 or 30 μM of taxifolin or cilostazol, respectively. In light of these results, the suppression of NF-κB by co-treatment with taxifolin and cilostazol provides a potential strategy for the management of AD by reducing neuroinflammation and Aβ generation.
Taken together, these results indicate co-treatment with taxifolin and cilostazol inhibits amyloidogenesis by suppressing P-JAK2/P-STAT3-coupled NF-κB-linked BACE1 expression via the up-regulation of SIRT1 in a synergistic manner, and suppresses proinflammatory reactions and increases cell viability. These results may provide a rationale for combinatorial therapy with taxifolin and cilostazol in clinical trials targeting AD.

Cell culture
Mouse neuroblastoma N2a wild-type cells and N2a Swe cells were donated by Dr. Takeshi Iwatsubo (Department of Neuropathology and Neuroscience, Graduate School of Pharmaceutical Sciences, University of Tokyo) [20] and grown in DMEM medium containing 10% fetal bovine serum (FBS). To evoke endogenous Aβ overproduction, the culture medium for N2a and N2a Swe cells was switched from DMEM containing 10% fetal bovine serum (FBS) to DMEM containing 1% FBS, and cells were then cultured for 3~24 hr. BV-2 cells were obtained from the American Type Culture Collection (Rockville, Maryland) and grown in DMEM containing 10% FBS. By Western blot analysis, we confirmed the purity of BV2 cells: the cells highly expressed CD11b and CD68 proteins, indicating microglial cell line (data not shown).

NF-κB p65 transcription factor assay
To assess nuclear NF-κB transcription levels, cell lysates were extracted using a nuclear extraction kit (Enzo Life Sciences, Farmingdale, NY). Specific DNA binding activities of NF-κB p65 in nuclear extracts were determined using NF-κB p65 transcription factor assay kit (Active Motif, Carlsbad, CA).

Western blot analysis
Proteins (30 μg) were loaded onto 8~15% SDS-polyacrylamide electrophoresis gels, electrophoresed, and transferred to nitrocellulose membranes (Amersham Biosciences, Inc., Piscataway, NJ), which were then incubated with appropriate antibodies. Immunoreactive bands were visualized using the chemiluminescent reagent in the Supersignal West Dura Extended Duration Substrate Kit (Pierce Chemical, Rockford, IL). Signals from bands were quantified using a GS-710 calibrated imaging densitometer (Bio-Rad, Hercules, CA).

Immunocytochemistry
Cells were cultured in cover glass-bottomed dishes, fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100 in PBS. Expression of NF-κB p65 was detected using anti-NF-κB p65 antibody. Cells were incubated with primary antibodies for 4 hrs, and then with Cy3-conjugated secondary antibodies for 1 hr. Fluorescent images were obtained using a confocal microscope (OLYMPUS FV-1000, Tokyo).

Quantitative real-time reverse transcription
Total RNA was extracted from cells using TRIzol reagent (Invitrogen, San Diego, CA). BACE1 gene expression was determined by real-time PCR using a LightCycler 96 system (Roche Molecular Biochemicals, Mannheim, Germany) instrument and LightCycler DNA Master SYBR Green I (Roche) using 1 ng of reverse-transcribed total RNA. PCR was performed using 95˚C for 10 min followed by 50 amplifications of 95˚C for 10 s, 50˚C for 10 s, and 72˚C for 10 s. The following oligonucleotide primers specific for BACE1 were used: murine BACE1: 5 0 -GCATGATCATTGGTGGTATC-3 0 (sense) and 5 0 -CCATCTTGAGATCTTGACCA-3 0 (antisense). Data were analyzed using LightCycler 96 Software (Roche Molecular Biochemicals).

NF-κB DNA binding assay
To measure NF-κB transcription factor levels in nuclear extracts, cell lysates were extracted using a nuclear extraction kit (Chemicon International, Temecula, CA). NF-κB p65 DNA binding activities were determined using a colorimetric NF-κB p65 transcription factor assay kit (Rockland Immunochemicals, Gilbertsville, PA), according to the manufacturer's instructions.
Determination of nitric oxide (NO) production BV2 microglial cells (1.5 × 10 5 cells/ml) were grown in 96-well plates and then incubated with or without LPS (1 μg/ml) in the absence or presence of different concentrations of taxifolin for 24 hours. Nitrite accumulations in supernatant were assessed using a Griess assay. Briefly, 50 μl of culture supernatant was mixed with an equal volume of Griess reagent [0.1% N-(1-naphthyl)-ethylenediamine and 1% sulfanilamide in 5% phosphoric acid] and incubated at room temperature for 10 minutes. Absorbance at 540 nm was measured using a microplate reader (Thermo Electron Corporation, Marietta, OH) using sodium nitrite as a standard.

SIRT1 deacetylation assay
The in vitro deacetylase assay was performed using a fluorometric SIRT1 Assay Kit (CS1040; Sigma), according to the manufacturer's protocol. Briefly, the reaction was carried out at 37˚C for 30 min and deacetylase activity was detected as a fluorescent emission at 450 nm, using an excitation wavelength of 360 nm. The fluorescence intensities of compounds at 450 nm were subtracted from initially determined assay values. The assay was performed in duplicate at least three times.

Statistical analysis
Results are presented as means ± SEMs. One-way analysis of variance (ANOVA) followed by Tukey's post hoc multiple comparisons method was used to determine the significances of intergroup differences. The student's t-test was used to determine the significances of differences means of untreated and inhibitor treated cells. Statistical significance was accepted for P values of < 0.05.