Neuroprotection by aripiprazole against β-amyloid-induced toxicity by P-CK2α activation via inhibition of GSK-3β

Psychosis is reported over 30% of patients with Alzheimer's disease (AD) in clinics. Aripiprazole is an atypical antipsychotic drug with partial agonist activity at the D2 dopamine and 5-HT1A receptors with low side-effect profile. We identified aripiprazole is able to overcome the amyloid-β (Aβ)-evoked neurotoxicity and then increase the cell viability. This study elucidated the mechanism(s) by which aripiprazole ameliorates Aβ1-42-induced decreased neurite outgrowth and viability in neuronal cells. Pretreatment with aripiprazole increased Brain-derived neurotrophic factor (BDNF) mRNA and protein expressions in N2a cells. Additionally, phosphorylated casein kinase 2α at Y 255 (P-CK2α) was increased in time- and concentration-dependent manners. Furthermore, Aβ1-42-induced decreased BDNF and P-CK2α expression were increased over control level by aripiprazole. Subsequently, Aβ1-42-induced decreased levels of phosphorylated glycogen synthase-3β at Ser9 (P-GSK-3β) and nuclear P-β-catenin (Ser675) were elevated by aripiprazole, which were inhibited by K252A (inhibitor of BDNF receptor) and tetrabromocinnamic acid (TBCA, CK2 inhibitor), indicating that BDNF and P-CK2α activation are implicated in the aripiprazole effects. Expressions of cyclin D1 and insulin-like growth factor 2 (IGF2) mRNA were increased by aripiprazole; even in the presence of Aβ1-42, which was blocked by K252A and TBCA. In CK2α gene-silenced N2a cells, aripiprazole failed to increase P-GSK-3β and P-β-catenin expressions. Consequently, aripiprazole ameliorated Aβ1-42-induced attenuation of neurite elongation in HT22 cells, and this effect was blocked by both TBCA and imatinib. Decreased viability induced by Aβ1-42 was recovered by aripiprazole. These findings provide evidence supporting that aripiprazole can provide an effective therapeutic strategy against Aβ-induced neurotoxicity in AD-associated psychosis.

In addition, AD patients show several neuropsychiatric symptoms such as depression, agitation and psychosis (delusions, hallucinations), which have a negative impact on cognition [4].
Brain-derived neurotrophic factor (BDNF), the most abundant neurotrophin in the brain, has pivotal roles in synaptic plasticity and cognition [5]. Moreover, BDNF was demonstrated to inhibit GSK-3β activity through increased phosphorylation at serine 9 in cerebellar granule cells and human neuroblastoma SH-SY5Y cells [6]. The activation of the PI3K/Akt pathway by BDNF leads to inactivation of GSK-3β by phosphorylation at serine 9 [7]. Recently, aripiprazole was demonstrated to increase the BDNF level in the hippocampus of rats subjected to immobilization stress [8]. Furthermore, CK2 (casein kinase 2), a highly conserved tetrameric serine/ threonine kinase, plays an essential role in stimulation of the β-catenin/Tcf-LEF pathway [9,10]. In addition, many researchers have reported a reduction in pro-BDNF levels in brains of patients with AD [11,12].
On the other hand, neuronal morphogenesis involves the formation and differentiation of neurites into axons and dendrites [13]. NGF-stimulated axonal elongation is occurred by activation of p75 NTR in cultured hippocampal neurons through inhibition of GSK-3β activity [14]. Reportedly, when β-catenin is stabilized, it translocates to nuclei, where it acts over Tcf/LEF sites and induces transcriptional activation [15]. β-catenin accumulation activates transcription of insulin-like growth factor (IGF)2 and cyclin D1 (a protein that promotes cell cycle entry), because the promoters of β-catenin have Tcf/LEF motifs [16]. Several studies have shown that antidepressants increase expression of IGF1 [7] and IGF2 [17]. In addition, IGF2 shows the increase in BDNF and IGF1 [18]. IGF2 mRNA level was reported to be declined in the frontal cortex of AD patients in early stages of neuropathology [19].
Schneider and Dagerman [25] have reported that prevalence of psychosis is estimated in patients with AD: range from 10 to 73% (median of 34%) within clinic populations. It has been known that declining of cognitive function in AD patients is associated with a high prevalence of psychotic symptoms [26] and Results are represented as mean ± SEM of duplicates each pooled 4 -5 independent experiments. * P < 0.05, ** P < 0.01 vs. None; # P < 0.05, ## P < 0.01 vs. Aβ1-42 alone. www.impactjournals.com/oncotarget behavioral disturbances [27]. De Deyn et al. [4,28] have reported that in patients with psychosis associated with AD, aripiprazole-treatment showed significantly greater improvements in psychiatric rating scale compared to placebo and modest efficacy in the treatment of ADrelated psychosis.
Given that aripiprazole is able to overcome the Aβevoked inactivation of Wnt/β-catenin by increasing the phosphorylated GSK-3β (Ser 9) through activation of CK2α, we hypothesized that aripiprazole might increase P-GSK-3β level and nuclear translocation of β-catenin with enhanced expression of cyclin D and IGF2 mRNA through increased BDNF production-linked activation of P-CK2α and thereby it can enhance neurite outgrowth.

Aripiprazole increases expression of BDNF mRNA and protein in N2a cells
BDNF has been shown to have important roles in hippocampal synaptic plasticity [29] and memory function [30]. We assessed the increase in BDNF mRNA transcription and protein expression levels after treatment with aripiprazole in N2a cells. Following application of aripiprazole (3 μM) in N2a cells, the expression of BDNF mRNA was significantly elevated by 2.01 ± 0.38 fold (P < 0.05) at 3 hr, and subsequently declined at 6 -24 hr after treatment ( Figure 1A). Accordingly, the expression of BDNF protein after treatment with aripiprazole (3 μM) significantly increased in a time-dependent manner (0 -48 hr), and reached a plateau at 24 -48 hr (P < 0.05) ( Figure  1B). The expression of BDNF protein at 24 hr also was elevated with increased concentration of aripiprazole (1 -10 μM) (P < 0.05) ( Figure 1C). Some studies have reported a reduction in pro-BDNF levels in the brains of patients with AD [11,12]. Cells that were previously exposed to Aβ1-42 (10 μM) for 3 hr were treated with aripiprazole (1 -10 μM) for 24 hr. As shown in Figure 1D, Aβ1-42 exposure significantly decreased the expression of BDNF protein (up to 0.68 ± 0.11 fold, P < 0.01), and this decrease was prevented and rather elevated over the control by aripiprazole (3 and 10 μM) treatment to 1.88 ± 0.25 fold (P < 0.01) and 2.16 ± 0.30 fold (P < 0.01), respectively.

Increase in cyclin D1 and IGF2 in N2a cells
Cyclin D1 is an important regulator of G1/S phase cell cycle progression, and it is known to play an essential role in NGF-mediated differentiation [35]. We determined whether cyclin D1 is necessary to exert the proliferative effect of β-catenin signaling, since cyclin D1 has a role linked to the target genes of β-catenin. Following treatment with aripiprazole (3 μM), the expression of cyclin D1 mRNA was assessed over time (0 -36 hr). As shown in Figure 5A, the mRNA of cyclin D1 was maximally induced at 24 hr to 2.97 ± 0.59 fold (P < 0.001) and thereafter declined. Further, cyclin D1 mRNA expression was significantly suppressed by Aβ1-42 (10 μM) to 0.56 ± 0.09 fold (P < 0.05). Upon treatment with aripiprazole (3 μM) in the presence of Aβ1-42 (10 μM), cyclin D1 mRNA expression was significantly increased to 2.94 ± 0.65 fold (P < 0.01), but the increase was completely attenuated by K252A (100 nM, P < 0.05) and TBCA (10 μM, P < 0.05) ( Figure 5B).

Effect on cell viability
N2a cells were treated with different concentrations of aripiprazole for 24 hr without Aβ1-42, after which the cell viability/cytotoxicity assay was performed. There was little change in cell viability up to 10 μM of aripiprazole, but 30 μM of aripiprazole caused significant decrease in viability to 38% (P < 0.001), indicating aripiprazole is relatively safe drug ( Figure 7A). The cytotoxic effect of exogenously applied Aβ1-42 in N2a cells was assessed using a cell viability/cytotoxicity assay. Exposure of N2a cells to Aβ1-42 (10 μM) for 24 and 48 hr resulted in a significant decline in cell viability by 71.4 ± 3.0% (P < 0.001) and 71.8 ± 5.9% (P < 0.001), respectively. The decreased viability induced by Aβ1-42 was recovered by aripiprazole (3 μM) to marginally at 24 h, and significantly to 81.3 ± 3.2% (P < 0.05) at 48 h, which was blocked by K252A (100 nM, tropomyosin receptor kinase B (TrkB) receptor inhibitor) [43] and by TBCA (10 μM, CK2 inhibitor) [44] (Figure 7B & 7C). These results suggest BDNF and CK2 activation are involved in the aripiprazole-stimulated cell viability.

DISCUSSION
The results of this study demonstrates that aripiprazole enhances neurite outgrowth and cell viability in the presence of Aβ1-42 by enhancing BDNF production and suppressing Aβ-induced GSK-3β activation, and thereby promoting nuclear translocation of P-β-catenin and increasing expression of cyclin D1 and IGF2 in the nucleus via enhancement of P-CK2α activation.
BDNF, a neurotrophin family member, has important roles in hippocampal synaptic plasticity [29]  and memory function [30]. Some researchers have reported a reduction in pro-BDNF levels in AD brains [11,12]. Wnt signal activation has a role in rescuing neurons from degeneration and improves animal behavioral impairments induced by β-amyloid fibril [39,40]. We observed the increase in BDNF mRNA transcription and protein expression after treatment with aripiprazole in N2a cells. Even though application of Aβ1-42 significantly decreased the expression of BDNF, the decreased BDNF level overwhelmingly surpassed the control levels by  treatment with aripiprazole. BDNF has critical functions in promoting survival and differentiation of neural stem cells via activation of Wnt/β-catenin signaling molecules [41].
Given that BDNF increases CK2 activity, we assessed the increase of P-CK2α (Y 255) expression after aripiprazole treatment. The expression of P-CK2α was significantly increased time-and concentrationdependently in N2a cells by aripiprazole without changing total CK2α expression. As Lee et al. [32] have indicated, P-CK2α expression was significantly decreased in response to Aβ1-42 in this study: Aβ1-42-induced decreased P-CK2α level was significantly recovered over the control value (by 1.3 ~ 1.7 fold) under pretreatment with aripiprazole. It is widely known that GSK-3β is inhibited via phosphorylation at specific serine residues (e.g., serine 9 for GSK-3β) [33], and accumulation of active GSK-3β has been implicated in neurofibrillary degeneration in AD [42]. As predicted, the level of P-GSK-3β (Ser 9) was significantly decreased to ~ 0.24 fold (P < 0.001) by Aβ1-42, but following treatment with aripiprazole, the decreased P-GSK-3β level was elevated. The increased P-GSK-3β levels were significantly blocked by K252A (BDNF receptor inhibitor) [43] and by TBCA (a CK2 inhibitor) [44]: these findings indicate that BDNF and CK2 activation are involved in aripiprazole-stimulated P-GSK-3β levels. CK2 is also implicated in Wnt signaling, where it acts as a positive regulator by phosphorylation of β-catenin, thereby leading to resistance to degradation by the proteasome and increased co-transcriptional activity [45]. As CK2α inhibits GSK-3β by phosphorylation at Ser 9, it was hypothesized that aripiprazole must stabilize and translocate β-catenin to the nucleus. Balaramana et al. [46] suggested that Wnt/β-catenin activity was notably low in AD patients' brain. Consistent with this report, upon exposure of N2a cells to Aβ1-42, the level of P-β-catenin (Ser 675) was significantly decreased in the nuclear compartments. Interestingly, decreased nuclear P-β-catenin level was significantly elevated by aripiprazole, and these increases were completely blocked by K252A and TBCA. These results strongly suggest that aripiprazole-promoted nuclear translocation of P-β-catenin implies activation of BDNF and CK2α. These results support those reported by Sinha et al. [17] showing that inhibition of GSK3β activation is important for maintaining viability and activating the Wnt pathway.
Previous reports have shown that β-catenin activates the transcription of cyclin D1 (indicative of a promitogenic cell response) through TCF-binding sites within the promoter, which has a direct effect on cell proliferation [16] and through IGF2, a potent proliferative signaling protein [17]. In the present study, aripiprazole significantly increased the expressions of cyclin D1 and IGF2 mRNA, which had been suppressed by Aβ1-42. These increased mRNA expressions were blocked by both K252A and TBCA, indicating that aripiprazole-stimulated expression of cyclin D1 and IGF2 mRNA implies activation of BDNF and CK2α.
The postulation that aripiprazole-stimulated elevations of P-GSK-3β (Ser 9) and P-β-catenin (Ser 675) expressions are mediated via CK2α activation was further confirmed using N2a cells transfected with CK2α siRNA. After silencing the CK2α gene, the expressions of P-GSK-3β and P-β-catenin were not induced by aripiprazole, whereas negative control cells were obviously responsive to aripiprazole. It has been demonstrated that activation of CK2 by NGF enhances neurite extension in PC12 cells [47]. In addition, depletion of CK2 by antisense oligonucleotide has been reported to inhibit neuritogenesis in neuroblastoma cells, indicative of the importance of CK2α activation in neurite elongation [14]. In the present study, HT22 cells, mouse hippocampal neuronal cell line, were used instead of N2a cell, because HT22 cells phenotypically resemble neuronal precursor cells expressing BDNF receptor TrkB, and lack functional ionotropic glutamate receptors [37,48], thus it was possible to exclude excitotoxicity as a cause for neurite outgrowth damage by glutamate other than Aβ1-42.
The Aβ1-42-induced decrease in neurite length in HT22 cells was prevented by aripiprazole, and the recovered neurite elongation was blocked by TBCA (CK2 inhibitor) and imatinib (β-catenin inhibitor), these findings indicating that activation of CK2α and β-catenin is importantly implicated in aripiprazole-stimulated neurite outgrowth in HT22 cells. Considering that IGF2 increases hippocampal levels of NGF, BDNF, and NT3 to varying degrees in animal model AD [18], it is suggested that BDNF is importantly involved in the aripiprazolestimulated neurite outgrowth in support of critical roles in the function and survival of neurons.
It is known that aripiprazole's mechanism of action is pharmacologically ascribed to a combination of partial agonistic activity at D 2 and 5-HT 1A receptors and antagonistic activity at 5-HT 2A receptors. Shioda et al. [49] have proposed that nuclear calcium/calmodulindependent protein kinase II (CaMKII) functions in transcriptional activation in the neurotrophin BDNF through the phosphorylation of diverse nuclear proteins, including CREB. However, it remains undefined as to the mechanism by which aripiprazole stimulates BDNF synthesis is related to D 2 dopamine receptors, and/or to agonistic activity of 5-HT 1A receptors or antagonistic activity of 5-HT 2A receptors. This goes beyond the scope of the current study.
Considering these results are related to pharmacological inhibition and genetic blockade of CK2α, it is concluded that the activation of BDNF-coupled P-CK2α (Y 255) by aripiprazole stimulates expression of cyclin D1 and IGF2 mRNA through mediation of P-GSK-3β (Ser 9) and nuclear P-β-catenin (Ser 657), thereby contributing to neurite outgrowth and cell viability, even in the presence of Aβ1-42.

Cell culture
The N2a, wild-type cells, a mouse neuroblastoma cell line, were provided by Dr. Takeshi Iwatsubo (Department of Neuropathology and Neuroscience, Graduate School of Pharmaceutical Sciences, The University of Tokyo) and cultured in DMEM supplemented with 10% FBS. N2a cells (neuroblastoma cell line) has been described to produce low levels of tyrosine hydroxylase and dopamine, and differentiate into dopamine neurons. Both tyrosine hydroxylase and dopamine levels were significantly enhanced by cAMP responsive element binding protein (CREB) [50]. HT22 cells, a murine hippocampal cell line, were donated by Dr. H.T. Chung (Ulsan University, Ulsan, Korea) and cultured in DMEM supplemented with 10% FBS.

RT-qPCR analysis
Total RNA was isolated from cells by using TRIzol reagent (Invitrogen). cDNA was synthesized from 1 μg of total RNA. Gene expressions were measured by performing real-time PCR using a LightCycler 96 system (Roche Molecular Biochemicals, Mannheim, Germany) equipped with LightCycler DNA Master SYBR Green I (Roche Molecular Biochemicals). PCR was performed under the following conditions: 95 °C for 10 min followed by 50 amplification cycles of (95 °C for 10 s, 50 °C for 10 s, and 72 °C for 10 s). Primers sequences are detailed in Table 1. Quantification was performed by using LightCycler 96 Software (Roche Molecular Biochemicals).

Small interfering RNA preparation and transfection
CK2α small interfering (si)RNA oligonucleotide (GenBank accession No. NM_009974.2) was synthesized by Bioneer (Daejeon, Korea). siRNA molecules were transfected into cells by using X-tremeGENE siRNA transfection reagent (Roche, Indianapolis, IN) according to the manufacturer's instructions.

Quantification of neurite elongation
To observe neurite elongation, HT22 cells, a stable murine hippocampal cell line, were plated at a density of 1,000 cells per cm 2 on sterile, coated, 18 × 18-mm cover slips in a six-well culture plate. HT22 cells were incubated with Aβ1-42 (10 μM) alone or with aripiprazole (10 μM) in the absence and presence of inhibitors for 5 days. For the morphometric analysis, cells were fixed in 4% paraformaldehyde for 15 min at room temperature and then incubated with SMI-312 antibody (Cat. No. SMI312R, Covance, Princeton, NJ) for 1 hr. After a series of washes with PBS, secondary antibody conjugated to Alexa Fluor 488 (Invitrogen, Carlsbad, CA) was applied for 1 hr. All fluorescent images were magnified at ×400 by using an Axiovert 200 fluorescence microscope (Zeiss, Oberkochen, Germany). The length of the main neurite of each cell was measured in five independent experiments that were performed in duplicate.

Cell viability
Cell viability was evaluated using the Cyto XTM cell viability assay kit (LPS solution, Daejeon, Korea). For viability assay, cells were treated with 10 % Cyto XTM per well, and again incubated at 37˚C in a 5% CO2 incubator for 3 hr. Sample absorbance was determined at 450 nm using an ELISA (BioTek Inc., Winooski, VT).

Statistical analysis
Results are expressed as mean ± SEM values. The significances of results were determined by performing one-way analysis of variance (ANOVA) followed by Tukey's Multiple Comparison Test. The Student's t-test was used to determine the significances of treatment effects. P values of < 0.05 were considered significant.