Inhibitory effects of aripiprazole on interferon-γ-induced microglial activation via intracellular Ca2+ regulation in vitro

Authors


Address correspondence and reprint requests to Akira Monji, Department of Neuropsychiatry, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi Higashi-ku, Fukuoka 812-8582, Japan. E-mail: amonji@hf.rim.or.jp

Abstract

The activation of the inflammatory/immunological response system is suggested to be related to the pathophysiology of schizophrenia. Aripiprazole is a novel atypical antipsychotic, which is a high-affinity dopamine D2 receptor partial agonist. Atypical antipsychotics, all of which have dopamine D2 receptor antagonism, have recently reported to have significantly inhibitory effects on interferon (IFN)-γ-induced microglial activation in vitro. In the present study, we investigated whether or not aripiprazole also has anti-inflammatory effect on IFN-γ-induced microglial activation. Not quinpirole, dopamine D2 full agonist, but aripiprazole significantly inhibited the generation of nitric oxide (NO) and tumor necrosis factor (TNF)-α from IFN-γ-activated microglia and suppressed the IFN-γ-induced elevation of intracellular Ca2+ concentrations ([Ca2+]i) in murine microglial cells. Increased [Ca2+]i has been reported to be required, but by itself not sufficient, for the release of NO and certain cytokines. As a result, we can speculate that aripiprazole may inhibit IFN-γ-induced microglial activation through the suppression of IFN-γ-induced elevation of [Ca2+]i in microglia. Our results demonstrated that not only antipsychotics which have dopamine D2 receptor antagonism but also aripiprazole have anti-inflammatory effects via the inhibition of microglial activation. Antipsychotics may therefore have a potentially useful therapeutic effect on patients with schizophrenia by reducing the microglial inflammatory reactions.

Abbreviations used
[Ca2+]i

intracellular Ca2+ concentration

BAPTA

1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid

BIM I

bisindolylmaleimide I

DMSO

dimethyl sulfoxide

ERK

extracellular signal regulated kinase

IFN-γ

interferon-γ

LPS

lipopolysaccharide

MAPK

mitogen-activated protein kinase

MTT

3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetra-zolium bromide

NO

nitric oxide

PKC

protein kinase C

TNF-α

tumor necrosis factor-α

Microglial activation plays an important role in the pathophysiology of neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease via the release of proinflammatory cytokines, nitric oxide (NO) or reactive oxygen species (McGeer et al. 1993; Stoll and Jander 1999). Recent neuroimaging studies suggest the possibility that even schizophrenia may be a kind of neurodegenerative diseases like Alzheimer’s disease and Parkinson’s disease (Lieberman 1999; Lieberman et al. 2005; Salisbury et al. 2007). Post-mortem brain studies using class II human leucocyte antigen (HLA-DR) have revealed microglial activation in the brains of patients with schizophrenia (Bayer et al. 1999; Radewicz et al. 2000; Steiner et al. 2006, 2008). The activation of the inflammatory/immunological response has been increasingly reported in patients with schizophrenia (Lin et al. 1998; Zhang et al. 2004; Bernstein et al. 2005; Drzyzga et al. 2006). Immunomodulatory drugs such as cyclooxygenase-2 inhibitors and minocycline, a potent inhibitor of microglial activation, have shown beneficial effects on symptoms of schizophrenia (Akhondzadeh et al. 2007; Miyaoka et al. 2007). Moreover, atypical antipsychotics, such as clozapine and risperidone, have reported to decrease serum levels of IL-2, IL-6 and tumor necrosis factor (TNF) –α in patients with schizophrenia (Lu et al. 2004).

For long time, dopamine D2 receptor antagonism has been believed to be the primary therapeutic target for schizophrenia because almost all antipsychotics have dopamine D2 receptor antagonism (Kapur and Mamo 2003; Miyamoto et al. 2005). Aripiprazole is a novel atypical antipsychotic, which is a high-affinity dopamine D2 receptor partial agonist (Burris et al. 2002; Shapiro et al. 2003). In spite of a different pharmacological profile, aripiprazole is effective against the positive and negative symptoms of patients with schizophrenia like other antipsychotics with lower side effects (Potkin et al. 2003; Tandon et al. 2006; Kane et al. 2007; Taylor et al. 2007).

We recently demonstrated that atypical antipsychotics such as risperidone, perospirone, quetiapine and ziprasidone, all of which have D2 antagonism, have a significantly inhibitory effect on interferon (IFN) -γ-induced microglial activation in vitro (Kato et al. 2007; Bian et al. 2008). In the present study, we therefore investigated whether or not aripiprazole also has an inhibitory effect on IFN-γ-induced microglial activation. Intracellular Ca2+ regulation plays an important role in microglial activation (Hoffmann et al. 2003). We therefore also investigated the effect of aripiprazole on the intracellular Ca2+ regulation in IFN-γ-activated microglia.

Materials and methods

Materials

Aripiprazole was kindly donated by Otsuka Pharmaceutical Co., Ltd. (Tokushima, Japan). Recombinant IFN-γ and mouse granulocyte macrophage colony stimulating factor were purchased from R&D systems (Minneapolis, MN, USA). Bisindolylmaleimide I (BIM I) was purchased from BIOMOL (Plymouth Meeting, PA, USA). Quinpirole, SB203580, PD98059, 1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) AM and all other main chemicals were purchased from Sigma (St. Louis, MO, USA).

Aripiprazole was initially dissolved into 20 mM with dimethyl sulfoxide (DMSO) and then diluted to 1 mM with phosphate-buffered saline (PBS) for experiments. The final concentrations of aripiprazole were 5–20 μM. DMSO at the highest concentration (0.1%) under the experimental conditions was not toxic to cells.

Cell cultures

The murine microglial cell line, 6-3, was kindly gifted from Prof. M. Sawada of Nagoya University. The 6-3 cells were established from neonatal C57BL/6J (H-2b) mice using a non-enzymatic and non-virus-transformed procedure (Kanzawa et al. 2000). The 6-3 cells closely resemble primary cultured microglia (Sawada et al. 1998; Kanzawa et al. 2000). The 6-3 cells were cultured in Eagle’s minimal essential medium, 0.3% NaHCO3, 2 mM glutamine, 0.2% glucose, 10 g/mL insulin and 10% fetal calf serum, and then were maintained at 37°C in a 10% CO2 and 90% air atmosphere. One ng/mL mouse recombinant granulocyte macrophage colony stimulating factor was supplemented in the culture medium to maintain the 6-3 cells because these cells stopped proliferating in its absence (Kanzawa et al. 2000). The culture media were renewed twice per week.

Primary mixed cells were prepared from the whole brain of the 3-day-postnatal Sprague-Dawley rats, using Cell Strainer (BD Falcon, Franklin Lakes, NJ). Primary rat microglial cells were selected after attachment to Aclar film (Nisshin EM, Tokyo, Japan) for 2 h in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (10% FBS/Dulbecco’s modified Eagle’s medium). Aclar films were slightly washed by phosphate-buffered saline and then transferred to fresh 10% FBS/Dulbecco’s modified Eagle’s medium, and the fresh microglia was expanded for 1–2 days. The purity of the isolated microglia was assessed by immunocytochemical staining for microglial marker, Iba-1, and > 99% of cells were stained positively.

Rat pheocromocytoma PC12 cells were cultured in Eagle’s minimal essential medium, 0.3% NaHCO3, 2 mM glutamine, 0.2% glucose, 10 g/mL insulin and 10% fetal calf serum, and then were maintained at 37°C in a 10% CO2 and 90% air atmosphere.

Reverse transcription-polymerase chain reaction (RT-PCR)

Total RNA was isolated from the 6-3 murine microglial cells using the RNeasy Mini Kit (Quiagen). RT-PCR was performed with RNA and gene-specific primer for dopamine D2 receptors (forward: 5′-GCAGCCGAGCTTTCAGGGCC-3′ and reverse: 5′-GGGATGTTGCAGTCACAGTG-3′) as previously described by Lemmer et al. (Lemmer et al. 2002) and reagents included in the SuperScript III RT-PCR System (invitrogen). Reverse transcription and Amplification was carried out in a gradient cycler (Biometra). The reaction mixture was incubated at 94°C for 2 min to fully activate the Taq DNA polymerase, then followed by a touchdown protocol of denaturing at 94°C for 15 s, annealing from 94°C down to 66°C for 30 s, and extension at 68°C for 1 min in 30 cycles. Finally, a 5-min extension at 68°C was conducted. PCR products were resolved by electrophoresis in 2% agarose gels, stained with ethidium bromide, and photographed. The predicted size was checked by a 100 bp DNA ladder (Biobeer).

NO and TNF-α release assessment

The 6-3 cells were plated on 96-well tissue culture plates at 1 × 105 cells per 200 μL per well and then were pre-incubated in the presence or absence of aripiprazole or D2 receptor full agonist, quinpirole for 12 h and then incubated in the presence or absence of 50 U/mL IFN-γ or 1 μg/mL LPS at 37°C. After 48 h, the collected media were assayed for NO or TNF-α accumulation. NO or TNF-α release into the culture medium was measured using a Griess reaction assay kit (Dojindo, Kumamoto, Japan) or a mouse TNF-α enzyme-linked immunosorbent assay (ELISA) kit (Biosource International, Camarillo, CA, USA), respectively. The absorbance of Griess reaction or ELISA was read at 540 nm or 450 nm, respectively, using a plate reader (Labsystems Multiscan MS, Frankfurt, Germany). Rat primary microglial cells were plated on 96-well tissue culture plates at 1 × 104 cells per 100 μL per well and then were pre-incubated in the presence or absence of aripiprazole or quinpirole for 12 h and then were incubated in the presence or absence of 50 U/mL IFN-γ or 1 μg/mL LPS at 37°C. After 48 h, the collected media were assayed for NO accumulation as described above.

Intracellular Ca2+ imaging

The experiments were performed in HEPES buffer (150 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose and 10 mM HEPES, pH 7.4 with Tris-OH) at room temperature (25°C). Intracellular Ca2+ concentration ([Ca2+]i) was monitored using fura-2 (AM) (Grynkiewicz et al. 1985; Mizoguchi et al. 2002). The 6-3 cells plated on glass-base dish were loaded with 5 μM fura-2 AM (Dojindo, Wako, Japan) for 20 min and washed three times with HEPES buffer before the measurement. During the measurement using an inverted microscope (20×; Olympus IX70-22FL, Olympus Co. Tokyo, Japan), external HEPES buffer was constantly perfused (10 mL/min). For fura-2 excitation, the cells were illuminated with two alternating wavelengths, 340 and 380 nm using a computerized system. The emitted light was collected at 510 nm using a cooled CCD camera (C4742-95ER, Hamamatsu Photonics, Hamamatsu, Japan) and images were stored every 5 s. These series of sequential data were analyzed using the AquaCosmos software package (Hamamatsu photonics, Hamamatsu, Japan). The [Ca2+]i was calculated from the ratio (R) of fluorescence recorded at 340 and 380 nm excitation wavelengths for each pixel within a cell boundary (AquaCosmos software). Calibrations (conversion of R340/380 values into calcium concentrations) were performed as described previously (Grynkiewicz et al. 1985). Basal [Ca2+]i was determined from the initial 10 images of each cell recording. A [Ca2+]i signal was defined as an increase in R 340/380 with clear time correlation to the application of IFN-γ.

Cell viability

Cell viability was determined by colorimetric measurements of the reduction product of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetra-zolium bromide (MTT). After treatment with or without aripiprazole, the original medium was removed from the 96-well plates, and the cells were incubated for 2 h at 37°C in the presence of phenol red free minimum essential medium (Invitrogen Corporation, NY.) containing 0.5 mg/mL MTT. A 100 μL MTT lysis buffer (5% sodium dodecylsulfate and 5 mM HCl) was then added to each well, and the plates were incubated at 37°C overnight to dissolve the formazan that had formed in the wells. MTT is reduced to formazan in the mitochondria of living cells. Reduced MTT was measured by means of a plate reader (Labsystems Multiscan MS, Frankfurt, Germany) at a wavelength of 570 nm.

Statistics

All data are represented as the means ± SEM and they were analyzed by a one-way analysis of variance (anova) followed by Fisher’s PLSD post hoc test for specific comparisons. Significance was established at a level of p < 0.05.

Results

The effect of aripiprazole on NO and TNF-α release by IFN-γ-activated microglia

We found the PCR products of the dopamine D2L receptors mRNA, thus indicating that the 6-3 murine microglial cells express dopamine D2 receptors (Fig. 1).

Figure 1.

 The expression of dopamine D2 receptor mRNA in murine microglial cells. PCR using cDNA from the murine 6-3 microglial cells was performed. Agarose gel electrophoresis of PCR amplified products of dopamine D2 receptors. The band corresponds to mRNA of the dopamine D2L receptors (529 bp).

Interferon-γ significantly induced NO and TNF-α release from the 6-3 murine microglial cells as which is consistent with previous reports (Kato et al. 2007). The 6-3 cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or D2 receptor full agonist, quinpirole (5, 10 and 20 μM) for 12 h, then the cells were treated with each drug and IFN-γ (50 U/mL) for 48 h. Aripiprazole significantly inhibited the NO release dose-dependently in comparison with the positive control (DMSO + IFN-γ group), while quinpirole did not inhibit the NO release (Fig. 2a). In order to confirm whether these effects are specific to IFN-γ–induced microglial activation or not, we measured the effect of aripiprazole on LPS-induced microglial activation. Aripiprazole significantly inhibited NO release, while, quinpirole did not inhibit such NO release by LPS-activated 6-3 microglia (Fig. 2b). The degree of inhibition of NO release was less in LPS- than in IFN-γ– activated 6-3 microglia. These results suggest that the inhibitory effects of aripiprazole were not specific to IFN-γ receptor-mediated signaling. In addition, we also prepared rat primary microglial cells in order to confirm the relevance of our results in these cells. Similar to what was observed in the 6-3 murine microglial cells, aripiprazole significantly inhibited NO release from LPS-activated rat primary microglia, while, quinpirole did not inhibit NO release by LPS-activated rat primary microglia (Fig. 2c). The treatments of aripiprazole or quinpirole with or without LPS did not have any effect on the cell viability in these experiments (data not shown). These results suggest that the inhibitory effects of aripiprazole on microglial activation were not specific to the 6-3 microglia.

Figure 2.

 The effect of aripiprazole on NO release by activated microglia (a) & (b). The 6-3 microglial cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 12 h, then the cells were treated with each drug and IFN-γ (50 U/mL) (a) or LPS (1 μg/mL) (b) for 48 h. NO release were determined using the Griess assay. The results were expressed as the percentage values taking the IFN-γ or LPS + DMSO treatment group as 100%. All Data are represented as the means (SEM) of three independent experiments (n = 15). ###p < 0.001 in comparison to the control DMSO treatment group. *p < 0.05; **p < 0.01; ***p < 0.001 in comparison to the IFN-γ or LPS + DMSO treatment group. (c) Rat primary microglial cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 12 h, then the cells were treated with each drug and LPS (1 μg/mL) for 48 h. NO release were determined using the Griess assay. The results were expressed as the percentage values taking the LPS + DMSO treatment group as 100%. All Data are represented as the means (SEM) of three independent experiments (n = 15). ###p < 0.001 in comparison to the control DMSO treatment group. *p < 0.05; ***p < 0.001 in comparison to the LPS + DMSO treatment group.

The 6-3 cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 12 h, then the cells were treated with each drug and IFN-γ (50 U/mL) for 48 h. Aripiprazole strongly inhibited the release of TNF-α dose-dependently. On the other hand, quinpirole had no inhibitory effect on the release of TNF-α (Fig. 3). Aripiprazole or quinpirole itself had no effect on releasing NO and TNF-α without IFN-γ treatment (data not shown).

Figure 3.

 The effect of aripiprazole on TNF-α release by IFN-γ-activated microglia. The 6-3 microglial cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 12 h, then the cells were treated with each drug and IFN-γ (50 U/mL) for 48 h. TNF-α release were determined using the ELISA, respectively. The results were expressed as the percentage values taking the IFN-γ + DMSO treatment group as 100%. All Data are represented as the means (SEM) of three independent experiments (n = 15). ###p < 0.001 in comparison to the control DMSO treatment group. ***p < 0.001 in comparison to the IFN-γ + DMSO treatment group.

The intracellular signaling mechanism in NO and TNF-α release from IFN-γ-activated microglia

The 6-3 murine microglial cells were pre-treated with 1 μM BIM I (PKC inhibitor), 20 μM SB203580 (p38 MAPK protein inhibitor) or 20 μM PD98059 (MAPK protein inhibitor, for ERK) for 1 h, then the cells were treated with each drug and IFN-γ (50 U/mL) for 24 h. BIM I, SB203580 and PD98059 significantly inhibited NO and TNF-α release from IFN-γ-activated microglia (Fig. 4a and b). In comparison to the positive controls (DMSO), the specific inhibition of PKC, p38 MAPK and ERK was observed to lead to a decrease in the NO release to 27.3 ± 2.84%, 60.6 ± 5.29% and 61.7 ± 6.14%, respectively, while the specific inhibition of PKC, p38 MAPK and ERK was found to lead to a decrease in TNF-α release to 65.3 ± 2.72%, 66.0 ± 1.48% and 71.9 ± 2.25%, respectively. Each inhibitor treatment did not have any effect on the cell viability (data not shown). These results suggest that the PKC, p38 MAPK and ERK pathways all play key roles in IFN-γ-induced activation of the 6-3 murine microglial cells.

Figure 4.

 The effect of inhibitors on NO and TNF-α release by IFN-γ-activated microglia. NO (a) and TNF-α (b) release were determined using the Griess assay and ELISA, respectively. The 6-3 microglial cells were pre-treated with 1 μM BIM I (PKC inhibitor), 20 μM SB203580 (p38 MAPK protein inhibitor) or 20 μM PD98059 (MAPK protein inhibitor, for ERK) for 1 h, then the cells were treated with each drug and IFN-γ (50 U/mL) for 24 h. The results were expressed as the percentage values taking the IFN-γ + DMSO treatment group as 100%. All Data are represented as the means (SEM) of three independent experiments (n = 15). ###p < 0.001 in comparison to the DMSO control group. ***p < 0.001 in comparison to the IFN-γ + DMSO treatment group.

Aripiprazole attenuates the mobilization of intracellular Ca2+ induced by IFN-γ-activated microglia

In human microglia, IFN-γ rapidly induces a progressive increase in [Ca2+]i and IFN-γ acts solely through influx of Ca2+ (Franciosi et al. 2002) and [Ca2+]i is very important for the regulation of microglial activation including the release of NO and cytokines (Hoffmann et al. 2003).

We measured the effect of 12 h treatment with quinpirole (10 μM) or aripiprazole (5 μM) on the mobilization of intracellular Ca2+ induced by IFN-γ application in the 6-3 cells. As shown in Fig. 5(a), IFN-γ (50 U/mL) rapidly increased [Ca2+]i in the 6-3 cells (pretreated with 0.025% DMSO for 12 h; n = 10 cells). Once the intracellular Ca2+ rose, it gradually increased without attenuation. The increase in intracellular Ca2+ was sustained > 40 min after the washout of IFN-γ until the end of recording. Therefore, IFN-γ induced sustained intracellular Ca2+ elevation in the murine microglia as previously shown in human microglia (Franciosi et al. 2002).

Figure 5.

 The mobilization of intracellular Ca2+ induced by IFN-γ-activated microglia Sustained elevation of intracellular Ca2+ induced by IFN-γ application in the murine 6-3 microglial cells. Microglial cells were pre-treated with DMSO (0.025%) (a), quinpirole (10 μM) (b) or aripiprazole (5 μM) (c) for 12 h and then activated by IFN-γ (50 U/mL) for 5 min. Each panel shows the average trace of representative 15 traces of [Ca2+]i in each condition. The increase in the [Ca2+]i in response to IFN-γ application was calculated as the difference between basal [Ca2+]i and values obtained at 10 min after a 5-min treatment of IFN-γ (d). Data are represented as the means (SEM) of each experiment. ###p < 0.001 in comparison to the DMSO treatment group. The 6-3 microglial cells were pre-treated with BAPTA-AM (5 μM and 20 μM), then the cells were treated with IFN-γ (50 U/mL) for 24 h (e). The results were expressed as the percentage values taking the IFN-γ treatment group as 100%. Data are represented as the means (SEM) of three independent experiments (n = 15). *p < 0.05; **p < 0.01 in comparison to the IFN-γ treatment group.

In the 6-3 cells pre-treated with quinpirole (10 μM) or aripiprazole (5 μM) for 12 h, IFN-γ (50 U/mL) also induced sustained intracellular Ca2+ elevation (Fig. 5b and c; n = 23 for quinpirole and n = 28 for aripiprazole). However, the amplitude of [Ca2+]i increase induced by IFN-γ was very different between the cells pre-treated with quinpirole or aripiprazole. The treatment of quinpirole (10 μM for 12 h) did not affect the amplitude of increase in [Ca2+]i induced by IFN-γ in the cells (230.4 ± 26.8 nM in DMSO vs. 219.1 ± 34.8 nM in quinpirole; p = 0.76). In contrast, pre-treatment with aripiprazole (5 μM for 12 h) significantly reduced the amplitude of increase in [Ca2+]i at 10 min after a 5-min treatment of IFN-γ (DMSO vs. 80.0 ± 7.50 nM in aripiprazole; p < 0.001) (Fig. 5d).

The 6-3 cells were pre-treated with or without the membrane-permeable intracellular Ca2+ chelator, BAPTA-AM (5 μM and 20 μM), then the cells were treated with IFN-γ (50 U/mL) for 24 h. BAPTA significantly decreased NO release from IFN-γ-activated microglia (Fig. 5e) without cytotoxicity (MTT data not shown), which was the same as the previously reported findings of Hoffmann et al. (Hoffmann et al. 2003).

These results suggest that 12-h treatment of aripiprazole attenuates the mobilization of intracellular Ca2+ induced by IFN-γ application in the 6-3 murine microglial cells. The similar results were observed in rat primary microglial cells (data not shown).

Cell viability

The 6-3 cells were treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 60 h, or the 6-3 cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 12 h and then the cells were treated with each drug and IFN-γ (50 U/mL) for 48 h. Aripiprazole and quinpirole showed no significant microglial cytotoxicity at a concentration of less or equal 10 μM (Fig. 6a and b). IFN-γ treatment alone did not have any effect on the cell viability as previously demonstrated (Kato et al. 2007). The PC12 cells were treated with DMSO (0.1%) or aripiprazole (5, 10 and 20 μM) for 60 h. Aripiprazole showed no significant neuronal cytotoxicity under a concentration of 20 μM (Fig. 6c).

Figure 6.

 Cell viability. Cell viability was determined by colorimetric measurements of the reduction product of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetra-zolium bromide (MTT). The 6-3 microglial cells were treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 60 h (a), or the 6-3 microglial cells were pre-treated with DMSO (0.1%), aripiprazole (5, 10 and 20 μM) or quinpirole (5, 10 and 20 μM) for 12 h and then the cells were treated with each drug and IFN-γ (50 U/mL) for 48 h (b). The results were expressed as percentage values while taking the DMSO treatment group as 100% (a) or the DMSO plus IFN-γ treatment group as 100% (b), respectively. The PC12 cells were treated with DMSO (0.1%) or aripiprazole (5, 10 and 20 μM) for 60 h (c). The results were expressed as percentage values while taking the DMSO treatment group as 100% (c). All Data are represented as the means (SEM) of three independent experiments (n = 15), respectively. ***p < 0.001 in comparison to the DMSO treatment group.

Discussion

In the present study, the generation of NO and TNF-α from IFN-γ-activated microglia was significantly inhibited by aripiprazole. In addition, aripiprazole suppressed the IFN-γ-induced elevation of [Ca2+]i in microglia. Aripiprazole was not toxic either to microglial cells or neuronal cells at the concentrations where the effects described above were observed. The generation of NO and TNF-α from IFN-γ-activated microglia was significantly inhibited by specific inhibitors of PKC, p38 MAPK and ERK, respectively. In addition, aripiprazole significantly inhibited NO release from both LPS-activated 6-3 microglia and LPS-activated rat primary microglia while quinpirole did not inhibit the NO release by LPS-activated 6-3 microglia or LPS-activated rat primary microglia.

Interferon-γ, which is one of the typical activators of microglia along with lipopolysaccharide (LPS), has recently been reported to induce Ca2+ influx in human microglia (Franciosi et al. 2002). We also observed that IFN-γ induced sustained [Ca2+]i elevation in murine microglia. Hoffmann et al. demonstrated that the treatment of BAPTA can inhibit the release of NO and cytokines from LPS-activated microglia while ionomycin, an ionophore elevating Ca2+, has no effect on the release of NO or cytokines from LPS-activated microglia. They thus indicated that an increased amount of [Ca2+]i is required, but by itself is not sufficient, for the release of NO and certain cytokines from activated microglia (Hoffmann et al. 2003). We observed that the pre-treatment of aripiprazole attenuated the mobilization of intracellular Ca2+ induced by IFN-γ in murine microglia. Intracellular Ca2+ is one of the endogenous activators of PKC. Phorbol 12-myristate- 13-acetate, an activator of PKC, induces the activation of microglia (Nikodemova et al. 2006). In microglia, PKC has been reported to be an important initiator of the MAPK signaling pathway in the CNS. The activation of PKC affects MAPK cascade proteins including ERK 1/2 and p38 MAPK (Schonwasser et al. 1998). p38 MAPK plays a major role in the LPS-activated BV2 microglia while ERK 1/2 plays a major role in the IFN-γ activated BV2 microglia (Kim et al. 2004; Park et al. 2005). Based on these results, we can speculate that aripiprazole may inhibit IFN-γ-induced microglial activation through the suppression of IFN-γ-induced elevation of [Ca2+]i in microglia. The effect of aripiprazole on Ca2+ regulation shown in the present study is interesting because Ca2+ signaling dysfunction is proposed for the central unifying molecular pathology in schizophrenia (Lidow 2003). However, aripiprazole may modulate the expression of other factors that act upstream of calcium to dampen IFN-γ- induced signaling.

Labuzek et al. reported that chlorpromazine and loxapine, antipsychotics with D2 receptor antagonism, had inhibitory effect, while, quinpirole, D2 receptor full agonist, had little effect on the inflammatory cytokine release from LPS-activated rat primary microglia. They therefore suggested that microglia might not have functional dopaminergic receptors (Labuzek et al. 2005). However, Faber et al. provided the first evidence for the existence of functional dopaminergic receptors on rat primary microglia. In their study, quinpirole inhibited the release of NO from LPS-activated microglia (Farber et al. 2005). Hou et al. reported that even antipsychotics, all of which have D2 antagonism, had different effects on LPS-induced mouse N9 microglial activation. In their study, not either haloperidol or clozapine, but only olanzapine had an inhibitory effect on LPS-induced microglial activation. We recently reported that not haloperidol but several atypical antipsychotics, all of which have D2 antagonism, demonstrated a significantly inhibitory effect on IFN-γ–induced 6-3 microglial activation (Kato et al. 2007; Bian et al. 2008). In the present study, RT-PCR revealed the existence of dopamine D2 receptors while not quinpirole but aripiprazole demonstrated an inhibitory effect on NO release from IFN-γ- or LPS- activated 6-3 microglia. These results seem to suggest that the dopamine D2 receptors may not be involved in IFN-γ- or LPS- induced microglial activation. Aripiprazole has recently been reported to exert D2-receptor-mediated MAPK phosphorylation in transfected CHO cells (Bruins Slot et al. 2006; Urban et al. 2007). Therefore, the expression of the protein levels of dopamine D2 receptors as well as the functions of these receptors should be investigated to confirm the above speculations. Aripiprazole is also known to be both 5HT1A agonist and 5HT2A antagonist (Burris et al. 2002). Serotoninergic receptors might, therefore, be involved in the inhibitory effect of aripiprazole on IFN-γ-induced microglial activation. Microglia is known to express many kinds of neurotransmitter receptors including glutamatergic, GABAergic, purinergic, dopaminergic, cholinergic, adrenergic, and cannabinoid receptors (Pocock and Kettenmann 2007). However, to the best of our knowledge regarding serotonin receptors, there have been no reports regarding the expression of 5HT1A or 5HT2A receptors in the microglia, only one report for the existence of 5-HT7 in human microglial cells (Mahe et al. 2005). Atypical antipsychotics can have a positive effect on cell growth and survival through unique signaling pathways (Lu and Dwyer 2005). Therefore, the pharmacological basis for their effects does not appear to be directly related to only the dopaminergic and serotoninergic receptors.

Lipopolysaccharide is usually used as an activator of the microglia. LPS is a major constituent of the cell wall in gram negative bacteria and it is thus suitable to provide especially bacterial inflammatory reactions. However, there has been little evidence so far regarding the relationship between schizophrenia and bacterial infections. On the other hand, there have been some reports that suggested the relationship between schizophrenia and IFN-γ, a major immuno-activator in the CNS. IFN-γ is released by infiltrating T cells as well as from activated microglia in the CNS (Kawanokuchi et al. 2006). The most important immunological studies in schizophrenia have shown a shift from Th1-like cellular to Th2-like humoral immune reactivity and these studies have suggested a blunted IFN-γ signal in schizophrenia (Schwarz et al. 2001a,b). However, Rothermundt et al. have argued that the reduced IFN-γ production in vitro may reflect an increased production in vivo, as observed in several autoimmune disorders (Rothermundt et al. 2001). Furthermore, the serum levels of IL-2 and IFN-γ, and the production of these cytokines from peripheral blood mononuclear cells stimulated by phytohemagglutinin has been reported to be significantly higher in patients with schizophrenia than in controls (Cazzullo et al. 2001, 2002). We thus used IFN-γ as wells as LPS as an activator of microglia. Takeuchi et al. recently demonstrated that IFN-γ induced microglial-activation-induced cell death for the chronic treatment (Takeuchi et al. 2006). However, under the present experimental conditions, IFN-γ treatment did not have any effect on the cell viability of microglia.

The typical dose range of aripiprazole is 10–30 mg/day and the typical serum concentration or plasma range of aripiprazole is 0–1000 ng/mL (Alexopoulos et al. 2004; Chew et al. 2006). Antipsychotics are known to accumulate in brain tissue to levels that are 25–30 fold higher than serum levels (Baumann et al. 2004). Therefore, in spite of no evidence that the effect of a drug in cell culture could be compared to the effect of the same drug at a brain tissue level even in the same range of concentration, the concentrations of aripiprazole used in the present study might thus not be substantially different from the brain tissue levels for aripiprazole.

A recent post-mortem brain study has indicated a close relationship between schizophrenia and neurogenesis (Reif et al. 2006). Clozapine, olanzapine and risperidone have been reported to stimulate neurogenesis in vivo (Wakade et al. 2002; Halim et al. 2004; Wang et al. 2004). However, the direct effect of antipsychotics on the neurosphere have recently been reported to be limited (Councill et al. 2006). Inflammatory cytokines and NO have reported to have an inhibitory effect on neurogenesis (Monje et al. 2003; Cardenas et al. 2005; Iosif et al. 2006; Kaneko et al. 2006). Aripiprazole may thus indirectly stimulate neurogenesis through the inhibition of the microglia-derived release of inflammatory cytokines and NO.

Structural imaging studies, as well as gene expression studies and evidence for the dysfunction of myelin and oligodendrocyte, have suggested the presence of abnormalities of white matter in schizophrenia (Hakak et al. 2001; Davis et al. 2003; Uranova et al. 2004; Schlosser et al. 2007). Microglial activation in the CNS has been implicated in the pathogenesis of white matter disorders. Activated microglia has reported to induce cytotoxicity of oligodendrocytes via the release of NO, peroxynitrite and inflammatory cytokines such as TNF-α (Merrill et al. 1993; Buntinx et al. 2004; Li et al. 2005). Therefore, the present results suggest that aripiprazole may ameliorate white matter disorders via inhibiting microglial activation in the brain of patients with schizophrenia.

The present study demonstrated that not only atypical antipsychotics which have dopamine D2 receptor antagonism but also aripiprazole have significantly anti-inflammatory effects via the inhibition of microglial activation. Our results might therefore shed some new light on the understanding of the pathophysiology of schizophrenia and the therapeutic strategies for the treatment of schizophrenia with anti-inflammatory/immunosuppressive agents. For further studies, the more detailed molecular mechanism of the inhibitory effect of aripiprazole on microglial activation should be clarified while in vivo studies to confirm the present results should also be performed.

Acknowledgements

The present study was supported partly by a grant-in-aid from the Japan Society for the Promotion of Science. The authors thank Prof. Makoto Sawada of Nagoya University for providing us with the microglial cell line, 6-3. The authors also thank Dr. Makiko Kido, Shuji Fukagawa and Leo Gotoh of Kyushu University for valuable technical advices.

Contributors

All authors contributed substantially to the scientific process leading up to the writing of the present paper. AM, the principal investigator of the present research, and TK made the conception and design of the project and wrote the protocol. The performance of experiments and the analysis and interpretation of data were done by TK, YM, HH, SS, YS and SH. TK wrote the first draft of the manuscript. The critical revision of the manuscript was made by IT and SK. All authors contributed to and have approved the final manuscript.

Ancillary