N-Acetylcysteine Suppresses Microglial Inflammation and Induces Mortality Dose-Dependently via Tumor Necrosis Factor-α Signaling

N-acetylcysteine (NAC) is an antioxidant that prevents tumor necrosis factor (TNF)-α-induced cell death, but it also acts as a pro-oxidant, promoting reactive oxygen species independent apoptosis. Although there is plausible preclinical evidence for the use of NAC in the treatment of psychiatric disorders, deleterious side effects are still of concern. Microglia, key innate immune cells in the brain, play an important role in inflammation in psychiatric disorders. This study aimed to investigate the beneficial and deleterious effects of NAC on microglia and stress-induced behavior abnormalities in mice, and its association with microglial TNF-α and nitric oxide (NO) production. The microglial cell line MG6 was stimulated by Escherichia coli lipopolysaccharide (LPS) using NAC at varying concentrations for 24 h. NAC inhibited LPS-induced TNF-α and NO synthesis, whereas high concentrations (≥30 mM) caused MG6 mortality. Intraperitoneal injections of NAC did not ameliorate stress-induced behavioral abnormalities in mice, but high-doses induced microglial mortality. Furthermore, NAC-induced mortality was alleviated in microglial TNF-α-deficient mice and human primary M2 microglia. Our findings provide ample evidence for the use of NAC as a modulating agent of inflammation in the brain. The risk of side effects from NAC on TNF-α remains unclear and merits further mechanistic investigations.


Introduction
Inflammation involves homeostasis and disease processes, including diabetes, cancer, heart disease, arthritis, neurological disease, and psychiatric disorders [1,2]. The relevance of inflammation in those conditions has been proposed to be linked with alterations of production in cytokines and reactive oxygen species (ROS) [1]. N-acetylcysteine (NAC) is an acetylated variant and precursor of the amino acid L-cysteine [3], has excellent effects on inflammation [4] and ROS [5], and has been used for decades as a nutritional supplement and low-cost medication for various ailments [6]. In the central nervous system (CNS), NAC is an important antioxidant [7] that promotes the inhibition of the nitric oxide synthase (iNOS) enzyme, resulting in a decrease in the production of tumor necrosis factor

Effect of LPS on Microglial Activation and Viability
To determine the dose-response relationship and time course of E. coli LPS-induced microglial TNFα production, MG6 cells were stimulated with 0, 10, 100, and 1000 ng/mL LPS over a time course of 0, 6, 12, and 24 h. The protein levels of the major inflammatory cytokine TNF-α in the MG6 culture supernatant were determined. Two-way ANOVA revealed significant effects of LPS treatment (F 3, 20 = 1987; p < 0.0001) and culture time (F 1.756, 35.12 = 992.4; p < 0.0001) with a significant interaction (treatment × time: F 9, 60 = 383.9; p < 0.0001). In subsequent Bonferroni post-hoc analyses, the LPS groups with three different doses (10, 100, and 1000 ng/mL) significantly induced higher TNF-α production than the control groups of the three respective periods (6,12, and 24 h) ( Figure 1A). Among the varying doses of LPS, 1000 ng/mL induced microglial TNF-α production over 700 times more than the controls.
We used microglial cells as a model system, which produce TNF-α and NO synthesis upon LPS stimulation. This study aimed to: (1) investigate the dose-response and time course of the effect of LPS on TNF-α and NO production, and (2) identify the beneficial and deleterious effects of NAC using varying doses that are commonly used in in vitro and in vivo research. Our findings highlighted the potential applications of NAC in the treatment of psychiatric diseases.

Effect of LPS on Microglial Activation and Viability
To determine the dose-response relationship and time course of E. coli LPS-induced microglial TNFα production, MG6 cells were stimulated with 0, 10, 100, and 1000 ng/mL LPS over a time course of 0, 6, 12, and 24 h. The protein levels of the major inflammatory cytokine TNF-α in the MG6 culture supernatant were determined. Two-way ANOVA revealed significant effects of LPS treatment (F3, 20 = 1987; p < 0.0001) and culture time (F1.756, 35.12 = 992.4; p < 0.0001) with a significant interaction (treatment × time: F9, 60 = 383.9; p < 0.0001). In subsequent Bonferroni post-hoc analyses, the LPS groups with three different doses (10, 100, and 1000 ng/mL) significantly induced higher TNF-α production than the control groups of the three respective periods (6,12, and 24 h) ( Figure 1A). Among the varying doses of LPS, 1000 ng/mL induced microglial TNF-α production over 700 times more than the controls.  dose-response relationship of NAC with LPS-induced Il1b mRNA expression in MG6 cells (n = 6). * p < 0.05, *** p < 0.001, vs. LPS only. (G) The dose-response relationship of NAC with LPS-induced IL-1 synthesis in the medium of MG6 cells (n = 6) after 24 h. *** p < 0.001, vs. LPS only. (H) The dose-response relationship of NAC with LPS-induced Il10 mRNA expression in MG6 cells after 24 h (n = 6). * p < 0.05, vs. control. All data are presented as MEAN ± SEM. Furthermore, we assessed the viability of MG6 cells under LPS stimulation. After 24 h of incubation, the viability of MG6 cells was significantly decreased at 1000 ng/mL (p < 0.05), but not at lower doses (10 and 100 ng/mL) (F 3, 20 = 3.744; p = 0.0278) ( Figure 1B). Since LPS at 100 ng/mL showed a tendency to increase mortality of MG6 (p = 0.101), the 10 ng/mL of LPS was used in further experiments.

High Concentration of NAC Increased Microglial Mortality without LPS Challenge
Furthermore, we investigated the effect of NAC and LPS on the viability of microglia. MG6 cells were treated with LPS (10 ng/mL) either alone or in combination with NAC (5, 10, 20, 30, and 60 mM). After 24 h of NAC treatment, a one-way ANOVA followed by Bonferroni post-hoc revealed a significant reduction in viable cells at 30 mM (p < 0.05) and 60 mM (p < 0.001) NAC (F 7, 40 = 5.277; p = 0.01) (Figure 2A,B). Although the levels of LPS-induced inflammatory cytokines were decreased at 30 and 60 mM NAC, a high dose of NAC increased MG6 mortality with LPS (Figure 2A,B). After 24 h coincubation of LPS with NAC at 60 mM, the protein levels of cellular TNF-α were not detected in MG6 by western blotting, according to the highest mortality rate.
Furthermore, C57BL/6J mice were injected intraperitoneally with 20 mM or 30 mM NAC for two days, which is lower than in previous studies [30,47]. On day 3 following treatment with 30 mM NAC, there was a significant decrease in Iba-1 immunoreactivity (p < 0.05, Figure 2C) (various magnifying power 10×, 20×, and 60×) in the prefrontal cortex (PFC) compared to the saline group, but not in 20 mM NAC-treated group. Figure 2C shows that following the administration of two doses of 30 mM NAC, the microglial cell bodies in PFC disappeared and their dendritic branches were degraded (60×). Furthermore, C57BL/6J mice were injected intraperitoneally with 20 mM or 30 mM NAC for two days, which is lower than in previous studies [30,47]. On day 3 following treatment with 30 mM NAC, there was a significant decrease in Iba-1 immunoreactivity (p < 0.05, Figure 2C) (various magnifying power 10×, 20×, and 60×) in the prefrontal cortex (PFC) compared to the saline group, but not in 20 mM NAC-treated group. Figure 2C shows that following the administration of two doses of 30 mM NAC, the microglial cell bodies in PFC disappeared and their dendritic branches were degraded (60×).

Role of Nitric Oxide in NAC-Induced Microglial Mortality
The inorganic anions nitrite (NO 2 − ) and nitrate (NO 3 − ) are end-products of nitric oxide (NO) metabolism [48]. Several studies have demonstrated the potential protective effect of NO [49]. A previous study demonstrated that NAC at a dose of 1000 mg/kg significantly increased the serum NO 2 − /NO 3 − in rats [30]. Here, we examined NO synthesis from MG6 cells by measuring NO x (NO 2 − /NO 3 − ). NO x synthesis was significantly increased by LPS (p < 0.0001) but not by NAC alone in high concentrations (30 and 60 mM) in comparison to the control (p > 0.05) (F 11, 36 = 60.54; p < 0.0001) ( Figure 2D). Furthermore, NAC (5, 10, 20, 30, and 60 mM) significantly inhibited LPS-induced NO x synthesis (p < 0.0001) ( Figure 2D), indicating that NO may not be associated with microglial mortality.

Effects of NAC on Acute and Chronic Stress-Induced Behavior
Although NAC at 30 mM decreased the numbers of microglia in the PFC, pretreatment with NAC at 32.64 mg/kg (20 mM) or 48.96 mg/kg (30 mM) did not affect freezing behavior without footshock (p > 0.05; Figure 3A) (F 8, 45 = 34.40; p < 0.0001). Acute footshock stressinduced significantly longer freezing times after short-duration re-exposure (FS3) (p < 0.001; Figure 3A). In contrast, foot-shocked mice with long-duration re-exposure (FS30) had a significantly longer freezing time than those without footshock (vs. Con; p < 0.001) but shorter than re-exposed for a short time after footshock (vs. FS3; p < 0.001) ( Figure 3A). Pretreatment with 20 mM or 30 mM NAC had no effect on short-or long-duration freezing behaviors (p > 0.05) ( Figure 3A). Furthermore, we investigated the effect of NAC on chronic social defeat stress (SDS)-induced depressive behaviors in mice ( Figure 3B). Following chronic SDS exposure, immobility time in the forced swimming test was significantly longer than in the controls (p < 0.01; Figure 3C) (F 5, 60 = 11.63; p < 0.0001), and the sucrose preference was decreased as well (p < 0.05; Figure 3D) (F 5, 54 = 7.975; p < 0.0001). NAC pretreatment did not prevent chronic SDS-induced increased immobility time in forced swimming and decreased sucrose preference at both 20 mM and 30 mM (vs. SDS; p > 0.05; Figure 3C,D).
of NAC pretreatment (20 and 30 mM) on sucrose preference after chronic social defeat stress. * p < 0.05, ** p < 0.01, vs. non-stressed control mice. All data are presented as MEAN ± SEM.

Discussion
In this study, we tested various doses of NAC for their ability to inhibit LPS-induced synthesis of cytokines and NO in mouse and human microglial cells. LPS significantly increased the transcription of pro-inflammatory cytokines TNF-α and IL-1 but not the anti-inflammatory cytokine IL-10. Moreover, the increased TNF-α and NO levels caused by the low dose of LPS did not affect cellular viability. NAC inhibited LPS-induced TNFα, IL-1, and NO synthesis in a dose-dependent manner. Low NAC concentrations inhibited LPS-induced TNF-α, IL-1, and NO synthesis in MG6. However, high concentration of NAC (≥30 mM) induced cellular TNF-α aggregation in MG6 and caused cell death in MG6, SK-N-SH, U-87 MG, and PFC microglial cells, but not in human primary M2 microglial and mouse TNF-α-deficient microglial cells ( Figure 6).

Discussion
In this study, we tested various doses of NAC for their ability to inhibit LPS-induced synthesis of cytokines and NO in mouse and human microglial cells. LPS significantly increased the transcription of pro-inflammatory cytokines TNF-α and IL-1 but not the anti-inflammatory cytokine IL-10. Moreover, the increased TNF-α and NO levels caused by the low dose of LPS did not affect cellular viability. NAC inhibited LPS-induced TNF-α, IL-1, and NO synthesis in a dose-dependent manner. Low NAC concentrations inhibited LPS-induced TNF-α, IL-1, and NO synthesis in MG6. However, high concentration of NAC (≥30 mM) induced cellular TNF-α aggregation in MG6 and caused cell death in MG6, SK-N-SH, U-87 MG, and PFC microglial cells, but not in human primary M2 microglial and mouse TNF-α-deficient microglial cells ( Figure 6). NAC has been widely used in clinics for over 50 years [50]. NAC is the N-acetyl derivative of the amino acid L-cysteine. It is rapidly absorbed, reaching a peak plasma concentration of 50% after 4 h of oral administration. Due to its short half-life (5.58 h), it is recommended that dosing be divided into two daily doses [51]. L-cysteine rapidly oxidizes to free cystine and regulates glutamate antiporter activity, which is considered the key to therapeutic efficacy in the brain [52]. These effects probably occurred directly in the brain, as NAC crosses the blood-brain barrier and accumulates in the brain [45]. In the majority of previous studies, NAC has been used to inhibit the production of proinflammatory cytokines and reduce cytotoxic levels of NO by inhibiting the synthesis of TNF-α and iNOS. However, the multiple efficacies and toxicological effects of NAC must be highlighted in clinical and animal research. For instance, NAC treatment may induce gastrointestinal and dermatological disorders [41]. NAC treatment can also cause hepatic damage, renal failure, and death in patients with acute carbon tetrachloride poisoning [50]. When compared to controls, NAC treatment (10 mg/kg) for one week after eccentric exercise-induced muscle damage significantly increased tissue damage and oxidative stress [42]. Chronic systemic administration of NAC caused pulmonary arterial hypertension in mice [43]. Continuous infusion of high-dose NAC during a lipopolysaccharide challenge increased mortality in rats [44].
In murine models, recent studies found that pre-exposure to NAC at 2.5 mM for 24 h in N9 microglia cells could entirely prevent the Hg 2+ -induced transcription of Tnf, Il1b, and Nos2 [53]. In vitro, NAC concentrations less than 2.5 mM affected voltage-gated sodium and potassium channels, whereas high doses of NAC inhibited the action potential of rat sciatic nerve fibers [54]. The molecular mechanism by which NAC provides neuroprotection at low doses but causes neurotoxicity at high doses is unknown. According to one theory, a higher dose of NAC could eliminate all the available ROS required for the normal functioning of the cells [54]. However, an in vivo study revealed that pretreatment with 1000 mg/kg NAC elevated serum NO2 − /NO3 − , indicating that NAC has direct scavenging effects on ROS and beneficial effects in sodium taurocholateinduced acute pancreatitis in rats [30]. Our findings revealed that treatment with NAC (≥30 mM) alone had no effect on NO2 − /NO3 − synthesis but did cause microglia death. Taken together, NO combined with ROS is considered either toxic or protective, depending on the circumstances.
TNF-α is a major proinflammatory cytokine that plays a critical role in both homeostatic and pathophysiological states in epithelial cells, macrophages, microglia, NAC has been widely used in clinics for over 50 years [50]. NAC is the N-acetyl derivative of the amino acid L-cysteine. It is rapidly absorbed, reaching a peak plasma concentration of 50% after 4 h of oral administration. Due to its short half-life (5.58 h), it is recommended that dosing be divided into two daily doses [51]. L-cysteine rapidly oxidizes to free cystine and regulates glutamate antiporter activity, which is considered the key to therapeutic efficacy in the brain [52]. These effects probably occurred directly in the brain, as NAC crosses the blood-brain barrier and accumulates in the brain [45]. In the majority of previous studies, NAC has been used to inhibit the production of proinflammatory cytokines and reduce cytotoxic levels of NO by inhibiting the synthesis of TNF-α and iNOS. However, the multiple efficacies and toxicological effects of NAC must be highlighted in clinical and animal research. For instance, NAC treatment may induce gastrointestinal and dermatological disorders [41]. NAC treatment can also cause hepatic damage, renal failure, and death in patients with acute carbon tetrachloride poisoning [50]. When compared to controls, NAC treatment (10 mg/kg) for one week after eccentric exercise-induced muscle damage significantly increased tissue damage and oxidative stress [42]. Chronic systemic administration of NAC caused pulmonary arterial hypertension in mice [43]. Continuous infusion of high-dose NAC during a lipopolysaccharide challenge increased mortality in rats [44].
In murine models, recent studies found that pre-exposure to NAC at 2.5 mM for 24 h in N9 microglia cells could entirely prevent the Hg 2+ -induced transcription of Tnf, Il1b, and Nos2 [53]. In vitro, NAC concentrations less than 2.5 mM affected voltage-gated sodium and potassium channels, whereas high doses of NAC inhibited the action potential of rat sciatic nerve fibers [54]. The molecular mechanism by which NAC provides neuroprotection at low doses but causes neurotoxicity at high doses is unknown. According to one theory, a higher dose of NAC could eliminate all the available ROS required for the normal functioning of the cells [54]. However, an in vivo study revealed that pretreatment with 1000 mg/kg NAC elevated serum NO 2 − /NO 3 − , indicating that NAC has direct scavenging effects on ROS and beneficial effects in sodium taurocholate-induced acute pancreatitis in rats [30]. Our findings revealed that treatment with NAC (≥30 mM) alone had no effect on NO 2 − /NO 3 − synthesis but did cause microglia death. Taken together, NO combined with ROS is considered either toxic or protective, depending on the circumstances.
TNF-α is a major proinflammatory cytokine that plays a critical role in both homeostatic and pathophysiological states in epithelial cells, macrophages, microglia, astrocytes, and dendritic cells [55]. Overexpression of TNF-α induces apoptosis and necroptosis in the pathogenesis of inflammatory diseases [56]. Apart from pathogenic effects, TNF has been shown to maintain homeostatic expansion and protect against pathogens. For instance, physiological levels of glial TNF-α are required for synaptic scaling, which adjusts the strength of the synapse [57]. TNF desensitizes macrophages to the deleterious effects of secondary inflammatory challenges through tolerization [58]. In the CNS, TNF-α promotes the proliferation of oligodendrocytes progenitors and neuronal remyelination [59]. TNFα-deficient displayed improved spatial memory and learning abilities [60,61]. Microglial TNF-α has been implicated in lowering cocaine sensitivity via dopamine receptors [62]. We found that a high dose of NAC (≥30 mM) with LPS reduced medium TNF-α release, whereas increased cellular TNF-α aggregation caused MG6 mortality. Those NAC-induced mortalities were alleviated in TNF-α deficient human and mouse microglia. Here, we present molecular mechanisms underlying the homeostatic functions of microglial TNF-α in NAC-induced cell death that were not associated with NO synthesis (Figure 6). Additionally, NAC (30 mM, i.p.) also caused microglial mortality in the brain without altering acute stress-induced freezing behavior or chronic stress-induced depressive-line behaviors. Although microglia are essential for normal brain development, Cx3cr1-deficient [63] and microglia knockout mice [64] did not exhibit severe abnormal behaviors. Further research should discuss microglia and microglial TNF-α deficiency in regular behavior changes in the future. Taken together, our findings established the effective dose of NAC for protecting inflammation in the CNS and demonstrated that the homeostatic function of TNF-α is associated with NAC-induced microglial mortality.

Materials and Methods
All experimental procedures were performed in accordance with the Guidelines for the Care of Laboratory Animals of Tohoku University Graduate School of Medicine (Sendai, Japan).

Chemicals and Treatment
To determine the dose-response relationship and time course of Escherichia coli (E. coli) LPS-induced microglial TNFα production, MG6 cells were treated with LPS from Escherichia coli O111:B4 (Sigma-Aldrich, St. Louis, MO, USA, Product Number: L2630) at 0, 10, 100, and 1000 ng/mL to stimulate the release of TNF-α in the cell supernatants were measured by enzyme-linked immunosorbent assay (ELISA) at 0, 6, 12, 24 h, respectively. Furthermore, MG6, human primary M2 microglia, SK-N-SH, and U87-MG cells were incubated in a medium alone (control) or with 10 ng/mL E. coli LPS. The effects of NAC (FUJIFILM Wako, Tokyo, Japan, CAS RN ® : 616-91-1) (0, 5, 10, 30, and 60 mM) on the viabilities of each cell, transcription, and protein levels of cytokines, and NO productions from MG6 and human primary M2 microglia cells were assessed with coincubation with LPS (10 ng/mL) for 24 h. The non-LPS stimulated group received the same volume of phosphate-buffered saline (GIBCO, Carlsbad, CA, USA). Since NAC is a highly acidic compound (pH 2.2), we adjusted the pH of the NAC solution and culture medium to 8.0.
Tamoxifen (Cayman Ann Arbor, MI, USA) 20 mg were suspended in 100 µL 100% ethanol (FUJIFILM Wako), then dissolved in 900 µL corn oil (FUJIFILM Wako) at a concentration of 20 mg/mL by shaking 95 • C 1 min and 37 • C 1 h [65]. Tamoxifen was injected intraperitoneally (i.p.) for 5 doses of 5 mg with a separation of 48 h between doses. TNF fl/fl Cx3cr1-Cre ER mice were i.p. with NAC (20 or 30 mM) for 2 days after 25 days of the last tamoxifen treatment [66].

Animals
All experiments were conducted on male C57BL/6J mice aged 8 to 12 weeks. Mice were purchased from Japan SLC, Inc. (Shizuoka, Japan) and were individually housed and kept on a 12:12 h light/dark cycle with ad libitum access to food and water throughout the experimental period. The animals were acclimated in our animal facility for one week. Microglial TNF-deficient TNF fl/fl Cx3cr1-Cre ER mice were used in the current study. Floxed TNFα mice were procured from S. Nedospasov [67] and crossed with Cx3cr1tm2.1(cre/ERT2)Litt/WganJ (RRID:IMSR_JAX:021160)) mouse [68]. TNF fl/fl Cx3cr1-Cre ER mice lines were generated at Tohoku University for more than ten generations. After weaning on postnatal days (PNDs) 21-28, all mice were socially housed in same-sexed groups in a temperature-controlled environment with a 12:12 h light/dark cycle (lights on at 09:00 h) with ad libitum access to water and food. Genomic DNA extracted from the tails of mice was used for the standard PCR genotyping.

Acute Stress (Contextual Fear Conditioning Tests)
C57BL/6J mice were given saline or NAC (20 mM or 30 mM, pH 7 in saline; i.p.) 1 h before the first contextual fear conditioning training, at doses similar to those used in vitro experiments (33 or 49 mg/kg), but lower than the previous safety dose (204 mg/kg) [47]. Mice were then placed in the training chamber (17.5 × 17.5 × 15 cm), which had a stainlesssteel rod floor that was used to deliver footshocks (Ohara & Co., Ltd., Tokyo, Japan, year 2015). Each mouse was transferred from its home cage to the training chamber (from 10:00 a.m. to 12:00 a.m.) and allowed to explore it for 148 s before receiving a single 2-s footshock (0.4 mA) and consecutive 30-s exposures. Mice were re-exposed to the training chamber without receiving footshocks for 3 min (FS3) or 30 min (FS30) and then transferred to their home cages 24 h after conditioning. FS3 resulted in the retention of fear memory, whereas FS30 facilitated the extinction of fear memory. To validate the effect of NAC on the retention or extinction of fear memory, the percentage of time mice exhibited freezing behavior 24 h after a re-exposure session was measured [21,34]. The percentage of time the mice exhibited freezing behavior during a 5 min exposure (freezing time) was calculated as an indicator of the behavioral outcome of fear memory.

Chronic Social Defeat Stress (SDS)
The chronic SDS procedures were performed as previously described [69] [69], the C57BL/6 mice were treated with saline or NAC (32.64 mg/kg (20 mM) or 48.96 mg/kg (30 mM); pH 7.0; i.p.) per day and were exposed to a different CD1 aggressor mouse for 10 min per day for 10 consecutive days by removing the clear, perforated plexiglass divider. After the last exposure in each session, all C57BL/6 mice were housed individually. The stress-induced behaviors were tested from day 11 to day 12, including the sucrose preference test (days [11][12] and the forced swim test (day 13).

Sucrose Preference Test (SPT)
The SPT was performed as previously described [20]. It employed a two-bottle, freechoice sucrose consumption paradigm. For two days, the mice were habituated to drinking water from two tubes with stoppers fitted with ball-point sippers (Ancare, Bellmore, NY, USA). Following habituation, they were then exposed to 1% sucrose or drinking water for three consecutive days. The weights of the water-or sucrose-containing bottles were measured before and after this period. Sucrose preference was determined using the following equation: Sucrose pre f erence = (sucrose day 1 − sucrose day 2)/((sucrose day 1 − sucrose day 2) + (water day 1 − water day 2)) × 100

Forced Swim Test (FST)
The FST was performed as previously described [70]. Mice were individually placed in an inescapable transparent cylindrical tank filled with water (24 • C) for 6 min. The last 4 min of the test were examined. The behavioral activity was recorded using a video camera. Mobility and immobility times were automatically measured using ANY-Maze video-tracking software Version 6.10 (Stoelting Co., Wood Dale, IL, USA).

RNA Extraction and Quantitative Real-Time PCR
AllPrep ® DNA/RNA/Protein Mini Kit (QIAGEN, Ltd.-UK, Crawley, UK) was used to extract total RNA from MG6, mouse primary microglia, U-87 MG, SK-N-SH, and human primary M2 microglia. SuperScript™ VILO™ cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) was used to synthesize cDNA. The relative copy number of each transcript in each cDNA sample was determined using specific primers and iQ™ SYBR ® Green Supermix (Bio-Rad Inc., Hercules, CA, USA). A standard curve was created for each assay to adjust for differences in the amplification efficiency of the primer sets. 18S rRNA was used as an internal control for normalization. The forward and reverse primers for murine 18S were 5 -GTAACCCGTTGAACCCCATT-3 and 5 -CCATCCAATCGGTAGTAGCG-3 , respectively. The forward and reverse primers for human 18S were 5 -GAGGATGAGGTGGAACGTGT-3 and 5 -TCTTCAGTCGCTCCAGGTCT-3 , respectively. The forward and reverse primers for murine Tnf were 5 -AGCCCCCAGTCTGTATCCTT-3 and 5 -CTCCCTTTGCAG AACTCAGG-3 , respectively. The forward and reverse primers for murine Il1b were 5 -GCCCATCCTCTGTGACTCAT-3 and 5 -AGGCCACAGGTATTTTGTCG-3 , respectively. The forward and reverse primers for murine IL10 were 5 -CCAAGCCTTATCGGAAATGA-3 and 5 -TTTTCACAGGGGAGAAATCG-3 , respectively. The forward and reverse primers for human TNF were 5-TCCTTCAGACACCCTCAACC-3 and 5-AGGCCCCAGTTTGA ATTCTT-3, respectively.

Statistical Analysis
All assays were performed on three distinct occasions. Data are expressed as mean ± S.D. For parametric data, all comparisons were made using the Student's t test or oneway ANOVA, followed by a post hoc Turkey's test. Nonparametric data (ELISA of the primary microglia) were analyzed using the ANOVA test, followed by the Bonferroni test for post hoc analyses. SPSS software version 13.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analyses, and p < 0.05 was considered significant.

Informed Consent Statement: Not applicable.
Data Availability Statement: The datasets and materials generated and/or analyzed during the current study are available upon reasonable request from the corresponding author.