NADPH oxidase and reactive oxygen species contribute to alcohol-induced microglial activation and neurodegeneration

Eight-week male (20-22g) C57BL/6 mice were purchased from Jackson Laboratories (Bar Harbor, Maine). NF-κB enhanced GFP mice were gift from Dr. Christian Jobin's lab. All protocols in this study were approved by the Institutional Animal Care and Use Committee and were in accordance with the National Institute of Health regulations for the care and use of animals in research.

Human post-mortem brain tissue was obtained from the New South Wales Tissue Resource Center in Australia [ethics committee approval number: X11-0107]. Paraffin sections of orbitofrontal cortex (OFC) were used in this study. The detailed patients' medical history is presented in Table 1. Human alcoholic patients averaged with lifetime consumption of over 500 Kg of ethanol were compared to moderate drinkers who averaged less than 1 drink per day with lifetime consumption of less than 100 Kg of ethanol. Alcoholic neurodegeneration is associated with chronic high levels of alcohol consumption, whereas, there is no neurodegeneration associated with moderate drinkers [24‐26]. Only individuals with alcohol dependence not complicated by liver cirrhosis or nutritional deficiencies were included in this study. The main causes of death were cardiovascular disease for both groups. Complete life-style, medical histories and brain function as well as post-mortem interval (the time between death and the brain was removed from the body); causes of death and alcohol consumption are documented. Tobacco smoking history is also documented. Tobacco smoking is common in alcoholism and often confounds studies of alcoholism. Within our human subjects 6 of 8 controls and alcoholics had a history of smoking, although 3 controls were ex-smokers whereas all 6 smoking alcoholics continued to smoke. All psychiatric and alcohol use disorder diagnoses are confirmed using the Diagnostic Instrument for Brain Studies that is compliant with the Diagnostic Statistical Manual of Mental Disorders and has demonstrated reliability [27].

Cleaved caspase-3 (Asp 175) antibody was from Cell Signaling Technology (Danvers, MA). Fluoro-Jade B and mouse Neu-N antibody were from Chemicon international (Temecula, CA). Rabbit anti-Iba1 antibody was purchased from Wako Pure Chemical Industries, Ltd. (1-2Doshomachi 3-Chome Chuo-ku Osaka 540-8605, Japan). Monoclonal anti-mouse gp91^phox was from Transduction Laboratories (Lexington, KY). Rabbit polyclonal anti-gp91^phox IgG was purchased from Upstate cell signaling solutions (Temecula, CA). Goat polyclonal gp91^phox (C-15) antibody was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Rabbit polyclonal microtubule associated protein 2 (MAP2) antibody was purchased from Abcam (Cambridge, MA). Polyclonal Rabbit anti-Glial Fibrillary Acidic Protein was from DakoCytomation (Glostrup, Denmark). Hydroethidine was from Invitrogen Molecular Probes (Eugene, OR). All other reagents came from Sigma Chemical Co. (St. Louis, MO).

Sixty male C57BL/6 mice were randomly assigned to water control group (30 mice) and ethanol group (30 mice). The mice were treated intragastrically with water (control) or ethanol (5 g/kg, i.g., 25% ethanol w/v), with volumes matched, daily for 10 days. The average blood alcohol concentration at 1 hour after the first ethanol treatment and the last ethanol treatment was 302 mg/dl ± 12 (w/v, n = 10) and 297 mg/dl ± 11 (w/v, n = 10), respectively. The blood ethanol level is high and considered to model binge drinking [28]. Twenty mice from each group were sacrificed at 24 hr after the last dose of ethanol for mRNA and histochemistry. Ten mice from each group were sacrificed at 1 week after the last dose of ethanol for NOX gp91^phox immunostaining. For diphenyleneiodonium (DPI) treatment, forty male C57BL/6 mice were randomly assigned to 4 groups: control, EtOH, DPI and EtOH plus DPI (10 mice per group). The mice in EtOH and EtOH plus DPI groups were treated intragastrically with ethanol (5 g/kg, i.g., 25% ethanol w/v) daily for 10 days. The mice in control and DPI groups were gavaged with water daily for 10 days. DPI (3 mg/kg, i.p.) was given to mice at 0.5 hr and 24 hr after the last dose of ethanol. In both water and ethanol groups, mice were injected with saline, with volumes and time matched. Mice were sacrificed 3 hr after the last dose of DPI. For NF-κB transcription study, twenty male NF-κB enhanced GFP mice, a transgenic mouse expressing the enhanced GFP under the transcriptional control of NF-κB cis elements (cis-NF-κB^EGFP) [22, 23], were treated intragastrically with water (control, 10 mice) or ethanol (5 g/kg, i.g., 25% ethanol w/v, 10 mice), daily for 10 days. The mice were sacrificed 24 hr after the last dose of ethanol. For ROS analysis, male C57BL/6 or NF-κB enhanced GFP mice (10 mice per group) were treated intragastrically with water (control) or ethanol (5 g/kg, i.g., 25% ethanol w/v), daily for 10 days. Mice were injected with dehydroethidium (10 mg/kg, i.p.) in 0.5% carboxymethyl cellulose at 23.5 hr after the last dose of ethanol. Brains were harvested 30 min later and frozen sections (15 μm) were examined for hydroethidine oxidation product, ethidium accumulation, by fluorescence microscopy. All experiments were repeated 2 to 3 times.

Total RNA was extracted from the brain samples of C57BL/6 mice treated with ethanol or water, and reverse transcribed as described previously [29]. The primer sequences used in this study were as follows: NF-κB p65 (essential modulator), 5'-GGC GGC ACG TTT TAC TCT TT-3' (forward) and 5'-CCG TCT CCA GGA GGT TAA TGC-3' (reverse); β-actin, 5'- GTA TGA CTC CAC TCA CGG CAA A-3' (forward) and 5'-GGT CTC GCT CCT GGA AGA TG-3' (reverse). The SYBR green PCR master mix (Applied Biosystems, Foster City, CA) was used for real-time PCR analysis. The relative differences in expression between groups were expressed using cycle time (Ct) values normalized with β-actin, and relative differences between control and treatment groups were calculated and expressed as relative increases setting control as 100%.

Mouse brains were fixed with 4% paraformidehide in Phosphate Buffered Saline (PBS) and processed for immunostaining as described previously [29]. Human postmortem brains were processed to Paraffin sections for immunohistochemistry. Microglia were stained with rabbit anti-Iba1 antibody. Mouse NOX membrane subunit gp91^phox was immunostained with monoclonal anti-mouse gp91^phox or rabbit polyclonal anti-gp91^phox IgG. Human gp91^phox was immunostained with goat polyclonal gp91^phox antibody. Caspase-3 was immunostained with polyclonal anti-cleaved caspase-3 antibody. Neurons were stained with Neu-N or MAP2 antibody. Astrocytes were labeled with GFAP antibody. Immunolabeling was visualized by using nickel-enhanced 3,3'-diaminobenzidinne (DAB) or Alexa Fluor 488 (green) or 555 (red) or 633 (blue) dye.

In situ visualization of O₂ ^- and O₂ ^- - derived oxidant production was assessed by hydroethidine histochemistry [19, 30]. Mice were injected with dehydroethidium (10 mg/kg, i.p.) in 0.5% carboxymethyl cellulose at 23.5 hrs after the last dose of ethanol. Brains were harvested 30 min later and frozen sections (15 μm) were examined for hydroethidine oxidation product, ethidium accumulation, by fluorescence microscopy (excitation 510 nm; emission 580 nm).

Brain sections were immunostained with mouse Neu-N (a neuronal marker) antibody. Immunolabeling was visualized by using Alexa Fluor 555 dye. Sections were rinsed three times with PBS and one time with water before performing Fluoro-Jade B procedure. Sections stained with Neu-N were mounted on superfrost/plus microscope slides and air dried overnight. The sections were rinsed in distilled water for 2 min to rehydrate and transferred to a solution of 0.06% potassium permanganate for 10 min. The sections were then rinsed in distilled water for 2 min and placed in a 0.0004% Fluoro-Jade B solution made by adding 4 ml of a 0.01% stock solution of Fluoro-Jade B to 96 ml of 0.1% acetic acid. After 20 min in the Fluoro-Jade B staining solution, the stained slides were thoroughly washed in distilled water, dehydrated and coverslipped.

Immunoreactivity of mouse gp91^phox and fluorescent intensity of Fluoro-Jade B and ethidium were quantified using Bioquant Image Analysis Software (Nashville, TN). Images were captured on an Olympus BX51 microscope and Sony DCX-390 video camera at 40X. Light levels were normalized to preset levels and the microscope, camera, and software were background corrected to ensure reliability of image acquisition [31]. In each region (cortex and dentate gyrus), six random images from each brain sample were captured within a standard ROI (Region of Interest), the density of immunostaining and fluorescence was measured in pixels within this area (pixels/mm²). Subsequently, the average of the six measurements was used to represent the immunoreactivity or fluorescence intensity of each sample. When measuring fluorescence intensity in the cells, we eliminated the background by adjusting threshold to avoid background staining. For +IR cells counting, a modified stereological method was used to quantify cells within regions of interest following immunostaining of brain sections using the CAST stereological system [32, 33]. Specifically, cell density (N _v ) of caspase-3 and gp91^phox + immunoreactivity (+IR) was determined following the optical disector method [34, 35], which was calculated as follows:

Where ∑Q is the sum of the caspase-3 or gp91phox + IR cells counted from each disector frame, ∑disector is the sum of the number of disector frames counted, A(fr) is the known area associated with each disector frame, and h is the known distance between two disector planes (we used 10 μm).

For colabeling study, double or triple stained sections were digitally photographed with Leica SP2-AOBS confocal microscope and analyzed with Leica SP2 LCS software.

The data are expressed as mean ± SEM and statistical significance was assessed with an ANOVA followed by Bonferroni's t-test using the StatView program (Abacus Concepts, Berkeley, CA). A value of P < 0.05 was considered statistically significant.

To determine the effect of ethanol exposure on neurodegeneration in mice, immunohistochemistry for cleaved caspase-3 [36] and Fluoro-Jade B histochemistry were performed on C57BL/6 mouse brain sections treated with water or ethanol (5 g/kg, i.g.) daily for 10 days. Ethanol-treated mice showed an increase in activated caspase-3 immunoreactivity 24 hours after the last dose of ethanol treatment, compared to water controls (Figure 1A). The number of activated caspase-3+immunoreactive (IR) cells increased 3.1 fold in cortex and 3.5 fold in dentate gyrus in ethanol-treated mice (Figure 1B). To determine if caspase-3+immunoreactivity (+IR) was neuronal, double immunohistochemistry for cleaved caspase-3 and Neu-N, a neuronal marker was used. Confocal microscopy indicated that most activated caspase-3 +IR cells colocalize with Neu-N+IR cells (Figure 1C), suggesting chronic ethanol exposure causes neuronal cell death. Fluoro-Jade B, another cell death marker, was also used to assess ethanol-induced neurotoxicity [37]. Brain sections from control animals showed little or no Fluoro-Jade B staining. However, mouse brains exposed to chronic ethanol increased 10 fold in cortex and 7.6 fold in dentate gyrus in intensity of Fluoro-Jade B positive staining 24 hours after the last dose of ethanol treatment compared to water control group (Figure 2A). Confocal microscopy indicated that most Fluoro-Jade B positive cells were colocalized with Neu-N+IR (Figure 2B). Both cleaved caspase-3+IR and Fluoro-Jade B markers suggest chronic ethanol induce neuronal cell death in C57BL/6 adult mice. The orbitofrontal cortex (OFC) of human post-mortem brain was assessed using Fluoro-Jade B. The OFC of human moderate drinking control brain showed few Fluoro-Jade B cells whereas the OFC of alcoholic brain showed more labeled cells (Figure 3). Confocal microscopy found that Fluoro-Jade B positive cells in human brain were mostly colocalized with Neu-N, suggesting increased neuronal cell death in human post-mortem alcoholic brain.

Previous studies have linked activation of microglia, production of proinflammatory factors and reactive oxygen species (ROS) to neurodegeneration [16, 29, 38]. Our previous research found that 10 daily doses of ethanol significantly increased levels of brain proinflammatory genes (TNFα and MCP-1, etc.) [9]. To investigate proinflammatory responses in this experiment sections were immunostained with Iba1 microglial antibody. In the water control group, microglia have a resting morphology. Ethanol treated mouse brains showed activated microglia morphology in multiple brain regions, including cortex and dentate gyrus of hippocampus (Figure 4A) 24 h after the last dose of ethanol. Microglia activation following ethanol treatment is indicated by increased cell size, irregular shape, intensified Iba1 staining, and an altered ameboid morphology. Thus, Iba1+ IR morphological assessment indicate ethanol causes microglial activation.

Astrocyte activation was assessed by morphology using GFAP, an astrocyte-specific intermediate filament protein [39, 40]. Chronic ethanol treatment increased GFAP + IR in cortex and dentate gyrus (Figure 4B) 24 h after the last dose of ethanol. In addition to these two brain regions, astroglial activation was also notably observed in other brain areas, such as substantia nigra and forceps minor corpus callosum in the ethanol treated-mice (data not shown). Thus, the data together with increased cell death markers (caspase-3 and Fluoro-Jade B) by chronic ethanol treatment suggest that astroglial activation mediate ethanol-induced neurodegeneration.

Previous studies have suggested ethanol activates nuclear factor κB (NF-κB) transcription inducing expression of proinflammatory genes [14, 41]. To investigate effect of ethanol on NF-κB mRNA and protein expression, C57BL/6 and NF-κB-GFP reporter mice were treated intragastrically with water (control) or ethanol (5 g/kg, i.g., 25% ethanol w/v) daily for 10 days as before and sacrificed 24 hrs after the last dose of ethanol. Chronic ethanol significantly increased NF-κB-p65 gene expression in C57BL/6 mouse brain (Figure 5A). In NF-κB GFP reporter mice ethanol treatment markedly increased GFP Fluorescence in multiple brain regions, such as dentate gyrus (Figure 5B).

Cell phenotype for NF-κB activation and ROS production was examined using histochemical markers. NF-κB enhanced GFP reporter mice showed green fluorescence. ROS were detected by red hydroethidine histochemistry and cell type markers, e.g. Iba1 microglial antibody, MAP2 neuronal antibody and GFAP astroglial antibody. Confocal microscopy found that NF-κB and ROS were triple-labeled with Iba1 or MAP2, but not with GFAP (Figure 5C), indicating ethanol-induced NF-κB activation and ROS production mostly occurred in microglia and neurons.

NADPH oxidase (NOX) is the enzyme complex responsible for the respiratory burst in phagocytes. Activation of this enzyme in microglia causes the production of ROS leading to dopaminergic neurotoxicity [29]. To determine whether NOX is involved in chronic ethanol induced neurotoxicity we investigated the expression of NOX gp91^phox, the catalytic subunit of phagocytic oxidase commonly associated with proinflammatory responses. Cortex and dentate gyrus of ethanol treated mouse brain had significantly more gp91^phox +IR cells, compared to those of water control brain 24 hrs after ethanol treatment (Figure 6A). Quantification of gp91^phox +IR indicated that ethanol induced gp91^phox +IR 4.3 fold in cortex and 3.4 fold in DG (Figure 6B). Chronic ethanol-induced increases in gp91^phox expression remained elevated 1 week after ethanol treatment (Figure 6C). Thus, ethanol treatment of mice increases NOX gp91^phox expression that persists long after cessation of drinking.

Human postmortem alcoholic brain histochemistry showed significant increases in the number of gp91^phox + IR cells, compared to human moderate drinking control brain (Figure 7A, B). Double antibody studies with cell specific markers and confocal microscopy reveal that gp91^phox +IR cells in human postmortem alcoholic brain are colocalized with a neuronal marker (MAP-2), a microglial marker (Iba-1), and an astroglial marker (GFAP) (Figure 7C). These results indicate that alcohol increases NOX gp91^phox expression in both ethanol-treated mouse brain and human alcoholics consuming large amounts of alcohol.

To investigate whether induced NOX produces superoxide in brain, we used in situ visualization of reactive oxygen species, e.g. O₂ ^- and O₂ ^--derived oxidant production of ethidium from hydroethidine in vivo [19, 30]. In vehicle-treated mice, there was little to no detection of O₂ ^- and O₂ ^--derived oxidant production of ethidium. In ethanol-treated mice, a significant production of O₂ ^- and O₂ ^--derived oxidants was observed 24 hrs after 10 daily doses of ethanol exposure (Figure 8A). Quantification of the ethidium fluorescence indicates that ethanol increased O₂ ^- and O₂ ^--derived oxidants more than 7 fold in cortex and DG, compared to controls (Figure 8B). These findings indicate that ethanol increases expression of NOX gp91^phox and increases the formation of reactive oxygen species. Human alcoholics show a similar increase in NOX+IR consistent with chronic alcohol abuse in humans increasing proinflammatory oxidative stress in brain.

To determine if ROS are found in cells expressing NOX gp91^phox, we performed triple labeling using gp91^phox+IR together with antibodies against the Iba1, a marker of microglia or Neu-N, a marker of neurons, or GFAP, a marker of astrocytes on brain sections from the mice injected with dehydroethidium, an agent detecting ROS. Confocal microscopy indicated that gp91^phox (blue) and O₂ ^- (red) labeling was not detectable in water control mouse brains (picture not shown). However, intense fluorescence of gp91^phox and O₂ ^- was observed in ethanol treated mouse brains (Figure 9, first and second panels) 24 h after the last dose of ethanol. Triple-labeled cells are white due to gp91 (blue), O₂ (red) combining with microglial (Iba1, green) or neuronal (Neu-N, green) marker proteins (Figure 9), but not with astrocyte GFAP (pink). These results indicate that chronic ethanol-induced activation of NOX and production of O₂ ^- predominantly occurs in microglia and neurons in mouse brain.

In an effort to discern the role of ROS in ethanol-induced neurotoxicity, we used a NOX inhibitor, Diphenyliodonium (DPI) [42, 43]. As the resident innate immune cells in the brain, microglia are a predominant source of pro-inflammatory factors [TNFα, MCP-1, IFN-γ and oxidative stress (NO, H₂O₂, O₂ ^- and OHOO^-/ONOOH)], which are toxic to neurons [44]. To examine whether ethanol-induced microlgial activation is associated with ROS and the consequent neurotoxicity, C57BL/6 mice were injected with DPI. As shown in Figure 10, after 10 daily doses of ethanol treatment, microglia appear activated: increased cell size, irregular shape, and intensified Iba1 staining. Treatment with DPI significantly reduced microglial activation by ethanol exposure. In DPI and ethanol co-treatment group, microglia showed the resting morphology similar to those in the water control group. These results suggest an important role of NOX in ethanol-induced microglia activation. To analyze the effects of DPI treatment on ROS, hydroethidine histochemistry was performed. Exposure of mice to ethanol for 10 days was found to significantly increase ROS generation (Figure 11A, B), again providing confirmation that ethanol can elicit ROS generation in adult C57BL/6 mouse brain. DPI treatment caused 51% decrease in fluorescence intensity of O₂ ^- and O₂ ^--derived oxidants, compared with ethanol-treated group (Figure 11B), suggesting that NOX-generated ROS contribute to chronic ethanol-induced microglial activation.

Chronic ethanol increased the number of activated caspase-3+IR cells by 104% compared to water control group. DPI alone did not show any effect on caspase-3 immunolabeling compared to water controls. Co-treatment with DPI and ethanol reduced ethanol-induced increase in caspase-3+IR cells (Figure 12A). Also, DPI significantly reduced ethanol increased Fluoro-Jade B fluorescence intensity by about half (Figure 12B). DPI is a NOX inhibitor. DPI treatment reduced ROS and cell death markers (caspase-3 and Fluoro-Jade B), so the data suggest that NOX-generated ROS contribute to chronic ethanol-induced neurotoxicity.