Six weeks-old male wt, adiponectin knock-out (ApN^−/−) or adiponectin-receptor 2 knock-out (AdipoR2^−/−) mice with the same C57BL/6J genetic background were housed at 22°C with a 12-h light-dark cycle (lights on at 07:00) with free access to drink and chow (A04, SAFE). All experiments were conducted in compliance with the Institutional Animal Care and Use Committee of the University of Nice-Sophia Antipolis (permission number 010344.01 from the French “Ministère de l'Enseignement Supérieur et de la Recherche”).

Lipopolysaccharide (LPS) from Escherichia coli 0111:B4 was purchased from Sigma-Aldrich and freshly dissolved in sterile saline prior to i.p. injection at a dose of 1 mg/kg.

Wt mice were anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (10 mg/kg). The mouse was placed on a stereotaxic apparatus (Kopf) on heating mat and the skull was exposed. I.c.v injections (0.3 μg of ApN in 2 μl of NaCl 0.9% or vehicle alone) were performed using a Hamilton micro-syringe (cerebral lateral ventricle coordinates: −0.22 mm posterior and 1 mm lateral to the bregma; depth 2.0 mm) at 0.2 μl/min.

Mice were anesthetized with an i.p. injection of sodium pentobarbital; then CSF (2–5 μl) was obtained by cisternal puncture. Multiplex measurement of proinflammatory cytokines in mouse CSF was performed with the MSD 96-well multi-array mouse cytokine assay (Mesoscale). For comparison, data were normalized relative to the CSF volume.

Mice were deeply anesthetized with a lethal injection of pentobarbital. Immune brain cells were isolated from whole-brain homogenates using a protocol adapted from Cardona et al. (2006), as previously described in Cazareth et al. (2014). Mice were transcardially perfused with ice-cold HBSS containing 1 mg/ml EDTA. Brains were collected and roughly homogenized in PBS, resuspended in PBS containing 3 mg/ml collagenase D (Roche Diagnostics) and incubated 20 min at 37°C. After incubation, brain homogenates were filtered on 70 μm pore size cell strainers (BD Biosciences), centrifuged (10 min, 2,000 rpm), washed and resuspended in 6 ml of 38% isotonic Percoll (GE Healthcare), before centrifugation (20 min, 2,000 rpm, 4°C). The surface ring containing myelin and debris was discarded. Cell pellets containing brain immune cells were collected, washed with PBS containing 0.5% BSA and 2.5 mM EDTA and labeled for subsequent cell sorting and/or flow cytometry analysis.

Staining of brain immune cell surface antigens was performed as previously described (Cazareth et al., 2014). Briefly, Fc receptors were blocked with 2.4G2 antibody. Microglia was identified according to the labeling of CD11b-PercP-Cy5.5 and CD45-APC-Cy7 conjugated antibodies (BD Biosciences). Inflammatory monocytes were defined as CD11b⁺CD45^hiLy-6G^negLy-6C^hi cells (Ly-6G and Ly-6C antibodies were purchased from BD Biosciences). Cells were washed and resuspended in PBS containing 0.5% BSA and 2.5 mM EDTA for analysis and cell sorting with FACS Aria III (BD Biosciences).

Brain-sorted microglia was seeded at a density of 2 × 10⁴ cells/well in 96-well tissue culture plates (Falcon) in Dulbecco's Modified Eagle's Medium (DMEM) culture media (Gibco) containing 10% fetal bovine serum (FBS, Gibco). Cells were cultured at 37°C with 5% CO₂ and saturated humidity for 24 h. Fifteen minutes after cell plating, gApN (10 μg/ml, Phoenix Pharmaceuticals) was added, 1 h prior stimulation with LPS (0.5 μg/ml). LPS from Escherichia coli 0111:B4 was purchased from Sigma-Aldrich and freshly dissolved in sterile saline prior to cell treatment. Media were then collected and centrifuged at 1,200 rpm for 10 min at 4°C.

Mice were deeply anesthetized with pentobarbital and transcardially perfused with PBS and then 3.2% paraformaldehyde. Brains were postfixed in 3.2% paraformaldehyde for 48 h. Fixed brains were sectioned (40 μm) using a vibratome (Leica). Sections were blocked with 3% goat serum. Sections were washed and incubated with a rabbit anti-mouse ionized calcium binding adaptor molecule 1 (Iba-1) antibody (Wako Chemicals, USA), overnight at 4°C, followed by incubation with the appropriate fluorescent-conjugated secondary antibody (Invitrogen). Slices were then mounted on glass slices with Fluoroprep (Dako) containing 1 mg/ml Hoescht, a fluorescent specific DNA dye (Invitrogen). Each fluorochrome was independently captured with an FV10i scanning confocal microscope (Olympus, France). Quantification of Iba-1 staining was performed using ImageJ software (NIH). The software generated fluorescence intensity values by tracing the region-of-interest (ROI). Arbitrary units were defined in terms of strength of fluorescent signal. The final intensity values were calculated by subtracting the area of the selected region multiplied by the background fluorescence from the fluorescence intensity of the region-of-interest (ROI): fluorescence intensity (arbitrary units) = fluorescence intensity of ROI − (area of selected region × mean fluorescence of background).

Primary microglial cell cultures were prepared from the brains of newborn C56BL/6J mice (postnatal days 1–2; Marella and Chabry, 2004). The brains were removed and placed in DMEM, then the meninges and blood vessels were removed under a microscope. After mechanical dissociation, aliquots of the cell suspension were plated in a poly-D-lysine coated flask (Falcon). Cultures containing mainly astrocytes (90%) and microglial cells (10%) were maintained at 37°C in a humidified 5% CO2 atmosphere for 10 days, with the change of the medium every 3–4 days. The cultures were then shaken on an orbital shaker at 150 rpm for 60 min at 37°C. Floating cells in the supernatant were collected, centrifuged at 250 g for 5 min, and resuspended in the culture medium. This fraction contained the microglial cells (around 93% of purity). The adherent cell population (i.e., astrocytes) was grown in DMEM containing 10% FCS, split every week and used after two passages. Astrocytes cultures contained around 99% of GFAP+ positive cells.

BV2 mouse microglia cells were grown and maintained in DMEM (4.5 g/l glucose, 4 mM L-glutamine, without sodium pyruvate from Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco), 100 U/ml penicillin and 100 μg/ml streptomycin (Gibco). Cells were grown in an atmosphere of 5% CO2 at a temperature of 37°C. AdipoRon (2-(4-benzoylphenoxy)-N-[1-(phenylmethyl)-4 piperidinyl]acetamide) was purchased from Sigma-Aldrich and dissolved in 2.5% DMSO.

AdipoR1 or AdipoR2 interfering RNA (siRNA) or scrambled siRNA (Qiagen) were transfected into BV2 cells using RNAiMax Lipofectamine (Invitrogen) according to the manufacturer's recommendations. After various experimental tests, 48 h post-transfection time was selected as the optimal duration to allow maximal decrease of targeted protein expression. Therefore, in all experiments, cells were used 48 h after siRNA transfection.

BV2 cells were grown on poly-D-lysine-coated glass coverslips. After treatment, cells were fixed with 3.2% paraformaldehyde, then permeabilized in PBS pH 7.4, containing 0.3% Tween and 3% BSA. Cells were then incubated overnight with the appropriate primary antibody directed against CD11b (Abcam), iNOS (Merck Millipore), AdipoR1 or AdipoR2 (Sigma Aldrich) in the same buffer. At the end of the incubation time, cells were washed three times with PBS and incubated with either Alexa488-conjugated or Alexa594-conjugated secondary antibodies (Invitrogen) and Hoechst fluorescent stain (Invitrogen) for 1 h at room temperature. For NF-κB p65 subunit nuclear translocation study, cells were pre-treated with 10 μg/ml gApN or saline for 1 h before addition of 0.5 μg/ml LPS for 3 additional hours. Cells were then fixed and permeabilized as previously. Primary antibody against p65 was from Cell Signaling. Images were captured with an FV10i scanning confocal microscope (Olympus, France) and analyzed using Image J software (“National Institutes of Health” Image J Software™).

Total RNA from hippocampus, hypothalamus, amydgala, prefrontal cortex, microglia or astrocytes were isolated using the Trizol® RNA extraction kit (Invitrogen) according to the manufacturer recommendations followed by a RQ1 DNAse (Promega) treatment. First-strand cDNA were synthesized from 2 μg of total RNA with 200U of SuperScript III reverse transcriptase (SuperScriptIII, Invitrogen) in the appropriate buffer in the presence of 25 μg/ml random primers, 0.5 mM desoxyribonucleotide triphosphate mix, 5 mM dithiothreitol, 40U RNAsin (Promega). The reaction was incubated 5 min at 25°C, then 50 min at 50°C then inactivated 15 min at 70°C. Quantitative PCR was performed using the SYBRgreen method (Roche) with the LightCycler 480 sequence detector system (Roche Diagnostics). β-actin and GAPDH were used as reference genes for normalization. Primers were purchased from QIAGEN (QuantiTect primer assay, QIAGEN).

Supernatants from ex vivo microglia or BV2 cells were harvested and the concentration of secreted cytokines (TNFα, IL-1β, and IL-6) was detected using a cytometric bead array according to the manufacturer's instructions (BD Biosciences) or with the MSD 96-well multi-array mouse cytokine assay (Mesoscale), as specified for each set of experiments. For comparison, data were normalized relative to the cell number.

BV2 cells were seeded in 24-well plates as described above and then pre-treated with 10 μg/ml gApN for 1 h, then LPS (0.5 μg/ml) was added. Fifteen hours later, 2.7-dichlorofluorescin diacetate (DCFH-DA) (10 μM; Sigma-Aldrich) was added and incubated at 37°C, 5% CO₂. DCFH-DA is a redox-sensitive dye which is readily taken up by the cells. Within the cell, esterases cleave the acetate groups on DCFH-DA, thus trapping the reduced form of the probe (DCFH), intracellularly. ROS in the cells oxidize DCFH, yielding the fluorescent product DCF. After 60 min, culture media were removed and kept at 4°C for subsequent NO production assay. Cells were rinsed with PBS and harvested in Trypsin/EDTA (Gibco). Cells were pelleted and resuspended in PBS containing 0.5% BSA and 2.5 mM EDTA for FACS analysis (Fortessa BD Biosciences).

NO2⁻ was measured in BV2 culture supernatants to assess the NO production in microglial cells. Fifty microliters of the culture media were mixed with 50 μl of Griess' reagent for nitrite from Promega in a 96-well plate and incubated at 25°C for 10 min. The absorbance at 570 nm was measured by a microplate reader. Sodium nitrite (NaNO2) was used as the standard to calculate the NO2⁻ concentrations.

For total protein extraction, BV2 cells were rinsed with PBS and homogenized in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA containing 0.5% Triton X-100 and 0.5% sodium deoxycholate). Equal amounts of total proteins (40 μg) determined by the Bradford method (BioRad) were separated onto 12% SDS-PAGE, then transferred on nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany). Blots were incubated in PBS 0.1% Tween, 1% milk with antibodies directed against phospho-AMPKα (Thr172), total AMPK (Cell Signaling), iNOS/NOSII (Merck Millipore), or COX-2 (Chemicon), and then with the appropriate HRP-secondary antibody. Blots were developed using an enhanced chemoluminescence system (Immobilon, Merck Millipore) with a Fusion detector (Vilber). To correct for any loading artifact, same blots were re-probed with anti-actin antibody (Abcam). Densitometry analyses were performed with a “National Institutes of Health” Image J Software™ on the immuno-positive bands. Results were expressed as a percentage of control load. For NF-κB translocation study, cells were pre-treated for 1 h with 10 μg/ml gApN or saline before addition of 0.5 μg/ml LPS for 3 additional hours. Cells were then rinsed with PBS and detached with trypsin/EDTA (Gibco). After centrifugation, cells were homogenized in lysis buffer [10 mM Hepes, pH 7.4, 10 mM KCl, 0.1 mM EDTA, 0.5% NP-40, 1 mM dithiothreitol (DTT), and protease inhibitor cocktail from Roche]. Cells were incubated 20 min on ice and then centrifuged for 10 min at 12,000 g. The supernatants contained cytosolic proteins. For nuclear protein extraction, nuclear pellets were homogenized in extraction buffer (20 mM Hepes pH 7.4, 400 mM NaCl, 1 mM EDTA, 1 mM DTT, and protease inhibitor cocktail). After 30 min on ice, samples were centrifuged 15 min at 12,000 g, 4°C. Supernatants contained nuclear extracts. Equal amounts of protein of nuclear extracts were loaded on 10% SDS-PAGE then transferred to nitrocellulose as previously described. Blots were probed with antibody against NF-κB p65 subunit from Cell Signaling. Histone H3 immunoreactivity (Cell Signaling) was used as a nuclear loading control.

Statistical analysis was performed using the XLSTAT software. Significant differences between two groups of data were determined using a Mann & Whitney test for non-parametric data. Alternatively, Kruskal–Wallis analysis followed by a Conover-Iman test for multiple comparisons with Bonferroni correction between two independent groups, or a two-way ANOVA was applied. Results from data analysis were expressed as mean ± standard error of the mean (SEM). Statistical significance was set at ^*P < 0.05 and ^**P < 0.01, ^***P < 0.001.

In order to explore how ApN influences brain inflammation, we first evaluated the impact of ApN deficiency on mouse brain susceptibility to neuroinflammation. We analyzed the effect of a single LPS i.p. administration on pro-inflammatory cytokine mRNA expression in different brain regions of wild-type (Wt) and ApN^−/− mice. Brain regions were selected according to their implication in emotional stress information (amydgala, hippocampus, prefrontal cortex) or integration (hypothalamus). While IL-6 mRNA levels were not modified 24 h after LPS treatment in Wt mice regardless the brain area, IL-1β and TNFα mRNA levels were elevated in all the analyzed brain regions of Wt and ApN^−/− mice, although to different extents depending on the area (Figure 1A). After the LPS challenge, IL-1β, IL-6, and TNFα mRNA levels were statistically higher in all brain regions from ApN^−/− mice than in those from Wt mice, except from IL-6 mRNAs in prefrontal cortex (Figure 1A). We also investigated the levels of pro-inflammatory cytokines IL-1β, IL-6, and TNFα in the CSF of Wt and ApN^−/− mice, 24 h after saline or LPS ip administration (Figure 1B). As expected, pro-inflammatory cytokines levels were extremely low in NaCl-injected Wt and ApN^−/− mouse CSF samples. LPS induced a strong increase of IL-1β, IL-6, and TNFα levels in the CSF of Wt and ApN^−/− mice that was statistically higher in ApN^−/− mice as compared to Wt mice. Anti-inflammatory cytokine IL-10 was also detectable in the CSF of all mouse groups, but its levels were similarly increased after LPS challenge in Wt and ApN^−/− CSF samples (data not shown).

Microglia is defined as being CD11b⁺/CD45^low+ while the heterogeneous population of CNS-associated phagocytes is CD11b⁺/CD45^high+. The gating strategy for their correct identification is presented in Figure 1C. Three different populations were identified in the brain immune cell suspensions: CD11b⁺/CD45^high+ (CNS-associated phagocytes), CD11b⁺/CD45^low+ (microglia), and CD11b⁻/CD45 ^high+ cells (Figure 1C).

CNS infiltration of CD11b⁺CD45^high+Ly-6C^hi monocytes is a hallmark of CNS inflammation, including that caused by a peripheral injection of LPS. Flow cytometry analysis revealed that 24 h post LPS injection, the number of CD11b⁺CD45^high+Ly-6C^hi inflammatory monocytes was statistically higher in ApN^−/− mice as compared to Wt mice. The number of microglia collected from whole brains was similar in both mouse groups (Figure 1D).

Altogether, these results indicated that ApN deficiency triggers a stronger global neuroinflammatory response after a peripheral LPS challenge as compared to Wt mice, suggesting anti-inflammatory properties of ApN within the CNS.

Microglia, the immune-like cells of the brain, plays an important role in inflammatory responses in the CNS. Previously, we have shown that EE mediates anti-inflammatory effects in the brain by targeting microglia activation profile in an ApN-dependent manner (Chabry et al., 2015). Thus, we investigated whether the greater susceptibility of ApN^−/− mice toward neuroinflammation may be due to a different responsiveness of microglia caused by ApN deficiency. To answer this question, we subjected Wt and ApN^−/− mice to a single LPS i.p. injection (1 mg/kg) and analyzed microglia reactivity by immunohistochemistry, protein analysis and qRT-PCR. Figure 2A shows Iba-1 labeling of microglia in hippocampus and hypothalamus of Wt and ApN^−/− mouse brains, 24 h after LPS ip administration. Quantification of images shows that in both brain regions, Iba-1 fluorescence appeared more intense in ApN^−/− challenged mice than in Wt mice (Figure 2B).

Microglia was sorted out of LPS-challenged brains from Wt and ApN^−/− mice (Figure 2C). Protein analysis of the secretion media of brain-sorted microglia revealed that LPS-stimulated microglia sorted from ApN^−/− mice produced statistically more IL-1β, IL-6, and TNF-α pro-inflammatory cytokines than that from Wt mice (Figure 2D). Similarly, qRT-PCR analysis of pro-inflammatory cytokine mRNA expression in Wt and ApN^−/− microglia after an in vivo LPS challenge revealed higher levels of IL-1β, IL-6, and TNFα mRNAs in ApN^−/− microglia as compared to Wt microglia (Figure 2E). Altogether, these results indicated that ApN deficiency leads to a greater sensitivity of microglia toward an inflammatory challenge triggered by LPS without significant microglia proliferation.

We next conducted a series of experiments aimed at evaluating the effects of ApN on microglia inflammatory profile. We first examined the in vivo effects of ApN on microglia (Figures 3A–C). CBA measurement of cytokines in the secretion media of brain sorted microglia revealed that IL-1β, IL-6, and TNFα levels were significantly lower in the secretion media from microglia sorted out from gApN injected mouse brains as compared to microglia from saline-injected mouse brains (Figure 3B). In agreement with this, qRT-PCR analysis showed that IL-1β, IL-6, and TNFα mRNA levels were significantly lower in microglia sorted out from gApN injected mouse brains as compared to microglia from saline injected mouse brains (Figure 3C). These results suggested that i.c.v injection of gApN significantly, directly or indirectly, prevents the LPS-induced production of pro-inflammatory cytokines by microglia.

To deepen these results, we wondered if gApN could have direct anti-inflammatory effects on microglia. We first tested the ability of gApN to modulate the cytokine production by in vivo activated microglia (Figure 3D). Analysis of pro-inflammatory cytokines in the secretion media revealed that microglia sorted out from brains of LPS-treated mice released more IL-1β, IL-6, and TNFα than microglia from saline-treated mice (Figure 3D). In vitro treatment of microglia with 1 μg/ml gApN significantly reduced IL-1β and TNFα production, but failed to change IL-6 production (Figure 3D). On the other hand, a higher dose of gApN (10 μg/ml) promoted a greater inhibition of IL-1β and TNFα and also significantly reduced IL-6 release by LPS-activated microglia (Figure 3D). Next, we tested the ability of gApN pre-treatment to block the cytokine production by microglia that has been activated in vitro by LPS (Figure 3E). As expected, in vitro LPS treatment promoted a strong increase of IL-1β, IL-6, and TNFα mRNAs in microglia that was almost totally abolished by gApN pre-treatment (Figure 3E). Together, these results indicated that gApN has direct anti-inflammatory actions on microglia. Moreover, gApN treatment nonetheless significantly reduced the production of pro-inflammatory cytokines by activated microglia but also efficiently prevented microglia activation when used prophylactically in pre-treatment experiments.

Although, our results indicated that gApN exerted direct anti-inflammatory actions on microglia, we wondered whether gApN could also affect cytokine production by astrocytes, as previously mentioned (Wan et al., 2014). To answer this question, we compared the effects of gApN treatment on primary cultures of microglia, astrocytes, or mixed glial cells obtained from postnatal mouse brains. Microglia or astrocytes were treated with gApN (10 μg/ml) alone, LPS (0.5 μg/ml) alone, or pre-treated with gApN for 1 h before addition of LPS for 15 additional hours (Figure 4). Cells were collected and pro-inflammatory cytokine mRNA levels were analyzed by qRT-PCR. GApN treatment significantly reduced basal levels of IL-1β and IL-6 by microglia but did not modify TNFα mRNA level of unchallenged microglia. Moreover, pre-treatment of microglia with gApN also significantly blocked the increase of IL-1β, IL-6 and TNFα mRNAs elicited by LPS treatment (Figure 4A). In contrast, gApN did not reduce the basal or LPS-induced levels of pro-inflammatory cytokine mRNAs by astrocytes (Figure 4B). Interestingly, gApN treatment on astrocytes rather produced pro-inflammatory effects, since it significantly increased LPS-induced production of IL-1β and TNFα mRNAs (Figure 4B). We also assessed in both cell types the level of expression of IκB-α mRNA. IκB-α is an immediate-early gene induced by LPS challenge and involved in the NF-κB signaling pathway. Here, we confirmed that LPS increased the expression of IκB-α mRNA in both microglia and astrocytes (Figures 4A,B). GApN by itself had no effect on basal expression level of IκB-α mRNA in microglia while it significantly increased it in astrocytes. When used as a pre-treatment, gApN significantly blocked LPS-induced IκB-α mRNA increase in microglia but failed to change it in astrocytes (Figures 4A,B). We also examined gApN effects on primary cultured mixed glial cells containing microglia and astrocytes. Our results showed that gApN (10 μg/ml) successfully limited IL-1β, IL-6, and TNFα mRNA production by mixed glial cells stimulated by LPS (0.5 μg/ml; Figure 4C), suggesting that microglia/astrocytes crosstalk may favor global anti-inflammatory response to gApN. Altogether, these results suggested that gApN exerts different or even opposite effects on microglia and astrocytes but that in a cellular environment containing both microglia and astrocytes (brain, see Figure 1A, or mixed glial cell cultures), gApN treatment elicits rather overall anti-inflammatory effects.

In order to characterize in more detail the anti-inflammatory properties of ApN on microglia, we chose to conduct the following experiments on a more convenient cell model, i.e., the BV2 cell line. Cells were pre-treated for 1 h with saline or gApN (0.5, 2.5, or 10 μg/ml for qRT-PCR experiments, Figure 4D; 1, 10, and 50 μg/ml for CBA assays, Figure 4E) before subsequent LPS addition (0.5 μg/ml), for 15 h. BV2 cells were then collected and cytokine mRNA levels and cytokine secretion were analyzed by qRT-PCR (Figure 4D) and CBA (Figure 4E), respectively. On this cell model, gApN alone had no effect on basal expression levels of any examined mRNA or secreted cytokine. However, we confirmed that a pre-treatment with 10 μg/ml of gApN significantly limited the LPS-induced increase of IL-1β, IL-6, TNFα mRNA expression and protein release (Figures 4D,E). GApN also limited the LPS-induced increase of IκB-α mRNAs (Figure 4D). Our results indicated that gApN dose-dependently decreased the expression of pro-inflammatory markers characteristic of the M1 phenotype (IL-1β, IL-6, TNFα, and IκB-α) of microglia. Therefore, we wondered if gApN could simultaneously promote the expression of anti-inflammatory markers characteristics of the M2 phenotype. We then analyzed Arginase 1 and IL-10 mRNA levels. Our results indicated that neither LPS nor gApN alone or in combination with LPS altered the expression of these markers in the BV2 cell line (Figure 4D). These findings suggested that gApN may block M1 phenotype of microglia rather than promote a M2 phenotype.

When activated, microglia produces and releases not only inflammatory cytokines but also reactive oxygen species (ROS) and nitric oxide (NO). To assess whether ApN could regulate ROS and NO induction, BV2 cells were stimulated with LPS only or subsequently to a 1 h pre-treatment with gApN. GApN by itself significantly inhibited the basal level of ROS production in BV2 cells and had no effect on NO basal release. As expected, LPS treatment induced great increases of both ROS and NO production. Both were significantly blocked by gApN pre-treatment (Figures 5A,B). In microglial cells, NO is synthesized from iNOS (Minghetti and Levi, 1998). We thus investigated by immunocytochemistry iNOS immunoreactivity in BV2 cells upon gApN treatment only, LPS only or gApN pre-treatment followed by LPS addition. CD11b labeling was also examined as a control of microglial activation. Images quantification indicated that gApN alone had no effect but reduced both LPS-induced intracellular iNOS and CD11b surface immunoreactivities (Figures 5C,D).

Our next goal was to define downstream pathways that may explain ApN anti-inflammatory, anti-nitrosative, and anti-oxidative actions on microglia cells. In periphery, gApN was suggested to directly regulate glucose metabolism and insulin sensitivity through its action on AMPK (Berg et al., 2001; Tomas et al., 2002; Yamauchi et al., 2002). Moreover, it has been recognized that AMPK signal regulates the inflammatory responses induced by NF-κB (Salminen et al., 2011). However, nothing is known yet on ApN-mediated signaling pathways in microglia. To fill this gap, we analyzed by Western-blotting AMPKα phosphorylation on Thr172 in BV2 cells upon gApN treatment, with or without LPS challenge (Figure 6A). Phosphorylation of AMPKα was significantly enhanced by gApN in basal and LPS-stimulated conditions. In endothelial cells, HMW ApN has been shown to activate AMPK and eNOS, leading to NF-κB dependent signaling pathway regulation (Hattori et al., 2008). We thus analyzed NF-κB activation by studying the nuclear translocation of the p65 subunit in BV2 cells. As expected, nuclear p65 immunoreactivity was increased by LPS treatment. GApN pre-treatment before LPS administration partially blocked the nuclear translocation of p65 (Figure 6B). These results collectively suggest that gApN regulates inflammatory responses by modulating the activity of NF-κB via an AMPK signaling pathway.

The next step was to study which ApN receptor was responsible for the anti-inflammatory actions of ApN on microglia. Notably, we first noticed by qRT-PCR on microglia sorted out from Wt mouse brains that microglia expressed both AdipoR1 and AdipoR2 mRNAs, at equivalent levels (data not shown). We first subjected Wt, ApN^−/− and AdipoR2^−/− mice to a single i.p. injection of LPS (1 mg/kg). Twenty-four hours later, microglia was sorted out from mouse brains, plated and treated in vitro with saline or gApN (10 μg/ml), for 15 additional hours (Figure 7A). Cytokine level analysis of the secretion media revealed, as previously showed, that LPS-activated microglia from ApN^−/− mice produced more IL-1β, IL-6, and TNFα cytokines than that from Wt mice, although differences were not statistically significant for TNFα (Figure 7B). Interestingly, it appeared that LPS-activated microglia from AdipoR2^−/− mice also produced significantly more pro-inflammatory cytokines IL-1β and IL-6 than that from Wt mice, suggesting that AdipoR2 deficiency led to a higher susceptibility of microglia toward LPS challenge. However, our findings also revealed that in vitro gApN treatment efficiently inhibited IL-1β, IL-6, and TNFα production by Wt, ApN^−/−, and AdipoR2^−/− microglia, showing that AdipoR2 deficiency did not affect gApN anti-inflammatory actions on microglia (Figure 7B).

In periphery, ApN effects have been shown to depend on the molecular structure of the adipokine (globular or full-length) and the expression of the ApN receptor subtypes R1 and/or R2 in target cells (Yamauchi et al., 2002, 2003). However, nothing is known yet about microglial expression of ApN receptors. By qRT-PCR and immunocytochemistry analyzes, we studied AdipoR1 and AdipoR2 expression in BV2 cells. Our results showed that BV2 cells expressed both AdipoR1 and AdipoR2 mRNAs and comparable levels of corresponding proteins (Figure 7C, same results were obtained on brain sorted microglia, data not shown). To deepen AdipoR1 and AdipoR2 involvement in ApN actions on microglia, AdipoR1 or AdipoR2 expression was downregulated with specific siRNAs, cells were pre-treated with saline, 1 or 10 μg/ml of gApN and then treated with LPS (0.5 μg/ml). By qRT-PCR and immunocytochemistry, we showed that siRNA-mediated AdipoR1 or AdipoR2 downregulation led to a very efficient and specific reduction of AdipoR1 or AdipoR2 mRNA and protein expression with only minimal compensatory expression of the other receptor (Figure 7C).

qRT-PCR analyzes of pro-inflammatory cytokine mRNA expression showed that AdipoR1 downregulation blocked the inhibitory actions of gApN on LPS-induced IL-1β, IL-6, and TNFα mRNA upregulation by microglia while AdipoR2 downregulation did not (Figure 7D). It is nonetheless noteworthy to remark that, in agreement with our previous results on microglia sorted out from AdipoR2^−/− mouse brains AdipoR2 downregulation in BV2 cells seemed to increase the microglial susceptibility toward LPS challenge, as it expressed more IL-1β, IL-6, and TNFα mRNAs than that transfected with scrambled siRNAs (Figure 7D). Altogether, these findings suggested that AdipoR1 may mediate the anti-inflammatory properties of gApN on microglia, although AdipoR2 may also be involved in global microglial sensitivity toward inflammatory challenges.

We next studied AdipoR1 and AdipoR2 involvements in gApN-dependent AMPKα phosphorylation and downstream intracellular molecular targets. By Western-blotting, we showed that AdipoR1 downregulation totally blocked the gApN-induced phosphorylation of AMPKα, while AdipoR2 downregulation did not change it (Figure 8A).

NF-κB signaling pathways are activated in response to extracellular stimuli, including LPS, leading to the induction of various genes, including iNOS and COX-2, in microglia (Dang et al., 2014). Thus, we studied iNOS and COX2 immunoreactivities in BV2 cells transfected with scrambled, AdipoR1 or AdipoR2 siRNAs, and treated with LPS only or pre-treated with gApN (1 or 10 μg/ml) before addition of LPS (0.5 μg/ml) for 15 h. Results revealed that, as expected, LPS induced a great increase of both iNOS and COX2 expression in scrambled siRNA-transfected BV2 cells that was significantly reduced by the highest dose (10 μg/ml) of gApN (Figure 8B). GApN inhibitory effects were totally abolished in cells where AdipoR1 was downregulated whereas they remained significant in cells transfected with siRNA-targeted AdipoR2 (Figure 8B). In agreement with these results, LPS-stimulated NO release was reduced by gApN in cells transfected with scrambled or AdipoR2-targeted siRNAs while it was not modified in cells transfected with AdipoR1-targeted siRNAs (Figure 8C).

We also investigated gApN effect on p65 NF-κB nuclear translocation by immunocytochemistry in siRNA-transfected cells. As previously showed by Western-blotting of nuclear fractions (Figure 6B), analysis of immunocytochemistry images revealed that percentage of p65 positive nuclei was increased by LPS treatment in scrambled-, AdipoR1 siRNA-, AdipoR2 siRNA- transfected cells. GApN pre-treatment before LPS administration partially blocked the percentage of p65 positive nuclei in scrambled- and AdipoR2 siRNA-transfected cells but not in AdipoR1 siRNA-transfected cells (Figures 8D,E). Globally, our findings indicated that gApN anti-inflammatory and anti-oxidant actions on microglia are mediated through an AdipoR1/AMPK/NF-κB dependent signaling pathway (Figure 9).

Next, we tested whether an AdipoR agonist, AdipoRon, could also exert anti-inflammatory effects on brain and microglia. Wt mice were chronically treated with vehicle or 1 mg/kg AdipoRon for 15 consecutive days and then administered with a single i.p. injection of saline or LPS (Table 1). As expected, LPS injection induced a significant increase in IL-1β, IL-6, and TNFα protein levels in prefrontal cortex, hippocampus and hypothalamus but AdipoRon chronic pre-treatment failed to prevent these pro-inflammatory effects (Table 1), although it crosses blood-brain barrier (unpublished observations).

We also examined whether AdipoRon could exert direct anti-inflammatory actions on BV2 cells (Figure 10). AdipoRon pre-treatment (10⁻⁶ M) did not statistically modify IL-1β, IL-6, and TNFα mRNA expression levels in LPS-treated BV2 cells as compared to vehicle pre-treatment, whatever the incubation period considered (Figure 10). Dose-response experiments were also performed but none of the concentrations tested limited the LPS-induced pro-inflammatory cytokines mRNA levels increase (data not shown).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.