Pro-inflammatory gene expression and neurotoxic effects of activated microglia are attenuated by absence of CCAAT/enhancer binding protein β

Abstract

Background: Microglia and astrocytes respond to homeostatic disturbances with profound changes of gene expression. This response, known as glial activation or neuroinflammation, can be detrimental to the surrounding tissue. The transcription factor CCAAT/enhancer binding protein β (C/EBPβ) is an important regulator of gene expression in inflammation but little is known about its involvement in glial activation. To explore the functional role of C/EBPβ in glial activation we have analyzed pro-inflammatory gene expression and neurotoxicity in murine wild type and C/EBPβ-null glial cultures.

Methods: Due to fertility and mortality problems associated with the C/EBPβ-null genotype we developed a protocol to prepare mixed glial cultures from cerebral cortex of a single mouse embryo with high yield. Wild-type and C/EBPβ-null glial cultures were compared in terms of total cell density by Hoechst-33258 staining; microglial content by CD11b immunocytochemistry; astroglial content by GFAP western blot; gene expression by quantitative real-time PCR, western blot, immunocytochemistry and Griess reaction; and microglial neurotoxicity by estimating MAP2 content in neuronal/microglial cocultures. C/EBPβ DNA binding activity was evaluated by electrophoretic mobility shift assay and quantitative chromatin immunoprecipitation.

Results: C/EBPβ mRNA and protein levels, as well as DNA binding, were increased in glial cultures by treatment with lipopolysaccharide (LPS) or LPS + interferon γ (IFNγ). Quantitative chromatin immunoprecipitation showed binding of C/EBPβ to pro-inflammatory gene promoters in glial activation in a stimulus- and gene-dependent manner. In agreement with these results, LPS and LPS+IFNγ induced different transcriptional patterns between pro-inflammatory cytokines and NO synthase-2 genes. Furthermore, the expressions of IL-1β and NO synthase-2, and consequent NO production, were reduced in the absence of C/EBPβ. In addition, neurotoxicity elicited by LPS+IFNγ-treated microglia co-cultured with neurons was completely abolished by the absence of C/EBPβ in microglia.

Conclusions: These findings show involvement of C/EBPβ in the regulation of pro-inflammatory gene expression in glial activation, and demonstrate for the first time a key role for C/EBPβ in the induction of neurotoxic effects by activated microglia.

Figure 1

Basic characterization of C/EBPβ^-/- mixed glial cultures. Secondary mixed glial cultures from C/EBPβ^+/+ (white bars) and C/EBPβ^-/- (black bars) show similar total cell numbers and microglial density in control conditions and after 16 h of LPS or LPS+IFNγ. A. C/EBPβ^+/+ and C/EBPβ^-/- total cell density was estimated by Hoechst-33258-positive nucleus counting. No significant differences were observed between genotypes. Wild-type cultures show a statistically significant increase of cell density after 16 h of LPS and LPS+IFNγ treatment compared to control; C/EPBβ-null cultures show no difference after treatments. Two-way ANOVA, followed by Bonferroni's test was applied. *p < 0.05; compared to C/EBPβ^+/+ control. (n = 4). B. Microglia as a percentage of total cells was estimated by CD11b-positive cell counting in C/EBPβ^+/+ and C/EBPβ^-/- cultures after 16 h treatments with LPS, LPS+IFNγ or vehicle. Significant differences among treatments groups or genotypes are not observed. Two-way ANOVA, followed by Bonferroni's test was applied. (n = 4). C. Secondary mixed glial cultures were immunostained for CD11b (green). Nuclei are stained with Hoechst-33258 (blue). Microglial cell numbers were similar for C/EBPβ^+/+ and C/EBPβ^-/- cultures. LPS and LPS+IFNγ induced morphological changes in microglial cells in both genotypes. Bar = 50 μm. D. A representative western blot shows levels of GFAP in C/EBPβ^+/+ and C/EBPβ^-/- mixed glial protein extracts 16 h after vehicle (control), LPS and LPS+IFNγ treatments. β-Actin was used for normalization. E. Densitometric analysis was used to quantify GFAP protein levels versus β-actin in 4 independent western blots in arbitrary units (a.u.). Changes in GFAP protein levels are not observed. Two-way ANOVA, followed by Bonferroni's test was applied. (n = 4). F. Secondary mixed glial cultures were immunostained for GFAP (red) showing a confluent astrocytic layer. Overlapping of astroglial cell bodies makes counting very difficult and imprecise. No differences in astroglial morphology or density among genotypes are observed. Nuclei are stained with Hoechst-33258 (blue). Bar = 50 μm. E. Lack of NOS2 expression in activated astrocytes. C/EBPβ^+/+ and C/EBPβ^-/- secondary mixed glial cultures were immunostained for GFAP (green) and NOS2 (red), and stained for Hoechst 33258 (blue), after 16 h of LPS+IFNγ treatment. A marked reduction in number of NOS2-positive cells is seen in C/EBPβ-null cultures. The representative merge images show clearly that NOS2-positive cells do not colocalize with GFAP-positive cells. Bar = 50 μm.

Figure 2

C/EBPβ expression in activated mixed glial cultures. Effect of 100 ng/mL LPS alone or in combination with 0.1 ng/mL IFNγ on C/EBPβ expression in secondary mixed glial cultures. A. C/EBPβ mRNA expression is upregulated in glial activation. Cultures were treated with LPS and LPS+IFNγ for 6 h and mRNA was analyzed by qPCR. Results are expressed as relative fold units (r.f.u.) of ΔΔCt values between C/EBPβ and actin + Rn18s as reference genes. One-way ANOVA followed by Dunnett's test is applied. *p < 0.05; **p < 0.01 compared to control. (n = 3). B. LPS and LPS + IFNγ (24 h) increase nuclear C/EBPβ immunostaining (red) in secondary mixed glial cultures. In the right upper corner, a detail shows overlapping between Hoechst 33258 nuclear staining and C/EBPβ. Images are representative of 5 independent experiments. Bar = 50 μm. C. A western blot shows levels of C/EBPβ in secondary mixed glial cultures treated with LPS or LPS + IFNγ for 24 h. The C/EBPβ isoforms are identified as Full/LAP and LIP. β-Actin is used for normalization. This experiment was done 4 times with similar results. D. Full/LAP (grey bars) upregulation after LPS and LPS+IFNγ is statistically significant compared to control. LIP (dashed bars) upregulation is statistically significant only for LPS treatment. One-way ANOVA, followed by Dunnett's test is applied. *p < 0.05; **p < 0.01 compared to respective control. (n = 4-5)

Figure 3

Binding of C/EBPβ to proinflammatory gene promoters in activated mixed glial cultures. A. C/EBPβ DNA binding activity was analyzed by gel shift and supershift assays. Nuclear proteins were extracted from secondary mixed glial cultures treated with vehicle (lanes 1, 4, 7), LPS (lanes 2, 5, 8) or LPS+IFNγ (lanes 3, 6, 9) for 2 h. The first lane represents the probe without nuclear extract incubation (free probe). Arrows indicate four shifted complexes. Complex IV is a C/EBPβ independent complex. Lanes 1 to 3 show C/EBPs shifting complexes in wild type condition. Supershift with anti-C/EBPβ antibody (lanes 4 to 6) shows the presence of C/EBPβ in I-III complexes in all treatments. Rabbit IgG (lanes 7 to 9) is used as negative control for the supershift assay. This image is representative of four independent experiments. B. Quantitative analysis of C/EBPβ binding to NOS2, IL-1β, IL-6 and TNFα promoters by qChIP in mixed glial cultures. The sequences and positions of every C/EBPβ binding site and the primers used for qPCR are found in table 2. IL-10 was used as positive control. The qChIP assay was carried out after 2 h of LPS, LPS+IFNγ or vehicle (control) treatment. The IgG bars represent the means for IgG/Control, IgG/LPS and IgG/LPS+IFNγ PCR values for each gene. Input refers to total DNA. % of input represents the percentage of qChIP/Input ratio. One-way ANOVA, followed by Bonferroni's multiple comparison test is applied. **p < 0.01; ***p < 0.001 compared to control. #p < 0.05; ##p < 0.01; ###p < 0.001 compared to LPS. (n = 3)

Figure 4

Reduced proinflammatory gene expression in C/EBPβ^-/- mixed glial cultures. Expression of pro-inflammatory (NOS2, IL-1β, IL-6 and TNFα) and anti-inflammatory (IL-4 and TGFβ1) genes in C/EBPβ^+/+ (white bars) and C/EBPβ^-/- (black bars) mixed glial cultures. Cultures were treated with LPS or LPS+IFNγ for 6 h and then mRNA levels were analyzed by qPCR. In wild type cultures LPS and LPS+IFNγ induce expression of the four pro-inflammatory genes studied but do not affect mRNA levels of the anti-inflammatory genes IL-4 and TGFβ1. Absence of C/EBPβ results in significant decreases in LPS-induced expression of NOS2 and in LPS+IFNγ-induced expression of NOS2 and IL-1β. Results are expressed as relative fold units of ΔΔCt value between gene of interest and actin + Rn18s as reference genes. Two-way ANOVA, followed by Bonferroni's test was applied. *p < 0.05, ***p < 0.001 compared to respective C/EBPβ^+/+ condition. ^#p < 0.05; ^##p < 0.01; ^###p < 0.001 compared to respective control. ^%p < 0.05; ^%%%p < 0.001 compared to respective LPS condition.

Figure 5

Absence of C/EBPβ dramatically decreases microglial NOS2 protein expression and NO production in activated mixed glial cultures. A. NOS2 protein levels in total protein extracts from mixed glial cultures were analyzed by western blot, followed by densitometry. Data is expressed as NOS2 versus β-actin band intensities. Cultures were treated for 16 h with LPS, LPS+IFNγ or vehicle. In C/EBPβ^+/+ cultures (white bars) NOS2 protein levels were detected after LPS treatment, but LPS+IFNγ induced a clear upregulation, in agreement with mRNA expression levels. In C/EBPβ^-/- mixed glial cultures (black bars), NOS2 protein levels decreased in LPS and LPS+IFNγ compared to C/EBPβ^+/+ cultures. Two-way ANOVA, followed by Bonferroni's test was applied. ***p < 0.001 compared with C/EBPβ^+/+ condition. ^###p < 0.001 compared with respective control condition. (n = 5). A representative western blot is shown in B. C. NO production is decreased in activated C/EBPβ^-/- glial cultures. NO levels were measured by colorimetric analysis 48 h after treatments and normalized per cell number. Values are reported as micromolar concentration ×10⁶ cells. NO levels in C/EBPβ^+/+ (white bars) cultures were upregulated after LPS and LPS+IFNγ treatments compared to controls. In C/EBPβ^-/- glial cultures (black bars), NO production is reduced in LPS+IFNγ treatment compared to wild-type NO levels. Two-way ANOVA, followed by Bonferroni's test was applied. ***p < 0.001 compared to C/EBPβ^+/+ condition. ^###p < 0.001 compared to respective control condition. ^%%%p < 0.001 compared to respective LPS condition. (n = 7-9). D. NOS2 is expressed by activated microglia. C/EBPβ^+/+ and C/EBPβ^-/- secondary mixed glial cultures were immunostained for CD11b (green) and NOS2 (red), and stained for Hoechst 33258 (blue) after 16 h of LPS+IFNγ treatment. The C/EBPβ^+/+ merged image shows colocalization of NOS2-positive cells and CD11b-positive cells. Bar = 50 μm.

Figure 6

Lack of C/EBPβ in activated microglia completely abolishes the neurotoxic effects of activated microglia in neuronal-microglial co-cultures. A. MAP2 immunostaining of wild-type neurons co-cultured with C/EBPβ^+/+ or C/EBPβ^-/- microglia was performed 48 h after LPS+IFNγ treatment. MAP2 staining shows a clear decrease of network fibres caused by activated C/EBPβ^+/+ microglia, but not by C/EBPβ^-/- microglia or by vehicle-treated microglia. Images are representative of 5 independent experiments. (Bar = 100 μm). B. Evaluation of neuronal viability by MAP2/ABTS/ELISA assay 48 h after treatment with LPS+IFNγ or vehicle (control). Results are presented as % of MAP2 immunostaining in control cultures. Treatment with LPS+IFNγ reduces MAP2 immunostaining in neurons cocultured with C/EBPβ^+/+ microglia, but not with C/EBPβ^-/- microglia. Two-way ANOVA, followed by Bonferroni's test was applied. ^##p < 0.01 compared with C/EBPβ^+/+ control; *p < 0.05 compared to C/EBPβ^+/+ LPS+IFNγ. (n = 5).