A recent comparison of the murine BV-2 cell line with primary mouse microglia cells revealed highly overlapping gene expression profiles upon stimulation with LPS, although the response in BV-2 was generally less pronounced . Approximately 90% of the genes induced by LPS in BV-2 cells were also induced in primary microglia, and 50% of the genes were also affected in hippocampal microglia following in
vivo stimulation of mice by intracerebroventricular LPS injection. These observations indicate that the BV-2 cell line is a suitable model of murine microglia to study neuroinflammatory parameters.
In the present study, we employed BV-2 cells as a macrophage/microglia cell model to characterize the impact of endogenous and synthetic corticosteroids on NF-κB activation and on IL-6 and TNFR2 expression. We found that BV-2 cells functionally express MR, GR, and 11β-HSD1, and our results emphasize the importance of a well-balanced activity of MR, which stimulates pro-inflammatory mediators, and GR that counteracts these effects. The selective MR ligand aldosterone exclusively resulted in NF-κB activation and upregulation of IL-6 and TNFR2 expression, whereas the selective GR ligand dexamethasone had opposite effects. The concentration of aldosterone to activate MR in BV-2 cells was approximately one order of magnitude higher than anticipated based on its Kd of about 0.5 nM. The effect on IL-6 protein levels was more pronounced than that on mRNA expression, indicating increased translation and/or protein stability. Similar observations were made using dexamethasone, for which Kd values for GR of 3–4 nM have been reported [5, 35]. While 100 nM dexamethasone readily suppressed IL-6 protein it had no effect on mRNA expression. As a possible explanation an efflux pump expressed in BV-2 cells may lower the actual intracellular concentration of aldosterone and dexamethasone.
Corticosterone and 11-dehydrocorticosterone (upon conversion to corticosterone by 11β-HSD1) stimulated the expression of IL-6 and TNFR2 and activated NF-κB at low/moderate concentrations by acting through MR, whereas higher concentrations exerted suppressive effects by acting through GR. Over 95% of the circulating corticosterone is bound to transcortin and albumin. Under normal conditions peak corticosterone concentrations in mice and rats in the unstressed state range between 250 nM and 500 nM, thus assuming that the free fraction in plasma reaches concentrations up to 25 nM. The intracellular concentrations may differ significantly from this value depending on uptake and 11β-HSD-dependent metabolism. In the presented experiments, we used 25 nM as a low and 250 nM as a high corticosterone concentration in culture medium containing 10% FBS. Despite the reduced amount of serum proteins present in the culture medium, the capacity should be sufficient to bind most of the 25 nM corticosterone added, probably resulting in a free steroid concentration below 2 nM. Nevertheless, the MR with a Kd of 0.5 nM for corticosterone is expected to be occupied, whereas the GR with a Kd of 5 to 10 nM is probably not activated, thus reflecting the situation under normal physiological conditions. In contrast, at 250 nM the binding capacity of the FBS present in the culture medium is probably saturated, resulting in high unbound corticosterone levels, reflecting levels reached during stress conditions and leading to occupation of GR.
Upon further increasing corticosterone from concentrations needed for maximal induction of IL-6 expression, a rapid decline was observed (Figure 3A, B). This may be explained by the higher expression levels of GR compared with MR, suggesting that occupation of few GR molecules may be sufficient to suppress MR activity. It is not clear at present whether GR suppresses MR function by competing for coactivators/corepressors, by competing for binding sites on the promoter of a given target gene, or by formation of heterodimers. Interestingly, RU-486 enhanced IL-6 and TNFR2 expression in the absence of added steroids (Figure 3E, Figure 8B). In preliminary experiments using transfected cells, we observed ligand-independent MR activity that was lowered upon co-expressing GR. RU-486 might act as an inverse agonist and induce a conformational change upon binding to GR, which may prevent heterodimer formation or, alternatively, GR may compete with MR for a corepressor protein, thereby increasing MR activity.
Nevertheless, the results suggest a tightly controlled and coordinated action of MR and GR in the regulation of NF-κB activity and the production of and sensitivity to pro-inflammatory cytokines in microglia cells. Pro-inflammatory cytokines such as TNF-α, IL-1β, and IL-6, and subsequent activation of NF-κB, lead to elevated expression and activity of 11β-HSD1, which results in enhanced local levels of active glucocorticoids (this study, and ). The fact that both GR and MR promote 11β-HSD1 expression may represent an important feed-forward regulation of glucocorticoid activation in order to increase the intracellular concentration of active glucocorticoids and to shift the balance from an initially predominantly MR-mediated stimulation to a GR-mediated suppression of inflammation. The enhanced local glucocorticoid activation may be necessary for the resolution of inflammation. The suppression of pro-inflammatory cytokines and NF-κB activity upon activation of GR should then allow normalization of 11β-HSD1 expression and activity.
Evidence for a stimulation of inflammatory parameters by low to moderate glucocorticoid concentrations was also obtained from recent experiments with murine 3 T3-L1 adipocytes . Decreased adiponectin mRNA and increased IL-6 mRNA were observed upon incubation of 3 T3-L1 adipocytes with 100 nM corticosterone or cortisol. These effects were partially reversed by treatment with the MR antagonist eplerenone but not by GR antagonist RU486, suggesting that MR activation was responsible for the observed effects. Furthermore, 100 nM of the glucocorticoids increased NADPH oxidase subunit p22 mRNA levels and decreased catalase mRNA levels, which were reversed by co-treatment with eplerenone. However, the authors did not test the effects of various concentrations, nor was the role of 11β-HSD1 addressed.
Other investigators observed elevated 11β-HSD1 mRNA and activity in 3 T3-L1 adipocytes that were treated with LPS, TNF-α, or IL-1β . Pharmacological inhibition of 11β-HSD1 diminished the TNF-α-induced activation of NF-κB and MAPK signaling. Their study, however, did not assess the role of MR and the impact of higher glucocorticoid concentrations. At high glucocorticoid levels, pharmacological inhibition of 11β-HSD1 might abolish GR activation, thereby promoting MR-mediated pro-inflammatory effects. Since actual glucocorticoid concentrations in intact microglia cannot be measured, it will be important to assess the effects of 11β-HSD1 inhibitors on neuroinflammation in vivo in future studies. Nevertheless, these observations are in line with our findings in microglia cells that low/moderate glucocorticoid concentrations mainly act through MR, thereby promoting inflammatory parameters.
During acute inflammation, TNF-α acts through membrane-bound TNF-receptors on macrophage and microglial cells, leading to activation of transcription factors such as NF-κB and AP1, which can induce a second wave of pro-inflammatory cytokines, including TNF-α, IL-1β and IL-6. TNFR2 is highly expressed on microglia cells and plays an important role in the regulation of innate immune response following brain injury on infection . An elevated expression of TNFR2 upon MR activation, as observed in the present study, is expected to result in a higher sensitivity and more pronounced response to external pro-inflammatory stimuli.
Disruption of MR- and GR-mediated regulation of gene transcription and interaction with other transcription factors can occur at several levels. Reduced GR activity and/or enhanced MR activity, thus exacerbating inflammation, may be caused by the presence of xenobiotics differentially modulating receptor activity, post-translational receptor modifications, altered function of receptor-associated proteins, or altered protein stability . The pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 were shown to activate the HPA axis , thereby enhancing circulating glucocorticoids and exerting suppressive effects through GR activation. However, high levels of TNF-α have been associated with glucocorticoid resistance . Upon excessive HPA activation, a downregulation of GR activity, probably caused by altered phosphorylation of the receptor and reduced protein stability [38, 39], with concomitant glucocorticoid resistance has been observed, which may cause a shift from GR- to MR-mediated glucocorticoid effects. GR blockade by administration of RU486 or elimination of glucocorticoids by adrenalectomy sensitized C57BL/6 mice to low-dose TNF-α . Moreover, hepatic GR-deficient mice showed significantly higher levels of IL-6 in response to TNF-α treatment.
Glucocorticoid-resistance represents a major problem in chronic inflammation, including rheumatoid arthritis, ulcerative colitis, Crohn’s disease, atherosclerosis, cystic fibrosis, and chronic obstructive pulmonary disease . An impaired suppression by GR may lead to chronically enhanced MR activity. It remains to be investigated whether MR antagonists may prove beneficial in these diseases.
Increasing evidence indicates that neuroinflammation contributes to neuronal degeneration and the progression of Parkinson’s disease [41, 42]. Activated microglial cells and increased expression of pro-inflammatory mediators have been found in the substantia nigra of patients. Interestingly, elevated circulating cortisol levels were measured in Parkinson’s disease patients together with decreased GR expression in the substantia nigra . Selective ablation of GR in macrophage/microglia exacerbated the loss of dopaminergic neurons induced by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), enhanced the production of pro-inflammatory parameters, and diminished the expression of anti-inflammatory mediators. Based on the findings of the present study, we hypothesize that the potentiation of neuroinflammation in GR-deficient states is due to an impaired balance of pro- and anti-inflammatory mediators as a result of a dysbalance of MR and GR activity. The role of MR in Parkinson’s disease and whether MR antagonists may prove useful in the treatment of this disease remain to be investigated.