Inhibition of COX-2-mediated eicosanoid production plays a major role in the anti-inflammatory effects of the endocannabinoid N-docosahexaenoylethanolamine (DHEA) in macrophages

Abstract

Background and purpose: N-docosahexaenoylethanolamine (DHEA) is the ethanolamine conjugate of the long-chain polyunsaturated n-3 fatty acid docosahexaenoic (DHA; 22: 6n-3). Its concentration in animal tissues and human plasma increases when diets rich in fish or krill oil are consumed. DHEA displays anti-inflammatory properties in vitro and was found to be released during an inflammatory response in mice. Here, we further examine possible targets involved in the immune-modulating effects of DHEA.

Experimental approach: Antagonists for cannabinoid (CB)1 and CB2 receptors and PPARγ were used to explore effects of DHEA on NO release by LPS-stimulated RAW264.7 cells. The possible involvement of CB2 receptors was studied by comparing effects in LPS-stimulated peritoneal macrophages obtained from CB2 (-/-) and CB2 (+/+) mice. Effects on NF-κB activation were determined using a reporter cell line. To study DHEA effects on COX-2 and lipoxygenase activity, 21 different eicosanoids produced by LPS-stimulated RAW264.7 cells were quantified by LC-MS/MS. Finally, effects on mRNA expression profiles were analysed using gene arrays followed by Ingenuity(®) Pathways Analysis.

Key results: CB1 and CB2 receptors or PPARs were not involved in the effects of DHEA on NO release. NF-κB and IFN-β, key elements of the myeloid differentiation primary response protein D88 (MyD88)-dependent and MyD88-independent pathways were not decreased. By contrast, DHEA significantly reduced levels of several COX-2-derived eicosanoids. Gene expression analysis provided support for an effect on COX-2-mediated pathways.

Conclusions and implications: Our findings suggest that the anti-inflammatory effects of DHEA in macrophages predominantly take place via inhibition of eicosanoids produced through COX-2.

Linked articles: This article is part of a themed section on Cannabinoids 2013 published in volume 171 issue 6. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.2014.171.issue-6/issuetoc.

© 2014 The British Pharmacological Society.

Figures

Figure 1

Inhibitory effect of DHEA on IL-6 release from LPS-stimulated C57Bl/6 mouse peritoneal macrophages (A) and RAW264.7 cells (B). Cells were pre-incubated with DHEA for 30 min before LPS stimulation and IL-6 was measured after 16 h (peritoneal macrophages) or 24 h (RAW264.7 cells) of incubation with DHEA or vehicle in the presence of LPS. Data are expressed as % of the IL-6 release with LPS and solvent only (=100%). Bars represent means ± SEM from four separate mice experiments (three measured in triplicate and one in duplicate; A), and four separate experiments with the RAW264.7 cell line (measurements in duplicate; B). Asterisks indicate significant difference from the control (one-way anova, Dunnett's t-test; **P < 0.01). IL-6 levels (ng·mL−1) + SEM; for RAW264.7 cells, control 0; LPS 1 μg·mL−1 31.4 + 2.6; DHEA 10 μM 17.5 + 1.9; for peritoneal macrophages, control 0; LPS 0.1 μg·mL−1 3.1 + 1.6; DHEA 10 μM 1.7 + 0.7.

Figure 2

DHEA reduces NO release in both the MyD88-dependent and the MyD88-independent pathway. Dose–response graph showing poly-IC- and LPS-induced NO release modulated by DHEA in RAW264.7 mice macrophages. RAW264.7 macrophages were pre-incubated for 30 min with DHEA and stimulated for 48 h with either LPS or poly-IC (1 μg·mL−1) in the presence of DHEA. Data are expressed as % of the NO release with LPS and solvent only (=100%). Bars represent means ± SEM from three separate experiments with duplicate measurements. Asterisks indicate significant difference from the control [one-way anova, Dunnett's t-test (one-sided); *P < 0.05 and **P < 0.01]. [nitrite levels (μM) ± SEM: control 1.1 ± 0.7, poly-IC 1 μg·mL−1 26.3 ± 5.6, DHEA 0.01 μM 29.8 ± 6.5, DHEA 0.1 μM 24.7 ± 3.9, DHEA 1 μM 24.9 ± 5.6, DHEA 5 μM 21.9 ± 4.5, DHEA 10 μM 18.6 ± 3.6 *; n = 3, *P < 0.05 vs. poly IC alone]. [nitrite levels (μM) ± SEM: control 1.1 ± 0.3, LPS 1 μg·mL−1 28.3 ± 2, DHEA 0.01 μM 29.8 ± 2.8, DHEA 0.1 μM 28.9 ± 1.9, DHEA 1 μM 23.2 ± 3, DHEA 5 μM 16.3 ± 3**, DHEA 10 μM 13.1 ± 2.1**; n = 3, **P < 0.01 vs. LPS alone].

Figure 3

(A) Effect of DHEA on TNFα-induced NF-κB activity measured in a HEK293 NF-κB lacZ luciferase reporter assay containing a promoter sequence with five NF-κB binding sites. Cells were stimulated for 4 h with TNFα (50 ng·mL−1) in the presence of different concentrations of DHEA (0.1, 1, 5, 10 and 15 μM). CAPE (3 μg·mL−1), a synthetic NF-κB inhibitor was used as positive control. A 24 h time point gave similar results. (B) Effect of DHEA on IFN-β production of poly-IC stimulated RAW264.7 macrophages. Cells were stimulated for 24 h with poly-IC (1 μg·mL−1) in the presence of 1 or 10 μM of DHEA. Experimental details are described in the Methods. Data are expressed as %, where TNFα (A) or poly-IC (B) stimulation (containing vehicle) was set at 100%. Data represent means ± SEM of three separate experiments (each done in duplicate). Asterisks indicate significant difference from the control (one-way anova, Dunnett's t-test; **P < 0.01, ***P < 0.001). IFN-β levels (ng·mL−1) + SEM; control 0; poly-IC 1 μg·mL−1 1.05 + 0.2; DHEA 10 μM 1.27 + 0.3.

Figure 4

NO reduction induced by DHEA in LPS-stimulated peritoneal macrophages from CB2+/+ and CB2−/− mice did not differ. Cells were pre-incubated for 30 min with each ligand (0.1, 1 or 10 μM) and NO levels were measured after 48 h stimulation with LPS (1 μg·mL−1) in the presence of the respective ligand and concentration. Further details are described in the Methods. Data are expressed as a %, where LPS stimulation (containing vehicle) was set at 100%. Data represent means ± SEM of n = 4 CB2+/+ and n = 5 CB2−/− mice (each done in triplicate). No significant difference (n.s.) was found between CB2+/+ and CB2−/− (repeated measures). A significant overall treatment effect was found for DHEA.

Figure 5

DHEA-induced reduction of NO production does not seem to involve the CB1 or CB2 receptors or PPARγ. NO reduction induced by DHEA in LPS-stimulated RAW264.7 macrophages was not significantly reversed by rimonabant, SR144528, a combination of rimonabant and SR144528, or GW9662 (P > 0.05 for all combinations). Cells were incubated for 30 min with 1 μM of the respective antagonist after which the regular pre-incubation with DHEA (10 μM) was started (see Methods for further details). NO levels were determined after 48 h LPS (1 μg·mL−1) stimulation in the presence of the respective antagonist and ligand. Data are expressed as %, where LPS stimulation (containing vehicle) was set at 100%. Data represent means ± SEM (n = 3–4 experiments; each done in duplicate). Statistical analysis was performed by one-way anova, Dunnett's t-test, with *P < 0.05 (with respect to LPS control), **P < 0.01 (with respect to LPS control), ***P < 0.001 (with respect to LPS control). DHEA + antagonists (all combinations) were significantly different with respect to LPS control (not depicted in figure).

Figure 6

(A) DHEA reduced PGE2 production in LPS-stimulated RAW264.7 macrophages. Cells were pre-incubated for 30 min with 1 or 10 μM DHEA and stimulated for 24 h with LPS (0.1 μg·mL−1) in the presence of the respective ligand and concentration. Data represent means ± SEM of three separate experiments (each done in duplicate). The right figure shows the mean (± SEM) of another set of three separate experiments. Here, cells were stimulated with 1 μg·mL−1 LPS and treated with 10 μM DHA or 10 μM DHEA. All other conditions were similar as mentioned for the left figure. Experimental details are described in the Methods. Data are expressed as percentage, where LPS stimulation (containing vehicle) was set at 100%. Asterisks indicate significant difference from the control (one-way anova, Dunnett's t-test; *P < 0.05, **P < 0.01 and ***P < 0.001). PGE2 levels (pg·mL−1) + SEM; for A, control 19.7 + 9.5; LPS 0.1 μg·mL−1 204.1 + 81; DHEA 10 μM 49.6 + 10.1. (B) DHEA does not influence COX-2 at gene expression level. COX-2 expression was assessed using microarray analysis. Cells were pre-incubated for 30 min with 10 μM DHEA and stimulated for 24 h with LPS (0.1 μg·mL−1) in the presence of DHEA. As control non–LPS-stimulated cells were used. Graph shows means ± SEM of three separate experiments (each performed in duplo), no significant effect was found (n.s.); Q value > 0.05. (C) DHEA partly influences COX-2 at the protein level. Upper, representative example of Western Blot analysis showing COX-2 protein in RAW264.7 macrophages after 24 h stimulation with 1 μg·mL−1 LPS with or without DHEA. Antibodies were used against COX-2 or the control actin. Lower panel shows densitometry of COX-2 protein expression, corrected for actin expression. Data show mean ± SEM of six independent experiments (each performed in duplicate). Asterisks indicate significant difference from the control (one-way anova, Dunnett's t-test); *P < 0.05 ***P < 0.001.

Figure 7

Effect of DHEA on eicosanoid synthesis of LPS-stimulated RAW264.7 macrophages. DHEA reduced medium levels of several COX-derived arachidonic acid eicosanoids, and increased the levels of 19,20-DiHDPA, a DHA metabolite, after LPS (1 µg.ml-1) stimulation. Data represent means ± SEM of three independent experiments. Asterisks indicate significant difference from the control [one-way anova, Dunnett's t-test (one-sided)]; *P < 0.05 **P < 0.01 ***P < 0.001.

Figure 8

IPA showing activation Z-scores of upstream regulator mediators found to have highest overlapping opposite changes in gene-expression profiles compared with DHEA-induced gene-expression profiles. Activation Z-scores are shown for profiles significantly overlapping with P < 0.01 and cut off value of – 2.