Anti-inflammatory effects of minocycline are mediated by retinoid signaling

Abstract

Background: Minocycline is a lipophilic tetracycline of increasing appeal in neuroscience as it inhibits microglial activation, a mechanism involved in numerous neuropsychiatric disorders. Own data point towards retinoid-mediated effects of minocycline in murine brain and skin, and towards a vicious cycle of neuroinflammation which is driven by microglial activation-induced breakdown of local retinoids such as retinoic acid (RA). We therefore sought to study minocycline's anti-inflammatory effects on human microglial-like monocyte-derived cells in the context of retinoid signaling.

Results: As hypothesized, minocycline exposure resulted in a substantial increase of RA levels in the human monocytic cell line THP-1. While pro-inflammatory stimulation with lipopolysaccharides resulted in increased tryptophane-degrading indoleamine-2,3-dioxygenase IDO-expression and TNF-α levels in primary human monocyte-derived microglial-like cells, this effect was attenuated by minocycline only in the presence of retinoids. The anti-inflammatory effects of minocycline on TNF-α expression were completely abolished by a pharmacological blockage of retinoic acid receptors (RARs) using BMS-493 and unaffected by selectively blocking retinoid-X-receptors using UVI-3003.

Conclusions: Our data indicate for the first time a RA-dependent, anti-inflammatory effect for minocycline in human microglial-like cells via inhibition of local RA turnover. The RA-dependent mode of action for minocycline appears to be predominantly mediated through RAR-signaling.

Keywords: Cytokines; Microglia; Minocycline; Neuroinflammation; Retinoic acid.

Fig. 1

Minocycline inhibits RA-degradation in human monocytes. a Representative chromatographs of retinoic acid (RA)-measurements in cell culture supernatants from THP1-cells, each chromatograph representing only a single experiment. Degradation of RA is substantially inhibited in the presence of minocycline (MINO). All-trans-RA elutes with a retention time of ~ 9 min. b Quantitative analysis revealing an inhibition of RA-degradation in monocyte cells. While pretreatment with 10 µM minocycline reveals nearly no effect on RA levels, treatment with 100 µM minocycline results in a striking inhibition of RA-degradation. Results are presented as mean ± SEM, assessed in 4 independent experiments

Fig. 2

Minocycline prevents LPS-induced increase of RA-degradation in activated macrophages. LPS-stimulation significantly reduces retinoic acid (RA)-levels in THP1-cells. This effect can be reduced via minocycline: While pretreatment with 10 µM minocycline (MINO) results in a trend towards attenuating RA degradation, treatment with 100 µM minocycline results in a significant inhibition of LPS-induced RA-degradation. Results are presented as mean ± SEM, assessed in 3 independent experiments

Fig. 3

Expression of IDO and KMO after LPS-induced pro-inflammatory stimulation of human microglial-like cells. LPS induces an increase of the expression of the kynurenine pathway-enzymes IDO (a) and KMO (b). Minocycline (MINO) potently inhibits stimulation of indoleamine-2,3-dioxygenase (IDO) expression only in the presence of retinol (ROL) (a). For Kynurenine-3-monooxygenase (KMO), the same trend can be observed, however reaching no statistical significance (b). Results are given as fold over vehicle (VEH), mRNA expression was calculated relative to GAPDH housekeeping gene expression (ΔΔCT-Method). Results are presented as means +/- SEM from 4 healthy controls, assessed in 4 independent experiments

Fig. 4

TNF-α and IL-6 synthesis after LPS-induced pro-inflammatory stimulation of human microglial-like cells. LPS-stimulated cytokine expression of human microglial-like cells: Minocycline inhibits the synthesis of pro-inflammatory cytokines TNF-α (a, b) and IL-6 (c, d) only in combination with retionids. Cytokine levels are presented relative to protein concentration. Tested conditions included vehicle (VEH), retinol (ROL), minocycline (MINO) and a combination of ROL and MINO, given as differences between LPS-stimulated and the unstimulated conditions. Values are given as mean ± SEM from at least four healthy controls in independent experiments

Fig. 5

Pharmacological blockage of RA receptors RXR and RAR. Co-treatment with pan-retinoid-X-receptor (RXR) antagonist UVI-3003 (UVI) does not affect the anti-inflammatory effect of the combination of retinoic acid (RA) and minocycline (MINO) on TNF-α synthesis while pan-retinoic acid receptor (RAR) antagonist BMS-493 (BMS) completely impairs this anti-inflammatory mechanism. Results are presented as LPS-stimulated TNF-α, normalized for protein concentration. Values are given as mean ± SEM from at least four healthy controls in independent experiments