Background
Nuclear receptors are a large group of transcription factors that comprises 49 family members which enable the organism to quickly adapt to environmental changes by inducing the appropriate gene program [
1]. Nuclear receptors are activated by lipophilic ligands including hormones, fatty acids, bile acids and oxysterols, homo- and heterodimerize upon activation and regulate the metabolism of glucose, fatty acids, triglycerides and lipoproteins [
1]. Besides these metabolic effects, numerous immune-modulatory effects by nuclear receptors have been described, including modulation of central nervous system (CNS) autoimmunity [
2‐
4].
Interestingly, several members of the nuclear receptor family are also involved in molecular mechanisms regulating oligodendroglial differentiation. Downregulation of retinoid X receptor gamma (RXRγ) expression or application of RXRγ antagonists significantly impaired oligodendroglial differentiation. Similarly, mice lacking RXRγ showed delayed remyelination after induction of demyelination in vivo [
5]. In oligodendrocytes, RXRγ heterodimerizes with the vitamin D receptor (VDR) to induce oligodendroglial differentiation, and accordingly, vitamin D-induced activation accelerates oligodendroglial differentiation [
6]. Other nuclear receptors, such as thyroid receptors (TR) and liver X receptors (LXR), are well expressed in oligodendroglial lineage cells and modulate oligodendroglial differentiation and myelination. Patients with congenital hypoparathyroidism typically develop a hypomyelinating phenotype, and triiodothyronine (T3) is required for the maturation of oligodendroglial progenitor cells into mature myelinating oligodendrocytes (for review see [
7]). Mice lacking both LXRα and LXRβ display thinner myelin sheaths and reduced expression levels of myelin genes in the cerebellum. Furthermore, addition of pharmacological activators of LXRs to mixed glial cell cultures promotes oligodendroglial differentiation [
8].
The nuclear receptor farnesoid-X-receptor (FXR; NR1H4) is expressed not only in the liver, gut, kidney and adipose tissue, but also in the immune cells [
9,
10]. FXR heterodimerizes with other nuclear receptors, such as RXRs [
11,
12]. The best-characterized endogenous FXR ligands are lipophilic bile acids, namely cholic acid and chenodeoxycholic acid. In addition, the synthetic compound GW4064 serves as an agonist with high receptor specificity and efficacy [
11,
13,
14]. In accordance with its expression pattern, FXR activation has been shown to confer protection in several animal models of innate intestinal and hepatic inflammation [
11]. In experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis (MS), pharmacological activation of FXR significantly ameliorated the disease course by induction of anti-inflammatory macrophages [
15,
25]. Interestingly, FXR agonists are currently tested in clinical trials for treatment of alcoholic hepatitis, type 2 diabetes mellitus and primary biliary cirrhosis, demonstrating that FXR already represents an attractive pharmacological target in human metabolic diseases. The beneficial effect of FXR agonists in EAE suggests that FXR might represent a potential target in inflammatory-demyelinating CNS diseases, such as MS. Oligodendroglial lineage cells are not only a primary target in MS but also a prerequisite for remyelination, an endogenous repair mechanism. However, especially in chronic MS lesions, remyelination is limited at least partly due to a differentiation block inhibiting the differentiation of oligodendroglial progenitor cells (OPCs) into myelinating oligodendrocytes [
16‐
18]. Importantly, remyelination failure contributes to axonal damage and worsening of clinical signs in demyelinating animal models [
19,
20].
Given its recently illustrated relevance in macrophage-mediated protection from CNS autoimmunity and the functional role of several nuclear receptors for oligodendroglial differentiation, we addressed the functional relevance of FXR in different brain-resident cells, especially focusing on oligodendrocytes. Here, we show that FXR is expressed in murine oligodendrocytes. However, lack or activation of FXR did not influence oligodendroglial differentiation in vitro or in vivo. Furthermore, FXR activation did not affect key immune functions of astrocytes and microglial cells. These data demonstrate that, in contrast to other nuclear receptors, FXR is dispensable for oligodendroglial differentiation as well as for brain-resident astrocytes and microglia and hence suggest that pharmacological activation of FXR to downregulate CNS inflammation does not interfere with repair mechanisms, such as remyelination.
Materials and methods
Animals
FXR Ko mice [
21] were purchased from Jackson Laboratory and C57Bl/6 wild-type control mice from the animal facility in Münster, Germany. Animal handling and experiments were conducted according to the German Animal Welfare Act and approved by the responsible governmental authorities (LANUV Nordrhein-Westfalen; AZ-8.84.-02.05.20.12.286, 84.-02.04.2013.A029, AZ 84-02.04.2014.A221, AZ 84-02.04.2013.A374).
Primary oligodendroglial culture and treatment
Primary OPCs were isolated using the immunopanning method as described earlier [
22,
23]. OPCs/differentiating oligodendrocytes were incubated with 1 or 10 μM GW4064 (Tocris, 2473) for 24/48 h. Cell viability upon treatment was measured according to manufacturer’s protocol using CellTiter-Glo® Luminescent Cell Viability Assay (Promega, G7570). When treated with bone marrow-derived macrophage (BMM) supernatant, oligodendrocytes were incubated with 50% supernatant in culture medium for 48 h.
Cerebellar slice cultures and treatment
Sagittal, cerebellar slices from P1 mice were dissected as described earlier [
23]. After 12 days in culture, demyelination was induced by 16-h incubation with 0.5 mg/ml lysolecithin. During subsequent remyelination, the brain slices were incubated with 10 or 20 μM GW4064 for 14 days. The ex vivo slices were fixed and stained with anti-neurofilament 70 kDa (NFL; Dako, 1:500) and anti-MBP (Abcam, ab7349, 1:200). Confocal
z-stacks of whole slices were acquired with a Zeiss LSM 700 confocal microscope, and images were analysed using NIH ImageJ. The extent of myelination was quantified by determining the ratio between MBP and NFL staining. Two independent experiments were performed, and in each experiment, six brain slices and three z-stacks from six brain slices were analysed for each concentration of GW4064 or DMSO.
Macrophage and microglia culture and treatment
Preparation of bone marrow-derived macrophages and generation of primary astrocyte cultures via mixed glial culture preparation were performed as described previously [
2]. The murine embryonic stem cell-derived microglial precursor cells (ESdMs) were provided by Harald Neumann (University of Bonn) and cultured according to their publication [
2,
24]. Depending on the individual cell type, cells were treated with different concentrations of the synthetic FXR agonist GW4064 (Tocris, 2473); concentrations depended on individual cell type. Control groups were treated with the equivalent volume of solvent, i.e. DMSO. ESdMs and primary astrocytes were incubated for 24 h with the BMM-conditioned supernatant, before stimulation with LPS and IFNγ.
Generation of BMM supernatant
To produce BMM-conditioned supernatant, BMMs were treated with 15 μM GW4064 or DMSO and cultured for 72 h [
25].
Detection of TNFα and nitric oxide
TNFα in cell culture supernatants was quantified using ELISA Ready-SET-Go!® (murine TNFα; eBioscience) according to the manufacturer’s recommendations. Nitrite concentrations were measured in the supernatants using the Griess Reagent Kit (Invitrogen) according to the manufacturer’s instructions.
RT2 Profiler™ PCR Array
Treated ESdMs and astrocytes were stimulated with LPS and INFγ for 6 h. RNA isolation was performed with RNeasy Mini Kit (Qiagen) combined with the RNase-Free DNase Set (Qiagen) to digest DNA. For transcription to complementary DNA (cDNA), the RT2 FirstStrand Kit (Qiagen) was utilized. Afterwards, the RT2 Profiler™ PCR Array Inflammatory Response & Autoimmunity (#PAMM-077Z; Qiagen) was performed with RT2 Real-Time SYBR Green PCR Master Mix (SuperArray Bioscience) according to the manufacturer’s protocol. For normalization, the housekeeping gene Hsp90ab1 was taken as internal control. Analysis was performed at the RT2 Profiler PCR Array Data Analysis Web portal (provided from Qiagen).
RNA isolation and qRT-PCR
Total RNA from cells was isolated using peqGOLD Total RNA Kit (12-6634, PeqLab Biotechnologie GmbH). Messenger RNA (mRNA) was transcribed into cDNA by reverse transcription reaction (High Capacity cDNA Transcription Kit, Applied Biosystems), and cDNA was diluted to a final concentration of 0.75 ng/μl. qRT-PCR was performed using Power SYBR® Green PCR Master Mix (Applied Biosystems) and StepOne Plus real-time cycler (Applied Biosystems). The following primers were used: Fxr fw: GCCACAGATTTCCTCCTCGT, Fxr rev: CAGTCTCTCCCTGGTACCCA, Mbp fw: GTACAAGGACTCACACACGAGA, Mbp rev: GTTCGAGGTGTCACAATGTTCT, RPLP0 fw: CGACCTGGAAGTCCAACTAC and RPLP0 rev: ATCTGCTGCATCTGCTTG, and data were normalized using RPLP0 as internal control.
Immunocytochemistry (ICC)
OPCs were fixed directly after seeding or differentiated and fixed after 48 h as mature oligodendrocytes. Cells were permeabilised for 10 min in 0.5% Triton X-100 in PBS and blocked using 5% FCS/PBS for 30 min. The primary antibodies were rat anti-MBP (Abcam, ab7349, 1:200), rabbit anti-PDGFRα (Santa Cruz, SSC338, 1:300), rabbit-FXR NR1H4 (Abcam, ab28676, 1:200; or Santa Cruz sc-13063, 1:100) and mouse anti-Olig2 (Medac, 387M-16, 1:200). Incubation was performed at 4 °C over night. Secondary antibody staining was performed using Cy™3 Anti-Rat (1:500) (Jackson, 112-165-167) and anti-Rabbit Alexa Fluor® 488 conjugate (1:500) (Jackson, 111-545-144) for 2 h at RT before embedding with Roti®-Mount FluorCare DAPI (Carl Roth, HP20.1). Images were taken using the laser scanning microscope (LSM 700, Zeiss Jena). At least 200 cells were quantified, and the numbers of MBP+ and PDGFRα+ cells were assessed as percentage of total DAPI+ cells.
Immunohistochemistry (IHC)
Ten-day and 8-week-old WT or FXR Ko mice were sacrificed and intracardially perfused. The spleens, spinal cords and brains were removed and fixed in 4% PFA overnight. Paraffin sections (4 μm) were pretreated with citrate buffer (pH 6) and stained using an automated immunostainer (AutostainerLink 48, Dako). Primary antibodies were specific to FXR (NR1H4, rabbit, Abcam ab28676 1:200; or Santa Cruz sc-13063, 1:50), NogoA (mouse, 11c7, a generous gift from M.E. Schwab, Brain Research Institute, University of Zürich and Department of Biology, Swiss Federal Institute of Technology Zürich, Switzerland, 1:15,000) and Olig2 (rabbit, 18953, IBL, 1:150 and mouse, 387M-16, Medac, 1:200). Numbers of oligodendroglial cells were quantified in a blinded fashion in the corpus callosum, cerebellum and spinal cord in standardized microscopic fields of 10,000 μm2 each defined by an ocular morphometric grid.
Statistics
All cell culture experiments were performed in triplicates and replicated at least three times. All statistical analyses were performed using GraphPad Prism 5.03 (GraphPad Software Inc., San Diego, CA). In text and figures, results are provided as mean ± SEM. Multiple comparisons in the same data set were analysed by the Bonferroni-corrected (for selected groups) one-way ANOVA tests. Single comparisons to control were made using two-tailed Student’s t test. p values <0.05 were considered significant.
Discussion
Here, we demonstrate that FXR is expressed by OPCs and mature oligodendrocytes, like microglia and astrocytes [
25]. However, lack of FXR did not affect oligodendroglial differentiation in vitro or in vivo. Furthermore, activation of FXR using the synthetic agonist GW4064 did not affect oligodendroglial differentiation, ex vivo remyelination or pro-inflammatory activation of astrocytes or microglia.
Several nuclear receptors, such as RXRα, RXRβ, RXRγ, VDR, TR, LXRα and LXRβ, are expressed by oligodendrocytes and have been shown to affect oligodendroglial function [
5,
6,
8,
26]. Lack of TR or LXRα/LXRβ results in delayed oligodendroglial differentiation or hypomyelination in the cerebellum, respectively [
8,
26]. Similarly, blocking and/or downregulation of VDR, TRs or RXRγ impaired oligodendroglial differentiation [
5,
27]. In contrast, our results provide no indication that lack of FXR delays oligodendroglial differentiation in vitro or in vivo. FXR forms a heterodimer with RXR to induce expression of target genes, but FXR can also function as a repressor by suppressing gene expression [
13]. Our results suggest that, at least in oligodendrocytes, the function of FXR is redundant to or is compensated by other nuclear receptors.
Activation of several nuclear receptors such as RXRγ, TR, VDR and LXRα/LXRβ promotes oligodendroglial differentiation [
3,
5‐
8]. This is in contrast to our results, which show that activation of FXR by the synthetic agonist GW4064 had no effect on the number of MBP+ mature oligodendrocytes or expression levels of MBP. Oligodendrocytes display a region-dependent heterogeneity within the CNS [
28], and interestingly, lack of LXRα/β is associated with hypomyelination in the cerebellum [
8]. To determine whether FXR might play a more prominent role in the differentiation of cerebellar oligodendrocytes, we repeated our in vitro experiments using oligodendrocytes isolated specifically from the cerebellum. However, neither lack nor activation of FXR modulated the differentiation of cerebellar oligodendrocytes. Likewise, addition of FXR agonists to demyelinated cerebellar slice cultures did not exhibit an effect on remyelination in this ex vivo model in which exclusively CNS-resident cells are present.
It has recently been shown that activation of FXR in BMM induced an anti-inflammatory phenotype characterized by downregulation of pro-inflammatory and induction of anti-inflammatory genes [
25]. In light of the finding by Miron and colleagues who demonstrated an increased differentiation of rodent oligodendrocytes cultured in the presence of supernatants derived from anti-inflammatory microglial cells, we hypothesized that FXR-activated BMM might indirectly modulate oligodendrocyte differentiation [
29]. However, our results provide no evidence that activation of FXR in macrophages has a similar, indirect effect on oligodendroglial differentiation. Oligodendroglial differentiation and remyelination can also be accelerated in vivo by improved clearance of myelin debris by macrophages as it has been shown for the RXR agonist bexarotene [
30]. Whether activation of FXR in macrophages may potentially result in increased myelin phagocytosis is currently unknown. However, this could indirectly contribute to oligodendrocyte differentiation and remyelination.
The anti-inflammatory phenotype induced by GW4064-mediated activation of FXR in peripheral myeloid cells, such as BMMs, is well documented [
15,
25]. However, expression of FXR is not limited to myeloid cells, but is also present in microglia as well as astrocytes [
25]. Interestingly, FXR activation in microglia using GW4064 did not result in downregulation of pro-inflammatory key molecules, such as TNFα and NO, or altered expression of pro- and anti-inflammatory genes encoding for cytokines, interleukins or chemokines as well as their receptors. These data illustrate that activation of FXR via GW4064 has no influence on the immune responses of microglia and astrocytes. This contrasts with the activation of other nuclear receptors such as LXR, RXR and PPARα [
31,
32]. Considering previously published data [
25], our results suggest that FXR plays a pivotal role in downregulation of pro-inflammatory molecules in peripheral myeloid cells, but not in CNS-resident microglia suggesting heterogeneity between the two cell populations despite overlapping functions. This may be due to a considerable diversity of gene expression patterns between tissue macrophages from different mouse organs [
33] and could further explain the recent observation that blood-derived monocytes might mainly initiate demyelination in EAE whereas microglia potentially primarily clear myelin debris [
34].
Conclusion
In summary, we demonstrated that activation of FXR by GW4064 does not modulate oligodendroglial differentiation directly or indirectly via FXR activation in macrophages. Furthermore, there is no indication that GW4064 directly or indirectly influences the pro-inflammatory profiles of CNS-resident astrocytes or microglia. These data suggest that FXR agonists, such as GW4064, represent a potential therapeutic approach for MS which specifically targets peripheral immune cells including macrophages, but not brain-resident cells, such as oligodendrocytes, astrocytes or microglia.
Acknowledgements
The authors thank Harald Neumann (University of Bonn) for providing the murine embryonic stem cell-derived microglial precursor cells (ESdMs). Additionally, the authors thank Elke Hoffmann and Claudia Kemming for excellent technical assistance.