Identification of a selective glucocorticoid receptor modulator that prevents both diet‐induced obesity and inflammation

Abstract

Background and Purpose

High‐fat diet consumption results in obesity and chronic low‐grade inflammation in adipose tissue. Whereas glucocorticoid receptor (GR) antagonism reduces diet‐induced obesity, GR agonism reduces inflammation, the combination of which would be desired in a strategy to combat the metabolic syndrome. The purpose of this study was to assess the beneficial effects of the selective GR modulator C108297 on both diet‐induced weight gain and inflammation in mice and to elucidate underlying mechanisms.

Experimental Approach

Ten‐week‐old C57Bl/6 J mice were fed a high‐fat diet for 4 weeks while being treated with the selective GR modulator C108297, a full GR antagonist (RU486/mifepristone) or vehicle.

Key Results

C108297 and, to a lesser extent, mifepristone reduced body weight gain and fat mass. C108297 decreased food and fructose intake and increased lipolysis in white adipose tissue (WAT) and free fatty acid levels in plasma, resulting in decreased fat cell size and increased fatty acid oxidation. Furthermore, C108297 reduced macrophage infiltration and pro‐inflammatory cytokine expression in WAT, as well as in vitro LPS‐stimulated TNF‐α secretion in macrophage RAW 264.7 cells. However, mifepristone also increased energy expenditure, as measured by fully automatic metabolic cages, and enhanced expression of thermogenic markers in energy‐combusting brown adipose tissue (BAT) but did not affect inflammation.

Conclusions and Implications

C108297 attenuates obesity by reducing caloric intake and increasing lipolysis and fat oxidation, and in addition attenuates inflammation. These data suggest that selective GR modulation may be a viable strategy for the reduction of diet‐induced obesity and inflammation.

Abbreviations

BAT

brown adipose tissue

FFAs

Free fatty acids

GC

glucocorticoid

GR

glucocorticoid receptor

gWAT

gonadal white adipose tissue

HPA

hypothalamic–pituitary–adrenal

PL

phospholipid

TG

triglycerides

TC

total cholesterol

WAT

white adipose tissue

Tables of Links

TARGETS
Nuclear hormone receptors
Glucocorticoid receptor (GR; NR3C1)

LIGANDS
ACTH	Insulin
Corticosterone	LPS
Dexamethasone	MCP‐1 (CCL2)
Glycerol	Mifepristone (RU486)
IL‐1β	TNF‐α

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (Alexander et al., 2015).

Introduction

Obesity often coincides with chronic low‐grade inflammation, both being risk factors for the development of several diseases including type 2 diabetes and cardiovascular disease (Berg and Scherer, 2005). Compounds that reduce both diet‐induced obesity and inflammation are therefore of potential benefit. Glucocorticoids (GCs) are steroid hormones secreted by the adrenal cortex upon activation of the hypothalamic–pituitary–adrenal (HPA) axis. GCs bind to the intracellular glucocorticoid receptor (GR) and mineralocorticoid receptor, which upon ligand binding regulate gene transcription in a wide variety of tissues (Zhou and Cidlowski, 2005).

GCs can contribute to development of the metabolic syndrome, as evidenced by the side effects of their use as anti‐inflammatory drugs, such as weight gain, insulin resistance, hypertriglyceridaemia, hyperphagia and central obesity (Andrews and Walker, 1999; Macfarlane et al., 2008; Auvinen et al., 2013). Modulating cortisol activity with the classical GR antagonist RU486/mifepristone reduces body weight gain and fat mass in animals on a high‐fat diet and in Cushing's syndrome patients (Nieman et al., 1985; Asagami et al., 2011). Additionally, mifepristone may improve metabolic health by increasing the activity of energy‐combusting brown adipose tissue (BAT), in which the GR is also highly expressed, resulting in enhanced energy expenditure (Rodriguez and Palou, 2004; van den Beukel et al., 2014).

Although mifepristone improves the metabolic profile, as a GR antagonist it does not possess anti‐inflammatory properties. In addition, mifepristone also binds with high affinity to the progesterone receptor and weakly to the androgen receptor. Therefore, compounds acting at the GR that can achieve metabolic improvements (representing GR antagonism) but also inhibit inflammation (representing GR agonism) are of potential benefit.

Previous studies primarily aiming at potent anti‐inflammatory efficacy showed potential for these so‐called dissociated compounds or selective GR modulators (e.g. De Bosscher et al., 2010; Vandevyver et al., 2013); however, the potential beneficial effects for diet‐induced obesity have not yet been explored.

The aim of this study was to test the effects on metabolism and inflammation of the selective GR modulator C108297, a compound that combines agonistic and antagonistic effects and has no affinity for other steroid receptors (Clark et al., 2008; Zalachoras et al., 2013). This compound was previously shown to potently reduce weight gain in a diet‐induced obesity model (Asagami et al., 2011), but the mechanisms, as well as tissue‐specificity and effects on inflammation have not been investigated yet. Therefore, in this study, we investigated the effects of C108297 and mifepristone, compared with a vehicle control group, on metabolic parameters and inflammatory profile in a diet‐induced obesity model.

Methods

Animals and diet

All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals (Kilkenny et al., 2010; McGrath and Lilley, 2015). Ten‐week‐old male C57Bl/6 J mice (Jackson Laboratory, Bar Harbor, ME, USA) were housed in conventional cages with a 12:12 h light–dark cycle with free access to food and water. Mice were stratified into three groups based on body weight and plasma triglyceride levels to assure no differences in body weight and plasma triglyceride levels at the start of the study. To do so, before the start of the experiment, the mice were sorted based on their body weight and subsequently plasma triglyceride levels. Based on these two parameters, the mice were divided over the three different diet groups so that in the end, both body weight and plasma triglyceride levels were as close to equal between the diet groups. Mice were fed a high‐fat fructose diet consisting of high‐fat diet pellets (60% fat, 20% carbohydrates, 20% protein, Research Diets D12492) and 10% fructose water for 4 weeks while being housed at an ambient temperature of 21°C. The compounds C108297 (80 mg⋅kg⁻¹⋅day⁻¹), mifepristone (60 mg⋅kg⁻¹⋅day⁻¹) [doses were based on a previous study that showed weight‐reducing effects (Asagami et al., 2011)] or vehicle were mixed with the food. After 4 weeks of treatment, mice were killed by cervical dislocation and perfused for 5 min with ice‐cold PBS. Organs were harvested for further analysis. The experimenters were blind to the animal treatment. All group sizes started off with equal numbers (n = 10). Due to technical failures of the metabolic cages and loss of some samples during RNA isolation, group sizes ended up smaller in some cases. All animal experiments were approved by the institutional ethics committee on animal care and experimentation at Leiden University Medical Centre.

Indirect calorimetry

Indirect calorimetry was performed in fully automatic metabolic cages (LabMaster System, TSE Systems, Bad Homburg, Germany) in both the first and third week of treatment. After 20 h acclimatization, oxygen consumption (V˙O₂), carbon dioxide production (V˙CO₂) and caloric intake were measured for 5 consecutive days. Carbohydrate and fat oxidation rates were calculated from V˙O₂ and V˙CO₂ as described previously (Van Klinken et al., 2012). Total energy expenditure was calculated from VO ₂ and VCO ₂. Physical activity was measured using infrared sensor frames.

RNA isolation and Q‐RT‐PCR analysis

Total RNA was isolated with the Nucleospin^® RNA II Kit (Macherey‐Nagel) according to the manufacturer's instructions. One microgram of total RNA was reverse‐transcribed with iScript cDNA synthesis kit (Bio‐Rad), and the obtained cDNA was purified with Nucleospin Extract II kit (Macherey‐Nagel). Real‐time PCR was carried out on the IQ5 PCR machine (Bio‐Rad) using the Sensimix SYBR Green RT‐PCR mix (Quantace). Melt curve analysis was included to provide assurance that a single PCR product was formed. Expression levels were normalized using β2‐microglobulin and 36b4 as housekeeping genes. The mean of the high‐fat diet control group was then calculated, and all the individual test values were calculated as fold of the mean of the control group, and appropriate statistical analysis was conducted on these normalized values (Curtis et al., 2015). Primer sequences are listed in Supporting Information Table S1.

Plasma parameters

After 4 weeks of treatment, blood was collected from the tail vein of mice into chilled capillaries that were coated with paraoxon (Sigma‐Aldrich) to prevent ongoing lipolysis (Zambon et al., 1993). Capillaries were placed on ice and centrifuged, and plasma was assayed for glucose, insulin, triglycerides (TG), total cholesterol (TC), and phospholipid (PL) content using commercially available enzymatic kits for glucose, TG and TC (Roche Diagnostics) and PL (Instruchemie, Delfzijl, The Netherlands). Free fatty acids (FFAs) were measured using the NEFA C kit (Wako Diagnostics; Instruchemie) and insulin using the ultrasensitive mouse insulin ELISA kit (Chrystal Chem). The distribution of cholesterol over plasma lipoproteins was determined using fast protein liquid chromatography. Plasma was pooled per group, and 50 μL of each pool was injected onto a Superose 6 PC 3.2/30 column (Akta System, Amersham Pharmacia Biotech, Piscataway, NJ, USA) and eluted at a constant flow rate of 50 μL⋅min⁻¹ in PBS, 1 mM EDTA, pH 7.4. Fractions of 50 μL were collected and assayed for TC as described previously. am (08.30) and pm (17.30) blood samples for the determination of plasma corticosterone were collected at day 14 (am) and day 15 (pm), and all samples were collected within 90 s. Corticosterone levels were determined using ¹²⁵I RIA kit (ImmuChem, MP Biochemicals, Orangeburg, NY, USA).

Determination of adipocyte size

Gonadal fat pads were removed and kept in PBS. The tissues were minced and digested in 0.5 g⋅L⁻¹ collagenase in HEPES buffer (pH 7.4) with 20 g⋅L⁻¹ dialyzed bovine serum albumin (fraction V, Sigma, St Louis, USA) for 1 h at 37°C. The disaggregated adipose tissue was filtered through a nylon mesh with a pore size of 236 μm. For the isolation of mature adipocytes, cells were obtained from the surface of the filtrate and washed several times. The residue was used for stromal vascular fraction (SVF). Using direct microscopy, the diameter of ~100 adipocytes was determined from which mean cell diameter was calculated.

Isolation of stromal vascular fraction and flow cytometry

For the isolation of the SVF, the residue of the adipose tissue filtrate was centrifuged (10 min, 200× g), and the supernatant was discarded. The pellet was treated with erythrocyte lysis buffer, and the cells were subsequently counted using an automated cell counter (TC10, Bio‐Rad). The stromal vascular cells were fixed in 0.5% paraformaldehyde, stored in FACS buffer (PBS, 0.02% sodium azide, 0.5% FCS) in the dark at 4°C until antibody staining and flow cytometry analysis. Cells were stained for 30 min with fluorescently labelled primary antibodies for leukocytes (CD45.2), T cells (CD3), T helper cells (CD4), cytotoxic T cells (CD8a), macrophages (F4/80, CD11B), pro‐inflammatory M1‐macrophages (CD11C) and anti‐inflammatory M2‐macrophages (CD206). SVF was analysed by flow cytometry with LSRII flowcytometer and Diva software.

Adipocyte lipolysis

SVF was isolated from subcutaneous adipose tissue of chow fed mice as described above and the pre‐adipocytes included in the SVF were differentiated into mature adipocytes in vitro as described in Vroegrijk et al. (2013) with the exception that dexamethasone was not added to the differentiation medium. The fully differentiated adipocytes were incubated with C108297 or mifepristone for 15 h, and glycerol concentrations (index of lipolysis) were determined using a commercially available free glycerol kit (Sigma) with the inclusion of the hydrogen peroxide‐sensitive fluorescence dye Amplex Ultra Red.

The antilipolytic effect of insulin was determined in primary mature white adipocytes of gonadal adipose tissue from C108297 or mifepristone‐treated mice by incubating them for 2 h at 37°C in DMEM/F12 with 2% BSA in combination with or without 1 mM 8‐bromo cAMP (8b‐cAMP; Sigma) and/or insulin (10⁻⁹ M). Glycerol concentrations were determined as described above.

Liver lipid extraction

Lipids were extracted from livers with a modified protocol from Bligh and Dyer (Bligh and Dyer, 1959). Small liver samples (∼50 mg) were homogenized in 10 μL of ice‐cold methanol per milligram of tissue. Lipids were extracted into an organic phase by addition of 1800 μL of CH₃OH:CHCl₃ (3:1, v/v) to 45 μL of homogenate and subsequent centrifugation. The lower, organic phase was dried and suspended in 2% Triton X‐100. Hepatic triglycerides (TG) and total cholesterol (TC) concentrations were measured using commercial kits, as explained below. Liver lipids were reported as mg^‐1 protein, as determined using the BCA protein assay kit (Thermo Scientific, Rockford, IL, USA).

LPS‐induced TNF‐α release in RAW 264.7 cells

RAW 264.7 cells, a murine macrophage cell line, were seeded into 24‐well plates (1 × 10⁶ cells per well) and cultured overnight at 37°C in DMEM, 10% FBS, 1% pen/strep. Cells were washed with DMEM and incubated with LPS from Escherichia coli O55:B5 (Sigma‐Aldrich; L‐2637) (10 ng⋅mL⁻¹) that was pre‐incubated (120 min 37°C) with or without various concentrations (10⁻⁶ to 10⁻⁹ M) of C108297 or dexamethasone (10⁻⁶ M) in DMEM supplemented with 0.01% human serum albumin (4 h at 37°C). The medium was collected, and TNF‐α was determined in the medium using the commercially available mouse TNF‐α‐specific OptEIA™ ELISA (BD Biosciences Pharmingen) according to the manufacturer's instructions.

Immunohistochemistry for CD45

Brains were stored in 4% paraformaldehyde at 4°C for 2 days, cryoprotected in 15% (1 day) and 30% (1 day) sucrose solution and stored at −80°C. Coronal slices of 30 μm were cut in a cryostat and collected and stored until processing in cryoprotectant (30% glycerol, 30% ethylene glycol, 40% 0.1 M PBS). Free‐floating sections were washed in PBS at least five times for a total of 60 min and were incubated in 0.5% Triton X‐100 for 30 min at room temperature. Sections were incubated in 2% donkey serum in PBS for 45 min and then overnight incubated at room temperature with rat anti‐CD45 (1:1000; AbD Serotec, Kidlington, UK) primary antibody diluted in 2% serum. The next day, slices were washed in PBS several times for a total of 60 min and incubated with donkey anti‐rat secondary antibody (Alexa 488, Abcam, Cambridge, UK) for 3 h at room temperature. After four washes in PBS for a total of 30 min at RT, sections were mounted on glass slides and embedded in mounting medium containing DAPI (Vectashield, Vector labs, Burlingame, CA, USA) and visualized under a Nikon Eclipse E800 fluorescence microscope.

Statistical analysis

The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015). All data are expressed as mean ± SEM. Data were analysed with one‐way ANOVA and, when significant, with post hoc Tukey test. Significance is set at P < 0.05.

Results

Both C108297 and mifepristone reduce body weight gain with differential effects on food intake and energy expenditure

Four weeks of 60% high‐fat diet feeding increased body weight in control mice, an effect attenuated by treatment with C108297 and mifepristone (Figure 1A). This was due to the fact that C108297 and, to a lesser extent, mifepristone reduced fat mass as well as lean mass as compared with control mice (Figure 1B and C). Reduced weight gain may be caused by reduced caloric intake, increased energy expenditure or both. We found that C108297 decreased total caloric intake (Supporting Information Figure S1A), which was due to a reduction in food intake (Figure 1D) and fructose intake (Figure 1E). Moreover, C108297 changed the timing of food intake and caused a slight but significant shift towards dark phase food intake compared with the control group (Figure 1F). Mifepristone showed a trend towards a reduction in food intake (Figure 1D) and total caloric intake (Supporting Information Figure S1A), but did not reduce fructose intake (Figure 1E) or timing of food intake (Figure 1F).

Figure 1.

C108297 decreases caloric intake, whereas mifepristone increases energy expenditure. Ten‐week‐old male C57Bl/6 J mice were fed a high‐fat fructose diet while being treated with C108297, a full GR antagonist (mifepristone (MIF)) or vehicle for 4 weeks. (A) Body weight. (B) Fat mass and (C) lean mass determined with EchoMRI™ after 4 weeks of diet treatment. (D) 60% high‐fat diet intake, (E) 10% fructose intake, (F) percentage of food intake during the dark phase, (G) fat oxidation, (H) carbohydrate oxidation and (I) energy expenditure were measured in automated metabolic cages in the first week of treatment. Values are means ± SEM of nine to ten mice per group. *P < 0.05 compared with control. Details on statistics are given in the Supporting Information (Table S1).

With respect to energy expenditure, C108297 increased fat oxidation (Figure 1G) and decreased carbohydrate oxidation (Figure 1H), whereas mifepristone increased light phase energy expenditure (Supporting Information Figure S1B) and tended to increase total energy expenditure (Figure 1I).

Thus, C108297 and mifepristone both reduce body weight gain. C108297 reduces caloric intake and enhances fat oxidation, while mifepristone predominantly increases energy expenditure. Details on statistics presented in Figure 1 are provided in the Supporting Information (Table S1).

Mifepristone, but not C108297, increases BAT activity

An increase in fat oxidation as induced by C108297 and increased energy expenditure by mifepristone may point to enhanced BAT activation (Berbee et al., 2015). Moreover, GR antagonism is associated with increased BAT activity (Rodriguez and Palou, 2004; van den Beukel et al., 2014). Therefore, we next investigated whether increased BAT activity may be involved in the reduced body weight causes by C108297 and mifepristone. C108297 did not affect BAT weight (Figure 2A), intracellular lipid droplet size (Figure 2B) nor expression of BAT‐specific markers (Figure 2C–F). In contrast, although mifepristone did not affect BAT weight (Figure 2A), it markedly decreased lipid droplet size (Figure 2B) and increased gene expression of the thermogenic marker Ucp1 (Figure 2C) and the mitochondrial biogenesis marker Pgc‐1α (Figure 2D). Moreover, mifepristone increased gene expression of Lpl (Figure 2E) and Cd36 (Figure 2F), both of which are involved in uptake of triglyceride‐derived fatty acids by the tissue. Thus, these data collectively show that C108297 does not affect BAT metabolic activity while mifepristone increases its activity.

Figure 2.

Mifepristone reduces intracellular lipid content and increases BAT thermogenic markers. (A) BAT weight and (B) BAT histology [haematoxylin and eosin (H&E) staining]. BAT gene expression of (C) Ucp‐1, (D) Pgc‐1α, (E) Lpl and (F) Cd36 measured by Q‐RT‐PCR. Values are means ± SEM of eight to ten mice per group. *P < 0.05 compared with control. Details on statistics are given in the Supporting Information (Table S2).

C108297 reduces adipose tissue volume by increasing lipolysis rate and lowers plasma glucose levels

Because glucocorticoids have substantial effects on white adipose tissue (WAT) lipolysis, we reasoned that the reduced fat mass following C108297 treatment may not solely be due to lower caloric intake but could also be caused by intrinsic effects of C108297 on lipolysis rate. We, therefore, measured plasma FFA levels and investigated physiology and lipolysis markers in gonadal WAT (gWAT). C108297 increased plasma FFA levels (Figure 3A) and decreased gWAT weight (Figure 3B), which was due to smaller adipocyte size rather than reduced adipocyte number (Figure 3C–E). Accordingly, C108297 enhanced mRNA levels of Atgl, the rate‐limiting enzyme for intracellular lipolysis, in gWAT (Figure 3F). This is likely mediated via induction of the transcription factor FoxO1 (Wang et al., 2012) (Figure 3G), which is a GR target (Waddell et al., 2008; Lee et al., 2011) and positively regulates Atgl gene expression levels (Chakrabarti and Kandror, 2009). All these effects were absent in mifepristone‐treated mice. To study whether C108297 has direct effects on WAT lipolysis, we performed studies with in vitro differentiated white adipocytes derived from subcutaneous WAT of C57Bl6/J mice fed a chow diet and incubated these with C108297, mifepristone or vehicle for 15 h. Indeed, we showed that solely C108297 increased glycerol release, indicating a direct effect of C108297 on intracellular lipolysis in vitro (Supporting Information Figure S2A). To determine if the increased lipolysis rate by C108297 was due to reduced insulin sensitivity, adipocytes derived from compound‐treated and control mice were examined ex vivo. Upon treatment with the lipolysis‐inducing agent 8‐bromo‐cAMP, all groups similarly exhibited increased glycerol release (Figure 3H). Co‐treatment with insulin, which has antilipolytic effects on adipocytes, reduced lipolysis in all groups (Figure 3H), although the inhibition was lower by mifepristone (Figure 3I). This indicates that mifepristone reduces insulin responsiveness and excludes insulin resistance as a mechanism for enhanced lipolysis induced by C108297. Thus, C108297 reduced adipose tissue weight and adipocyte size by increased lipolysis rate of WAT. Details on statistics are given in the Supporting Information (Table S3).

Figure 3.

C108297 reduces adipose tissue weight and adipose tissue volume and increases lipolysis rate. (A) Plasma free fatty acids (FFA), (B) gonadal white adipose tissue (gWAT) weight, (C) adipocyte size, (D) number of fat cells per fat pad and (E) gWAT histology. (F) Adipose triglyceride lipase (Atgl) and (G) FoxO1 mRNA expression in gWAT measured by Q‐RT‐PCR. (H) Ex vivo glycerol release (as a measure of lipolysis) in gWAT derived from mice treated with control, mifepristone or C108297. Lipolysis was further stimulated with 8‐bromo‐cAMP (8B) and inhibited by insulin (8B + i). (I) Insulin inhibition of lipolysis in gonadal adipocytes. Values are means ± SEM of seven to nine mice per group. *P < 0.05 compared with control. ^# P < 0.05 compared with vehicle. Details on statistics are given in the Supporting Information (Table S3).

We next assessed the effect of C108297 on plasma glucose and lipid levels. Both C108297 and mifepristone decreased plasma glucose levels and increased plasma cholesterol levels, which was mainly due to increased HDL levels (Supporting Information Figure S2B–D). Plasma insulin, triglycerides and phospholipids were not different between groups (Supporting Information Figure S2E–G). Moreover, the increased plasma FFA levels did not result in ectopic fat accumulation in liver, as liver weight and liver lipid content was not different between groups (Supporting Information Figure S2H–J).

C108297 decreases pro‐inflammatory M1 macrophages, leukocytes and cytokines in gonodal WAT

To investigate if the reduction in fat mass and adipocyte size was associated with lower adipose tissue inflammation, we next determined the inflammatory profile in gWAT by means of flow cytometry. C108297 reduced the number of leukocytes (CD45+ cells) (Supporting Information Figure S3A and B) and the number of macrophages (F4/80 cells) (Figure 4A), but not the percentage of macrophages within leukocytes (Supporting Information Figure S3C). Within the macrophage fraction, C108297 reduced the number of pro‐inflammatory M1 macrophages (Supporting Information Figure S3D) and increased the number of anti‐inflammatory M2 macrophages (Supporting Information Figure S3E), resulting in a markedly lower M1:M2 ratio (Figure 4B). These beneficial effects on inflammation in gWAT were absent in mifepristone‐treated mice. Both C108297 and mifepristone did not alter T‐cell and B‐cell populations (data not shown). The reduced number of pro‐inflammatory macrophages coincided with reduced expression levels of the monocyte chemokine Mcp‐1 and mannose receptor C type 1 (Mrc1), indicating reduced migration of monocytes towards gWAT (Figure 4C and D). Additionally, C108297 (but not mifepristone) reduced the mRNA expression of the pro‐inflammatory cytokines Il1β and Tnfα (Figure 4E and F), and increased the expression of the anti‐inflammatory gene glucocorticoid‐induced leucine zipper (Gilz) (Figure 4G) but not of map kinase phosphatase 1 (Mkp1) and lipocortin1 (Supporting Information Figure S3F and G), indicating that C108297 exerts anti‐inflammatory actions via transrepression of pro‐inflammatory genes, but also via a direct up‐regulation of anti‐inflammatory genes in a selective modulatory manner.

Figure 4.

C108297 decreases macrophages and pro‐inflammatory cytokines in gonadal white adipose tissue and decreases inflammation in vitro. (A) Macrophages, F4/80 cells and (B) ratio of pro‐inflammatory M1 and anti‐inflammatory M2 cells determined with FACS analysis in gonadal white adipose tissue. gWAT gene expression (C) Mcp1, (D) Mrc1, (E) Il1β, (F) Tnfα, (G) Gilz measured by Q‐RT‐PCR. (H) Four‐hour LPS‐stimulated TNF‐α protein expression determined in vitro in macrophage RAW 264.7 cells, pre‐incubated for 2 h with or without various concentrations (10⁻⁶ to 10⁻⁹ M) of C108297, compared with maximal suppressive dose of dexamethasone (10⁻⁶ M; dashed line). Values are means ± SEM of seven to nine mice per group. *P < 0.05 compared with control. Details on statistics are given in the Supporting Information (Table S4).

Given the strong anti‐inflammatory properties of GCs, we next investigated if the anti‐inflammatory effects of C108297 were solely due to reduction in fat mass or whether C108297 also exerts intrinsic anti‐inflammatory properties. Therefore, we tested the efficacy of C108297 in reducing LPS‐stimulated TNF‐α protein expression in vitro in the mouse macrophage cell line RAW 264.7. LPS strongly increased TNF‐α levels (Supporting Information Figure S3H) and C108297 dose dependently reduced LPS‐stimulated TNF‐α levels, with 56% efficacy relative to the full GR agonist dexamethasone (Figure 4H), indicating that C108297 also exerts intrinsic anti‐inflammatory properties.

Given that obesity is also associated with increased inflammatory markers in the hypothalamus (Thaler et al., 2012), we also investigated expression levels of the microglial marker CD45 in the hypothalamus. C108297 reduced CD45 protein expression in the PVN of the hypothalamus (Supporting Information Figure S4B and C), but not arcuate nucleus, whereas mifepristone animals did not show a difference in either hypothalamic region, despite reduced body fat mass.

C108297 predominantly exerts agonism on HPA‐axis parameters

The fact that C108297 has both anti‐obesity and anti‐inflammatory effects suggests that it exerts both agonistic and antagonistic effects. We next assessed the effects of C108297 on the HPA‐axis by measuring corticosterone levels in plasma. Corticosterone is secreted by the adrenals in a diurnal pattern with peak plasma concentrations before the dark phase (i.e. active phase) and trough concentrations before the light phase (i.e. inactive phase). However, high‐fat diet exposure, and subsequently obesity, attenuates the rhythm in corticosterone concentrations (Kohsaka et al., 2007). In this study, we showed that mice on the high‐fat diet control group show equal am (morning) and pm (evening) corticosterone concentrations (Table 1), indicating a disruption of circadian corticosterone rhythm. C108297 reduced morning (am) and evening (pm) corticosterone values, indicating GR agonism on the HPA axis. Mifepristone also reduced AM corticosterone concentrations with a trend for a reduction in PM corticosterone concentrations. ACTH levels were consistently low throughout the circadian cycle. C108297 and mifepristone did not further suppress ACTH levels (Table 1), suggesting that part of the lowered plasma corticosterone levels may also be mediated at the adrenal level (Walker et al., 2015).

Table 1.

Effect of mifepristone and C108297 on thymus and spleen weight, and plasma corticosterone and ACTH levels, compared with vehicle control group

	Control	mifepristone	C108297
Corticosterone AM (nmol⋅L−1)	87.2 ± 11.3	47.2 ± 1.1*	47.6 ± 1.1*
Corticosterone PM (nmol⋅L−1)	106.6 ± 38.1	32.1 ± 11.3	26.3 ± 7.8*
ACTH AM (nmol⋅L−1)	6.2 ± 1.2	5.6 ± 0.6	5.2 ± 0.4
ACTH PM (nmol⋅L−1)	6.3 ± 0.8	4.5 ± 1.6	6.4 ± 1.7
Adrenal weight (mg)	1.34 ± 0.16	0.65 ± 0.13*	0.57 ± 0.15*
Star mRNA (relative to control)	1.00 ± 0.69	0.16 ± 0.09*	0.18 ± 0.10*
Thymus (mg)	46 ± 4	15 ± 2*	9 ± 3*
Spleen (mg)	79 ± 4	72 ± 2	57 ± 2*

Values are mean ± SEM.

^* P < 0.05 compared with control.

We also assessed adrenal weight and the corresponding gene expression levels in the adrenal of steroidogenic acute regulatory protein (StAR), which is the rate‐limiting step in the production of steroid hormones, including corticosterone. Both C108297 and mifepristone reduced adrenal weight and Star mRNA levels. In addition, C108297 markedly reduced thymus weight and spleen weight (Table 1), while mifepristone only reduced thymus weight. Thus, these data indicate that C108297 acts as an agonist on the HPA axis and immune tissues, while mifepristone is only partially agonistic for the GR (Meijer et al., 2005).

Discussion

This study aimed to characterize the metabolic and (anti)‐inflammatory effects of the selective GR modulator C108297 compared with mifepristone and vehicle. In the present study, we showed that both C108297 and, to a smaller extent, mifepristone reduced body weight gain and fat mass upon 4 weeks of HFFD feeding, whereby C108297 predominantly reduced caloric intake and mifepristone predominantly increased energy expenditure. In addition, C108297 increased lipolysis in WAT and FFA levels in plasma resulting in decreased fat cell size and increased fatty acid oxidation. Furthermore, C108297 decreased inflammation, evidenced by a decrease in pro‐inflammatory macrophage infiltration and cytokine expression levels in adipose tissue and decreased microglial cells in the hypothalamus. Importantly, mifepristone did not affect inflammatory markers.

The effects of C108297 and mifepristone on body weight gain observed in this study are consistent with previous studies in mouse models of diet‐induced obesity (Asagami et al., 2011; Beaudry et al., 2014). Metabolic cage data are helpful to determine dominant effects on metabolic and food intake measures. C108297 did not affect energy expenditure, whereas mifepristone significantly increased light phase energy expenditure (with a trend for total energy expenditure) and thermogenic markers in BAT, which adds to previous studies describing inhibitory effects of GCs and stimulatory effects of mifepristone on UCP‐1 expression and BAT thermogenesis (Rodriguez and Palou, 2004; van den Beukel et al., 2014).

C108297 reduced total caloric intake, which suggest antagonistic effects at the GR, which apparently were more potent than those of mifepristone (Dallman et al., 2004). Although mifepristone and C108297 virtually equally decreased food intake, C108297 also reduced fructose intake. Since particularly the combination of fat and sugar has detrimental metabolic effects (van den Heuvel et al., 2014), the simultaneous decrease in both the 60% high‐fat diet and fructose solution may have contributed to improved metabolic health. Moreover, the reduction in fructose solution argues against a taste aversion by C108297. Additionally, direct infusion in the stomach by oral gavage also resulted in reduction in body weight (Beaudry et al., 2014).

Lean control mice predominantly consume their food in the dark (active) phase of the day, with little food intake in the light (inactive) phase. High‐fat diet exposure, and subsequently obesity, results in disruption of this circadian rhythm in food intake (Kohsaka et al., 2007), which is consistent with the high‐fat diet control group in this study that equally consumed their calories in the light and dark phases. C108297 caused a slight, but significant, shift towards dark phase food intake, indicating improved circadian food intake pattern. Previous studies showed that disturbance of the circadian rhythm in food intake and particularly consumption of food during the light phase (i.e. resting phase) contributes to the development of obesity (Oosterman et al., 2014). Therefore, the (modestly) improved circadian food intake in C108297 together with decreased caloric intake has likely contributed to the reduction of body weight gain.

Another component that likely contributes to the reduced diet‐induced obesity by C108297 is increased fat oxidation. Due to enhanced lipolysis in WAT, plasma FFA levels were increased, which may have served as a source for the increased fat oxidation (Jung et al., 1980, Schiffelers et al., 2001). Importantly, the increased plasma FFA levels did not result in ectopic fat storage in the liver as liver weight and liver lipid content in mice treated with C108297 was not increased. The increased fat oxidation did not lead to increased energy expenditure, likely due to compensatory decreased carbohydrate oxidation and reduced diet‐induced thermogenesis (as a result of decreased food intake). Increased plasma insulin levels inhibit FFA release from adipose tissue, and increased FFA oxidation has been associated with insulin resistance (Guillaume‐Gentil et al., 1993). However, we showed that plasma insulin levels were not affected and that insulin reduced lipolysis in adipocytes derived from C108297‐treated mice. Therefore, the increased lipolysis and plasma FFA levels are not due to reduced insulin sensitivity. Likely, C108297 exerts direct agonistic effects on the GR in WAT leading to increased lipolysis and plasma FFA levels, thereby increasing fatty acid oxidation and reducing diet‐induced obesity.

We showed that C108297 reduced leukocyte and pro‐inflammatory macrophage infiltration in gonadal WAT. This is an important finding given that mifepristone, and GR antagonism in general is not associated with reduction of inflammation, despite a reduction in adiposity. The anti‐inflammatory properties of C108297 were further supported by the suppression of mRNA expression of the cytokines Il1β and Tnfα, indicating agonistic GR effects of C108297 in WAT. Diet‐induced obesity causes a shift in adipose tissue macrophages from an M2 polarized state to an M1 pro‐inflammatory state, and the number of M1 macrophages and the M1‐to‐M2 ratio are related to the development of insulin resistance (Fujisaka et al., 2009). We here showed that C108297 increased M2 and decreased M1 macrophages, leading to a favourable M1:M2 ratio. As previous studies reported improved glucose metabolism in mice treated with C108297 (Beaudry et al., 2014), the anti‐inflammatory effects of C108297 may be involved in reduction of the obesity‐induced inflammatory response and improvement of insulin sensitivity, which needs to be further investigated in future experiments.

C108297 also reaches the brain where it displays selective GR modulation on GR target genes and the stress system (Zalachoras et al., 2013). The reduction of CD45 by C108297 may point towards partial GR agonism and potential additional effects in reducing hypothalamic inflammation, although the effects of C108297 on (diet‐induced) neurological inflammatory processes remains to be further explored in future studies.

Conceptually, combining agonistic and antagonistic effects of GR may provide protection against metabolic disease. We previously showed that selective GR modulation by C108297 is based on a unique pattern of C108297‐induced recruitment of transcriptional co‐regulators by GR (Zalachoras et al., 2013). The efficacy in reducing inflammation that we observed suggests that C108297 also allows classical transrepression via interactions with pro‐inflammatory transcription factors (De Bosscher et al., 2010). The compound at this dose clearly had agonistic effects in negative feedback on the HPA axis, but so did mifepristone (which in some contexts has been shown to be a partial agonist) (Havel et al., 1996; Meijer et al., 2005). Effects on lipolysis also seem indicative of GR agonism (Macfarlane et al., 2008). In contrast, the effects of C108297 on food intake may be more indicative of antagonism of the GR (Tempel et al., 1993), analogous to some of the previously observed effects of this compound in the brain (Zalachoras et al., 2013).

In conclusion, C108297 not only attenuates obesity by reducing caloric intake and increasing lipolysis and fat oxidation but also attenuates inflammation in white adipose tissue. This suggests that C108297 or similar selective GR modulators may be beneficial in reducing both peripheral and central inflammation and reducing body weight gain. While the ideal mix of agonism and antagonism of GR targets remains unknown, our data show that novel ligands for classical receptors may bear promise for the treatment of obesity‐related disorders such as metabolic syndrome.

Author contributions

M.B., I.v.H. and E.P.v.d.P. performed the research. O.M., A.P. and P.R. designed the research study. H.H., J.K.B. and K.W.v.D. contributed essential reagents or tools. J.v.d.H., M.B., L.v.B. and V.v.H. analysed the data. J.v.d.H., M.B. and O.M. wrote the paper.

Conflict of interest

J.K.B. and H.H. are employees of Corcept Therapeutics, which develops glucocorticoid receptor ligands for clinical use. Corcept Therapeutics has provided compounds and financed part of the experiments.

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organisations engaged with supporting research.

Supporting information

Figure S1 10‐week old male C57Bl/6 J mice were fed a high‐fat fructose diet while being treated with C108297, a full GR antagonist (mifepristone (MIF)) or vehicle for 4 weeks. (A) Total caloric intake, (B) light phase and (C) dark phase energy expenditure, and (D) activity, measured in fully automated metabolic cages during first week of treatment. Values are means ± SEM of eight to ten animals per group. *P < 0.05 compared to control.

Figure S2 (A) In vitro stimulated glycerol release (as a measure of lipolysis) in chow‐derived in vitro differentiated subcutaneous white adipocytes stimulated with C108297, a full GR antagonist (mifepristone (MIF) or vehicle. (B‐J) 10‐week old male C57Bl/6 J mice were fed a high fat diet while being treated with C108297, a full GR antagonist (mifepristone (MIF) or vehicle for 4 weeks After the treatment period, blood samples from 4 h‐fasted mice were collected by tail vein bleeding and (B) glucose, (C,D) total cholesterol, (E) insulin, (F) triglycerides and (G) phospholipids determined. (H) Liver weight, (I) liver triglyceride content and (J) liver cholesterol content. Values are means ± SEM. *P < 0.05 compared to control.

Figure S3 10‐week old male C57Bl/6 J mice were fed a high fat diet while being treated with C108297, a full GR antagonist (mifepristone (MIF) or vehicle for 4 weeks. (A,B) Leukocytes (CD45+ cells), (C) F4/80 macrophages, (D) M1 macrophages, CD206‐CD11c + cells as % of F4/80 cells, and (E) M2 macrophages, CD206 + CD11c‐ cells as % of F4/80 cells, determined with FACS analysis in gonadal white adipose tissue. (F) Mkp1 mRNA and (G) Lipocortin1 mRNA measured by Q‐RT‐PCR. (H) Four hour with or without LPS‐stimulated TNFα protein expression determined in vitro in macrophage RAW 264.7 cells. Values are means ± SEM. *P < 0.05 compared to control.

Figure S4 10‐week old male C57Bl/6 J mice were fed a high fat diet while being treated with C108297, a full GR antagonist (mifepristone (MIF) or vehicle for 4 weeks (A) Immunohistochemical staining of microglial marker CD45 in the paraventricular nucleus of the hypothalamus in mice on a high‐fat fructose diet treated with vehicle, mifepristone or C108297. (B) Representative images of CD45 and the nuclear marker DAPI. 3 V, third ventricle. Values are means ± SEM. *P < 0.05 compared to control.

Table S1 Overview primers used for RT‐‐qPCR.

Table S2 Statistics for Figure 1.

Table S3 Statistics for Figure 2.

Table S4 Statistics for Figure 3.

Table S5 Statistics for Figure 4.

Table S6 Statistics for Table 1.

Table S7 Statistics for Supplemental figure S1.

Table S8 Statistics for Supplemental figure S2.

Table S9 Statistics for Supplemental figure S3 and S4.

Supporting info item

Supporting info item

Supporting info item

Supporting info item

Supporting info item

van den Heuvel, J. K. , Boon, M. R. , van Hengel, I. , Peschier‐van der Put, E. , van Beek, L. , van Harmelen, V. , van Dijk, K. W. , Pereira, A. M. , Hunt, H. , Belanoff, J. K. , Rensen, P. C. N. , and Meijer, O. C. (2016) Identification of a selective glucocorticoid receptor modulator that prevents both diet‐induced obesity and inflammation. British Journal of Pharmacology, 173: 1793–1804. doi: 10.1111/bph.13477.