Ghrelin Receptor Signaling Is Not Required for Glucocorticoid-Induced Obesity in Male Mice

Abstract

Chronically elevated levels of glucocorticoids increase food intake, weight gain, and adiposity. Similarly, ghrelin, a gut-secreted hormone, is also associated with weight gain, adiposity, and increased feeding. Here we sought to determine if corticosterone-induced metabolic and behavioral changes require functional ghrelin receptors (GHSR). To do this, we treated male C57BL mice with chronic corticosterone (CORT) mixed in their drinking water for 28 days. Half of these mice received the GHSR antagonist JMV2959 via osmotic minipumps while treated with CORT. In a second experiment, we gave the same CORT protocol to mice with a targeted mutation to the GHSR or their wild-type littermates. As expected, CORT treatment increased food intake, weight gain, and adiposity, but contrary to expectations, mice treated with a GHSR receptor antagonist or GHSR knockout (KO) mice did not show attenuated food intake, weight gain, or adiposity in response to CORT. Similarly, the effects of CORT on the liver were the same or more pronounced in GHSR antagonist-treated and GHSR KO mice. Treatment with JMV2959 did attenuate the effects of chronic CORT on glycemic regulation as determined by the glucose tolerance test. Finally, disruption of GHSR signaling resulted in behavioral responses associated with social withdrawal, potentially due to neuroprotective effects of GHSR activation. In all, we propose that blocking GHSR signaling helps to moderate glucose concentrations when CORT levels are high, but blocking GHSR signaling does not prevent increased food intake, weight gain, or increased adiposity produced by chronic CORT.

Glucocorticoids, in particular, cortisol in primates and corticosterone (CORT) in rodents, are steroid hormones associated with the stress response to metabolic and psychological stressors. Acutely, these hormones are essential to meet the challenges associated with these stressors. However, chronic exposure to glucocorticoids, as seen in patients with endocrine disorders such as Cushing’s syndrome or in patients that require chronic glucocorticoid treatment for the treatment for autoimmune disorders can result in increased weight gain, adiposity, insulin resistance, renal and cardiovascular disease, and metabolic syndrome (1–4). This is due to the multiple metabolic and behavioral effects of these steroid hormones. Indeed, rodents treated chronically with CORT show increased food intake, weight gain, and fat accumulation (5–7). The accumulation of fat is independent of caloric intake, as it is still seen in animals receiving chronic CORT but that are pair fed to control animals (6,8). Nevertheless, increased caloric intake also contributes and may have an impact on the development of obesity given that chronic CORT enhances preferences for calorie-dense foods (9–12). Additionally, glucocorticoids have known effects on glucose metabolism, with sustained glucocorticoid exposure resulting in hyperglycemia, impaired insulin sensitivity, and impaired pancreatic β-cell function (13–17). Given that under normal conditions glucocorticoids are hormones that are intimately associated with the stress response (17), these data suggest that glucocorticoids play a key role in the development of obesity in response to chronic stressors.

Ghrelin is an orexigenic peptide secreted by the stomach (18) and is also associated with metabolic changes produced by stress (19–21). Acutely, the active form of ghrelin (acyl-ghrelin) stimulates food intake and increases food-seeking behaviors, effects that are blocked by ghrelin receptor antagonists (22–26). Ghrelin concentrations vary in relation to energetic status, with higher ghrelin concentrations found in animals in a state of negative energy balance and, conversely, lower levels when animals are in a positive energy state (27). Chronically elevated levels of ghrelin result in increased weight gain in association with changes in substrate utilization favoring carbohydrate over fat metabolism, ultimately leading to the accumulation of fat in adipose tissues (22). Furthermore, these metabolic effects are initiated following ghrelin-mediated activation of growth hormone secretagogue receptors (GHSR) in cells within the hypothalamic arcuate nucleus. Within the arcuate nucleus, stimulation of GHSRs increases the activity of cells expressing the agouti-related peptide and neuropeptide Y, bot peptides that promote increased caloric intake, reduced energy expenditure, and increased adiposity (22,28).

While ghrelin is commonly studied within the context of feeding, recent data point to ghrelin as an important contributor to the metabolic adaptations required to meet the demands of environmental stressors. For example, acute food shortages and prolonged food scarcity both lead to increased plasma ghrelin concentrations (29–31). Additionally, other types of stressors including psychosocial stressors stimulate ghrelin secretion and have been associated with metabolic and behavioral changes that promote the accumulation of body fat (21,32–35). For instance, mice exposed to chronic social defeat show increased caloric intake, weight gain, increased carbohydrate utilization, fat accumulation, hyperleptinemia, and hyperinsulinemia, changes that are not seen in similarly stressed GHSR knock-out (KO) mice or in mice receiving infusions of a ghrelin receptor antagonist (21,34). Thus, elevated plasma levels of ghrelin during chronic stress seem to produce metabolic perturbations associated with obesity and insulin resistance that are similar to observed in animals receiving chronic CORT treatment.

Given these data, we hypothesized that ghrelin mediates the effects of chronic CORT on metabolism to ultimately produce excess body weight and body fat. We tested this hypothesis by exposing GHSR KO mice or mice treated with a GHSR antagonist to a chronic CORT treatment and comparing their body weight to wild type (WT) littermates or saline-treated control animals, expecting that the GHSR mutation or antagonism would attenuate the metabolic effects of the chronic CORT treatment. Our results demonstrate that, in contrast to our prediction, pharmacological or genetic ablation of the GHSR produces increased vulnerability to chronic CORT challenge.

Materials and Methods

Animals and housing

All procedures were approved by the Carleton University Animal Care Committee and followed the guidelines of the Canadian Council for Animal Care. Male mice were used in the current project. In experiment 1, we used male C57BL mice (n = 32) purchased from Charles River Farms and weighing 20 to 25 grams at the onset of the study. In experiment 2, we used 2- to 4-month-old male GHSR KO (n = 14) and WT littermates (n = 13) weighing 25 to 30 grams. These mice were derived from a stock originally generated at Regeneron Pharmaceuticals Inc. (Tarrytown, NY, US) and bred at the Carleton University Department of Neuroscience animal facility. The line was backcrossed onto a C57BL/6 mouse strain from Jackson Laboratories (Bar Harbor, MA, US). All mice were housed in shoebox cages with enrichment (chew sticks and wood blocks, nestlets, and nesting materials) and with free access to regular laboratory chow and water and using a 12-hour light/dark cycle (lights off at 8:00 pm) in a temperature controlled room (22^oC) for the duration of the study.

Corticosterone treatment

In the two experiments conducted, mice were single housed, and their food intake fluid intake and body weights were recorded daily throughout the study. Food and weight measures were also used to calculate feeding efficiency, defined as the amount of weight gained in milligrams per kilocalorie consumed. After a 2-week baseline, drinking water was replaced with a solution containing free corticosterone (100 μg/mL; Sigma, St. Louis, MO) or vehicle (1% ethanol [EtOH]). CORT was dissolved in 100% EtOH and then diluted to the final concentration of 100 μg/mL. Solutions were replaced daily at the time of weighing. All data recording occurred every day between 8:00 and 10:00 am. This procedure is effective in producing CORT-induced obesity in mice (6,36).

Procedure

Experiment 1: Effects of chronic CORT treatment on body weight and food intake in mice treated peripherally with the GHSR antagonist JMV2959.

Following a 2-week baseline, mice were weight matched and assigned to groups that received CORT or 1% EtOH and were treated with either saline or the GHSR antagonist JMV2959 delivered chronically through an osmotic minipump (Model 1004, Alzet) at a dose of 100 µg/day for 28 days. The minipumps were primed inside a vacutaner filled with sterile saline and heated to a temperature of 36°C for 12 hours before being implanted subcutaneously under anesthesia (5% isofluorane mixed with oxygen). The day in which the minipumps were implanted was considered day 1 of the experimental phase and also coincided with the onset of CORT or 1% EtOH treatment. Once the animal was anaesthetized, the intrascapular space was shaved and cleaned with iodine. A 5mm incision was made with small sharp scissors, and a small space was created under the skin using forceps to allow for the insertion of the minipump. Once the minipump was inserted, the wound was sutured, and the animal was given an injection of meloxicam (1 mg/kg) every 12 hours for 36 hours after the surgery to control postoperative pain. Experimental groups were as follows: (i) vehicle/EtOH, (ii) control/CORT, (iii) JMV2959/ EtOH, and (iv) JMV2959/CORT (n = 8 mice per group). A glucose tolerance test (see following discussion) was conducted on all animals on day 18 of treatment. On days 20 and 22 of treatment, mice were tested for the presence of changes in socioemotional behaviors and stress-coping responses using the open field test, social interaction test, and the forced swim test as described in the following text. The open field test and social interaction test were conducted on day 20 (social interaction test was conducted immediately after the open field test), and the forced swim test was conducted on day 22 of treatment. Finally, starting at 8:00 am on the morning of day 28 of CORT treatment, animals were killed by decapitation. Blood glucose was measured at the time of decapitation with a Contour blood glucose meter. Trunk blood was then collected in BD Vacutainer K2 ethylenediaminetetraacetic acid-coated glass tubes (BD Franklin Lakes NJ, US), placed on ice and centrifuged at 3500 rpm for 15 minutes at 4^oC. Plasma was removed and stored at –80^oC for later analyses. The brain and samples from the liver and visceral fat were collected, flash frozen, and stored at –80^oC.

Experiment 2: Effects of chronic CORT treatment on body weight and food intake in GHSR KO and WT mice.

Chronic CORT was administered to adrenally intact GHSR WT and KO male mice (n = 27) weighing between 25 and 30 grams at the onset of the study. At the end of the baseline period, GHSR KO and WT mice were weight matched and assigned to groups receiving CORT or vehicle to obtain 4 experimental groups (n = 8 mice per group): (i) WT EtOH, (ii) WT CORT, (iii) KO EtOH, and (iv) KO CORT. The CORT treatment lasted 28 days. On day 18 of treatment, glucose tolerance tests were conducted in all mice (see the following text for details). All mice were killed by decapitation starting at 8:00 am on the morning of day 28 of CORT administration, and tissues collected as described in Experiment 1.

Behavioral measures

In addition to metabolic dysregulation, chronic CORT treatment has been associated with behavioral changes that include decreased exploration, increase immobility, and social withdrawal (37,38). To determine if these effects were changed by a reduction in GHSR signaling, we tested mice from Experiment 1 on a battery of behavioral tests described as follows.

Open field test.

Mice were taken from their home cage and brought to a testing room where they were placed in an arena measuring 60 cm × 60 cm in width and 90 cm in height. The arena was subdivided into 2 target regions: the edge region of the arena and the center region of the arena. Animals were filmed for the next 7 minutes. After this, the mice were brought back to their home cage for an additional 10 minutes before being tested again for the social interaction test. Two independent investigators recorded the time the mice spent in the center of the arena versus the edges of the arena in that latter 5 minutes of the test session.

Social Interaction test.

Mice were placed in the same arena used for the open field test, this time with a transparent plastic cylinder (5 cm diameter). Experimental mice were allowed to habituate to the box for 5 minutes. At the end of this habituation period, an unfamiliar mouse was placed in the plastic cylinder, and experimental mice were placed on a corner of the testing box. The cylinder was perforated to allow for olfactory information to be detected and was secured at the center of the open field box to keep the unfamiliar stimulus mouse from getting out of the cylinder. Experimental mice were recorded for the next 5 minutes, and the videos were later scored to determine how long it took experimental mice to approach the unfamiliar mouse and how much time they spent sniffing the cylinder containing the unfamiliar mouse. After the videos were recorded, all mice were taken out of the box and returned to their cage in the colony room.

Forced swim test.

Mice were taken from their home room into a testing room where they were placed into a large plexiglass cylinder filled with water at a temperature of 25°C. A camera was used to record the behavior of these mice during the 5-minute test period. These videos were later used to determine how long the animals spent immobile in the water.

These behavioral tests were conducted between 8:00 and 10:00 am. Additionally, both of the behavioral tests were scored by 2 observers blind to the treatment conditions.

Glucose tolerance test

On day 17 of CORT or 1% EtOH treatment, all mice were fasted overnight and brought to a testing room on the morning of day 18 of treatment to do a glucose tolerance test. This time point was chosen because previous work shows that chronic CORT using the same dose used in the current study requires more than 15 days of treatment to influence glucose tolerance, yet these effects are certainly evident by day 28 of treatment (6). This same paper also shows that mice treated with chronic CORT reach asymptote in their weight gain between days 18 and 20 of treatment, indicating this as a potential time point in which CORT treated mice may begin showing alterations in glucose clearance (6). On the test day, mice were restrained gently and injected with an intraperitoneal bolus injection of a 20% glucose solution mixed in sterile isotonic 0.9% saline at a dose of 2 gm/kg of body weight. To determine glucose concentrations in whole blood, we swabbed the tail with alcohol, and we made a small incision on the tip using a 22 G needle to collect a 5 to 10 μL droplet of blood that was sufficient to monitor blood glucose using a Contour glucose meter. Blood glucose levels were monitored immediately after (Time 0) and at 15, 30, 60, and 120 minutes following glucose injection.

Plasma hormone analyses

Plasma insulin (39) and leptin (40) levels were measured using commercially available Multiplex Biomarker Immunoassays for Luminex xMAP technology (Millipore, MA, US; catalogue no. MMHMAG-44K) with sensitivity ranging from 68.5 to 50,000 pg/mL. All samples were run in duplicate, with results reported in pg/mL and had an inter-assay coefficient of variability less than 10%. Hormonal assays on plasma were conducted according to the specifications provided by the assay manufacturers.

Fat and liver analyses

Perigonadal, retroperitoneal, and subcutaneous white adipose tissue (WAT) and brown adipose tissue (BAT) were dissected and weighed for every mouse. Perigonadal WAT reflected the weight of WAT collected from the peritoneal cavity and epididymal adipose pad. Retroperitoneal fat reflected the total weight of WAT from the retroperitoneal space and the WAT surrounding the kidneys. Subcutaneous fat reflected the weight of inguinal WAT depots. The weight of BAT was obtained from the depot located in the intrascapular space.

Livers from all mice were rapidly extracted upon decapitation and flash frozen. Triglyceride stores were analyzed with a standard enzymatic hydrolysis quantification kit (Abnova, Taipei, Taiwan) as directed by the manufacturer’s instructions. All samples were analyzed in duplicate. The triglyceride assay produced an intra-assay coefficient of variation of less than 12% with a lower limit of detectability of 2 nmol/well. A small portion of the liver of each mouse was obtained from the original sample, weighed, and fixed in 4% paraformaldehyde for 48 hours. These samples were then cryoprotected in 30% sucrose solution overnight and sliced on a cryostat to obtain 10 µm thin sections that were thaw mounted on gelatin-coated slides. Tissue sections were stained using a standard hematoxylin and eosin stain to observe liver pathology. A trained lab technician blind to study conditions analyzed the liver sections to determine the degree of liver pathology. Scoring parameters were the NAFLD scoring system developed by Liang et al (41), due to its specificity to C57/BL6 mouse tissue, reproducibility, and correlation to human clinical models. In addition, hepatic sclerosis was scored using the system outlined by Ferrell (42), looking at potential ischemic damage as reflected in fibrosis. Macrosteatosis and microsteatosis were quantified using ImageJ, and fibrosis was scored using the ISHAK scale.

Quantitative real-time polymerase chain reaction

Chronic CORT leads to a downregulation of glucocorticoid receptors in the prefrontal cortex (PFC) and hippocampus, brain regions associated with feedback mechanisms important for the regulation of the hypothalamic–pituitary–adrenal (HPA) axis (43–46). This consequently leads to reduced synaptic inputs, decreased plasticity, and ultimately damage to these brain structures (45–47). Within the hypothalamus, CORT stimulates neuropeptide Y, promoting an increase in appetite and fat deposition (48,49). To determine the role of GHSR signaling in excess glucocorticoid mediated HPA axis dysregulation, we collected brain punches containing the PFC, hippocampus, and mediobasal hypothalamus using the Palkovitz method (50) and the Franklin and Paxinos mouse atlas as an anatomical guide (51). These samples were processed for quantitative real-time polymerase chain reaction (PCR) looking at the expression of glucocorticoid receptors (NR3C1) and genes associated with GHSR signaling and neuroprotection including the expression of carnitine palmitoyltranferase 1 (CPT-1), uncoupling protein 2 (UCP2) (52). Samples from livers were collected to determine differences between the groups in gene expression for genes associated with cell stress and inflammation. These included the analysis analyses of relative changes in the expression of interleukin 6 receptor, alkaline phosphatase, alanine aminotransferase, transforming growth factor α, and transforming growth factor α receptor 1. The expression of elongation factor 2 messenger ribonucleic acid (mRNA) was used as control for liver samples, and the expression of 18S ribosomal ribonucleic acid (RNA) was used as control for brain samples. Total RNA was extracted with TRIzol as detailed in the manufactures protocol (Life Technologies) from liver samples collected from mice both experiments and from brain punches from mice in Experiment 2.

The purity and concentration of RNA was determined using a Nanodrop spectrophotometer (Thermo Scientific). RNA integrity was examined via agarose gel electrophoresis (1% w/v agarose gel with ethidium bromide). RNA of adequate quality was reverse transcribed with iScript Reverse Transcription Supermix as detailed by the manufacturer (Bio-Rad). Reaction efficiencies of all primer pairs were tested using pooled complementary deoxyribonucleic acid from all mice involved in these experiments. The reaction efficiencies of all primer pairs fell between 95% and 110%. Individual samples were diluted to fit in the linear dynamic range of the standard curve for each primer pair. The sequences of the primers used for the target genes are found in Table 1. Quantitative PCR plates ran on a CFX Connect Real-Time PCR system (BioRad). All samples were run in triplicate, and the data were organized and extracted using Bio-Rad CFX Manager 3.1 and analyzed using the 2^-∆∆CT method (53,54).

Table 1.

Real-time PCR primer pair sequences

Target	Primer direction	Primer sequence 5’- 3’
Interleukin 6 receptor (IL-6R)	Forward Reverse	CATCCTGAGACTCAAGCAGAAATGGAAG GTGAGGAGAGGAACCAGAAGGAAG
Tumor necrosis factor receptor subfamily 1A (TNFRS1A)	Forward Reverse	CTTGCTAGGTCTTTGCCTTCTATCCTT GCTTTCCAGCCTTCTCCTCTTTGAC
Alanine amino transferase	Forward Reverse	GCCTTTATCACCAGGAGAGATGGT ACCAGGAGCTTCAGGATTGTAGAAA
Alkaline phosphatase	Forward Reverse	TGATGTGGAATACGAACTGGATGAGAAG TAGTGGGAATGCTTGTGTCTGGG
Elongation factor 2 (EGF-2)	Forward Reverse	GCCTTTGGTAGAGTGTTCTCTGG CATCAGAATGGTTCTCTGGATAGGC
Uncoupling protein 2 (UCP2)	Forward Reverse	TCTGGATACCGCCAAGGTGCT TTGTAGAGGCTGCGTGGA
Carnitine palmitoyltransferase CPT-1	Forward Reverse	TGT CCA AGT ATC TGG CAG TCG CAT AGC CGT CAT CAG CAA CC
Glucocoticoid receptor (Nr3C1)	Forward Reverse	CAAAGGCGATACCAGGATTCA GGGTCATTTGGTCATCCAGGT
18S ribosomal RNA	Forward Reverse	GAC TGT CTC GCC GGT GTC GGA GAG CCG GAA CGT CGA

Statistical analyses

Data were analyzed using IBM SPSS software using a 2 × 2 between group analysis of variance (ANOVA) design when looking at group differences with genotype or pharmacological condition (WT or GHS-R KO; saline or JMV 2959) and treatment (CORT or 1% EtOH) as between-group factors. Repeated measures of ANOVAs with genotype or pharmacological condition and treatment as the between group variables and treatment days as within group factors were conducted to evaluate data across the experiment. Significant interaction effects were followed up with Fischer’s least significance difference post hoc tests.

Results

Ghrelin receptor pharmacological blockade or gene deletion results in a poorer metabolic phenotype in mice receiving chronic CORT

In general, CORT treatment resulted in metabolic alterations that were evident after the first week of treatment. Unexpectedly, 6 mice died during the CORT treatment, and 1 reached end point and had to be euthanized. The data from these mice were not used for the analyses presented next.

As expected, CORT treated mice consumed more food and gained more weight regardless of whether they received saline or JMV2959 injections (main effect for CORT on food intake; F [1,26] = 166.1, P < .01; main effect for CORT on weight gain; F [1,26] = 20.3, P < .01). Also as expected, JMV2959 treatment produced an overall decrease in food intake regardless of whether the mice received CORT or not (main effect for JMV2959; F [1,26] = 5.37, P = .028; see Fig. 1A). While the effect of JMV2959 seemed to be mostly evident in CORT treated mice, there was no significant interaction effect (F [3,30] = 3.64, P = .067). Despite the decrease in food intake, mice treated with JMV2959 tended to gain more weight regardless of whether the mice received CORT or not (main effect for JMV2959; F [1,26] = 3.53, P = .07; see Fig. 1B). The fact that JMV2959-treated mice gained weight while decreasing their food intake was also reflected in a significant main effect in feeding efficiency (main effect for JMV2959 treatment; F [1,26] = 5.76, P = .02; see Fig. 1C). In contrast, there were no significant main effect of CORT or a significant CORT × JMV2959 interaction effect in feeding efficiency (P > .05)

Figure1.

Cummulative food intake and weight change in mice that received vehicle or JMV2959 (Experiment 1) or WT and GHSR KO mice (Experiment 2) given CORT mixed in their drinking water. As depicted in this figure, the GHSR antagonist JMV2959 was not effective in reducing the effects of chronic CORT (100 μg/mL) on food intake or weight gain (A, B). GHSR KO mice also showed a significant increase in caloric intake that was equivalent to that seen in their WT littermates and gained more weight than WT CORT treated mice (P < .05). *Significant main effect for CORT, P < .05. **Significant from CORT-treated WT mice, Fisher’s least significant difference posthoc, P < .05. #Significant main effect for JMV2959.

CORT treatment also increased food intake and weight gain in WT and GHSR KO mice. While there were no differences in food intake between CORT-treated WT and GHSR KO mice, GHSR KO mice gained significantly more weight than WT mice after 4 weeks of CORT treatment (significant interaction effect; F [3,30] = 6.79, P = .016; see Figs. 1D and 1E). This was also reflected in metabolic efficiency, where CORT treated GHSR KO mice had much higher feeding efficiency than mice in any of the other groups (significant interaction effect; F [3,30] = 5.98, P = .02; see Fig. 1F).

In line with results on weight gain, the ghrelin receptor antagonist JMV2959 did not attenuate CORT-induced adiposity in mice. Statistical analyses on the weight of different fat pads determined that while CORT resulted in higher perigonadal, retroperitoneal, subcutaneous, and brown fat weights, JMV2959 treatment was not effective in decreasing adipose tissue accumulation (see Fig. 2). If JMV2959 had any effects, these were of enhancing accumulation of fat in the retroperitoneal fat depots (main effect for JMV2959; F [1,26] = 3.89, P = .059). Furthermore, a near significant interaction suggested that brown fat depots from CORT-treated mice that also received JMV2959 tended to be higher than those from CORT-treated mice receiving only the vehicle (interaction effect; F [3,30] = 3.01, P = .089). These effects were more evident in CORT-treated mice lacking a functional GHSR. As seen in Fig. 2B, GHSR KO mice tended to have lower fat pad weights than WT mice, but the adipogenic effects of CORT were still evident in these mice as demonstrated by a significant main effect of CORT in all fat pad weight measures analyzed (main effects for total fat pad weight, F [1,26] = 12.8, P = .02; perigonadal, F [1,26] = 12.1, P = .002; subcutaneous, F [1,26] = 7.61, P = .01; retroperitoneal, F [1,26] = 8.26, P = .009; BAT, F [1,26] = 114.42, P < .001). Thus, regardless of mice expressing the GHSR or not, CORT treatment significantly increased total weight of fat pads (see Fig. 2).

Figure 2.

Fat pad weight from mice in Experiments 1 and 2. As depicted in this figure, CORT treatment resulted in increases in overall fat pad weight, and these increases tended to be more pronounced in retroperitoneal and brown fat from mice receiving the GHSR antagonist JMV2959. Similar results were observed in CORT-treated WT and GHSR KO mice where CORT treatment resulted in increases in overall fat pad weight and individual fat pads. While GHSR KO mice had lower fat pads when treated with EtOH, they showed marked increases in all fat pads when treated with CORT, ultimately accumulating as much fat as WT CORT-treated mice. These increases were more pronounced in fat pads from GHSR KO mice. *Significant main effect for CORT, P < .05.

CORT treatment also caused a profound increase in plasma insulin and leptin concentrations. While insulin concentrations were not any higher in JMV2959-treated or GHSR KO mice compared to controls, leptin concentration were significantly lower in CORT-treated mice that also received JMV2959 compared to CORT-treated controls (leptin interaction effect, F [3,30] = 7.44, P = .01; see Fig. 3A). In contrast, leptin concentrations of GHSR KO mice treated with CORT were significantly higher than those of CORT treated WT mice (leptin interaction effect, F [3,24] = 7.10, P = .014; see Fig. 3B). CORT also significantly increased plasma glucose concentrations (main effect for CORT experiment 1, F [1,26] = 110.66, P < .001; main effect for CORT experiment 2, F [1,26] = 19.2, P < .001), but neither blocking the GHSR with JMV2959 nor the lack of a functional GHSR resulted in lower fasting plasma glucose levels following CORT treatment. A significant repeated measures ANOVA showed that mice lacking GHSR, however, were better at clearing glucose in the glucose tolerance test regardless of whether they received CORT or not as determined by a significant genotype × time interaction effect (F [1,26] = 3.16, P = .017; see Fig. 3C). A main effect for genotype on measures of area under the curve also confirmed that GHSR KO mice were able to clear glucose compared to WT animals regardless of whether they received CORT or not (F[1.26] = 4.97, P = .035).

Figure 3.

Plasma leptin, insulin, and fasting glucose levels in mice that received vehicle or JMV2959 (Experiment 1) or WT and GHSR KO mice (Experiment 2) given CORT mixed in their drinking water. The figure also shows plasma glucose concentrations following the glucose tolerance test. As seen in this figure, the effects of chronic CORT treatment on leptin, insulin, or glucose concentrations were not attenuated by JMV2959 treatment nor by deletion of the GHSR. Glucose clearance was better in GHSR KO mice regardless of whether they received CORT or not. *Significant main effect for CORT, P < 0.05. **Significant from CORT-treated WT mice, Fisher’s least significant difference posthoc, P < .05.

Chronic CORT treatment has been associated with the accumulation of lipids in the liver ultimately leading to hepatic steatosis (55,56). Consistent with this, our results show that chronic CORT treatment influenced liver adiposity and signs of hepatic damage. Nevertheless, no significant CORT × JMV2959 or CORT ×y genotype interaction effects were detected, suggesting that decreasing GHSR signaling did not attenuate the effects of CORT on lipid accumulation in the liver. Thus, regardless of whether mice received JMV2959 or not, CORT treatment led to increased lipid deposition and triglyceride content in the liver (main effect for CORT, F [1,26] = 75.6, P < .0001). Similarly, the liver of CORT-treated GHSR KO and WT mice showed increased lipid deposition and triglyceride content (main effect for CORT, F [1,22] = 26,6, P < .0001; see Fig. 4). Further histological analyses of liver samples showed that GHSR blockade with JMV2959 did not prevent CORT-induced liver steatosis (macrosteatosis; main effect for CORT in JMV2959 study, F [1,26] = 34.32, P < 0.0001; microsteatosis; main effect for CORT, F [1,26] = 12.33, P < .001). In the same line, CORT increased liver steatosis in both WT and GHSR KO mice (macrosteatosis main effect for CORT, F [1,22] = 22.86, P < .0001; microsteatosis main effect for CORT, F [1,22] = 16.7, P < .001; see Fig. 4). In fact, quantitative PCR analyses of liver samples showed that GHSR KO mice exposed to chronic CORT treatment had higher levels of alanine transferase expression compared to controls and CORT-treated WT mice (interaction effect, F [1,22] = 5.54, P = .03). While chronic CORT treatment also increased mRNA expression of other markers of liver inflammation including interleukin 6 and tumor necrosis factor alpha receptors, genetic deletion or pharmacological antagonism of the GHSR did not significantly alter mRNA expression of these targets (see Fig. 5). Hepatic sclerosis was not observed in any mice regardless of treatment received.

Figure 4.

Effects of chronic CORT treatment on liver triglyceride content and histopathology in control and in mice treated with JMV2959 or with a targeted deletion of the GHSR gene. (A) Representative images of histological samples collected from animals in Experiments 1 and 2, and processed for Periodic acid–Schiff and red oil red staining to analyze the effects of chronic CORT on the integrity of the liver in mice that received vehicle or JMV2959 or WT and GHSR KO mice given CORT mixed in their drinking water. These images demonstrate that chronic CORT treatment caused significant damage to the liver as demonstrated by increased fibrosis, inflammation, and lipid accumulation. These effects were not mitigated by genetic deletion or pharmacological reduction of GHSR signaling. As seen in this figure, blocking GHSR signaling with JMV2959 (B) or with a genetic deletion of the GHSR gene (C) did not attenuate the effects of chronic CORT on fat accumulation in the liver, nor the structural damage observed in liver sample pathological assessments. *Significant main effect for CORT, P < .05.

Figure 5.

Effects of chronic CORT treatment on markers for compromised liver function and inflammation. Neither treatment with the GHSR antagonist JMV2959 nor GHSR deletion prevented increases in alanine aminotransferase and alkaline phosphatase, enzymes associated with compromised liver function. Decreased GHSR signaling was also not effective in reducing the increased expression of interleukin 6 receptors in observed after chronic CORT. *Significant main effect for CORT, P < .05. **Significant from CORT-treated WT mice, Fisher’s least significant difference posthoc P < .05.

Treatment with JMV2959 produces behavioral responses that are independent from elevated CORT

Analyses on the data from behavioral screens for emotional and coping behaviors demonstrated that, regardless of drug administered, chronic CORT treatment resulted in increased immobility in the forced swim test (F [3,30] = 20.42, P < .001, see Fig. 6). Similarly, chronic CORT decreased locomotor activity as measured by line breaks in the open field test (F [3,30] = 5.36, P = .03; see Fig. 6) but did not alter the amount of time spent in the center of the open field box (P > .05). Interestingly, and regardless of whether mice were receiving CORT or not, JMV2959-treated mice had longer latencies to approach a stranger (main effect for drug, F [3,30] = 4.58, P = .042; see Fig. 6) suggesting that this GHSR antagonist may have anxiogenic or locomotor effects that are not associated with chronic CORT.

Figure 6.

Effects of chronic CORT treatment anxiety and depressive-like behaviors in mice treated with vehicle or JMV2959. Results show that a reduction in GHSR signaling was not only ineffective in reducing some of the behavioral consequences of being exposed to chronic CORT including reduced locomotor activity and increased helplessness but also was associated with increased anxiety-like behaviors independently from CORT treatment. *Significant main effect for CORT, P < .05. **Significant from vehicle treated mice, P < .05.

Chronic CORT results in alterations in mRNA expression in the PFC and hippocampus, and some of these changes are more evident in GHSR KO mice

Analyses of mRNA expression in the PFC show that chronic CORT treatment resulted in overall increases in UCP2 mRNA expression and decreased NR3C1 mRNA expression in the PFC (main effect for CORT on UCP2 expression, F [1,24] = 33.38, P < .001; main effect for NR3C1, F [1,24] = 6.1, P = .021). Importantly, the increase in UCP2 expression was attenuated in GHSR KO mice (interaction effect, F [3,28] = 4.204, P = .05; see Fig. 7). Within the hippocampus, chronic CORT treatment produced a significant decrease in the expression of NR3C1 mRNA, but this effect was independent of the genotype of the mice (main effect for CORT, F [1,23] = 38.01, P < .001). Similarly, GHSR KO mice tended to have higher mRNA expression of NR3C1 regardless of CORT treatment; this effect did not attain statistical significance (main effect for genotype, F [1,23] = 4.04, P = .056). Chronic CORT treatment did not change the expression of CPT-1, UCP2, or NR3C1 expression in the hypothalamus (see Fig. 7).

Figure 7.

Effects of chronic CORT on the expression of transcripts associated with ghrelin signaling and metabolic function (CPT-1, UCP2) or HPA function (NR3C1) in the PFC, hippocampus, and hypothalamus of WT and GHSR KO mice. Chronic CORT treatment resulted in increased expression of UCP2 in the PFC, and this effect was attenuated in GHSR KO mice. Chronic CORT treatment was also associated with decreased NR3C1 expression in the hippocampus, but this effect was not affected by the absence of GHSR. *Significant main effect for CORT, P < .05. **Significant from WT CORT-treated mice, Fisher least significant difference, P < .05.

Discussion

In the present study, we examined the role of GHSR signaling in mediating the obesogenic effects of chronic CORT treatment. We hypothesized that ghrelin signaling through the GHSR would be important for the obesogenic effects of chronic CORT exposure, in a similar manner to that which we have observed in response to chronic stress (34) and that pharmacological blockade or genetic ablation of the receptor would improve the metabolic phenotype of mice exposed to chronic CORT. Our results indicate that, in contrast to our hypothesis, the GHSR is not required for the obesogenic effects of chronic CORT exposure.

As expected, chronic treatment with CORT reliably produced metabolic disturbances consistent with previous reports (6,7,36). For instance, mice given chronic CORT developed excessive caloric intake, weight gain, adiposity, and hepatic steatosis. The effects of CORT, however, were not diminished by JMV2959 treatment nor in GHSR KO mice. Thus, while JMV2959 treatment reduced cumulative food intake in CORT-treated mice, these mice gained more weight per milligram of food consumed than CORT-treated control mice. Similarly, weight gain of GHSR KO mice treated with CORT was higher than the weight gained by WT mice treated with CORT. Chronic treatment with CORT also resulted in increased adiposity that was not prevented by decreased GHSR signaling. In fact, mice treated with the GHSR antagonist or mice lacking GHSR accumulated more retroperitoneal fat and BAT in response to CORT than control mice treated with CORT. The accumulation of adiposity also resulted in increased plasma leptin concentrations that were attenuated in JMV2959-treated mice and exaggerated in GHSR KO mice in comparison with CORT-treated controls. In all, these data support the notion that GHSR signaling, while commonly associated with increased feeding and adiposity, may in fact have an opposite effect in situations where CORT is chronically elevated. While these data are paradoxical when one considers previously reported lipolytic effects of GHSR receptor antagonists or GHSR deletion, GHSR receptor mutations worsen the metabolic phenotype of leptin deficient mice (57). More important, however, is the fact that these data suggest that the weight gain, adiposity, and increased food intake that results from chronically elevated levels of glucocorticoids do not require ghrelin receptor signaling.

Chronic CORT treatment also resulted in the accumulation of fat in the liver and in the expression of markers for liver stress and damage as indicated by histological examination of liver samples, as well as by changes in the expression of genes encoding for proteins and enzymes associated with liver stress. These effects were severe in all mice treated with CORT, and no differences were observed in histological samples of livers from CORT GHSR KO mice nor in GHSR receptor antagonist-treated mice compared to samples from control mice. GHSR KO mice treated with CORT did show increased mRNA expression for alanine transferase, an enzyme commonly used as a marker for liver stress in clinical assessments of liver function. These data again not only show that the effects of CORT on liver function, lipid accumulation, and liver stress are independent from ghrelin receptor signaling, but also that reduced or deleted ghrelin receptor signaling may amplify the pathological effects of CORT on liver function. These data also support previous reports suggesting that ghrelin has protective effects against liver damage in response to chronic exposure to a high fat diet or ischemia/perfusion (58,59).

Ghrelin has been implicated in the regulation of glucose homeostasis, and mice with mutations to the gene encoding for ghrelin, GHSR, or double KO for these genes show improved glucose tolerance (60). In this study, we observed that mice treated with CORT also showed high levels of fasting glucose, and these were independent of whether the mice had functional ghrelin receptors or were treated with the GHSR antagonist JMV2959 by the end of the experiment. In addition, our data shows that by day 18 of treatment, GHSR KO mice showed lower plasma glucose concentrations during the glucose tolerance test than their WT littermates, and these effects were independent from CORT treatment, suggesting that blocking ghrelin receptor signaling can improve glucose tolerance even when CORT levels are high. Nevertheless, it is important to note that the time period we chose to do this test may represent a time where CORT has not completely disrupted glucose regulation in GHSR KO or WT mice.

Chronic treatment with glucocorticoids for prolonged periods of time are associated with behavioral changes similar to those observed following chronic stress, and these include decreased locomotor activity, increased immobility, and changes in the motivation to interact with conspecifics (61–64). In our study, we observed that CORT treatment also resulted in reduced locomotor activity in the open field box and longer immobility time in the forced swim test. While these tests may suggest that CORT treatment is producing effects that result in negative affect, they also suggest that the metabolic phenotype of these animals may be due in part to decreased locomotor activity. Moreover, this reduction in locomotion was observed in CORT-treated mice regardless of whether they were treated with JMV2959 or not, indicating that blocking ghrelin receptors does not alter locomotor behaviors in these tests. Importantly, and regardless of whether mice received CORT or not, JMV2959-treated mice showed decreased motivation to interact socially with a conspecific as measured by an increased latency to approach a stranger mouse in the social interaction test. This would be in line with data showing that GHSR KO mice are more susceptible to chronic stress-induced social interaction deficits using a similar task (65). This is also interesting because it suggests that GHSR signaling may be associated with social behaviors.

The effects of CORT manipulation on affective states is thought to occur via the actions of CORT on a number of brain regions that include the hippocampus, prefrontal cortex, and hypothalamus, all important in the regulation of the stress response (36,66–68). Chronically elevated CORT, however, leads to changes in the ability of these regions to regulate the stress response and ultimately can cause cellular changes that include decreased cell proliferation in the hippocampus, decreased dendritic spines in both the hippocampus and prefrontal cortex, and ultimately decreased volume in these structures (69–71). Furthermore, there is evidence for the actions of glucocorticoids on mitochondrial processes that lead to increased oxidative stress and altered cellular function in various tissues including the brain (72). Interestingly, there is evidence supporting the notion that ghrelin has neuroprotective and neurotrophic effects that could attenuate stress-induced insults (52). Moreover, these effects are associated with the recruitment of UCP2, a protein that has neuroprotective effects in cells by removing free radicals derived from the oxidation of fuels in the mitochondria (52,73). Here we reasoned that chronic CORT would increase UCP2 expression in brain regions like the hippocampus, PFC, and hypothalamus. Furthermore, we expected that if GHSR signaling was important for protection of these regions, we would also observe attenuated UCP2 expression in the hippocampus, PFC, and hypothalamus of CORT-treated GHSR KO mice. Our results show that, as we expected, CORT-treated mice showed significant increases in UCP2 expression in the PFC, an effect that was marginally attenuated in CORT-treated GHSR KO mice. The expression of UCP2 in the hypothalamus or the hippocampus was not significantly affected by CORT nor by the absence of the GHSR. In contrast, and as expected, the expression of the glucocorticoid receptor in the PFC and hippocampus was significantly reduced in CORT-treated mice, but these changes were not different between GHSR KO and their WT littermates. Ultimately, these data suggest that the adaptations associated with the regulation of the HPA axis are independent of GHSR signaling, but that GHSR signaling, at least in the PFC, may play a role in mitigating the negative impact of chronic CORT.

The paradigm used in the current study is a common model to study the effects of chronic exposure to CORT in a number of systems, but caution should be used to parallel this model with models using stressors to examine the effects of CORT. Exposure to chronic CORT has an immediate impact on the HPA axis, and because the animals receive the CORT in their drinking water, it is difficult for them to engage allostatic systems that reduce CORT. Additionally, CORT treatment induced a hyperglycemic state causing subsequent polydipsia that increased CORT dosing. Not surprisingly, these animals show a dramatic reduction in glucocorticoid receptor expression in the hippocampus and PFC that is independent of intact GHSR signaling. In contrast, chronic stress paradigms allow for glucocorticoid induced negative feedback mechanisms that lower CORT levels, an adaptation that is critical for reducing the impact of chronic stressors. A number of chronic stress paradigms including the social defeat stress model result in increased caloric intake and, in some cases, increased weight gain and adiposity and decreased energy expenditure that parallel those seen in chronic CORT experiments (although less dramatic) (11,34,74,75). Chronic social defeat stress, however, does not increase CORT levels to the levels seen in chronic CORT-treated mice (34). Moreover, these stress induced alterations are accompanied by increased ghrelin secretion and can be attenuated by central delivery of GHSR antagonists or in GHSR KO mice. In all, these data would suggest that, while ghrelin and CORT are released during stress responses, the metabolic and neuroprotective effects of these hormones are independent of each other.

Finally, results derived from these data point to a potential protective effect of GSHR signaling in the severe metabolic and behavioral phenotype induced by chronic CORT. The effects of CORT on weight gain, adiposity, and hyperleptinemia were amplified by compromised GHSR signaling. These data would suggest that ghrelin can act to regulate metabolism by mitigating metabolic imbalance that can be derived by exposure to chronic elevated levels of glucocorticoids. Furthermore, CORT-treated mice in Experiment 1 also showed a number of behavioral alterations that included increased immobility and decreased locomotor activity as reported in previous studies (37). Surprisingly, mice treated with JMV2959 showed a deficit in the social interaction test, and this effect was evident regardless of whether the animals were given chronic CORT or not. These data suggest that decreases in GHSR signaling may have an impact in social aspects of behavior as hinted by previous studies using GHSR KO mice (21,32).

Chronic exposure to glucocorticoids is associated with decreased synapses, altered synaptic transmission and ultimately with brain pathology (43,44,67). These pathological changes are thought to be mediated in part by a downregulation of glucocorticoid receptors and consequently a dysregulation of the HPA axis (10,76). In contrast, ghrelin has neuroprotective effects, and these are mediated by the activation of proteins that eliminate oxidative stress including UCP2 (52,73). In this study, we report that, as expected, chronic CORT resulted in decreased glucocorticoid receptor mRNA expression in the PFC and hippocampus and these effects were seen regardless of a functional GHSR. Also as expected, we observed increased UCP2 in the PFC of CORT-treated mice, an effect that was attenuated in GHSR KO mice. These data would support the idea that, at least in the PFC, GHSR might have a neuroprotective role, although this needs to be further confirmed. Nevertheless, should this be the case, one could argue that the limited expression of UCP2 seen in GHSR KO mice in the PFC in response to CORT may reflect increased vulnerability to the effects of chronic CORT on depressive and anxiety-like behaviors that were observed in JMV2959-treated mice in this study and that have been reported in previous work with GHSR KO mice (21). Taken together, these data suggest that ghrelin and glucocorticoids influence metabolism through relatively independent pathways and demonstrate that the ghrelin system is not a viable target to curb glucocorticoid-induced metabolic alterations.

Additional Information

Disclosure summary: The corresponding author declares no existing conflicts of interest on behalf of all authors in this manuscript.