Ghrelin influences a variety of metabolic functions through a direct action at its receptor, the GhrR (GhrR-1a). Ghrelin knockout (KO) and GhrR KO mice are resistant to the negative effects of high-fat diet (HFD) feeding. We have generated several classes of small-molecule GhrR antagonists and evaluated whether pharmacologic blockade of ghrelin signaling can recapitulate the phenotype of ghrelin/GhrR KO mice. Antagonist treatment blocked ghrelin-induced and spontaneous food intake; however, the effects on spontaneous feeding were absent in GhrR KO mice, suggesting target-specific effects of the antagonists. Oral administration of antagonists to HFD-fed mice improved insulin sensitivity in both glucose tolerance and glycemic clamp tests. The insulin sensitivity observed was characterized by improved glucose disposal with dramatically decreased insulin secretion. It is noteworthy that these results mimic those obtained in similar tests of HFD-fed GhrR KO mice. HFD-fed mice treated for 56 days with antagonist experienced a transient decrease in food intake but a sustained body weight decrease resulting from decreased white adipose, but not lean tissue. They also had improved glucose disposal and a striking reduction in the amount of insulin needed to achieve this. These mice had reduced hepatic steatosis, improved liver function, and no evidence of systemic toxicity relative to controls. Furthermore, GhrR KO mice placed on low- or high-fat diets had lifespans similar to the wild type, emphasizing the long-term safety of ghrelin receptor blockade. We have therefore demonstrated that chronic pharmacologic blockade of the GhrR is an effective and safe strategy for treating metabolic syndrome.
Ghrelin is a 28-amino acid peptide synthesized in the stomach and pancreas of mammals. Although the ghrelin receptor (GhrR; aka GHSR-1a) was originally cloned as the growth hormone secretagogue receptor (Howard et al., 1996), ghrelin has been shown to affect a variety of metabolic functions including increased food intake (FI), fat storage, gastrointestinal motility, and growth hormone release. Uniquely, ghrelin is post-translationally modified with an octanoyl side chain on serine position 3, which is required for its activity at the GhrR (Smith, 2005). Removal of the acyl group renders the molecule completely inactive at the cognate GhrR. The GhrR is a member of the GPCR super family and its localization is consistent with its known biological functions. The highest concentrations of the receptor mRNA are in the arcuate nucleus of the hypothalamus and the pituitary gland; lower but significant amounts are found in the peripheral neuraxis such as the nodose ganglia, nucleus tractus solitarius, and pancreas (Zigman et al., 2006).
A growing body of evidence suggests that a loss of GhrR signaling improves the health of animals that have been fed a “Westernized” diet that is rich in fats and carbohydrates. GhrR knockout (KO) mice fed high-fat diets (HFDs) accumulated less adipose tissue but retained their lean tissue mass (Zigman et al., 2005; Longo et al., 2008). We have reported that GhrR KO mice resisted diet-induced hepatic steatosis and had higher insulin sensitivity and lower hyperinsulinemia (Longo et al., 2008). Lower FI, and not changes in energy expenditure, seemed to partially explain the mechanism by which these mice resisted diet-induced obesity (Longo et al., 2008). It is noteworthy that GhrR KO mice exhibited greater metabolic flexibility and a lower rate of intestinal dietary lipid absorption/secretion, which suggested that alterations in fuel usage and partitioning, as well as gastrointestinal effects, contributed to their relatively healthy phenotype (Longo et al., 2008). A subsequent analysis of multiple cohorts of HFD-fed GhrR mice confirmed the improved insulin sensitivity of GhrR KO mice and showed how this may be independent of decreases in body weight (Qi et al., 2011). Furthermore, ghrelin KO mice also had a better metabolic profile when fed a Western diet (Wortley et al., 2004, 2005). Thus, antagonism of the GhrR may constitute a unique and robust approach to managing the metabolic syndrome.
Indeed, reports have been published describing classes of ghrelin antagonist that impart improvements in metabolic function. These include peptide-based inhibitors (Asakawa et al., 2003) and small molecules (Esler et al., 2007; Moulin et al., 2007; Rudolph et al., 2007) that affect body weight gain, FI, and energy expenditure. In this work, we present evidence that antagonism of the GhrR with orally bioavailable small molecules provides a new therapeutic modality for the simultaneous treatment of obesity and insulin resistance in mice. The results corroborate those observed in KO mice and provide further support that ghrelin signaling is important in the control of metabolic stress.
Materials and Methods
All animal studies were approved by Elixir Pharmaceuticals' animal care and use committee. Starting at 5 weeks of age, male C57BL/6 mice (Taconic Farms, Germantown, NY) were fed a HFD (5.24 kcal/g, or 60% kcal, from fat; Research Diets, New Brunswick, NJ) for 14 to 16 weeks, with food and water available ad libitum. Mice were group-housed in ventilated cages (Thoren Caging Systems, Hazelton, PA) with enrichment (Igloos, Bioserv, Frenchtown, NJ; Enviro-dri, PharmaServ, Framingham, MA) in a controlled environment (72°F, ∼40% humidity, 12-h light-dark cycles). Wild-type (WT) and GhrR KO mice on a C57BL/6 background were bred from a single male founder at Charles River Laboratories Inc. (Wilmington, MA) (Sun et al., 2004). Mice were genotyped by a polymerase chain reaction assay (Transnetyx, Cordova, TN). In addition, loss of functional GhrR signaling in these mice was verified by demonstrating lack of ghrelin-stimulated FI, growth hormone release, and insulin resistance (Supplemental Fig. 1). Mice were weaned at 3 weeks of age and then maintained on PicoLab rodent diet 20 (LFD; Purina, St. Louis, MO) until 8 weeks of age. Starting at 8 weeks of age, some male GhrR KO and WT mice were placed on a HFD as described above. All experimental mice were housed individually for at least 5 days before study initiation. Except where noted mice were tested in groups of 8 [glucose tolerance test (GTT)] or 10 (FI and body weight studies).
Clonal Chinese hamster ovary cells expressing the human GhrR-aequorin system were obtained from Euroscreen (Brussels, Belgium). Cells were cultured in complete Ham's F12 media containing 10% fetal bovine serum, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml fungizone antimycotic in 0.85% saline, 400 μg/ml G418 (Geneticin) and 250 μg/ml phleomycin. Cells were cultured as monolayers in 75-cm2 cell culture flasks at 37°C with 5% CO2, split, and fed every 2 to 3 days. Cells grown to midlog phase were removed from flasks by a gentle wash with PBS containing 5 mM EDTA (PBS-EDTA), followed by a 10-min incubation at 37°C in PBS-EDTA. Cells were recovered by centrifugation, then counted and resuspended at a density of 5 × 106 cells/ml in BSA medium (Dulbecco's modified Eagle's medium/Ham's F12 with HEPES, without phenol red, containing 0.1% BSA). Coelenterazine h was added to the cells at 5 μM final concentration. The cells were protected from light and incubated at room temperature for 4 to 16 h. The cells were diluted in BSA medium (1/10, v/v) and incubated with stirring for 1 h at room temperature. Human ghrelin (Phoenix Pharmaceuticals, Inc., Burlingame, CA) was diluted in BSA medium and dispensed into black 96-well plates (50 μl/well). Test compounds were diluted to 100× final concentration in 100% dimethyl sulfoxide, and 1 μl was added to each well. We injected 50 μl of stirring cell suspension into each well using a Luminoskan luminescent plate reader equipped with injectors (Thermo Fisher Scientific, Waltham, MA). Light emissions were recorded for 30 s, integrated, and analyzed using Luminoskan Ascent software, resulting in one value representing the intensity of emitted light per test well. We generated IC50 curves and Ki values for each compound with XLfit software (IDBS, Alameda, CA).
Effect of GhrR Antagonist on Ghrelin-Induced FI.
Individually housed, 10-week-old, ad libitum LFD-fed, male C57BL/6 mice were used in these studies. On the morning of the experiment, food was temporarily removed from the cages and weighed. Mice then received an intraperitoneal injection (5 ml/kg) of either vehicle-1 (4% dimethyl sulfoxide, 10% β-OH-cyclodextran) or compound D (CpdD) at 3, 10, or 30 mg/kg. One minute later, vehicle-treated mice received a second intraperitoneal injection of either vehicle-2 (double distilled H2O) or ghrelin (5 mg/kg). Immediately after the second intraperitoneal injection, mice were returned to their home cages and their food was returned. FI was measured 6 h after the injections. In a separate experiment, continuous, overnight endogenous FI was measured in LFD-fed GhrR KO and WT mice dosed intraperitoneally with vehicle or CpdD (30 mg/kg, dosed 3 and 1 h before lights off) using an Oxymax system (Columbus Instruments, Inc., Columbus, OH).
Repeated Treatment of HFD-Fed Mice with GhrR Antagonists.
One week before study initiation, mice that exceeded 42 g b.wt. were individually housed and acclimated to oral dosing with vehicle (5 ml/kg b.i.d.) for 1 week at 9:00 AM and 5:00 PM. Body and food weights were recorded during afternoon dosing. Mice that lost ≥10% of their original body weight during the acclimation period were eliminated from the study. Next, mice were sorted into four groups of equivalent mean body weight (n = 10/group). We then commenced vehicle and compound dosing for 7, 14, 28, or 56 days, followed by a GTT (see below). To determine whether constant dosing was required for observed effects of GhrR antagonist treatment an additional 7-day compound dosing study was performed in which one group of mice received all doses of compound except for the final pre-GTT dose. Likewise, in the 56-day study an additional group was included in which mice received only a single pre-GTT dose of compound or vehicle.
In some experiments, additional analyses were performed. The percentage of glycated hemoglobin (%HbA1c) was determined using a DCA2000+ analyzer (Bayer, Indianapolis, IN). Plasma triglycerides, lipoproteins, and cholesterol were determined using a Cholestech LDX analyzer (Cholestech, Hayward, CA). Mouse IGF-1 was measured by enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN). Plasma free fatty acids were measured using a kit from Sigma-Aldrich (St. Louis, MO). Additional biochemical markers in plasma were assessed by AniLytics (Gaithersburg, MD). Insulin resistance was calculated using the homeostasis model assessment of insulin resistance (HOMA-IR) using fasting plasma insulin (milliunits per liter) and fasting blood glucose (millimolar) in the following equation: HOMA-IR = (fasting plasma insulin × fasting blood glucose)/22.5 (Matthews et al., 1985). Liver lipid content was determined by chloroform/methanol (2:1) extraction (Folch et al., 1957).
Glucose Tolerance Test.
On the evening before the experiment, body weight was measured in all mice followed by an overnight (16 h) fast. On the morning of the GTT, a 40-μl blood sample was collected by tail nick for determination of fasted blood glucose and plasma insulin. One hour later, mice were dosed with glucose (dose and route given in legends for Figs. 2, 4, and 5). Mice in the pharmacology studies received their normal dose of compound after the initial tail bleed 1 h before being given glucose. Tail blood was sampled at 15, 30, 60, and 120 min after the glucose dose for the determination of blood glucose and plasma insulin. Blood glucose was measured using Ascencia Elite glucometers (Bayer), and plasma insulin was measured by enzyme-linked immunosorbent assay (CrystalChem, Downer's Grove, IL). In one experiment, the capacity of mice to secrete insulin was tested using the insulin secretagogue repaglinide (Sigma-Aldrich).
HG clamps were performed on mice that had been dosed orally for 7 days with either vehicle or CpdB (45 mg/kg b.i.d.). The HG clamp methodology has been described previously (Qi et al., 2011).
Groups of mice were administered vehicle or CpdB, and body weight and FI were recorded daily for 8 days. A third group was administered vehicle, and each animal within this group was fed daily with an amount of food equivalent to the mean daily FI of the CpdB-treated group (PF to CpdB). A fourth group of food restriction (FR) (to 60% vehicle) was administered vehicle and fed daily with an amount of food equivalent to 60% of the mean daily FI of vehicle-dosed mice.
Liver and pancreas were fixed in 10% buffered formalin, embedded in paraffin, and cut into 5-μm sections (Mass Histology, Worcester, MA). Sections were dewaxed in xylene, hydrated, and stained with hematoxylin and eosin (H&E) for examination of liver morphology. Images of liver were acquired using a 12,5-megapixel camera (DP70; Olympus, Tokyo, Japan). Pancreas sections were dewaxed in xylene, hydrated, and boiled for 10 min in a microwave oven in 10 mM sodium citrate, pH 6 for antigen retrieval. Primary antibodies used for pancreas staining were guinea pig anti-insulin (Millipore Corporation, Billerica, MA) diluted 1:1000 and rabbit antiglucagon diluted 1:200 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Biotinylated anti-guinea pig or anti-rabbit secondary antibodies were used and detected with the Vectastain ABC kit and diaminobenzidine tetrahydrochloride (Vector Laboratories, Burlingame, CA). Sections were then counterstained with H&E. Stained sections were permanently mounted with clarion mounting media (Sigma-Aldrich). Images were acquired with a Spot digital camera (Micro Video Instruments, Avon, MA).
Statistical Analyses of Data.
Data collected over time were analyzed by repeated measures two-way analysis of variance with Bonferroni post hoc tests. End-of-study plasma values were compared using Student's t test. Unless otherwise noted, data were considered statistically significant at P < 0.05. All analyses were performed using Prism version 4.03 for Windows (GraphPad Software Inc., San Diego, CA).
We generated multiple chemical series of GhrR antagonists to test whether the favorable metabolic phenotype of the GhrR KO mouse fed a high-fat diet could be recapitulated by pharmacologic blockade of the receptor. Several representatives of these series were chosen for studies in vivo. These compounds were screened for their ability to inhibit ghrelin-stimulated activation of the human GhrR using a luminescence-based reporter assay system. Ki values for each compound were calculated from IC50 values using the Cheng-Prusoff equation (Cheng and Prusoff, 1973) and confirmed through 125I-ghrelin binding studies using membranes from cells that overexpressed the human or rodent GhrRs. Compounds showed a strong correlation of aequorin assay Ki at human and rodent receptors (r = 0.7; p < 0.05; data not shown). Properties of select competitive antagonist molecules are shown in Supplemental Table 1; oral bioavailability is also shown for several compounds. Compound structures are shown in Supplemental Fig. 2. In addition, compounds were routinely counterscreened against 20 to 100 other drug-like targets (GPCRs, transporters, ion channels, enzymes) and found to be 100- to >2000-fold selective for the human GhrR (data not shown).
Blockade of Ghrelin-Induced Responses.
We have shown that ghrelin administration stimulates FI and growth hormone release, responses that are absent in GhrR KO mice (Supplemental Fig. 1, A and B). In addition, the ability of a single intraperitoneal injection of ghrelin to induce acute, transient glucose intolerance was absent in HFD-fed GhrR KO mice (Supplemental Fig. 1C). Thus, a series of experiments was carried out to determine whether antagonist compounds could block ghrelin-induced responses in a dose-dependent and selective manner. We initially used FI assays to determine the acute efficacy of ghrelin antagonist compounds. Consistent with previous reports in a variety of species, ghrelin administration (5 mg/kg i.p.) stimulated a significant increase in FI in mice over a 6-h period, and concomitant intraperitoneal dosing with CpdD was able to reduce significantly the ghrelin-stimulated FI in a dose-dependent manner (Fig. 1A). In this in vivo assay format we have been able to show that numerous compounds with in vitro antagonist activity are able to block ghrelin-induced FI (data not shown).
To examine the selectivity of antagonists for effects on the GhrR we monitored the effect of antagonist treatment on spontaneous feeding using a metabolic chamber apparatus that records continual food consumption. Eight-week-old WT and GhrR KO mice with access to a normal chow diet were dosed orally with either vehicle or CpdD at 3 and 1 h before the beginning of the dark cycle. We found that mice treated with CpdD had decreased nocturnal FI during the 12-h dark-phase period (Fig. 1B). It is noteworthy that the spontaneous FI in GhrR KO mice administered vehicle was lower compared with the WT and equal to WT mice treated with CpdD. However, administration of CpdD to KO mice showed no further effect on FI, showing that the effects of the compound on FI were selective for the GhrR (Fig. 1B).
Repeated Treatment of HFD-Fed Mice with GhrR Antagonists.
We and others have previously shown that GhrR KO mice resist many of the negative effects of HFD, including the development of insulin resistance. To determine whether GhrR antagonist treatment of HFD-fed mice could recapitulate this phenotype in WT mice, a potent antagonist CpdB (Supplemental Table 1) was administered to HFD-fed mice and a variety of parameters was measured. HFD-fed mice were dosed for 7 days with either vehicle or CpdB (30 or 60 mg/kg b.i.d. p.o.), and the effects were evaluated. CpdB administration resulted in a dose-dependent decrease in body weight and FI (Fig. 2, A and B, respectively). In this study, CpdB had only modest, nonsignificant effects on glucose disposal (Fig. 2C); however, the insulin required to dispose of the glucose was significantly reduced (Fig. 2D). We have observed the insulin sparing effects of both genetic and pharmacologic blockade of GhrR signaling in GTTs using either oral or intraperitoneal glucose loads (Supplemental Fig. 3), indicating that the incretin effect of the oral glucose load (Pratley and Gilbert, 2008) is intact in the context of GhrR blockade.
To further analyze the results of CpdB in terms of an insulin sensitivity index and specificity for the GhrR we transformed the data from the GTT using the HOMA-IR. HOMA-IR transformation shows that mice treated with CpdB have a lower insulin resistance score compared with vehicle, as expected (Fig. 2E). In a separate study, we compared the effect of 7-day oral CpdB (60 mg/kg b.i.d.) versus vehicle treatment in GhrR KO mice to demonstrate the selectivity of the response. As shown in Fig. 2F, GhrR KO mice showed a lower HOMA-IR score compared with the WT when placed on the HFD, corroborating previous reports of increased insulin sensitivity in the KO. In addition, we found that CpdB had equivalent HOMA-IR scores relative to both vehicle and CpdB-treated HFD-fed GhrR KO mice, establishing its GhrR selectivity. However, in this experiment the differences between GhrR KO mice and the CpdB-treated WT mice failed to reach statistical significance relative to the vehicle-treated WT mice. Nevertheless, this study, coupled with the observations in Fig. 1B, support that CpdB is likely acting specifically through the GhrR to improve insulin sensitivity.
Repeated dosing of GhrR antagonists is often associated with decreases in body weight and FI. However, the effects during a 7-day treatment period can be subtle and body weight changes are not always observed (Supplemental Fig. 4, A and B). Despite this, improvements in glucose homeostasis are consistently observed (Supplemental Fig. 4, C and D). Nevertheless, to more fully investigate the contributions of decreased body weight and FI to the increased insulin sensitivity response to GhrR antagonist treatment we performed a pair-feeding study, comparing vehicle and CpdB in the various feeding paradigms in a 2 × 2 design. The results show that, although pair-fed animals mirrored the body weight changes induced by CpdB as expected, they did not show any improvement in insulin or glucose homeostasis (see Fig. 3), indicating that the effect of CpdB on apparent insulin sensitivity was not caused by a loss in body mass, rather they suggest an alteration of metabolic set point, much like what was observed in GhrR KO mice on HFD.
To evaluate whether the insulin-sensitizing effect of GhrR antagonist treatment is sustained after dosing is discontinued, HFD-fed mice were treated twice daily for 7 days with CpdB (60 mg/kg p.o.) as above except the last pre-GTT dose was eliminated. In these animals the effect on insulin sensitivity was still manifest but the magnitude of the response was diminished (Supplemental Fig. 5A). In a related experiment we dosed animals with only a single pre-GTT dose of CpdB and found that there was no efficacy with respect to improved glucose homeostasis (Supplemental Fig. 5B). Taken together, these result suggest that relatively constant blockade of the receptor is required for maintained efficacy.
To confirm the improved glucose homeostasis evident in the GTT assays, CpdB-treated, HFD-fed mice were evaluated in an HG clamp. Figure 4, A and B shows that there were no differences in first-phase insulin release after the initial glucose load in CpdB and vehicle-treated mice. However, during the second glycemic clamp (300 mg/dl) phase of the experiment (Fig. 4C), CpdB-treated mice displayed significant reduction (AUC; p < 0.05) in the insulin required to maintain the clamp (Fig. 4D). Furthermore, the glucose infusion rate was increased in CpdB-treated animals consistent with improved insulin sensitivity (Fig. 4F). Thus, pharmacologic blockade of GhrR signaling is associated with increased insulin sensitivity and this is consistent with HG clamp studies obtained in GhrR KO mice.
We next addressed the effects of longer-term treatment with a ghrelin antagonist on metabolic homeostasis. Mice were treated with vehicle or CpdB (60 mg/kg p.o.) for 28 and 56 days. As observed with shorter-term (7 days) treatment, CpdB caused a transient decrease in FI that was completely resolved by 7 days (Fig. 5A). It is noteworthy that animals lost 9 to 10% of their body weight by 7 days and maintained this level of weight loss throughout the 56-day study (Fig. 5B). After 28 or 56 days of treatment an oral GTT was performed. With 28- and 56-day treatment periods we noticed greater efficiencies in glucose disposal (Fig. 5, C and E, respectively) and even more remarkable reductions in the second-phase plasma insulin response (Fig. 5, D and F).
That the flat insulin response observed after long-term treatment was the result of CpdB-induced toxicity is unlikely because glucose disposal was greatly improved after 28 and 56 days. In addition, immunohistochemical staining for insulin (Fig. 6, C and D) or glucagon (Fig. 6, E and F) revealed no discernable differences in vehicle- or CpdB-treated mice in this study. In addition, HG clamp experiments have demonstrated that first-phase insulin release does indeed occur in animals dosed with GhrR antagonists (Fig. 4A). Nevertheless, to further investigate this issue we performed a pancreas challenge study using the short-acting insulin secretagogue, repaglinide, to determine whether the ability of the pancreas to release insulin was affected by antagonist treatment. Mice treated for 14 days with CpdB (30 mg/kg b.i.d. p.o.) were given repaglinide (2 mg/kg) at the time of the last dose of CpdB, and the insulin and glucose responses were recorded during a GTT. As expected, mice dosed chronically with CpdB were capable of secreting insulin in response to repaglinide, comparable with the vehicle-treated mice (Supplemental Fig. 6). Taken together, these experiments further demonstrate the safety of the compound on pancreas function.
Upon further examination of organs from long-term GhrR antagonist-treated mice we observed two very striking features. First, the amount of white fat in these animals was significantly reduced (Table 1). We found that subcutaneous fat was not significantly changed but omental fat was reduced by 25 to 45% relative to vehicle-treated controls. These changes probably contributed substantially to the sustained, reduced total body weight of the animals because the mass of different muscle beds was not changed (Table 1). Pancreas wet weight was not significantly altered; however, brown fat pads were slightly reduced with CpdB treatment (Table 1). We and others have shown in GhrR KO mice that white adipose tissue and total body weight are significantly reduced relative to WT controls during chronic HFD feeding. Thus the effects of chronic CpdB treatment replicate the effects of long-term genetic blockade of GhrR signaling.
A second remarkable observation was the extent to which CpdB treatment diminished the fat content of the liver of HFD-fed mice. Livers from CpdB-treated mice had a more normal, reddish appearance compared with the pink color characteristic of steatotic livers of vehicle-treated mice on HFD. Upon gross examination, livers from vehicle-treated animals were highly enlarged compared with those of CpdB-treated mice, as determined by wet weight measurement (Table 1). H&E staining of liver sections revealed widespread steatosis in the vehicle-treated mice fed the HFD, whereas mice treated with CpdB were remarkably devoid of the characteristic lipid inclusions (Fig. 6, A and B, respectively). Finally, biochemical measurement of total fat showed that the livers of CpdB-treated mice had 50% reduced fat content compared with controls (Table 1).
Plasma analysis revealed other important observations in CpdB-treated mice (Table 1). Accompanying the improved insulin sensitivity was a slight, but significant, decrease in HbA1c of 0.46%. We also observed significant decreases in triglycerides and total cholesterol (TC)/HDL cholesterol ratios; no significant effects were observed for HDL or FFA levels, however. It is noteworthy that steady-state plasma insulin-like growth factor-1 levels were not significantly reduced by GhrR antagonist treatment (Table 1). Therefore improvement in several metabolic parameters and functions were evident with long-term pharmacologic blockade of GhrR signaling.
We also assessed the safety profile of the mice after long-term GhrR antagonist treatment by making note of the outward appearance of the mice and measuring a series of plasma analytes. Other than the obvious body weight effect, there was no detectable difference in the overall appearance of CpdB-treated mice versus those treated with vehicle. Supplemental Table 2 shows that electrolytes and several metabolites were not changed. There was, however, a significant decrease in serum alanine aminotransferase and aspartate aminotransferase, in keeping with a protective effect on liver function and morphology (i.e., prevention of steatosis). Lactate dehydrogenase, a marker of cardiac and skeletal muscle damage, was also significantly reduced. Overall, the health of the animals treated with CpdB was, if anything, improved and reflects the point that GhrR antagonism does not seem to be deleterious. In support of this we have performed lifespan studies with male and female GhrR KO and WT mice, fed either a HFD or LFD, and showed that the animals had similar maximal lifespans (Supplemental Fig. 7; Supplemental Table 3). Taken together, the results show an overall lack of organ toxicity and damage or electrolyte imbalance after chronic ghrelin antagonist treatment.
We describe the development and assessment of potent, orally bioavailable small-molecule antagonists of the GhrR. These molecules have low nanomolar affinity for the GhrR and show a high degree of selectivity against a broad panel of other GPCRs and molecular targets. GhrR antagonists were able to block both ghrelin-stimulated and spontaneous feeding in mice. Repeat oral dosing resulted in variable and transient decreases in FI and body weight during the first week of dosing. However, a consistent observation was that repeated exposure of HFD-fed mice to these GhrR antagonists for 7 or more days resulted in dramatic improvements in insulin sensitivity that was accompanied by significantly decreased insulin secretion, independent of effects on body weight. Collectively, these results show that pharmacological GhrR inhibition recapitulates the phenotype of GhrR KO mice fed a HFD (Wortley et al., 2004, 2005; Zigman et al., 2005; Longo et al., 2008; Qi et al., 2011).
Experimental evidence gathered over the last several years has shown that ghrelin, an acylated peptide secreted by the stomach, is a unique regulator of energy balance and metabolic function. Ghrelin acts centrally as a potent orexigen in rodents and humans and promotes positive energy balance and weight gain when administered chronically (Tschöp et al., 2000; Nakazato et al., 2001; Wren et al., 2001; Cowley et al., 2003). Conversely, both ghrelin and GhrR KO mice had lower body weight compared with controls, after several weeks of high-fat diet feeding (Wortley et al., 2005; Zigman et al., 2005; Longo et al., 2008). Lower total FI and equal total energy expenditure in GhrR KO mice has confirmed that these animals were in negative energy balance (Zigman et al., 2005; Longo et al., 2008). It is noteworthy that ghrelin KO mice had similar total FI and total energy expenditure, suggesting that there was a net loss of energy through decreased nutrient absorption in these animals (Wortley et al., 2005). In fact, GhrR KO mice have also demonstrated higher fecal excretion of lipid and a lower intestinal absorption/secretion rate of dietary lipid, indicating that a loss of ghrelin signaling limits the rate of systemic energy absorption (Longo et al., 2008). Ghrelin also promotes gastric acid secretion in the stomach, gastric motility, and gastric emptying (Masuda et al., 2000; Levin et al., 2005, 2006; Charoenthongtrakul et al., 2009). Specific antagonism of the GhrR with [d-Lys-3]-GHRP-6 was shown previously to lower the rate of gastric emptying in mice (Asakawa et al., 2003). A delayed systemic exposure to ingested nutrients might confer some protection against postprandial hyperglycemia in these mice, in a manner similar to that which has been described for glucagon-like peptide-1, or its analog, Exenatide (Wettergren et al., 1993; Willms et al., 1996; Kolterman et al., 2003). Transgenic mouse studies have also revealed the potential therapeutic benefit of blocking ghrelin signaling in improving insulin sensitivity (Wortley et al., 2004, 2005; Zigman et al., 2005; Longo et al., 2008; Qi et al., 2011). It has therefore been proposed that loss of ghrelin signaling or antagonism of the GhrR may prevent the development of diabetes in mice and humans. Results of pharmacological studies with ghrelin antagonists in HFD-fed mice presented here substantiate this hypothesis.
The GhrR antagonist treatment-induced improvements in glucose homeostasis were also confirmed by hyperglycemic clamp experiments (Fig. 4). These studies revealed that although there were no initial differences in glucose-stimulated insulin release in HFD-fed mice dosed for 7 days with CpdB or vehicle CpdB-treated mice required significantly less insulin to sustain the clamped glucose levels. The glucose infusion rates were also significantly higher in GhrR antagonist-treated mice. Thus, the clamp data recapitulated the observations made during the GTTs of compound-treated mice as well as similar clamp analyses of HFD-fed GhrR KO mice (Qi et al., 2011).
To determine the long-term effects of chemical antagonism of the GhrR, we performed extended dosing with CpdB for 56 days. We noted again that the effects of this antagonist on FI were transient and restricted to the first 6 days of dosing (Fig. 5A). Despite a return to control FI levels after 7 days, CpdB-treated mice had a lower body weight that was sustained throughout the dosing period (Fig. 5B). In response to a GTT, mice dosed with CpdB for 28 and 56 days (Fig. 5, C and E, respectively) showed a significant improvement in glucose disposal that was associated with remarkable reductions in glucose-induced insulin secretion (Fig. 5, D and F). These results provide further demonstration that extended CpdB treatment promoted a much greater insulin sensitivity in these animals. If antagonist treatment resulted in lower insulin secretion without also dramatically improving insulin sensitivity the glucose excursions in these mice would be expected to dramatically increase, not decrease. The pancreatic islets from the CpdB-treated mice were similar in terms of their morphology and insulin content compared with control mice (Fig. 6). GhrR antagonist treatment also had no effect on the incretin response or a secretagogue challenge with repaglinide, further substantiating that the pancreas of these animals was not compromised by antagonist treatment.
Long-term dosing with CpdB had several other metabolic benefits. Antagonist-treated mice had lower %HbA1c relative to controls, consistent with their improved insulin sensitivity (Table 1). Plasma triglycerides and TC were lowered significantly, with a trend of decreased hepatic very-low-density lipoprotein production. Furthermore, the TC/HDL-C ratio was reduced significantly in the compound-treated mice. The collective reversal of diet-induced dyslipidemia in these mice suggests that GhrR antagonism may have cardioprotective benefits. Perhaps most striking was the dramatic decrease in hepatic lipid, steatosis, and markers of liver dysfunction such as alanine aminotransferase and aspartate aminotransferase (Table 1; Supplemental Table 2; Fig. 6, A and B). Although we did not explore directly insulin signaling in the liver, the inverse relationship between steatosis and hepatic insulin sensitivity has been well established (den Boer et al., 2004). In turn, the body weight loss observed in the antagonist-treated animals probably is attributed to reductions in both liver and adipose tissue mass, but not skeletal muscle mass (Table 1). Thus, chemical antagonism of the GhrR limited the accumulation of adipose tissue and fatty liver, which would normally contribute to peripheral insulin resistance, emphasizing the point that blockade of ghrelin signaling has more widespread effects beyond energy, insulin, and glucose homeostasis.
The duration and timing of GhrR antagonist dosing was critical to the insulin-sparing effect of these compounds. GhrR antagonist dosing for 7 or more days was shown repeatedly to improve insulin sensitivity. However, mice that did not receive their final morning dose of CpdB before a GTT had glucose-induced insulin levels that were intermediate between those of the vehicle control mice and the mice that did receive their final dose (Supplemental Fig. 5A). Therefore, sustained tonicity in terms of chemical antagonism of the GhrR may be required to maintain the insulin sparing effect. On the other hand, an acute dose of the antagonist CpdB 1 h before a GTT led to higher glucose-induced insulin secretion (Supplemental Fig. 5B). This result is consistent with the observations that the GhrR antagonist [d-Lys3]-GHRP-6 stimulated glucose-induced insulin release during a GTT (Dezaki et al., 2004) or in perfused pancreas (Dezaki et al., 2006), whereas the acute administration of ghrelin opposed glucose-induced insulin release (Dezaki et al., 2004, 2006; Sun et al., 2006). In contrast, chronic administration of [d-Lys3]-GHRP-6 was shown to decrease both blood glucose and plasma insulin levels, whereas chronic administration with ghrelin had the opposite effect (Asakawa et al., 2003).
Although it remains unclear why the effects of chronic versus acute GhrR antagonism on glucose-induced insulin secretion differ, these results highlight the complex role of ghrelin signaling in islet function (reviewed in Dezaki et al., 2008). It is important to note that although a previous study indicates that ghrelin KO mice had lower glucose and higher insulin during a GTT the mice used in that experiment were 8 weeks old and fed a low-fat chow diet (Sun et al., 2006). However, ghrelin KO mice that were fed HFD for several weeks had lower fasted glucose and insulin (Wortley et al., 2005). In addition, HFD-fed GhrR KO mice were euglycemic and had lower insulin during a GTT (Longo et al., 2008). Based on these results, it is clear that diet macronutrient composition affects significantly the role of ghrelin signaling in the regulation of insulin secretion. Given that some of the highest levels of ghrelin receptor are found in hypothalamic regions involved in energy regulation (Zigman et al., 2006), we postulate that genetic deletion or chronic chemical antagonism of the GhrR has centrally mediated effects on both peripheral insulin sensitivity and islet function that are separable from, and opposite to, their acute or localized effects in the pancreas.
In summary, these experiments inform us of several things: 1) continuous exposure of a GhrR antagonist is an effective means of improving a broad range of pathologic features of metabolic syndrome, such as insulin resistance, obesity, and hepatic steatosis; 2) the effects of these pharmacologic antagonists are not acutely therapeutic and require continuous dosing to maintain benefits; 3) the reduced insulin secretion upon extended exposure to these GhrR antagonists is not the result of pancreas dysfunction; 4) decreased body weight is not required for the observed improvements in insulin sensitivity; and 5) blockade of ghrelin receptor signaling is safe and well tolerated in metabolically stressed animals. Furthermore, deletion of the GhrR does not compromise lifespan regardless of whether animals are given a HFD or LFD, supporting the safety of blocking this mechanism for extended periods. Therefore, the pharmacologic GhrR antagonists described here are able to improve glucose disposal while decreasing glucose-stimulated insulin secretion and may therefore represent an insulin-sparing therapeutic strategy for treating type 2 diabetes.
Participated in research design: Longo, Govek, Nolan, Hixon, Kelder, Kopchick, Saunders, Navia, Curtis, DiStefano, and Geddes.
Conducted experiments: Longo, Govek, Nolan, McDonagh, Charoenthongtrakul, Giuliana, Morgan, Hixon, Zhou, Kelder, Navia, and Geddes.
Contributed new reagents or analytic tools: Saunders.
Performed data analysis: Longo, Govek, Nolan, McDonagh, Giuliana, Hixon, Kelder, Kopchick, Saunders, Navia, Curtis, DiStefano, and Geddes.
Wrote or contributed to the writing of the manuscript: Longo, Saunders, DiStefano, and Geddes.
We thank Tim Morrison for his superb oversight of Elixir's animal facility.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- ghrelin receptor
- food restricted
- food intake
- glucose tolerance test
- high-fat diet
- low-fat diet
- pair fed
- wild type
- total cholesterol
- G protein-coupled receptor
- phosphate-buffered saline
- bovine serum albumin
- compound D
- compound B
- percentage of glycated hemoglobin
- homeostasis model assessment of insulin resistance
- hematoxylin and eosin
- growth hormone-releasing peptide-6
- high-density lipoprotein cholesterol
- area under the curve
- Received May 10, 2011.
- Accepted July 19, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics