| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Bayer Research Center, Bayer Healthcare, West Haven, Connecticut 06516
Address all correspondence and requests for reprints to: William Esler, Department of Cardiovascular and Metabolic Diseases, Pfizer Global Research and Development, MS 8220-3336, Eastern Point Road, Groton, Connecticut 06340. E-mail: william.esler{at}pfizer.com; or Joachim Rudolph, Genentech Inc., Medicinal Chemistry, 1 DNA Way, South San Francisco, California 94080. E-mail: rudolph.joachim{at}gene.com.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The peptide hormone ghrelin plays a central role in regulation of energy homeostasis (5). Ghrelin signaling through its receptor, type 1a GH secretagogue receptor (GHS-R1a) acts as a meal initiator and decreases energy expenditure and fat utilization (5). A variety of experimental data support the role of ghrelin and GHS-R1a in meal initiation. Plasma ghrelin levels peak before meal initiation and decrease shortly after food consumption in humans (6, 7, 8). Fasting dramatically increases plasma ghrelin levels (9, 10, 11, 12, 13), whereas overeating over a period of weeks blunts plasma levels of the hormone (13, 14). Additionally, ghrelin administration leads to a robust but transient increase in food consumption in both rodents (9, 15, 16) and humans (17). Consistent with its proposed role as a nutrient sensor, caloric intake but not simple distention of the stomach suppresses plasma ghrelin levels (18).
A growing body of data supports the hypothesis that ghrelin, in addition to its other roles in modulating metabolism, directly regulates glucose homeostasis through action on GHS-R1a in the pancreas. Ghrelin and GHS-R1a are both expressed in pancreas and islets (19, 20, 21, 22, 23, 24, 25). Ghrelin has been shown to suppress insulin secretion in vitro (24, 26) and in vivo and causes an increase in plasma glucose both in rodents (24, 27) and humans (28, 29, 30, 31, 32, 33). Furthermore, chronic treatment with small-molecule GHS-R1a agonists induces hyperglycemia and insulin resistance in humans (34, 35, 36). Recently, GHS-R1a deficiency in ob/ob mice was shown to improve glucose tolerance and enhance insulin secretion (37). These observations taken together suggest that GHS-R1a antagonists may have utility for the treatment of type 2 diabetes. To explore the potential value of GHS-R1a antagonists for regulation of glucose homeostasis, we undertook evaluation of a novel class of potent and selective small-molecule GHS-R1a antagonists identified in our laboratories. These compounds promote glucose-dependent insulin secretion likely by direct action on GHS-R1a in the pancreas and suppress food intake and promote body weight loss through selective loss of fat mass.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Synthesis of GHS-R1a antagonists
Materials and Methods for the synthesis of YIL-781 and YIL-870 can be found in the supplemental material (published as supplemental data on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org).
GHS-R1a pharmacology
Affinity for the GHS-R1a was determined using a radioligand binding assay. Test compound and 50 pM [125I]ghrelin were mixed in buffer [25 mM HEPES (pH 7.4), 5 mM magnesium chloride, 1 mM calcium chloride, 1 mM EDTA, 0.1% BSA). A suspension of membrane from HEK293S cells overexpressing ovine GHS-R1a was preincubated for 30 min with 5 mg/ml wheat germ agglutinin SPA beads and then added to the wells containing test compound and [125I]ghrelin. Nonspecific binding was defined using 1 µM ghrelin peptide. After a 2-h incubation at room temperature, the amount of [125I]ghrelin bound to the membranes was measured.
Functional activity of compounds was measured using GTP
S binding to membranes prepared from HEK293S cells overexpressing recombinant GHS-R1a. GHS-R1a agonism was measured as stimulation of GTPS binding above basal levels by compound. Compounds (tested at 10 µM) that failed to stimulate GTPS binding over basal levels (E/Emax < 10%) were evaluated for their ability to serve as antagonists. Ghrelin dose-response curves were performed in the absence and presence of a fixed concentration of the putative antagonist. Membrane from HEK293S cells overexpressing GHS-R1 was mixed with test compound and/or ghrelin in buffer (25 mM HEPES, 50 mM sodium chloride, 10 mM magnesium chloride, 1 µM GDP, and 0.1% BSA, pH 7.4) and 200 pM [35S]GTPS. After a 60-min room temperature incubation, bound and free [35S]GTPS were separated by filtrating through GF/B membranes. After multiple washes, the amount of [35S]GTPS on the filtered membranes was determined. Nonspecific binding was defined in the presence of 10 µM GTPS.
In vitro insulin secretion assay
Islets of Langerhans were prepared from the pancreas of male Sprague Dawley rats by collagenase digestion and isolation on a Ficoll gradient. Dispersed islet cells prepared by treatment of the islets with trypsin, were seeded into 96-well V-bottom plates and cultured overnight in media (38, 39). After overnight incubation, cells were incubated with Krebs-Ringer-HEPES buffer containing 3 mM glucose in the presence or absence of ghrelin, des-acyl ghrelin (10–12 to 10–6 M), and/or YIL-781 (1 µM) for 30 min at 37 C. The cells were subsequently challenged with 20 mM glucose (30 min at 37 C). Insulin content in the supernatant was measured via antiinsulin scintillation proximity assay.
Glucose tolerance tests in rats
Male Wistar rats were fasted overnight (16–18 h) and then given ghrelin antagonist, liraglutide (40) (positive control), or vehicle (polyethylene glycol/10 mM methanesulfonic acid 80:20) by oral gavage. Five hours after dosing, the fasting blood glucose level was measured from tail-tip blood using a Glucometer (Bayer Corp., Mishawaka, IN), and the animals were given 2 g/kg glucose by ip injection (ip glucose tolerance test or IPGTT). Blood glucose was measured again after 15, 30, and 60 min for the IPGTT. Additional blood glucose measurements were made after 90 and 120 min for the oral glucose tolerance test or after 90 min when an IPGTT was performed with diet-induced obese (DIO) rats. The area under the glucose curve (AUC) from 0–60, 0–90, or 0–120 min was calculated using the trapezoidal method, and the effect of the compound on the AUC was expressed as a percentage of the AUC for the vehicle-treated group. When insulin levels were measured, 0.05 ml blood was collected in a capillary tube at the same time periods as described. Insulin was determined using an ELISA (Alpco Diagnostics, Windham, NH). For the IPGTT experiment using DIO rats, male Wistar rats were fed a high-fat diet (D12451 from Research Diets containing 45% calories from fat, 35% from carbohydrate, and 20% from protein) upon arrival for a period of 4–6 wk. Before compound evaluation, the rats were prescreened for elevated glucose AUC in an IPGTT and grouped to ensure homogeneity in the test animals. For compound evaluation, the IPGTT experiment was run the same as with lean rats except that the glucose excursion was measured over 90 min.
Insulin tolerance test in rats
Fed male Wistar rats were given vehicle (polyethylene glycol/10 mM methanesulfonic acid 80:20) or 0.3 mg/kg YIL-781, and food was withheld from the rats 2 h later. Five hours after dosing, the rats were given an ip dose of vehicle (PBS), 0.2 U/kg insulin, or 1 U/kg insulin. Blood glucose was measured from tail-tip blood just before the ip dose and 10, 20, 30, 40, and 50 min afterward. The change in blood glucose from time 0 was calculated, and the slope of the change in blood glucose vs. time curve was determined.
Assessment of gastric emptying in mice
Vehicle or YIL-870 was administered by oral gavage to fasted male C57BL6 mice. Five hours after dosing, 1 ml of a semisolid test meal (10% charcoal, 5% gum arabic, and 1% carboxymethylcellulose) was given by oral gavage. The mice were water-restricted 1 h before the meal, and 0.5 h after the test meal, the mice were euthanized with CO2 gas and a laparotomy was performed. The stomach was removed, quickly weighed, opened and the contents removed, and weighed again. The difference between the initial stomach weight and the final weight is the weight of the meal remaining in the stomach. Exendin-4 (3 µg/kg), which is known to delay gastric emptying (41), was used as positive control and was given by sc injection 20 min before the test meal. To rule out effects on gastric secretion, experiments were performed identically with the exception that the test meal was not administered.
DIO mouse model
Male C57BL/6 mice were fed a high-fat diet (D12451 from Research Diets containing 45% calories from fat, 35% from carbohydrate, and 20% from protein) for 16 wk before the start of the studies. The animals included in the studies had an average body weight greater than 4 SD of the mean body weight of mice that were fed a standard low-fat diet (5% calories from fat). After 16 wk on the high-fat diet, GHS-R1a antagonists, rimonabant, or vehicle was administered orally approximately 0.5–1.0 h before the feeding phase (dark cycle). Rimonabant was selected as a positive control for comparison with our GHS-R1a antagonists because it promotes body weight via an anorexigenic mechanism (42). Daily body weight and food intake were recorded during treatment. Fat and lean masses were measured in each animal at the start and end of the studies using NMR imaging with a Bruker Minispec instrument. Body composition was assessed while animals were conscious. Pair-feeding was performed by measuring the 24-h food intake of compound-treated mice each day and limiting food consumption to that amount in a group of vehicle-treated mice.
Mouse fasted refed model
Male C57BL/6 mice (two mice per cage) were fasted overnight and dosed the following day, 1 h before refeeding. Cumulative food intake was recorded for a period of 24 h after food return.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), a piperidine-substituted quinazolinone derivative, was evaluated for the ability to bind to GHS-R1a in membranes from cells overexpressing recombinant GHS-R1a. The compound was found to be a potent GHS-R1a ligand with a Ki of 17 nM (Fig. 1B). To determine whether YIL-781 functioned as an agonist or antagonist at GHS-R1a, the ability of the compound to stimulate [35S]GTPS binding was examined (Fig. 1C). Unlike ghrelin, which stimulated a dose-dependent activation of GHS-R1a, YIL-781 showed no propensity to activate GHS-R1a. YIL-781 was found to right-shift the ghrelin dose-response curve and thus acts as a competitive GHS-R1a antagonist with a Kb of 11 nM.
|
10 µM; see supplemental Table 1).
The effect of ghrelin on insulin secretion in vitro was explored using a dispersed rat islet insulin secretion assay (Fig. 1
, D–F) (38, 39). The response of rat dispersed islets to a 20 mM glucose challenge was measured in the presence or absence of ghrelin ranging in concentration from 10–12 to 10–7 M (Fig. 1D). Ghrelin produced a dose-dependent suppression of insulin secretion with an IC50 of 0.9 nM and a maximal effect of –23%.
The posttranslational octanoylation of des-acyl ghrelin at Ser-3 to produce mature ghrelin is a requirement for the peptide to bind and activate GHS-R1a (44). Des-acyl ghrelin, however, is not devoid of biological activity and has recently been shown to mediate a variety of functions through an unidentified receptor distinct from GHS-R1a (43, 44). To ascertain whether the effects of ghrelin on insulin secretion are consistent with the pharmacology of GHS-R1a, ghrelin and des-acyl ghrelin were compared for their effects on insulin secretion (Fig. 1E). Although ghrelin produced 19 and 23% suppression in insulin secretion when tested at 10 and 100 nM, respectively, des-acyl ghrelin had no significant effect even at concentrations of 1 µM.
We then examined the ability of the selective GHS-R1a antagonist YIL-781 to block the effect of ghrelin on insulin secretion. Although 10 nM ghrelin produced a 25% suppression of insulin secretion in response to a 20 mM glucose challenge, coincubation with 1 µM YIL-781 fully blocked this effect. Under these conditions, 1 µM YIL-781 had no effect in the absence of ghrelin (Fig. 1F).
The effects of YIL-781 on insulin secretion and glucose homeostasis in vivo were evaluated using an IPGTT in rats (Fig. 2, A–E). YIL-781 or vehicle was dosed orally in fasted male Wistar rats at 10 mg/kg. Five hours after dosing, a glucose challenge (2 g/kg) was administered by ip injection. Blood glucose was measured immediately before glucose challenge (defined as time zero) and at various time points over 1 h after the glucose challenge (Fig. 2A). YIL-781 did not affect blood glucose levels at the fasted stage (time zero). However, upon glucose administration, the compound caused a 23% decrease in the glucose excursion relative to the vehicle-treated animals. This effect was similar in magnitude to the effect of parenterally administered NN2211 (liraglutide), a long-lasting incretin mimetic in clinical development for the treatment of diabetes (40). To further explore the in vivo efficacy of YIL-781 in this model, a dose-response experiment was performed (Fig. 2B). YIL-781 was found to promote significant reductions in glucose excursion ranging from 0.3 mg/kg to the highest dose tested (10 mg/kg). At the minimal efficacious dose of 0.3 mg/kg, the plasma concentration of YIL-781 was found to range between 1 and 9 nM (supplemental Table 2 and supplemental Fig. 1). These concentrations are in line with the potency of YIL-781 for GHS-R1a (Kb = 11 nM).
|
, C–E). At the minimal efficacious dose of 0.3 mg/kg, YIL-781 trended to increase the insulin levels (Fig. 2C) and insulin-glucose ratio (Fig. 2D) over the time course of the excursion. The effect on insulin glucose ratio reached significance at the 1-h time point (Fig. 2D). Overall, the compound produced a 54% increase (P < 0.05) in insulin glucose ratio over the course of the IPGTT (Fig. 2E). The observation that the compound increased insulin secretion in vivo is consistent with the in vitro results indicating that YIL-781 could fully overcome the inhibitory effects of ghrelin on insulin secretion. The acute impact of YIL-781 on insulin sensitivity was evaluated using an insulin tolerance test model (supplemental Fig. 2). At a dose of 0.3 mg/kg, YIL-781 did not alter the effect of a submaximal dose of insulin (0.2 U/kg) on blood glucose levels. This result, in combination with the effect of the compound on insulin secretion, demonstrates that, at least acutely, the GHS-R1a antagonist YIL-781 improves glucose tolerance by promoting insulin release rather than enhancing insulin sensitivity.
One limitation of many insulin secretagogues is the propensity to cause hypoglycemia, i.e. promote insulin secretion even at the state of low blood glucose. We therefore examined the effect of 30 mg/kg YIL-781 on blood glucose levels in fasted animals (Fig. 2F). In contrast to nateglinide, which caused a dramatic reduction in fasting blood glucose, YIL-781 at 100 times its efficacious dose did not cause any physiologically relevant changes in blood glucose over the 5-h experiment. Comparison of the glucose AUCs indicates that YIL-781 was indistinguishable from the vehicle control, whereas nateglinide at a 100 mg/kg dose produced a 19% decrease.
To evaluate whether GHS-R1a antagonists could improve glucose tolerance in a disease model, YIL-781 was tested in the insulin-resistant DIO rat (Fig. 2G). A 3 mg/kg oral dose of YIL-781 was found to cause a 19% reduction in glucose excursion in this model.
In addition to its role in regulating insulin secretion, there is compelling evidence that ghrelin plays a critical role in modulating body weight through control of food intake and/or regulation of fuel substrate efficiency (5). To examine the effect of YIL-781 on body weight, the compound (at 3, 10, and 30 mg/kg) or vehicle was administered by oral gavage to DIO mice once daily for 9 d (Fig. 2H). The highest-dose group (30 mg/kg) produced a 5% decrease in body weight. Because ghrelin has been hypothesized to regulate feeding through a central mechanism, brain concentrations of YIL-781 were determined in pharmacokinetic studies and were found to be 20% of plasma.
A second GHS-R1a antagonist, YIL-870 (Fig. 3A), was also examined in these animal models. As observed for YIL-781, YIL-870 significantly reduced glucose excursion in the IPGTT model when dosed orally at 10 mg/kg (Fig. 3B). The magnitude of glucose reduction was somewhat lower with YIL-870 (–17%) relative to YIL-781 (–23%) at this dose, likely resulting from the difference in GHS-R1a potency between the two compounds.
|
), the compound caused about a 15% drop in body weight relative to vehicle-treated mice. As was observed for the cannabinoid 1 (CB1) receptor antagonist rimonabant (42), a rapid decrease in body weight occurred over the first few days. However, although the body weight progression curve of the rimonabant-treated animals began to parallel the vehicle-treated group after d 4, the group treated with YIL-870 continued to diverge from the vehicle-treated group (albeit at a slower rate than initially) throughout the course of the experiment.
Food intake was measured daily throughout the course of the experiment (Fig. 3D). Rimonabant produced the typical anorexigenic pattern of reduced food intake; food intake dropped dramatically on the first few days of the study, then quickly rebounded and was superimposable with vehicle for the remainder of the study. In contrast, the GHS-R1a antagonist produced a more complex pattern. YIL-870 produced a rapid drop in food intake. However, the effect was more sustained and did not reach a minimum until d 10. After d 10, the reduction in food intake diminished in magnitude but never recovered to the levels observed in the vehicle group. The fact that food intake for YIL-870-treated animals never rebounded to the levels seen with the vehicle-treated mice likely indicates why body weight remained divergent between the two groups.
Comparison of the body composition of the animals at d 0 and 20 of this study revealed that YIL-870 produces a dramatic reduction in fat mass and no reduction in lean mass (Fig. 3E). Although rimonabant also reduced fat mass, the magnitude of the reduction was only roughly 25% of that observed with the GHS-R1a antagonist.
A dose-response experiment with YIL-870 revealed that its effect on body weight in DIO mice was both dose dependent and roughly dose proportional (Fig. 3F). To examine the relative contribution of reduced food intake to the observed weight loss, a group pair-fed to the highest dose (10 mg/kg) was also included. Pair-feeding produced a nearly superimposable weight-loss curve relative to the GHS-R1a antagonist-treated group, especially at the early time points. From d 6–10, there was a slight trend toward greater weight loss in the compound-treated group, but this did not reach significance (Fig. 3F). These results indicate that reduction in food intake is the primary root cause for the body weight loss affected by the GHS-R1a antagonist YIL-870.
The brain exposure of YIL-870 was also examined. When dosed in mice, YIL-870 was found to have significantly greater brain exposure than YIL-781 (supplemental Fig. 3).
The effect of YIL-870 on gastric emptying was evaluated in a mouse gastric emptying model (supplemental Fig. 4). At the highest dose tested (10 mg/kg), YIL-870 caused a 23% increase in stomach contents remaining 30 min after administration of the test meal. The magnitude of this effect was comparable to the positive control the glucagon-like peptide-1 analog exendin-4 (3 µg/kg), which is known to delay gastric emptying (41). Lower doses of YIL-870 had no effect on gastric emptying.
To help examine the role of central nervous system (CNS) exposure for the effects of ghrelin antagonists on food intake, the effects of a series of potent GHS-R1a antagonists with comparable plasma exposure but differing brain exposure was examined in the mouse fasted-refed model (Table 1). Although compounds with appreciable brain exposure produced significant decreases in 24-h food consumption, a potent antagonist with no detectable brain exposure had no impact on food intake despite high plasma exposure. A compound with modest brain exposure produced a weak response in this model. In contrast, all of the compounds tested demonstrated efficacy in a glucose tolerance test model (Table 1).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Numerous laboratories have demonstrated that ghrelin suppresses insulin secretion in vitro from static preparations of rodent islets (45, 46), pancreatic cell lines (45, 47), and perfused rat pancreas (26). We confirmed these findings in our lab using a dispersed rat islet insulin secretion assay amenable for use in drug discovery (38, 39). In this assay, rat pancreatic islets are partially purified, dispersed into cells and then reaggregated into pseudo-islets in 96-well plates. Insulin response to glucose stimulation is then measured using an antiinsulin scintillation proximity assay. This assay has been shown to retain the response of whole islets for a variety of secretagogues but has higher throughput and lower variability (38, 39).
As has been previously reported for static islet incubations (45, 46), ghrelin dose-dependently suppressed glucose-stimulated insulin secretion in the dispersed islet assay with an IC50 of 0.9 nM. This potency is consistent with the Ki of ghrelin for GHS-R1a (0.2–2 nM) and corresponds with literature data for ghrelin-mediated suppression of insulin secretion in static islets (45, 46) and in Min-6 cells (47).
Des-acyl ghrelin, which has no affinity to GHS-R1a (48), has been shown to share some of the biological effects of ghrelin, leading in turn to the hypothesis that ghrelin may exert some of its activities through unidentified receptors distinct from GHS-R1a (43, 44). The present data, however, suggest that the impact of ghrelin on insulin secretion is mediated by GHS-R1a. Unlike ghrelin, which dose-dependently suppressed insulin secretion, des-acyl ghrelin at equal or 10-fold higher concentrations had no impact on insulin secretion. Furthermore, the potent and selective GHS-R1a antagonist YIL-781 fully blocked the effect of ghrelin on insulin secretion but had no impact in the absence of the peptide.
The observation that YIL-781 had no effect on insulin secretion in the absence of added ghrelin conflicts with observations from one group that reported that a peptidic GHS-R1a antagonist suppressed insulin secretion in the absence of exogenous ghrelin (24). The authors interpreted these results as an indication that the peptidic GHS-R1a antagonist was blocking the effect of endogenous ghrelin produced by islets. There are multiple possible explanations for this discrepancy. Although mature ghrelin was produced endogenously in the dispersed islet assay, the total ghrelin content in our assay (0.005–0.01 fmol per islet) is almost certainly too low to yield concentrations in the assay well required to activate GHS-R1a (IC50 = 0.5–2 nM). This value is consistent with the reported ghrelin content of islets (0.004 fmol per islet) (24). Thus in this system, endogenous ghrelin produced may not exert sufficient tone for an antagonist to show an effect in the absence of exogenous ghrelin. It is also possible that differences in assay conditions between our experiments and those of Dezaki et al. (24), such as glucose concentration, may contribute to this discrepancy. In the present study, the inhibitory effects of ghrelin on insulin secretion were measured under conditions where insulin secretion was maximally stimulated by glucose. Dezaki et al. (24) examined the effects of GHS-R1a antagonists at half-maximal glucose concentrations.
GHS-R1a antagonists were also evaluated for their impact in vivo. YIL-781 and YIL-870 were found to improve glucose tolerance in the IPGTT model in lean Wistar rats and insulin-resistant DIO rats. Because a single oral administration of the antagonist in fasted animals was sufficient to improve glucose tolerance, GHS-R1a antagonists appear to have a direct impact on glucose homeostasis impendent of the potential additional benefits that may arise from chronic dosing. In addition, the observation that GHS-R1a antagonists improve glucose tolerance in the IPGTT model where glucose is administered parenterally would argue that the glucose-lowering effects are not a consequence of delayed gastric emptying reducing glucose absorption.
Consistent with the observation that YIL-781 blocked the inhibitory effect of ghrelin on glucose-stimulated insulin secretion in vitro, the compound appears to at least acutely improve glucose homeostasis by promoting insulin secretion in vivo. At the minimal efficacious dose of 0.3 mg/kg, YIL-781 stimulated insulin secretion in the rat IPGTT model but had no acute effect on insulin sensitivity as assessed in an insulin tolerance test. However, we cannot exclude that chronic treatment would also lead to improved insulin sensitivity. Indeed, improvements in insulin sensitivity would be expected given the weight loss observed in DIO mice after chronic dosing with GHS-R1a antagonists (see below).
Many insulin secretagogues such as nateglinide show a propensity to cause hypoglycemia. To ascertain whether ghrelin receptor antagonists cause hypoglycemia, the effect of YIL-781 on fasting blood glucose was examined. YIL-781 at 30 mg/kg (100 times the minimal efficacious dose in the rat IPGTT) showed no tendency to cause hypoglycemia.
The ghrelin receptor antagonists YIL-781 and YIL-870 were also evaluated for effects on body weight in DIO mice. After once-daily oral administration, both compounds produced weight loss in the DIO mouse model. However, the dose of the compound required for promoting weight loss and the magnitude of the effect was dramatically different for the two compounds. At the highest dose tested (30 mg/kg), YIL-781 led to a 5% decrease in body weight over 9 d. Although plasma levels of the compound were roughly comparable in rats and mice, the two lower doses (3 and 10 mg/kg), which had promoted significant improvements in glucose homeostasis, failed to produce any effects on body weight. In contrast, YIL-870 produced dramatic effects on weight loss. At 10 mg/kg, YIL-870 produced a 15% reduction in body weight over the course of the study. In addition, the 3 mg/kg dose also appeared efficacious. A group pair fed to the 10 mg/kg group indicated that all of the weight loss could be attributed to the anorectic effect of the compound. However, we cannot rule out that over a longer time course an energy expenditure component could contribute to additional weight loss. Exploration of the effect of YIL-870 on body composition after the weight loss study revealed that weight loss could be attributed to selective loss of fat mass with no decrease in lean mass.
Because ghrelin has been hypothesized to mediate its effects on feeding by acting on the hypothalamus (5), one likely explanation for the difference in weight loss efficacy and potency observed between YIL-781 and YIL-870 could relate to differences in CNS penetration. YIL-870 (Fig. 3A) has decreased polar surface area and reduced molecular flexibility compared with YIL-781 (supplemental Table 3), parameters important for CNS penetration (49). Although plasma concentrations of the two compounds upon oral dosing were similar, brain levels of YIL-870 were considerably higher.
Also consistent with this hypothesis, in a series of potent GHS-R1a antagonists with comparable plasma exposure but differing brain exposure, the ability to suppress feeding appeared to roughly correlate with brain exposure. Because pair-feeding experiments had suggested suppressed feeding was the immediate cause of the weight loss observed with YIL-870, weight loss effects are likely to be CNS driven.
Interestingly, when the test compounds with varying brain exposure were evaluated in the IPGTT model, all of the compounds were active regardless of apparent CNS penetration. This result is consistent with in vitro studies demonstrating that GHS-R1a antagonists had a direct effect on insulin secretion in islet preparations. Thus, antagonizing ghrelin action on GHS-R1a in pancreatic islets is likely responsible for increased insulin secretion, whereas blocking ghrelin activity in the CNS likely drives the impact on body weight. Consequently, it may be possible to functionally separate the effects of GHS-R1a antagonists on insulin secretion and food intake by modulating CNS exposure. Additional experiments including intracerebroventricular administration of non-CNS-penetrating compounds would be required for a more detailed investigation of this hypothesis.
Ghrelin has been reported to have prokinetic (i.e. accelerating) effects on gastric emptying in animals (10) and humans (50). The magnitude of this effect is comparable to that observed with the glucagon-like peptide-1 analog exendin-4 (3 µg/kg) and is consistent with a moderate delay in gastric emptying. At 3 mg/kg, the GHS-R1a antagonist YIL-870 did not produce a significant increase in stomach contents in this model despite the fact that this dose caused 4% decrease in body weight and decreased food intake in the DIO mouse model. Thus, although the exact contribution of delayed gastric emptying to the anorexigenic effects of GHS-R1a antagonists is not clear, it is apparently not an absolute requirement. Given the nontraditional pattern of food intake inhibition observed at the 10 mg/kg dose, it is tempting to speculate that this pattern is a composite of an anorectic effect and delayed gastric emptying. An anorectic effect could account for the dramatic and transient suppression of food intake, whereas the impact of delayed gastric emptying could account for the sustained moderate reduction in food intake.
One of the initial concerns raised about the therapeutic potential of GHS-R1a antagonists is the lower plasma ghrelin level observed in obese individuals (51). Like obese humans, DIO rodents have suppressed plasma ghrelin levels (52). In the present study, GHS-R1a antagonists were shown to both improve glucose tolerance and promote weight loss in DIO rodents, indicating that sufficient ghrelin tone is present in the obese state for GHS-R1a antagonists to have a therapeutic effect.
Recently, mouse models deficient in either ghrelin peptide (37) or GHS-R1a (53) have been shown to have improved glucose tolerance when crossed into the leptin-deficient ob/ob mouse model or challenged with a high-fat diet, respectively. In the case of the former model, improvements in glucose homeostasis were at least in part attributed to an enhancement of insulin secretion (37). In contrast to the present studies, however, some of these genetic models did not show the impact on food intake (28, 37, 54) predicted by the numerous studies implicating ghrelin as a key orexigenic hormone (5). The observation that some ghrelin or GHS-R1a knockout models do not show an anorexigenic phenotype may result from compensatory mechanisms regulating hypothalamic control of feeding. Consistent with this hypothesis, targeted ablation of the hypothalamic neuropeptide Y/agouti-related protein axis in neonatal mice has no impact on feeding (55). In contrast, disruption of this feeding center in adult mice leads to rapid starvation (55). This difference is almost surely a result of compensatory changes that occurred when the neuropeptide Y/agouti-related protein neurons were disrupted in neonatal mice, which were no longer possible when the feeding nucleus was disrupted in adults (55). Because ghrelin regulates feeding through action on GHS-R1a in this population of neurons (5, 56), compensatory changes resulting from genetic ablation of either ghrelin or GHS-R1a would not be unexpected. Furthermore, in the case of ghrelin-deficient models, unintentional ablation of obestatin, which is encoded by the ghrelin gene and has been shown to suppress food intake, may further mask the potential impact of ghrelin on feeding (57).
The intricate interrelationship between type 2 diabetes and obesity coupled with the endemic surges in the incidence of both disorders suggest that an agent that can directly improve glucose homeostasis (without causing hypoglycemia) and promote weight loss through selective reductions in fat mass would have considerable clinical utility. The present study demonstrates that GHS-R1a antagonists directly improve glucose homeostasis by enhancing glucose-stimulated insulin secretion and promote weight loss through selective loss of fat mass in rodents. GHS-R1a antagonists, which enhance insulin secretion and suppress appetite, have potential to open a new front in the battle against diabesity.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online July 26, 2007
Abbreviations: AUC, Area under the glucose curve; CNS, central nervous system; DIO, diet-induced obese; GHS-R1a, type 1a GH secretagogue receptor; IPGTT, ip glucose tolerance test.
Received February 23, 2007.
Accepted for publication July 17, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-cells of humans and rats and stimulates insulin secretion. Diabetes 51:124–129This article has been cited by other articles:
 |
|
 |
|
  P. S. Petersen, D. P. D. Woldbye, A. N. Madsen, K. L. Egerod, C. Jin, M. Lang, M. Rasmussen, A. G. Beck-Sickinger, and B. Holst In Vivo Characterization of High Basal Signaling from the Ghrelin Receptor Endocrinology, November 1, 2009; 150(11): 4920 - 4930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. H. Moran and M. J. Dailey Gut Peptides: Targets for Antiobesity Drug Development? Endocrinology, June 1, 2009; 150(6): 2526 - 2530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-C. Lu, J. Xu, N. Chinookoswong, S. Liu, S. Steavenson, C. Gegg, D. Brankow, R. Lindberg, M. Veniant, and W. Gu An Acyl-Ghrelin-Specific Neutralizing Antibody Inhibits the Acute Ghrelin-Mediated Orexigenic Effects in Mice Mol. Pharmacol., April 1, 2009; 75(4): 901 - 907. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. V. Mayorov, N. Amara, J. Y. Chang, J. A. Moss, M. S. Hixon, D. I. Ruiz, M. M. Meijler, E. P. Zorrilla, and K. D. Janda Catalytic antibody degradation of ghrelin increases whole-body metabolic rate and reduces refeeding in fasting mice PNAS, November 11, 2008; 105(45): 17487 - 17492. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Lanfranco, L. Bonelli, M. Baldi, E. Me, F. Broglio, and E. Ghigo Acylated Ghrelin Inhibits Spontaneous Luteinizing Hormone Pulsatility and Responsiveness to Naloxone But Not That to Gonadotropin-Releasing Hormone in Young Men: Evidence for a Central Inhibitory Action of Ghrelin on the Gonadal Axis J. Clin. Endocrinol. Metab., September 1, 2008; 93(9): 3633 - 3639. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S. Ahima Antagonism of Ghrelin for Glycemic Control in Type 2 Diabetes Mellitus? Endocrinology, November 1, 2007; 148(11): 5173 - 5174. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
