|Home | About | Journals | Submit | Contact Us | Français|
We and others have demonstrated previously that ghrelin receptor (GhrR) knock out (KO) mice fed a high fat diet (HFD) have increased insulin sensitivity and metabolic flexibility relative to WT littermates. A striking feature of the HFD-fed GhrR KO mouse is the dramatic decrease in hepatic steatosis. To characterize further the underlying mechanisms of glucose homeostasis in GhrR KO mice, we conducted both hyperglycemic (HG) and hyperinsulinemic-euglycemic (HI-E) clamps. Additionally, we investigated tissue glucose uptake and specifically examined liver insulin sensitivity.
Consistent with glucose tolerance-test data, in HG clamp experiments, GhrR KO mice showed a reduction in glucose-stimulated insulin release relative to WT littermates. Nevertheless, a robust 1st phase insulin secretion was still achieved, indicating that a healthy β-cell response is maintained. Additionally, GhrR KO mice demonstrated both a significantly increased glucose infusion rate and significantly reduced insulin requirement for maintenance of the HG clamp, consistent with their relative insulin sensitivity. In HI-E clamps, both LFD-fed and HFD-fed GhrR KO mice showed higher peripheral insulin sensitivity relative to WT littermates as indicated by a significant increase in insulin-stimulated glucose disposal (Rd), and decreased hepatic glucose production (HGP). HFD-fed GhrR KO mice showed a marked increase in peripheral tissue glucose uptake in a variety of tissues, including skeletal muscle, brown adipose tissue and white adipose tissue. GhrR KO mice fed a HFD also showed a modest, but significant decrease in conversion of pyruvate to glucose, as would be anticipated if these mice displayed increased liver insulin sensitivity. Additionally, the levels of UCP2 and UCP1 were reduced in the liver and BAT, respectively, in GhrR KO mice relative to WT mice.
These results indicate that improved glucose homeostasis of GhrR KO mice is characterized by robust improvements of glucose disposal in both normal and metabolically challenged states, relative to WT controls. GhrR KO mice have an intact 1st phase insulin response but require significantly less insulin for glucose disposal. Our experiments reveal that the insulin sensitivity of GhrR KO mice is due to both BW independent and dependent factors. We also provide several lines of evidence that a key feature of the GhrR KO mouse is maintenance of hepatic insulin sensitivity during metabolic challenge.
Administration of exogenous acyl-ghrelin peptide causes insulin resistance in humans . Consistent with this observation, mice with genetic blockade of ghrelin signaling resist diet-induced obesity and show evidence of improved glucose homeostasis under this metabolic stress [2-5]. Pharmacologic antagonism of the ghrelin receptor suppresses appetite, promotes weight loss and improves glucose tolerance [6,7]. Overall, these studies provide compelling evidence that ghrelin and its receptor play an important role in regulation of glucose homeostasis. However, the mechanism[s] linking improved insulin sensitivity and ghrelin signaling are still under investigation.
Circulating ghrelin levels are modulated by changes in nutritional status, such as food deprivation  or exposure to HFD [9,10]. Similarly, the physiological consequences of manipulating ghrelin signaling may vary according to metabolic status. For example, Sun et al.  demonstrated that ghrelin-deficient mice fed a low-fat diet (LFD) respond to a glucose challenge with improved glucose disposal resulting from increased insulin secretion, relative to WT littermates. In contrast, under conditions of metabolic stress such as high fat diet (HFD)-feeding, ghrelin-deficient mice have improved glucose homeostasis characterized by significantly lower fasting serum insulin levels . Consistent with this latter observation, we have previously observed that HFD-fed GhrR KO mice have improved insulin sensitivity corresponding with a striking reduction in the insulin required for glucose disposal in response to a glucose challenge .
In the current series of experiments we performed hyperglycemic clamps in HFD-fed GhrR KO mice in order to examine both insulin sensitivity and the dynamics of insulin secretion in response to a glucose challenge. To further understand the mechanism by which blockade of ghrelin receptor signaling improves insulin sensitivity under conditions of metabolic stress, we also performed a hyperinsulinemic-euglycemic clamp in GhrR KO mice fed a HFD and determined tissue glucose dynamics. Our results clearly demonstrate that loss of signaling through the GhrR improves insulin sensitivity under conditions of HFD-induced metabolic stress. This insulin sensitivity is characterized by a decreased insulin requirement in the face of a glucose challenge. Additionally, our results reveal a role of hepatic insulin sensitivity in the phenotype of GhrR KO mice.
In preliminary evaluations, GhrR KO mice showed improved insulin sensitivity on both low and HFD [2,4]. The fasting blood glucose was 65.5 ± 4.5 vs. 78.4 ± 3.9 mg/dl (p < 0.05) on LFD, and 107.7 ± 3.9 vs. 118.2 ± 7.4 mg/dl on HFD in GhrR KO mice and WT mice, respectively (data not shown). The corresponding fasting plasma insulin was 0.66 ± 0.07 vs. 0.77 ± 0.08 ng/ml on LFD, and 0.76 ± 0.09 vs. 1.41 ± 0.32 ng/ml (p < 0.05) on HFD in GhrR KO and WT mice, respectively (data not shown).
To evaluate the effects of HFD on islet ß-cell function we performed the HG clamp. This assay allows for the evaluation of glucose-stimulated insulin secretion (GSIS) to an initial glucose challenge. Initial experiments comparing diet-induced obese (DIO) vs lean mice confirmed the ability of this assay to reveal the relative insulin resistance of DIO mice relative to lean controls (Figure (Figure1).1). The pre-study fasted body weight of the DIO mice (45.57 ± 0.96) was significantly greater than the chow-fed littermates (33.46 ± 1.33). Using the identical experimental paradigm, we observed no significant differences in GSIS of GhrR KO mice compared with WT animals at either 1 min (2.09 ± 0.41 vs 1.31 ± 0.15 ng/mL) or 5 min (1.88 ± 0.40 vs 1.20 ± 0.15 ng/mL) after the initial glucose challenge (Figure (Figure2A).2A). The HG clamp experiments thus highlight that in GhrR KO mice the ß-cell function is intact relative to WT controls. However, during the second phase of the experiment WT mice required more insulin to maintain steady state hyperglycemia relative to GhrR KO mice, reaching statistical significance at the 70 min (5.72 ± 1.40 vs 2.41 ± 0.32 ng/mL) and 90 min (6.14 ± 1.50 vs 2.90 ± 0.34 ng/mL) time points (Figure (Figure2B).2B). Moreover, the GIR was significantly higher in KO mice compared to WT (7.76 ± 1.65 vs 5.68 ± 1.52 mg/kg/min, Figure Figure2C),2C), indicative of increased insulin sensitivity of the GhrR KO mice. GhrR KO mice, therefore, have increased insulin sensitivity characterized by a decreased insulin requirement for glucose disposal relative to WT controls. Prior to initiating the clamp experiment, there was a statistically significant difference in fasted body weight (BW) of GhrR KO mice versus WT controls (40.99 ± 1.28 vs 45.07 ± 1.27, respectively). Thus, in this experiment the reduced BW of the GhrR KO mice will have contributed to the overall improvement in their insulin sensitivity.
We next evaluated hepatic insulin sensitivity and insulin-mediated glucose disposal HFD- and LFD-fed male GhrR KO and WT mice (Figure (Figure3).3). Prior to the study, fasted BW of LFD-fed GhrR KO mice (24.86 ± 0.67) was significantly less than WT controls (29.7 ± 1.63). However, in the HFD-fed groups, while there was a trend toward reduced BW in the GhrR KO mice (43.90 ± 1.21), it was not statistically significant versus WT (47.91 ± 1.37). Significant increases in the GIR of LFD-fed GhrR KO mice (41.0 ± 1.4 mg/kg/min) vs WT littermates (33.1 ± 2.1 mg/kg/min) as well as HFD-fed GhrR KO mice (29.1 ± 2.1 mg/kg/min) vs WT (22.0 ± 2.3 mg/kg/min) were observed (Figure (Figure3A).3A). Consistent with this, glucose disposal (Rd) was significantly enhanced in GhrR KO mice (Figure (Figure3B).3B). HGP (Figure (Figure3C)3C) was significantly reduced in GhrR KO vs WT littermates on either a LFD (0.12 ± 0.09 vs. 0.57 ± 0.15 mg/kg/min) or HFD (-0.14 ± 0.32 vs. 1.2 ± 0.41 mg/kg/min). Furthermore, the enhancement of Rd seen in GhrR KO mice was associated with significantly increased glucose uptake in 6 out of 8 tissues tested (WAT, BAT, gastrocnemius, soleus, diaphragm and cerebral cortex; see Figure Figure4).4). The trend of increased glucose uptake in tissues of GhrR KO mice relative to WT controls failed to reach statistical significance for hypothalamus and heart. These results are consistent with the enhanced insulin sensitivity previously observed in conscious HFD-fed GhrR KO mice . While differences in BW will have contributed to the insulin sensitivity observed in the LFD-fed KO mice, based on the data obtained in the HFD-fed group it is equally clear that BW is not the only factor involved in the insulin sensitivity of GhrR KO mice.
To confirm the reduced HGP of HFD-fed GhrR KO mice observed in the clamp study, we conducted a pyruvate tolerance test. There were no significant differences in fasted BW between GhrR KO and WT mice fed a HFD (47.63 ± 1.36 and 48.78 ± 1.28, respectively). HFD-fed GhrR KO mice had lower plasma glucose levels measured at 15 and 30 minutes after administration of the gluconeogenic substrate pyruvate (Figure (Figure5,5, p < 0.05), indicative of reduced hepatic glucose output.
UCP2 is a fatty acid-responsive mitochondrial inner membrane carrier protein showing wide tissue distribution, but with a substantially increased presence in fatty liver. In comparison with lean animals, hepatic gene expression of UCP2 is increased in ob/ob mice and DIO rats [11,12]. Increased UCP2 expression in steatotic liver appears to play a reactive and protective role in limiting oxidative damage associated with increased fatty-acid oxidation [13-16]. In the present study GhrR KO mice fed a HFD showed significantly reduced UCP2 expression relative to HFD-fed WT mice (Figure (Figure6A),6A), consistent with the decreased hepatic steatosis in HFD-fed GhrR KO mice .
UCP1 functions in BAT to uncouple substrate oxidation from ATP production leading to the generation of heat . We measured BAT UCP1 expression to determine if the increased BAT glucose uptake may lead to increased thermogenesis. In contrast to liver-UCP2, there was no change in BAT-UCP1 mRNA expression in HFD-fed GhrR KO mice (Figure (Figure6B6B).
Ghrelin and its receptor play an important role in the regulation of glucose homeostasis. Both ghrelin KO and GhrR KO mice demonstrate a lower fasting blood glucose with reduced corresponding plasma insulin relative to wild type littermates when fed a HFD [2,5], suggesting that ablation of ghrelin signaling improves insulin sensitivity. In the present studies, we have further characterized the degree and nature of the insulin sensitivity in GhrR KO mice in both hyperglycemic and hyperinsulinemic-euglycemic clamp assays. Obviously, decreases in body weight will lead to improvements in glucose homeostasis. However, while there was a tendency for the HFD-fed GhrR KO mice to have reduced BW (~5 g) relative to WT controls, this was not always the case. Our lab has evaluated dozens of HFD-fed GhrRKO mouse cohorts over the course of roughly 5 years of experimenting with this model. In so doing we have documented a degree of variability in the BW response to HFD from cohort to cohort such that we occasionally observe non-significant differences between groups. That variability is reflected in the data described in the present work and most likely indicates physiological variability one would expect to see upon repeated testing over the course of a year. However, despite the variability of the BW responses in GhrR KO mouse cohorts, the reported pattern of insulin sensitivity characterized by significantly reduced insulin release coupled with improved glucose disposal is always present. Thus, while not investigated explicitly in these experiments, our data reveal that there is a BW-independent component to the improved insulin sensitivity of GhrR KO mice. Further, we provide evidence from HI-clamp, PTT and expression data that suggests an hepatic mechanism for GhrR KO insulin sensitivity. Clearly, decreases in BW contribute to the overall improved insulin sensitivity of HFD-fed GhrR KO mice, but it is not the only factor.
Our lab has demonstrated previously that GhrR KO mice fed HFD require less insulin for glucose disposal relative to WT controls during a GTT . We therefore conducted HG clamps in order to examine GSIS in these animals in more detail. The GhrR KO mice fed a HFD responded to the initial glucose priming dose with a robust first-phase release of insulin that was not significantly different than WT controls, indicating no deficit in their ability to secrete insulin upon challenge . However, a relative trend of decreased insulin requirement of GhrR KO mice was maintained throughout the 2nd phase of insulin release , and reaching its maximal effect at the point of HG clamp. This pattern is consistent with previous data obtained during a GTT in HFD-fed GhrR KO mice (2). Indeed, GhrR KO mice fed a HFD are more insulin sensitive compared to WT mice as determined by their significantly increased GIR during the HG clamp. Furthermore GhrR KO mice showed decreased HGP, increased glucose disposal and increased tissue glucose uptake in HI-E clamp studies. Remarkably, these effects could be discerned regardless of whether mice were fed a LFD or HFD. In contrast, LFD-fed ghrelin-deficient mice responded to a glucose challenge with increased insulin secretion . The reasons for this apparent dichotomy may be a reflection of differences in islet glucose sensing as a result of chronic exposure to high plasma FFA  caused by the HFD. Thus, unlike glucose responsiveness, the underlying improvements of insulin sensitivity resulting from blockade of GhrR signaling do not appear to be affected by diet.
In order to understand the mechanism underlying improved hepatic insulin sensitivity, we evaluated hepatic insulin sensitivity of HFD-fed GhrR KO mice with a pyruvate tolerance test (PTT), which measures the capacity of the liver to convert pyruvate to glucose, a process that would normally be inhibited by insulin. In comparison to WT control mice, GhrR KO mice had a significantly lower gluconeogenic response to the pyruvate challenge, consistent with their relative insulin sensitivity. Therefore, blockade of GhrR signaling does not appear to interfere with gluconeogenesis, but rather may decrease HGP via an alternative mechanism, possibly as a secondary result of decreased hepatic lipid burden (2).
Hyperinsulinemia is associated with hepatic steatosis and hyperlipidemia in humans and animal models [20-22], and treatment with insulin-sensitizing drugs ameliorates these conditions [21,23]. Increased fatty acid levels lead to elevations in hepatic UCP2 expression in rats , and the circulating free fatty acid concentration correlates with UCP2 expression in white fat and skeletal muscle , suggesting that UCP2 is important for the metabolic adaptation of these tissues to excessive fatty acid. Both GhrR and ghrelin KO mice have improved plasma lipid profiles relative to WT mice on a HFD [2,5]. GhrR KO mice have higher fatty acid oxidation and lower lipogenesis, as evidenced by improved hepatic steatosis and lower intestinal triglyceride secretion rate when exposed to HFD . Ghrelin deficiency has been associated previously with a decrease in mitochondrial UCP2 mRNA expression in the livers of chow-fed mice . Likewise, in the present study hepatic expression of UCP2 mRNA was lower in HFD-fed GhrR KO mice. Taken together, our results suggest that lower expression of UCP2 mRNA of liver reflects improved lipid metabolism in HFD-fed GhrR KO mice and likely contributes to the overall insulin sensitivity in these animals.
Our observation that GhrR KO mice had increased glucose uptake into BAT suggested that increased glucose utilization in this tissue may contribute to the improved metabolic phenotype of these mice via increased fatty acid oxidation. BAT is a thermogenic organ in which increased expression of UCP1 decreases mitochondrial energy efficiency, leading to the generation of heat . However, the impact of GhrR signaling on thermogenesis is not yet clear. To investigate whether the improved glucose uptake in BAT is associated with increased thermogenesis in GhrR KO mice consuming HFD, we measured UCP1 mRNA expression in BAT. We previously reported that GhrR deletion in mice had no effect on total energy expenditure , which was consistent with the finding that injection of ghrelin had no effect on energy expenditure in rats . In another study, however, central administration of ghrelin suppressed energy expenditure and thermogenesis in BAT via an inhibitory effect on BAT sympathetic nerve activity [26,27]. In the current study, BAT UCP1 mRNA expression was unaffected by genotype in HFD-fed mice, which indicated that diet-induced thermogenesis was not impacted by the loss of ghrelin signaling.
In summary, these data provide further compelling evidence that blockade of the GhrR improves insulin sensitivity. This effect in HFD-fed GhrR KO mice has now been demonstrated using the GTT, PTT, HG-clamp and HI-E clamp assays. Furthermore, the characteristic decrease in the insulin required for glucose disposal in GhrR KO mice, originally demonstrated in GTT experiments, has been confirmed in HG clamps. The insulin sensitivity of these mice was confirmed by the observation of increased glucose uptake by peripheral tissue during HI-clamp. Thus, we provide evidence for hepatic insulin sensitivity as a contributing factor to the overall insulin sensitivity observed in the GhrR KO mice. Whether hepatic insulin sensitivity contributes to, or is secondary to reduced hepatic steatosis was not determined in these studies. GhrR KO mice have decreased HGP in both the HI-clamp and PTT. Consistent with these observations we found elevated UCP2 expression levels in the livers of HFD-fed GhrR KO mice. These findings demonstrate further the broad metabolic improvements associated with blockade of GhrR signaling and substantiate the therapeutic potential of GhrR antagonists in the treatment of metabolic diseases, such as type 2 diabetes.
The GhrR KO animals used in these studies were bred from a single founder mouse at the Charles River Laboratories (Wilmington, MA) . GhrR +/- mice on the C57BL/6 genetic background (N4) were bred to produce homozygous null and wild type littermate controls. Mice were genotyped as described . Mice were housed in controlled environment rooms (72°F, ~40% humidity, 12 h/12 h in-phase light/dark cycles at 6 am/6 pm) in ventilated racks (Thoren; Hazelton, PA). All mice had ad libitum access to normal chow from weaning (PicoLab rodent diet 20, Purina; St. Louis, MO). Diet-induced obesity (DIO) was generated in mice by giving ad libitum access to a 60% kcal fat diet (D12492, Research Diets; Rahway, NJ) starting at eight weeks of age, for at least 4 months. In all experiments, pre-study BW was measured the morning after an overnight fast (maximum 16 hours). The work described herein involved the use of several independent cohorts of GhrR KO and WT mice over the course of two years. Experimental procedures were in accordance with regulations of the Elixir Pharmaceuticals Institutional Animal Care and Use Committee.
The HG clamps were performed over a 90 min period as previously described , with the following modifications. During the entire procedure, the animals were kept warm using a heating pad. Tail blood samples (25 μl) were removed at 0 (baseline), 1 and 5 min after a priming intravenous injection of 50% glucose (0.25 g/kg), for the measurement of first phase blood glucose and plasma insulin. Immediately after the 5 min sampling time point animals received an infusion of 20% glucose. Tail blood glucose readings were taken at 5-10 min intervals thereafter and the 20% glucose infusion rate was adjusted in order to reach, and then maintain, blood glucose levels at approximately 300 mg/dl. Twenty-five μl blood samples were collected via the carotid artery at 20, 50, 70 and 90 min after baseline for plasma insulin determination. During the 50-90 min time points, steady state hyperglycemia was established (3 consecutive readings of 300 mg/dl) at which time the GIR was determined. Preliminary experiments were performed comparing diet-induced obese (DIO) vs lean mice prior to conducting experiments on HFD-fed GhrR KO vs WT mice (Figure (Figure11).
30-week-old HFD- and LFD-fed male mice were fasted overnight (16-18 h), anesthetized with sodium pentobarbital (50 mg/kg, i.p.), before undergoing a hyperinsulinemic-euglycemic clamp as previously described . During the entire procedure, the animals were kept warm using a heating pad. Thirty min after the right internal jugular vein cannulation, mice underwent the hyperinsulinemic-euglycemic (HI) clamp for 120 minutes. A second dose of anesthetic (10 mg/kg, i.p.) was delivered 30 minutes after the start of the clamp and mice were maintained under anesthesia for the duration of the experiment. A priming dose of human insulin (100 mU/kg, iv; Humulin R; Eli Lilly, Indianapolis, IN) was administered, followed by continuous iv infusion at 50 mU/kg/min. Tail blood samples (4 μl) were collected at 10 min intervals for measurement of glucose (Ascensia Elite glucometer, Bayer, Indianapolis, IN), and 20% glucose was infused to maintain blood glucose between 110 and 140 mg/dl. Insulin-stimulated whole body glucose disposal was assessed using a priming injection of 5-μCi HPLC-purified [3-3H] glucose (Perkin Elmer, Boston, MA) followed by continuous infusion at 0.05 μCi/min throughout the study; Once steady state was attained for 30 minutes, a bolus of 10 μCi 2-deoxy-D-[1-14C] glucose (Perkin Elmer, Boston, MA) was injected intravenously (minute 75) to determine the insulin-stimulated glucose uptake of various tissues. Insulin was measured in 10 μl blood samples drawn before cannulation surgery and at the end of clamps. The mice were killed at the end of the experiment, and muscles, perigonadal white adipose tissue, and liver were harvested, frozen in liquid nitrogen, and stored at -80°C until processing. The glucose infusion rate (GIR), HGP, rate of glucose disposal (Rd), and tissue glucose uptake were determined as previously described [31,32].
30-week-old, HFD-fed male mice were fasted overnight (16-18 h), before undergoing a PTT as previously described . GhrR KO mice and wild type littermates each received an i.p. injection of 1 g/kg sodium pyruvate dissolved in water. Tail-vein blood samples were assessed for glucose immediately before injection (time 0) and at 15, 30 and 60 minutes post-injection.
Total RNA of flash-frozen liver and BAT from overnight fasted mice was extracted using TRIzol reagent (Invitrogen). The mRNA of PEPCK, G-6-Pase and UCP-2 in liver, and mRNA of UCP-1 in BAT were analyzed by real time PCR and normalized to a housekeeping gene (36B4) (Light Cycler 480 SYBR Green, Roche Diagnostics, Indianapolis, IN).
Plasma insulin was measured by a homogeneous time-resolved fluorescence immunoassay (HTRF) (Cisbio-US, Inc, Bedford, MA).
All values are expressed as the mean ± SE. Changes in various parameters were analyzed by two-way ANOVA and pair-wise differences assessed using Bonferroni post-hoc tests (GraphPad Prism, San Diego, CA). We compared end-of-study values using the Student's t-test. P < 0.05 was considered significant.
YQ designed and conducted HG and HI-E clamps and PTT assays and assisted in the writing/editing of the manuscript. KAL assisted in conducting clamp and PTT assays and in writing/editing the manuscript. TM and AN assisted in conducting PTT and expression assays. EG assisted in conducting PTT and expression assays as well as writing/editing the manuscript. DJG, SG, CZ, KM, and JH assisted in conducting clamp and PTT assays. JOS assisted in writing/editing manuscript. PSD was involved in study design and assisted in writing/editing manuscript. BJG was involved in study design and wrote the manuscript. All authors read and approved the final manuscript.
We would like to thank Tim Morrison for his excellent work as manager of Elixir's animal facility.