Introduction

Ghrelin is a small peptide (28 amino acids) that is primarily produced by the stomach and binds to the growth hormone secretagogue (GHS) receptor (GHS-R) [1, 2]. It circulates in two major forms: n-octanoyl-ghrelin (ghrelin), which contains an n-octanoyl modification at serine-3, and des-octanoyl or unacylated ghrelin (des-ghrelin) [3]. Although des-ghrelin does not bind to the GHS-R [1, 4], it may be biologically active [5].

Ghrelin is a component of the gut-brain axis and is involved in the control of energy balance. It stimulates food intake in rodents and man [5–8], where peak levels in plasma occur just before each meal and fall rapidly after refeeding, suggesting that it serves to initiate food consumption [9]. To enhance appetite, ghrelin acts in the hypothalamus, where it promotes neuropeptide Y (NPY) and orexin gene expression and inhibits the production of pro-opiomelanocortin/α-melanocyte stimulating hormone (-MSH) via activation of the GHS-R [10–16]. We recently showed that intracerebroventricular infusion of NPY hampers insulin’s capacity to suppress endogenous glucose production (EGP) [17], whereas central injection of melanotan II, a synthetic analogue of α-MSH, reinforces insulin action on glucose disposal [18]. Also, a number of studies have demonstrated an inverse relationship between fasting plasma ghrelin and insulin concentrations and insulin resistance in adults and children [19–27]. These findings led us to hypothesise that ghrelin might adversely affect insulin sensitivity through activation of the GHS-R.

To test this hypothesis, we intravenously administered vehicle, ghrelin or growth hormone releasing peptide 6 (GHRP-6), which is a specific agonist of the GHS-R, and measured EGP and glucose disposal by 14C-glucose dilution during a hyperinsulinaemic clamp in mice.

Until recently, acylation of the serine-3 residue of the ghrelin molecule was believed to be required for bioactivity. However, recent data suggest that des-ghrelin may counteract ghrelin in the control of energy balance [28], and in vitro experiments have revealed opposing effects of ghrelin and des-ghrelin on glucose production by primary hepatocytes [29]. To further explore the role of des-ghrelin in the regulation of fuel flux, we studied the effects of this peptide in the same experimental context.

Materials and methods

Animals

Male 3-month-old C57BL/6J mice (originating from The Jackson Laboratory, Bar Harbor, ME, USA, and bred in our own animal facility) were housed in a temperature-controlled room on a 12-h light/12-h dark cycle, with free access to water and chow diet. Body weight of the mice is reported in Table 1. All animal experiments were performed in accordance with the regulations of Dutch law on animal welfare, and the institutional ethics committee for animal procedures approved the protocol.

Table 1 Body weight and plasma parameters in mice that received ghrelin (n=9), des-ghrelin (n=8), ghrelin and des-ghrelin together (n=7), GHRP-6 (n=9) or vehicle (n=8)
Full size table

Hyperinsulinaemic–euglycaemic clamp study

Mice were fasted for 4 h, with food withdrawn at 05.00 h. Hyperinsulinaemic–euglycaemic clamp studies were started at 09.00 h as described earlier [30, 31]. During the experiments, mice were sedated with 6.25 mg/kg acepromazine (Sanofi Sante Animale, Libourne Cedex, France), 6.25 mg/kg midazolam (Roche, Mijdrecht, the Netherlands) and 0.3125 mg/kg fentanyl (Janssen–Cilag, Tilburg, the Netherlands).

Vehicle, ghrelin or GHRP-6 (peptides from PhoenixEurope, Karlsruhe, Germany) was administered in a primed (vehicle, 0 μg/nmol; ghrelin, 0.60 μg or 0.18 nmol; GHRP-6, 0.16 μg or 0.18 nmol), continuous (vehicle, 0 μg h−1 nmol−1 h−1; ghrelin, 1 μg/h or 0.30 nmol/h; GHRP-6, 0.26 μg/h or 0.30 nmol/h) i.v. infusion during the whole experiment (basal and hyperinsulinaemic periods). The doses of ghrelin and GHRP-6 were based on previous reports [6, 7, 32]. First, basal rates of glucose production/uptake were measured by giving a primed continuous infusion of 14C-glucose (primed, 26 kBq; continuous, 44 kBq/h) (Amersham, Little Chalfont, UK) for 80 min. Subsequently, insulin was administered in a primed (4.1 mU), continuous (6.8 mU/h) i.v. infusion for 2–3 h aimed at a steady-state circulating insulin level of approximately 6 ng/ml. A variable infusion of 12.5% d-glucose was used to maintain euglycaemia (measured by tail bleeding; Freestyle TheraSense; Disetronic Medical Systems, Vianen, the Netherlands). Blood samples (75 μl) were taken during the basal period (after 60 and 80 min) and during the clamp period (when glucose levels were stable and 20 and 40 min later) for determination of plasma glucose, NEFA, insulin and 14C-glucose specific activities.

To assess tissue-specific insulin-mediated glucose uptake, 2-deoxy-D-[3H]glucose (2-DG; Amersham) was administered as a bolus (37 kBq) 40 min before the end of the clamp experiment. At the end of the clamp, mice were killed and skeletal muscle (hindlimb) and adipose tissue were isolated and frozen in liquid nitrogen for subsequent analysis. There were eight (vehicle) or nine (ghrelin, GHRP-6) mice in each group during the hyperinsulinaemic–euglycaemic clamp study. Because of technical problems, data on the tissue-specific, insulin-mediated glucose uptake of three mice were lost in the vehicle group.

Additional studies with des-ghrelin

In addition, a hyperinsulinaemic–euglycaemic clamp study was performed while des-ghrelin or a combination of ghrelin and des-ghrelin was administered. Des-ghrelin or the combination of ghrelin and des-ghrelin were administered in a primed (des-ghrelin, 0.60 μg or 0.19 nmol; combination of ghrelin and des-ghrelin, 0.60 μg or 0.18 nmol ghrelin and 0.60 μg or 0.18 nmol des-ghrelin), continuous (des-ghrelin, 1.0 μg/h or 0.31 nmol/h; combination of ghrelin and des-ghrelin, 1.0 μg/h or 0.31 nmol/h ghrelin and 1.0 μg/h or 0.30 nmol/h des-ghrelin) i.v. infusion during the whole experiment (basal and hyperinsulinaemic periods). The doses of des-ghrelin and the combination of ghrelin and des-ghrelin were based on a previous report [32]. Hyperinsulinaemic–euglycaemic clamps were performed as stated above and randomised with these other groups.

Analytical procedures

Plasma levels of glucose and NEFA were determined using commercially available kits (Instruchemie, Delfzijl, the Netherlands and Wako, Neuss, Germany). Plasma insulin concentration was measured with a mouse insulin ELISA (Mercodia, Uppsala, Sweden). Total plasma 14C-glucose was determined in 7.5 μl plasma and in supernatants after trichloroacetic acid (20%) precipitation and water evaporation to eliminate tritiated water. There were seven (combination of ghrelin and des-ghrelin) or eight (des-ghrelin) mice in each group during the hyperinsulinaemic–euglycaemic clamp study.

Tissue analysis

For determination of tissue 2-DG uptake, the homogenate of muscle and adipose tissue was boiled, and the supernatant was subjected to an ion-change column to separate 2-DG-6-phosphate from 2-DG as described previously [30, 33].

Calculations

The turnover rate of glucose (μmol min−1 kg−1) was calculated during the basal period and in steady-state clamp conditions as the rate of tracer infusion (disintegrations per minute [dpm] infused per min [dpm/min]) divided by the plasma-specific activity of 14C-glucose (dpm/μmol). The ratio was corrected for body weight. During the hyperinsulinaemic clamp, EGP was calculated as the difference between the tracer-derived rate of glucose appearance and the glucose infusion rate.

Tissue-specific glucose uptake in muscle and adipose tissue was calculated from tissue 2-DG content, corrected for plasma-specific activity and expressed as μmol/g tissue.

Statistical analysis

Differences between groups were determined with the Kruskal–Wallis non-parametric test for k independent samples. When significant differences were found, the Mann–Whitney non-parametric test was used as a follow-up test to determine differences between experimental groups and the control group. A p value of <0.05 was considered statistically significant. All values shown represent the mean ± SD.

Results

Plasma parameters

Body weight, plasma glucose, NEFA and insulin concentrations in basal and hyperinsulinaemic conditions are shown in Table 1. Body weight did not differ between the groups of mice. Basal glucose levels were significantly lower in the group that received GHRP-6 than in the control group. Basal insulin and NEFA levels were not different between groups. Moreover, in steady-state hyperinsulinaemic conditions, plasma glucose, NEFA and insulin concentrations were not different between groups.

Glucose turnover

Basal EGP, glucose infusion rate, insulin-mediated glucose disposal, inhibition of EGP and tissue-specific glucose uptake are shown in Table 2. In basal conditions, EGP (and thereby whole body glucose disposal) was significantly lower in animals that received GHRP-6 than in control animals. During the hyperinsulinaemic period, the rate of glucose infusion necessary to maintain euglycaemia was significantly higher in animals that received ghrelin than in control mice. Accordingly, the glucose disposal rate was significantly higher in animals that received ghrelin compared with the control group (Fig. 1). In contrast, hyperinsulinaemia suppressed EGP to a significantly lesser degree in animals that received ghrelin than in control animals (Fig. 2).

Table 2 Glucose turnover and tissue-specific glucose uptake in mice that received des-ghrelin (n=8), ghrelin and des-ghrelin together (n=7) or vehicle
Full size table
Fig. 1
figure 1

Insulin-mediated glucose disposal (μmol kg−1 min−1) in mice as indicated. Vehicle-treated, n=8; ghrelin-treated, n=9; GHRP-6-treated, n=9. Data are mean and SD. *p<0.05

Full size image
Fig. 2
figure 2

Inhibition of EGP by insulin in mice as indicated. Vehicle-treated, n=8; ghrelin-treated, n=9; GHRP-6-treated, n=9. Data are mean and SD. *p<0.05

Full size image

Tissue-specific glucose uptake

Insulin-mediated 2-DG uptake in muscle tissue was higher in ghrelin-treated animals than in the control group. Insulin-mediated 2-DG uptake in adipose tissue tended to be higher in ghrelin-treated animals than in the control group, although this difference did not reach statistical significance (p=0.09) (Fig. 3a,b).

Fig. 3
figure 3

Muscle-specific glucose uptake (μmol/g) (a) and adipose tissue-specific glucose uptake (μmol/g) (b) under hyperinsulinaemic conditions in mice as indicated. Vehicle, n=5; ghrelin, n=9; GHRP-6, n=9. Data are mean and SD. *p<0.05

Full size image

Additional studies with des-ghrelin

Plasma parameters

Body weight, plasma glucose, NEFA and insulin concentrations in basal and hyperinsulinaemic conditions are shown in Table 1. Body weight did not differ between the groups of mice. Basal glucose levels were significantly lower in the group that received the combination of ghrelin and des-ghrelin compared with the control group. Basal insulin and NEFA levels were not different between groups. Moreover, in steady-state hyperinsulinaemic conditions, plasma glucose, NEFA and insulin concentrations were not different between groups.

Glucose turnover and tissue-specific glucose uptake

Basal EGP, glucose infusion rate, insulin-mediated glucose disposal, inhibition of EGP, and tissue-specific glucose uptake are shown in Table 2. During basal conditions, EGP (and thereby whole body glucose disposal) was significantly lower in animals that received the combination of ghrelin and des-ghrelin compared with the control animals. During the hyperinsulinaemic period, the rate of glucose infusion necessary to maintain euglycaemia, insulin-stimulated glucose disposal and tissue-specific glucose uptake did not differ between the groups. In contrast, hyperinsulinaemia suppressed EGP to a significantly lesser degree in animals that received des-ghrelin than in control animals.

The effects of the combination of ghrelin and des-ghrelin were not identical to the sum of the individual peptides’ effects. Glucose infusion rate and the rate of insulin-mediated glucose disposal were significantly increased during ghrelin infusion, whereas des-ghrelin did not affect these parameters. Interestingly, co-administration of the two peptides restored the glucose infusion rate to control values. Although ghrelin and des-ghrelin alone caused hepatic insulin resistance, the combination of the two peptides abolished the individual effects of the peptides on hepatic insulin sensitivity.

Discussion

This study shows that intravenous administration of ghrelin reinforces the action of insulin on glucose disposal, whereas it hampers insulin’s capacity to inhibit EGP. In contrast, administration of GHRP-6 did not affect insulin action. Des-ghrelin hampered insulin’s capacity to inhibit EGP, whereas it did not affect glucose disposal. The restraining effects of des-ghrelin and ghrelin on hepatic insulin action were abolished by simultaneous administration of both peptides.

The glucose infusion rate required to maintain euglycaemia during insulin infusion was significantly higher in ghrelin-treated animals, indicating that this peptide enhances the whole-body insulin sensitivity of glucose metabolism. In particular, ghrelin clearly stimulated insulin-mediated glucose disposal, as determined by the tracer dilution method, an observation that was consistent with enhanced 2-DG uptake in muscle and adipose tissue during hyperinsulinaemia in ghrelin-treated animals (although statistical significance was reached only for muscle). In contrast, ghrelin hampered inhibition of EGP by insulin. As previous papers have reported that ghrelin stimulates NPY release and inhibits pro-opiomelanocortin/α-MSH neuronal activity in the arcuate nucleus [34], we expected to find that ghrelin would compromise the insulin sensitivity of both liver and muscle/adipose tissue (we recently showed that central NPY hampers insulin’s action on EGP, whereas melanocortin pathways promote insulin-mediated glucose disposal [17, 18]). However, the mechanism whereby ghrelin affects insulin action remains to be established. Intravenously administered peptide may act through central pathways, but also through the activation of receptors on peripheral tissues (the GHS-R is present in various tissues, such as pancreas and liver [15, 35]). Indeed, ghrelin blocks insulin’s inhibitory effect on the gene expression of key gluconeogenic enzymes in a hepatoma cell line [35]. These data corroborate our own in vivo observation of ghrelin hampering the action of insulin on EGP. In contrast, i.v. administration of ghrelin appears to enhance insulin-mediated glucose uptake, where we had hypothesised a priori that it would block this action (through the down-regulation of hypothalamic α-MSH signalling [18]). It is conceivable that ghrelin acts on myocytes and/or adipocytes directly to stimulate glucose disposal. We are not aware of any in vitro or in vivo data documenting ghrelin’s effects on insulin action in muscle or adipose tissue. Thus, further studies are required to establish if ghrelin affects insulin action via central or peripheral mechanisms.

Administration of GHRP-6, an agonist of the GHS-R, did not affect the action of insulin. This observation is in line with other reports indicating that ghrelin stimulates glucose production by primary hepatocytes [29], promotes adipogenesis in bone marrow [36] and inhibits preadipocyte cell proliferation [37], whereas GHS-R agonists do not. Together, these data suggest that ghrelin does not affect insulin action and metabolism via the GHS-R, but via an as-yet unidentified mechanism.

Des-ghrelin blocked the action of insulin on EGP as much as ghrelin did. In contrast, it did not affect insulin-mediated glucose disposal. These data support the emerging view that n-octanoyl modification of the serine-3 residue of ghrelin is not absolutely required for bioactivity [28, 29, 32]. However, they apparently contradict a report suggesting that ghrelin and des-ghrelin have opposing (stimulatory and inhibitory, respectively) effects on glucose production by hepatocytes [29]. The contradiction could be explained by the fact that in vitro and in vivo administration of des-ghrelin may have different effects, as des-ghrelin, like ghrelin, can act in the brain as well as in peripheral tissues [28, 36]. Co-administration of des-ghrelin abolished the effect of ghrelin on hepatic insulin sensitivity, which accords with an in vitro study showing similar results in isolated hepatocytes [29]. Our findings corroborate mounting evidence indicating that des-ghrelin has potentially important biological effects. The receptor mediating des-ghrelin bioactivity remains to be identified.

The physiological relevance of the observed effects of ghrelin in decreasing hepatic insulin sensitivity and increasing peripheral insulin sensitivity remains to be established. Ghrelin secretion by (primarily) the stomach is significantly enhanced during fasting to stimulate appetite and initiate meal consumption [9]. In a previous study, we showed that fasting enhances insulin-mediated glucose uptake in mice [38]. The data we present here allow us to hypothesise that the increase in plasma ghrelin levels during fasting is involved in the physiology of this phenomenon. However, des-ghrelin concentrations clearly dominate the plasma profile of ghrelin-like peptides, at least in the human [39, 40]. In view of the current evidence suggesting that des-ghrelin has metabolic effects of its own and indeed appears to interact with ghrelin in the control of metabolism, further studies are required to establish the role of ghrelin peptides in the regulation of energy balance and fuel flux. Moreover, the receptor(s) mediating the metabolic signals conveyed by (des-)ghrelin need to be identified.

In conclusion, ghrelin differentially affects the tissue-specific action of insulin, hampering its ability to inhibit EGP while reinforcing its effect on glucose disposal. These effects occur acutely and they are not secondary to ghrelin’s well-known influence on feeding and body weight. The GHS-R is not likely to mediate ghrelin’s metabolic effects. Des-ghrelin also appears to modulate fuel flux and may counteract ghrelin in the control of glucose metabolism.