Exendin-4 Potently Decreases Ghrelin Levels in Fasting Rats

RESEARCH DESIGN AND METHODS

Animal housing.

Adult male Sprague-Dawley rats (250–325 g) were maintained with free access to water and standard feed (A04; Panlab, Barcelona, Spain) with a 12-h light:dark cycle (lights on 800–2000) and controlled room temperature (20–21°C). All experimental procedures were carried out in accordance with the European Union regulations regarding the protection of animals used for experimental purposes (Council Directive CEE 86/609).

Drugs and peptides.

Ex-4, Ex-3, Ex-(9–39), GLP-1(7–36)-NH₂, GLP-1(1–37), and sodium pentobarbital were all provided by Sigma-Aldrich (Alcobendas, Spain). GLP-2, Ser8-GLP-1(7–36)-NH₂, Ex-(3–39), and Lys-Pyrrolidide (H-Lys[4-nitro-Z]-pyrrolidide-HCl) were obtained from Bachem (Bubendorf, Switzerland).

Intraperitoneal administration.

Rats were maintained in standard conditions before fasting, and animals were handled on a daily basis to avoid stress. On the day of the experiment, rats were intraperitoneally administered the test drug dissolved in sterile 0.9% saline or vehicle alone, and they were returned to their cages. Trunk blood samples were collected after decapitation in tubes cooled on ice. The blood was centrifuged, and the serum obtained was frozen and stored at −20°C until assayed.

Intracerebroventricular administration.

A permanent polyethylene cannula (PE 20; Intramedic) was stereotaxically implanted into the lateral ventricle of rats anesthetized with sodium pentobarbital (50 mg/kg i.p.) using the coordinates established by Paxinos and Watson (27). After surgery, the rats were left to recover for at least 7 days in individual cages and were handled daily to avoid stress. On the day of the experiment, the treatments were intracerebroventricularly administered to the rats in a volume of 5 μl [Ex-4 (0.5, 1, or 5 μg); GLP-1(7–36)-NH₂ (0.1, 1, or 10 μg); GLP-1(1–37) (10 μg)] or vehicle, and trunk blood was collected 1 or 2 h later. The correct placement of the cannula was confirmed by postmortem injection of 10 μl Trypan blue (0.5%) (Sigma-Aldrich), and only rats with clear staining of the third ventricle were included in the study.

Freely moving experimental protocol.

One week before the experiment, male Sprague-Dawley rats (275–325 g) were implanted with an intracerebroventricular cannula and a permanent silastic indwelling atrial cannula introduced through the right jugular vein under pentobarbital anesthesia. Rats were allowed to recover for 3 days in individual cages, and they were acclimatized to the experimental conditions in metabolic cages for a further 4 days. All animals included in the study recovered their body weight to the values before surgery, and their 24-h accumulated food intake was at least 24 g (means ± SE 25.5 ± 1.5 g). From the 8th day after implantation, the cannula rats were fasted for 48 h. The peptides assayed were administered through the intracerebroventricular cannula as follows: Ex-4 (0.5 μg/rat), GLP-1 (10 μg/rat), or saline in a volume of 5 μl. Thirty minutes later, a blood sample (300 μl) was obtained immediately after the injection of a heparin bolus (0.3 ml, 1,000 IU/ml) (Rovi, Madrid, Spain) through an extension tube attached to the free end of the venous cannula. Then, rats were given access to pulverized standard feed (time 0), and additional blood samples were obtained at 30 min and at 1, 2, 4, 8, and 24 h, when the volume extracted was replaced with saline. Partial food intake was measured at the time of blood sampling. The accumulated food intake for each interval was presented, as well as the food intake per 100 g body wt, and over each interval (minute) was considered as the weighted food intake.

Hormone measurements.

Total ghrelin was measured using a specific radioiummunoassay (RIA) kit following the manufacturer’s instructions (Phoenix Pharmaceuticals). In our hands, the intrassay coefficents of variation (CVs) were 2.5% (8 pg/tube), 2.0% (32 pg/tube), and 4.5% (128 pg/tube), whereas the interassay CVs were 7.2, 4.3, and 6.2, respectively. Leptin levels were determined using a specific rat RIA (Linco Research) with intrassay CVs of 2.4% (1.6 ng/ml) and 4.8% (11.6 ng/ml), and interassay CVs of 4.6 and 5.7%, respectively. Insulin levels were also measured with a sensitive RIA assay (DRG Systems, Marburg, Germany), with intrassay CVs of 2.3% (0.2 ng/ml) and 3.9% (1 ng/ml) and interassay CVs of 4.7 and 5.9%, respectively. C-peptide levels were determined using a specific rat RIA (DRG Systems), presenting intrassay CVs of 0.8% (200 pmol/l) and 1.9% (800 pmol/l) and interassay CVs of 1 and 2.1%, respectively.

Statistical analysis.

The data are presented as the means ± SE in all graphs. Student’s t test for independent samples was used throughout for comparison between the two groups. One-way ANOVA followed by Tukey’s post hoc test was used for comparisons between multiple independent groups.

RESULTS

Ex-4 (5 μg/kg) significantly reduced the levels of ghrelin in unrestrained, nonanesthetized animals 30 min after a single intraperitoneal injection compared with control rats administered saline (Fig. 1). Indeed, in fasted animals, higher doses of Ex-4 (20 μg/kg i.p.) produced a 72% reduction in ghrelin levels 2 and 4 h postinjection, and since ghrelin levels were still significantly reduced after 8 h, the effect of Ex-4 appears to be potent and to have a long-lasting effect (Fig. 2A). Fasted animals displayed low levels of leptin that remained unchanged after Ex-4 administration (Fig. 2B). Ex-4 (5 μg/kg i.p.) reduced ghrelin levels in animals fasted for 12–72 h but not in rats fed ad libitum (Fig. 2C). As expected, leptin levels diminished rapidly after the onset of fasting and they fell 88% after 12 h. Similarly, leptin levels were unaffected by Ex-4 administration at any time point examined (Fig. 2D).

The intraperitoneal administration of Ex-4 diminished the levels of ghrelin in fasted animals in a dose-dependent manner (Fig. 3A), with maximal inhibition at 5 μg/kg. Ex-4 reduced ghrelin levels by up to 74% with respect to saline-administered fasted rats, which represents 38% below that of the ghrelin found in rats fed ad libitum. In contrast, the inhibitory effect of Ex-4 was not reproduced by other members of the GLP-1 family, such as the endogenous insulinotropic active peptide GLP-1(7–36)-NH₂, even at higher peptide doses (20 μg/kg) (Fig. 3B). Likewise, GLP-2, another active anorexigenic product of the proglucagon peptide also failed to mimic the effect of Ex-4 (28), as did the inactive GLP-1 (1–37) assayed here as a negative control (11,29) . These peptides may be rapidly inactivated by the enzyme DPP-IV, especially the native GLP-1(7–36)-NH₂, which has a very short half-life in plasma (<2 min) (21). However, no effects of GLP-1(7–36)-NH₂ (20 μg/kg i.p.) on ghrelin levels were observed over shorter periods of time (5, 15, 30, and 60 min) (Fig. 3C).

Fasting levels of ghrelin were also inhibited in a dose-dependent manner by intracerebroventricularly administered Ex-4 (Fig. 3D). Interestingly, GLP-1(7–36)-NH₂ at doses previously (0.1, 1, and 10 μg) (12,13) and here (10 μg) effectively reduced food intake and did not affect the ghrelin levels in fasting rats when intracerebroventricularly given (Figs. 3E and F).

Ex-(9–39) (100 μg/kg i.p.) is a specific antagonist of the GLP-1r (30) that completely blocked the increase in insulin induced by Ex-4 (Fig. 4A) but not the Ex-4–induced decrease in ghrelin (Fig. 4B). Indeed, Ex-(9–39) itself significantly reduced the fasting levels of ghrelin (>25%). In addition, the inactivation of DPP-IV by administration of the long-term blocker Lys-Pyrrolidide (31) markedly increased the circulating levels of C-peptide by preventing the enzymatic degradation of the endogenous GLP-1 (Fig. 4C). In contrast, inactivating DPP-IV did not modify ghrelin levels over the period studied (0, 30, and 60 min) (Fig. 4D). Moreover, the GLP-1 analog Ser8-GLP-1(7–36)-NH₂, which has a half-life of ∼1 h in body fluids (32), failed to reduce ghrelin levels when intraperitoneally administered. However, Ex-3, which differs from Ex-4 in two amino acids being one of those a serine at position two of the peptide, displayed a similar profile of ghrelin inhibition as Ex-4 (Fig. 4E). Interestingly, the Ex-4 fragment that lacks the first two amino acids [Ex-(3–39)], failed to inhibit ghrelin, indicating that those amino acids are important for Ex-4 to influence circulating ghrelin levels (Fig. 4E).

We also studied whether the decrease in ghrelin following intracerebroventricular administration of Ex-4–modified food intake in freely moving rats (Fig. 5). The elevated levels of ghrelin provoked by fasting over the previous 24 h progressively diminished on the initiation of feeding, reaching a minimum after 4 h in saline-treated rats (Fig. 5C). Administration of Ex-4 (0.5 μg i.c.v.) 30 min before allowing the rats free access to food almost completely blocked the food intake for 8 h in fasted rats, with a marked reduction still detected at 24 h (Fig. 5A). The weighted food intake values were also reduced by 95% at 30 min and >80% at 1 h, and they remained very low for up to 8 h after Ex-4 administration (Fig. 5B). It is important to emphasize that the kinetics of ghrelin circulating levels in the Ex-4–treated rats after intracerebroventricular administration were very similar to those described following intraperitoneal administration. The levels of ghrelin start to fall after 30 min, and they remain very low for at least 8 h, correlating with the marked reduction in food intake (Fig. 5C). Accordingly, insulin levels also remained low for up to 8 h after Ex-4 administration (Fig. 5D).

In similar experiments, GLP-1 (10 μg) was much less effective, and it reduced the accumulated food intake by 58% at 30 min, 63% at 1 h, 64% at 2 h, and 62% at 4 h. Thereafter, GLP-1–treated rats showed a catch-up phenomenon to produce a similar food intake to controls over the entire 24-h period (Fig. 5A). GLP-1(7–36)-NH₂ also reduced weighted food intake values after 30 min (50%) and 1 h (60%) (Fig. 5B), and this was accompanied by a similar reduction in ghrelin levels to that found in saline-treated rats over this period (Fig. 5C). Insulin levels progressively increased from the 1st h onwards, apparently reflecting the food intake (Fig. 5D).

Finally, we found that, while still very potent, Ex-4 (5 μg/kg) was less effective at reducing food intake in fasted rats when intraperitoneally administered, although it still reduced ghrelin levels (Fig. 6). Again, GLP-1(7–36)-NH₂ (50 μg/kg i.p.) did not reduce ghrelin levels at the beginning of the experiment and accordingly did not modify food intake. Furthermore, prior administration of Ex-(9–39) (100 μg/kg i.p.) neither blocked the effect of Ex-4 on ghrelin levels nor modified the reduction in food intake (Fig. 6).

DISCUSSION

We have identified the first compound that strongly diminishes circulating levels of ghrelin. Here we provide evidence that Ex-4, a peptide postulated as an agonist of the GLP-1r (19,28,36), reduces the circulating levels of ghrelin for long periods in fasted rats in a dose-dependent manner. It is noteworthy that a single intraperitoneal injection of Ex-4 reduces ghrelin levels for at least 8 h. Indeed, a reduction of 74% of the control fasting levels of ghrelin was observed, and this reduced level was 38% below the ghrelin levels in rats fed ad libitum. In this respect, no other signal or hormone has been reported to be as potent as Ex-4. However, the effects of Ex-4 on ghrelin are restricted to the fasting state, since it does not significantly reduce the concentration of ghrelin in rats feeding ad libitum.

It is thought that the reduction in food intake induced by Ex-4 may be due to its anorexigenic effects exerted over the hypothalamus (12,13,33). On the other hand, leptin appears to interfere with the generation of appetite-antagonizing ghrelin effects in the hypothalamus (34) and to reduce ghrelin levels in the rat (35). Therefore, leptin might influence the effects of Ex-4 on ghrelin inhibition. However, central administration of Ex-4 reduces food intake in both lean and obese Zucker rats, suggesting that this effect is not dependent on the circulating levels of leptin (36). Accordingly, we also found a decrease in leptin levels following fasting that was unaffected by Ex-4 administration. Since Ex-4 strongly reduced circulating ghrelin levels, it may be inferred that the effects of Ex-4 on ghrelin are independent of the changes in leptin and that the anorexigenic effects of Ex-4 could be due to its ability to reduce the levels of the orexigenic hormone ghrelin.

The effects of Ex-4 on ghrelin are evident after both peripheral (intraperitoneal) and central (intracerebroventricular) administration. Ex-4 can cross the blood-brain barrier in a manner that is not affected by fasting but that is limited by dose (37). Thus, Ex-4 activity might be explained by a single central mechanism that would also be activated by peripheral administration if sufficient Ex-4 crosses the blood-brain barrier (37). While this hypothesis seems reasonable at present, additional studies will be necessary to ascertain whether this is in fact the case.

We also analyzed the effects of other peptides of the GLP-1 family, such as GLP-1(7–36)-NH₂, the endogenous GLP-1r agonist, or GLP-2. Intriguingly, none of these were able to reduce circulating ghrelin levels at any time point (minutes to hours) irrespective of the route of administration. These peptides had essentially the same effects as saline or the inactive GLP-1 analog GLP-1(1–37) used as a negative control (11,35). It could be argued that the short half-life of GLP-1(7–36)-NH₂ might justify this lack of activity; however, no responses were observed even after 5 or 15 min. Moreover intracerebroventricular administration appears to augment GLP-1 anorexigenic activity precluding the degradation of GLP-1 (12,13,33). Whereas we have used doses that are effective in reducing food intake both here (Fig. 5) and in other similar studies (12,13), no significant change in ghrelin levels was produced by administration of GLP-1. The Ser8-GLP-1(7–36)-NH₂ analog that has a much longer half-life than the native peptide also failed to affect ghrelin. In contrast Ex-3, which differs in two amino acids with respect to Ex-4 [being one of those the same serine as Ser8-GLP-1(7–36)-NH₂, also strongly diminished ghrelin. Moreover, increasing the endogenous GLP-1 activity by blocking DPP-IV with Lys-Pyrrolidide did not produce a significant change in ghrelin levels. Furthermore, a specific antagonist of the GLP-1r, Ex-(9–39), did not block the inhibitory effect of Ex-4 on ghrelin levels when used in the pharmacologically active range; rather, it had a partial mimetic effect reducing ghrelin levels by >25%. This could not be explained by a loss of Ex-(9–39) activity, as it completely blocked the Ex-4–induced increase of insulin in the same experiment, showing that it effectively blocked GLP-1r–mediated effects (30). Together, these data suggest that ghrelin inhibition appears to correspond to an Exendin-specific mechanism, which might be independent of GLP-1r activation.

In recent years, several studies have provided evidence that the endocrine effects of Ex-4 cannot be adequately explained by activation of the GLP-1r. Indeed, Ex-4 but not GLP-1 increases insulin sensitivity and glucose transport in cultured L6 myotubules in vitro (38), and the effects of Ex-4 and Ex-(9–39) on cAMP levels in 3T3-L1 adipocytes are not mediated by the GLP-1r (39). Significantly, GLP-1 activation was unable to modify energy expenditure (measured as Vo₂) or the respiratory exchange ratio in mice (40), whereas Ex-4 clearly reduced both these parameters. Moreover, other endocrine effects elicited by Ex-4 are not reproduced by GLP-1, such as the reduction of thyrotropin secretion in rats (41). Hence, we can infer that Ex-4 exerts some of its effects through a mechanism that does not involve the activation of the GLP-1r. Our data also suggest that there may be a specific pathway through which Ex-4 acts, although whether exendins may have their own specific independent receptor remains to be seen. In this sense, it would be interesting to repeat some of these studies in the GLP-1r knockout mice, although the basic phenotype of this model does not differ greatly from that of the wild-type littermates (42), and unexpectedly, these mice do not become diabetic (43). Apparently, these mice compensate for the lack of GLP-1 with an increase in the levels of the complementary incretin glucose-dependent insulinotropic peptide (44).

Although ghrelin appears to exert important biological effects in a variety of tissues, its physiological role in health and disease is still far from clear. Deletion of the ghrelin gene in mice does not impair either growth or appetite (45), although it has been shown that the constitutive absence of ghrelin prevents the obesity induced by a high-fat diet (46). Furthermore, transgenic rats expressing an antisense growth hormone secretagogue receptor mRNA that selectively attenuates growth hormone secretagogue receptor protein expression in the arcuate nucleus display a lower body weight and less adipose tissue than control rats (47). Furthermore, obestatin, a peptide generated from the ghrelin gene, inhibits food intake in normal mice, apparently by binding to the orphan receptor gpr39 (48). Therefore, the information generated by the ghrelin knockout mice might be limited by the fact that several peptides can be derived from the ghrelin-gene. Given that these may have different and even opposing biological effects, their simultaneous inactivation may make it difficult to determine the influence of ghrelin on energy balance, food intake, and body weight.

We show here that a single intracerebroventricular injection of Ex-4 strongly reduces ghrelin levels and food intake for at least 8 h without any significant changes in the levels of insulin (Fig. 5). The plasma ghrelin profile in response to intracerebroventricularly administered Ex-4 in fasted freely moving rats appears to be similar to that after Ex-4 intraperitoneal administration. Interestingly, the effects of Ex-4 on ghrelin and food intake were closely correlated following intracerebroventricular or intraperitoneal administration of Ex-4. The effects of Ex-4 on ghrelin levels seem to be exerted by a mechanism not mimicked by GLP-1 or blocked by the prior administration of Ex-(9–39). However, the influence of GLP-1r on this aspect of Ex-4 activity cannot be excluded. Although less potent, GLP-1(7–36)-NH₂ (intracerebroventricular) diminished the accumulated food intake despite the fact that it does not alter ghrelin levels. These data indicate that the activation of the GLP-1r has its own role in regulating food intake, which might be independent of the variation in ghrelin expression. In fact, we believe that activation of the GLP-1r might be partially responsible for some of the effects exerted by Ex-4 on food intake.

These results could be particularly relevant to the pathophysiology of certain types of obesity, such as the PWS. Indeed, it has been established that the levels of circulating ghrelin are strongly elevated in the plasma of patients with PWS due to an increase in ghrelin-producing cells in the stomach (7,8,49). Thus, it has been proposed that ghrelin might play a significant or indeed crucial role in the pathogenesis of this still incurable disease (8). The only rodent model for PWS currently available does not survive beyond the neonatal stage due to a failure to thrive (50). We have tested the effect of Ex-4 in extensively fasted (48 and 72 h) rats with intense hyperghrelinemia, a model representative of this pathology. Ex-4 strongly inhibits ghrelin, and it may therefore be a candidate drug for the treatment of PWS or even for obesity in general. Since Ex-4 has recently been approved for the treatment of type 2 diabetes, this information is of immediate clinical relevance and utility. The assessment of ghrelin levels in human subjects treated with Ex-4 are eagerly awaited.

In conclusion, Ex-4 reduces fasting levels of ghrelin for long periods in a dose-dependent manner, both at peripheral and central sites. Ex-4 is the most potent peptide to exert this effect to date. The Ex-4–induced reduction in ghrelin levels is sufficiently potent as to open new avenues in the study of the physiological regulation of ghrelin, as well as to clarify the pathophysiological role of ghrelin in the hyperphagia of type 2 diabetes and obesity. This effect may be especially relevant for the treatment of PWS, which to date is characterized by an absence of medical treatment other than bariatric surgery (7).