Date: January 21st, 2010

Speaker's Name and Affiliation:

• Wolfgang Langhans
• Swiss Federal Institute of Technology, Zurich

Seminar Title: "Peripheral GLP-1 and Satiation"

Chair: Harry R. Kissileff, Ph.D.

Rapporteur: Kathleen L. Keller, Ph.D.

Attendees and their Affiliation:

Kathleen Keller	Columbia/Obesity Research Center
Harry Kissileff	Obesity Research Center
Jennifer Nasser	St. Luke's
Majella O'keefe	Columbia/St. Luke's
Miranda Johnson	NJMS
Marie-Pierre St-Onge	Columbia
Gerry Smith	WCMC
Ralph Norgren	Penn State
Steven Zukerman	Brooklyn College
John Glendenning	Barnard College
John G. Kral	SUNY Downstate
Michael Lewis	Hunter/Columbia
Karen Ackroff	Brooklyn College
JA Grinker	Univ Michigan
Susan Carnell	NYORC
Mousumi Bose	NYORC
Joe Vasselli	NYORC
Carol Maggio	NYORC
Tung Fong	Forrest
Allan Geliebter	Columbia/SLRH

Summary: (Prepared by the Rapporteur)

Peripheral GLP-1 is a typical gut/brain peptide. In the brain, GLP-1 is expressed in some neurons of the NTS, in the gut it is released from enteroendocrine L-cells, primarily in the distal jejunum and ileum, in response to luminal nutrient (carbohydrate and fat) stimulation. Receptors for GLP-1 are widely distributed in the brain and in peripheral organs such as pancreatic islets, the whole gastrointestinal tract, and various parts of the brain, including areas that are implicated in the control of food intake and energy balance. GLP-1 has several effects including a stimulatory effect on insulin release and an inhibitory effect on eating. Peripheral GLP-1 is implicated in satiation because 1) exogenous GLP-1 has been shown to inhibit eating by primarily reducing meal size and 2) because IP injection of the specific GLP-1 receptor antagonist exendin 9-39 (Ex-9) stimulated eating under some conditions. The site and mechanism of action 2 of peripheral GLP-1 to inhibit eating is largely unknown. We recently started experiments to address these issues. Gut hormones can signal the brain in an endocrine fashion or through a paracrine effect on afferent nerves. Our studies address both possibilities.

Effect of intravenously administered GLP-1 on eating (possible endocrine effect)

To mimic the meal-induced release of endogenous GLP-1 into the hepatic portal vein (HPV) as closely as possible, we equipped adult male rats with HPV or HPV and vena cava (VC) catheters and performed remotely controlled infusions of various doses of GLP-1 (0.33, 1.0, 3.0 nmol/kg body weight, BW) or vehicle during the first spontaneous nocturnal meal. Under these conditions GLP-1 infusions into the HPV (1.0 and 3.0 nmol/kg BW) specifically reduced the size and duration of the ongoing meal, without significantly affecting other parameters of spontaneous eating, subsequent meals or cumulative food intake. This suggests that GLP-1 specifically enhances satiation, i.e. the processes that terminate meals. The effect was not significantly dose- related. While these findings are consistent with a possible satiating effect of endogenous GLP-1, they do not support a role of GLP-1 in the graded satiating effect of food. Also, intrameal HPV and VC infusions of GLP-1 (1 nmol/kg BW) produced similar effects on eating, suggesting that the eating-inhibitory effect of circulating GLP-1 is not mediated by GLP-1 receptors located in the HPV or liver.

We also examined whether the eating-inhibitory effect of circulating GLP-1 is mediated by abdominal vagal afferents. Adult male rats were again equipped with HPV catheters and underwent subdiaphragmatic vagal deafferentation (SDA) or sham- surgery. SDA is the most specific technique available to lesion all abdominal vagal afferents, but leaves about 50% of the vagal efferents intact. After recovery, intrameal infusions of GLP-1 (0.25, 0.5, 1.0 nmol/kg BW) or vehicle were performed as before, and eating behavior was recorded. GLP-1 (0.5 and 1.0 nmol/kg BW) reduced meal size and duration similarly in SDA and sham-operated rats, indicating that abdominal vagal afferents are not necessary for the eating-inhibitory effect of circulating GLP-1.

In order to identify central nervous system circuits that may be involved in the eating-inhibitory effect of circulating GLP-1, we measured c-Fos peptide, the product of the early gene c-fos in some hindbrain (area postrema [AP], nucleus of the solitary tract [NST]) and forebrain structures (central area of the amygdala [CeA] as well as in the hypothalamic paraventricular and arcuate nuclei [PVN and Arc]) after HPV GLP-1 (1 nmol/kg BW) infusion. The conditions described above were modified for these experiments, i.e., rats were adapted to a 19 h feeding 5 h deprivation (0900-1400) schedule. Infusions were given at dark onset (1200). Preliminary experiments revealed that GLP-1 also reduced meal size under very similar conditions. GLP-1 infused via the HPV in a dose (1 nmol/kg BW) that reduced meal size increased c-Fos expression in the NST, AP and CeA, suggesting a role for these brain areas in the eating-inhibitory effect of circulating GLP-1. Whether these effects reflect a direct activation of these areas related to satiation, aversion, or both requires further research. Most recently, we found that infusion (2 ¦Ìl/min) of the GLP-1 receptor antagonist Ex-9 (10 ¦Ìg) into the 4th ventricle blocked the eating-inhibitory effect of HPV GLP-1 (1 nmol/kg BW), indicating that GLP-1 receptors that are accessible from the 4th ventricle are involved in mediating this effect.

As mentioned above, the rationale for infusing comparatively low doses of GLP-1 during spontaneous meals into the HPV was to mimic the physiological release of GLP-1 during a meal. Yet, the release of endogenous GLP-1 during meals in rats has scarcely 3 been measured. In another set of experiments we therefore equipped rats with HPV and VC catheters and took blood samples in parallel from both blood vessels in relation to a spontaneous nocturnal chow meal to measure active GLP-1, insulin, and glucose in plasma. This experiment revealed that endogenous GLP-1 increased during the meal in the HPV, but not in the VC. The peak plasma concentration of GLP-1 in the HPV occurred at 6 min after meal onset. HPV GLP-1 had returned to baseline values at 20 min after meal onset. VC GLP-1 did not increase. These findings suggest that under these conditions endogenous GLP-1 may have paracrine effects in the intestine or endocrine effects in the hepatic portal vein or liver, but not systemic endocrine effects. The data also argue against a role of systemic increases in GLP-1 in satiation under these conditions, i.e., in rats eating chow.

Effect of intraperitoneally administered GLP-1 on eating (possible paracrine effect)

Based on these and other findings we also examined the effect of intraperitoneal (IP) GLP-1 infusions on eating under the same conditions as described for the HPV infusions, i.e., adult male rats were equipped with IP instead of HPV or VC catheters and the effects of remotely controlled, intrameal GLP-1 infusions (doses: 2.5, 5.0, 10.0 nmol/kg BW) on eating were assessed. Similar to its effect after HPV infusion, GLP-1 (10 nmol/kg BW) selectively reduced meal size in these experiments, although the threshold dose for a reliable reduction of meal size and duration was about ten times higher than for HPV infusion. Surprisingly, SDA blocked the eating-inhibitory effect of IP but not HPV GLP-1 infusions. All animals in this experiment carried HPV and IP catheters and had undergone either SDA or sham surgery. These data suggest that IP and HPV infusions of GLP-1 activate different pathways to inhibit eating and that vagal afferent signaling is necessary for the full eating-inhibitory effect of IP, but not HPV, GLP-1. IP administered peptides appear to accumulate in the intestinal lymph and intestinal lymph closely resembles the composition of the interstitial fluid in the lamina propria of the mucosa, where lymphatic ducts and vagal afferents terminate. The findings are therefore consistent with the idea that IP administered GLP-1 acts on intestinal vagal afferents to inhibit eating. IP GLP-1 administration may therefore mimic a paracrine satiating effect of endogenous GLP-1 during meals. Further research is necessary to critically examine this hypothesis.

To more directly examine the possible physiological relevance of endogenous intestinal GLP-1, we next infused Ex-9 IP under comparable conditions. The currently available reports on the eating effects of peripheral GLP-1 antagonism are discrepant. While Williams et al. reported that IP injections of Ex-9 increased food intake under some, but not all, conditions, Kim et al. found that HPV or jugular vein infusion of another GLP-1 receptor antagonist did not increase food intake in rats. Because some evidence indicates that Ex-9 may more effectively stimulate eating after a carbohydrate preload or when the animals have already eaten, we infused Ex-9 during the second (instead of during the first, as for GLP-1) nocturnal meal. Under these conditions 10 and 30 nmol/kg Ex-9 did not increase meal size. Thirty nmol/kg Ex-9, however, effectively blocked the eating-inhibitory effect of IP infused GLP-1 (1 nmol/kg BW), indicating that the Ex-9 could reach the receptors that mediate the eating-inhibitory effect of exogenous GLP-1 in a sufficiently high concentration and early enough to antagonize this effect. Therefore, the most parsimonious interpretation of the failure of Ex-9 alone to alter food intake is that endogenous GLP-1 is not necessary for satiation under these conditions. Alternative explanations are that Ex-9 can block only pharmacological effects of GLP-1 under our conditions or that the dose of the antagonist that we used was still too low to block the effect of the endogenous peptide, i.e. that a higher dose of Ex-9 may be 4 necessary to increase normal meal size than to block the satiating effect of exogenous GLP-1. That peptide receptor antagonists block the eating-inhibitory actions of exogenous peptides more effectively than they block the actions of endogenous peptides has been reported in tests of CCK1 receptor antagonism.

Williams et al. reported an increase in 1 h food intake after IP injection of 100 ¦Ìg/kg BW Ex-9, which is a dose very similar to our 30 nmol dose. There are several possible explanations for this discrepancy. First, the failure of Ex-9 to increase meal size under our conditions may be a pharmacokinetic artifact, i.e., it is possible that Ex-9 reached the receptors too late or in too small a concentration to influence the effect of endogenous GLP-1. Another possible explanation is related to the different experimental set ups. Further studies are necessary to clarify this point.

Finally as GLP-1 has been shown to synergize with glucose to stimulate insulin release from the pancreas and as this effect appears to be mediated by vagal afferents originating in the hepatic portal area, we started to examine whether GLP-1 may also interact with glucose to inhibit eating. We found that subthreshold doses of GLP-1 and glucose combined to reduce meal size after intrameal IP infusion, suggesting that the synergistic stimulatory effect of GLP-1 and glucose on insulin release may extend to satiation. The physiological relevance of this observation warrants further investigation.

Summary and Conclusions

These findings show that intra-meal infusions of GLP-1 via different routes selectively reduce meal size and that abdominal vagal afferents are necessary for the full eating-inhibitory effect of IP, but not IV infused GLP-1. This is consistent with the hypothesis that circulating GLP-1 acts in the hindbrain, whereas IP administered GLP-1 acts on intestinal vagal afferents, to inhibit eating. At present, our data argue against a role of systemic increases in endogenous GLP-1 in satiation under our conditions. Further studies should therefore examine the physiological relevance of endogenous intestinal GLP-1 for satiation under our conditions as well as its mechanism and site of action.

Question & Answers:

Q. What is the pathway from the L-cell to the blood vessels?
A. In the small intestine, the L-cells release GLP-1 into the interstitial fluid of the lamina propria, from where GLP-1 can eventually enter the lymph vessels and capillaries.

Q. Yes, but how does the molecule (GLP-1) penetrate the capillary? Has this been determined yet?
A. I am not aware of such measurements, but would assume that GLP-1 can easily penetrate the capillary wall through its large pores..

Q. Do you mean that there's no carry over effect at the next meal at this dose of GLP-1?
A. That is correct.

Q. If you were interested in meal initiation, it seems to me that you'd want to look at the inter-meal interval.
A. I agree, but doing that is tricky because we cannot predict when a rat will spontaneously take its next meal. I wouldn't claim from these data that we fully know the effect of GLP-1 on the inter-meal interval.

Q. Would you presume that the lack of effect on inter-meal interval under your conditions is due to the quick degradation of GLP-1?
A. Yes, this is certainly one possible explanation.

Q. Is there any way to influence the inter-meal interval?
A. I am aware of two situations in which the inter-meal interval is affected: 1) Administration of 2-DG to antagonize glucose utilization or mercaptoacetate to block fatty acid oxidation. In response to both metabolic inhibitors rats will initiate a meal, i.e. shorten the inter-meal interval It is questionable, however, whether these effects are physiologically relevant for the mechanisms of spontaneous meal initiation. 2) Administration of bacterial lipopolysaccharide also lengthens the inter-meal interval.

Q. Data from Margolskee have shown that there are GLP-1 receptors co-expressed with taste receptors, and these might be important in the development of diabetes. Can you comment?
A. We cannot address the physiological relevance of the GLP-1 receptors on taste cells with our model.

Q. Does anyone know the time course of GLP-1 release into the brain.
A. I think it is unknown whether there is a short-term release of central GLP-1 similar to what we observed in the periphery.

Q. Is it possible that food intake is already at a maximum in your Ex-9 experiments?
A. The observed meal sizes are within a reasonable range for meal sizes at the beginning of the dark phase and I don't think that this was the reason for the lack of effect of Ex-9 on meal size.

Q. Do you do any experiments during the daylight? Would you expect similar results?
A. Rats mostly eat during the night, so we do most of our experiments during the dark cycle. I would, however, expect similar results during daylight.

Q. Have you looked at the effect of GLP-1 on specific macronutrient intake?
A. Not yet. We wanted to look at a natural feeding paradigm and not at changes after intragastric nutrient administration. But, we are in the process of looking at the effects of specific macronutrient meals on GLP-1 release now.

Q. How long would you expect it to take food to get to the ileum in the rat?
A. I'm not sure. Perhaps an hour or more?

Q. Are you saying that GLP-1 is broken down in the liver?
A. Yes; DPP-IV activity in the liver is very high. It is estimated that about half of the GLP-1 that reaches the liver is inactivated there.

Q. Is it possible that there is a source of GLP-1 in the brain?
A. Yes, GLP-1 is expressed in neurons of the NST.

Q. What happens to GLP-1 in the brain as a result of a chow meal?
A. I don't know and I'm not sure if someone has looked at that yet.

Q. Where are the receptors for GLP-1 in the brain?
A. They are in several brain areas, including areas that are implicated in the control of eating, such as the NST, the area postrema (AP), and the arcuate (Arc) and the paraventricular (PVN) nuclei of the hypothalamus..

Q. Is it possible that an IP injection of glucose could trigger GLP-1 release from the intestine?
A. I would assume so because GLP-1 release was also shown from isolated small intestine in response to vascular glucose perfusion and because IP administered glucose may gain access to the interstitial fluid of the lamina propria and might enter the L-cells from the serosal side through GLUT-2.

Q. Is GLP-1 implicated in any other behaviors?
A. Central GLP-1 is implicated in aversion.

Q. There were early studies showing that GLP-1 affected the CTA. What is the status on that?
A. I haven't seen any recent follow-ups to those studies.

Q. Is there a human application for GLP-1?
A. I'm not sure if I'm the right person to be asked, but as far as I know, GLP-1 agonists are presently being tested in clinical trials for the treatment of type-2- diabetes and obesity.

Q. Are there any indications that MUFA and PUFA (different fatty acids) have different effects on the release of GLP-1?
A. It's possible and I think there is some evidence for differential effects. Different fatty acid transporters and receptors (CD36 and GPR120) have been identified in the L-cells, but the effect of different fatty acids on GLP-1 release could of course also be indirect and mediated through some other gut peptide (e.g. CCK).