Glucagon-like peptide-1 (GLP-1) modulates glucose levels following a meal, including by inhibition of gastric emptying and intestinal transport. Intra-arterial injection of GLP-1 into the gastric corpus, antrum, or pylorus of anesthetized dogs had no effect on the contractile activity of the resting or neurally activated stomach. GLP-1 injected intra-arterially inhibited intestinal segments when activated by enteric nerve stimulation but not by acetylcholine. Isolated ileum segments were perfused intra-arterially, instrumented with strain gauges to record circular muscle activity and with subserosal electrodes to stimulate enteric nerves. GLP-1 caused concentration-dependent inhibition of nerve-stimulated phasic but not tonic activity. This was absent during TTX-induced activity and partly prevented byN G-nitro-l-arginine. Exendin-(9—39), the GLP-1 antagonist, had no intrinsic activity and did not affect the actions of GLP-1. Capsaicin mimicked the effects of GLP-1 and may have reduced the effect of subsequent GLP-1. GLP-1 may mediate paracrine action on afferent nerves in the canine ileal mucosa using an unusual receptor.
- gastric emptying
- intestinal motility
- afferent nerves
- paracrine action
glucagon-like peptide (GLP)-1 is produced in L-type endocrine cells of the ileal and colonic mucosa, where it is usually colocalized with peptide YY and GLP-2 (8, 20, 21, 24, 42, 49, 53, 68). It contains sequence similarities with glucagon and is derived by alternate processing of a common precursor, proglucagon (amino acids 78–107). It is released by glucose, fatty, and peptone meals (3, 20, 21, 42, 49), which is constrained by somatostatin (41). GLP-1 acts as an incretin (51, 60, 62) with multiple actions on pancreatic β-cells: it raises cAMP, inhibits the ATP-sensitive K+ channel, depolarizes, enhances Ca2+ entry, and promotes insulin release (40,46). It also inhibits release of glucagon from pancreatic α-cells, possibly by increasing release of somatostatin by D cells (20, 21, 42, 49). When the action of GLP-1 is inhibited, glucose tolerance is impaired (11). Receptors for GLP-1 are present in the periphery as well as the central nervous system, in which it may act to modulate fluid and food intake (7,10, 15, 18, 19, 39, 48, 52, 72). GLP-1 receptors are also present on afferent nerves, and several of its actions are postulated to involve activation of vagal afferents or those from the dorsal roots (45, 54, 70).
Until now, most receptors have been characterized by their susceptibility to agonists and antagonists and in several cases by cloning and expression of the receptor in cell lines. Only one cloned GLP-1 receptor has been found (7, 10, 15, 18, 39, 52, 72), but there is one previous case in which receptors have distinctly different pharmacology, i.e., insensitivity to the antagonist exendin-(9—39), as previously described (54).
GLP-1 has important effects on gastrointestinal motility. It inhibits gastric emptying and intestinal transit (4, 7, 48, 57,66, 67) and is postulated to be involved in ileal braking of gastric emptying and intestinal transit following glucose (5, 22,47, 50, 71), fat (37, 38, 59, 61), and peptone (38) meals. The locations of receptors for these actions have not been identified, but they are known to require an intact vagus nerve (36, 44, 70). We demonstrated in awake instrumented dogs that GLP-1 inhibits gastric emptying of a saline meal by decreasing the number and volume of flow pulses through the pylorus. These changes were associated with inhibition of antropyloric pressure waves, stimulation of isolated pyloric pressure waves, and an increase in pyloric tone (4).
The half-life of GLP-1 in the plasma is short (1–2 min) as it is rapidly degraded by dipeptidyl peptidase IV, and it is uncertain whether the active peptide reaches all of its receptors in an intact form (6, 9, 16, 17, 43, 55, 56, 63). Therefore, we hypothesize that it might have a local action in the ileum. The objectives of this study were to determine whether GLP-1 acted on the canine ileum, a site at which it is released in the canine intestine, and to determine the site and nature of any such actions.
Random source dogs were studied. Animals of both sexes were used if they were in good health and had been treated in our animal facility for parasites. They were anesthetized with 40 mg/kg pentobarbital sodium and maintained at a level of anesthesia in which they did not respond to pain and had no eye-blink reaction to touching the cornea. They were respired artificially throughout the experiment. They were euthanized after surgery and completion of the experiment with intraveous hypertonic KCl. These procedures were approved by the McMaster University Animal Care Committee.
Preparation of isolated segments.
This has previously been described in detail (13, 14, 32-35,38, 45, 69) and is summarized briefly. The abdomen was opened with a midline incision, and the ileum was brought to the exterior. A segment of 10–15 cm was isolated completely from the body by cannulating a terminal artery and vein, perfusing the artery at 2.7 ml/min with physiological saline, collecting the venous effluent, cutting all connection by mesentery, nerves, and vasculature to the body, and separating the segment from the intestine by double ties at each end. At the distal end, a drainage cannula was inserted into the lumen through a stab wound and tied in place with a purse-string suture. Luminal drainage was by gravity. Two strain gauges were sewed to the serosa to record circular muscle contraction, one 2–3 cm from the proximal and the other 2–3 cm from the distal end. Two pairs of 1-cm-long silver stimulating electrodes were inserted subserosally oriented in the circumferential axis. Each pair was 0.5–1 cm distal to the strain gauges to initiate ascending excitation. The exposed ileal segment was wrapped in paraffin-soaked gauze and warmed with a lamp. The animal was warmed with a heater in the operating table.
Close intra-arterial perfusion.
The techniques for perfusing the gastric corpus, antrum, pylorus, and duodenum have been previously described (2, 25-27,29-31). Briefly, in animals anesthetized as described above, an artery close to the organ was selected and cannulated with a small cannula containing heparinized physiological solution. The area perfused was delineated with a perfusion of physiological solution, a strain gauge was attached to record circular muscle contractions, and electrodes pairs were inserted 0.5–1 cm below in cirumferential orientation.
In all cases, a recovery period of 30 min was followed by perfusion of increasing concentrations of acetylcholine until a maximum response was attained. Other contractions were evaluated in terms of the percent maximal response to acetylcholine. Acetylcholine was also infused intermittently and after completion of the experiment to ascertain that contractility remained unimpaired. If the response to a maximal dose was decreased to <80%, the experiment was discarded. After a further rest period of 10–15 min, the tissues were stimulated with electrical field stimulation (EFS) of the electrodes distal to the strain gauges (3 pulses/s, 40 V/cm, 0.5-ms duration). In isolated segments, EFS was continued for 6 min, with a change in the polarity of the stimulating electrodes after 3 min. The two sets of electrodes (below the proximal and distal strain gauges) were each stimulated alternately for 6 min throughout the experiment, and data from both electrode sites were combined to study the time of effects of GLP-1. EFS was carried out initially until stable responses were obtained. GLP-1 infusions in physiological solution, by way of a T tube in the intra-arterial cannula, were found in preliminary experiments to have a delayed the onset of action, partly due to the dead space in the perfusion tubing but mostly due to the inherent action of the agent. In experiments with close intra-arterial infusion, actions of GLP-1 were studied on responses to EFS as described above or on responses to repetitive infusions of a submaximal dose of acetylcholine (35). The physiological solution contained (in mM) 115.5 NaCl, 21.9 NaHCO3, 4.6 KCl, 1.16 MgSO4 · 7H2O, 1.16 NaH2PO4 · H2O, 2.5 CaCl2, and 11.1 glucose, maintained at 37°C and bubbled with 95% O2-5% CO2.
Records of the second minute of each recording at a given polarity of stimulation were analyzed (during the initial minute, contractile activities were unstable). Tonic tension data were analyzed as area under the contraction curve as previously described (13). Phasic tension increments were analyzed separately from tonic increments by summing each contraction during the second minute and averaging the value. To obtain data on the duration of action of GLP-1, responses both from proximal and distal electrodes and at both polarities of stimulation were combined.N G-nitro-l-arginine (l-NNA) infusion increased tonic and phasic activity. The effect of l-NNA on responses to GLP-1 was studied by evaluating the percent decrease in the increment of activity caused by EFS after l-NNA vs. that in the increment beforel-NNA.
Data were usually analyzed by paired comparisons with Dunnett's correction when multiple comparisons were made (Prism 3 software). When the effects of l-NNA on responses to GLP-1 were assessed, the comparison was between the percent inhibition of responses to the same concentration of GLP-1 before l-NNA compared with responses after l-NNA.
Drugs and reagents.
GLP-1 and exendin-(9—39) amide were obtained from Bachem Bioscience (Philadelphia, PA). C capsaicin and the capsaicin receptor antibody were from Calbiochem-Novabiochem (San Diego, CA). Other drugs and chemicals were from Sigma (St. Louis, MO).
In vivo GLP-1 caused marked inhibition of gastric emptying in dogs fed a saline meal. This resulted from increased pyloric tone and inhibition of propulsive antral contractions (4). Therefore, we initially evaluated the effects of intra-arterial infusion of GLP-1 (0.1–100 ng) into the antral and pyloric regions of anesthetized dogs at rest and during stimulation of enteric and vagal nerves as previously described (2, 26). There were no effects (data from 6 experiments not shown). We also infused GLP-1 intravenously while activating enteric nerves, again without effect. These results shifted our focus to the ileum.
When GLP-1 (10−11–10−9 M) was infused intra-arterially into the ileum (n = 6) during stimulation from serosal electrodes placed 0.5–1 cm distal to the strain gauge over the perfused site (to elicit the ascending excitatory reflex), there was persistent inhibition of the phasic activity elicited by EFS (Fig. 1 B). A similar GLP-1 infusion had no significant effect on responses to acetylcholine given intra-arterially every 2 min. These studies (n = 4) also showed that, under resting conditions, GLP-1 perfusions were without significant contractile effect (Fig.1 A). We observed that recovery of phasic activity in response to EFS after intra-arterial infusions of GLP-1 was delayed for long periods.
To understand the mechanisms involved, we used isolated arterially perfused ileal segments instrumented with two strain gauges, each with a distal pair of serosal electrodes, stimulated alternately as described in methods. Infusion of GLP-1 inhibited responses to EFS of these electrodes applied alternately to the proximal and distal pair. Figure 2 shows an example of the concentration-dependent (10−8–3 × 10−7 M) effects of GLP-1 infusions. Infusions were carried out for 8 min and studied for at least 6 min after infusions were stopped on responses to EFS. Maximum inhibition was obtained with 10−7 M GLP-1 (10−9 M had no consistent effect). Figure 3 summarizes experiments of this type, demonstrating that GLP-1 primarily inhibited phasic but not tonic responses to EFS.
To define the duration of action of GLP-1, experiments were carried out using a concentration of 10−7 M for 8 min. EFS was carried out alternately at both proximal and distal strain gauges, and the effects were assembled (Fig. 4). The inhibition of phasic activity extended for >22 min (>15 min after perfusion was stopped), whereas there were no consistent effects on tonic contractile activity.
Exendin-(9—39) amide (3 × 10−8 M) infused for 14 min before and during infusion of either 3 × 10−8 or 10−7 M GLP-1 had no ability to interfere with the inhibitory action of GLP-1 on phasic activity in response to EFS (Fig.5). GLP-1 infused after exendin-(9—39) had no effect on tone as usual (not shown). Exendin-(9—39) itself had no effects on phasic responses to EFS (Fig.6) or tonic responses (not shown).
When TTX (10–20 μg) was infused and responses to EFS were abolished, persistent phasic and tonic activity ensued as previously described (12, 35). GLP-1 had no effect on this activity in concentrations through 3 × 10−7 M (data not shown), indicating that it had no effect on myogenic activity, consistent with its lack of inhibitory effect on responses to intra-arterial acetylcholine. Perfusion of l-NNA (10−4 M) also induced persistent tonic and phasic activity, but responses to excitatory stimulation of proximal enteric nerves were preserved, as previously described (12). GLP-1 had a reduced inhibitory effect on these responses, suggesting that nitric oxide release mediated its effects in part (Fig.7), and the normalized results were significantly less reduced than in the same tissues perfused with GLP-1 before l-NNA (P < 0.03).
Capsaicin infused at 10−6 or 3 × 10−6 M had no effects on responses to EFS, but at 10−5 M, it, similar to GLP-1, caused significant and persistent inhibition of phasic but not tonic responses to EFS (Fig.8). We attempted to achieve desensitization of responses to capsaicin by subsequent infusions of capsaicin after an infusion of 10−5 M, but we were unable to obtain persistent lack of effect to a subsequent infusion; i.e., when the phasic responses recovered sufficiently to test whether GLP-1 could inhibit it, capsaicin was also capable of reducing it. After capsaicin, GLP-1 still causes inhibition of the residual phasic responses to EFS (Fig.9), but the magnitude appeared to be less (nonsignificantly). There was no effect of GLP-1 on tone after capsaicin (not shown).
The main findings of this study are that GLP-1 acts locally in the ileal region of the gastrointestinal tract near its release sites to inhibit activity of the enteric nervous system initiated by local activation of the ascending excitatory peristaltic reflex. Under the conditions of our experiments, using an isolated-perfused ileal segment, no extrinsic neural or hormonal influences are involved. GLP-1 causes concentration-dependent persistent inhibition of reflex neural activity, abolished by TTX, reduced by l-NNA, and without any effect on responses to exogenous acetylcholine.
The most surprising and interesting finding was that these effects of GLP-1 were unaffected by prior infusion of exendin-(9—39) amide, an antagonist of the actions of GLP-1 at the known receptor (7, 10,15, 18, 39, 52, 72). This receptor mediates most of the known actions of GLP-1, e.g., actions to enhance pancreatic β-cell growth and release insulin (22, 23, 40, 46, 50, 51, 60,62), inhibition of gastric emptying, secretion, and intestinal motility (37, 38, 44, 59, 66, 67), central actions on the dorsal vagal nucleus and other central nervous system sites (58,65). In the current literature, there is only one reported action of GLP-1 that was exendin-(9—39) insensitive, the activation of hepatic vagal afferent nerves by intraportal infusion of GLP-1 or exendin-4 (54). Thus our data as well as those of Nishizawa et al. (54) suggest that a neural receptor class for the action of GLP-1 may exist. One caveat is that we did not demonstrate that our exendin-(9—39) was bioactive. In anesthetized animals, there is no readily available test of bioactivity. Thus, although three different batches from the supplier, Bachem, behaved the same in our experiments, it is possible that they were not bioactive. Another caveat is that we did not go to higher concentrations of exendin-(9—39) because of the very high costs that would have resulted from perfusion before and during the exposure to GLP-1 in multiple experiments. However, we believe that our concentrations were sufficient, at least, to affect the actions of GLP-1 based on reports in the literature (e.g., Ref. 54).
Interestingly, the action of GLP-1, persistent inhibition of phasic activity, was mimicked by capsaicin in a concentration-dependent manner. Capsaicin may have reduced the inhibitory action of subsequent GLP-1, but it clearly did not abolish it, and the change was not significant. This was probably because under our conditions, we were unable to demonstrate persistent desensitization of responses to capsaicin; i.e., when the phasic responses to EFS recovered after capsaicin enough to test the inhibitory effects of GLP-1, the inhibitory response to capsaicin had also recovered. Thus the interpretation of these results must be restrained. Suggestive evidence was that some L cells containing GLP-1 were near sites of immunoreactivity to the capsaicin receptor, presumably on sensory nerves (unpublished observations).
The site of action of GLP-1 released from ileal L cells is unclear, because the peptide is rapidly degraded by a peptidase located in the intestine and on endothelial cells (1, 6, 9, 16, 17, 43). Whether or not GLP-1 released at this site should be considered as having a hormonelike action under physiological conditions is unclear. It may function as a paracrine agent. Thus it is possible that a local site of action near the L cell, where GLP-1 would be present in higher concentrations before degradation, may participate in the physiological role of GLP-1 and may do so by activation of a unique GLP-1 receptor. Clearly, further study is required to evaluate it importance. In this connection, the possibility that the action is on sensory nerves, because it appears that all GLP-1 actions on gastrointestinal motor function operate through neural connections and are not local to the sites of the response, such as the stomach and small intestine, needs to be considered.
We acknowledge the valuable technical assistance of C. Gill and Z. Woskowska.
This research was supported, in part, by the Medical Research Council of Canada and by the Father Sean O'Sullivan Research Fund.
Address for reprint requests and other correspondence: E. E. Daniel, Dept. of Pharmacology, Univ. of Alberta, 9–70 Medical Sciences Bldg., Edmonton, AB, Canada T6G 2H7 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
May 15, 2002;10.1152/ajpgi.00110.2002
- Copyright © 2002 the American Physiological Society