Peptide YY, Glucagon-Like Peptide-1, and Neurotensin Responses to Luminal Factors in the Isolated Vascularly Perfused Rat Ileum

Abstract

Exposure of the ileum to nutrients markedly inhibits several upper gastrointestinal functions. Hormonal peptides of the ileal wall, i.e. peptide YY (PYY), glucagon-like peptide-1 (GLP-1), and neurotensin (NT), are thought to play a role in this negative feedback mechanism. The present study was conducted to comparatively assess the secretion of PYY, GLP-1, and NT upon luminal infusion of a variety of individual luminal factors in the isolated vascularly perfused rat ileum preparation. PYY, GLP-1, and NT were measured in the portal effluent with specific RIAs. Glucose (250 mm) induced a pronounced release of the three peptides, whereas a physiological concentration of 5 mm did not induce peptide secretion. Peptone (5%, wt/vol) evoked a sustained release of PYY, GLP-1, and NT. Only NT secretion was increased upon luminal administration of 100 mm sodium oleate. Short chain fatty acids (20 mm) evoked an early and transient release of the three peptides. In contrast, taurocholate (20 mm) induced a sustained release of PYY, GLP-1, and NT, but the threshold concentration for peptide release was lower for NT than for PYY or GLP-1. Cellulose or pectin (0.5%, wt/vol) did not modify peptide secretion. In conclusion, glucose and peptone are potent stimulants of PYY, GLP-1, and NT release. Only NT is released upon oleic acid stimulation. Finally, taurocholate is a potent stimulant of the release of the three peptides. Overall, PYY, GLP-1, and NT may participate cooperatively in the ileal brake. As relatively high concentrations of the various stimulants were required to elicit peptide release, it seems likely that this mechanism operates in cases of maldigestion or malabsorption.

UNABSORBED nutrients reaching the ileum influence the functions of the upper gut to increase the efficiency of digestion and absorption. For example, unabsorbed fat or protein in the ileum delays the passage of material through the small intestine. As a delay in small bowel transit increases the contact time between luminal contents and absorptive epithelium, this mechanism may serve to increase the absorption of a meal. This negative feedback mechanism referred to as the ileal brake presumably involves hormonal intermediates. As the ileal mucosa contains several distinct populations of endocrine cells, many studies focused on the effects of the products of these cells on gastrointestinal functions. Neurotensin (NT), the major product of the so-called N cells, and peptide YY (PYY) and glucagon-like peptide 1 (GLP-1), which are cosynthesized in the so-called L cells, have been implicated in these feedback mechanisms.

Circulating levels of NT, PYY, and GLP-1 rapidly increase in response to oral ingestion of a mixed meal, thus suggesting that their release is triggered in part by hormonal and/or neural signals originating from the upper small intestine. As NT, PYY, and GLP-1 are contained in open-type cells, nutrients making contact with the ileal mucosa are also capable of eliciting peptide secretion. Indeed, high concentrations of glucose in the ileum stimulated the release of the three peptides (1–4). However, the influence of physiological concentrations of carbohydrate on hormone secretion is less documented. Conflicting data are available about the release of NT and PYY upon ileal infusion of fat. A strong release of these peptides was observed in some studies, whereas other reports indicated no effect (5–8). Only one recent study reported a stimulatory effect of ileal perfusion of fat on GLP-1 secretion (9). Similarly, the secretory activity of ileal N and L cells upon stimulation with undigested proteins is poorly documented. Bile salts in the ileum stimulate the release of NT and PYY (10, 11), but GLP-1 secretion has not been investigated. The effect of the ileal administration of short chain fatty acids (SCFAs) was only studied on PYY secretion, and no release was observed (12). Finally, large amounts of fibers reach the ileum, but their ability to stimulate the secretory activity of ileal N and L cells has not yet been explored. Overall, little is known about the comparative responsiveness of these two cell types upon stimulation with the individual components of the intestinal chyme. This was the purpose of the present study conducted with the isolated vascularly perfused rat ileum preparation (13). This model provides a unique opportunity to study the secretion of NT, PYY, and GLP-1 in response to well defined luminal stimuli in a manner that eliminates the influences potentially encountered in vivo.

Materials and Methods

Materials

BSA was purchased from Biovalori (Cassen, France). Azonutril 25, a mixture of amino acids, was obtained from Laboratoires Roger Bellon (Neuilly-sur-Seine, France). This solution consists of 3.4% (wt/vol) isoleucine, 9.3% leucine, 8.5% lysine, 6.3% methionine, 8.3% phenylalanine, 3.4% threonine, 1.7% tryptophan, 8.4% valine, 2.7% aspartic acid, 3.4% glutamic acid, 6.4% alanine, 16.8% arginine, 1% cysteine, 6% glycine, 3.4% histidine, 5.4% proline, 0.9% serine, 0.2% tyrosine, 2% citrulline, and 1.5% ornithine. The total amino acid content was 14.8 g/100 ml mixture. The following reagents were obtained from Sigma Chemical Co. (Saint-Quentin-Fallavier, France): cellulose, pectin from apple, propionate, n-butyrate, and taurocholate. Oleic acid and glucose were purchased from Merck (Darmstadt, Germany). ¹²⁵Iodine was supplied as sodium iodine from Amersham (Les Ulis, France). Peptone, purchased from Sigma, was an ovalbumin enzymatic hydrolysate containing 31% free amino acids and 69% peptides. The molecular weight of the peptides ranged from 120-5000.

Surgical preparation

The surgical steps and functional viability of the isolated vascularly perfused rat ileum have been previously reported in detail (13). Male Wistar rats (250–300 g) were anesthetized with sodium pentobarbitone (50 mg/kg, ip), and the abdomen was opened with a midline incision. The right and middle colic veins and arteries were tied and cut between ligatures near the serosa of the colon to free the upper small intestine from the hindgut. A cannula was inserted and tied into the terminal ileum, and another one was inserted into the ileum 10 cm proximal to the first. The gut lumen was flushed twice with 10 ml isotonic saline, then the remaining jejunum and duodenum were removed after their respective arteries and veins were ligated. A metal cannula and a silicone elastomer tubing were then quickly inserted in the superior mesenteric artery and portal vein, respectively. The arterial perfusion started immediately at a rate of 2.5 ml/min with an oxygenated Krebs-Henseleit buffer (pH 7.4) containing 25% washed bovine erythrocytes, 3% BSA, 5 mm glucose, and 1% Azonutril (vol/vol). The perfusion pressure, continuously recorded with a mercury manometer, ranged from 40–60 mm Hg. The preparation was removed and transferred to a bath at 37 C. Venous blood effluent was collected as 2-min fractions in tubes containing 250 μl 300 mm EDTA. The supernatant was rapidly separated from erythrocytes by centrifugation and frozen as 1-ml fractions at −20 C for subsequent determinations of PYY and GLP-1. Before the determination of NT, portal supernatant (1 ml) was treated with 2 ml ethanol. The ethanol extracts were dried and kept at −30 C.

Experimental protocol

The experiments consisted of a 20-min control basal period during which isotonic saline was infused into the lumen at a rate of 250 μl/min. This was followed by a 30-min period of stimulation of peptide release and a subsequent 10-min control period. Each luminal component was administered first as a bolus of 2 ml followed by a slow infusion rate of 250 μl/min for 29 min. The lumen was then flushed out with air followed by an infusion of isotonic saline at a rate of 200 μl/min. The pH of each infused compound was adjusted to 7–7.5, and the osmolarity was increased when required to 300 mosmol/kg H₂O by addition of appropriate amounts of sodium chloride. The amino acid mixture Azonutril 25 was diluted 5-fold in water to obtain a final osmolarity of 300 mosmol/kg H₂O. Oleic acid was infused as a soap after adjusting the pH of the solution to 7.8 with NaOH.

RIAs

NT-like immunoreactivity (NT-LI) in the portal effluent was measured with an antiserum specific for intact NT, as previously described (13, 14). Briefly, antiserum 29G was obtained in a rabbit after repeated injection of NT conjugated to BSA and was used at a final dilution of 1:200,000. The antiserum cross-reacted 100% with intact NT, 75% with NT-(4–13), 45% with NT-(6–13), and less than 0.1% with N-terminal fragments 1–12, 1–11, 1–10, 1–8, and 1–6 of NT. The radioactive ligand was mono-iodo-[¹²⁵I-Tyr³]NT, labeled and purified as described by Holst-Pedersen et al. (15). The sensitivity and ID₅₀ were 0.6 and 3.0 fmol/tube, respectively. HPLC analysis of portal supernatants followed by RIA with antiserum 29G revealed a single peak coeluting with NT-(1–13) (13).

The RIA for PYY in portal effluent was performed as previously described with antiserum A4D obtained from a rabbit after repeated injection of synthetic porcine PYY conjugated to BSA through ethylcarbodiimide condensation (16). This antiserum, which cross-reacted less than 0.1% with bovine pancreatic polypeptide and NPY, was used in the assay at a final dilution of 1:800,000. The synthetic peptide was iodinated with carrier-free Na¹²⁵I by means of the chloramine-T reagent and was purified by reverse phase HPLC as previously described (16). The minimum detectable concentration of PYY and the ID₅₀ of the assay were 1 and 7 fmol/tube, respectively. Portal supernatants run on a Sephadex G-50 column revealed a single immunoreactive peak coeluting with the synthetic peptide.

The GLP-1 assay was performed as recently described (17, 18). Briefly, antiserum against GLP-1-(7–36) amide was obtained in a rabbit by immunization with synthetic GLP-1-(7–36) amide conjugated to BSA and was used at a final dilution of 1:300,000. The reactivity of the antiserum 199D was 100% for GLP-1-(7–36) amide, 84% for GLP-1-(1–36) amide, and less than 0.1% for GLP-1-(1–37), GLP-1-(7–37), GLP-2, glucagon, secretin, vasoactive intestinal peptide, and GIP. The synthetic GLP-1-(7–36) amide was radioiodinated using the chloramine-T method and purified by reverse phase HPLC. The detection limit and ID₅₀ were 0.6 and 4.5 fmol/tube, respectively. Gel filtration on a Sephadex G-50 column revealed in the portal effluent the presence of a single peak corresponding to the positions of synthetic GLP-1-(7–36) amide and GLP-1-(1–36) amide, which were indistinguishable in the present system.

Calculations and statistics

Data in all figures are presented as the mean ± se and are expressed as femtomoles per 2 min. The integrated responses of immunoreactive material released by a given stimulus were calculated by subtraction of the basal immunoreactivity produced during a given period from the immunoreactivity released upon stimulation during the same period. For statistical analysis of the data, Student’s t test for paired or unpaired values was used where appropriate.

Results

Release of PYY, GLP-1, and NT by nutrients

Luminal infusion of 250 mm glucose induced a prompt release of PYY-LI (peak of 150.0 ± 8.6 fmol/2 min at 2 min from a basal level of 28.7 ± 6.9 fmol/2 min; P < 0.05) followed by a sustained secretion at a mean plateau value of 120 fmol/2 min (Fig. 1). Similarly, the concentration of GLP-1-LI in the portal effluent was markedly increased upon luminal stimulation with 250 mm glucose (mean plateau value of 60 fmol/2 min from a basal level of 9.7 ± 1.6 fmol/2 min; P < 0.05; Fig. 1). The pattern of NT-LI release was characterized by an early peak secretion (65.7 ± 11.5 fmol/2 min at 2 min from a basal level of 11.9 ± 2.5 fmol/2 min; P < 0.05) followed by a return to a low plateau value of 30 fmol/2 min (P < 0.05; Fig. 1). Luminal infusion of 5 mm glucose only induced a transient release of PYY and NT (Fig. 1).

Figure 1.

Effects of luminal administration of 5 mm (n = 6) or 250 mm (n = 6) glucose on release of PYY (top), GLP-1 (middle), and NT (bottom) in the portal effluent of the isolated rat ileum preparation. Results are expressed as femtomoles of PYY-, GLP-1-, and NT-LI released per 2 min ± se.

Ileal infusion of peptones (5%, wt/vol) rapidly elevated PYY levels in the portal effluent (early peak of 140.6 ± 42.1 fmol/2 min from a basal level of 26.9 ± 2.2 fmol/2 min; P < 0.05), and this was followed by a release at a high plateau value of about 80 fmol/2 min (Fig. 2). The pattern of GLP-1 release upon administration of peptones was similar to that of PYY: a peak of 88.1 ± 21.8 fmol/2 min from a basal level of 27.3 ± 7.0 fmol/2 min (P < 0.05) followed by a gradual rise to a high level (maximal value of 72.9 ± 14.9 fmol/2 min at the end of the stimulation period; Fig. 2). Peptones also induced a marked increase in the NT-LI concentration in the portal effluent (peak at 1000% of the basal level followed by a plateau value of 550% of basal; P < 0.05; Fig. 2). The mixture of amino acids (total concentration of 250 mm) was a weak stimulant of PYY, GLP-1, and NT release [integrated responses, 158 ± 67 (P < 0.1), 148 ± 43 (P < 0.05), and 90 ± 17 fmol/30 min (P < 0.05), respectively; n = 5 in each set of experiments].

Figure 2.

Effects of luminal administration of peptones (5%, wt/vol) on the release of PYY, GLP-1, and NT in the portal effluent of the isolated vascularly perfused rat ileum (femtomoles per 2 min ± se of seven experiments).

Luminal infusion of oleate (100 mm) did not elicit any significant release of PYY and GLP-1 (Fig. 3). In contrast, the level of NT-LI in the portal effluent increased early after luminal placement of oleate (peak of 32.9 ± 10.0 fmol/2 min at 2 min from a basal level of 9.3 ± 0.8 fmol/2 min; P < 0.05). A sustained secretion of NT-LI was observed thereafter, reaching a plateau value of about 20 fmol/2 min (P < 0.05; Fig. 3). A lower concentration of oleate (20 mm) did not modify the basal levels of PYY, GLP-1, and NT (data not shown).

Figure 3.

Effects of luminal administration of sodium oleate on the release of PYY, GLP-1, and NT. Results are the mean ± se of seven experiments.

Release of PYY, GLP-1, and NT by taurocholate, SCFAs, and fibers

At a concentration of 10 mm, taurocholate, the major bile salt in rats, did not induce a significant release of PYY and GLP-1 (Fig. 4). In contrast, the level of NT-LI in the portal effluent was significantly elevated upon administration of the same dose of taurocholate. The basal level of NT-LI (4.2 ± 0.2 fmol/2 min) reached a plateau value of 20 fmol/2 min (P < 0.05) 8 min after the start of the stimulation (Fig. 4). When 20 mm taurocholate was administered, portal PYY, GLP-1, and NT promptly increased to plateau values of 670%, 220%, and 520% of basal, respectively (P < 0.05; Fig. 4). At the end of the stimulation period, portal PYY, GLP-1, and NT concentrations rapidly decreased to near-basal values.

Figure 4.

Release of PYY-, GLP-1-, and NT-LI evoked by luminal administration of 10 mm (n = 6) or 20 mm (n = 6) taurocholate in the isolated vascularly perfused rat ileum. Results are the mean ± se.

At a concentration of 20 mm, butyrate induced an early and transient release of PYY, GLP-1, and NT (peaks at 400%, 220%, and 730% of the basal levels, respectively) with a subsequent decline to basal values at 4 min (Fig. 5). The same pattern of PYY, GLP-1, and NT secretion was observed upon administration of propionate at the same concentration (Fig. 6). In contrast, 5 mm butyrate or propionate did not induce a significant release of PYY, GLP-1, and NT (data not shown).

Figure 5.

Release of PYY-, GLP-1-, and NT-LI induced by luminal administration of 20 mm butyrate. Results are the mean ± se of four experiments.

Figure 6.

Effects of luminal administration of 20 mm propionate on PYY-, GLP-1-, and NT-LI secretion. Results are the mean ± se of five experiments.

Neither cellulose (0.5%, wt/vol) nor pectin (0.5%, wt/vol) significantly modified the basal levels of portal PYY, GLP-1, and NT immunoreactivities (data not shown).

Discussion

Glucose in the ileum induced a marked release of PYY, GLP-1 and NT. These results are in agreement with those obtained from in vivo studies and from experiments with the isolated vascularly perfused ileum preparation (1–4). However, the concentrations of luminal glucose that are required to elicit peptide responses are supraphysiological. Indeed, the luminal glucose concentration in the distal part of the small intestine is 0.6–1.2 mm under physiological conditions (19). Thus, it is likely that glucose in the ileum does not contribute significantly to NT, PYY, and GLP-1 release under physiological circumstances.

Completion of fat absorption by the proximal small intestine has been widely accepted since the pioneering studies of Borgstrom et al. (20). However, recent data indicate that, even after usual meals, absorption of fat is not complete by midgut (21). Fatty acids reaching the distal small intestine may therefore modulate the secretory activity of the L and N cells. Only one recently published study focused on the effect of ileal administration of fatty acids on GLP-1 release (9). This study, performed in humans, indicated that ileal lipid perfusion had stimulatory effects on GLP-1 release. In contrast, the present work showed that oleic acid did not modify basal GLP-1 immunoreactivity in the isolated rat ileum system. Together, these data suggest an indirect mechanism rather than a direct stimulation of the ileal GLP-1-containing cells by luminal fats in vivo unless species to species variations may account for this divergence. Similarly, intraileal administration of sodium oleate in rats promptly released PYY (5, 12), whereas no stimulatory effect was observed here, thus strengthening the hypothesis that luminal placement of fatty acids in the ileum induces PYY release through an indirect hormonal and/or neural pathway.

In contrast to the lack of stimulatory effect of sodium oleate on the ileal L cells, the present study showed a significant release of NT in the portal effluent of the isolated rat ileum upon oleate challenge. The effect of fatty acid administration in the distal small intestine on NT release in vivo is not clear. A study performed in the dog showed no effect of intraileal sodium oleate on NT release (8), whereas another work in the same species demonstrated NT release after luminal placement of oleic acid in micellar aqueous solution (22). Finally, oleic acid induced NT secretion from isolated canine ileal N cells (23). As NT inhibits gastric emptying and slows intestinal transit, it seems likely that this peptide may participate in the fat-induced ileal brake.

Little is known about the influence of protein hydrolysates on the secretory activity of ileal L and N cells. A potent stimulatory effect of chicken egg hydrolysate on GLP-1 secretion is here described for the first time in the isolated vascularly perfused rat ileum. Additionally, PYY and NT were released. In contrast, plasma PYY levels were elevated by ileal perfusion of casein hydrolysate in dogs, whereas plasma NT concentrations were not modified (24). In humans, another study showed that plasma levels of PYY and GLP-1 were not altered after ileal perfusion with peptone (9). As the doses of peptones used in these different studies were similar, it may be speculated that these discrepancies result from species to species differences.

Pectin, a polygalacturonic acid polymer, was shown to be a potent stimulant of GLP-1 and PYY release in the isolated vascularly perfused rat colon (25, 26). In contrast, no stimulatory effect of pectin was detected in the present study with the ileal preparation, thus suggesting regional differences in the sensitivity of L cells to pectin. On the contrary, cellulose, another dietary fiber, was inefficient at stimulating PYY and GLP-1 release in the isolated rat ileum as well as in the isolated rat colon (25, 26).

The luminal bile salt concentration in the rat duodenum is about 10 mm. Although passive absorption occurs in the jejunum (27), the concentration of bile salts increases to 20–30 mm in the distal jejunum and proximal ileum (28). Ileal active transport reduces the luminal bile salt concentration to 2–3 mm in the terminal part of the small intestine (28). Our previous work performed with the isolated vascularly perfused rat jejuno-ileum showed that bile salts administered at the concentration found in the distal jejunum are potent stimulants of NT release (10). The threshold concentration of taurocholate required for NT release was approximately 10 mm. The present study additionally revealed that taurocholate potently stimulated the release of both PYY and GLP-1 in the ileum. However, the threshold concentration of taurocholate required for GLP-1 and PYY release was approximately 2-fold higher than that required for significant NT release, thus suggesting a slight difference in the sensitivity of L cells vs. that of N cells to taurocholate. Interestingly, luminal bile in the isolated perfused rabbit terminal ileum increased the release of PYY (11). The physiological meaning of the bile salt-induced PYY and GLP-1 release from the distal small intestine is unknown. As 1) bile salts stimulate water and electrolyte secretion, and 2) PYY displays potent inhibitory effects of stimulated rat water flux in the small intestine (29, 30), it is tempting to speculate that locally released PYY upon bile salt challenge could restrain the output of water and electrolytes. Additional experiments with the isolated vascularly perfused rat small intestine are required to validate this hypothesis.

SCFAs are produced by bacterial anaerobic fermentation of carbohydrates in the forestomach of ruminants and in the hindgut of monogastric animals. They accumulate in concentrations up to 150 mm in the human colon. These acids are also present in the small intestine, but their concentration seldom exceeds 1–5 mm. Over this range of concentration, butyrate or propionate did not elicit any significant release of NT, GLP-1, or PYY in the present study. A solely transient release of the three peptides upon stimulation with 20 mm butyrate or propionate was observed. Our results agree with a previous in vivo study that reported that a mixture of SCFAs injected into the rat ileum released insignificant amounts of PYY, whereas it was potent in the colon (12). For comparison, SCFAs caused a release of PYY from the isolated vascularly colon of rats and rabbits (26, 31). Interestingly, none of these SCFAs produced any release of GLP-1 in the rat colonic preparation (25).

In conclusion, we observed with the isolated vascularly perfused rat ileum preparation a systematic corelease of PYY, GLP-1, and NT. Moreover, these findings support the idea that a variety of chyme components making contact with mucosa containing L and N cells stimulates the release of GLP-1, PYY, and NT. These peptides may, in turn, cooperatively participate in the late postprandial regulation of several functions of the upper part of the gut to improve digestion and absorption of nutrients mainly in cases of maldigestion or malabsorption.