INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

GLUCAGON-LIKE PEPTIDE-1 (GLP-1) is an intestinal hormone released after nutrients from L cells located in the ileum and colon (24). GLP-1 participates in the ileal brake mechanism causing inhibition of gastric emptying and proximal intestinal motility (41), inhibits gastric acid secretion (10, 36), and is one of the potent incretin hormones enhancing glucose-dependent insulin secretion (7, 18). Thus GLP-1 plays an important role in the absorption and assimilation of nutrients. Several factors have been implicated in the stimulation of GLP-1 secretion including direct effects of nutrients, predominantly glucose and fat (15, 32); neuropeptides, including gastrin-releasing peptide (GRP) (8); and in rodents, circulating hormones, such as glucose-dependent insulinotropic peptide (29). Much less is known about the mechanisms involved in counterregulation of GLP-1 secretion. Somatostatin has been reported to inhibit GLP-1 secretion in certain species (14, 20, 21). However, within the gastrointestinal tract, differential posttranslational processing leads to two principle bioactive molecular forms of somatostatin: somatostatin-14 (S-14) and somatostatin-28 (S-28) (37). S-14 is localized to the foregut and enteric nervous system, whereas S-28 is found predominantly in the mucosa of the ileum and colon (28), similar to GLP-1. The mechanisms regulating S-28 secretion (2, 12, 13) and certain of its inhibitory actions (9, 42) are distinct from S-14, suggesting that the physiological roles of the two peptides are different. The biological effects of S-14 and S-28 are mediated by five G protein-coupled receptors [somatostatin receptor (SSTR) subtype 1 to SSTR5], expressed in most tissues including the gastrointestinal tract (17). Although S-28 and S-14 bind with similar affinity to SSTR1 to SSTR4, SSTR5 is characterized by its preferential affinity for S-28 (22, 23, 27). The recent availability of agonists showing relatively selective affinities for the SSTRs have provided new insights into the somatostatin receptor subtypes regulating the secretion of certain hormones (11, 38, 39) and digestive functions (5, 11, 35, 43).

The proximity of D cells producing S-28 and L cells containing GLP-1 in the ileum suggests that S-28 acting through SSTR5 may participate in the direct regulation of GLP-1 secretion. Furthermore, because GLP-1 stimulates S-28 and S-14 release (1), a negative feedback may be present whereby GLP-1 regulates its own secretion by activating S-28 release. In the present study, a rat intestinal cell culture system that secretes proglucagon-derived peptides (16) was used to examine the roles of S-28, S-14, and SSTRs in the regulation of GLP-1 secretion.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell isolation and culture. Fetal rat intestinal cells were placed into monolayer cultures as described in detail previously (2). In brief, small and large intestines from 19- to 21-day-old fetal Wistar rats were dissected free of gastric and pancreatic tissues, and the cells were dispersed by incubation with collagenase (blend type H; Sigma Chemical, St. Louis, MO), hyaluronidase (type II), and deoxyribonuclease-1 (Sigma Chemical) and placed into monolayer culture for 24 h at a density of 0.625 fetal rat intestines/60-mm dish. Cells were then washed with Hanks' balanced salt solution and incubated with test agents for 2 h in DMEM containing 0.5% (vol/vol) FBS, 1 g/l glucose, 20 µU/ml insulin, 50 IU/ml penicillin, 50 µg/ml streptomycin, 10 µM diprotinin A, 10 µM amastatin, and 1 µM phosphoramidon. Groups of two dishes were used for all experiments except for the gel permeation chromatography studies in which groups of 10 dishes were used. The secretion experiments were repeated four to six times, and separate gel chromatography studies were performed in triplicate. After the incubation period, medium samples were centrifuged to remove any floating cells and made to 0.1% (vol/vol) trifluoroacetic acid. Cells were homogenized in 1 N HCl containing 5% (vol/vol) HCOOH, 1% (vol/vol) trifluoroacetic acid, and 1% (wt/vol) NaCl. The peptides contained in the media and cell samples were then collected separately by passage through a cartridge of C₁₈ silica (Sep-Pak; Waters Associates, Milford, MA), which affords a >95% recovery of exogenously added peptides, as reported previously (2). Aliquots of each extract were dried in vacuo, and the samples were stored at 20°C for subsequent RIA determinations. Animal protocols were approved by the University of Toronto animal care committee according to Canadian Counsel on Animal Care standards.

Somatostatin analogs and test agents. The analogs NC8-12, BIM-23268, and BIM-23058 were gifts from Dr. J. E. Taylor (Biomeasure, Milford, MA). The analog L-362,855 and two structurally related compounds L-372,587 and L-372,588 were gifts from Dr. R. M. Freidinger (Merck Research Laboratories, West Point, PA). The chemical and pharmacological characteristics of the analogs are shown in Table 1. The analogs were dissolved in distilled water at 1 mM, lyophilized, and stored at 20°C until they were used. Gastrin-releasing peptide (GRP), GLP-1-(7-36)-NH₂, and S-28 were obtained from Bachem (Torrance, CA), and S-14 was obtained from Peninsula Laboratories (Belmont, CA). Phorbol 12-myristate 13-acetate (PMA), forskolin, IBMX, diprotinin A, amastatin, and phosphoramidon were obtained from Sigma Chemical.

RIA and chromatography. Immunoreactive GLP-1 was measured by RIA using antiserum RA 7168 (Peninsula), as described in detail previously (10). The reactivity of the antiserum was 100% for GLP-1-(7-36)-amide, 42% for GLP-(1-36)-amide, 0.4% for GLP-1-(7-37), and <0.2% for GLP-1-(1-37), GLP-2, and other members of the glucagon-secretin group of peptides. Gel permeation chromatography of sample aliquots has previously indicated that the RIA detects a single peak corresponding to the position of synthetic GLP-1-(7-36)-amide (10). The detection limit of the assay is 0.4 fmol/tube or 2.0 fmol/ml, and the sensitivity (IC₅₀) is 10.0 fmol/tube or 50.0 fmol/ml. Somatostatin-like immunoreactivity (SLI) was determined by RIA, as described previously (12), using an antiserum that detects both S-14 and S-28 with equal affinity. The detection limit of the assay is 0.3 fmol/tube or 1.5 fmol/ml, and the sensitivity (IC₅₀) is 9.5 fmol/tube or 47.5 fmol/ml. Gel permeation chromatography of dried sample aliquots (100 fmol/sample) was performed using a 9 × 1,000-mm Sephadex G-50 superfine column, as described previously (12). Columns were calibrated with dextran blue (void volume), cytochrome c (mol wt 12,384), synthetic S-28, synthetic S-14, and Na¹²⁵I (total volume). Elution was carried out at 6 ml/h and 4°C with 125 mM NH₄HCO₃, pH 9.0, containing 100 mM NaCl and 0.1% (wt/vol) BSA. Synthetic S-28 and S-14 elute at 53 ml [coefficient of distribution (K_av) = 0.68)] and 67 ml (K_av = 1.02), respectively, under these conditions, and recovery exceeds 95%. The recovery of experimental SLI added to the column was 94 ± 2%. In the present study in control medium, S-28 and S-14 levels per 10 dishes were 26 ± 5 and 21 ± 4 fmol, respectively; in control cells, S-28 and S-14 levels per 10 dishes were 549 ± 112 and 508 ± 103 fmol, respectively (n = 6).

Data and statistical analysis. Secretion of GLP-1 was determined as a percentage of the total cell content (TCC) of GLP-1 (100× medium GLP-1/medium plus cellular peptide) at the end of the incubation period. Synthesis of GLP-1, determined as a function of the total GLP-1 content of media and cells, did not change under all test conditions used in the present study, consistent with previous findings (16). GLP-1 suppression by each analog was calculated as a percentage of the control GLP-1 secretion (GRP, PMA, or forskolin alone) above basal, in the same experiment. In the gel chromatography studies, the release of S-28 and S-14 was determined as the sum of SLI in each fraction under the respective peaks. Statistical comparisons of means within a group were evaluated by Student's paired t-test, and differences between groups were tested by ANOVA followed by a multiple comparisons test using Sigma-Stat (Jandel Scientific, San Rafael, CA). Results are expressed as means ± SE; P  0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effects of S-14, S-28, and analogs on GLP-1 secretion. As shown in Fig. 1, treatment of intestinal cell cultures for 2 h with GRP (100 nM) stimulated GLP-1 to 225 ± 4% of paired basal control values (P < 0.001). S-28 and S-14 caused concentration-dependent inhibition of GRP-stimulated GLP-1 secretion; however, S-28 was more potent compared with S-14 (P < 0.001). The half-maximal effective concentrations (EC₅₀) for S-28 and S-14 were 0.01 and 5.8 nM, respectively, and maximal inhibition to basal values occurred at 1 nM S-28 and 1 µM S-14.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Secretion of glucagon-like peptide-1 (GLP-1) in the basal state and in response to treatment with 100 nM gastrin-releasing peptide (GRP) alone and with different concentrations of somatostatin-14 (S-14) or somatostatin-28 (S-28). Cells were incubated with test agents for 2 h, after which peptides were extracted separately from cells and cell media. Secretion is expressed as a percentage of total cell content. Results are means ± SE of 6 experiments. * P < 0.05, ** P < 0.001 compared with GRP alone.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   GLP-1 response to treatment with 100 nM GRP alone and with different concentrations of S-14, S-28, and the analogs BIM-23268, L-372,588, L-362,855, and L-372,587. GLP-1 inhibition to each analog is expressed as a percentage of the GLP-1 response to GRP alone above basal and is the mean ± SE of 4-6 experiments.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   GLP-1 response to treatment with 100 nM GRP alone and with different concentrations of S-14, S-28, and the analogs NC8-12 and BIM-23058. GLP-1 inhibition to each analog is expressed as a percentage of the GLP-1 response to GRP alone above basal and is the mean ± SE of 4-6 experiments.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Secretion of GLP-1 in the basal state and in response to treatment with 1 µM phorbol 12-myristate 13-acetate (PMA) alone and PMA plus S-14, S-28, BIM-23268, L-372,588, and NC8-12. Secretion is expressed as a percentage of total cell content. Results are means ± SE of 4-6 experiments. * P < 0.05, ** P < 0.001 compared with PMA alone.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Secretion of GLP-1 in the basal state and in response to treatment with 10 µM each of forskolin (FSK)/IBMX alone and to FSK/IBMX plus S-14, S-28, BIM-23268, L-372,588, and NC8-12. Secretion is expressed as a percentage of total cell content. Results are means ± SE of 4-6 experiments. * P < 0.05, ** P < 0.01 compared with FSK/IBMX alone.

Effect of analogs on GLP-1-stimulated S-14 and S-28 secretion. Previous studies with our model have shown that treatment with GLP-1-(7-36)-NH₂ dose-dependently stimulates elevations of S-14 and S-28 secretion, as the culture system is comprised of a heterogeneous cell population that includes D cells producing somatostatin (2). To determine whether GLP-1 and S-28 secretion are regulated via a feedback loop through SSTR5, gel permeation chromatography was used to characterize S-28 and S-14 responses after GLP-1-(7-36)-NH₂ treatment with and without the addition of SSTR analogs. Control media and cells contained two predominant peaks of somatostatin that coeluted with S-28 (K_av = 0.68) and S-14 (K_av = 1.02) (Fig. 6A). Treatment with GLP-1-(7-36)-NH₂ (1 µM) increased secretion of S-28 from 26 ± 5 to 73 ± 7 fmol/10 dishes or threefold above paired controls (P < 0.001) and secretion of S-14 from 21 ± 4 to 109 ± 10 fmol/10 dishes or fivefold above paired controls (P < 0.001) (Fig. 6B), similar to responses reported previously (1). The proportions of S-28 and S-14 in the cells were not altered from controls. The SSTR5-preferring analog BIM-23268 suppressed to baseline the S-28 response after GLP-1-(7-36)-NH₂, without altering S-14 (Fig. 6C). In contrast, the S-28 and S-14 responses to GLP-1-(7-36)-NH₂ after addition of the SSTR2 analog NC8-12 (S-28, 68 ± 9 fmol/10 dishes; S-14, 114 ± 11 fmol/10 dishes; n = 3) were not different from values observed after GLP-1-(7-36)-NH₂ treatment alone.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Characterization by gel chromatography of SLI peptides contained in cell media from intestinal cultures treated for 2 h under control conditions (A; n = 3), with 1 µM GLP-1-(7-36)-NH2 (B; n = 3), or with 1 µM GLP-1-(7-36)-NH2 plus BIM-23268 (C; n = 3). The elution positions of dextran blue [void volume (Vo)], cytochrome c (CC), synthetic S-28, and synthetic S-14 are indicated.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study shows that regulation of GLP-1 secretion by somatostatin occurs via S-28 and primarily through activation of SSTR5. Although both bioactive somatostatin molecular forms caused dose-dependent inhibition of GLP-1 secretion from rat intestinal cell cultures when stimulated by the neurotransmitter GRP, S-28 was markedly more effective than S-14 with an inhibitory action at an IC₅₀ that is within the range for receptor binding (30). Our study further shows that GLP-1-stimulated S-28 secretion is autoregulated by activation of SSTR5. Coupled with the structural localization in the ileum of D cells producing S-28 in proximity to L cells secreting GLP-1, these findings suggest the presence of an interrelationship between S-28 and GLP-1 mediated through SSTR5.

Recent investigations with the isolated perfused porcine ileum similarly showed that S-28 is a potent inhibitor of GLP-1 secretion, whereas S-14 is ineffective (14). The concordance of these findings in an integrated model (14) with our results in a cell culture system therefore lends further support to the notion of a relationship between S-28 and GLP-1. However, five somatostatin receptor subtypes mediate the biological actions of S-28 and S-14 (26). Although all receptor subtypes are expressed in the gastrointestinal tract (17), SSTR5 is the only subtype that preferentially binds S-28 (22, 23, 27). Our results obtained with agonists that provide relatively selective affinities for different somatostatin receptor subtypes show that activation of SSTR5 modulates GLP-1 suppression by S-28. BIM-23268, an analog with high selectivity for SSTR5 (6, 38), caused dose-dependent inhibition of GLP-1 secretion similar to S-28 and more potently than S-14. The present data also support and extend previous investigations (4, 25, 40), indicating the importance of single hydroxyl groups in ligand binding to SSTR5. L-362,855, with two phenylalanine residues in its structure, is a peptide with relatively high affinity for rat and human SSTR5 (6, 23, 30), yet did not influence GLP-1 secretion. However, the structurally related analog L-372,588 with the conversion of Phe to Tyr at position 7 caused inhibition equipotent to S-28, whereas L-372,587 with the conversion of Phe to Tyr at position 2 was ineffective. Similar differences in the agonist effects of these compounds were shown on inhibition of L-type Ca²⁺ channel current conductance in AtT-20 cells and on inhibition of forskolin-stimulated cAMP accumulation in Chinese hamster ovary (CHO-K₁) cells expressing rat SSTR5 (40). Thus the presence of a hydroxyl group at position 7 markedly enhances the agonist properties of this group of peptides at SSTR5. Although the SSTR3 analog showed only a weak effect, NC8-12, an analog that binds to the two isoforms of the SSTR2 receptor with high affinity (35, 43), also caused inhibition of GLP-1 secretion marginally more potent than S-14 but less effective than S-28. These results suggest that the two functional SSTR receptor subtypes SSTR5 and SSTR2 mediate regulation of GLP-1 secretion by S-28, although activation of SSTR-5 predominates.

GRP potently stimulated GLP-1 release, consistent with results from in vivo studies (8). Signal transduction pathways modulating the biological effects of GRP include influx of extracellular Ca²⁺ through L-type voltage-gated channels, elevations of intracellular Ca²⁺, and activation of protein kinase C-dependent pathways (19). Although nitrendipine does not influence GLP-1 release in our model, PMA, an activator of diacylglycerol-sensitive protein kinase C isozymes, stimulated GLP-1 secretion to levels comparable to those observed after GRP. The patterns of suppression after PMA-stimulated GLP-1 secretion by S-28 and the SSTR5 analogs were consistent with the results obtained after GRP. These results suggest that SSTR5 and SSTR2 couple to protein kinase C, similar to findings demonstrated for S-28 inhibition of smooth muscle cells (5). However, our results differ from observations obtained with SSTR5-transfected cells where activation of this receptor subtype stimulated phosphoinositide metabolism (44). These divergent findings may reflect differences between normal and transfected cells where a higher number of receptors cause more variable coupling to this effector pathway.

Activation of protein kinase A-dependent pathways is also a potent stimulus for GLP-1 secretion (16), whereas somatostatin inhibits adenylyl cyclase and cAMP formation (26). The more effective suppression of forskolin-stimulated GLP-1 secretion by S-28 compared with S-14 accords with the preferential role for S-28 in the regulation of GLP-1 secretion. The equipotent inhibition by BIM-23268 and NC8-12 of GLP-1 stimulated by forskolin is consistent with suppression of adenylyl cyclase described in SSTR5- and SSTR2-transfected cells (26) and implicates involvement of both receptor subtypes in the regulation of protein kinase A-dependent GLP-1 secretion. In contrast to its potent inhibition of PMA-stimulated GLP-1 secretion, L-372,588 did not alter forskolin-stimulated GLP-1 secretion, suggesting suppression of protein kinase C-dependent pathways may be a relatively selective action of this SSTR5 analog.

That more than one SSTR subtype is expressed in the same cell type and individually contributes to the functions of somatostatin is well recognized (38). However, recent studies indicate that in certain settings there may be the requirement for the combined action of two SSTRs, as exemplified by the activation of both SSTR5 and SSTR2 for inhibition of platelet-derived growth factor-induced proliferation via Ras-dependent pathways (3). That two SSTRs may undergo heterodimerization to alter various functional properties of the receptors, including agonist regulation, may account for this finding (33). Although we found SSTR5 and SSTR2 coupled individually to protein kinase C and protein kinase A, additional L-cell functions may be modulated by the combined actions of these two receptor subtypes and requires further study.

Certain lines of evidence point to the presence of a regulatory feedback loop between L cells and D cells producing S-28 in the ileum. Studies undertaken with the isolated perfused porcine ileum indicate that nearly all the somatostatin immunoreactivity secreted in the basal state is comprised of S-28, and somatostatin immunoneutralization causes a marked rise of GLP-1 secretion (14). Our results compliment these findings by showing that GLP-1 potently stimulated somatostatin release which included elevations of S-28 and S-14, and coadministration of the SSTR5 agonist BIM-23268 caused preferential inhibition of S-28. Thus SSTR5 modulates not only inhibition of GLP-1 secretion, but also autoregulates S-28 secretion when stimulated by GLP-1. Although the culture model employed in this study is comprised of a heterogeneous cell population that also includes cells secreting peptide YY and cholecystokinin, neither of these peptides modulates somatostatin or GLP-1 release when administered to the cultures (2, 16). Our findings, coupled with in vivo results (14), therefore suggest the presence of an interrelationship between ileal L cells and D cells whereby, via paracrine or endocrine mechanisms, GLP-1 stimulates S-28 secretion, and both S-28 and GLP-1 release are restrained through activation of SSTR5.

GLP-1 is a potent stimulus of glucose-dependent insulin secretion (7, 18), and inhibition of insulin secretion by somatostatin is also mediated by S-28 through SSTR5. Studies of glucose-stimulated insulin secretion in vitro with pancreatic islets and in vivo demonstrated that SSTR5-selective analogs decreased insulin secretion, whereas SSTR2 analogs were ineffective (9, 39). These observations together with our present results on GLP-1 secretion lend further support to the idea that S-28 via SSTR5 plays an important role as a decretin in the counterregulation of the enteroinsular axis.

In conclusion, our studies indicate that somatostatin inhibition of GLP-1 secretion is mediated by S-28 mainly through activation of SSTR5, with a lesser effect by SSTR2. Both receptor subtypes modulate GLP-1 responses to activation of protein kinase C- and protein kinase A-dependent pathways. SSTR5 also autoregulates S-28 secretion when stimulated by GLP-1-(7-36)-NH₂. Single hydroxyl group substitutions modify the agonist properties of SSTR5 analogs. The findings suggest the presence of an interrelationship between L and D cells in the ileum whereby regulation of GLP-1 and S-28 secretion is modulated through SSTR-5.