INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

GLUCAGON-LIKE peptide-1 (GLP-1) is an incretin hormone secreted from L cells of the intestine in response to nutrient ingestion (6). GLP-1 helps to regulate plasma glucose levels by enhancing insulin secretion from the pancreatic -cells (14) as well as diminishing blood levels of both glucagon and somatostatin (11). The insulinotropic effects of GLP-1 are mediated via the pancreatic GLP-1 receptor, which is linked to activation of adenylyl cyclase (26). Recent evidence suggests that GLP-1 may also have peripheral effects to enhance glucose utilization in insulin-sensitive tissues (i.e., fat, muscle, and liver) (3, 5). Recent studies have shown that GLP-1 stimulates glycogen synthesis and increases glycolysis and glucose oxidation in isolated rat soleus muscle and liver (24, 25). These observations were strengthened by evidence for specific GLP-1 binding sites on membranes of rat myocytes (4). However, it has not been clear whether the response to GLP-1 in insulin target tissues occurs via the pancreatic GLP-1 receptor.

In this study, we used L6 myotubes as a model of skeletal muscle (2, 12, 19) to investigate whether GLP-1 has an insulinomimetic effect in muscle and whether the effect is mediated by the pancreatic GLP-1 receptor isoform. We demonstrate that the insulinomimetic effects of GLP-1 on parental L6 myotubes are mediated via a receptor that is functionally different from the pancreatic GLP-1 receptor.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Plasmid constructs. Full-length rat pancreatic GLP-1 receptor clone (Ref. 22; a gift of Dr. B. Thorens, University of Lausanne, Switzerland) was subcloned in a plasmid in which GLP-1 receptor and neomycin phosphotransferase (G418 resistance) genes were each driven by individual mouse RNA polymerase II promoters (pPol2GLPR).

Cell culture and transfection. Rat L6 myoblasts were grown as described previously (2) in a humidified 95% air-5% CO₂ incubator at 37°C, in alpha minimum essential medium (AMEM; Paragon Biotech, Baltimore, MD) containing 10% fetal bovine serum (FBS; GIBCO BRL, Gaithersburg, MD) and supplemented with 50 U/ml penicillin, 50 µg/ml streptomycin, and 2 mM glutamine. The medium was changed every 2 days. After reaching confluence, the cells were induced to differentiate by exposure to 2% FBS in AMEM for 2 days.

Intracellular cAMP levels. Cells grown in 12-well plates were washed 3 times and incubated with 1 ml of Krebs-Ringer phosphate buffer (KRP) containing 128 mM NaCl, 5 mM KCl, 1.3 mM MgCl₂, 1.2 mM CaCl₂, 25 mM HEPES (pH 7.4), 1 mM NaPO₄ (pH 7.4), 2.5 mM glucose, and 0.1% BSA for 4 h at 37°C. Cells were then exposed to KRP containing various concentrations of GLP-1-(736) amide (Bachem, Torrance, CA) for 30 min at 37°C, with or without 1 mM IBMX. After treatment, cells were washed three times with ice-cold Dulbecco's phosphate buffer saline (DPBS) and lysed with ice-cold 0.6 M perchloric acid for 5 min. Aliquots of cell lysate were transferred to microfuge tubes, and the pH was adjusted to 7.0 with 5 M K₂CO₃. After 5 min of centrifugation (2,000 g), the supernatant was vacuum dried and redissolved in 200-500 µl of RIA buffer (500 mM Tris, pH 7.5, and 4 mM EDTA, pH 8.0). After addition of 0.15 mM sodium bicarbonate (20-50 µl) and 0.15 mM zinc sulfate (20-50 µl), samples were incubated for 15 min on ice and then centrifuged for 5 min at 2,000 g. Aliquots of the supernatant were assayed using the [³H]cAMP assay kit (Amersham). Data are presented in picomoles or nanomoles of cAMP per milligram of protein. Each well contained comparable amounts of cellular protein (data not shown).

Glycogen synthesis. A modification of the method of Myers (17) was used. In brief, L6 myotubes in 12-well dishes were washed twice with KRP and incubated in KRP containing 2.5 mM glucose for 3 h at 37°C. The buffer was then replaced with fresh KRP containing D-[U-¹⁴C]glucose (0.29 mCi/mmol; Amersham), with or without GLP-1 or other peptide analogs and/or porcine insulin (Calbiochem, San Diego, CA), for 30 min at 37°C. The cells were then washed three times with ice-cold DPBS and lysed in 400 µl of 20% KOH. The lysate was transferred to microfuge tubes containing 100 µl of carrier glycogen (type VII from muscle, Sigma, St. Louis, MO) at a final concentration of 1 mg/ml. The lysate was boiled and centrifuged, and the glycogen was precipitated with 2.5 volumes of 100% ethanol with overnight incubation at 20°C. After one repetition of the precipitation step, the glycogen pellet was dissolved in 200 µl of water and counted in 10 ml of Ecoscint A (National Diagnostics, Atlanta, GA).

Glucose oxidation and production of lactate and other glycolytic intermediates. L6 myotubes grown in Nunc 25 cm² flasks were washed three times with KRP and incubated in KRP for 1 h at 37°C. The buffer was then replaced with fresh KRP containing D-[U-¹⁴C]glucose (1.8 mCi/ml final concentration) with and without GLP-1 and/or insulin. Each flask was capped with fitted plastic caps containing GF/C filters (Whatman, Maidstone, UK) to trap CO₂ (16). After 1 h of incubation at 37°C, 200 µl of 1 M hyamine hydroxide was injected onto the filter and 200 µl of concentrated perchloric acid was added to the cells. After 90 min of shaking, the filter was added to 10 ml of Ecoscint A containing 10 µl of glacial acetic acid (to reduce chemiluminescence) and counted. For measurements of lactate and other glycolytic intermediates, cells were scraped from the flask and transferred to 15-ml tubes. After the lysate had been centrifuged (2,500 g), the perchloric acid-containing supernatant was transferred and neutralized (10 µl 0.5% phenol red indicator and ~300 µl 10 M KOH) and then incubated overnight at 20°C. The lysate was clarified by centrifugation, and the supernatant was loaded onto an anion-exchange column (AG-X8, Bio-Rad, Richmond, CA) and washed three times with distilled water. The resin was counted in 10 ml of Ecoscint A with 30 µl of glacial acetic acid to reduce color quench. The anion exchange column binds all negatively charged metabolites, including lactate, pyruvate, and other glycolytic intermediates. We consider that the majority of the glycolytic products will be lactate. However, copurification of other glycolytic intermediates was not excluded.

GLP-1 binding. L6 myotubes grown in 12-well dishes were washed and incubated with serum-free AMEM for 4 h at 37°C. The cells were then washed twice with binding buffer (10 mM Tris, pH 7.4, 120 mM NaCl, 1.2 mM MgSO₄, 5 mM KCl, and 15 mM sodium acetate) and incubated at 4°C overnight with 0.5 ml of binding buffer containing 2% BSA, 500 U/ml aprotinin, 25,000 counts/min (cpm) ¹²⁵I-labeled GLP-1 (2,000 Ci/mmol; Amersham), and a range of concentrations of unlabeled GLP-1. We only used freshly prepared ¹²⁵I-GLP-1, within 3 wk of the reference date. The buffer was then removed, and cells were washed three times with ice-cold DPBS and lysed with 0.5 ml of 0.5 N NaOH-0.1% SDS. The radioactivity in the lysate was counted in a gamma counter (10/600 Plus, ICN, Costa Mesa, CA). Specific binding was determined by subtraction of the radioactivity associated with cells incubated with a large excess of unlabeled GLP-1 (0.5 µM).

Ca²⁺ measurements. Cells attached to coverslips were loaded with fura 2 by incubation for 1 h at 37°C with 6 µM fura 2-AM. The coverslips were mounted in a microscope chamber and perfused at 35-37°C with KRP containing various additives. The temperatures of the chambers and microscope objective were kept constant by jacketed water circulation. Cells were studied on a Zeiss Axiovert microscope with a water immersion ×50 Leitz objective, numerical aperture 1.0. Excitation light was provided by a 75-W xenon lamp attenuated to 20% power by neutral density filters. Under these conditions, photobleaching of fura 2 was not detectable. Fura 2 fluorescent emission was measured at 420-570 nm in response to alternating excitation wavelengths of 350 ± 10 nm and 380 ± 10 nm by a computer-controlled filter wheel. Pairs of images were collected every 6 s. Fluorescence was detected by a Hamamatsu intensifier charge-coupled device camera operating at constant intensifier and camera gain. Data were collected as four-frame averages (128 ms/image), which were eight-bit digitized and analyzed by a Perceptics image processor (Knoxville, TN). In each experiment, up to 20 cells in the camera field could be selected for real-time analysis. Data from each of the selected cells in the field were collected, stored separately, and treated as independent observations.

Statistical analysis. Data are means ± SE. We used the Student's two-tailed t-test, and differences were considered significant when P < 0.05.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

GLP-1 regulates glycogen synthesis in parental L6 myotubes. To investigate the effects of GLP-1 on glycogen synthesis in muscle, L6 cells were differentiated to myotubes and then treated with or without insulin in the presence or absence of GLP-1 for 30 min. Insulin alone (1.0 or 100 nM) significantly increased glycogen levels (1.41 ± 0.08-fold and 2.40 ± 0.28-fold, respectively; n = 13, P < 0.001). As shown in Fig. 1, in the absence of insulin or at 0.1 nM insulin, GLP-1 stimulation of glycogen synthesis was slight (1.2-fold, P < 0.05 and P = 0.06, respectively). GLP-1 (10 nM) enhanced insulin-mediated glycogen synthesis at 1 or 100 nM insulin (P < 0.01 and P < 0.05, respectively). The results suggest that GLP-1 has insulinomimetic effects in L6 myotubes.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Effects of glucagon-like peptide-1 (GLP-1) on basal and insulin-mediated glycogen synthesis in L6 cells. Cells were incubated in Krebs-Ringer phosphate buffer (KRP) containing D-[14C]glucose in presence (solid bars) or absence (open bars) of GLP-1 (10 nM) and with a range of concentrations of insulin (0-100 nM) for 30 min, and [14C]glycogen synthesis was measured as described in EXPERIMENTAL PROCEDURES. Data are normalized as percent of response to 100 nM insulin in absence of GLP-1 (0.7 ± 0.1 nmol glycogen/mg protein) and are means ± SE from triplicate measurements in 13 separate experiments. Significant differences between presence and absence of GLP-1: * P < 0.05, ** P < 0.01. Basal glycogen synthesis was significantly different from glycogen synthesis in presence of 1 or 100 nM insulin (P < 0.001). Glycogen synthesis in presence of 0.1 nM insulin was not significantly different from basal.

¹²⁵I-GLP-1 binding in L6 myotubes. As shown in Fig. 2, ¹²⁵I-GLP-1 binding was saturable, and the data were well fitted with a single binding site having a dissociation constant (K_d) of 1.84 ± 0.35 nM (n = 7) and a maximum receptor number of 5,444 ± 1,472 receptors/cell (Table 1). Specific binding, as determined by subtraction of the nonspecific binding (662 ± 94 cpm) from the total binding (812 ± 103 cpm, n = 11), was 150 ± 22 cpm. The specific binding was 0.58 ± 0.08% (n = 11; the amount of specific binding was significantly greater than zero, P < 0.001) of total radioactivity added and 20 ± 2.8% of total binding. These values are similar to the values previously obtained from 3T3-L1 adipocytes (15).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Saturation binding isotherm of 125I-GLP-1 in L6 cells. 125I-GLP-1 binding was measured in intact L6 cells in presence of various concentrations of unlabeled GLP-1. Specific binding was calculated as total binding minus nonspecific binding (latter measured in presence of 500 nM unlabeled GLP-1). The results were converted to fmol bound after calculation of specific activity of labeled GLP-1, and results in each experiment were normalized to amount bound at 10 nM GLP-1 (10.4 ± 3.0 fmol, mean ± SE, n = 9). Curve is best fit for single binding site with a dissociation constant (Kd) of 1.84 ± 0.35 nM. Data are means ± SE from triplicate measurements in 9 separate experiments.

Signal transduction via GLP-1 receptor in L6 myotubes. To investigate the signal transduction mechanisms activated by GLP-1 in L6 myotubes, we first measured intracellular cAMP levels. GLP-1 (10 nM) alone had no significant effect on cAMP levels (Fig. 3A). However, in the presence of 1 mM IBMX (a phosphodiesterase inhibitor), GLP-1 (10 nM) lowered cAMP levels to 83 ± 4% of IBMX control (n = 3, P < 0.05). This small but significant decrease in cAMP levels by GLP-1 is opposite to the known effect of the activated form of the pancreatic GLP-1 receptor (26).

View larger version (28K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effects of GLP-1 on intracellular cAMP (A) and intracellular free Ca2+ levels (B) in L6 cells. A: cells were incubated in KRP containing 100 nM insulin in presence or absence of 10 nM GLP-1, with or without IBMX (1 mM), for 30 min. Data are percentage of control (1 mM IBMX in absence of GLP-1, 6.7 ± 1.4 pmol cAMP/mg protein) and are means ± SE from 7 separate experiments performed in duplicate. * P < 0.05. B: intracellular free Ca2+ was measured using a fluorescence-imaging microscope, and cells attached to coverslips were loaded with fura 2 dye. Ratio of fluorescence at 350-nm excitation to that at 380-nm excitation is shown on y-axis vs. time. Boxed areas indicate periods when cells were exposed to either 100 nM GLP-1 or 10 µM carbachol (carb). A representative experiment (averaged responses of 20 cells) is shown. Similar results were obtained in 4 monolayers examined in 2 separate experiments.

GLP-1 regulates glycolysis and glucose oxidation in parental L6 myotubes. To examine the potential role of GLP-1 in the regulation of cAMP-independent processes, we examined glycolysis and glucose oxidation in L6 cells by measuring the production of lactate and CO₂, respectively. These processes have been shown to be cAMP-independent in muscle (9, 18). As shown in Fig. 4, 100 nM insulin for 1 h induced significant increases in both CO₂ and lactate production (2.0-fold over basal). Similarly, GLP-1 (10 nM) significantly increased basal glycolysis and glucose oxidation by ~1.5-fold. In the presence of 1 nM insulin, glycolysis and glucose oxidation were increased by 154 ± 13.5 and 155.5 ± 15.7% (n = 3), respectively. In the presence of 1 nM insulin, GLP-1 (10 nM) had a small additive effect on CO₂ production (117 ± 3.5% vs. insulin alone, P < 0.05, n = 3), whereas the same effect on lactate production did not reach significant levels (112.8 ± 12.3% vs. insulin alone, P > 0.05, n = 6). These results suggest that GLP-1 has insulinomimetic effects in L6 myotubes that may also involve cAMP-independent signaling pathways.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effects of GLP-1 on glucose oxidation in L6 cells. Cells were incubated in KRP containing D-[14C]glucose in presence of GLP-1 (solid bars) or in absence of GLP-1 (10 nM) with (hatched bars) or without (open bars) insulin (100 nM) for 1 h at 37°C. [14C]CO2 (A) or [14C]lactate (B) was measured as described in EXPERIMENTAL PROCEDURES. Data are presented in counts/min (cpm) per mg protein and are means ± SE from duplicate measurements of 3 separate experiments. * P < 0.05; ** P < 0.01.

Transfection of L6 myotubes with pancreatic GLP-1 receptor. The physiological responses of L6 cells to GLP-1, (i.e., decreased cAMP and increased glycogen synthesis, lactate, and CO₂ production) could not be explained by our current understanding of early postreceptor pathways of pancreatic GLP-1 receptor signaling. The results suggest that in L6 myotubes a novel receptor is expressed that responds to GLP-1 (with or without expression of the pancreatic type GLP-1 receptor) or that the pancreatic GLP-1 receptor isoform uses different signal transduction mechanisms in L6 cells and in pancreas. The transfection of the cloned pancreatic GLP-1 receptor into L6 cells allowed the direct comparison of receptor signaling between the overexpressed pancreatic GLP-1 receptor and the endogenous GLP-1 receptor. As shown in Fig. 5, both GLP-1 receptor-transfected L6 cell lines and the parental L6 line differentiated to myotubes. This suggests that all transfectants (L6/pancGLPR) retained the intact differentiation properties of parental cells. These results are important for further analysis of pancreatic GLP-1 receptor function in transfected cells, since it allows comparison of GLP-1 function within a cellular milieu similar to parental L6 cells.

View larger version (202K): [in this window] [in a new window]   Fig. 5.   Visualization of various differentiated L6 cell lines. Cells were viewed by differential interference contrast imaging in a Zeiss Axiovert microscope with a ×40 objective. Parental (A) and GLP-1 receptor-transfected L6 cells [mixed population (B), clone 5 (C), and clone 2 (D)]. All GLP-1 receptor-transfected cells (B-D) differentiated to myotubes normally (similar to A, parental L6 cells), which suggests that transfected cells conserved physiological features of parental L6 cells. Scale in D applies to A-D.

Binding characteristics of various L6 cell lines. GLP-1 binding was analyzed in four independent L6/pancGLPR cell lines and compared with parental L6 cells (Table 1). Scatchard analysis revealed that the number of GLP-1 receptor binding sites varied among the different transfectant cells, ranging from 3 times (clone 2) to 10 times (mixed population) higher than the number of endogenous GLP-1 receptor sites. The binding affinity for GLP-1 was similar among parental cells and L6 transfectant clones.

Signal transduction in GLP-1 receptor-transfected L6 cells. We measured cAMP levels in transfected L6 myotubes expressing the pancreatic GLP-1 receptor (Fig. 6A). In contrast to parental L6 cells, 1 nM GLP-1 increased cAMP 7- to 100-fold in transfectant populations. These results are unlikely to be due to cell cloning artifacts because similar results were observed in both the mixed population and the cloned transfectants. The concentration that elicited half-maximal response was 10 nM GLP-1 for the mixed population, with a cAMP response of 311.9 ± 62.4 pmol/mg protein (n = 3). The results show that the pancreatic GLP-1 receptor is linked to increases in cAMP levels in transfected L6 myotubes, suggesting that sufficient G_s, the G protein that activates adenylyl cyclase, was available for receptor coupling. We also performed measurements of intracellular cAMP in the presence of GLP-1 and the phosphodiesterase inhibitor IBMX, to eliminate any variation in phosphodiesterase activity in various cell lines. In these conditions, a positive correlation (r = 0.985, P < 0.05) was observed between GLP-1-stimulated levels of cAMP and the number of GLP-1 binding sites (Fig. 6B). Our data suggest that heterologous expression of the pancreatic GLP-1 receptor in L6 myotubes did not saturate the signal transduction mechanism.

View larger version (14K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Fig. 6.   Effects of GLP-1 on cAMP levels in L6 cells transfected with with pancreatic GLP-1 receptor. A: intracellular cAMP levels were measured after 30-min exposure of GLP-1 receptor-transfected L6 cell lines to various GLP-1 concentrations. Data are in pmol/mg protein and are means ± SE of 3 independent experiments performed in duplicate. B: correlation between cAMP levels after 30-min treatment of GLP-1 receptor-transfected L6 cell lines with 0.1 nM and in presence of 1 mM IBMX (n = 4). GLP-1 receptor numbers were derived from Table 1. Coefficient of correlation is r = 0.985 (P < 0.05).

GLP-1 inhibits glycogen synthesis in L6/pancGLPR myotubes. As shown in Fig. 7, insulin increased glycogen synthesis in the mixed population of L6/pancGLPR cells, indicating that transfectants respond normally to insulin. However, in contrast to parental L6 cells, 0.1 nM GLP-1 significantly inhibited basal and insulin-stimulated glycogen synthesis. Similar results were observed with all clones of L6/pancGLPR cells (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Effects of GLP-1 on glycogen synthesis in L6 cells transfected with pancreatic GLP-1 receptor. Glycogen synthesis was measured in a mixed population of GLP-1 receptor-transfected L6 cells treated for 30 min with (solid bars) or without (open bars) GLP-1 (0.1 nM) in presence of a range of concentrations of insulin (0-100 nM). Data are normalized as percent of response to 100 nM insulin in absence of GLP-1 (0.82 ± 0.09 nmol glycogen/mg protein) and are means ± SE from triplicate measurements in 3 separate experiments. * P < 0.05 represent significant differences between presence vs. absence of GLP-1. Basal glycogen synthesis was significantly different from glycogen synthesis in presence of 1 or 100 nM insulin (P < 0.05 and P < 0.01, respectively). Glycogen synthesis in presence of 0.1 nM insulin was not significantly different from basal.

Exendin-4-(939) differentially affects parental L6 and L6/pancGLPR myotubes. An antagonist of pancreatic GLP-1 receptor was used to confirm that effects in transfectants were due to pancreatic GLP-1 receptor. As shown in Fig. 8, exendin-4-(939) (0.1 µM; a generous gift of Dr. John Eng, Veterans Affairs Medical Center, Bronx, NY) reversed the inhibitory effect of GLP-1 (0.03 nM) on glycogen synthesis in L6/pancGLPR cells. However, in parental L6 myotubes (Fig. 9), 10 nM exendin-4-(939) stimulated basal and insulin-mediated glycogen synthesis. This suggests that this antagonist of the pancreatic GLP-1 receptor is an agonist in parental L6 myotubes, since it mimics the effect of GLP-1 in this cell line.

View larger version (39K): [in this window] [in a new window]   Fig. 8.   Effect of exendin-4-(939) on glycogen synthesis in L6 cells transfected with pancreatic GLP-1 receptor. Glycogen synthesis was measured for 30 min in GLP-1 receptor-transfected cells in presence or absence of 100 nM insulin. GLP-1 (0.03 nM) was added in presence or absence of 0.1 µM exendin-4-(939). * P < 0.05, *** P < 0.001. Data are normalized as percent of response to 100 nM insulin alone. Data are means ± SE from triplicate measurements in 3 separate experiments.

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Effect of exendin-4-(939) on glycogen synthesis in parental L6 cells. Glycogen synthesis was measured for 30 min in L6 parental cells in presence (solid bars) or absence (open bars) of 10 nM exendin-4-(939), with or without 100 nM insulin. * P < 0.05, *** P < 0.001 for presence vs. absence of exendin-4-(939). Data are percent of response in presence of 100 nM insulin alone and are means ± SE from triplicate measurements in 3 separate experiments.

Effects of different analogs of GLP-1 in L6 myotubes. To further characterize the receptor that binds GLP-1 in parental L6 myotubes, a number of peptides related to GLP-1 were used in parental L6 myotubes to displace ¹²⁵I-GLP-1 binding at 1 and 500 nM. Table 2 shows that 1 nM exendin-4-(939), exendin-4-(139), GLP-1-(136), and GLP-1 itself displaced ¹²⁵I-GLP-1 binding by 30-50%. GLP-2 and glucose-dependent insulinotropic peptide were without effect, demonstrating the peptide specificity of the results. These data suggest that the agonist effect of exendin-4-(939) and GLP-1 on glycogen synthesis is likely to be mediated via a common receptor, albeit with properties different from those of the pancreatic GLP-1 receptor (compare Figs. 8 and 9). We also performed glycogen synthesis experiments in the presence of 10 nM GLP-1-(136) and exendin-4-(139) (Table 3). Both of these peptides were agonists and enhanced insulin-mediated glycogen synthesis but had no effects on basal glycogen synthesis (data not shown). These data suggest a peptide specificity in L6 cells that is different from what is known for pancreatic GLP-1 receptor. Table 3 also shows that the response of cells to GLP-1 was dose dependent; 1 and 10 nM GLP-1 increased insulin-mediated glycogen synthesis. However, 0.1 nM GLP-1 was without effect (data not shown). Surprisingly, GLP-1 at 100 nM had no significant effect on glycogen synthesis. An explanation for this could be that GLP-1 at this high concentration acts also via the glucagon receptor, shown by RT-PCR to be present in these cells (data not shown). Experiments confirmed that basal and insulin-mediated glycogen synthesis were inhibited in the presence of 10 nM glucagon, by 22.6 ± 5.75 and 19.3 ± 7.6%, respectively (n = 7, P < 0.05), which is in accord with a classical inhibitory action of glucagon on glycogen synthesis.

View this table: [in this window] [in a new window]   Table 2.   Inhibition of specific binding of 125I-GLP-1-(736) amide by various peptides in parental L6 myotubes

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

An effect of GLP-1 to enhance peripheral glucose utilization in muscle would be beneficial to patients with either type I or type II diabetes. However, controversial results exist in the literature as to a direct effect of GLP-1 on muscle (7, 25). In addition, there is no published report exploring a potential synergistic effect of GLP-1 on insulin-mediated glycogen synthesis, glycolysis, or glucose oxidation. Recent evidence from pancreatic GLP-1 receptor knockout mice showed that glucose levels after an oral glucose tolerance test were not modified in the presence or absence of GLP-1 (20) and that whole body glucose utilization was not modified during a hyperinsulinemic euglycemic clamp (21). However, other contraregulatory hormones such as glucagon, somatostatin, and glucocorticoid hormones could have masked a GLP-1 effect in extrapancreatic tissues of the knockout mice. GLP-1 has recently been shown to have effects on hormone secretions from the hypothalamus-pituitary-adrenal axis (1, 13).

To investigate the effect of GLP-1 on a homogenous muscle cell preparation in which conditions can be precisely controlled, we used L6 cultured cells that differentiate to myotubes. Glucose metabolism responds to insulin in L6 myotubes. We show that GLP-1 increased glycolysis and glucose oxidation as well as glycogen synthesis. Our data support previous observations of glycogenic properties of GLP-1 in muscle (25) and reveal that these effects of GLP-1 are additive to those of insulin.

In L6 myotubes, GLP-1 activates signal transduction pathways different from those activated by pancreatic GLP-1 receptor. Activation of the rat pancreatic GLP-1 receptor has been shown to increase intracellular cAMP; when overexpressed in COS cells (26), this receptor can also increase intracellular Ca²⁺. Our results in L6 myotubes are in accord with previously published data (4) showing GLP-1 has no effect on basal adenylyl cyclase activity in muscle. However, in L6 myotubes, we were able to show that an IBMX-stimulated elevation in cAMP was diminished in the presence of GLP-1. Because intracellular Ca²⁺ was not mobilized by GLP-1, the results show clear differences from the pancreatic GLP-1 receptor.

The results also suggest that GLP-1 has some physiological effects on L6 myotubes via cAMP-independent mechanisms. GLP-1 stimulated glycolysis (lactate production) and glucose oxidation (CO₂ production), two metabolic pathways predicted to be cAMP-independent in muscle (9, 18). The alternative signaling pathway(s) may involve inositolphosphoglycans and diacylglycerol levels, as shown for BC₃H-1 myocytes and liver (8, 23). Therefore, different rate-limiting steps may control these various physiological responses.

Our evidence suggests that the effects of GLP-1 in L6 muscle cells are probably mediated by a receptor distinct from pancreatic GLP-1 receptor. The difference in signal transduction linked to GLP-1 action could be due either to the pancreatic GLP-1 receptor isoform expressed in L6 myotubes but coupled to different signal transduction or to a distinct receptor expressed in muscle that responds to GLP-1. To distinguish between these two possibilities, we transfected the pancreatic GLP-1 receptor isoform cDNA into L6 myoblasts. The transfected myoblasts were able to differentiate into myotubes, so comparison of signal transduction and function between transfected and parental L6 cells was made between comparable differentiated cells. In stable transfectants, activation of pancreatic GLP-1 receptor gave the predicted response of increasing cAMP levels and inhibiting basal and insulin-mediated glycogen synthesis. These data suggest that endogenous receptors that respond to GLP-1 in L6 cells are distinct from the pancreatic GLP-1 receptor isoform. This effect was not simply due to overexpression of the GLP-1 receptor, since clones that contain 3 times (clone 2) to 10 times (mixed population) more GLP-1 binding sites than the endogenous receptor gave qualitatively similar results and did not saturate the cellular signal transduction pathways.

Finally, exendin-4-(939), a specific antagonist of pancreatic GLP-1 receptor (10), inhibited GLP-1-mediated glycogen synthesis in L6 cells transfected with pancreatic type GLP-1 receptors but had opposite effects in parental L6 cells. Similarly, GLP-1-(136), an inactive peptide on pancreatic GLP-1 receptor, acted as an agonist and increased glycogen synthesis in parental L6 myotubes. Experiments in which these peptides were used to compete against binding of ¹²⁵I-GLP-1 suggest that the agonistic nature of exendin-4-(939), exendin-4-(139), GLP-1-(136), and GLP-1-(736) is most likely mediated via a common receptor. Similar results were obtained in 3T3-L1 adipocytes (15) and rat muscle cells (4).

The results strongly suggest that parental L6 cells express a distinct receptor that responds to GLP-1, other than the one expressed in pancreas, since the peptide responsiveness of this receptor is different from that of the pancreatic GLP-1 receptor. Because neither Western nor Northern blots detect a protein with homology to pancreatic GLP-1 receptor and attempts to amplify GLP-1 receptor mRNA using RT-PCR and primers derived from various regions of pancreatic GLP-1 receptor have failed (data not shown), other strategies will be necessary to resolve this new receptor. Although the L6 muscle cell line allows the identification of a receptor with novel properties, the presence of this novel receptor in normal muscle cells needs to be investigated. Furthermore, isolation of this new putative receptor will be needed to further characterize its primary ligand.

In summary, our data show that GLP-1 acts in L6 cells via a receptor that has signal transduction and ligand specificity different from those of the pancreatic receptor and that acts in part via cAMP-independent pathways.