Predominant role of active versus facilitative glucose transport for glucagon-like peptide-1 secretion
- 1.8k Downloads
Several glucose-sensing pathways have been implicated in glucose-triggered secretion of glucagon-like peptide-1 (GLP-1) from intestinal L cells. One involves glucose metabolism and closure of ATP-sensitive K+ channels, and another exploits the electrogenic nature of Na+-coupled glucose transporters (SGLTs). This study aimed to elucidate the role of these distinct mechanisms in glucose-stimulated GLP-1 secretion.
Glucose uptake into L cells (either GLUTag cells or cells in primary cultures, using a new transgenic mouse model combining proglucagon promoter-driven Cre recombinase with a ROSA26tdRFP reporter) was monitored with the FLII12Pglu-700μδ6 glucose sensor. Effects of pharmacological and genetic interference with SGLT1 or facilitative glucose transport (GLUT) on intracellular glucose accumulation and metabolism (measured by NAD(P)H autofluorescence), cytosolic Ca2+ (monitored with Fura2) and GLP-1 secretion (assayed by ELISA) were assessed.
L cell glucose uptake was dominated by GLUT-mediated transport, being abolished by phloretin but not phloridzin. NAD(P)H autofluorescence was glucose dependent and enhanced by a glucokinase activator. In GLUTag cells, but not primary L cells, phloretin partially impaired glucose-dependent secretion, and suppressed an amplifying effect of glucose under depolarising high K+ conditions. The key importance of SGLT1 in GLUTag and primary cells was evident from the impairment of secretion by phloridzin or Sglt1 knockdown and failure of glucose to trigger cytosolic Ca2+ elevation in primary L cells from Sglt1 knockout mice.
SGLT1 acts as the luminal glucose sensor in L cells, but intracellular glucose concentrations are largely determined by GLUT activity. Although L cell glucose metabolism depends partially on glucokinase activity, this plays only a minor role in glucose-stimulated GLP-1 secretion.
KeywordsGlucagon-like peptide-1 (GLP-1) Glucokinase KATP channel L cells SGLT1
Bacterial artificial chromosome
Cyan fluorescent protein
Förster resonance energy transfer
Glucose-dependent insulinotropic polypeptide
ATP-sensitive potassium channel
Red fluorescent protein
Sodium-dependent glucose transporter 1
Tandem red fluorescent protein
Yellow fluorescent protein
Glucagon-like peptide-1 (GLP-1) is an incretin hormone secreted from intestinal L cells, located throughout the gut epithelium particularly in the ileum and colon . GLP-1 augments insulin secretion in a glucose-dependent manner, and, together with glucose-dependent insulinotropic polypeptide (GIP), is responsible for up to 70% of the insulin response to food intake in healthy individuals [2, 3]. As GLP-1 additionally inhibits glucagon secretion, slows gastric emptying and enhances satiety, it is an attractive target for pharmaceutical interventions. Recent therapies that increase GLP-1 activity, by the use of degradation-resistant mimetics or inhibition of GLP-1 cleavage by dipeptidyl peptidase IV, improve glycaemia in patients with type 2 diabetes, demonstrating the success of targeting the GLP-1 axis. Current interest in studying pathways underlying GLP-1 release could lead to new therapeutic strategies for increasing endogenous GLP-1 secretion.
GLP-1 release is triggered by ingestion of carbohydrates, fats and protein, and is believed to reflect, at least in part, the direct sensing of luminal nutrients via the apical processes of L cells. The release of GLP-1 and GIP after glucose ingestion accounts for the incretin effect, characterised by enhanced insulin release triggered by oral compared with intravenous glucose [4, 5]. This difference can most easily be explained by a ‘luminal’ glucose sensor, shielded from variations in the plasma glucose concentration. The identity of this ‘sensor’ for detecting luminal sugars has been much debated but not fully elucidated.
Initial studies using the GLUTag cell line suggested that L cells may use the classical glucose-sensing machinery used by the pancreatic beta cell, involving glucose phosphorylation by glucokinase, enhanced glycolytic and mitochondrial metabolism, and closure of ATP-sensitive potassium (KATP) channels . More recent studies confirmed that the KATP channel subunits, Kir6.2 and SUR1, and Glucokinase are expressed at high levels in purified mouse L cells and that the proteins are detectable by immunostaining in human L cells [7, 8, 9]. Electrophysiological and secretion studies have demonstrated that KATP channels are functional in murine L cells and that sulfonylureas can stimulate GLP-1 secretion from primary colonic cultures .
An element of glucose sensing by L cells is, however, clearly independent of metabolism, as GLP-1 secretion is also stimulated by non-metabolisable sugars such as methyl-α-glucopyranoside (αMG) and 3-O-methylglucose in whole animals , intestinal preparations , GLUTag cells and primary intestinal cultures [7, 12]. Early in vivo studies showed that luminal sugar stimulation of GLP-1 secretion is Na+ dependent and that the specificity of the response paralleled the sugar specificity of Na+-coupled glucose transport [11, 13]. This led to the identification of a distinct glucose-sensing pathway in L cells, resulting from the activity of Na+-coupled glucose transporters (SGLTs) . These concomitantly carry Na+ ions for each glucose molecule transported, thereby generating small depolarising currents sufficient to trigger electrical activity and Ca2+ entry and consequent increased GLP-1 secretion.
As the relative importance of SGLTs, intracellular glucose levels and metabolism for glucose-dependent GLP-1 secretion is unclear, we aimed to further investigate the roles of these mechanisms in determining L cell glucose fluxes and GLP-1 release.
Animal procedures were approved by the local ethics committee and conformed with UK Home Office regulations. Sglt1−/− mice  on a C57BL/6 background were crossed with GLU-Venus transgenic mice . Sglt1−/− and Sglt1+/+ littermates received a glucose/galactose-reduced diet (Altromin, Lage, Germany).
Labelling of intestinal L cells with a red fluorescent protein (RFP) was achieved by crossing Rosa26tdRFP reporter mice  with mice expressing Cre recombinase under the control of the proglucagon promoter (GLU-Cre12 mice). GLU-Cre12 mice were created using a construct based on the bacterial artificial chromosome (BAC) RP23-343C17 (Children’s Hospital Oakland Research Institute, Oakland, CA, USA) in which the sequence between the proglucagon start codon in exon 2 and the stop codon in exon 6 was replaced by iCre using Red/ET recombination technology (Genebridges, Heidelberg, Germany) (see electronic supplementary material [ESM Methods/Table 1] for more details).
Intestines from 3- to 6-month-old mice were collected, and the epithelial cells cultured as described previously . The upper (∼top third) small intestine (SI) comprised a 10 cm length distal to the stomach, and the colon was taken distal to the ileocolic junction. Aliquots were plated on to 24-well plates or 35 mm glass-bottomed dishes (MatTek, Ashland, MA, USA) coated with Matrigel (BD Biosciences, Oxford, UK) for 24–48 h and incubated at 37°C in 5% CO2. GLUTag cells were cultured as described previously [16, 17].
Intracellular glucose measurements
The FLII12Pglu-700μδ6 Förster resonance energy transfer (FRET) glucose sensor  was cloned into pShuttle-CMV (Qbiogene, Carlsbad, CA, USA) for generation of adenoviruses . GLUTag cells were transfected with pcDNA3.1 containing the glucose sensor under cytomegalovirus (CMV) promoter control (Addgene, Cambridge, MA, USA), using Lipofectamine 2000 (Invitrogen, Paisley, UK). Cells were then seeded on to Matrigel-coated glass-bottomed dishes and imaged 24–48 h later. Before imaging, GLUTag cells were incubated in saline buffer (see below) for 10 min at room temperature. Two-day-old primary colonic cultures from GLU-Cre-×−tdRFP mice were transduced with adenovirus encoding the FLII12Pglu-700μδ6 glucose sensor and imaged 72 h later. Before each experiment, colonic cultures were incubated in forskolin (10 μmol/l) and 3-isobutyl-1-methylxanthine (100 μmol/l) for 30 min at 37°C. L cells were identified by their RFP fluorescence and characteristic morphology. FRET imaging was performed using an inverted fluorescence microscope (Nikon Eclipse TE2000-S or Olympus IX71) with a × 40 oil immersion objective. The FRET probe was excited every 5 s at 434/10 nm using a 75 W xenon arc lamp and monochromator (Cairn Research, Faversham, UK) controlled by MetaFluor software (Molecular Devices, Wokingham, UK). Emission was recorded with a CCD camera (QuantEM Photometrics, Tucson, AZ, USA or Orca-ER, Hamamatsu Photonics, Welwyn Garden City, UK) behind an Optosplit II image splitter (Cairn Research) equipped with cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) emission filter sets. Fluorescence was recorded from individual cells, background corrected and expressed as the ratio YFP/CFP. For analysis, data were averaged over 15–20 s, and peak responses normalised by dividing by the pretreatment baseline.
Tissues were fixed with 4% (w/v) paraformaldehyde for 24 h, cryoprotected in 30% sucrose and embedded in Optimal Cutting Temperature compound (OCT; CellPath, Newton, UK). Tissue sections (8 μm) were permeabilised with 0.05% Tween, blocked with 3% BSA for 1 h, and incubated with 1:300 diluted glucagon antibody (catalogue no. sc-13091, Santa Cruz Biotechnology, Santa Cruz, CA, USA) or 1:50 diluted Glut2 antibody (catalogue no. sc-7580; Santa Cruz) overnight at room temperature. Tissues were then incubated for 1 h at room temperature with Alexa 488-conjugated goat anti-rabbit antibody or donkey anti-goat Alexa Fluor 633 (1:300 dilution; Invitrogen) and Hoechst stain. Tissue samples stained with secondary antibody alone served as controls. Images were captured using either an inverted fluorescence (Olympus IX71) or confocal (Zeiss LSM510) microscope and processed using Volocity and MetaFluor software.
Assay of glucokinase activity
Glucokinase activity was determined using an adapted pyridine nucleotide-coupled assay . Briefly, ∼3 × 107 GLUTag cells were homogenised in buffer containing (in mmol/l) 25 Hepes, 150 KCl, 2 dithiothreitol and 1 EDTA (pH 7.4), and the resultant supernatant fraction was mixed with assay buffer containing (in mmol/l) 100 Hepes, 6 MgCl2, 5 ATP, 0.1% BSA, 150 KCl, 1 dithiothreitol, 1 β-NAD, 0.045 5-thio-d-glucose 6-phosphate, 1 3-O-methyl-N-acetylglucosamine (Axxora, Exeter, UK), 2.5 IU/ml glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides and glucose (0–60 mmol/l) in a 96-well plate. NADH fluorescence (excitation 320 nm, emission 460 nm) was measured every 1 min for 2 h at 30°C using a Fluoroskan Ascent plate reader (Thermo Scientific, Basingstoke, UK).
NAD(P)H imaging in GLUTag cells
NAD(P)H levels were imaged in GLUTag cells cultured for 24–48 h on Matrigel-coated glass-bottomed dishes. Cells were washed with and incubated in saline buffer for ∼15 min, mounted on an Olympus IX71 microscope with ×40 oil-immersion objective, and imaged using an Orca-ER CCD camera and Metafluor software. Cells were perfused with saline buffer (plus glucose as indicated) at room temperature. Autofluorescence at 360/15 nm excitation, 510/80 nm emission, was measured every 10 s, background corrected, and averaged over 60 s periods. Mean fluorescence under test conditions was normalised by dividing by the mean of baselines measured before addition, and after washout, of the test reagent.
Analysis of primary L cells by flow cytometry
Single-cell digests were fixed with 4% paraformaldehyde in PBS for 30 min at room temperature and blocked with PBS/10% goat serum at 4°C overnight. Cells were permeabilised with 0.1% vol./vol. Triton X-100 in PBS/10% goat serum for 30 min at room temperature, and then incubated with or without primary antibody detecting proglucagon (source as above; 1:200 dilution) in PBS/10% goat serum at room temperature for 3 h. Cells were rinsed three times in PBS/10% goat serum and incubated for 1 h with secondary antibody (Alexa Fluor 488; Invitrogen, Eugene, OR, USA; catalogue no. A-11034 at 1:300 dilution). After three washes with PBS, cells were analysed using a BD LSRFortessa analyser (BD Biosciences, San Jose, CA, USA) equipped with 488 nm and 561 nm lasers for excitation of Alexa Fluor 488 and RFP, respectively. Data were analysed using FlowJo 7.6 software (Tree Star, Ashland OR, USA).
For NAD(P)H measurements from primary L cells, SI tissue from GLU-Venus mice was digested to single cells  and analysed on the same day. Cells were washed and resuspended in saline buffer containing 0.1% BSA and various glucose concentrations (0–30 mmol/l) 10–15 min before analysis. From each condition, 106 events were assessed using an LSRFortessa analyser and FlowJo 7.6 software. Venus-positive L cells were selected on the basis of their characteristic fluorescence when excited at 488 nm, and autofluorescence was assessed by excitation at 355 nm (emission 450 nm). The geometric mean of the 355 nm fluorescence intensity distribution from each cell population was calculated and normalised to that of the zero glucose control measured on the same day.
Small interfering RNA (siRNA) knockdown
GLUTag cells were transfected with 50 nmol/l scrambled or Sglt1 siRNA (Qiagen, Crawley, UK) for 24 h using Lipofectamine 2000. Cells were reseeded into 24-well plates and used 24 h later. Knockdown efficiency was assessed using Taqman assays for Slc5a1 (sglt1) and Actb (β-actin) (Applied Biosystems, Paisley, UK) [7, 21].
Secretion studies on GLUTag and primary intestinal cultures were performed 24–36 h after plating in Matrigel-coated 24-well plates. Cultures were incubated with test reagents in saline buffer containing 0.1% fatty acid-free BSA for 2 h at 37°C. Cell lysates were then collected from primary cultures as described previously . Supernatant fractions and lysates were assayed using either a GLP-1-active ELISA kit (Millipore, Watford, UK) or a total GLP-1 assay (MesoScale Discovery, Gaithersburg, MD, USA). For primary cells, GLP-1 secretion was expressed as a fraction of the total hormone content per well, normalised to basal secretion measured in parallel. For GLUTag cells, supernatant fraction concentrations were normalised to basal levels in parallel control wells.
Experiments were performed on 5- to 8-day-old cultures from Sglt1−/− mice and wild-type littermates, crossed into the GLU-Venus background. Cells were loaded in 7 μmol/l fura2-AM (Invitrogen, UK) and 0.01% pluronic F127, and incubated in saline buffer containing 1 mmol/l glucose and 300 μmol/l eserine, for 30 min. Experiments were performed using the Olympus IX71 imaging system described above. Fura2 was excited at 340 and 380 nm (emission 510/80 nm), and Venus at 475 nm (emission 535/50 nm). Fura2 fluorescence measurements were taken every 2 s, background corrected, and expressed as the 340/380 nm ratio. Mean fluorescence ratios were determined over 20 s, and responses expressed as the maximum ratio achieved during stimulation divided by the mean of the ratios measured before and after washout. Cells were included in the analysis if they responded to the positive controls, 10 mmol/l glutamine and 30 mmol/l KCl.
Saline buffer contained (mmol/l) 4.5 KCl, 138 NaCl, 4.2 NaHCO3, 1.2 NaH2PO4, 2.6 CaCl2, 1.2 MgCl2 and 10 HEPES (pH 7.4, NaOH). In some experiments, KCl was increased with equivalent reduction in NaCl, in others Na+ was replaced by N-methyl-d-glucamine (NMDG). Where possible, solutions were prepared as ×1,000 stock. Glucokinase activator (GKA50)  (AstraZeneca, Macclesfield, UK) was dissolved in DMSO. Chemicals were supplied by Sigma Aldrich (Poole, UK) unless otherwise stated.
Data are presented as means ± SEM. Significance was assessed using Student’s t test (Microsoft Excel) or by one- or two-way ANOVA followed by Bonferroni or Dunnett’s post hoc tests (Graph Pad Prism Software, San Diego, CA, USA). p < 0.05 was considered significant.
Glucose uptake pathways in L cells
Glucose-stimulated changes in L cell metabolism
Role of glucokinase in L cells
Effect of intracellular glucose on GLP-1 secretion
Role of SGLT1 in glucose-triggered GLP-1 secretion
As the impaired Ca2+ response and GLP-1 secretion of L cells in this SGLT1-deficient model may still arise from a defect in glucose uptake into neighbouring cells rather than the enteroendocrine cells themselves, we also knocked down Sglt1 in GLUTag cells using siRNA, which decreased mRNA expression by ∼60% (Fig. 6c). Sglt1 knockdown largely abolished the effects of both glucose (10 mmol/l) and αMG (100 mmol/l) on GLP-1 release (Fig. 6d).
The mechanism underlying glucose sensing by L cells is a topic of recent debate, with sweet taste receptors (Tas1R2/3) [23, 24], SGLT1  and KATP channels  each suggested to play a role. We demonstrate here that genetic or pharmacological interference with SGLT1 abolishes glucose-triggered Ca2+ responses and GLP-1 secretion from L cells in primary culture. This is consistent with the inhibition of GLP-1 secretion by phloridzin from perfused intestinal preparations  and the more recent observation of impaired glucose-triggered GLP-1 secretion in Sglt1 knockout mice . The latter observations are incompatible with the concept of an apically located glucose receptor on the surface of the L cell, as inhibition of glucose absorption would, if anything, tend to increase exposure of such a receptor to luminal glucose. Although they could be explained by a basolaterally expressed receptor which is exposed to elevated glucose concentrations after SGLT1-dependent absorption through enterocytes, the current findings, in which both sides of the cells are exposed to glucose, argue against any major role of an extracellular receptor. This is consistent with our previous observation that artificial sweeteners, at concentrations that saturate Tas1R2/3 receptors, did not trigger GLP-1 release from primary intestinal cultures . A metabolic sensing mechanism downstream of SGLT1-mediated glucose uptake could also be envisioned. Monitoring intracellular glucose levels, however, allowed us to dissociate the stimulatory action of Na+-coupled glucose uptake from possible downstream metabolic effects, as SGLT inhibition had only minor effects on intracellular glucose levels, whereas GLUT inhibitors largely abolished glucose uptake in L cells but did not significantly impair GLP-1 secretion. The finding that Sglt1 knockdown in GLUTag cells impaired both glucose- and αMG-triggered secretion argues for a mechanism intrinsic to the enteroendocrine cells, rather than involving coupling through neighbouring enterocytes. It also demonstrates a dominant role of SGLT1 over SGLT3, which is also produced in GLUTag and primary L cells . We thus conclude that the electrogenic uptake itself via apically localised SGLT1  is the major glucose-sensing mechanism in L cells.
To enable the use in primary tissues of FRET-based sensors employing YFP and CFP, we generated a new BAC transgenic mouse model expressing the gene encoding Cre recombinase under control of the proglucagon promoter, targeting enteroendocrine L cells as well as pancreatic alpha cells unlike the shorter promoter constructs used previously to drive Cre expression in alpha cells . Quantification by FACS analysis revealed that, although the majority of proglucagon-stained cells in the colon also contained the Cre reporter in GLU-Cre12×tdRFP mice, ∼30% did not. This suggests that a small but significant number of L cells may escape Cre recombination, and should be taken into account when GLU-Cre12 mice are used in future conditional gene knockout experiments. Cre-mediated activation of RFP production was also evident in some cells that did not stain for proglucagon, consistent with the observation that a small proportion of the red fluorescent cells in primary colonic cultures did not exhibit morphology typical of L cells. These cells were readily identifiable and could be avoided in single-cell imaging experiments, but are likely to reflect transient transgene activation in a different cell population during development. Similar findings in the other founder GLU-Cre strains (see ESM Table 2) suggest that this is not merely an artefact of the particular transgene integration site.
Previous analysis detected Glut1, Glut2 and Glut5 expression in primary murine L cells, with Glut2 evident in L cells from the SI, and Glut1 in those from the colon . GLUTag cells notably lack Glut2 , but express Glut3, as determined by Affymetrix microarrays (data not shown). The role of GLUTs in incretin hormone secretion is unclear, as their pharmacological inhibition had no significant effect on glucose-stimulated GLP-1 secretion in primary cultures, although mice lacking GLUT2 showed reduced plasma GLP-1 concentrations following oral glucose, and a lower intestinal GLP-1 content . Whereas SGLT1 is apically located on L cells , GLUT2 appeared localised to the basolateral surface of L cells and enterocytes, suggesting that intracellular glucose concentrations would reflect basolateral rather than luminal glucose levels. Whereas GLP-1 secretion is predominantly stimulated by oral rather than systemic glucose delivery, GLP-1 release from the perfused pig intestine was found to be influenced also by vascular glucose levels , possibly through alteration of the intracellular glucose concentration in L cells.
Whether intracellular glucose metabolism plays any role in determining GLP-1 secretion remains uncertain. In GLUTag but not primary L cells, we observed a strong inhibition of secretion when glucose uptake was completely blocked and an amplifying action of glucose under depolarising conditions. The balance between the metabolic and electrogenic effects of glucose is thus slightly different between the cell line and primary culture, with a more dominant role for SGLT1-based glucose sensing in the latter. NAD(P)H autofluorescence measurements suggest that GLUTag and primary L cells increase their metabolic rate in response to extracellular glucose elevation, consistent with our previous observation that ATP concentrations in GLUTag cells are elevated upon exposure to 1 mmol/l glucose . NAD(P)H changes occurring at glucose concentrations above ∼1 mmol/l would be consistent with the recruitment of Glucokinase, which is known to be expressed in enteroendocrine cells [7, 9, 29, 30, 31]. Glucokinase activity was demonstrable in GLUTag cell extracts, and was responsive to the glucokinase activator, GKA50. The observed S0.5 value (∼3 mmol/l) in the absence of GKA50 is lower than the expected value of ∼5-8 mmol/l , possibly reflecting incomplete inhibition of hexokinases I–III or additional regulation of enzyme activity by unknown factors in GLUTag cell extracts. GKA50 significantly affected NAD(P)H autofluorescence at 3 mmol/l glucose in GLUTag cells, demonstrating that glucokinase exhibits at least some control over the metabolic flux in L cells, but had only a small effect on glucose-stimulated GLP-1 secretion from GLUTag cells and no effect on secretion from primary cultures.
The present study demonstrates that metabolism plays at best a minor role in glucose-stimulated GLP-1 secretion in primary cultures, consistent with the finding that non-metabolisable glucose analogues such as αMG are effective stimuli of GLP-1 release in vivo and in vitro [10, 11, 16, 32]. Although phloretin abolished glucose accumulation in GLUTag and primary L cells, its effect on glucose-stimulated GLP-1 secretion was restricted to the cell line, suggesting that glucose metabolism does not enhance secretion in the context of a predominant SGLT1-mediated stimulus. Future work should address whether the glucokinase/KATP channel machinery exerts longer-term effects on L cells or enables modulation of GLP-1 secretion by neurohormonal or alternative nutritional stimuli.
We would like to thank M. Bowen and M. Gratian (Cambridge Institute for Medical Research [CIMR]) and A. Petrunkina and V. Romashova (CIMR) for help with microscopy and flow cytometry, respectively. D. Drucker (Samuel Lunenfeld Research Institute, Toronto, Canada) kindly provided the GLUTag cell line. R. Sprengel (Max Planck Institute for Medical Research, Heidelberg, Germany) and H.J. Fehling (University of Ulm, Ulm, Germany) kindly provided a plasmid containing iCre sequence and the Rosa26-tdRFP reporter mice, respectively. The glucokinase activator, GKA50, was a gift from AstraZeneca.
This work was funded by Wellcome Trust grants to F.M. Gribble and F. Reimann (WT088357 and WT084210) and by the Deutsche Forschungsgemeinschaft to H. Koepsell (Grant SFB 487/C1). GLP-1 assays and mouse husbandry were supported by the MRC Centre for Obesity and Related Metabolic Diseases (Cambridge).
Duality of interest
AstraZeneca was not involved in the research described in this manuscript beyond the gift of GKA50, but approved the manuscript prior to submission. The authors declare that there is no other duality of interest associated with this manuscript.
HEP, AA, GR and PR designed and performed experiments, HK generated and provided the Sglt1 knockout mice and contributed to design and interpretation of experiments involving these mice. HEP, FR and FMG designed the study and drafted the manuscript. All authors contributed to critical revision of the manuscript and approved the final version.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
- 8.Nielsen LB, Ploug KB, Swift P et al (2007) Co-localisation of the Kir6.2/SUR1 channel complex with glucagon-like peptide-1 and glucose-dependent insulinotrophic polypeptide expression in human ileal cells and implications for glycaemic control in new onset type 1 diabetes. Eur J Endocrinol 156:663–671PubMedCrossRefGoogle Scholar
- 32.Shima K, Suda T, Nishimoto K, Yoshimoto S (1990) Relationship between molecular structures of sugars and their ability to stimulate the release of glucagon-like peptide-1 from canine ileal loops. Acta Endocrinol (Copenh) 123:464–470Google Scholar