Abstract
One aspect of the effects of metformin on glucagon-like peptide (GLP)-1 might be associated with the mechanism by which the cross talk between insulin and Wnt signaling enhances GLP1 secretion, due to the action of metformin as an insulin sensitizer. However, this remains completely unknown. In this study, we have investigated the mechanisms of the action of metformin on cross talk between insulin and Wnt signaling. GLP1 enhancement by meformin was determined in human NCI-H716 intestinal L-cells and hyperglycemic db/db mice treated with metformin (0.25 and 0.5 mM and/or 12.5 mg/kg body weight) for 24 h and 2 months. Metformin increased GLP1 secretion in L-cells and db/db mice. Metformin stimulated the nuclear translocation of β-catenin and TOPflash reporter activity, and gene depletion of β-catenin or enhancement of mutation of transcription factor 7-like 2 binding site offset GLP1. In addition, insulin receptor substrate 2 gene depletion blocked metformin-enhanced β-catenin translocation. These effects were preceded by an increase in glucose utilization and calcium influx, the activation of calcium-dependent protein kinase, and, in turn, the activation of insulin signaling, and the inhibition of glycogen synthase kinase 3β, a potent inhibitor of β-catenin. Furthermore, high blood glucose levels were controlled via GLP1 receptor-dependent insulinotropic pathways in db/db mice, which were evidenced by the increase in GLP1 and insulin levels at 30 min after oral glucose loading and pancreatic insulinotropic gene expression. Our findings indicate that the cooperation between Wnt and its upstream insulin signaling pathways might be a novel and important mechanism underlying the effects of metformin on GLP1 production.
Introduction
Glucagon-like peptide (GLP)-1 is an intestinal hormone that exerts its effects in the regulation of glucose metabolism, the stimulation of pancreatic insulin secretion, proinsulin gene expression, and the proliferation and anti-apoptosis of β-cells as well as the inhibition of glucagon release, gastric emptying, and food intake. The insulinotropic effect of GLP1 has led to the development of GLP1 analogs for the treatment of type 2 diabetes (Baggio & Drucker 2007, Kim & Lee 2010).
Metformin, a biguanide that was introduced as an anti-diabetic agent, acts through inhibition of hepatic gluconeogenesis and stimulation of glucose uptake into peripheral tissues (Sarabia et al. 1992, Radziuk et al. 1997). Recently, metformin has been increasingly used in combination with new incretin-based therapies such as GLP1 analogs and dipeptidyl peptidase (DPP)-4 inhibitors (D'Alessio 2011), but direct interaction between metformin and the incretin axis has also been suggested; the administration of metformin to obese subjects with or without type 2 diabetes or DPP4-deficient F344/DuCrj rats lead to increase in GLP1 levels (Yasuda et al. 2002, Mannucci et al. 2004). Furthermore, metformin selectively increased plasma levels of GLP1 in mice without alteration in another incretin hormone, glucose-dependent insulinotropic polypeptide (GIP), or peptide YY, a gastrointestinal hormone that is colocalized with GLP1 in L-cells (Maida et al. 2011). Collectively, these observations indicate that the action of metformin on the gut endocrine system may be L-cell-specific and, more precisely, GLP1-specific, and may be distinct from DPP4 inhibition.
Recently, it has been reported that the insulin and Wnt signaling pathways converge their signals to produce GLP1 (Yi et al. 2008). The binding of Wnt ligands to the frizzled receptor and its coreceptor, lipoprotein-related protein (LRP)5/6, leads to the translocation of β-catenin into the nucleus and combination of β-catenin with transcription factor 7-like 2 (TCF7L2), which has its binding site in the G2 enhancer element of proglucagon (Glu), a precursor of GLP1 (Akiyama 2000). Insulin enhances GLP1 production via inhibition of glycogen synthase kinase (GSK)-3β-dependent degradation of β-catenin (Yi et al. 2008). Furthermore, in human hepatocytes, granulosa cells, and C2C12 myotubes, metformin quickly leads to insulin receptor phosphorylation and enhances insulin signaling, e.g., phosphoinositide 3 kinase (PI3K), both in the presence or absence of insulin (Kumar & Dey 2002, Gunton et al. 2003, Rice et al. 2011). We therefore hypothesized that metformin could increase GLP1 production and secretion via the bipartite transcription factor β-catenin/TCF7L2, the major effector of the canonical Wnt signaling pathway, which might be controlled by its upstream insulin signaling.
To investigate our hypothesis, we treated NCI-H716 cells, a cellular model for GLP1 secretion (Reimer et al. 2001), with metformin (0.25, 0.5, 1, and 2 mM) for 24 h. We also investigated the effects of metformin in vivo using db/db mice, an animal model of type 2 diabetes, by administering metformin (12.5 and 50 mg/kg body weight, BW) for 2 months.
Materials and methods
Chemicals and reagents
Metformin, LY294002, and nifedipine were purchased from Sigma–Aldrich and Calbiochem (San Diego, CA, USA). Metformin was dissolved in 0.25% carboxymethylcellulose for oral administration.
Cell culture and experimental animals
NCI-H716 cells (cell lines ≤10 passages, ATCC, Manassas, VA, USA) were maintained in a suspension culture as recommended by the supplier. Two days before the experiment, 200×104 cells were seeded into six-well culture plates precoated with Matrigel, as described previously (Reimer et al. 2001). On the day of the experiment, L-cells were treated with metformin (0.25 or 0.5 mM) for 24 h. Mouse studies were conducted with the approval of and in accordance with the guidelines of the Animal Care and Use Committee of the Samsung Biomedical Research Institute (SBRI No. C-A9-228-1, Seoul, Korea), which comply with national and international laws and policies (National Institutes of Health Guide for Care and Use of Laboratory Animals, 7th edition, 1996). Metformin (10 ml/kg BW) at a dose of 12.5 or 50 mg/kg BW per day was administered to 12-week-old db/db male mice (Central Lab Animal, Seoul, Korea) by oral gavage for 8 weeks. Animals were housed individually in the animal facility with a 12 h light:12 h darkness cycle.
Oral and intraperitoneal glucose tolerance tests
Oral and intraperitoneal glucose tolerance tests (OGTT and IPGTT) were performed after oral and/or i.p. administration of glucose (1 g/kg BW) as described previously (Wheatcroft et al. 2007). Blood glucose concentrations were measured using a glucose analyzer (APEC, Danvers, MA, USA) at 0, 30, 60, 90, and 120 min. Also, serum GLP1 and insulin levels were measured at 30 min after oral glucose challenge to determine the relationship between the glucose-lowering effect and GLP1 enhancement by metformin.
Biochemical assays
ELISA kits were used to quantify serum GLP1 (EDI, San Diego, CA, USA), insulin (Mercodia, Uppsala, Sweden), glucagon (R&D Systems, Minneapolis, MN, USA), C-peptide (Cusabio Biotech, Wuhan, China), and pancreatic homogenate cAMP (Enzo Life Sciences International, Plymouth Meeting, PA, USA).
Measurement of intracellular ATP
After stimulation with 16.7 mM glucose, cells were harvested and 2.5% trichloroacetic acid (TCA) was added to extract ATP. The TCA-containing supernatant was neutralized with Tris-acetate (pH 7.75). ATP content was measured using a luciferin/luciferase method (Promega) and determined using a Wallac Trilux 1450 Microbeta liquid scintillation counter (Wallac, Turku, Finland).
Glucose uptake assay
After treatment of cells with metformin for 24 h and glucose deprivation for 5 min, 30 μM 6-(N-(7-nitrobronz-2-oxal-1,3-diazol-4-yl)amino)-6-deoxyglucose (2-NBDG; Molecular Probes, Eugene, OR, USA), a fluorescent and non-metabolizable glucose analog (Kumar & Dey 2002, Rice et al. 2011), in Krebs–Ringer Bicarbonate (KRB) buffer was added for 10 min. The cells were washed twice with PBS and resuspended in 1 ml PBS. Glucose uptake by L-cells was determined using a FACSort flow cytometer (Becton-Dickinson, Franklin Lakes, NJ, USA) at an excitation wavelength of 488 nm. Vehicle-treated cells were used as controls.
Measurement of intracellular calcium using Fluo-4 staining
After treatment with metformin for 24 h, cells were stimulated for 1 h with 16.7 mM glucose in KRB buffer and treated with 3 mM Fluo-4 (Molecular Probes) in Ca2+-free Krebs–Ringer–HEPES buffer supplemented with 1% BSA for 1 h without light. Cells that displayed Ca2+ influx were counted by using the software Image J (http://rsbweb.nih.gov/ij/).
Enzyme activity assay
Mitochondrial complex I activities were expressed by measuring kinetic substrate (NADH) oxidation (MitoScience, Eugene, OR, USA). The mitochondria of cells and intestinal tissue were extracted by lysis and homogenization followed by low- (600 g) and high- (11 000 g) speed centrifugation using a mitochondria isolation kit (Sigma–Aldrich). DPP4 activity was determined by fluorometrically monitoring the conversion of Gly-Pro-4-oMe-2-NA into 7-amino-4-methyl-coumarin for 1 min (Calbiochem).
TOP/FOPflash transfection
cDNA TOPflash and FOPflash reporter plasmids were purchased from Millipore (Billerica, MA, USA). In these plasmids, expression of a luciferase reporter gene is driven by a minimum TK promoter fused with three copies of the TCF7L2-binding site and its mutant-binding site. The L-cells (50×104) were seeded into 24-well plates for 24 h and transfected with 100 ng TOPflash or FOPflash using Lipofectamine (Invitrogen). After transfection, cells were treated with metformin for 24 h and harvested for luciferase reporter gene analysis (Promega).
RNA interference
siRNA against human β-catenin, insulin receptor substrate 2 (IRS2), and scrambled RNA (as control) were purchased from Dharmacon Research (Lafayette, CO, USA). The L-cells (200×104) were seeded into six-well plates for 24 h and then transfected with 50 nM siRNA using Lipofectamine 2000 (Invitrogen). After transfection, cells were treated with metformin for 24, 48, or 72 h.
Immunohistochemical and immunofluroscent staining
For immunohistochemistry (IHC), isolated mouse pancreas and intestinal tissue were fixed with 10% formalin, embedded in paraffin, mounted on slides, deparaffinized in xylene and ethanol, and blocked with 10% normal horse serum, as detailed previously (Brown et al. 1996). Tissues were incubated with primary antibodies (1:200) overnight at 4 °C followed by the addition of biotinylated anti-rabbit IgG and then avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, CA, USA). Nuclei were stained with hematoxylin (Sigma–Aldrich). Tissues for immunofluorescence were incubated with primary antibody (1:200) followed by secondary antibody (1:1000) labeled with FITC (Invitrogen). Nuclei were stained with Hoechst 33258 (0.5 mM in PBS; Sigma–Aldrich). Images were captured using an Eclipse TE 200 confocal microscope (Nikon, Tokyo, Japan) and quantified by using the software Image J (http://rsbweb.nih.gov/ij/) for the measurement of GLP1R fluorescence intensity (Fig. 1B) or islet area (Supplementary Figure S7, see section on supplementary data given at the end of this article).
Western blotting
Cells were lysed in lysis buffer (20 mM Tris–HCl; pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate) containing NaF, phenylmethanesulfonylfluoride, and Na3VO4, and centrifuged at 12 000 g for 10 min at 4 °C. Twenty micrograms of protein were separated by 10% SDS–PAGE and transferred onto nitrocellulose membranes. Proteins were visualized with primary antibody (1:1000), followed by HRP-labeled secondary antibody (1:5000). Western blot bands were quantified using the software Image J (http://rsbweb.nih.gov/ij/) and corrected by β-actin (Supplementary Figure S1, see section on supplementary data given at the end of this article). Relative activity of each protein was calculated by the formula: phosphorylated protein band/total protein band (Supplementary Figure S2).
Total RNA extraction and real-time PCR
Total RNA was isolated and reverse-transcribed using TRIzol reagent and a reverse transcriptase kit (Invitrogen) according to the manufacturer's instructions. Real-time PCR was conducted using a Biosystem 7500 real-time PCR system (ABI, Foster City, CA, USA). Five hundred nanograms of cDNA were amplified in a 25 μl SYBR Green PCR master mix (ABI) with specific primer pairs (Supplementary Table S1, see section on supplementary data given at the end of this article). The relative expression levels of target genes were calculated from the change in the threshold cycle (ΔCt) as 2−ΔΔCt, where ΔΔCt=ΔCttreatment−ΔCtcontrol and ΔCt=ΔCttarget gene=ΔCtβ actin.
Microarray analysis
Gene expression profiles were assessed using a Ref-8 v3 Expression BeadChip (Illumina, San Diego, CA, USA), which contains more than 46 000 individual human genes per piece. Each gene was derived from the NCBI Human Reference Sequence (RefSeq) database. For microarray analysis, the isolated RNA were subjected to cDNA synthesis and transcription-biotinylated cRNA synthesis; they were then hybridized, scanned, and quantified using an Illumina bead array reader and software (Illumina Beadstation 500; Illumina) as detailed previously (Kim et al. 2010). Target sequence signals with detection P values of ≤0.05 were quantile normalized using the program R (version 2.4.1). Biological processes were analyzed via Protein Analysis Through Evolutionary Relationships (PANTHER; http://www.pantherdb.org). The results of two-way hierachical clustering analysis of gene expression profiles are shown in Supplementary Figure S6, see section on supplementary data given at the end of this article.
Statistical analyses
All experiments were conducted at least in triplicate (n=3), and the results were expressed as the mean±s.e.m. Statistical analyses were conducted via ANOVA, followed by Duncan's multiple range test or Student's t-test using SAS Proprietary Software Release 8.2 (SAS Institute, Cary, NC, USA). Data with P values <0.05 were considered significant.
Results
Metformin increases expression and content of GLP1 as well as Glu
Metformin increased both the contents and the protein expression levels of GLP1, as well as the expression of Glu in L-cells compared when with controls (P<0.05, Fig. 2A, C, and D). Also, db/db mice treated with meformin had higher contents and expression levels of GLP1, as well as higher expression of Glu in serum and the intestines, than mice not treated with metformin (P<0.05, Fig. 2B, C, and D). The increment in serum GLP1 was larger in the postprandial state (90 min) than in the fasting state in db/db mice (P<0.05, Fig. 2B). IHC revealed that the expression of GLP1 (green color) was more abundant in sections of the ileum from metformin-treated mice than controls (Fig. 2E). Interestingly, at low concentrations of metformin (0.25 and 0.5 mM), GLP1 production was increased but the phosphorylation of AMPK and its upstream liver kinase B (LKB) was not stimulated, whereas at high concentrations (1 and 2 mM and/or 50 mg/kg BW), the phosphorylation of AMPK and LKB was increased without increase in GLP1 production (Fig. 2D and Supplementary Figure S1A and S2, see section on supplementary data given at the end of this article). The latter might be attributable to an inhibitory effect by the excessive amount of metformin, which was proven by MTT assay (Fig. 2F). Thus, we chose the doses (0.25 and 0.5 mM and/or 12.5 mg/kg BW) of metformin, which did not decrease cellular viability or BW (Supplementary Figure S3) and enhanced GLP1 production, to investigate the mechanisms. These findings indicate that the stimulatory effects of metformin on GLP1 production are not mediated through an AMPK-dependent pathway.
Metformin enhances GLP1 levels via the Wnt signaling pathway which is controlled by upstream insulin signaling
In L-cells treated with metformin, the expression and nuclear translocation of β-catenin were increased compared with the controls (Fig. 3A and Supplementary Figure S1B). GLP1 enhancement by metformin was closely associated with β-catenin nuclear translocation (Fig. 3B). In the absence of β-catenin by using siRNA, the stimulatory effects of metformin on GLP1 secretion disappeared, but were later restored concomitantly with increased β-catenin expression (72–96 h post-transfection of β-catenin siRNA into L-cells, Fig. 3B). Metformin also increased the transcription of a luciferase reporter gene containing three copies of the TCF7L2-binding sites (TOPflash) in Glu, but did not alter the transcription of a luciferase reporter gene containing the mutant TCF7L2-binding sites in L-cells (FOPflash, P<0.05, Fig. 3C). In this regard, GLP1 enhancement by 0.5 mM metformin was offset by transfection of FOPflash, but not TOPflash (P<0.05, Fig. 3D). The microarray data for L-cells also demonstrated that metformin increased the expression of cyclin D1, a Wnt-responsive gene, but decreased the expression of GSK3β, a β-catenin inhibitor, compared with controls (Table 1 and Supplementary Figure S5). Interestingly, depletion of IRS2 by siRNA blocked metformin-induced translocation of β-catenin into the nucleus and enhancement of GLP1 in the L-cells (Fig. 3E and F and Supplementary Figure S1C). In this regard, these findings indicate that β-catenin/TCF7L2 may play a primary role in GLP1 enhancement by metformin, and the insulin signaling pathway is upstream of Wnt signaling. In the intestines of metformin-treated mice, both expression and nuclear translocation of β-catenin were increased in a similar pattern to that observed in L-cells (Fig. 3A and Supplementary Figure S1B).
Gene expression profiles in L-cells exposed to 0.5 or 2 mM metformin. The fold changes are represented as the mean gene expression ratios from each of two independently repeated microarray experiments comparing cells treated with metformin (0.5 or 2 mM)
Genbank ID | Gene name | Fold change | ||
---|---|---|---|---|
Gene symbol | Gene description | 0.5 mM | 2 mM | |
Wnt signaling | ||||
| PRAD1 | Cyclin D1 (CCND1), mRNA | 1.51 | −1.65 |
| GSK3β | Glycogen synthase kinase 3 beta (GSK3B), mRNA | 1.43 | 2.07 |
Enzymes involved in glycolysis and TCA | ||||
| PFKM | Phosphofructokinase, muscle (PFKM), mRNA | 1.92 | 1.38 |
| PDPR | Pyruvate dehydrogenase phosphatase regulatory subunit (PDPR), mRNA | 1.48 | −1.04 |
| IDH3A | Isocitrate dehydrogenase 3 (NAD+) alpha (IDH3A), nuclear gene encoding mitochondrial protein, mRNA | 1.56 | −1.01 |
| IDH1 | Isocitrate dehydrogenase 1 (NADP+), soluble (IDH1), mRNA | 1.58 | −1.42 |
| MDH1 | Malate dehydrogenase 1, NAD (soluble) (MDH1), mRNA | 1.41 | −1.07 |
| ATP1B1 | ATPase, Na+/K+ transporting, beta 1 polypeptide (ATP1B1), transcript variant 1, mRNA | 2.22 | 1.42 |
Activation of the insulin signaling pathways occurs through the increase in glucose utilization and calcium mobilization by metformin
In the L-cells, metformin increased the expression of insulin signaling molecules such as IRS2, phospho-PI3K, and phospho-protein kinase B (PKB) and its upstream genes including calcium-calmodulin-dependent kinase 2 (CAMK2) and cAMP response element binding protein (CREB), and the phosphorylation of its downstream gene GSK3β at serine residue 9 as compared with controls (Fig. 4A and Supplementary Figure S1, E and F, and S2). Interestingly, this insulin signaling axis was regulated by glucose utilization and calcium mobilization. FACS and luciferin/Luc analyses, the methods for glucose uptake and ATP contents, revealed that metformin increased the uptake of 2-NBDG, a non-matabolizable glucose analog, into the L-cells, and increased intracellular ATP levels, an indicator of glucose utilization in L-cells, as compared with controls (P<0.05, Fig. 4B and C), which were followed by an increase in the influx of calcium (stained by Fluo-4 dye) through L-type calcium channels, which was evidenced by the addition of 10 nM nifedipine, an L-type calcium channel blocker. The addition of nifedipine resulted in a decrease in intracellular calcium levels as compared with metformin alone in L-cells (P<0.05, Fig. 4D). Taken together, the glucose utilization and a subsequent calcium mobilization are thought to stimulate the phosphorylation of CAMK2 and CREB, calcium-dependent protein kinases (Fig. 4A and Supplementary Figure S1, E and F, and S2).
Enhancement of glucose uptake into L-cells was associated with an increased expression of sodium/glucose cotransporter 1 (SGLT1) and its upstream molecule Hur (P<0.05, Fig. 4E and Supplementary Figure S1D). Furthermore, increased ATP content in L-cells was associated with an increase in gene expression, including expression of the genes encoding the enzymes required for glycolysis and the tricarboxylic acid (TCA) cycle, such as phosphofructokinase, isocitrate dehydrogenase 3, and malate dehydrogenase 1 (Table 1 and Supplementary Figure S5), as well as mitochondrial complex I activity (P<0.05, Fig. 4F and Supplementary Figure S4). The participation of these signaling factors in GLP1 production and secretion was demonstrated by a variety of inhibitors, including nifedipine, LY294002, and SGLT1 siRNA (P<0.05, Fig. 4G). SGLT1 expression was dramatically decreased by 48 h-post transfection of SGLT1 siRNA into L-cells.
Animal study also revealed increases in the expression of genes involved in insulin signaling axis (Fig. 4A), intracellular levels of ATP (Fig. 4C), the expression of SGLT1 and Hur (Fig. 4E), and the activity of complex I (Fig. 4F) by metformin administration.
Elevated GLP1 levels and its incretin effect contribute to the glucose-lowering effect of metformin
Mice treated with metformin exhibited lower blood glucose levels and AUC values after oral and/or intraperitoneal glucose compared with mice that had not received metformin (P<0.05) (Fig. 1A and Supplementary Figure S6). To determine whether this glucose-lowering effect might be related to the elevated GLP1 levels and a subsequent improvement in insulin secretory response, serum GLP1 and insulin level were measured at 30 min after glucose challenge. As expected, the metformin-treated mice showed higher GLP1 and insulin levels in serum than controls (Fig. 1A middle and right panel) in contrast to blood glucose levels. These results were associated with higher expression of GLP1 receptor in islets and the increase in molecules and/or genes involved in GLP1-dependent insulinotropic mechanisms, such as cAMP, phospho-CREB, and pancreas/duodenum homeobox 1 (PDX1; Fig. 1B, C, D, and E and Supplementary Figure S1G and S2) in the pancreas. Activation of this insulinotropic mechanism led to the elevation of serum insulin and C-peptide levels but not glucagon level (P<0.05, Table 2) and the increase in number and mass of β-cells of metformin-treated mice compared with controls (Fig. 1F and Supplementary Figure S7, see section on supplementary data given at the end of this article). Collectively, the elevated levels of GLP1 might be one of the important mechanisms underlying the glucose-lowering effects of metformin (Fig. 5).
Biochemical parameters involved in glucose homeostasis. Insulin, C-peptide, and glucagon were measured in sera from db/db mice (n=3) after overnight fasting. Data are expressed as mean±s.e.m. (n=3–5)
Insulin (μg/l) | C-peptide (ng/ml) | Glucagon (ng/ml) | |
---|---|---|---|
Control | 1.84±0.01 | 6.32±0.32 | 6.932±0.13 |
Met (12.5 mg/kg BW) | 3.68±0.62* | 15.80±0.79* | 4.88±0.33* |
*Significant differences between two groups (Student's t-test; P<0.05).
Discussion
We observed that the glucose-lowering effects of metformin may be at least associated with the elevated GLP1 levels and incretin effect on the pancreas, which is supported by the increase in serum GLP1 and insulin levels and the activation of pancreatic insulinotropic mechanisms after glucose challenge. Until recently, this stimulatory effect of metformin on GLP1 was reported to be associated with the inhibition of DPP4, an enzyme that abolishes the biological activity of GLP1 (D'Alessio 2011). However, in our study we identified a novel mechanism by which metformin directly stimulated GLP1 production and secretion through cross talk between the insulin and Wnt signaling pathways. Indeed, metformin increased the expression and contents of GLP1 as well as the activation of the insulin and Wnt signaling pathways in L-cells and in the intestines of db/db mice. However, this effect of metformin on GLP1 was blocked by β-catenin and Irs2 gene depletion, and/or PI3K inhibitor, and interestingly β-catenin nuclear translocation by metformin was offset by Irs2 gene depletion. In this regard, GLP1 enhancement by metformin occurred through β-catenin and its upstream insulin signaling, and metformin increased β-catenin nuclear translocation and the transcription of the Wnt-responsive gene Glu, a precursor of GLP1, through the activation of the insulin signaling pathway (IRS2/phospho-PI3K/phospho-PKB) followed by the inhibition of GSK-3β.
These findings are consistent with recent studies, which showed that insulin enhances intestinal GLP1 production through cooperation with Wnt signaling pathways (Yi et al. 2008) and metformin activates the insulin signaling pathways such as IRS2 and PI3K in the hepatocytes, granulosa cells, and C2C12 myotubes (Kumar & Dey 2002, Gunton et al. 2003).
On the other hand, research into the action of metformin on the insulin signaling pathway has been focused on IRS1/2 and/or PI3K. However, we found a new pathway, i.e., glucose utilization and calcium mobilization. Metformin therapy has long been known to lower postprandial blood glucose concentrations via inhibition of intestinal glucose absorption, but this hypothesis has been disputed (Jackson et al. 1987). In our study, metformin increased the intestinal absorption of glucose by increasing the expression of SGLT1, an energy-dependent glucose transporter in the brush border membrane of the lumen. Increased SGLT1 expression may be associated with an increase in Hur protein, as Hur protein has been demonstrated to bind to the UTR of SGLT mRNA and upregulate SGLT1 expression (Loflin & Lever 2001). In addition, our microarray data showed that metformin increased the expression of sodium–potassium ATPase, which might lead to a lower gradient of sodium ions in enterocytes and consequently stimulate the uptake of sodium ions by upregulating the expression of SGLT1. Interestingly, metformin increased the intestinal utilization of absorbed glucose, which was estimated by measuring intracellular ATP levels, through the upregulation of genes involved in glycolysis, the TCA cycle, and mitochondrial complex I. These results were in agreement with other studies which showed that metformin significantly increased SGLT1 (SLC5A1) gene expression as well as glucose utilization in the duodenum and jejunum (Lenzen et al. 1996) and metformin administration reversed the decreased activity of PFK, a key enzyme of glycolysis, in the skeletal muscle, liver, and adipose tissue of diabetic mice (Da Silva et al. 2010). However, metformin has been reported to suppress NADH-dependent dehydrogenase complex I activity, because it induces NADH-dependent lactic acidosis in cases of excessive dosage (Batandier et al. 2006, Piwkowska et al. 2010). Indeed, higher concentrations (1 and/or 2 mM) of metformin suppressed complex I activity (Supplementary Figure S4) but lower concentrations of metformin (0.5 mM and 12.5 mg/kg BW) did not. This discrepancy between dosages warrants further study.
An increase in ATP could close ATP-sensitive potassium channels and a subsequent membrane depolarization leads L-type calcium channels to open followed by calcium mobilization (Reimann et al. 2005, Bailey et al. 2008). The opening of L-type calcium channels was determined using nifedipine, an L-type calcium channel blocker. Addition of nifedipine to metformin reversed the increase in calcium levels in L-cells treated with metformin. Also, the calcium mobilization stimulated the phosphorylation of CAMK2 and CREB and the induction of IRS2 as shown in previous studies (Sheng et al. 1991, Hook & Means 2001). In this regard, glucose utilization and calcium mobilization might be the primary mechanism by which metformin activates insulin signaling pathways in the intestinal cells.
Metformin decreased blood glucose levels concurrently with the elevation of GLP1 and insulin levels. In addition, mice treated with metformin showed higher expression of GLP1 receptor in islets, and higher cAMP levels and phospho-CREB and PDX1 expression in the pancreas than mice not treated with metformin. These results are consistent with the evidence of other studies which showed that metformin enhanced the expression of the genes encoding the receptors for both GIP and GLP1 in mouse islets and also increased the effects of GIP and GLP1 on insulin secretion from β-cells (Maida et al. 2011). Thus, these results indicate the possibility that the blood glucose-lowering effects of metformin are associated with the elevated GLP1 levels, because GLP1 binds to the N-terminal extracellular region of the GLP1 receptor in β-cells (Montrose-Rafizadeh et al. 1999) increasing intracellular cAMP levels via adenylate cyclase, and subsequently inducing the phosphorylation of CREB (Holz 2004) and the expression of PDX1 (Wang et al. 1999). Activation of this insulinotropic mechanism led to the elevation of insulin and C-peptide levels during fasting as well as glucose loading in mice treated with metformin twice daily for 4–5 days. Our findings could be interpreted as relevant to humans, as intestinal and/or oral metformin concentrations (0.25, 0.5 mM, and/or 12.5 mg/kg BW) are within the range recorded in the jejunum, which had about 30–300 times higher concentrations than plasma (8–24 μM), in patients who took metformin at the dosage of 850 mg once daily for 2–3 weeks (Tucker et al. 1981, Bailey et al. 2008) and the db/db mice have similar phenotypes to human type 2 diabetic patients.
Taken together, these results indicates that metformin seems to enhance GLP1 production and secretion via the cooperation between Wnt and its upstream insulin signaling pathways, and that glucose utilization and calcium mobilization play a prime role in the activation of insulin signaling pathways. Furthermore, metformin could exert glucoregulatory actions via the modulation of the incretin axis, which provides the novel function of metformin as a mediator of endocrine communication between the gut and pancreatic islets, and which could be a novel target for drug development.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/JOE-13-0381.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Department of Health and Welfare Grant (A084065) of Korea and Samsung Biomedical Research Institute Grant (C-A9-228-1) of Korea.
References
Akiyama T 2000 Wnt/β-catenin signaling. Cytokine & Growth Factor Reviews 11 273–282. (doi:10.1016/S1359-6101(00)00011-3)
Baggio LL & Drucker DJ 2007 Biology of incretins: GLP-1 and GIP. Gastroenterology 132 2131–2157. (doi:10.1053/j.gastro.2007.03.054)
Bailey CJ, Wilcock C & Scarpello JH 2008 Metformin and the intestine. Diabetologia 51 1552–1553. (doi:10.1007/s00125-008-1053-5)
Batandier C, Guigas B, Detaille D, El-Mir MY, Fontaine E, Rigoulet M & Leverve XM 2006 The ROS production induced by a reverse-electron flux at respiratory-chain complex 1 is hampered by metformin. Journal of Bioenergetics and Biomembranes 38 33–42. (doi:10.1007/s10863-006-9003-8)
Brown D, Lydon J, McLaughlin M, Stuart-Tilley A, Tyszkowski R & Alper S 1996 Antigen retrieval in cryostat tissue sections and cultured cells by treatment with sodium dodecyl sulfate (SDS). Histochemistry and Cell Biology 105 261–267. (doi:10.1007/BF01463929)
D'Alessio DA 2011 What if gut hormones aren't really hormones: DPP-4 inhibition and local action of GLP-1 in the gastrointestinal tract. Endocrinology 152 2925–2926. (doi:10.1210/en.2011-1385)
Da Silva D, Zancan P, Coelho WS, Gomez LS & Sola-Penna M 2010 Metformin reverses hexokinase and 6-phosphofructo-1-kinase inhibition in skeletal muscle, liver and adipose tissues from streptozotocin-induced diabetic mouse. Archives of Biochemistry and Biophysics 496 53–60. (doi:10.1016/j.abb.2010.01.013)
Gunton JE, Delhanty PJ, Takahashi S & Baxter RC 2003 Metformin rapidly increases insulin receptor activation in human liver and signals preferentially through insulin-receptor substrate-2. Journal of Clinical Endocrinology and Metabolism 88 1323–1332. (doi:10.1210/jc.2002-021394)
Holz GG 2004 Epac: a new cAMP-binding protein in support of glucagon-like peptide-1 receptor-mediated signal transduction in the pancreatic β-cell. Diabetes 53 5–13. (doi:10.2337/diabetes.53.1.5)
Hook SS & Means AR 2001 Ca2+/CaM-dependent kinases: from activation to function. Annual Review of Pharmacology and Toxicology 41 471–505. (doi:10.1146/annurev.pharmtox.41.1.471)
Jackson RA, Hawa MI, Jaspan JB, Sim BM, Disilvio L, Featherbe D & Kurtz AB 1987 Mechanism of metformin action in non-insulin-dependent diabetes. Diabetes 36 632–640. (doi:10.2337/diab.36.5.632)
Kim M-H & Lee M-K 2010 The incretins and pancreatic β-cells: use of glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide to cure type 2 diabetes mellitus. Korean Diabetes Journal 34 2–9. (doi:10.4093/kdj.2010.34.1.2)
Kim MH, Park JS, Seo MS, Jung JW, Lee YS & Kang KS 2010 Genistein and daidzein repress adipogenic differentiation of human adipose tissue-derived mesenchymal stem cells via Wnt/β-catenin signalling or lipolysis. Cell Proliferation 43 594–605. (doi:10.1111/j.1365-2184.2010.00709.x)
Kumar N & Dey CS 2002 Metformin enhances insulin signalling in insulin-dependent and-independent pathways in insulin resistant muscle cells. British Journal of Pharmacology 137 329–336. (doi:10.1038/sj.bjp.0704878)
Lenzen S, Lortz S & Tiedge M 1996 Effect of metformin on SGLT1, GLUT2, and GLUT5 hexose transporter gene expression in small intestine from rats. Biochemical Pharmacology 51 893–896. (doi:10.1016/0006-2952(95)02243-0)
Loflin P & Lever JE 2001 HuR binds a cyclic nucleotide-dependent, stabilizing domain in the 3′ untranslated region of Na+/glucose cotransporter (SGLT1) mRNA. FEBS Letters 509 267–271. (doi:10.1016/S0014-5793(01)03176-3)
Maida ALB, Cao X & Drucker DJ 2011 Metformin regulates the incretin receptor axis via a peroxisome proliferator-activated receptor α-dependent pathway in mice. Diabetologia 54 339–349. (doi:10.1007/s00125-010-1937-z)
Mannucci E, Tesi F, Bardini G, Ognibene A, Petracca MG, Ciani S, Pezzatini A, Brogi M, Dicembrini I & Cremasco F et al. 2004 Effects of metformin on glucagon-like peptide-1 levels in obese patients with and without type 2 diabetes. Diabetes Nutrition & Metabolism 17 336–342.
Montrose-Rafizadeh C, Avdonin P, Garant MJ, Rodgers BD, Kole S, Yang H, Levine MA, Schwindinger W & Bernier M 1999 Pancreatic glucagon-like peptide-1 receptor couples to multiple G proteins and activates mitogen-activated protein kinase pathways in Chinese hamster ovary cells. Endocrinology 140 1132–1140. (doi:10.1210/en.140.3.1132)
Piwkowska A, Rogacka D, Jankowski M, Dominiczak MH, Stepinski JK & Angielski S 2010 Metformin induces suppression of NAD(P)H oxidase activity in podocytes. Biochemical and Biophysical Research Communications 393 268–273. (doi:10.1016/j.bbrc.2010.01.119)
Radziuk J, Zhang Z, Wiernsperger N & Pye S 1997 Effects of metformin on lactate uptake and gluconeogenesis in the perfused rat liver. Diabetes 46 1406–1413. (doi:10.2337/diab.46.9.1406)
Reimann F, Maziarz M, Flock G, Habib AM, Drucker DJ & Gribble FM 2005 Characterization and functional role of voltage gated cation conductances in the glucagon-like peptide-1 secreting GLUTag cell line. Journal of Physiology 563 161–175. (doi:10.1113/jphysiol.2004.076414)
Reimer RA, Darimont C, Gremlich S, Nicolas-Metral V, Ruegg UT & Mace K 2001 A human cellular model for studying the regulation of glucagon-like peptide-1 secretion. Endocrinology 142 4522–4528. (doi:10.1210/en.142.10.4522)
Rice S, Pellatt LJ, Bryan SJ, Whitehead SA & Mason HD 2011 Action of metformin on the insulin-signaling pathway and on glucose transport in human granulosa cells. Journal of Clinical Endocrinology and Metabolism 96 E427–E435. (doi:10.1210/jc.2010-2060)
Sarabia V, Lam L, Burdett E, Leiter LA & Klip A 1992 Glucose transport in human skeletal muscle cells in culture. Stimulation by insulin and metformin. Journal of Clinical Investigation 90 1386–1395. (doi:10.1172/JCI116005)
Sheng M, Thompson MA & Greenberg ME 1991 CREB: a Ca2+-regulated transcription factor phosphorylated by calmodulin-dependent kinases. Science 252 1427–1430. (doi:10.1126/science.1646483)
Tucker GT, Casey C, Phillips PJ, Connor H, Ward JD & Woods HF 1981 Metformin kinetics in healthy subjects and in patients with diabetes mellitus. British Journal of Clinical Pharmacology 12 235–246. (doi:10.1111/j.1365-2125.1981.tb01304.x)
Wang X, Cahill CM, Pineyro MA, Zhou J, Doyle ME & Egan JM 1999 Glucagon-like peptide-1 regulates the β cell transcription factor, PDX-1, in insulinoma cells. Endocrinology 140 4904–4907. (doi:10.1210/en.140.10.4904)
Wheatcroft SB, Kearney MT, Shah AM, Ezzat VA, Miell JR, Modo M, Williams SC, Cawthorn WP, Medina-Gomez G & Vidal-Puig A et al. 2007 IGF-binding protein-2 protects against the development of obesity and insulin resistance. Diabetes 56 285–294. (doi:10.2337/db06-0436)
Yasuda N, Inoue T, Nagakura T, Yamazaki K, Kira K, Saeki T & Tanaka I 2002 Enhanced secretion of glucagon-like peptide 1 by biguanide compounds. Biochemical and Biophysical Research Communications 298 779–784. (doi:10.1016/S0006-291X(02)02565-2)
Yi F, Sun J, Lim GE, Fantus IG, Brubaker PL & Jin T 2008 Cross talk between the insulin and Wnt signaling pathways: evidence from intestinal endocrine L cells. Endocrinology 149 2341–2351. (doi:10.1210/en.2007-1142)