151 Uncoupling 2 negatively regulates glucose-induced glucagon-like peptide 1 secretion

Hongjie Zhang1,2, Jing Li1, Xiangying Liang1, Yun Luo1, Ke Zen1 and Chen-Yu Zhang1 1State Key Laboratory of Pharmaceutical Biotechnology, School of Biological Sciences, Jiangsu Diabetes Center, Nanjing University, Nanjing, Jiangsu 210093, People’s Republic of China

2First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210009, People’s Republic of China

(Correspondence should be addressed to C-Y Zhang; Email: [email protected]; K Zen; Email: [email protected])

Abstract

It is known that endogenous levels of the incretin hormone glucagon-like peptide 1 (GLP1) can be enhanced by various secretagogues, but the mechanism underlying GLP1 secretion is still not fully understood. We assessed the possible effect of 2 (UCP2) on GLP1 secretion in mouse intestinal tract and NCI-H716 cells, a well- characterized human enteroendocrine L cell model. Localization of UCP2 and GLP1 in the gastrointestinal tract was assessed by immunofluorescence staining. Ucp2 mRNA levels in gut were analyzed by quantitative RT-PCR. Human NCI-H716 cells were transiently transfected with siRNAs targeting UCP2. The plasma and ileum tissue levels of GLP1 (7–36) amide were measured using an ELISA kit. UCP2 was primarily expressed in the mucosal layer and colocalized with GLP1 in gastrointestinal mucosa. L cells secreting GLP1 also expressed UCP2. After glucose administration, UCP2-deficient mice showed increased glucose-induced GLP1 secretion compared with wild-type littermates. GLP1 secretion increased after NCI-H716 cells were transfected with siRNAs targeting UCP2. UCP2 was markedly upregulated in ileum tissue from ob/ob mice, and GLP1 secretion decreased compared with normal mice. Furthermore, GLP1 secretion increased after administration of genipin by oral gavage. Taken together, these results reveal an inhibitory role of UCP2 in glucose-induced GLP1 secretion. Journal of Molecular Endocrinology (2012) 48, 151–158

Introduction (Herrmann et al. 1995, Ritzel et al. 1997, Rorsman 1997, Gribble et al. 2003, Reimann et al.2006, It is widely recognized that oral ingestion of glucose Margolskee et al. 2007, Rozengurt & Sternini 2007). triggers more insulin release than intravenous delivery However, the underlying mechanism of GLP1 release of glucose, which results in a similar plasma glucose in vivo is still debated both in terms of relative profile – a phenomenon known as the ‘incretin effect’ contributions of direct vs indirect glucose-sensing (Elrick et al. 1964, McIntyre et al. 1965, Perley & Kipnis pathways and identity of the signaling pathways within 1967). This is mainly attributed to two hormones, L cells. Thus, we sought to determine which glucose- glucose-dependent insulinotropic polypeptide (GIP) sensing mechanism in the gut lumen might underlie and glucagon-like peptide 1 (GLP1) (McIntyre et al. this L-cell response. 1964, Thorens 1995). GLP1 is a product of post- It is known that one mechanism for sensing glucose translational cleavage of proglucagon (proG) peptide; in pancreatic endocrine b-cells is uncoupling protein 2 the majority of it is produced in L cells of the ileum and (UCP2)-mediated insulin secretion. UCP2 is a member colon. Although cleavage of proG results in several of the inner mitochondrial membrane anion carrier different forms of GLP1, GLP1 (7–36) amide has been superfamily (Ricqier & Bouillaud 2000, Rousset et al. reported to be the most abundant form secreted from 2007). Proton leak activity of UCP2 has an important L cells (Goke et al. 1991, Fehmann et al. 1995). role in negative regulation of insulin secretion via its Carbohydrates in the gut lumen elicit release of inhibition of ATP synthesis in b-cells (Jezek 2002). GLP1 from L cells and GIP from K cells (Hirasawa et al. Pancreatic endocrine b-cells are glucose-sensing cells 2005). It appears that glucose within the lumen of the equipped with a KATP channel consisting of two gut acts on the luminal surface to stimulate GLP1 subunits, SUR1 and Kir6.2 (Donley et al.2005). secretion (Meier & Nauck 2005). A variety of signaling Intestinal endocrine L cells secreting GLP1 are also mechanisms have been proposed to explain how L cells glucose-sensing cells and coexpress the KATP channel might sense glucose, including ATP-sensitive potassium subunits in vivo (Nielsen et al.2007). Ucp2 mRNA is (KATP) channel closure, sodium glucose widely expressed, including in the gastrointestinal tract activity, and activation of sweet taste receptors (Nedergaard & Cannon 2003, Mattiasson & Sullivan

Journal of Molecular Endocrinology (2012) 48, 151–158 DOI: 10.1530/JME-11-0114 0952–5041/12/048–151 q 2012 Society for Endocrinology Printed in Great Britain Online version via http://www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access 152 H ZHANG and others . UCP2 negatively regulates GLP1 secretion

2006). However, the physiological function of Ucp2 in RNA isolation and quantitative RT-PCR the gastrointestinal tract remains unknown. Combining The reagents used for RNA isolation and quantitative these pieces of evidence raises the question of whether RT-PCR included TRIzol (Invitrogen), oligo(dT) there is any relationship between UCP2 and GLP1 18 primers and AMV reverse transcriptase (Invitrogen), secretion in the gastrointestinal tract. and primers and TaqMan probes (Shanghai Shinegene In this study, we investigated whether L cells secreting GLP1 also expressed UCP2, and whether UCP2 may Molecular Biotechnology Co., Ltd., Shanghai, China). regulate GLP1 secretion in L cells in a similar manner Ileum tissue samples were dissected and immediately to its regulation of insulin secretion in b-cells. We frozen in liquid nitrogen. Total cellular RNA was measured plasma GLP1 levels, ileum tissue GLP1 levels, extracted using TRIzol reagent. The purity and concen- tration of RNA were determined by measuring the and plasma glucose levels before and after oral O . administration of glucose in UCP2-deficient mice, absorbance at 260 and 280 nm, where A260/280 1 7 wild-type littermates, normal C57BL/6J mice, and was considered as sufficient purity. One microgram of ob/ob mice. We also used NCI-H716 cells, a well- RNA was reverse transcribed into cDNA using characterized human L-cell model, to further verify oligo(dT)18 primers and AMV reverse transcriptase at the effect of UCP2 on secretion of GLP1. 42 8C for 1 h. Quantitative RT-PCR was performed with an ABI Prism 7000 sequence detection system (Applied Biosystems, Foster City, CA, USA) using TaqMan probes (Shanghai Shinegene Molecular Biotechnology); Materials and methods threshold cycle numbers were obtained by ABI Prism 7000 SDS software version 1.0. The primers and probe Animals and sample preparation sequences used in this study were as follows: Ucp2 (sense primer, 50-CCAATGTTGCCCGWAATG-30; anti- 0 0 Animals and treatment sense primer, 5 -TGAGGTTGGCTTTCAGG AG-3 ; and probe, 50-FAMCCTGGTGACCTATGACCTCATCA UCP2-deficient mice on a C57BL/6J background and AAG-30)andb-actin (sense primer, 50-GGCACCACACYTT- their wild-type littermates, ob/ob mice on a C57BL/6J CTACAATG-30; antisense primer, 50-GGGGGTGTTGA- background, and C57BL/6J mice were obtained from AGGTCTCAAAC-30;andprobe,50-FAMCTGTGGCCCC- the animal center at Nanjing University. Mice were TGAGGAGCACCC-30). Amplification conditions were maintained and the experiments were performed one cycle at 95 8C for 5 min followed by 40 cycles at according to the National Institutes of Health Guide 95 8C for 30 s and 60 8C for 1 min, and one final cycle at for Care and Use of Laboratory Animals. All mice 72 8Cfor4min. (males, aged 8–10 weeks) were maintained at 23 8C with a regular 12 h light:12 h darkness cycle. Following overnight fasting with free access to water, UCP2- Western blot deficient mice (nZ20) and wild-type littermates (nZ20) were given glucose (2 g/kg) by oral gavage Primary antibodies used for western blot analysis were and killed 15, 30, or 60 min thereafter. We used ob/ob polyclonal antibody-recognizing UCP2 (C-20, sc6525, mice (nZ5) and C57BL/6 mice (nZ5) to investigate Santa Cruz Biotechnology, Santa Cruz, CA, USA), diluted the expression of UCP2 in the gut, and plasma levels of 1:700 and a polyclonal anti-cytochrome c antibody (6H2, GLP1, in the state of pathological high blood glucose. sc13561, Santa Cruz Biotechnology), diluted 1:1000. Ob/ob mice and C57BL/6 mice were fasted for 2 h, and Mitochondrial protein from the NIC-H716 cells was isolated using 1 ml extraction buffer (250 mM sucrose, basal plasma glucose and GLP1 levels in ileum tissue . were then measured. We also administered genipin to 1 mM EDTA, and 10 mM Tris, pH 7 4) supplemented ob/ob mice via oral gavage, collected samples 1 h after with protease inhibitors (1 mmol/l phenylmethyl- ! genipin administration, and detected GLP1 levels in sulfonyl fluoride, 2 mmol/l leupeptin, and 1 aprotinin). ileum tissue and glucose in the blood. The mixture was centrifuged at 800 g for 10 min. The resulting supernatant was centrifuged at 10 000 g for 10 min and the mitochondrial pellet resuspended Tissue samples in 25 ml TES buffer. Mitochondrial protein concen- Samples from ileum tissue were frozen in liquid tration was determined colorimetrically using a BCA nitrogen and kept for RNA preparation, fixed in Protein Assay (Pierce, Rockford, IL, USA). Mito- 4% paraformaldehyde for immunohistochemistry, or chondrial (15 mg) were mixed with 3! sample immediately taken for preparation of fresh mito- buffer (0.5 M phosphate buffer, pH 7.0, 30% (w/v) chondrial protein. Dipeptidyl peptidase IV inhibitor glycerol, 7.5% (w/v) SDS, and 0.75 mM bromophenol (Linco Research, St Charles, MO, USA) was added to blue), boiled for 5 min, and separated by electro- the blood samples collected for the detection of GLP1. phoresis on a 12.5% SDS–PAGE gel. After separation,

Journal of Molecular Endocrinology (2012) 48, 151–158 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access UCP2 negatively regulates GLP1 secretion . H ZHANG and others 153 proteins were transferred to Immobilon PVDF mem- FITC (green) goat anti-rabbit (1:200) was added for branes (Millipore Corporation, Bedford, MA, USA). 30 min at RT in the dark. Hoechast was used as a nuclear UCP2 proteins were detected with a polyclonal counterstain for 45 min. Following a final wash with PBS antibody-recognizing UCP2 (Santa Cruz Bio- and addition of Slow Fate equilibration buffer (Molecu- technology) at a dilution of 1:700 followed by lar Probes cat #S2828), slides were mounted with 10 ml incubation with HRP-conjugated monkey anti-goat Slow Fade (Molecular Probes cat #S2828). IgG secondary antibody (Santa Cruz) at a dilution of 1:2000 and detected with enhanced chemi- Culture of NCI-H716 cells and secretion studies luminescence (ECL detection system; NEN, Boston, MA, USA). To validate equal protein loading among Human enteroendocrine NCI-H716 cells were main- various lanes, PVDF membranes were stripped and tained in suspension culture as described by the reprobed with a polyclonal anti-cytochrome c antibody American Type Culture Collection (Manassas, VA, (Santa Cruz Biotechnology) at a dilution of 1:1000. USA). Two days before the experiments, cells were seeded into 24-well culture plates precoated with Matrigel Localization of UCP2 and GLP1 in mouse ileum tissue as described previously (Reimer et al.2001). On the day of and NCI-H716 cells the experiments, supernatants were replaced with PBS containing 1 mM CaCl2 and dipeptidyl peptidase IV Localization of UCP2 and GLP1 in mouse ileum inhibitor. The solutions were adjusted to pH 7.2. Cells were incubated at 37 8C for 1 h either without glucose or The antibodies used for tissue immunostaining were with a range of glucose concentrations and RNA polyclonal rabbit anti-GLP1 (bs-0038R; Beijing Bio- interference targeting UCP2 (final concentration synthesis Biotechnology Co., Beijing, China) and poly- 300 nM, as described in the next section). GLP1 was clonal goat anti-UCP2 antibody (C-20, sc6525, Santa measured by ELISA and normalized to protein content. Cruz Biotechnology). For double-labeling experiments involving a combination of GLP1 and UCP2 antibodies, tissue sections were incubated with the primary siRNA preparation and NCI-H716 cell transfection antibodies (anti-UCP2, 1:50) at 4 8C overnight and The siRNA sequences targeting UCP2 and a nonspecific then washed with PBS. For the secondary antibody, siRNA control were purchased from Shanghai Shine- FITC (green) goat anti-rabbit (1:200) was added for gene Molecular Biotechnology. Subconfluent differen- 30 min at RT in the dark and then washed with PBS. tiated NCI-H716 cells were transiently transfected with Slides were fixed in 4% paraformaldehyde in PBS for siRNAs using Lipofectamine 2000 according to the 15 min and then washed with PBS. Anti-GLP1 (1:100) manufacturer’s protocol (Life Technologies). The was added and incubated at 4 8C overnight, followed by entire mixture was then added to the cells in one dish, a further wash with PBS, application of the secondary resulting in a final siRNA concentration of 300 nM. Cells antibody (Cy3 (Red) donkey anti-goat (1:200), for were usually examined 48 h after transfection. 30 min at RT in the dark), and a final wash with PBS. Slow Fade equilibration buffer (Molecular Probes cat #S2828) was applied, and the slides were mounted with GLP1 measurement 10 ml Slow Fate (Molecular Probes cat #S2828), covered, and sections viewed with a fluorescence microscope. The level of serum GLP1 (7–36) amide was measured using an ELISA kit (Linco Research). This assay relies on a monoclonal antibody fixed in a coated micro-well plate Localization of UCP2 and GLP1 in NCI-H716 cells binding to the N-terminal region of active GLP1. The Antibodies used for immunofluorescence staining in concentration of active GLP1 is proportional to the fluore- NCI-H716 cells were anti-GLP1 (C-17; Santa Cruz) and scence generated by umbelliferone, which is produced by anti-UCP2 (LS-B1911; LifeSpan BioSciences). Cells were alkaline phosphatase-catalyzed hydrolysis of methyl grown on coverslips, fixed in 4% paraformaldehyde in umbelliferyl phosphate (conjugated with anti-GLP1 PBS for 10 min, washed with PBS, and cooled with 100% monoclonal antibodies). Samples (100 mlofeachsample) methanol at K20 8C for 20 min. Thereafter, cells were were added to each assay well. ELISA has a working range washed with PBS and permeabilized with 0.1% Triton of 2–100 pM. Extraction of GLP1 (7–36) amide from ileum X-100 for 10 min. After blocking with Dako blocking was carried out with ethanol/acid (5:1 v/v) solution solution, primary antibody (anti-GLP1, 1:100) was (5 ml/g tissue). Samples were homogenized at added and incubated at 4 8C overnight. For the 24 000 r.p.m. and kept for 24 h at 4 8C. Homogenates secondary antibody, Cy3 (Red) donkey anti-goat were centrifuged (2000 g) and the supernatant was (1:200) was applied for 30 min at RT in the dark decanted and diluted in saline. Concentrations of GLP1 followed by a PBS wash. Anti-UCP2 (1:50) was added and (7–36) amide in tissue extracts were measured using an incubated at 4 8C overnight. For the secondary antibody, ELISA method (GLP1 ELISA kit; Linco Research). www.endocrinology-journals.org Journal of Molecular Endocrinology (2012) 48, 151–158

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access 154 H ZHANG and others . UCP2 negatively regulates GLP1 secretion

A UCP2 GLP1 Merge localization of UCP2 and GLP1 as studied in slides of NCI-H716 cells. Immunofluorescence-positive staining for GLP1 was observed in the cytoplasm of NCI-H716 cells (Fig. 1B, green staining). Immunofluorescence- positive staining for UCP2 was also observed in the cytoplasm of human NCI-H716 cells (Fig. 1B, red staining). Both GLP1 and UCP2 granules were mainly expressed in the cytoplasm. This result suggests that L cells secreting GLP1 can also express UCP2. B UCP2 UCP2 UCP2-deficient mice have higher plasma GLP1 levels after glucose loading As shown in Fig. 2, the initial blood glucose level (basal plasma level of glucose) in wild-type littermates was slightly higher compared with UCP2-deficient mice, but there were no significant differences (see left panel, UCP2-deficient mice vs wild-type littermates: 5.3G0.2vs . G . O . C GLP1 UCP2 6 4 0 4, P 0 05). There was no significant difference in basal plasma levels of GLP1 between UCP2-deficient mice and wild-type littermates (see right panel, UCP2- deficient mice vs wild-type littermates: 3.7G0.1vs3.5 G0.2, PZ0.063). Following administration of glucose (2.0 g/kg) via oral gavage, blood glucose increased in both UCP2-deficient mice and wild-type littermates and peaked at 15 min after administration. The blood glucose levels in wild-type littermates were much higher Figure 1 Localization of UCP2 and GLP1 in mouse ileum and NCI-H716 cells. (A) Immunofluorescence staining using than those in UCP2-deficient mice after glucose anti-GLP1 (red) and anti-UCP2 (green) antibodies demonstrated loading (Fig. 2, left panel). However, plasma GLP1 localization of UCP2 and GLP1 in certain cells in mouse ileum levels in UCP2-deficient mice were significantly higher ! (400 magnification). (B) Immunofluorescence staining using than those in wild-type littermates at 15, 30, and 60 min anti-UCP2 (green) antibody demonstrated localization of UCP2 in . G . . certain cells in UCP2-deficient mouse and wild-type mouse ileum after oral glucose loading (15 min: 9 4 1 0vs69 (400! magnification). (C) Immunofluorescence staining with G0.3 pM, P!0.001; 30 min: 7.9G0.3vs5.6G0.4 pM, anti-GLP1 and anti-UCP2 antibodies in NCI-H716 cells, an P!0.001; 60 min: 4.9G0.1vs3.38G0.10 pM, P!0.01; L-cell model. This showed anti-UCP2 antibody-positive Fig. 2, right panel). staining (green) and anti-GLP1 antibody-positive staining (red) (400! magnification).

Statistical analyses 25 12 Data are expressed as meanGS.E.M. Statistical signi- *** 20 ** 10 ficance was calculated by one-way ANOVA or unpaired *** 8 *** two-tailed t-test. P values !0.05 were regarded as 15 6 significant differences. 10 ** ** 4 Wild-type Glucose (mmol/l) GLP1 level (pM) GLP1 level Wild-type 5 2 UCP2–/– UCP2–/– 0 0 Results 0 15 30 60 0 15 30 60 Time (min) Time (min) Localization of UCP2 and GLP1 in mouse ileum tissue Figure 2 UCP2-deficient mice have higher serum GLP1 and lower and NCI-H716 cells plasma glucose levels. Blood glucose (left panel) and plasma GLP1 (right panel) levels were assessed at time 0 and 15, 30, and . As shown in Fig. 1A, immunofluorescence staining 60 min after an oral glucose load (2 0 g/kg). UCP2-deficient mice (total number: 20, each time point nZ5) and wild-type mice (total using anti-GLP1 and anti-UCP2 antibodies demon- number: 20, each time point nZ5) **P!0.01, ***P!0.001. Full strated that UCP2 was colocalized with GLP1 in certain colour version of this figure available via http://dx.doi.org/10.1530/ cells in the mouse ileum. Figure 1B shows the JME-11-0114.

Journal of Molecular Endocrinology (2012) 48, 151–158 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access UCP2 negatively regulates GLP1 secretion . H ZHANG and others 155

A 350 350 300 * 300 * * * 250 250 * * * * 200 200 150 150 100 100 50 50 Relative GLP1 level (%) GLP1 level Relative Relative GLP1 level (%) GLP1 level Relative 0 0 Control5 10203060 Control 5 15 30 60 120 Glucose (mM) Time (min) BC

160 1·0 * UCP2 34 kDa 0·8 120 0·6 Cytochrome c 15 kDa 0·4 80

level of UCP2 level ** Relative mRNA Relative 0·2 40 0 si-UCP2 Untransfected (%) GLP1 level Relative 0 Mock transfected si-UCP2 si-UCP2 Untransfected Mock transfected Untransfected Mock transfected Glucose 30 mM +++ Figure 3 Downregulated UCP2 increased secretion of GLP1 in L cells after glucose stimulation. (A) Time course experiments revealed release of GLP1 from NCI-H716 cells into the supernatant: NCI-H716 cells were incubated with 30 mM glucose and the supernatants were collected at 0 (control), 5, 15, 30, 60, and 120 min to measure GLP1 levels. In the dose- dependent experiment, NCI-H716 cells were incubated with 0 (control), 5, 10, 20, 30, and 60 mM glucose for 1 h, and the levels of GLP1 in the supernatants were determined. *P!0.05 vs control. (B) The knockdown efficiency of si-UCP2 in NCI-H716 cells. NCI-H716 cells were left untransfected (blank), mock transfected (no siRNA), or transfected with si-UCP2. After 48 h, the mRNA levels of Ucp2 were determined by quantitative qRT-PCR with normalization to b-actin, and the expressions of UCP2 protein were detected by western blotting. Cytochrome c was used as a loading control. **P!0.01 vs untransfected (blank). (C) NCI-H716 cells treated with si-UCP2 increased glucose-mediated GLP1 secretion. NCI-H716 cells were incubated with growth medium only as a normal control (blank), or 300 nM si-UCP2 for 48 h plus 30 mM glucose for 1 h. GLP1 concentrations in the supernatant after 1 h culture were determined using a GLP1 assay kit. *P!0.05 vs untransfected (blank). Full colour version of this figure available via http://dx.doi.org/10.1530/JME-11-0114.

Downregulated UCP2 increased secretion of GLP1 levels of GLP1 in ileum tissue, we used ob/ob mice in L cells after glucose stimulation (nZ5) and C57BL/6 mice (nZ5) to investigate the expression of UCP2. As shown in Fig. 4, Ucp2 mRNA Time course experiments revealed that release of GLP1 levels in ileum tissue from ob/ob mice were increased from NCI-H716 cells into medium reached a peak at (Fig. 4A), whereas GLP1 ileum tissue levels were 30 min, and a stable plateau at 60 min, after glucose significantly lower compared with those in control loading. Furthermore, stimulation of NCI-H716 cells C57BL/6 mice (Fig. 4B). with %30 mM glucose led to concentration-dependent release of GLP1 into the medium, and a stable plateau was reached at higher concentrations (Fig. 3A). RNA Acute inhibition of UCP2 by genipin can improve interference targeting UCP2 decreased the expression GLP1 secretion in ob/ob mice of UCP2 in NCI-H716 cells (Fig. 3B) accompanied by In ob/ob mice following administration of genipin (an increased glucose-mediated GLP1 secretion (Fig. 3C). aglycone known to block UCP2 function) by oral gavage, GLP1 in ileum tissue increased compared with Ucp2 mRNA in ileum tissue increased and GLP1 animals administrated saline (Fig. 5B). However, there secretion inhibited in ob/ob mice was no effect on blood glucose in ob/ob mice as a result of genipin treatment (Fig. 5A), suggesting that acute To evaluate the effect of pathological high blood inhibition of UCP2 by genipin can improve GLP1 glucose on the expression of UCP2 in the gut and secretion in ob/ob mice. www.endocrinology-journals.org Journal of Molecular Endocrinology (2012) 48, 151–158

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access 156 H ZHANG and others . UCP2 negatively regulates GLP1 secretion

A 2·5 producing more insulin or releasing more GLP1 after 12 *** 2·0 * 10 glucose loading. 1·5 8 The finding that there are many more L cells in the 6

mRNA 1·0 distal (ileum and colon) vs proximal (duodenum and 4

Relative Ucp2 Relative 0·5 2 proximal jejunum) gut led to the suggestion that the Glucose (mmol\l) 0 0 early rapid rise in plasma GLP1 after glucose load was WT ob/ob WT ob/ob not due to direct stimulation of L cells. ‘Proximal- B to-distal’ models proposed that indirect neurally 120 mediated signaling, initiated by glucose-sensing 100 K cells or other non-L cells of the proximal gut, leads 80 ** to release of GLP1 from L cells of the distal gut 60 level (%) level (Knudsen et al. 1975, Holst 2007). However, there is 40 Relative GLP1 Relative 20 considerable evidence that luminal glucose directly 0 leads to GLP1 secretion from the proximal gut. First, WT ob/ob although in humans there are more L cells in the gut Figure 4 Ucp2 mRNA is upregulated and GLP1 secretion distally than proximally, there are still many L cells in inhibited in ob/ob mice. (A) Left panel: the levels of Ucp2 mRNA the duodenum and jejunum (Theodorakis et al. 2006). in ileum of ob/ob mice were determined by quantitative RT-PCR. Secondly, the time course (onset, peak, and duration) Ucp2 mRNA levels in ileum increased in ob/ob mice compared with C57BL/6 mice, *P!0.05. Right panel: blood glucose was of glucose-elicited release of GLP1 in humans is assessed with a high level of blood glucose in ob/ob mice consistent with that of glucose reaching the proximal compared with C57BL/6 mice, ***P!0.001. (B) The GLP1 levels intestine (Schirra et al.1996, Balks et al.1997, in ileum tissue in ob/ob mice were determined using a GLP1 Theodorakis et al. 2004, Kim et al. 2005). Thirdly, ELISA kit; the levels of GLP1 in ileum tissue from ob/ob mice were lower than those in C57BL/6 mice. Set the level of GLP1 in applying small amounts of glucose solution by catheter C57BL/6 mice (control) as relative GLP1 level Z100%. into the duodenum in humans leads to increased **P!0.01, the level of GLP1 in ob/ob vs C57BL/6 mice. Ob/ob plasma GLP1 levels (Vilsboll & Holst 2004). Finally, mice (nZ5) and C57BL/6 mice (nZ5). many animal studies have shown that glucose directly stimulates GLP1 secretion, for example, in an isolated perfused porcine ileum experiment (Hansen 2004). Discussion To investigate the potential pathophysiological rel- evance of UCP2 to GLP1, we conducted experiments The ‘incretin effect’ of GLP1 has been known for many with ob/ob mice. Ob/ob mice have blood glucose levels in years, and its role in the overall regulation of insulin the diabetic range, indicating a defect in their b-cell release in vivo is well established. However, regulation of response, as well as significantly increased levels of GLP1 secretion itself is not well understood. The Ucp2 mRNA and protein in their pancreatic islets. presence of a UCP2 response to glucose in enteroendo- UCP2 plays an important role in the pathogenesis of crine L cells may identify a new mechanism by which type 2 diabetes by inhibiting b-cell insulin secretion intestinal hormone secretion is regulated. Our data

indicate that glucose-induced secretion from L cells is AB partly mediated via UCP2. 250 Compared with wild-type littermates, UCP2-deficient 12 200 * mice had higher plasma levels of GLP1 after adminis- 10 150 tration of glucose; however, there was no significant 8 6 100 difference in the basal level of GLP1 (UCP2-deficient 4 50 . G . . G . 2

mice vs wild-type littermates: 3 7 0 1vs35 0 2, Glucose (mmol/l) 0 (%) GLP1 level Relative 0 PZ0.063). This result indicates that UCP2 had no ob/ob ob/ob+genipin ob/ob ob/ob+genipin effect on baseline GLP1 secretion but impacted on Figure 5 Effect of genipin on glucose and GLP1 homeostasis in postprandial secretion of GLP1. In vitro, decreased ob/ob mice. (A) Blood glucose values were assessed in ob/ob UCP2 expression in NCI-H716 cells by siRNA mice with or without genipin administration. The blood glucose interference resulted in increased glucose-mediated levels were detected after administration of genipin (100 mg/kg). There was no significant difference between the ob/ob mice with GLP1 secretion. These results indicate that UCP2 is genipin and without genipin (PO0.07). (B) The GLP1 levels in intimately involved in glucose-stimulated secretion of ileum tissues were assessed in ob/ob mice after administration of GLP1 from L cells. genipin. Following 2 h fasting, mice were treated with genipin In our study, we found that blood glucose levels in (100 mg/kg). Ob/ob mice treated with saline served as controls. Mice were killed 1 h after administration of genipin, and ileum wild-type littermates were much higher than those in tissue samples were collected to detect the level of GLP1. UCP2-deficient mice after glucose loading (Fig. 2, left Set the ob/ob mice treated with saline (control) as relative GLP1 panel). This may be due to UCP2-deficient mice level Z100%. Both ob/ob mice and C57BL/6 mice, nZ5*P!0.05.

Journal of Molecular Endocrinology (2012) 48, 151–158 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access UCP2 negatively regulates GLP1 secretion . H ZHANG and others 157

(Lameloise et al. 2001, Zhang et al. 2001), but it is not draft of the manuscript. C-Y Z designed the study. K Z participated in clear whether UCP2 levels are also increased in L cells its design and coordination and helped to draft the manuscript. All in the small intestine and colon in obesity-induced type authors read and approved the final manuscript. 2 diabetes. We compared Ucp2 mRNA levels in ileum tissue from ob/ob mice and C57BL/6 mice and found that Ucp2 mRNA levels in ileum from ob/ob mice were increased in the fasting state, whereas GLP1 levels in References ileum tissues from ob/ob mice were decreased. This Balks HJ, Holst JJ, von zur Muhlen A & Brabant G 1997 Rapid result suggests that chronic high glucose induces high oscillations in plasma glucagon-like peptide-1 (GLP-1) in humans: levels of Ucp2 mRNA in ob/ob mouse intestine and cholinergic control of GLP-1 secretion via muscarinic receptors. impaired the GLP1 secretion. Journal of Clinical Endocrinology and Metabolism 82 786–790. Genipin can rapidly inhibit UCP2-mediated proton (doi:10.1210/jc.82.3.786) leak. In pancreatic islet b-cells, genipin increases Dhanvantari S & Brubaker PL 1998 Proglucagon processing in an islet cell line: effects of PC1 overexpression and PC2 depletion. mitochondrial membrane potential and ATP levels Endocrinology 139 1630–1637. (doi:10.1210/en.139.4.1630) while closing KATP channels in a UCP2-dependent Donley VR, Hiskett EK, Kidder AC & Schermerhorn T 2005 manner, stimulating insulin secretion. Importantly, ATP-sensitive potassium channel (KATP channel) expression in acute addition of genipin to isolated islets reverses the normal canine pancreas and in canine insulinomas. high glucose- and obesity-induced b-cell dysfunction BMC Veterinary Research 1 8. (doi:10.1186/1746-6148-1-8) Elrick H, Stimler L, Hlad CJ Jr & Arai Y 1964 Plasma insulin response (Zhang et al. 2006). Our results indicate that to oral and intravenous glucose administration. Journal of Clinical acute inhibition of UCP2 by genipin can improve Endocrinology and Metabolism 24 1076–1082. (doi:10.1210/jcem-24- GLP1 secretion in ob/ob mice, suggesting that 10-1076) UCP2 negatively regulates GLP1 secretion in chronic Fehmann HC, Goke R & Goke B 1995 Cell and molecular-biology of high-glucose state. the incretin hormones glucagon-like peptide-1 and glucose- dependent insulin releasing polypeptide. Endocrine Reviews 16 It is well known that tissue-specific post-translational 390–410. (doi:10.1210/edrv-16-3-390) processing of proG results in the production of a Goke R, Fehmann HC & Goke B 1991 Glucagon-like peptide-1 (7–36) diversity of peptides in the intestine and pancreas. amide is a new incretin/enterogastrone candidate. European Glicentin, oxyntomodulin, GLP1, and GLP2 are the Journal of Clinical Investigation 21 135–144. (doi:10.1111/j.1365- 2362.1991.tb01802.x) proG-derived peptides produced in the L cells of the Gribble FM, Williams L, Simpson AK & Reimann F 2003 A novel intestine, whereas glucagon is the major product of glucose-sensing mechanism contributing to glucagon-like peptide-1 processing in the A cells of the pancreas (Dhanvantari secretion from the GLUTag cell line. Diabetes 52 1147–1154. (doi:10. & Brubaker 1998). Whether UCP2 affects the proG- 2337/diabetes.52.5.1147) derived peptides produced by influencing post- Hansen L 2004 Glucagon-like peptide-1 secretion is influenced by perfusate glucose concentration and by a feedback mechanism translational processing of proG in the L cell of the involving somatostatin in isolated perfused porcine ileum. intestine is unclear and will be explored further in Regulatory Peptides 118 11–18. (doi:10.1016/j.regpep.2003.10.021) future studies. Herrmann C, Go¨ke R, Richter G, Fehmann HC, Arnold R & Go¨oke B In summary, our in vivo and in vitro results suggest 1995 Glucagon-like peptide-1 and glucose dependent insulin- that UCP2 inhibits GLP1 secretion. Our findings releasing polypeptide plasma levels in response to nutrients. Digestion 56 117–126. (doi:10.1159/000201231) provide evidence that UCP2 may serve as a negative Hirasawa A, Tsumaya K, Awaji T, Katsuma S, Adachi T, Yamada M, regulator of GLP1 secretion in the gastrointestinal tract. Sugimoto Y, Miyazaki S & Tsujimoto G 2005 Free fatty acids regulate gut incretin glucagon-like peptide-1 secretion through GPR120. Nature Medicine 11 90–94. (doi:10.1038/nm1168) Declaration of interest Holst JJ 2007 The physiology of glucagon-like peptide 1. Physiological Reviews 87 1409–1439. (doi:10.1152/physrev.00034.2006) The authors declare that there is no conflict of interest that could be Jezek P 2002 Possible physiological roles of mitochondrial uncoupling perceived as prejudicing the impartiality of the research reported. proteins – UCPn. International Journal of Biochemistry & Cell Biology 34 1190–1206. (doi:10.1016/S1357-2725(02)00061-4) Kim BJ, Carlson OD, Jang HJ, Elahi D, Berry C & Egan JM 2005 Peptide Funding YY is secreted after oral glucose administration in a gender-specific manner. Journal of Clinical Endocrinology and Metabolism 90 6665–6671. (doi:10.1210/jc.2005-0409) This work was supported by grants from the National Natural Science Knudsen JB, Holst JJ, Asnaes S & Johansen A 1975 Identification of Foundation of China (nos 30225037, 30471991, and 30771039). cells with pancreatic-type and gut-type glucagon immunoreactivity in the human colon. Acta Pathologica et Microbiologica Scandinavica 83 741–743. (doi:10.1111/j.1699-0463.1975.tb01407.x) Author contribution statement Lameloise N, Muzzin P, Prentki M & Assimacopoulos-Jeannet F 2001 Uncoupling protein 2: a possible link between fatty acid excess and H Z, J L, X L, and Y L carried out the field collection and/or impaired glucose-stimulated insulin secretion? Diabetes 50 803–809. preparation of the samples. H Z and J L performed cell culture and (doi:10.2337/diabetes.50.4.803) RNAi. X L performed animal experiments. Y L participated in the Margolskee RF, Dyer J, Kokrashvili Z, Salmon KS, Ilegems E, Daly K, statistical analysis and interpretation of the data. H Z prepared the first Maillet EL, Ninomiya Y, Mosinger B & Shirazi-Beechey SP 2007 www.endocrinology-journals.org Journal of Molecular Endocrinology (2012) 48, 151–158

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access 158 H ZHANG and others . UCP2 negatively regulates GLP1 secretion

T1R3 and gustducin in gut sense sugars to regulate expression of Rorsman P 1997 The pancreatic b-cell as a fuel sensor: an C Na -glucose cotransporter 1. PNAS 104 15075–15080. (doi:10. electrophysiologist’s viewpoint. Diabetologia 40 487–495. (doi:10. 1073/pnas.0706678104) 1007/s001250050706) Mattiasson G & Sullivan PG 2006 The emerging functions of UCP2 in Rousset S, Mozo J, Dujardin G, Emre Y, Masscheleyn S, Ricquier D & health, disease, and therapeutics. Antioxidants & Redox Signaling Cassard-Doulcier AM 2007 UCP2 is a mitochondrial transporter 8 1–38. (doi:10.1089/ars.2006.8.1) with an unusual very short half-life. FEBS Letters 581 479–482. McIntyre N, Holdsworth CD & Turner DS 1964 New interpretation of (doi:10.1016/j.febslet.2007.01.010) oral glucose tolerance. Lancet 41 20–21. (doi:10.1016/S0140- Rozengurt E & Sternini C 2007 Taste receptor signaling in the 6736(64)90011-X) mammalian gut. Current Opinion in Pharmacology 7 557–562. (doi:10. McIntyre N, Holdsworth CD & Turner DS 1965 Intestinal factors in the 1016/j.coph.2007.10.002) control of insulin secretion. Journal of Clinical Endocrinology and Schirra J, Katschinski M, Weidmann C, Schafer T, Wank U, Arnold R & Metabolism 25 1317–1324. (doi:10.1210/jcem-25-10-1317) Goke B 1996 Gastric emptying and release of incretin hormones Meier JJ & Nauck MA 2005 Glucagon-like peptide 1 (GLP-1) in biology after glucose ingestion in humans. Journal of Clinical Investigation and pathology. Diabetes Metabolism Research and Reviews 21 91–117. 97 92–103. (doi:10.1172/JCI118411) (doi:10.1002/dmrr.538) Theodorakis MJ, Carlson O, Muller DC & Egan JM 2004 Elevated Nedergaard J & Cannon B 2003 The ‘novel’‘uncoupling’ proteins plasma glucose-dependent insulinotropic polypeptide associates UCP2 and UCP3: what do they really do? Pros and cons for with hyperinsulinemia in impaired glucose tolerance. Diabetes Care suggested functions Experimental Physiology 88 65–84. (doi:10.1113/ 27 1692–1698. (doi:10.2337/diacare.27.7.1692) eph8802502) Theodorakis MJ, Carlson O, Michopoulos S, Doyle ME, Juhaszova M, Nielsen LB, Ploug KB, Swift P, Ørskov C, Jansen-Olesen I, Chiarelli F, Petraki K & Egan JM 2006 Human duodenal enteroendocrine cells: Holst JJ, Hougaard P, Po¨rksen S, Holl R et al. 2007 Co-localisation of source of both incretin peptides, GLP-1 and GIP. American Journal of the Kir6.2/SUR1 channel complex with glucagon-like peptide-1 Physiology. Endocrinology and Metabolism 290 E550–E559. (doi:10. and glucose-dependent insulinotrophic polypeptide expression in 1152/ajpendo.00326.2004) human ileal cells and implications for glycaemic control in new Thorens B 1995 Glucagon-like peptide-1 and control of insulin- onset type 1 diabetes. European Journal of Endocrinology 156 663–671. secretion. Diabe`te &Me´tabolisme 21 311–318. (doi:10.1530/EJE-06-0756) Vilsboll T & Holst JJ 2004 Incretins, insulin secretion and type 2 Perley MJ & Kipnis DM 1967 Plasma insulin responses to oral and diabetes mellitus. Diabetologia 47 357–366. (doi:10.1007/s00125-004- intravenous glucose. Journal of Clinical Investigation 46 1954–1962. 1342-6) (doi:10.1172/JCI105685) Zhang CY, Baffy G, Perret P, Krauss S, Peroni O, Grujic D, Hagen T, Reimann F, Ward PS & Gribble FM 2006 Signalling mechanisms Vidal-Puig AJ, Boss O, Kim YB et al. 2001 Uncoupling protein-2 underlying the release of glucagon-like peptide-1. Diabetes 55 negatively regulates insulin secretion and is a major link between (Suppl 2) S78–S85. (doi:10.2337/db06-S010) obesity, beta cell dysfunction, and type 2 diabetes. Cell 105 745–755. Reimer RA, Darimont C, Gremlich S, Nicolas-Metral V, Ruegg UT & (doi:10.1016/S0092-8674(01)00378-6) Mace K 2001 A human cellular model for studying the regulation of Zhang CY, Parton LE, Ye CP, Krauss S, Shen R, Lin CT, Porco JA Jr & glucagon-like peptide-1 secretion. Endocrinology 142 4522–4528. Lowell BB 2006 Genipin inhibits UCP2-mediated proton leak and (doi:10.1210/en.142.10.4522) acutely reverses obesity- and high glucose-induced b cell dysfunc- Ricqier C & Bouillaud F 2000 The uncoupling protein homologues: tion in isolated pancreatic islets. Cell Metabolism 3 417–427. (doi:10. UCP1, UCP2, UCP3, StUCP and AtUCP. Biochemical Journal 345 1016/j.cmet.2006.04.010) 161–179. (doi:10.1042/0264-6021:3450161) Ritzel U, Fromme A, Ottleben M, Leonhardt U & Ramadori G 1997 Release of glucagon-like peptide-1 (GLP-1) by carbohydrates in the Received in final form 3 January 2012 perfused rat ileum. Acta Diabetologia 34 18–21. (doi:10.1007/ Accepted 18 January 2012 s005920050059) Made available online as an Accepted Preprint 18 January 2012

Journal of Molecular Endocrinology (2012) 48, 151–158 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 10/02/2021 02:01:12PM via free access