Duodenal transposition in a rodent model

Sixteen adult male Wistar rats were randomized to either sham operation or duodenal transposition. Rats were anaesthetized with a mixture of Hypnorm® (0.35 mL) and diazepam (0.35 mL), which was injected i.m. before the procedure. Duodenal transposition involved excision of the segment of duodenum containing the hepatopancreatic ampulla, which allows drainage of both bile and pancreatic juices into the gut, followed by the anastomosis of the transected segment to the ileum, 10 cm proximal to the caecum. The sham operation consisted of transection of the proximal and distal parts of the duodenum followed by re-anastomosis in order for the rats to have the same surgical insult and time under anaesthesia. Rats were housed individually and received a normal diet of chow and water ad libitum. Animals were killed 4 weeks post-operatively. Mixed arterial and venous blood was collected into lithium-heparin tubes containing 5000 kallikrein inhibitor units aprotinin, immediately spun down at 1600×g and stored at −70°C until analysis.

Primary murine intestinal cultures

The 2- to 6-month-old C57BL6 mice were killed by cervical dislocation and the gut collected into ice-cold Leibovitz-15 (L-15) medium. The intestine was opened longitudinally, rinsed in PBS, and chopped into 1–2 mm pieces. Upper small intestinal cultures contained tissue from the top 10 cm of the gut distal to the stomach and colon cultures consisted of tissue distal to the ileocolic junction. Tissue was digested with 0.4 mg·mL⁻¹ Collagenase XI, centrifuged at 300×g, and resuspended in Dulbecco's modified Eagle's medium (DMEM; 25 mM glucose) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 U·mL⁻¹ penicillin and 0.1 mg·mL⁻¹ streptomycin. Aliquots of tissue were plated and incubated at 37°C, 5% CO₂.

Cell line culture

GLUTag cells were cultured in DMEM (5.6 mM glucose) supplemented with 10% (v/v) FBS, 10 µg·mL⁻¹ penicillin/streptomycin and 2 mM glutamine. They were incubated at 37°C, 5% CO₂ and the culture medium was changed every 3 days. When reaching ~70% confluence (~every 5 days), cells were trypsinized with 1X trypsin-EDTA, diluted and reseeded.

siRNA transfection

GLUTag cells plated on 6 cm Matrigel-coated plastic plates were transfected with either scrambled siRNA (control) or siRNA targeting murine GPBA (Qiagen, Crawley, UK) for 24 h. Transfection was performed using OptiMEM (Invitrogen, Paisley, UK) media, 12 µL Lipofectamine 2000 (Invitrogen) and 50 nM siRNA as instructed by the manufacturer. GLUTag cells were re-seeded on 24-well plates and allowed to recover for 24 h before being used in secretion experiments.

GLP-1 and GLP-2 secretion assay

Secretion studies on GLUTag and primary intestinal cultures were performed 24–36 h after plating on 24-well culture plates coated with Matrigel (BD Bioscience, Oxford, UK). Cultures were washed thoroughly and incubated with test reagents in saline solution (see below) containing 0.1% fatty acid-free BSA for 2 h at 37°C. Media was then collected and centrifuged to remove contaminating cells. In primary cell secretion experiments, the plated cells were treated with lysis buffer containing: 50 mM Tris-HCl, 150 mM NaCl, 1% IGEPAL-CA 630, 0.5% deoxycholic acid (DCA) and one tablet of complete EDTA-free protease inhibitor cocktail (Roche, Burgess Hill, UK). Plates were freeze-thawed and mechanically disrupted to extract intracellular peptides.

GLP-1 was assayed in the supernatant collected from GLUTag cells and in both the supernatant and cell extracts from primary intestinal cultures using an elisa specific for GLP-1 (7–36) amide and GLP-1 (7–37) (GLP-1 active ELISA-kit, Millipore, Billerica, MA, USA).

Total GLP-2 was measured in supernatant collected from GLUTag experiments and in rat plasma by radioimmunoassay using antiserum FT-17 which cross-reacts 100% with human and rat GLP-2, less than 0.1% with GLP-1 and not with any other known gastrointestinal or pancreatic hormone. The assay measures GLP-2 (1–33) and its degradation product GLP-2 (3–33) and major proglucagon fragment (Feltrin et al., 2006).

For the primary cell experiments, GLP-1 secretion in each well was expressed as a fraction of the total of that hormone measured in the same well, and was normalized to the basal secretion measured in parallel on the same day. For GLUTag cell experiments, secretion was normalized to the basal measured in parallel on the same day.

RNA extraction and qRT-PCR

Total RNA from FACS-sorted cells and GLUTag cultures was isolated using a micro scale RNA isolation kit (Ambion, Warrington, UK) and Tri-Reagent, respectively. A total of 100–150 ng of RNA was reverse-transcribed according to standard protocols using a Peltier Thermal Cycler-200 (MJ Research, Waltham, MA, USA). Quantitative RT-PCR was performed with 7900 HT Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). PCR reactions mix consisted of first-strand cDNA template, primers (TaqMan gene expression assays, Applied Biosystems) and PCR Master mix (Applied Biosystems). In all cases, expression was compared with that of β-actin measured on the same sample, in parallel, on the same plate, giving a cycle threshold difference (ΔCT) for β-actin minus the test gene. Mean, SEM and statistics were obtained for the ΔCT data, and only converted to relative expression levels (2^ΔCT) for presentation.

Ca²⁺ measurements

Cells were plated in Matrigel-coated glass bottom dishes (MatTek, Ashland, MA, USA) 1–3 days prior to use and loaded with fura-2 by incubation in 2 µM of the acetoxymethyl ester (Molecular Probes, Leiden, the Netherlands) for 30 min in saline solution containing 1 mM glucose at room temperature. Cells were then washed, and dishes mounted on an inverted fluorescence microscope (Olympus IX71, Southall, UK) with a ×40 oil immersion objective. Excitation at 340 and 380 nm was achieved using a 75 W xenon arc lamp with a monochromator (Cairn Research, Faversham, UK) controlled by MetaFluor software (Universal Imaging; Cairn Research) and emission was recorded with a charged-coupled device camera (Orca ER, Hammamatsu; Cairn Research). Background-subtracted fluorescence was normalized to a baseline average measured before application of the first test reagent and expressed as a 340/380 nm ratio, and the response to test reagents was defined as the maximum concentration (averaged over 20 s) achieved during their application.

cAMP FRET measurements

Single cell measurements of cAMP levels were made using the FRET-based sensor, Epac2-camps kindly donated by M. Lohse (Nikolaev et al., 2004). GLUTag cells were seeded on 6 cm Matrigel-coated plastic dishes and when 80–90% confluent transfected with 3 µg of Epac2-camps DNA probe and 15 µL Lipofectamine 2000 for 24 h. Cells were trypsinized and re-seeded on Matrigel-coated 35 mm glass-bottomed dishes for experiments 24–48 h later. Before each experiment, cells were washed with saline solution and dishes mounted on an inverted microscope and continually perfused with standard saline solution with or without test reagents. Cells were visualized with a ×40 oil immersion objective and excitation at 435 nm was achieved using a CFP/YFP filter set and Xenon Arc Lamp coupled to a monochromator (Cairn Research) and controlled by MetaFlour software (Universal Imaging; Cairn Research). Cyan fluorescent protein (CFP) emission was acquired at 470 nm and YFP emission was acquired at 535 nm using an Optosplit II image splitter and Hamamatsu Orca-ER digital camera (Cairn Research). For the analysis, all data were normalized to an average baseline FRET measured before application of the first test reagent. Average fluorescence ratios were determined over 10 s periods, before addition of a test agent, during its perfusion, and during perfusion with forskolin (fsk)/IBMX (10 µM of each). The maximal response of each cell to a test reagent was expressed as a fraction of the maximal response achieved with fsk/IBMX measured in the same cell at the end of each experiment, where 0 represents no response and 1 is the maximal response. Cells in which the response was less than 5% of the fsk/IBMX response were classed as non-responders and excluded from subsequent data analysis.

Single cells were monitored as either taurolithocholic acid (TLCA) or GPBAR-A was perfused into the bath chamber. With this probe, addition of either reagent resulted in a decreased FRET signal, consistent with an increase in intracellular cAMP.

Solutions and chemicals

The saline solution contained (in mM) 128 NaCl, 5.6 KCl, 4.2 NaHCO₃, 1.2 NaH₂PO₄, 2.6 CaCl₂, 1.2 MgCl₂, 10 HEPES. For experiments performed at a high potassium concentration, KCl was increased to 70 mM and NaCl decreased to 72.5 mM to maintain osmolarity. Unless stated, all drugs and chemicals were obtained from Sigma (Poole, UK). Where possible, drugs were made up as 1000× stock. The bile acids were tested at both 10 and 30 µM in the initial secretion studies, as indicated, and as both stimulated a similar response the results were pooled. For all other experiments thereafter, bile acids were used at 10 µM. The GPBA agonist, referred to as GPBAR-A, is ((4-[[3,5-bis(trifluoromethyl)phenyl]methyl]-6-(2-fluorophenyl)-4,5-dihydro-pyrido[3,2-f]-1,4-oxazepin-3(2H)-one) (Keitel et al., 2009) and was used at 3 µM.

Bile acids and a specific GPBA agonist stimulate GLP-1 secretion from primary intestinal cultures

Bile acid responses were further studied in mixed murine primary intestinal cultures, a model developed to mimic more closely the physiology of the native adult L-cell (Reimann et al., 2008; Friedlander et al., 2010; Rogers et al., 2010; Tolhurst et al., 2011). In primary colonic cultures, TLCA, LCA and DCA promoted GLP-1 secretion by 1.8-fold (P < 0.001, n = 24), 1.8-fold (P < 0.001, n = 12) and 2.4-fold (P < 0.001, n = 9), respectively (Figure 2A). The response to CDCA did not reach statistical significance. GPBAR-A stimulated a substantial, 4.2-fold (P < 0.001, n = 21), increase in GLP-1 release from colonic cell cultures. Upper small intestinal cultures exhibited a smaller, 2.6-fold (P < 0.01, n = 7), GLP-1 response to GPBAR-A, and were not responsive to TLCA in the absence of added nutrient (Figure 2B). In 10 mM glucose, however, TLCA triggered an additional 1.5-fold increment in secretion, suggesting a synergistic interaction between glucose and bile acid sensing pathways in these cells.

Open in a separate window

Figure 2

GLP-1 secretion from primary murine cultures. (A) Mixed murine intestinal colon cultures were incubated in saline solution containing bile acids (10 or 30 µM) or GPBAR-A (3 µM). (B) Mixed murine upper intestinal cultures were incubated with TLCA (10 µM) or GPBAR-A (3 µM) in the presence (solid columns) or absence (open columns) of glucose (10 mM). GLP-1 was measured in the supernatant and cell extracts and secretion expressed as a percentage of total GLP-1 relative to basal secretion measured in parallel on the same day. Error bars represent 1 SEM; the number of wells is shown above each column, and significance is shown relative to baseline, using anova followed by a single factor t-test: ns, not significant; **P < 0.01, ***P < 0.001; ††P < 0.01, †††P < 0.001. Data were collated from two to nine separate experiments.

GPBA mediated cAMP response

We employed a FRET-based cAMP sensor, Epac2-camps (Nikolaev et al., 2004), recently validated in our laboratory for use in GLUTag cells (Friedlander et al., 2010), to monitor the time course and relative magnitude of cAMP changes in individual GLP-1-producing cells following addition of either TLCA or GPBAR-A. Figure 3A shows an example of a FRET trace, in which the CFP/YFP ratio acts as a readout of cAMP. GPBAR-A and TLCA triggered an elevation of cAMP concentrations within a few minutes of their addition to the perfusion solution, as exemplified in Figure 3A. The response to TLCA was 38% (P < 0.05, n = 11) of the maximal response to fsk/IBMX and, consistent with the secretion results, GPBAR-A stimulated a greater increase in cAMP with a mean elevation to 57% (P < 0.001, n = 11) of the maximum (Figure 3B).

Open in a separate window

Figure 3

Intracellular cAMP changes in response to GPBA activation. (A) GLUTag cells transfected with a cytosolic Epac2-based FRET probe were perfused with GPBAR-A (3 µM), followed by forskolin (10 µM)/IBMX (10 µM) as indicated, while CFP and YFP emission was monitored. Each trace represents the CFP/YFP ratio of a single cell normalized to the starting value. (B) Mean changes in the CFP/YFP emission ratio in response to GPBAR-A or GPBA relative to the maximal response to fsk/IBMX, where 0 represents no change in the emission ratio and 1 is the maximal change. The number of cells that responded out of the total number of cells monitored is indicated above each column. The mean was calculated from responding cells only, from three to five separate experiments. Statistical significance was tested using Student's t-test; versus basal *P < 0.05,***P < 0.001, and GPBAR-A versus TLCA †P < 0.05. Error bars represent 1 SEM.

GPBA mediated Ca²⁺_i response

To investigate whether GPBA activation, and the associated more modest cAMP response, would affect intracellular calcium levels, we monitored the 340/380 nm fluorescence ratio in GLUTag cells loaded with the calcium indicator, fura-2 (Figure 4A). GLUTag cells were selected in preference to primary cells for this experiment, because of our previous experience that a relatively large cell number would be required to dissect the effects of cAMP. Using the definition of a calcium response as an increase in the fluorescence ratio of at least 1.2-fold, we found that GPBAR-A increased calcium in 56/149 GLUTag cells, with mean response of 1.3-fold across all cells (P = 0.01, n = 149). Glucose, by contrast, triggered an elevation in calcium in 132/149 cells, with a mean 2.2-fold increment above basal (P = 2.0 × 10⁻⁸, n = 149). When glucose was added in the presence of GPBAR-A, 148/149 cells responded (P = 0.0001 by χ²-test vs. glucose alone), and the mean calcium response was 2.6-fold above basal (P = 2.1 × 10⁻⁴⁰, n = 149).

Open in a separate window

Figure 4

Intracellular calcium responses to GPBA activation in GLUTag cells. Ca²⁺ monitored as the ratio of fluorescence at 340 and 380 nm in individual GLUTag cells loaded with fura-2. (A) Representative traces of the response observed to the application of GPBAR-A (3 µM) and glucose (Glc, 10 mM) added as indicated by the horizontal bars. (B) Mean 340/380 nm ratio averaged over 10 s periods (n = 149 cells from five separate experiments) divided into quartiles based upon the size of the initial glucose response. (i) The 25% least glucose-responsive cells, (ii) the next 25% least glucose-responsive, (iii) the third quartile of glucose responders and (iv) the top 25% of glucose-responsive GLUTag cells. Error bars indicate 1 SE. ns, not significant; **P < 0.01, ***P < 0.001.

To examine the explanation for the synergistic calcium responses triggered by glucose and GPBAR-A, we stratified the cells into quartiles based on the size of their initial glucose response (Figure 4B). Analysed in this way, it is evident that cells that were initially unresponsive to glucose became glucose-responsive in the presence of GPBAR-A. Thus, in the 25% least (non-) glucose-responsive cells, in which the mean initial glucose response was only 1.2-fold (P < 0.001, n = 37), the response to glucose in the presence of GPBAR-A was increased to 1.7-fold (P < 0.001, n = 37) relative to GPBAR-A. The second quartile of responders similarly showed an enhanced glucose response in GPBAR-A. By contrast, cells that exhibited a robust initial response to glucose (top 50% of glucose responders), on average, showed no significant enhancement of the glucose response in the presence of GPBAR-A (third quartile 2.3-fold vs. 2.4-fold, P = 0.1, n = 38 and fourth quartile 3.6-fold vs. 3.2-fold, P = 0.07, n = 37).

GPBA-mediated GLP-1 secretion is not dependent on K_ATP channel closure

To investigate the role of K_ATP channel closure as a critical event downstream of GPBA activation, we tested the effects of TLCA and GPBAR-A on primary colonic cultures in the presence of the sulphonylurea glibenclamide at a concentration predicted to maximally close K_ATP channels (100 nM). While glibenclamide itself triggered GLP-1 release, consistent with the known expression of K_ATP channel subunits in L-cells (Reimann et al., 2008), it had no effect on the ability of GPBAR-A and TLCA to stimulate GLP-1 secretion (Figure 5A). Similarly, both TLCA and GPBAR-A enhanced GLP-1 secretion in the presence of 70 mM KCl and the K_ATP channel opener diazoxide (340 µM), a combination predicted to depolarize the cell membrane towards the Nernst-potential for K⁺ and thus activate voltage-gated Ca²⁺ channels. In small intestinal and colonic cultures, 70 mM KCl + diazoxide triggered the release of GLP-1 by 2.1-fold (P < 0.05, n = 9) and 3.2-fold (P < 0.001, n = 6), respectively (Figure 5A, B), and GPBAR-A increased the GLP-1 secretory response by a further 1.5-fold (P < 0.05, n = 9) in small intestine cultures and 2.4-fold (P < 0.001, n = 6) in the colon. TLCA stimulated secretion a further 2.0-fold (P < 0.001, n = 5) under these depolarizing conditions in the colon.

Open in a separate window

Figure 5

GLP-1 secretion from primary murine cultures in response to GPBA activation in the presence of glibenclamide and under depolarizing conditions. (A) Mixed murine intestinal colon cultures were incubated in saline solution containing GPBAR-A (3 µM) or TLCA (10 µM) in either standard saline solution (open columns), high KCl (70 mM) saline solution containing diazoxide (Dz, 340 µM) or with glibenclamide (100 nM) in standard saline solution. (B) GLP-1 response from primary cultures of murine upper small intestine incubated with GPBAR-A (3 µM) or TLCA (10 µM) in either standard saline solution (open columns) or a high KCl (70 mM) saline solution containing diazoxide (340 µM). GLP-1 was measured in the supernatant and cell extracts and secretion expressed as a percentage of total GLP-1 relative to basal secretion measured in parallel on the same day. Error bars represent 1 SEM; the number of wells is shown above each column, and significance is shown relative to baseline and between conditions calculated by anova followed Student's t-tests: ns, not significant; †P < 0.05; **P < 0.01, ***/†††P < 0.001. Data were collated from two to four separate experiments.

The authors would like to thank Martin Lohse for the gift of the Epac-based cAMP sensors and the GLUTag cells were kindly provided by Dr D Drucker (Toronto). We also thank Professors MA Ghatei and SR Bloom for the GLP-2 assays. This work was supported by grants from the Wellcome Trust (#WT088357, #WT084210), and an MRC PhD studentship. ClR was funded by a NIHR clinician scientist award. Mouse work was supported by MRC-CORD.