Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of gastric inhibitory polypeptide in the regulation of pancreatic endocrine secretion Verchere, Cameron Bruce 1991

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A1 V47.pdf [ 12.44MB ]
Metadata
JSON: 831-1.0101044.json
JSON-LD: 831-1.0101044-ld.json
RDF/XML (Pretty): 831-1.0101044-rdf.xml
RDF/JSON: 831-1.0101044-rdf.json
Turtle: 831-1.0101044-turtle.txt
N-Triples: 831-1.0101044-rdf-ntriples.txt
Original Record: 831-1.0101044-source.json
Full Text
831-1.0101044-fulltext.txt
Citation
831-1.0101044.ris

Full Text

THE ROLE OF GASTRIC INHIBITORY POLYPEPTIDE IN THE REGULATION OF PANCREATIC ENDOCRINE SECRETION By C A M E R O N B R U C E V E R C H E R E B.Sc, The University of British Columbia, 1983 M.Sc, The University of British Columbia, 1987 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S P H Y in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Physiology We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A August 1991 © C a m e r o n Bruce Verchere, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date OCT. 7 > i^H I DE-6 (2/88) i i ABSTRACT Gastric inhibitory polypeptide (GIP) has often been referred to as glucose-dependent insulinotropic polypeptide because of its potent stimulatory effect on insulin secretion. A stimulatory action of GIP on the release of other islet hormones has also been suggested in certain studies. However, attempts to investigate the interaction of GIP with pancreatic islet cells at the molecular and cellular levels have been unsuccessful, presumably due to damage to the GIP receptor in the enzymatic isolation of pancreatic islets. The experiments presented here examined the hypothesis that GIP exerts a direct influence on the secretion of islet hormones via specific receptors. This was investigated primarily through the use of in vitro preparations of islets and /3-cells made more sensitive to the actions of GIP by tissue culture. Isolated rat islets, following two days of culture, responded to physiological concentrations of GIP (1 nM) with increased insulin, glucagon and somatostatin secretion. The stimulation of islet hormone secretion was dependent upon the concentrations of both GIP and glucose. The threshold glucose concentration for 1.0 n M GIP-stimulated insulin release from cultured islets was 8.9 m M and the maximum potentiation of GIP-stimulated insulin secretion was observed in the presence of 17.8 m M glucose. In 2.75 m M glucose, below the threshold for the insulinotropic action of GIP, the presence of arginine (10 mM) was sufficient to permit the stimulatory action of GIP on the 0-cell to occur, suggesting that non-glucose fuel stimuli can sensitize the /3-cell to the action of GIP. In contrast, glucagon release from isolated islets was stimulated by 1.0 n M GIP only in the presence of low glucose concentrations, as glucose levels greater than 2.75 m M attenuated the stimulatory effect of GIP on the a-cell. GIP-stimulated somatostatin secretion was not clearly glucose-i i i dependent, although the maximum stimulation of somatostatin release by GIP occurred in 17.8 m M glucose. Acetylcholine (ACh) strongly suppressed somatostatin secretion induced by 17.8 m M glucose plus 10 n M GIP. Together, these results suggest that GIP is involved not only in the regulation of insulin and glucagon secretion, as was shown in previous studies, but also that GIP may regulate pancreatic somatostatin secretion. Further, the demonstration that GIP stimulated the release of these islet hormones in vitro indicates that the peptide acts directly on islet cells and that its effects are not dependent on indirect neural or vascular pathways. GIP-stimulated islet hormone secretion appears to be strongly influenced by the presence of glucose, other fuels and neural inputs. GIP also stimulated the release of insulin from a preparation of /3-cells sorted to > 98 % purity using a fluorescence-activated cell-sorter (FACS), confirming that GIP acts directly on the /3-cell to exert its insulinotropic effect. The sensitivity of FACS-purified B-cells to GIP was considerably less than intact islets, but could be enhanced by the addition of 10 n M glucagon or by culturing the cells with a non-/S-cell fraction obtained from the F A C S that consisted of approximately 50 % a-cells. These results suggested that the mechanism of GIP-stimulated insulin secretion may partly involve intra-islet interactions, possibly with the islet a-cell. Other known insulin secretagogues increased secretion from FACS-purified j3-cells, including glucagon, glucagon-like peptide I (7-36-NH2), A C h and cholecystokinin-8 (CCK-8). GIP also stimulated insulin release from a mouse tumor 0-cell line (/3TC3). GIP-induced insulin secretion from /3TC3 cells was glucose-dependent and like FACS-purified /3-cells, was found to be approximately 100-fold less sensitive than isolated islets to the stimulatory effects of GIP. A mono-iodinated, biologically active form of 1 2 5 I -GIP was purified by reverse-phase high performance liquid chromatography (HPLC) for use in radioreceptor binding i v studies with j8TC3 eells and cultured rat islets. Binding of the HPLC-purified radioligand to 0TC3 cells and cultured islets was displaced by GIP in a concentration-dependent manner starting at concentrations of GIP as low as 1 n M . 1 2 5 I -GIP could not be displaced by numerous other peptides including those structurally related to GIP. The data strongly support the existence of a specific receptor for GIP on pancreatic j8-cells. A biotinylated form of GIP (B-GIP) was also produced and found to stimulate insulin release from the perfused rat pancreas and may therefore be a useful probe in future GIP receptor studies. Islets isolated from obese Zucker (fa/fa) rats were found to be more sensitive to the insulinotropic action of GIP than those of lean (Fa/?) rats. As had been observed in the perfused pancreas, the islets of obese but not lean rats responded to GIP (2.0 nM) in the presence of 4.4 m M glucose, below the normal threshold for the insulinotropic action of GIP. This could not be attributed to intra-islet interactions, since islet glucagon and somatostatin secretion from obese and lean rat islets were similar in 4.4 m M glucose plus GIP, and obese rat islets were more sensitive to the inhibitory effects of somatostatin. The results suggest the existence of a /3-cell defect in obese rats, possibly at the level of the GIP receptor or intracellular signal transduction. In 8.9 m M glucose, glucagon secretion from obese but not lean rat islets was suppressed compared to 4.4 m M glucose and enhanced by the addition of 2.0 n M GIP. The data indicate the presence of multiple alterations in hormone secretion from obese Zucker rat islets which remain apparent in culture, following removal of the islets from their in situ environment. These defects may contribute to the altered metabolic and obese state of fa/fa animals. GIP-stimulated insulin secretion from the perfused rat pancreas was potently suppressed by infusion of the pancreatic neuropeptide galanin. The inhibitory action of galanin appeared to be specific for GIP-stimulated insulin secretion, since the same V concentration of galanin (10 nM) had no effect on A C h - or CCK-8-stimulated insulin release and modestly suppressed the insulin response to arginine. The further demonstration that galanin suppressed GIP-induced insulin release from mixtures of FACS-purified B- and non-/3-cells suggests that galanin exerts its inhibitory influence on GIP-stimulated insulin release by a direct interaction at the level of the B-ce\\. Porcine galanin was without effect on somatostatin secretion from the perfused rat pancreas in the presence of GIP. In summary, these studies provided clear evidence that GIP directly interacts with the pancreatic islet to stimulate the release of insulin, glucagon and somatostatin. Moreover, it was shown that the stimulation of islet hormone secretion by GIP is subject to considerable modulation by glucose and other agents. The data further showed that alterations in islet sensitivity to GIP may contribute to the metabolic disturbances which exist in obese Zucker rats. Finally, a sensitive radioreceptor assay for GIP was developed and the presence of GIP receptors on rat pancreatic islets was demonstrated. v i TABLE OF CONTENTS Page A B S T R A C T ii LIST O F T A B L E S xiv LIST O F F I G U R E S xv A C K N O W L E D G E M E N T S xviii I N T R O D U C T I O N 1 M E T H O D S 50 I. E X P E R I M E N T A L P R E P A R A T I O N S A . Animals 50 1. WistarRats 50 2. Zucker Rats 50 B. In Vitro Preparations 51 1. Isolated Islets 51 a) . Materials 51 b) . Islet Isolation 51 c) . Hormone Secretion from Isolated Islets 53 2. FACS-Purified Islet Cells 54 a) . Materials 54 b) . Islet Cell Preparation 55 c) . Purification of /3- and Non-/3 Islet Cells 56 d) . Culture of /?-Cells and Non-/?-Cells 56 e) . Immunocytochemical Characterization of Cell 58 Fractions v i i f). Insulin Secretion Experiments 59 3. /3TC3 Tumor Cell Line 59 a) . Culture of /3TC3 Cells 61 b) . Insulin Secretion Experiments 61 4. Perfused Pancreas 62 a) . Apparatus 62 b) . Surgical Procedure 63 c) . Solutions 64 (i) . Perfusate 64 (ii) . Drugs and Peptides 64 d) . Perfusion Procedure 64 II. GIP R E C E P T O R STUDIES A . GIP Receptor Probes 65 1. 1 2 5 I -GIP 65 a) . Preparation of 1 2 5 I -GIP 65 b) . H P L C Purification of 1 2 5 I -GIP 68 c) . Specific Activity of 1 2 5 I -GIP 69 d) . Analysis of HPLC-Purifiied 1 2 5 I -GIP 70 e) . 1 2 7 / 1 2 5 I - G I P 72 (i) . Preparation of 1 2 7 / 1 2 5 I - G I P 72 (ii) . H P L C Purification of 1 2 7 / 1 2 5 I - G I P 73 2. Biotinylated GIP (B-GIP) 73 a) . Preparation of B-GIP 73 b) . H P L C Purification of B-GIP 74 v i i i B. GIP Receptor Binding Studies 1. Assay Buffer 2. Peptides 3. Binding Assays a) . )3TC3 Cells b) . Isolated Islets 4. Calculations III. P E P T I D E Q U A N T I F I C A T I O N A . Radioimmunoassay 78 1. Insulin 78 a) . Assay Buffer 79 b) . Antiserum 79 c) . 1 2 5I-Insulin 79 d) . Standards 80 e) . Controls 81 f) . Separation 81 g) . Procedure 81 h) . Calculations 82 2. Glucagon 82 a) . Assay Buffer 82 b) . Antibody 83 c) . 1 2 5I-Glucagon 83 d) . Standards 83 e) . Controls 84 74 75 75 75 76 76 77 ix f) . Separation 84 g) . Procedure 84 h) . Calculations 84 3. Somatostatin 85 a) . Assay Buffer 85 b) . Antibody 85 c) . 125I-Somatostatin . 85 d) . Standards 86 e) . Controls 86 f) . Separation 86 g) . Procedure 86 h) . Calculations 87 4. Gastric Inhibitory Polypeptide 87 a) . Assay Buffer 87 b) . Antiserum 87 c) . 1 2 5 I -GIP 88 d) . Standards 88 e) . Separation 88 f) . Procedure 88 B. Enzyme-Linked Immunosorbent Assay 89 1. Gastric Inhibitory Polypeptide 89 a) . Assay Buffer 90 b) . Antisera 90 c) . Standards 90 X d) . Procedure 90 e) . Controls 91 C. Protein Assay 92 IV. A N A L Y S I S O F D A T A A . Isolated Islets, FACS-Purified 0-Cells and /3TC3 Cell Secretion 92 B. Perfused Pancreas 92 1. Galanin Experiments 92 2. GIP Gradient Experiments 92 R E S U L T S 93 I. E F F E C T S O F GIP O N H O R M O N E S E C R E T I O N F R O M I S O L A T E D R A T ISLETS A . Insulin Secretion 94 1. Effect of Glucose Concentration on GIP-Stimulated 94 Insulin Secretion 2. Concentration-Dependency of GIP-Stimulated 94 Insulin Secretion in Low (2.75 mM) and High (17.8 mM) Glucose 3. Modulation of GIP-Stimulated Insulin Secretion by 95 Acetylcholine and Arginine B. Glucagon Secretion 96 1. Effect of Glucose Concentration on GIP-Stimulated 96 Glucagon Secretion 2. Concentration-Dependency of GIP-Stimulated 96 Glucagon Secretion in 2.75 m M Glucose 3. Modulation of GIP-Stimulated Glucagon Secretion 96 by Arginine C. Somatostatin Secretion 97 x i 1. Effect of Glucose Concentration on GIP-Stimulated 97 Somatostatin Secretion 2. Concentration-Dependency of GIP-Stimulated 97 Somatostatin Secretion in 17.8 m M Glucose 3. Modulation of GIP-Stimulated Somatostatin Secretion 97 by Acetylcholine II. E F F E C T S O F GIP A N D O T H E R INSULIN S E C R E T A G O G U E S O N INSULIN S E C R E T I O N F R O M R A T P A N C R E A T I C /3-CELLS PURIFIED B Y F L U O R E S C E N C E - A C T I V A T E D C E L L - S O R T I N G A . Effect of Glucose on Insulin Secretion 108 from FACS-Purified 0-Cells B. Effect of GIP on Insulin Secretion from 108 FACS-Purified /8-Cells C. Effect of Acetylcholine, Cholecystokinin-8, Glucagon 108 and Glucagon-Like Peptide-1 (7-36-NH2) on Insulin Secretion from FACS-Purified B-Cells D . Modulation of GIP-Stimulated Insulin Secretion from 109 FACS-Purified 0-Cells by Glucagon E . Effect of GIP on Insulin Secretion from Mixtures of 109 FACS-Purified B- and Non-0-Cells III. E F F E C T S O F GIP O N INSULIN S E C R E T I O N F R O M BTC3 C E L L S A . Hormone Content of BTC3 Cells 115 B. Effect of GIP on Insulin Secretion in the Presence of 115 4.4 and 17.8 m M Glucose C. Effect of GIP on Insulin Secretion in the Presence of 115 I B M X and Forskolin D . Effect of GLP-1 (7-36-NH2) on Insulin Secretion 116 IV. E F F E C T S O F GIP O N H O R M O N E S E C R E T I O N F R O M I S O L A T E D P A N C R E A T I C ISLETS O F L E A N A N D O B E S E Z U C K E R R A T S A . Hormone Content of Zucker Rat Islets Before and After Culture 119 B. Insulin Secretion from Lean and Obese Zucker Rat Islets 120 x i i 1. Effect of GIP on Insulin Secretion from Zucker Rat Islets 120 2. Effect of Somatostatin on GIP-Stimulated Insulin Secretion 121 from Zucker Rat Islets C. Glucagon Secretion from Lean and Obese Zucker Rat Islets 121 1. Effect of Glucose on Glucagon Secretion 121 from Zucker Rat Islets 2. Effect of GIP on Glucagon Secretion 121 from Zucker Rat Islets D . Somatostatin Secretion from Lean and Obese Zucker Rat Islets 122 1. Effect of GIP and Glucose on Somatostatin Secretion 122 from Zucker Rat Islets V . E F F E C T S O F G A L A N I N O N INSULIN A N D S O M A T O S T A T I N S E C R E T I O N S T I M U L A T E D B Y GIP A N D O T H E R INSULIN S E C R E T O G O G U E S A . Effects of Galanin on Insulin Secretion 128 1. Effect of Porcine Galanin on GIP-Stimulated Insulin 128 Secretion from the Perfused Rat Pancreas 2. Effect of Porcine and Rat Galanin on Insulin Secretion 128 from the Perfused Rat Pancreas Stimulated by Acetylcholine, Cholecystokinin-8 and Arginine 3. Comparison of the Inhibitory Effect of Rat and Porcine 129 Galanin on GIP-Stimulated Insulin Secretion from Rat 0-Cells B. Effects of Galanin on Somatostatin Secretion 129 1. Effect of Porcine Galanin on Somatostatin Secretion 129 from the Perfused Rat Pancreas in the Presence of GIP VI. INVESTIGATIONS INTO T H E E X I S T E N C E O F A GIP R E C E P T O R O N P A N C R E A T I C 0 - C E L L S A . Development of GIP Receptor Probes 135 1. Iodinated ( 1 2 : >I)-GIP 135 x i i i a) . H P L C Purification of 1 2 5 I -GIP 135 b) . Iodination State of Different Peaks 136 of HPLC-Purified 1 2 5 I -GIP c) . Binding of Different Peaks of HPLC-Purified 137 1 2 5 I -GIP to BTC3 Cells d) . Biological Activity of HPLC-Purified I-GIP 138 (i) H P L C Purification of 1 2 7 / 1 2 5 I - G I P 138 (ii) Effect of Peak 2 of HPLC-Purified 138 127/125J. g i p o n i n s m i n Secretion from the Perfused Rat Pancreas d) . Displacement of HPLC-Purified 1 2 5 I -GIP 139 from BTC3 Cells by GIP e) . Displacement of HPLC-Purified 1 2 5 I -GIP 139 from Isolated Islets by GIP B. Biotinylated GIP 1-30 140 1. H P L C Characterization of B-GIP 140 2. Biological Activity of B-GIP 141 DISCUSSION 155 R E F E R E N C E S 191 A P P E N D I X 220 I. C H E M I C A L S O U R C E S 220 II. LIST O F A B B R E V I A T I O N S 222 III. A N T I B O D Y S O U R C E S 223 IV. S Y S T E M E I N T E R N A T I O N A L (SI) UNITS 224 xiv LIST OF TABLES Number Title Page I Some Possible Physiological Actions of Gastric Inhibitory Polypeptide 6 II Possible Reasons for the Insensitivity of Isolated Islet Preparations 39 to the Insulinotropic Action of Gastric Inhibitory Polypeptide III Hormone Content of Zucker Rat Islets Before and After Culture 119 IV Iodination State of Different Peaks of HPLC-Purified 1 2 5 I -GIP 147 X V LIST OF FIGURES Number Title Page 1 Amino Acid Sequences of Porcine, Human, Bovine and Rodent 3 Gastric Inhibitory Polypeptide 2 Analysis of Dispersed Rat Islet Cells by Autofluorescence 57 and Light Scatter on a Fluorescence Activated Cell-Sorter 3 Immunocytochemical Staining of FACS-Purified /3-Cells. 60 4 Profile of Gastric Inhibitory Polypeptide Iodination Mixture 67 Eluted on Sephadex R G-15. 5 Self-Displacement of GIP by 1 2 5 I -GIP 71 6 Effect of Glucose Concentration on GIP-Stimulated Insulin 99 Secretion from Isolated Rat Islets 7 Concentration-Dependency of GIP-Stimulated Insulin Secretion 100 from Isolated Rat Islets 8 Modulation of GIP-Stimulated Insulin Secretion from 101 Isolated Rat Islets by Acetylcholine and Arginine 9 Effect of Glucose Concentration on GIP-Stimulated Glucagon 102 Secretion from Isolated Rat Islets 10 Concentration-Dependency of GIP-Stimulated Glucagon 103 Secretion from Isolated Rat Islets 11 Modulation of GIP-Stimulated Glucagon Secretion from 104 Isolated Rat Islets by Arginine 12 Effect of Glucose Concentration on GIP-Stimulated 105 Somatostatin Secretion from Isolated Rat Islets 13 Concentration-Dependency of GIP-Stimulated Somatostatin 106 Secretion from Isolated Rat Islets 14 Modulation of GIP-Stimulated Somatostatin Secretion from 107 Isolated Rat Islets by Acetylcholine 15 Effect of GIP on Insulin Secretion from FACS-Purified Rat /3-Cells 111 x v i 16 Effect of Glucagon and Glucagon-Like Peptide-1 (7-36-NH2) 112 on Insulin Secretion from FACS-Purified Rat /3-Cells 17 Effect of Acetylcholine and Cholecystokinin-8 on Insulin 113 Secretion from FACS-Purified Rat /3-Cells 18 Effect of GIP on Insulin Secretion from FACS-Purified Rat 0-Cells 114 in the Presence of Glucagon or FACS-Purified Non-/3-Cells 19 Effect of GIP on Insulin Secretion from )3TC3 Cells in 117 4.4 and 17.8 m M Glucose 20 Effect of GIP on Insulin Secretion from )STC3 Cells in 118 the Presence of I B M X and Forskolin 21 Insulin Secretion (per Islet) from Lean and Obese Zucker 123 Isolated Rat Islets in the Presence of Glucose and GIP 22 Effect of Glucose and GIP on Insulin Secretion from 124 Isolated Islets of Lean and Obese Zucker Rats 23 Effect of Somatostatin-14 on GIP-Stimulated Insulin Release 125 from Isolated Islets of Lean and Obese Zucker Rats 24 Effect of Glucose and GIP on Glucagon Secretion from 126 Isolated Islets of Lean and Obese Zucker Rats 25 Effect of Glucose and GIP on Somatostatin Secretion from 127 Isolated Islets of Lean and Obese Zucker Rats 26 Effect of Porcine Galanin on GIP-Stimulated Insulin Secretion 130 from the Perfused Rat Pancreas 27 Effect of Porcine and Rat Galanin on Acetylcholine-Stimulated 131 Insulin Secretion from the Perfused Rat Pancreas 28 Summary of the Effects of Porcine and Rat Galanin on 132 Acetylcholine-, Cholecystokinin-8-, and Arginine-Stimulated Insulin Secretion from the Perfused Pancreas 29 Comparison of the Inhibitory Effect of Porcine and Rat Galanin 133 on GIP-Stimulated Insulin Secretion from Mixtures of FACS-Purified Rat p- and non-0-Cells 30 Effect of Porcine Galanin on Somatostatin Secretion from the 134 Perfused Rat Pancreas in the Presence of GIP x v i i 31 H P L C Elution Profile of 1 2 5 I -GIP 1-42 in a Gradient of 142 32-38 % Acetonitrile 32 H P L C Purity Analysis of Collected Fractions of Peak 2 of 143 HPLC-Purified 1 2 5 I -GIP 33 H P L C Elution Profile of 1 2 5 I -GIP 1-42: Isocratic Elution 144 in 32 % Acetonitrile 34 H P L C Elution Profile of 1 2 5 I -GIP 1-30 in a Gradient of 145 32-38 % Acetonitrile 35 Analysis of Iodination State of HPLC-Purified 1 2 5 I -GIP (Peak 2) 146 36 Specific Binding of Different Peaks of HPLC-Purified 148 1 2 5 I -GIP to /3TC3 Cells 37 H P L C Elution Profile of 1 2 7 / 1 2 5 I - G I P 1-42 in a Gradient of 149 30-36 % Acetonitrile 38 Effect of HPLC-Purified 1 2 7 / 1 2 5 I - G I P (Peak 2) on Insulin 150 Secretion from the Perfused Rat Pancreas 39 Displacement of HPLC-Purified 1 2 5 I -GIP 151 from /3TC3 Cells by GIP 40 Displacement of HPLC-Purified 1 2 5 I -GIP from Cultured 152 Rat Islets by GIP 41 H P L C Analysis of Biotinylated GIP 1-30 153 42 Effect of Biotinylated GIP 1-30 on Insulin Secretion 154 from the Perfused Rat Pancreas x v i i i ACKNOWLEDGEMENTS I would first like to thank my supervisor, Dr. John C. Brown, for his expert guidance, support and friendship over the past several years. He has helped me develop as both a scientist and a person. I also thank Dr. Ray Pederson, for his constant support throughout my stay in this department as a supervisory committee member, graduate advisor, retreat host, first beer tent organizer (and patron), and for countless other duties listed neither here nor in his job description. I thank Dr. Yin Nam Kwok for sharing his unique insights on science and life, as well as for his help in receptor binding assays and for providing reagents for the somatostatin RIA. I thank Dr. Chris Mcintosh for his friendship and for helpful discussions on much of the work presented in this thesis, particularly on the production and purification of GIP receptor probes. Thanks also to Dr. Alison Buchan for her help in cell culture experiments and to Dr. Mark Meloche for his assistance in setting up the isolated islet and FACS-purified /3-cell preparations. I sincerely thank the members of my supervisory committee (Drs. J. A . Pearson, Brown, Pederson, K . G . Baimbridge, and R. Brownsey) for their advice in the planning and interpretation of experiments and their speed and skill in reading and editing the final version. I am very grateful for the skillful technical and secretarial assistance that people in the M R C group were always willing to provide. In particular, the following individuals made direct contributions to this thesis: Sue Aynsley, Herminia Sy, Connie Chisolm, Petra Sanderson, and Marie Langton. Generous support was also provided by the Phsyiology Department staff: John Sanker, Joe Tay, Dave Phelan, Debbie Yaschuk, Mary Forsyth, and Zaira Khan. xix Other graduate students in the M R C group helped make graduate studies fun and finishable. I especially thank the board of directors of B.T. Inc.: Rob Campos, Mike Wheeler, Andy Obenaus, and Eric Accili (for excellent back-up vocals). I will never forget what I learned during those stressful summers in the tent: "When it comes to serving beer, we're in tents". Thanks also to Carole McLean, Sue Curtis, Tim Kieffer and Jian "Duke" Wang for their contributions to this work. I gratefully acknowledge the following people for making available antibodies and peptides for use in these studies: Prof. N. Yanaihara (Shizuoka, Japan) for GIP 1-30; Dr. L. Morgan (Guildford, England) for GIP antiserum; and Dr. M . Gregor (Berlin, Germany) for glucagon antibody. I also thank Dr. Z . Huang ( M R C Group) for maintaining the /3TC3 cell line in culture. The financial support of the Medical Research Council of Canada and the Canadian Diabetes Association is also gratefully acknowledged. And finally, I thank my family, especially my parents for their tireless support of my studies, and my wife, Cindy, for her love and encouragement and excellent tracings of the chromatograms. For Cindy, who is always there and for Cito, who is there 162 days a year 1 INTRODUCTION Although it was originally isolated and named on the basis of its ability to inhibit gastric acid secretion, the gut hormone gastric inhibitory polypeptide (GIP) has been more widely studied for its role in the regulation of islet hormone secretion, particularly insulin (Brown et al, 1989). A potent, glucose-dependent insulinotropic action of GIP has been demonstrated both in vivo (Pederson et al, 1975) and in the perfused pancreas (Pederson and Brown, 1976), leading Brown and Pederson (1976) to suggest an alternative interpretation of the acronym GIP: glucose-dependent, insulinotropic polypeptide. Still, consolidation of the place of GIP among the growing list of gastrointestinal hormones with established physiological roles has been hampered by the failure to unequivocally demonstrate that the peptide acts directly on pancreatic /3-cells via specific receptors to stimulate insulin secretion. In particular, physiological concentrations of the peptide failed to stimulate islet hormone secretion from in vitro preparations of pancreatic islets (Schauder et al, 1975; Schafer and Schatz, 1979; Fujimoto, 1981; Szecowka et al, 1982a) and further, the presence of GIP receptors on normal pancreatic islet tissue has not been demonstrated. One of several reasons, proposed by Brown et al (1980), for the lack of an observable effect of GIP on insulin secretion in vitro was damage to GIP receptors caused during the enzymatic isolation of pancreatic islet tissue. This idea was recently supported by the demonstration that following short-term culture, isolated islets regained their sensitivity to physiological concentrations of GIP (Siegel and Creutzfeldt, 1985; 1988). In the studies presented in this thesis, isolated islets and islet cells in culture were used to examine the effect of GIP on the secretion of islet hormones under different conditions in vitro. The goal of these investigations was to help establish whether a physiological role for GIP exists in the regulation of islet 2 hormone secretion. This work addressed the hypothesis that GIP acts directly on specific islet cell receptors to stimulate the secretion of islet hormones and that these effects of GIP are influenced by other circulating factors and neural pathways. Further, alterations in islet cell sensitivity to GIP may play a role in the hyperinsulinemia of obesity. This was explored using islets from a genetic model of obesity, the Zucker rat. In 1930, Kosaka and Lim proposed the existence of a hormone, termed enterogastrone, which was secreted in response to fat in the intestinal lumen and inhibited gastric acid secretion. At that time, only two intestinal hormones had been isolated: secretin (Bayliss and Starling, 1902) and cholecystokinin (CCK) (Ivy and Oldberg, 1927). Both were enterogastrone candidates since preparations of these hormones, although now known to have been relatively impure, were found to have acid inhibitory activity in the vagally denervated gastric (Heidenhain) pouch (Gillespie and Grossman, 1964). Interestingly, C C K preparations which had been found to suppress histamine- or gastrin-stimulated acid secretion (Gillespie and Grossman, 1964; Brown and Magee, 1967) were also found to stimulate acid secretion under fasting conditions (Magee and Nakamura, 1966). The possibility that these disparate effects of C C K on acid secretion might be due to impurities in available preparations of the hormone was explored by Brown and Pederson (1970). It was found that increasing the purity of the C C K preparation enhanced its acid stimulatory activity in the fasting state and reduced its inhibitory activity on pentagastrin-stimulated acid secretion. It was hypothesized that a substance, inhibitory for acid secretion, was removed in the further purification of C C K . Using the canine Heidenhain pouch as a bioassay for the putative inhibitor, Brown et al (1969,1970) purified the active substance from extracts of hog duodeno-jejunal mucosa and named it gastric inhibitory polypeptide, or GIP (Brown, 1971). The complete 42 amino acid sequence of the porcine peptide, as determined by 3 Brown and Dryburgh (1971) and later corrected by Jornvall et al (1981), is shown in Figure 1. The bovine (Carlquist et al, 1984), human (Moody et al, 1984) and rodent (B. Chow, unpublished observations) forms of the hormone have since been sequenced and differ from the porcine peptide at only one (bovine) or two (human and rodent) sites (Figure 1). This high degree of conservation of the peptide among species suggests that GIP has an important regulatory role. The sequence of GIP, as well as the structure of its precursor peptide (Takeda et al, 1987) and the gene encoding it (Inagaki et al, 1989), indicates that GIP belongs to the glucagon family F I G U R E 1: Amino Acid Sequences of Porcine. Human. Bovine and Rodent  Gastric Inhibitory Polypeptide. 1 2 3 4 5 6 7 8 9 10 11 12 P O R C I N E : Tyr-Ala-Glu-Gly-Thr-Phe-Ile-Ser-Asp-Tyr-Ser-Ile-H U M A N : B O V I N E : R O D E N T : 13 14 15 16 17 18 19 20 21 22 23 24 P O R C I N E : -Ala-Met-Asp-Lys-Ile-Arg-Gln-Gln-Asp-Phe-Val-Asn-H U M A N : -His-B O V I N E : R O D E N T : 25 26 27 28 29 30 31 32 33 34 35 36 P O R C I N E : -Trp-Leu-Leu-Ala-Gln-Lys-Gly-Lys-Lys-Ser-Asp-Trp-H U M A N : -Asn-B O V I N E : R O D E N T : -Asn-37 38 39 40 41 42 P O R C I N E : -Lys-His-Asn-Ile-Thr-Gln H U M A N : B O V I N E : -Ile-R O D E N T : -Leu-From Jornvall et al, 1981 (porcine); Carlquist et al, 1984 (bovine); Moody et al, 1984 (human) and B. Chow, unpublished observations (rodent). For human, bovine and rodent GIP, only those amino acids which differ from porcine GIP are indicated. 4 of peptides, which share sequence homology as well as some biological activities (Bell, 1986). The family, which includes glucagon and the glucagon-like peptides (GLP-I and GLP-II), secretin, vasoactive intestinal peptide (VIP), peptide histidine methionine (PHM) and growth hormone releasing factor (GRF), probably arose from a common ancestral gene (Bell, 1986). These related peptides have been shown to share some biological activities, including the stimulation of insulin secretion via the activation of adenylate cyclase in the pancreatic /3-cell. The distribution of GIP, as determined by immunocytochemistry, appears to be limited to specific endocrine cells (K cells) of the duodenum and jejunum in man (Polak et al, 1973; Buchan et al, 1978; Buchan et al, 1982), extending to the terminal ileum in rat and dog (Buchan et al, 1982). Unlike many other peptide hormones, GIP has not been found to exist in any neural tissue nor in endocrine tissue outside of the intestine. Although some immunocytochemical studies indicated that GIP was colocalized with glucagon in the pancreatic a-cell (Alumets et al, 1978; Ahren et al, 1981), it has been suggested that these results were due to cross-reactivity of the GIP antisera used with glucagon or the glucagon-like peptides and their precursors (Brown et al, 1989; Buchan et al, 1978). In concurrence with this, R N A blot analysis detected the presence of human preproGIP m R N A only in samples of human intestine and in no other tissue, including pancreas (Takeda et al, 1987; Inagaki et al, 1989). Investigations into the possible physiological roles of GIP have not yet resolved whether the peptide is indeed an enterogastrone. The hormone has been shown to be released in response to ingestion of fat (Cleator and Gourlay, 1975; Pederson et al, 1975). Yet while the early experiments of Brown et al (1969, 1970) clearly showed that GIP suppressed gastrin- and histamine-stimulated acid secretion from the canine Heidenhain pouch, GIP was found to be a weak inhibitor of pentagastrin-stimulated acid secretion in man (Maxwell et al, 1980) and rat (El-5 Munshid et al, 1980). This discrepancy was clarified somewhat by the demonstration that GIP only suppressed acid secretion from denervated gastric pouches (Soon-Shiong et al, 1979). In addition, infusion of the cholinergic agonist bethanecol was found to abolish the acid inhibitory effects of GIP in the Heidenhain pouch, indicating that parasympathetic nervous activity diminished the acid inhibitory activity of GIP (Soon-Shiong et al, 1984). It has since been suggested that the effects of GIP on acid secretion may be mediated by the release of gastric somatostatin, a potent inhibitor of parietal cell secretion. Concentrations of GIP as low as 1 n M were shown to have a marked stimulatory effect on the release of somatostatin-like immunoreactivity (SLI) from the perfused rat stomach (Mcintosh et al, 1981). In addition, the GIP-stimulated SLI release was abolished by electrical stimulation of the vagus nerves, suggesting a mechanism whereby the parasympathetic innervation suppressed GIP inhibition of acid secretion. As proposed by Brown et al (1989), these studies suggested that GIP may act as a physiological enterogastrone in situations in which vagal activity is reduced, such as the interdigestive period, when GIP levels have been shown to remain elevated (Jorde et al, 1980; Salera et al, 1983). Numerous other effects of GIP on various physiological functions have been suggested by certain studies and are summarized in Table 1. These include several anabolic actions related to fat metabolism including the activation of lipoprotein lipase (Eckel et al, 1979), inhibition of glucagon-induced lipolysis (Dupre et al, 1976), and potentiation of insulin-stimulated triglyceride synthesis (Beck and Max, 1983). A role for GIP in the control of fat metabolism seems logical in consideration of the strong stimulatory effect of fat on GIP release (Beck, 1989). Other proposed actions of GIP include influences on gastrointestinal motility, intestinal secretion, mesenteric blood flow and anterior pituitary hormone secretion (Brown et al, 1989). The physiological significance of many of these putative actions 6 T A B L E 1: Some Possible Physiological Actions of Gastric Inhibitory Polypeptide 1. Inhibition of gastric acid secretion 2. Stimulation of gastric somatostatin secretion 3. Stimulation of insulin secretion 4. Stimulation of glucagon secretion 5. Stimulation of pancreatic somatostatin secretion 6. Inhibition of gastrin secretion 7. Inhibition of pepsin secretion 8. Stimulation of chylomicron clearance 9. Inhibition of lipolysis / stimulation of lipogenesis 10. Inhibition of lower esophageal sphincter pressure 11. Stimulation of mesenteric blood flow 12. Inhibition of intestinal absorption Consult text and Brown et al (1989) for references. of GIP remains to be established since in most of these studies, pharmacological concentrations of GIP were employed. The most widely accepted and frequently studied regulatory role of GIP is the regulation of hormone secretion from the pancreatic islets. The Islands of Langerhans, first identified by Langerhans in 1896 and now more commonly called pancreatic islets, are spherical clusters of several thousand endocrine cells scattered through the exocrine pancreas. The hormones secreted by the islets are generally involved in the maintenance of metabolic homeostasis (Unger and Dobbs, 1978). Four distinct endocrine cell types have been identified in the islet (Orci, 1982): the B-cel\ (containing insulin), the a-cell (glucagon), the 5-cell (somatostatin) and the PP-cell (pancreatic polypeptide). The presence of other peptides, possibly with hormonal roles, has also been demonstrated in these cells. For example, two additional products of glucagon gene processing, the 37 amino acid glucagon-like peptide-I (GLP-I) and the truncated C-terminal section of that peptide [GLP-I (7-36-NH 2)] have been shown to be produced by the pancreas and may be colocalized in the a-cell with glucagon (Mosjov et al, 1986; Manaka et al, 1987). In addition, islet amyloid polypeptide (LAPP or amylin) was recently isolated from islet amyloid 7 precipitates (Westermark et al, 1987; Cooper et al, 1987) and has been shown to coexist with insulin in /3-cell secretory granules (Lukinius et al, 1989). Immunocytochemical studies have demonstrated that the distribution of these cell types within the islet is non-random. In all species in which the distribution of islet cell types has been characterized, B-cells were found to make up the majority of the central part of the islet. Although the distribution of the other cell types varied among species, in rat and to a lesser extent man (Orci, 1982), a mixture of a-, B-, S- and PP-cells were found to make up the periphery of the islet and surround the /3-cell core. Interestingly, two distinct types of islets have been shown to exist in different sections of the pancreas. Islets rich in a-cells but containing few PP-cells were more commonly found in the dorsal (tail) section of the pancreas, while islets in the ventral (head) part of the gland were found to be rich in PP-cells with few a-cells (Baetens et al, 1979). This distribution may be related to the fact that the dorsal and ventral sections of the pancreas arise from different embryological origins, although the physiological significance of this difference, if any, is unknown. The close anatomical relationship between a- and 0-cells in the pancreatic islets led Haist (1965) to first speculate that glucagon and insulin may exert local control over each other's release. It has since been demonstrated that exogenous administration of islet hormones does indeed have profound effects on the release of other islet hormones. Specifically, glucagon has been shown to stimulate both B-and 8-cell secretion, insulin has been found to suppress a-cell secretion, and somatostatin was shown to potently inhibit the release of both insulin and glucagon (Samols et al, 1986). These studies have strengthened the idea that islet hormone secretion may be under considerable local control. One of several possible pathways by which intra-islet effects may be mediated is via the interstitial space (paracrine secretion). In addition, rat islets have been shown to possess a unique 8 islet microvasculature, in which B-cells in the core of the islet are perfused first followed by non-/3-cells in the heterocellular periphery of the islet (Bonner-Weir and Orci, 1982). Retrograde pancreas perfusion studies (Samols et al, 1988; Samols and Stagner, 1991) have recently provided support for this idea and further suggest that a specific order of perfusion exists in the islet of several different species, including man and rat, which is B- to a- to 5-cell. Transmission by either paracrine or intra-islet endocrine pathways is likely to expose the target cells to concentrations of hormone several fold that observed in the peripheral circulation. Samols and Stagner (1990) estimated that via the islet microvasculature, a- and S -cells would be exposed to concentrations of endogenous insulin 102 to 104 times peripheral levels. A third proposed mode of communication between islet cells is via gap junctions. Studies have shown that the cells within an islet are electrically coupled and that injected dyes rapidly pass between both homogeneous and heterogeneous cell types within an islet (Meda et al, 1984). It is therefore likely that islet cells share intracellular mediators such as calcium and c A M P and may recruit each other to response. Currently available techniques have limited study into these proposed intra-islet regulatory mechanisms and thus they remain speculative (Weir and Bonner-Weir, 1990). Still, the majority of studies on islet hormone secretion, including some of the experiments presented in this thesis, have employed preparations such as the perfused pancreas or isolated islets in which the islets were intact, and therefore possible intra-islet control mechanisms in the interpretation of results must be considered. The important physiological roles of insulin and glucagon in the regulation of carbohydrate metabolism and their implication in the pathophysiology of metabolic disorders, primarily diabetes mellitus, has led to considerable study into the regulation of islet hormone secretion. The control of insulin secretion from the pancreatic B-ce\\, in particular, has been the focus of more study than any other 9 hormone since the development of the first radioimmunoassay (RIA) for insulin in 1959 (Berson and Yalow). The regulation of insulin secretion is now known to be multifactorial, involving nutrient, neural and hormonal pathways which interact at the /3-cell to produce an appropriate level of insulin secretion in response to the prevailing metabolic condition (Gerich et al, 1976; Berthoud, 1984; Rasmussen et al, 1990). Since insulin is a primary regulator of carbohydrate and lipid metabolism, it follows that the metabolic fuels, and in particular glucose, are prominently involved in the control of /3-cell secretion. Indeed, D-glucose has been demonstrated to be a strong stimulus for insulin secretion in numerous studies in vivo and in vitro. Circulating insulin levels have been shown to rise abruptly in response to an intravenous or oral glucose challenge in man (Perley and Kipnis, 1967). A stimulatory effect of glucose on insulin secretion has also been demonstrated in the perfused pancreas (Grodsky et al, 1963), isolated islets in static incubation (Zawalich et al, 1977) or in a flow-through perifusion system (Zawalich et al, 1987), dispersed islet cells (Pipeleers et al, 1982) and purified /3-cells (Pipeleers et al, 1985b). Appropriately, the threshold glucose concentration for stimulation of insulin secretion in man (approximately 5 mM) is marginally above the normal fasting glucose concentration of 4.4 m M and is exceeded postprandially, when blood glucose levels rise to 6-7 m M (Tasaka et al, 1975). Yet when challenged with a glucose concentration similar to that observed following a meal, the perfused pancreas has been shown to produce a modest increase in insulin secretion compared to that observed in the presence of a supraphysiologic glucose stimulus (e.g. 17.8 mM) or neurohormonal stimuli such as acetylcholine (ACh) or GIP (Verchere et al, 1991). Since /3-cell stimuli other than glucose generally require the presence of a threshold glucose concentration (usually 5-6 mM) to be present to exert their insulinotropic activity, it has been suggested that the primary role of 10 glucose in the regulation of insulin secretion is to modify the responsiveness of the /3-cell to these neurohormonal agents (Rasmussen et al, 1990). The importance of neurohormonal interactions with glucose at the /J-cell was emphasized by the recent demonstration that glucose alone was a very poor stimulus for insulin secretion from purified single /3-cells, requiring the presence of glucagon or c A M P for stimulatory activity (Pipeleers et al, 1985b). In addition to glucose, other nutrients absorbed into the circulation following meal ingestion have insulin stimulating capabilities, including other carbohydrates, amino acids and fatty acids. The only carbohydrate shown to effectively stimulate insulin secretion from islets in vitro in the absence of glucose was the hexose sugar mannose, while fructose was found to weakly enhance insulin release in the presence of threshold glucose concentrations (Zawalich et al, 1977). Amino acids have been found to vary widely and between species in their ability to stimulate the /3-cell. Arginine, lysine and leucine were the most effective stimulators of insulin secretion in man (Floyd et al, 1966) and in the perfused rat pancreas (Gerich et al, 1974), while in dogs tryptophan, leucine, aspartate and isoleucine were most potent (Rocha et al, 1972). The effect of fats on insulin secretion has been less well studied. Although infusion of free fatty acid mixtures in dog (Crespin et al, 1969) and in man (Balasse and Ooms, 1973) have been shown to cause an increase in circulating insulin levels, it has not been established specifically which fats act on the /3-cell or if indeed, fats act directly on the /3-cell at all. Both fat- (Balasse and Ooms, 1973) and amino acid- (Floyd et al, 1970) stimulated insulin release have been shown to be glucose-dependent, having a weak or non-existent effect in the absence of glucose and being strongly potentiated in the presence of hyperglycemia. The mechanisms by which nutrient secretagogues, including glucose, act on the yS-cell are thought to be related. Studies in the perfused pancreas (Grodsky et al, 1963) and isolated islets (Malaisse et al, 1976; 1977; Zawalich et al, 1977) suggested that the ability of a sugar to stimulate insulin secretion correlated with its rate of metabolism in the B-ce\\. In 1979, Malaisse and colleagues described the "Fuel Hypothesis", which proposed that nutrient secretagogues for insulin are taken up by the jS-cell and rapidly metabolised and that the products generated by nutrient metabolism are coupled to the secretory process. Such a mechanism would allow the /3-cell to act as a fuel sensor organ, uniquely equipped to secrete insulin according to the energy state of the cell. This general scheme for nutrient-induced insulin release has become widely accepted, although the mechanisms coupling nutrient metabolism to secretion are still unclear (Prentki and Matschinsky, 1987; MacDonald, 1990). Adenine and guanine nucleotides, p H , altered oxidation state of pyridine nucleotides, inorganic phosphates, and acetyl- and acyl-CoA's have all been implicated as possible mediators in the stimulus-secretion coupling of nutrient secretagogues in the jS-cell (Prentki and Matschinsky, 1987). Of these, A T P generated by fuel catabolism is the strongest candidate, since it has been shown to close ATP-sensitive potassium channels on the /3-cell membrane, leading to depolarization and subsequent increases in intracellular calcium (Cook and Hales, 1984). The insulin response to a rapid-onset and sustained stimulus of glucose or other secretagogue such as GIP is typically biphasic, consisting of a rapid first peak immediately followed by a nadir and then a prolonged second phase which may plateau or continue to increase (Curry et al, 1968; Porte and Pupo, 1969). The physiological relevance of the biphasic insulin response to a rapid-onset stimulus has been questioned since it is not observed when a graded increase of the stimulus is presented, such as with blood glucose after a meal (Grodsky, 1972). However, it has been suggested that the first-phase burst of insulin secretion may serve to prime insulin target tissues, primarily the liver, to prevent secondary hypoglycemia due to the effects of insulin (Grodky, 1989; Luzi and DeFronzo, 1990). The mechanism 12 underlying the biphasic dynamics of insulin secretion is unknown. O'Connor et al (1980) proposed two models to explain this pattern: a storage-limited model, which supposes that either intracellular mediators or /3-cell secretory granules are segregated into two intracellular pools, one immediately available for rapid release of insulin; and a signal-limited model, which postulates an interaction between stimulatory and feedback inhibitory signals. Rasmussen et al (1990) recently described evidence for the former theory using pharmacological stimulators of different intracellular mediators of 0-cell secretion. It was suggested that activation of calcium-sensitive, calmodulin-dependent protein kinases mediated the first peak of insulin secretion, whereas the second phase was due to activation of protein kinase C. In any case, elucidation of the mechanisms underlying this pattern of insulin release may prove useful in understanding the pathophysiology of diabetes, since it has been known for many years that patients with non-insulin dependent diabetes mellitus (NIDDM) have deficient first phase insulin secretion (Seltzer et al, 1967). Several groups recently demonstrated the existence of a third phase of insulin secretion observed in perfused pancreas and isolated islet preparations subjected to prolonged (> 3 h) exposure to secretagogues (Bolaffi et al, 1986; Hoenig et al, 1986; Curry, 1986). A decline in secretory rate from elevated second phase levels was observed which lasted for the duration of the experiment (up to 50 h). Since after 20 h exposure to high glucose the islets could still be stimulated with forskolin, the diminished secretory rate could not be attributed to exhaustion of stored insulin or cell death and may have been related to desensitization of the /3-cell to the secretagogue (Bolaffi et al, 1986). It was suggested that this desensitization may parallel the hyperglycemia-induced desensitization observed in N I D D M (Grodsky, 1989). The anabolic effects of insulin in metabolic regulation are mirrored by the 13 effects of glucagon, the major counter-regulatory hormone of insulin (Unger and Dobbs, 1978). It follows that many of the control mechanisms which serve to enhance insulin secretion would conversely diminish the secretion of glucagon. Although the regulation of glucagon secretion has been less well studied than that of insulin, glucagon release also appears to be controlled by an interplay of metabolic, hormonal and neural pathways (Unger and Orci, 1981). Like insulin, ingested nutrients such as glucose and amino acids which reach the islets via the circulation have been shown to be important in controlling glucagon secretion. A n increase in plasma glucose concentration has been shown to suppress the basal secretion of glucagon (Ohneda et al, 1969) while low glucose levels stimulated a-cell secretion (Unger et al, 1962). Glucose regulation of a-cell secretion may be especially important during the interdigestive period and in stress ("fight or flight") situations, when increased glucagon levels are required for adequate hepatic fuel production (Unger and Orci, 1981). The observation that glucose modulated glucagon secretion from the perfused pancreas (Pederson and Brown, 1978; Unger, 1985) suggested that the hexose exerts its action directly on the a-cell. However, isolated islet (Leclerg-Meyer et al, 1971) and purified a-cell preparations (Pipeleers et al, 1985c) have been found to be relatively insensitive to changes in the glucose concentration of the incubation medium. This led to the suggestion that glucose regulation of glucagon secretion occurs at least in part via indirect mechanisms or requires the presence of other mediators of a-cell secretion. In vivo, insulin-induced hypoglycemia produced an abrupt increase in circulating glucagon levels; however, this may have been only partly due to a direct effect on the a-cell, since the autonomic nervous system is thought to play a major role in that response (Havel and Taborsky, 1989). Thus, glucose may act more as a modulator of the control of glucagon secretion by neurohormonal agents, although in contrast to the glucose-dependency of the insulin response to non-glucose secretagogues, the 14 a-cell has generally been found to be more responsive to neurohormonal stimulation under hypoglycemic conditions (Pederson and Brown, 1978; Pipeleers et al, 1985c). Alternatively, it has been suggested that the inhibition of glucagon secretion by glucose observed in vivo and in the perfused pancreas may have been largely due to intra-islet inhibitory effects of insulin, released from the /3-cell in response to glucose, and travelling from B- to a-cell via the islet microvasculature (Bonner-Weir and Orci, 1982; Unger, 1983; Samols and Stagner, 1990). With the exception of the branched chain amino acids, all amino acids have been shown to stimulate glucagon release to varying degrees (Rocha et al, 1972). The most potent stimulatory effect on a-cell secretion both in the dog (Rocha et al, 1972) and from purified rat a-cells (Pipeleers et al, 1985c) was observed in the presence of alanine or arginine. Ingestion of a meal high in protein but low in carbohydrate content thus evokes a considerable increase in glucagon levels in addition to insulin. The physiological function of this may be to diminish insulin-induced hypoglycemia while ensuring that insulin levels are high enough to metabolize the products of protein digestion (Unger et al, 1969). Infusion of free fatty acids in man was shown to lower serum glucagon concentrations (Madison et al, 1968). This effect, coupled with the observed rise in insulin secretion, would result in decreased lipolysis and increased deposition of lipids into adipocytes. The mechanisms of action of nutrient modulators of a-cell secretion are not known, as islets contain only 5-20% a-cells (Orci, 1982) and therefore isolated islets are not suitable for the study of intracellular mechanisms in a-cells. The powerful technique of fluorescence-activated cell sorting (FACS) has been shown to yield preparations of a-cells >95% pure and promises to yield new information on this subject (Pipeleers et al, 1985a). Wang and McDaniel (1990) recently showed that arginine was capable of inducing a rapid increase in intracellular calcium in F A C S -purified a-cells. 15 Somatostatin, first isolated by Brazeau et al (1973) as a 14 amino acid peptide from sheep hypothalamic extracts, has been found in nerves and endocrine cells throughout the body, including the pancreatic islets (Polak and Bloom, 1986). It has been shown to have an inhibitory effect on numerous physiological processes, particularly in the gut (Mcintosh, 1985), including the secretion of most gastroenteropancreatic hormones (Reichlin, 1983). Other molecular forms of somatostatin have also been identified, such as somatostatin-28, which is the predominant form of the peptide in the intestine (Mcintosh, 1985); however, pancreatic somatostatin has been shown to occur primarily in the 14 amino acid form (Patel et al, 1981). The physiological role of pancreatic somatostatin is not clear, since islet somatostatin probably makes an insignificant contribution to the circulating levels of the peptide (Taborsky and Ensinck, 1984). Two possible regulatory functions of pancreatic somatostatin have been proposed. The peptide may inhibit the secretion of other islet hormones via a paracrine mechanism (Orci and Unger, 1975) or may function as a local regulator of exocrine pancreatic secretion, via the insulo-acinar axis. The majority of the blood flow to the exocrine pancreas first courses through the islets, thus exposing the exocrine tissue to high concentrations of islet hormones, and somatostatin-14 has been shown to have potent inhibitory effects on the secretion of pancreatic enzymes and bicarbonate (Williams and Goldfine, 1985). Perhaps because of the unknown function of pancreatic somatostatin coupled with the difficulty of examining its secretion in vivo, the mechanisms controlling islet somatostatin secretion are not well understood. In general, fuels stimulatory for insulin secretion appear to have similar effects on islet somatostatin secretion. In the perfused pancreas of rat (Sorenson et al, 1980; Ostenson et al, 1990) and dog (Ipp et al, 1977) and in isolated rat islets (Schauder et al, 1976), glucose has been shown to have a stimulatory effect on somatostatin secretion suggesting a direct 16 action of the sugar on the S-cell. However, in some studies, no significant effect of glucose on somatostatin secretion from the perfused rat pancreas was observed (Dahl, 1983). These variations may be related to unexplained differences in the conditions or preparation used, species or strain of animal. In perfused pancreas studies, the possibility that endogenous insulin stimulated by glucose may suppress somatostatin secretion must be considered (Samols et al, 1988). The effects of amino acids on pancreatic somatostatin secretion have been less well studied, although arginine (Patton et al, 1976; Ostenson et al, 1990) and leucine (Ipp et al, 1977) have been shown to stimulate somatostatin secretion from the perfused pancreas. As early as 1849, the demonstration by Claude Bernard that puncture of the fourth ventricle in dogs caused hyperglycemia suggested that the nervous system was involved in glucose homeostasis. The autonomic nervous system is now known to have a prominent role in the regulation of carbohydrate metabolism, largely via exerting control over the secretion of insulin and other islet hormones (Woods and Porte, 1974; Miller, 1981; Edwards, 1984; Niijima, 1989). The pancreatic islets receive a rich innervation which was noted by Langerhans when he first described pancreatic islets in 1869. The innervation of the islets has been shown to include cholinergic, catecholaminergic and peptidergic nerve terminals impinging on both islet cells and vasculature (Miller, 1981; Ahren et al, 1986b). Parasympathetic preganglionic nerve fibers are carried by the vagus nerve, terminating on postganglionic fibers in pancreatic ganglia scattered throughout the exocrine pancreas and within islets. The parasympathetic postganglionic fibers terminating on islets have been found to be largely cholinergic (Amenta et al, 1983); levels of choline acetyltransferase in the islets were shown to be several fold that of the surrounding exocrine tissue (Godfrey and Matschinsky, 1975). Most preganglionic sympathetic efferents, carried in the splanchnic nerve, 17 synapse on postganglionic fibers in the celiac ganglion, which in turn travel to the islet via the mixed autonomic nerve. However, some sympathetic preganglionic fibers were recently shown to project directly to intrapancreatic sympathetic ganglia (Luiten et al, 1984), possibly giving rise to postganglionic, intramural sympathetic nerve fibers (Baetens et al, 1985). The major postganglionic sympathetic neurotransmitter is norepinephrine. In recent years, increasing numbers of peptides have been localised to pancreatic nerves. The majority of these peptides have been shown to have effects on insulin secretion and often the other islet hormones as well, suggesting that they are involved in the autonomic regulation of islet hormone secretion. Candidate neuropeptides for this role (i.e. those that have both been found in pancreatic nerves and shown to influence islet hormone release) include VIP, the C-terminal tetrapeptide of C C K (CCK-4), gastrin-releasing peptide (GRP), neuropeptide Y (NPY), calcitonin gene-related peptide (CGRP), galanin, substance P and met-enkephalin (Ahren et al, 1986b). Whether these peptides act as part of the parasympathetic or the sympathetic signal, exist in afferent or efferent nerve fibers, or are colocalized with other transmitters is largely unknown and presently the focus of considerable study. However, it is becoming increasingly clear that the pancreatic innervation is made up of a complex network of cholinergic, catecholaminergic and peptidergic intrinsic and extrinsic neurons, similar to the enteric innervation, and most likely critical to the integration of pancreatic endocrine function. Soon after the discovery of insulin in 1921, a stimulatory action of the parasympathetic nervous system on insulin secretion was suggested by the studies of Britton (1925), who showed that electrical stimulation of the vagus nerves in adrenalectomized cats caused a decrease in blood glucose. In 1927, elaborate cross-perfusion studies by Labarre (1927) provided further indication that this lowering of blood sugar was due to the effects of insulin, by demonstrating that blood from a 18 dog in which the vagus nerve was stimulated could reduce blood glucose in a recipient dog. Several groups have since confirmed this observation using R I A to measure an increase in circulating insulin levels following vagal stimulation in vivo (Daniel and Henderson et al, 1967; Frohman et al, 1967; Kaneto et al, 1967). Vagal stimulation of insulin release has also been demonstrated using in situ perfused pancreas preparations in dog (Bergman and Miller, 1973) and rat (Nishi et al, 1987). As blood-borne factors could not recirculate in this model, a direct effect of the vagus on the 0-cell was implied. Acetylcholine, the predominant neurotransmitter of the parasympathetic nervous system, has been shown to have a potent, glucose- and concentration-dependent stimulatory effect on insulin secretion from the perfused pancreas of rat (Loubatieres-Mariani et al, 1973; Verchere et al, 1991), pig (Hoist et al, 198 Id) and dog (Iversen, 1973) as well as isolated islets (Garcia et al, 1988), suggesting that the action of the vagus nerve on the /3-cell was mediated largely by this transmitter. In support of this, evidence for the presence of cholinergic receptors on the /3-cell has been presented. The muscarinic antagonist atropine has been shown to block A C h -stimulated insulin release in vitro (Iversen, 1973; Sharp et al, 1974; Verspohl et al, 1990) and the presence of muscarinic receptors, possibly of the M 3 subtype, has been demonstrated on isolated rat islets (Grill and Ostenson, 1983; Verspohl et al, 1990). However, the inhibitory effects of cholinergic antagonists on vagally induced insulin release have been found to vary between species, strongly suggesting that in some species, non-muscarinic mechanisms are involved. Specifically, it has been found without exception that infusion of the nicotinic antagonist hexamethonium during vagal nerve stimulation abolished the insulin response indicating that nicotinic ganglionic transmission was an essential part of the vagus pathway to the 8-cell (Hoist et al, 1981b; Ahren and Taborsky, 1986; Nishi et al, 1987). Further, infusion of atropine strongly but incompletely suppressed the vagally stimulated 19 insulin response in dogs (Ahren and Taborsky, 1986) and rat (Nishi et al, 1987) and was without effect in cats (Uvnas-Wallenstein and Nilson, 1978) and pigs (Hoist et al, 1981b). These studies suggested that at least part of the insulin response to vagal stimulation in most species was non-muscarinic and probably mediated via post-ganglionic peptidergic neurons. Which of the candidate neuropeptides listed previously mediates this effect in which species is still unclear. However, in the pig, both VIP and G R P have been found in intrapancreatic nerves, were shown to stimulate insulin secretion, and were released in response to vagal stimulation suggesting that these two peptides may fulfill that role in that species (Ahren et al, 1986b). Pancreatic glucagon secretion has also been shown to be increased in response to electrical stimulation of the vagal nerves in numerous species (Kaneto et al, 1974; Hoist et al, 1981a; Bloom and Edwards, 1981; Ahren and Taborsky, 1986; Nishi et al, 1987). Further, this effect could be mimicked by infusion of A C h , suggesting that a-cell muscarinic receptors were involved in this response (Iversen, 1973; Kaneto and Kosaka, 1974; Hoist et al, 1981d). However, evidence for non-muscarinic mechanisms in different species, similar to that observed with vagally stimulated insulin release, has been provided in studies using cholinergic antagonists. Vagally induced glucagon secretion was found to be abolished by infusion of hexamethonium in all species tested (Hoist et al, 1981b; Ahren and Taborsky, 1986; Nishi et al, 1987) whereas atropine only partially suppressed the response in rats (Nishi et al, 1987) and calves (Bloom and Edwards, 1981) and was without effect in dogs (Ahren and Taborsky, 1986) and pigs (Hoist et al, 1981b). Nonmuscarinic mechanisms may also be important in man, because infusion of the muscarinic agonist bethanecol was found to have no effect on glucagon secretion and atropine was without effect on the glucagon response to insulin-induced hypoglycemia (Palmer et al, 1979). Both VIP and G R P have been shown to 20 stimulate glucagon release and are therefore strong candidates to be peptidergic neurotransmitters mediating the vagal control of glucagon release. The physiological function of the dual increase in insulin and glucagon caused by parasympathetic stimulation is unclear and seems futile in view of the reciprocal actions of insulin and glucagon on blood glucose. Ahren et al (1986b) proposed that increases in the levels of these two hormones may be useful when it is desirable to increase glucose turnover; for example, during exercise, increased glucose transport into muscle is required with a parallel increase in hepatic glucose production to maintain euglycemia. The observation that GIP-stimulated somatostatin secretion from the perfused rat stomach was potently inhibited by stimulation of the vagus nerves (Mcintosh et al, 1981) tempted speculation about the possible existence of a similar mechanism at the pancreatic S-cell. Investigations into the parasympathetic regulation of pancreatic somatostatin secretion have produced rather equivocal results, although species differences appear to be important. In the dog, although a suppression of somatostatin secretion by A C h infusion has been reported (Samols et al, 1981), other studies found A C h (Hermansen, 1980) or vagal nerve stimulation (Ahren et al, 1986a) to enhance somatostatin release. In the perfused pig pancreas, an atropine-sensitive lowering of somatostatin release was caused by both vagal stimulation and A C h infusion (Hoist et al, 1983). More recently, Nishi et al (1987) observed a slight suppression of somatostatin output during electrical stimulation of the vagus in the perfused rat pancreas. Interestingly, infusion of atropine unmasked this inhibitory effect and a brisk stimulation of somatostatin secretion resulted. Clearly, the parasympathetic input to the islet S-cell appears to be complex and possibly species-dependent and requires further study. The primary role of the sympathetic nervous system in the control of carbohydrate metabolism is thought to be in mediating the increase in blood glucose 21 in response to hypoglycemic stress (Havel and Taborsky, 1989; Niijima, 1989). The effects of sympathetic nerve activation on insulin and glucagon secretion have generally been shown to be reciprocal, decreasing insulin and increasing glucagon concentrations and resulting in a rapid increase in blood glucose. Porte and Williams (1966) first demonstrated that infusion of norepinephrine in man produced a drop in circulating insulin levels. Stimulation of the splanchnic nerves, which carry the sympathetic input to the pancreas, was later shown to have a powerful inhibitory effect on insulin secretion both in vivo (Porte et al, 1973; Bloom and Edwards, 1975; Ahren et al, 1987) and in the perfused pancreas (Miller, 1975; Kurose et al, 1990) of several species. The repeated demonstration that this effect could be attenuated by the a-adrenergic antagonist phentolamine suggested that at least part of the inhibitory effects of the sympathetic nervous system was mediated by the release of norepinephrine from postganglionic sympathetic nerve terminals (Porte et al, 1973; Miller, 1975; Kurose et al, 1990). This idea was further supported in vitro by the administration of a-adrenergic agonists to isolated rat islets, which was shown to duplicate the effect of sympathetic nerve stimulation on insulin release, suggesting the involvement of /3-cell a-adrenergic receptors (Malaisse et al, 1967; Nakaki et al, 1980). Recent studies on the effects of various adrenergic agonists and antagonists on insulin release from isolated rat islets (Nakaki et al, 1980) and FACS-purified B-cells (Schuit and Pipeleers, 1986) have concluded that the a-adrenergic inhibitory effect was mediated by the a2-type adrenergic receptor. In 1967, Porte demonstrated the presence of a stimulatory effect of B-adrenergic receptor activation on insulin secretion during isoproterenol infusion in man. By administration of phentolamine during splanchnic nerve stimulation, it was further shown that it is possible to unmask this stimulatory effect (Miller, 1975; Roy et al, 1984). The physiological purpose of this B-stimulatory component remains obscure. Robertson and Porte (1973) showed that infusion of phentolamine or 22 propranolol in man induced a rise or fall, respectively, in plasma insulin levels. This suggested that sympathetic tone at rest may modulate insulin secretion via both adrenergic receptor subtypes. Yet in the rat, this mechanism may not be present since recent attempts to demonstrate a /3-adrenergic stimulatory effect in rat isolated islets (Lacey et al, 1990) and FACS-purified 0-cells (Schuit and Pipeleers, 1986) have failed. Additional mechanisms may also be involved in the sympathetic regulation of insulin secretion. Ahren et al (1987b) showed that norepinephrine infused directly into the pancreatic artery in dogs failed to reproduce the drop in pancreatic insulin output observed during sympathetic nerve stimulation. In addition, it was shown that concomitant a- and 0-adrenergic blockade by infusion of both propranolol and phentolamine attenuated but did not abolish the inhibitory effect of electrical stimulation of the sympathetic inputs to the pancreas in rat (Kurose et al, 1990) and dog (Dunning et al, 1988). These studies strongly suggested that nonadrenergic, possibly peptidergic mechanisms mediate at least part of the sympathetic inhibition of 0-cell secretion. A n increase in glucagon secretion in response to sympathetic nerve stimulation has been documented in several studies (Bloom and Edwards, 1975; Hoist et al, 1981c; Ahren et al, 1987; Kurose et al, 1990). The importance of this mechanism in the control of glucagon secretion was underlined by the demonstration that sympathetic nerve stimulation could enhance glucagon secretion even in the presence of glucose concentrations above normal physiological levels in calves (Bloom and Edwards, 1975) and in dogs (Girardier et al, 1976), although not in pigs (Hoist et al, 1981c). The effect appeared to be mediated via both a- and B-adrenergic receptors (Samols and Weir, 1979; Ahren et al, 1986b). Activation of the sympathetic nervous system and the subsequent stimulation of glucagon secretion to increase blood glucose levels has been suggested to be the predominant 23 mechanism in the response to hypoglycemic stress (Havel and Taborsky, 1989). However, combined a- and /3-adrenergic blockade has been shown to have no effect on the glucagon response to insulin-induced hypoglycemia in man (Rizza et al, 1979) and rat (Patel, 1984), suggesting either the presence of redundant (non-sympathetic) mechanisms in this response or the involvement of peptidergic pathways. Like insulin, islet somatostatin secretion has been shown to be suppressed by a-adrenergic agonists yet stimulated by /3-adrenergic agents (Samols and Weir, 1979). However, the predominant sympathetic effect on the 5-cell is clearly inhibitory, since in both rats (Kurose et al, 1990) and dogs (Dunning et al, 1988), electrical stimulation of the sympathetic nerves innervating the pancreas was shown to inhibit basal somatostatin secretion. In both species, this effect was only partially reversible by simultaneous a- and /3-adrenergic receptor blockade, again indicating the involvement of nonadrenergic mechanisms. Several candidate neuropeptides have been considered as mediators of the nonadrenergic component of the effect of the sympathetic nervous system on islet hormone secretion. Two of these peptides, N P Y and galanin, have been the focus of most study. N P Y has been localised to pancreatic nerves, often colocalized with norepinephrine (Pettersson et al, 1986; Dunning et al, 1987), and has been shown to be released into the pancreatic vein of the pig during electrical stimulation of the sympathetic nerves (Shiekh et al, 1988). Further, N P Y was found to have modest inhibitory effects on insulin secretion from the perfused pancreas and isolated islets in rat, suggesting a direct effect on the /8-cell (Moltz and McDonald, 1985), although no effect on islet hormone secretion was found in the dog (Dunning et al, 1987). Galanin, a 29 amino acid neuropeptide recently isolated by Tatemoto et al (1983), has also been demonstrated to be present in pancreatic nerves (Dunning et al, 1986). Numerous studies have shown that galanin possesses a potent inhibitory 24 effect on insulin secretion under a variety of conditions both in vivo (Dunning et al, 1986; Schneurer et al, 1990) and in vitro (Silvestre et al, 1987; Miralles et al, 1988; 1990). The presence of galanin receptors has been demonstrated in a hamster pancreatic /3-cell line, indicating that the neuropeptide acts directly on the j3-cell to mediate its inhibitory effect (Amiranoff et al, 1987). The effect of galanin on glucagon and somatostatin secretion has been less well studied; however, in the dog (Dunning et al, 1987) and in the perfused rat pancreas (Miralles et al, 1990) the peptide was shown to inhibit somatostatin and stimulate glucagon release under basal conditions and, in the rat, in the presence of various secretagogues. In further support of a role for galanin as a mediator of the nonadrenergic component of the sympathetic regulation of islet hormone secretion, Dunning and Taborsky (1989) showed that stimulation of the mixed pancreatic nerve, in which the postganglionic sympathetic input to the pancreas travels, resulted in an increase in the immunoreactive galanin concentration in the pancreatic vein. In addition, infusion of galanin into the pancreatic artery, at a concentration set to mimic that produced by nerve stimulation, produced a decrease in pancreatic insulin and somatostatin output similar to that produced by nerve stimulation, and a more modest increase in glucagon output. These studies led to the suggestion that galanin may be an integral component of the sympathetic neural pathway controlling islet secretion (Dunning and Taborsky, 1988). Studies in the rat (Silvestre et al, 1987; Miralles et al, 1988, 1990) and mouse (Lindskog and Ahren, 1987; 1988) have shown galanin to be a potent inhibitor of insulin secretion under a variety of conditions, suggesting that galanin interferes with an essential, common step in the insulin secretory mechanism (Silvestre et al, 1987; Ahren et al, 1988). In contrast, other studies using /3-cell lines (Amiranoff et al, 1988) or perfused rat pancreas (Yoshimura et al, 1989) have found that the inhibitory action of galanin was only observed in the presence of specific stimuli. 25 Thus, galanin suppressed GIP- but not carbamoylcholine-stimulated insulin secretion from Rinm5f /3-cells (Amiranoff et al, 1988), suggesting that the inhibitory effects of the neuropeptide may be more specific for GIP-stimulated insulin release. Alternatively, these results might be ascribed to differences in the responsiveness of the Rinm5f /3-cell line, which does not respond to glucose stimulation (Shimuzu et al, 1988), or to a diminished biological activity of porcine galanin on rat /3-cells. The possibility that the inhibitory effects of galanin are more marked in the presence of certain stimuli such as GIP is worthy of further examination, since insight into the mechanism of action and physiological function of these peptides could be gained. A third level of complexity in the regulation of islet hormone secretion is added by the consideration of the effects of other circulating hormones. In addition to the probable intra-islet effects of the islet hormones, numerous (neuro)endocrine substances including atrial natiuretic peptide, growth hormone, and growth hormone releasing factor have been shown to have various modulatory influences on the secretion of islet hormones. Whether the observed effects of many of these substances have physiological relevance is unknown. The focus by far of greatest study in the hormonal regulation of islet function has been the role of the gastrointestinal hormones. This pathway, wherein hormonal signals from the gut are released in response to the products of food ingestion in the intestinal lumen and travel to the pancreas via the circulation to influence islet hormone secretion, was termed the "enteroinsular axis" by Unger and Eisentraut in 1969. The idea that a humoral substance from the intestine might be involved in the regulation of carbohydrate metabolism arose soon after Bayliss and Starling (1902) discovered the first intestinal hormone, secretin. Moore et al (1906) proposed that the gut also contained a "chemical excitant for the internal secretion of the pancreas" and attempted to treat diabetes mellitus by injection of intestinal extracts. Zunz and LaBarre (1929) later prepared an intestinal extract, free of 26 secretin activity, which produced hypoglycemia in dogs. This result was attributed to the effects of a humoral substance in the extract which enhanced the secretion of the endocrine pancreas. Labarre (1932) termed this intestinal factor "incretin". The isolation of insulin, the internal secretion of the pancreas (Banting and Best, 1922), and its subsequent availability in a clinically adequate, purer form (Banting et al, 1922) enabled the effective treatment of patients with diabetes mellitus. This discovery, coupled with the studies of Loew et al (1940) which questioned the existence of the proposed intestinal factor, slowed the search for incretin. Interest in the existence of an incretin factor was not renewed until the development of RIA (Berson and Yalow, 1959), which allowed the measurement of insulin concentrations in plasma and more meaningful studies into the regulation of insulin secretion. It was soon shown that intravenous glucose was a weaker stimulus for insulin secretion than glucose given either orally (Elrick et al, 1964) or intrajejunally (Mclntyre et al, 1964). These studies suggested that a hormone, released in response to glucose in the intestinal lumen, acted on the pancreas to enhance glucose-stimulated insulin secretion, restoring interest in the search for incretin factors. Perley and Kipnis (1967) later suggested that approximately half of the insulin response to oral glucose was attributable to the effects of intestinal hormones. Several of the known gastrointestinal hormones have been shown to possess insulinotropic activity and their potential role as incretins has recently been evaluated (Brown, 1988; Creutzfeldt and Ebert, 1988). Creutzfeldt (1979) defined the criteria for acceptance as an incretin: firstly, the hormone must be secreted in response to nutrients, particularly carbohydrates, in the intestine; and secondly, physiological concentrations of the hormone must stimulate insulin secretion under hyperglycemic conditions. As will be discussed, these criteria exclude most of the known gastrointestinal hormones with the exception of GIP and the newest incretin 27 candidate, GLP-I (7-36-NH2). As recently as 1988, Creutzfeldt and Ebert stated that GIP was the "strongest incretin candidate". However, rejection of a gut hormone as an incretin candidate by this definition does not necessarily preclude a physiological role for the substance in the regulation of islet hormone secretion. Thus, numerous studies have suggested that certain enteric hormones are released in response to intestinal stimuli other than glucose and that some exert effects on the secretion of islet peptides other than insulin. In addition, some gut hormones potentiate the insulinotropic action of non-glucose fuels or interact synergistically at the islet with other neural and hormonal pathways originating in the gut. These additional elements led Creutzfeldt (1979) to suggest that the term "enteroinsular axis" be expanded to encompass all signals - nutrient, neural and hormonal - from the gut to the islet and their interactions in the regulation of insulin, glucagon, somatostatin and pancreatic polypeptide release. A n insulinotropic effect of secretin infusion in man was first demonstrated by Dupre (1964). Intraduodenal administration of acid, the primary stimulus for secretin secretion, was later shown to induce an increase in insulin levels under hyperglycemic conditions (Fahrenkrug et al, 1978). However, this effect has since been attributed to GIP, since GIP antiserum was shown to abolish the insulin response to intestinal infusion of acid in rats (Ebert et al, 1979). The studies of Bloom and colleagues, who showed that ingestion of glucose (Bloom, 1974) or a mixed meal (Bloom et al, 1975) was without effect on circulating secretin levels, strongly suggested that the hormone be excluded from consideration as an incretin. Still, secretin has been shown to potentiate glucose-induced insulin secretion both in vivo (Dupre et al, 1969) and in vitro using perifused mouse islets (Kofod et al, 1986). Yet the concentrations of secretin required to stimulate insulin secretion in those and other studies were at least 10 fold the maximum observed concentration of circulating immunoreactive secretin following intraduodenal administration of acid 28 (Chey et al, 1978). Further, in pigs, Lindkaer Jensen et al (1978) demonstrated that the concentration of secretin needed to produce an insulin response was approximately 100 times the concentration of secretin required to stimulate pancreatic bicarbonate output. These studies suggested that the insulinotropic effect of secretin was not physiologically relevant. Indeed, nanomolar concentrations of secretin were found to have no significant effect on insulin secretion from the perfused rat pancreas, although a slight, monophasic increase was consistently observed (Pederson and Brown, 1979). Lerner and Porte (1970) had earlier demonstrated a weak, monophasic effect of secretin in man, and suggested that only the first phase of insulin secretion was affected by the hormone. Several authors have thus concluded that secretin is not an incretin and probably does not play a significant role in the regulation of insulin secretion (Hoist et al, 1980; Brown, 1988; Creutzfeldt and Ebert, 1988). Although gastrin has been shown to be released in response to ingestion of a meal (Feldman et al, 1978), the major stimuli for gastrin secretion are the products of protein digestion (Elwin, 1974), whereas glucose has been shown to be a relatively weak stimulus for gastrin secretion (Rehfeld and Stadil, 1975). Although gastrin has been shown to stimulate insulin secretion (Dupre et al, 1969; Creutzfeldt et al, 1970), its insulinotropic effect was weak and transitory, and the concentration required to initiate /3-cell secretion far exceeded the plasma levels achieved following meal ingestion (Rehfeld and Stadil, 1975). Therefore, like secretin, gastrin has been excluded from the list of incretin candidates, and is most likely not involved in the regulation of insulin secretion. Dupre and Beck (1966) first demonstrated the presence of insulinotropic activity in a crude preparation of C C K . However, since GIP was later isolated from this preparation (Brown et al 1969; 1970), an insulinotropic effect of C C K was not verified until the preparation of GIP-free C C K (Pederson and Brown, 1979) and the 29 use of the synthetic C-terminal octapeptide form of C C K (CCK-8) (Hermansen, 1984). CCK-induced insulin release has been shown to be glucose-dependent in isolated islets (Verspohl et al, 1986a; Verspohl and Ammon, 1987) and in the perfused pancreas (Pederson and Brown, 1979), where it was found to be approximately 50% as potent as GIP on a molar basis. Several additional molecular forms of the peptide, including CCK-39, CCK-33 and CCK-4 have also been shown to stimulate insulin release in certain species (Okabayashi et al, 1989). Further, B-cell receptors specific for C C K have been demonstrated by autoradiography (Sakamoto et al, 1985) and radioreceptor binding studies (Verspohl et al, 1986a). Despite the unequivocal evidence for a stimulatory effect of C C K on the B-cell, whether the hormone contributes to the postprandial insulin response remains controversial (Brown, 1988; Creutzfeldt and Ebert, 1988; Okabayashi et al, 1989). The study of this problem has been limited by the lack of specific and sensitive assays for the measurement of circulating C C K concentrations until relatively recently (Jansen and Lamers, 1983; Cantor, 1986). Measurement of plasma C C K concentrations by a sensitive bioassay found that circulating C C K levels increased after ingestion of fat, glucose or amino acids, although the C C K response to oral glucose was weak and transient (Liddle et al, 1985). Other studies showed no effect of oral glucose administration on C C K measured by either bioassay (Wang and Grossman, 1951) or R I A (Reimers et al, 1988). Using RIA, it was recently found that plasma C C K levels in man following ingestion of a mixed meal increased from resting values of 1-3 p M to a peak value of 5-12 p M (Jansen and Lamers, 1983; Cantor, 1986; Hildebrand et al, 1991). By comparison, the minimum concentration of CCK-8 required to stimulate insulin secretion in the presence of hyperglycemia was 100 p M in the perfused rat pancreas (Otsuki et al, 1979) and 1 n M in isolated rat islets (Verspohl and Ammon, 1987). In man, infusion of exogenous CCK-8 at a rate which achieved circulating levels of 8-12 p M was without effect on insulin 30 secretion under hyperglycemic conditions (Rushkakoff et al, 1987; Reimers et al, 1988), while a similar concentration of CCK-8 was capable of stimulating insulin release during concomitant infusion of arginine or mixed amino acids (Rushkakoff et al, 1987; Hildebrand et al, 1991). Thus, while C C K has been shown not to contribute to the incretin effect during carbohydrate ingestion, it may contribute to the insulin response to a mixed meal, leading Rushkakoff et al (1987) to suggest that the incretin concept be extended to include those enteric hormones which are released in response to non-glucose nutrients and which are insulinotropic in the presence of hyperaminoacidemia. The development of specific C C K receptor antagonists has provided an additional tool for the study of the role of C C K in the insulin response to a meal. In rats, the C C K antagonist L364,718 attenuated the insulin increase observed after intraduodenal infusion of a glucose/casein mixture, but not during infusion of glucose alone. This confirmed the idea that C C K was not involved in the insulin response to glucose ingestion and supported the contention that the peptide does help mediate the increase in insulin secretion during protein ingestion. Another C C K antagonist, loxiglumide, potently suppressed the pancreatic polypeptide increase observed after ingestion of a mixed liquid meal or intestinal perfusion of an amino acid solution in man, but was without effect on the insulin response to these stimuli (Hildebrand et al, 1991). The results suggested that C C K is not involved in the postprandial insulin response, at least in man, although it may play a physiological role in the enteroinsular axis as a modulator of the secretion of other islet hormones. Such a role was supported by previous demonstrations that C C K stimulated glucagon and somatostatin secretion in vitro in various species (Ipp et al, 1977; Szecowka et al, 1982b; Hermansen, 1984; Verspohl and Ammon, 1987). Sequencing of the preproglucagon gene by two separate groups (Bell et al, 1983; Lopez et al, 1983) revealed that the precursor peptide contained, in addition 31 to glucagon and glicentin, the sequence of two glucagon-like peptides (GLP-I and GLP-II). Mosjov et al (1986) showed that post-translational processing of the preprohormone differed in pancreas and intestine: GLP-I was found to occur in both pancreas and gut in either a 37 amino acid form [GLP-I (1-37)] or in a C-terminal 30 residue form [GLP-I (7-36-NH2)]; GLP-II was found only in the intestine. It was soon shown that GLP-I (7-36-NH2), but not GLP-I (1-37) or G L P -II, was a potent stimulator of insulin secretion (Hoist et al, 1987; Kreymann et al, 1987; Mosjov et al, 1987). Increasing evidence now suggests that this newly discovered peptide warrants serious consideration as an incretin. Radioimmunoassay and chromatography of tissue extracts showed that GLP-I (7-36-N H 2 ) was present in the small intestine of man (Kreymann et al, 1987) and pigs (Hoist et al, 1987). Kreymann et al (1987) observed a sustained increase in plasma levels of GLP-I (7-36-NH2) following an oral glucose load or ingestion of a mixed meal in man. Further, under fasting conditions, infusion of the peptide at a dose which resulted in plasma concentrations similar to those observed postprandially produced a pronounced rise in circulating insulin levels. Finally, a potent glucose-dependent, stimulatory effect on the j3-cell has been demonstrated in the perfused rat pancreas (Weir et al, 1989; Suzuki et al, 1989). Thus, according to the few studies performed to date, GLP-I (7-36-NH2) appears to fulfill the criteria for a physiological incretin, although confirmation of this role awaits further investigation. Interestingly, of all the peptides derived from the preproglucagon gene, the sequence of GLP-I (7-36-NH2) shows the greatest homology to GIP. The peptide may also be involved in the regulation of the release of other islet hormones, since it was found to stimulate somatostatin secretion and inhibit glucagon secretion from the perfused pancreas of both rats (Suzuki et al, 1989) and pigs (Orskov et al, 1988). Rabinovitch and Dupre (1972) found that the insulinotropic action of a 10% pure C C K preparation could be removed by further purification of the hormone. 32 Since this resembled the loss of acid inhibitory activity previously observed by Brown and Pederson (1970) in the purification of GIP from C C K , Dupre et al (1973) reasoned that GIP might have insulin-releasing capabilities. They showed that concomitant infusion of porcine GIP (1.0 /xg/min) and glucose (0.5 g/min) in man produced a pronounced increase in circulating insulin levels and improved glucose tolerance compared to glucose infusion alone. The insulin response was sustained for the duration of the GIP infusion and was not observed in fasted (euglycemic) individuals. This glucose-dependent, insulinotropic action of GIP was soon confirmed by experiments in vivo in the dog (Pederson et al, 1975) and man (Elahi et al, 1979) and in the perfused rat pancreas (Pederson and Brown, 1976). Elahi et al (1979) used the glucose clamp technique to demonstrate that mild hyperglycemia (at least 1.4 m M above basal levels) was sufficient to initiate the insulinotropic action of GIP. This corresponded with the threshold glucose concentration of approximately 5.5 m M that had been observed in the perfused rat pancreas (Pederson and Brown, 1976). In that preparation, the maximum potentiating action of glucose on GIP-stimulated insulin release was observed at approximately 16 m M (Brown, 1982). In addition, the stimulatory effect of GIP on the /3-cell was found to be concentration-dependent. In the presence of 8.9 m M glucose, a concentration of 1 ng/ml (200 pM) GIP, when administered as a "square-wave" (rapid-onset) stimulus, was sufficient to induce insulin secretion (Pederson and Brown, 1976). However, Brown et al (1980) have suggested that the rapid onset of an infusion of GIP did not represent a physiological stimulus. Thus, in the presence of 17.8 m M , when GIP was presented to the perfused rat pancreas as a linear concentration gradient from 0 to 200 p M , the insulinotropic effect of GIP was initiated at concentrations as low as 70 p M (Pederson et al, 1982). Development of an RIA for the measurement of immunoreactive-GIP (IR-GIP) concentrations in plasma by Kuzio et al (1974) allowed the assessment of GIP 33 as a potential incretin. Serum IR-GIP concentrations in fasted individuals were determined to be approximately 50 p M , rising to about 250 p M following ingestion of a mixed meal. A n increase in circulating levels of IR-GIP was also observed in response to an oral glucose load in normal man (Cataland et al, 1974; Cleator and Gourlay, 1975; Andersen et al, 1978), dog (Pederson et al, 1975) and rat (Pederson et al, 1982). No change in IR-GIP concentration was observed when glucose was administered intravenously (Cataland et al, 1974; Andersen et al, 1978). The peak levels of plasma IR-GIP achieved following carbohydrate ingestion in these studies (approximately 250 pM) were comparable to those used to induce an insulin response in other studies, strongly suggesting that the hormone was an incretin. Several other nutrient secretagogues for GIP release have been identified. In addition to glucose, other carbohydrates such as galactose stimulated IR-GIP release from the intestine. Morgan (1979) suggested that sugar-stimulated IR-GIP release required the intestinal active transport of monosaccharides and this has been supported by other studies (Creutzfeldt and Ebert, 1977) . Ingestion of fat has also been shown to be a potent stimulus for insulin secretion in man (Brown, 1974; Cleator and Gourlay, 1975; Falko et al, 1975; Ross and Dupre, 1978) and dogs (Pederson et al, 1975). Interestingly, the IR-GIP response to fat ingestion was not accompanied by an increase in circulating insulin unless intravenous glucose was administered as well (Brown, 1974; Cleator and Gourlay, 1975; Ross and Dupre, 1978). This demonstrated that the glucose-dependency of GIP-stimulated insulin secretion provided an important safeguard against hypoglycemia by preventing the inappropriate stimulation of insulin release during a high fat, low carbohydrate meal. Whether proteins and amino acids are physiological secretagogues for GIP is uncertain. Although ingestion of a high protein meal was shown to have no influence on circulating IR-GIP concentrations (Brown, 1974; Cleator and Gourlay, 1975), intraduodenal administration of a mixture of basic amino acids was found to 34 stimulate IR-GIP release (Thomas et al, 1978). In contrast, a mixture of aromatic amino acids, known to be a potent stimulus for C C K release, had no effect on IR-GIP levels. Intraduodenal acidification has also been shown to be a stimulus for IR-GIP secretion (Ebert et al, 1979). Since the first GIP RIA was developed by Kuzio et al (1974), measurement of fasting and stimulated plasma IR-GIP levels in man in different laboratories have produced values that varied widely, although the relative increases in IR-GIP concentrations over basal in response to nutrients has been similar (Kuzio et al, 1974; Lauritsen and Moody, 1978; Morgan, 1979; Sarson et al, 1980; Burhol et al, 1980). For example, Sarson et al (1984) found that after oral glucose plasma IR-GIP increased from only 10 p M to 40 p M , a fourfold increase, yet absolute values of IR-GIP were approximately five times lower than what had been observed in other laboratories. When porcine GIP was infused in human volunteers to achieve circulating IR-GIP levels of 40 p M , in association with induced hyperglycemia, no insulinotropic effect was observed, leading these investigators to speculate that the reported effects of GIP were pharmacological and that GIP was not an important incretin. Jorde et al (1983b) attempted to resolve this disparity by measuring fasting and postprandial IR-GIP levels using seven different antisera. Considerable differences were observed; values ranged from 12-92 p M (fasting) to 35-235 p M (postprandial). Since the porcine peptide has invariably been used as standard and to raise antisera for RIAs to measure IR-GIP in man, these results could be attributed to the two amino acid differences between human and porcine GIP. The recent sequencing of human GIP and its successful production by both chemical synthesis (Yajima et al, 1985) and recombinant D N A techniques (Chow et al, 1990) will allow the proper evaluation of IR-GIP concentrations in man and its effectiveness as a stimulator of insulin secretion. Kreymann et al (1987) recently measured IR-GIP using human GIP as standards in the R IA and showed that the 35 circulating levels of IR-GIP attained following oral glucose or a mixed meal were sufficient, when mimicked by exogenous infusion of human GIP, to stimulate insulin secretion. That study further demonstrated that exogenous GLP-I (7-36-NH2) was more potent than GIP on a molar basis, but that after a mixed meal or oral glucose challenge, circulating IR-GIP concentrations attained a level several fold higher than GLP-I (7-36-NH2). At present no specific GIP antagonists have been developed to investigate the physiological role of GIP in the postprandial insulin response in man, but antisera to GIP have been produced which may suppress the biological activity of the hormone. In attempting to quantify the contribution of GIP to the incretin effect, Creutzfeldt and coworkers tested the effect of GIP antibody injection on the insulin response to various known intestinal stimuli for IR-GIP release in vivo in rats. During intraduodenal acid administration with intravenous glucose infusion, GIP antibodies strongly inhibited the initial increase in insulin levels, although after 20-30 min circulating insulin concentrations were the same in control and test rats (Ebert et al, 1979). Similarly, GIP antiserum only blocked the early phases of the insulin reponse to oral glucose (Ebert and Creutzfeldt, 1982). These results correlated well with the recent studies of Kreymann et al (1987), who found that when human volunteers consumed a test breakfast, plasma IR-GIP levels rose more rapidly than GLP-I (7-36-NH2), implying that GIP mediated the early insulin response to a meal while GLP-I (7-36-NH2) was more important in the late response. Little is known about the mechanism by which GIP stimulates insulin release from the j3-cell. How GIP interacts with other secretagogues at the /3-cell or indeed, whether or not GIP acts directly on the /3-cell is not clear. Investigation into the interactions of GIP has been hindered by the insensitivity of in vitro models of insulin secretion to GIP and the lack of availability of homogeneous preparations of 36 /3-cells until recently. Using a hamster pancreatic 0-cell line (In III), Amiranoff et al (1984) showed that GIP produced a concentration-dependent increase in c A M P content in the cells which was parallelled by stimulation of insulin release. The idea that GIP acted via stimulation of 0-cell adenylate cyclase was later substantiated by studies with human insulinoma tissue in vitro (Maletti et al, 1987) and cultured rat islets (Siegel and Creutzfeldt, 1985). However, the observed increase in c A M P in these studies was modest even in the presence of the phosphodiesterase inhibitor, 3-isobutyl-l-methylxanthine (IBMX); therefore, the possibility that GIP acted at least in part through other pathways must also be considered. In this regard, Lardinois et al (1990) recently showed that GIP-stimulated insulin release from neonatal rat pancreatic cell cultures could be suppressed by inhibitors of the intracellular enzymes lipoxygenase and cyclooxygenase. It has been further suggested that /3-cell calcium transport is an essential part of the mechanism of GIP-stimulated insulin secretion (Brunicardi et al, 1986a). It should also be noted that all studies to date have used either heterogeneous cell preparations or neoplastic B-cell lines which may not accurately depict the situation in the 0-cell in situ. Delineation of the mechanisms of GIP-stimulated insulin secretion will require the use of homogeneous preparations such as can be obtained from cell sorting, which was recently used to show that the structurally-related hormone glucagon was a potent stimulator of c A M P production in the yS-cell (Schuit and Pipeleers, 1985). The exact mechanism underlying the glucose-dependency of the insulin response to GIP is also unclear. D-glyceraldehyde, an intermediate in the glycolytic pathway, was able to potentiate the insulinotropic action of GIP in the absence of glucose from the perfused rat pancreas (Dahl, 1983). Mannoheptulose, which blocks glycolysis, abolished GIP-stimulated insulin secretion in that preparation (Muller et al, 1982). These studies strongly suggested that glucose metabolism was a prerequisite for GIP-stimulated insulin release to occur. Other insulin 37 secretagogues such as C C K (Verspohl and Ammon, 1987; Zawalich et al, 1987), A C h (Garcia et al, 1988; Verchere et al, 1991) and GLP-I (7-36-NH2) (Weir et al, 1989; Komatsu et al, 1989) have also exhibited glucose-dependency. A common mechanism whereby glucose metabolism in the /3-cell permits stimulation by non-glucose secretagogues has been proposed (Rasmussen et al, 1990; Zawalich and Rasmussen, 1990). This model suggests that if the /3-cell is exposed to moderate increases in glucose concentration, increased /0-cell glucose metabolism occurs, resulting in a partial depolarization of the /3-cell membrane via the closing of A T P -sensitive K + channels. In this state, insulin secretion is only modestly increased above basal levels but the /3-cell is highly sensitive to further depolarization, secondary to increases in phosphoinositol metabolism or c A M P production. In addition to the glucose-dependency of the insulinotropic action of GIP, several studies have suggested that GIP-stimulated insulin release is subject to considerable modulation by neural pathways and other circulating factors. Using the perfused rat pancreas, Pederson and Brown (1978) showed that the insulinotropic effect of GIP was attenuated by 20 m M arginine and this was later supported by an in vivo study in man (Elahi et al, 1982). These studies suggested that GIP and arginine acted on the /3-cell via a similar mechanism. However, in the presence of a glucose concentration below the threshold for GIP-stimulated insulin release (2.7 mM), arginine (5 or 10 mM) was able to potentiate the insulinotropic action of GIP (Pederson and Brown, 1978). This indicated that the presence of certain nutrients other than glucose might be sufficient for GIP-stimulated insulin secretion to occur. A potentiating interaction between GIP and both A C h (Verchere et al, 1991) and C C K (Sandberg et al, 1988) has also been demonstrated in the perfused rat pancreas, while the effects of concomitant infusion of GIP and GLP-I (7-36-NH2) were only additive (Fehmann et al, 1989) and VIP was found to suppress GIP-stimulated insulin secretion (Sandberg et al, 1988). Rasmussen and 38 his colleagues have suggested that such interactions between various insulin secretagogues occur not at the receptor level but rather at the level of signal transduction within the /3-cell (Zawalich and Rasmussen, 1990; Rasmussen et al, 1990). Thus, agonists that act via distinct intracellular pathways tend to potentiate the action of each other on the /3-cell, while those that act via the same intracellular pathway tend to have additive stimulatory effects. If GIP, as some studies have suggested, stimulates c A M P production in the 0-cell, then the presence of /3-cell secretagogues which act via the stimulation of /3-cell phosphoinositide metabolism (e.g. A C h and C C K ) would potentiate the insulinotropic action of GIP. It follows that other activators of adenylate cyclase, such as GLP-I (7-36-NH2) and VIP, would not potentiate the stimulatory effect of GIP. To properly examine the action of GIP at the level of the /3-cell and its interactions with glucose and other secretagogues, in vitro preparations of isolated islets or pure /3-cells are desirable. The perfused pancreas is sensitive to physiological concentrations of GIP (Pederson and Brown, 1976) and remains an acceptable model for the investigation of the actions of GIP. However, in that preparation the intrinsic innervation and vasculature of the pancreas is intact and must be considered as a possible influence in the interpretation of results (Pipeleers, 1984). In isolated islets, cell-cell contacts are maintained while the islets are removed from neural and vascular influences. Yet isolated islets have rarely been used to investigate the insulinotropic action of GIP. This has largely been due to the insensitivity of this preparation to concentrations of GIP thought to be in the physiological range (50 p M to 250 pM) (Schauder et al, 1975; Schafer and Schatz, 1979; Szecowka et al, 1982a). In 1980, Brown et al proposed four reasons to explain the weak insulin response to GIP observed in the islets (Table 2). While none of these reasons have yet been totally excluded, Siegel and Creutzfeldt (1985) recently demonstrated that isolated islets regained their sensitivity to GIP 39 T A B L E 2: Possible Reasons for the Insensitivity of Isolated Islet Preparations to  the Insulinotropic Action of Gastric Inhibitory Polypeptide 1. Enzymatic damage to /3-cell GIP receptors occurs during collagenase isolation of the islets. 2. Substances inhibitory for insulin release (e.g. somatostatin) accumulate in the incubation medium. 3. Secretagogues are delivered to the /3-cell via diffusion in isolated islets vs. directly (via the circulation) in the perfused pancreas. 4. A substance or pathway necessary for GIP-stimulated insulin release is removed during the isolation procedure. Adapted from Brown et al, 1980. following 24 h culture, suggesting that GIP receptors were damaged during the isolation procedure and were regenerated in culture. This had been previously proposed by Fujimoto et al (1978), who demonstrated that neonatal rat pancreatic cultures were sensitive to GIP concentrations as low as 0.2 n M . To eliminate possible intra-islet influences in the study of the regulation of insulin secretion, two types of homogeneous /3-cell preparations have been employed: /3-cell lines and FACS-purified /3-cells. Neoplastic /3-cell lines have been used more frequently because a large number of cells can be obtained relatively easily, thus making them suitable for examining numerous experimental conditions, purifying cellular proteins, or studying receptors. However, these cell lines may alter their response characteristics after numerous passages or may differ greatly from normal /3-cells in their response to certain secretagogues. For example, the Rinm5f cell, a frequently used /3-cell line, was found not to respond to glucose possibly due to a deficiency of the glycolytic enzyme glucokinase (Shimizu, 1988). GIP has been shown to induce an insulin response in two cell lines: hamster In III cells (Amiranoff et al, 1984) and Rinm5f cells (Amiranoff et al, 1988). A recently developed /8-cell line, the /3TC3 cell, responded normally to glucose as well as stimulators of the c A M P and phosphoinositol intracellular pathways, suggesting that 40 it closely resembled normal /3-cells in function and could be used as a pure /3-cell model of physiological insulin secretion (D'Ambra et al, 1990). The use of the F A C S to obtain highly pure preparations of /3-cells from normal tissue has provided a new and powerful technique for examining the effects of insulin secretagogues at the level of the /3-cell (Pipeleers et al, 1985a). After culture, this preparation has been shown to be sensitive to concentrations of glucagon as low as 1 n M (Pipeleers et al, 1985c). Whether or not GIP stimulates insulin secretion from this preparation has not been studied. The demonstration of binding sites for GIP in the In III 0-cell line (Amiranoff et al, 1984) and in a human insulinoma (Maletti et al, 1987) suggested that GIP acted directly on the pancreatic /3-cell via specific receptors to stimulate insulin release. Binding of 1 2 5 I -GIP in these studies was found to be saturable and could not be displaced by peptides structurally related to GIP. Both a high affinity (Kd = 7 n M ; 3000 binding sites/cell) and a low affinity (K<i = 800 n M ; 1.5 X 105 binding sites/cell) binding site for GIP was identified. The /3-cell membrane protein that bound GIP in these cells was partially characterized as a protein monomer of approximate molecular weight 59000 (Amiranoff et al, 1986). Still, decisive proof that GIP acts via /3-cell receptors requires the demonstration of such a receptor on nontumor /3-cells obtained from normal animals, since the density of GIP binding sites on neoplastic /3-cells might not accurately reflect the physiological distribution of these receptors in situ. To date, no such data has been presented. Fujimoto et al (1978) described the visualisation of GIP binding sites in neonatal pancreatic cell cultures by immunoperoxidase staining using anti-GIP serum, although only a small subpopulation of the cells were stained. Brown et al (1989) recently suggested two possible explanations for the lack of success in demonstrating GIP receptors in normal /3-cell tissue. First, as was described, the GIP receptor may be particularly susceptible to enzymatic damage by collagenase during islet isolation. It follows 41 that this procedure may make the presence of GIP receptors difficult to demonstrate. Second, iodination of GIP probably results in a heterogeneous population of iodinated peptides. This would include multi-iodinated forms of the peptide, since GIP contains two tyrosine residues, at positions 1 and 10. In addition, methionine (position 14) and tryptophan (positions 25 and 36) have been shown in other peptides to be susceptible to oxidative damage during the iodination procedure (Lambert et al, 1982). If alteration of these residues occurred at sites essential for the biological action of GIP, then these forms of 1 2 5 I -GIP might not bind to the receptor. Although the site of the GIP molecule responsible for insulinotropic activity is not fully known, synthetic GIP 1-30 has been shown to stimulate insulin secretion from the perfused rat pancreas as potently as GIP 1-42 (molar basis) suggesting that the C-terminal part of the molecule was not important (Pederson et al, 1990). Which N-terminal residues are essential for receptor binding is unclear, since bovine GIP 4-42 (Maletti et al, 1986), but not porcine GIP 3-42 (Dahl, 1983), retained full insulinotropic activity in the perfused rat pancreas. To circumvent these problems, several approaches are possible. First, if short-term culture of islet tissue does indeed replenish /3-cell GIP receptors following collagenase digestion, then cultured islets or /3-cells should be a useful preparation to investigate the presence of the GIP receptor. Second, purification of a homogeneous, biologically active form of 1 2 5 I -GIP could provide a suitable probe for the GIP receptor. In support of this idea, Maletti et al (1984) used reverse-phase high-performance liquid chromatography (HPLC) to produce [mono- 1 2 5I-Tyr 1 0 ]GIP which was insulinotropic in the perfused pancreas and was successfully used to demonstrate GIP receptor binding in tumor cells. A unique, recently developed approach in studying peptide receptors takes advantage of the high affinity binding constant (10"15 M) between the glycoprotein avidin and the vitamin biotin. Biotinylated peptides, produced by coupling biotin to 4 2 certain amino acid residues, have increasingly been used as probes for various peptide receptors (Wilchek and Bayer, 1988). Binding of the biotinylated peptide to the receptor can be visualized using an avidin-conjugated probe such as streptavidin-peroxidase or streptavidin-fluorescein. For example, this technique was used to demonstrate the presence of receptors for gonadotropin releasing hormone on pituitary monolayer cell cultures (Childs et al, 1983). In certain peptides, biotinylation may produce a more biologically active probe than iodination, since damage to the molecule by oxidation is avoided, and in some peptides the residues that are coupled to biotin may be outside of the active site of the molecule. In the case of GIP, for example, biotinylation of e-amine groups (lysine residues) would not be expected to alter any amino acid residues in the hydrophobic region of the peptide (amino acids 6-14 and 19-27), which has been suggested to be of possible importance in receptor binding (Brown et al, 1989). While the insulinotropic action of GIP has been unequivocally demonstrated, whether GIP has a role in the regulation of the secretion of other islet hormones has not yet been elucidated. A stimulatory effect of GIP on glucagon secretion has been suggested by the results of several studies. When infused at a concentration of 1 nM, GIP induced an increase in glucagon secretion from the perfused pancreas of rat (Pederson and Brown, 1978; Suzuki et al, 1990), dog (Adrian et al, 1978) and man (Brunicardi et al, 1990). This stimulatory effect was observed only in the presence of glucose concentrations equal to or below a threshold level of approximately 5.5 m M in the perfused rat pancreas, which coincided with the threshold concentration above which glucose potentiated the insulinotropic action of GIP in that preparation (Pederson and Brown, 1978; Suzuki et al, 1990). In addition, it was found that when arginine was added to the perfusate, the glucagon response to GIP became biphasic and was greatly enhanced (Pederson and Brown, 1978). Although the effect of GIP on glucagon secretion from isolated rat islets has 43 not yet been studied, in vitro preparations of neonatal rat pancreatic cell cultures (Fujimoto et al, 1978) and isolated mouse islets (Bailey et al, 1990) have both been shown to secrete glucagon in response to GIP. Interestingly, the stimulatory effect of GIP in those experiments was not diminished by increasing the glucose concentration to >16 m M . Together, these studies indicated that GIP may modulate a-cell secretion, although the influence of glucose and other amino acids on this mechanism is not fully understood. Further, the existence of this action of GIP remains to be verified in man, since infusion of porcine GIP failed to stimulate glucagon secretion in human volunteers under either euglycemic or hyperglycemic conditions (Elahi et al, 1979). The effects of GIP on pancreatic somatostatin secretion are even less clear. Ipp et al (1977) showed that, in the presence of 5 m M glucose, GIP could stimulate somatostatin secretion from the perfused dog pancreas, although the concentration of GIP used (approximately 12 nM) far exceeded the physiological range for the hormone. In the perfused rat pancreas, no effect of GIP on somatostatin release was observed unless GIP concentrations of at least 10 n M were used (Schmid et al, 1990); however, Dahl (1983) showed that 1 n M GIP was capable of stimulating somatostatin release if the phosphodiesterase inhibitor theophylline was added to the perfusate. This contrasted markedly with the situation in the perfused stomach, where GIP has been shown to potently stimulate gastric somatostatin secretion (Mcintosh et al, 1981). Bailey et al (1990) recently reported that 10 n M GIP could stimulate somatostatin secretion from isolated mouse islets in the presence of either 5.6 or 16.7 m M glucose. Thus, GIP may be able to stimulate the release of islet somatostatin under certain conditions and further study is warranted. It should also be noted that a stimulatory effect of GIP on pancreatic polypeptide secretion has been reported in isolated mouse islets (Bailey et al, 1990), the perfused dog pancreas (Adrian et al, 1978) and in human volunteers (Amland et al, 1985a). In 44 mouse islets, the effect was observed in the presence of 5.6 but not 16.7 m M glucose. The potent insulinotropic action of GIP has led to speculation that the hormone could be involved in the pathogenesis of certain diseases of carbohydrate metabolism. To date, however, no such role has been clearly defined (Creutzfeldt and Ebert, 1988; Krarup, 1988; Brown et al, 1989). In Type 1 diabetes, or insulin-dependent diabetes mellitus (IDDM), IR-GIP secretion in response to a test meal was found to be reduced in newly diagnosed diabetics although this response normalized following conventional insulin therapy (Krarup et al, 1985). As expected, the insulin response to exogenous GIP infusion in I D D M patients was reduced but this was attributed to the loss of functional /3-cell mass in this disease and not to a specific loss of sensitivity to GIP (Krarup et al, 1987). Type II diabetes (NIDDM) is characterized by normal or increased circulating insulin levels when compared to non-diabetic individuals, although insulin secretion is clearly impaired when corrected for the degree of hyperglycemia (Porte, 1991). The pathogenesis of N I D D M remains poorly understood but is probably multifactorial in etiology, involving both insulin resistance in peripheral tissues and defective /3-cell secretion. Since the early insulin response to oral glucose or a mixed meal has clearly been shown to be diminished in N I D D M (Bagdade et al, 1967; Nauck et al, 1986; Elahi et al, 1984; Jones et al, 1989), a defect in the enteroinsular axis has been implicated. Several studies have indicated that impaired GIP secretion in N I D D M could not account for this observation, since plasma IR-GIP concentrations were usually found to be greater than non-diabetic individuals following meal or glucose ingestion, although normal IR-GIP levels were sometimes observed in N I D D M patients (Ross et al, 1977; Elahi et al, 1984; Mazzaferri et al, 1985; Nauck et al, 1986; Krarup, 1988; Jones et al, 1989). The reason for the GIP hypersecretion in N I D D M has not been determined but might be a compensatory attempt to enhance /3-cell secretion. By comparing the insulin 45 and IR-GIP responses to oral glucose in diabetic and normal individuals during a hyperglycemic clamp, Elahi et al (1984) determined the /3-cell sensitivity to GIP and suggested that it was unchanged in N I D D M . Infusion of porcine GIP into N I D D M patients has been shown to induce an immediate increase in circulating insulin levels, although the insulin response was impaired compared to normal subjects (Jones et al, 1987; Krarup et al, 1987). It was not clear whether the diminished response to GIP was due to a generalized impairment in /3-cell function or a defect specific to GIP-stimulated insulin release. These studies have not supported the hypothesis that an alteration in GIP secretion or action is involved in the delayed postprandial insulin response in N I D D M . Hyperinsulinemia has been a consistent observation in obese man and in animal models of obesity in both the fasted and the postprandial state (Karam et al, 1963; Perley and Kipnis, 1967; Creutzfeldt et al, 1978; Chan et al, 1984). Perley and Kipnis (1967) demonstrated that the insulin response to oral glucose in obese individuals was prolonged and increased, implicating the enteroinsular axis as a possible contributor. Studies into the possible role of GIP in the hyperinsulinemia of obesity have produced equivocal results. Basal IR-GIP levels in obese individuals have been shown to be either normal (Creutzfeldt et al, 1978; Jorde et al, 1983a; Service et al, 1984; Amland et al, 1984) or greater than normal (Salera et al, 1982; Elahi et al, 1984; Mazzaferri et al, 1985). Creutzfeldt et al (1978) found that when obese subjects with normal glucose tolerance ingested a high calorie (1031 kcal) mixed meal, both IR-GIP and insulin secretion was markedly greater than normal subjects. Most other studies, however, have demonstrated normal postprandial IR-GIP responses in obesity (Jorde et al, 1983a; Service et al, 1984; Amland et al, 1984). While several investigators have demonstrated an exaggerated IR-GIP response to oral glucose in obese subjects (Salera et al, 1982; Elahi et al, 1984; Mazzaferri et al, 1985), others have found the IR-GIP response to oral glucose to be 4 6 similar to that observed in lean individuals (Creutzfeldt et al, 1978; Amland et al, 1984). Further, /3-cell sensitivity to GIP appears to be unaltered in obese humans. Intravenous infusion of porcine GIP during a hyperglycemic clamp produced a similar insulin response in both lean and obese volunteers (Amland et al, 1985b), and Elahi et al (1984) calculated /3-cell sensitivity to GIP from the insulin and IR-GIP response to oral glucose and found it to be normal in obese individuals. Still, studies in obese man must be interpreted cautiously because such individuals can vary widely in terms of degree and onset of obesity, diet, carbohydrate metabolism and genetic makeup. These factors can partly be controlled for by the use of genetic animal models of obesity, such as the Zucker fatty rat, in which obesity and hyperinsulinemia follow a relatively consistent and predictable course. Further, the development of obesity in these animals has been shown to resemble the onset and progression of the obese syndrome in juvenile onset obese man (Argiles, 1989). The Zucker obese rat, which arose spontaneously as a mutation in a cross between two normal rat strains (Zucker and Zucker, 1961), is now used frequently in the study of obesity. The obese mutation is transmitted as an autosomal, Mendelian recessive characteristic. Homozygous recessive (fa/fa) animals have been shown to become overtly obese at approximately 5 weeks of age, exhibiting a marked hyperplasia (Johnson et al, 1978) and hypertrophy (Bray, 1969) of adipocytes that were not seen in the lean (Fa/-) littermates of these animals. The obese phenotype was further characterized by hyperphagia (Zucker and Zucker, 1962; Bray and York, 1972), hyperlipemia (Zucker and Zucker, 1962; Cushman et al, 1978), and a striking hyperinsulinemia which became apparent as early as 2 weeks of age (Blonz et al, 1985). The cause of obesity in these animals is unknown. Evidence has been presented that the primary defect was hypothalamic in origin (Martin et al, 1986), although metabolic changes have been found to occur in various organs in these rats. Whether hyperinsulinemia plays a causative role in the 47 development of obesity or is secondary to other metabolic changes is unclear. Destruction of pancreatic /3-cells in obese rats by streptozotocin treatment, followed by insulin replacement, did not diminish the elevated adipose tissue lipogenesis in obese animals, but did reduce the hyperphagia (Stolz and Martin, 1982; Chan and Stern, 1982). Further, it has been suggested that the age of onset of adipose cell hypertrophy (5-7 days) (Boulange et al, 1979) may precede the onset of hyperinsulinemia (2 weeks) (Blonz et al, 1985). These results suggest that insulin hypersecretion may be a secondary manifestation of other metabolic alterations (e.g. increased peripheral insulin resistance), although it clearly contributes to the development and maintenance of the obese state. The mechanism underlying the /3-cell hypersecretion in obese Zucker rats is unknown. Alterations in both islet mass and function have been clearly demonstrated. Pancreatic insulin content of obese animals has been shown to be approximately 2 times that of lean animals at 5-6 weeks of age (Kuffert et al, 1988). This has been attributed to both a hypertrophy and hyperplasia of pancreatic /3-cells within the islet (Shino et al, 1973; Larsson et al, 1977; Chan et al, 1984). Enhanced /3-cell sensitivity to various secretagogues has also been demonstrated in numerous studies both in vivo and in vitro (Bryce et al, 1970; Hayek and Woodside, 1979; Rohner-Jeanrenaud et al, 1983; Chan et al, 1984; Blonz et al, 1985; Curry and Stern, 1985). Rohner-Jeanrenaud et al (1983) implicated a role for the parasympathetic nervous system in the hypersecretion of insulin from the pancreas of Zucker rats. In genetically preobese (17 day old) animals, intravenous glucose produced an augmented insulin response compared to lean rats that was reversible by atropine pretreatment. It was suggested that a hypothalamic defect resulting in increased parasympathetic activity produced a chronic overstimulation of insulin secretion. However, atropine-sensitive hypersensitivity to secretagogues persisted in the perfused pancreas of adult Zucker rats, suggesting that increased islet mass or /3-cell 48 sensitivity could account for the enhanced insulin response. Other studies have failed to confirm the involvement of the vagus nerves in mediating the hypersecretion of insulin (Chan, 1985). Chan et al (1984) investigated the contribution of the enteroinsular axis and GIP to the hyperinsulinemia of obese Zucker rats. In response to an oral glucose challenge, both lean and obese rats showed normal glucose tolerance, and basal and stimulated IR-GIP levels were similar in the two phenotypes. However, pre- and post-challenge insulin levels were markedly higher in the obese animals. The possibility that /3-cell sensitivity to GIP was altered in the obese animals was investigated using the perfused pancreas. It was found that the insulin response to GIP was enhanced in the pancreas of obese animals and, more interestingly, that the glucose threshold for the insulinotropic action of GIP was well below fasting levels in these rats. The presence of an unidentified 0-cell defect was suggested, possibly at the level of the GIP receptor, which uncoupled the glucose-dependency of the effect of GIP on /3-cell secretion. The authors further speculated that this mechanism contributed to the fasting hyperinsulinemia observed in these animals. A possible alteration of the GIP receptor in adipose tissue in obese Zucker rats was later demonstrated by Beck and Max (1987), who showed that GIP-induced fatty acid incorporation into adipose tissue was enhanced in obese animals; a modification in GIP receptor number or affinity was suggested as the probable cause. Since Chan et al (1984) used the perfused pancreas in their study, the possible influences of intrapancreatic nerves or other islet hormones on the insulin response to GIP could not be eliminated. Further, the results could be at least partly attributed to the increased pancreatic insulin content in obese animals rather than a defect at the level of the /3-cell. Clearly, in vitro approaches using isolated islets and FACS-purified /3-cells are warranted to investigate the mechanism underlying the hypersensitivity of the obese Zucker rat pancreas to GIP. 49 In summary, GIP appears to be an important component of the enteroinsular axis. The hormone is a potent stimulator of insulin secretion and this action is influenced by glucose, other fuels, hormones, and neural pathways. Further, the peptide may have a role in the regulation of the secretion of other pancreatic hormones. Finally, GIP may be involved in the pathophysiology of certain metabolic conditions, such as hyperinsulinemia in the obese Zucker rat. Yet many facets of the role of GIP in the regulation of islet hormone secretion remain poorly understood. Proper investigation into the effects of GIP, or any hormone, on its target cell(s) requires not only study in the whole animal and intact organ but also at the level of the target cell with in vitro preparations. Integration of the observations gained from the single cell to the whole organism provides the framework to fully understand the action of the peptide. In the case of GIP, in vitro preparations sensitive to the action of this hormone were unavailable until recently. Thus, little was known of the nature of the interaction of GIP at the level of the islet. Some aspects of the role of GIP in the regulation of islet hormone secretion were therefore addressed in the present work through the use of four different preparations frequently employed in the study of endocrine pancreatic secretion in vitro: isolated islets, FACS-purified /3-cells, neoplastic /3-cells (the / 3 T C 3 cell line), and the perfused pancreas. Using these techniques, this thesis examined the hypothesis that GIP influences the secretion of insulin, glucagon and somatostatin from the pancreatic islet via direct interaction with specific islet cell receptors. The goal of these studies was to gain a better understanding of the role of GIP in the regulation of the secretions of the endocrine pancreas. 50 METHODS I. E X P E R I M E N T A L P R E P A R A T I O N S Several different in vitro preparations were employed to examine the regulation of islet hormone secretion by GIP and the binding of GIP to pancreatic endocrine tissue. A Animals 1. WistarRats Male Wistar rats weighing 250-350 g were used in most experiments. These rats were supplied by Animal Care, U.B.C. or Charles River Laboratories (Quebec). Upon arrival, they were housed in either metal cages (3-5 rats/cage) or polyethylene cages with bedding (5-8 rats/cage) for at least 2 days before use. The room was temperature- and light-controlled (24 °C; 12 h cycle) and rats were allowed free access to laboratory food and tap water. 2. Zucker Rats The Zucker rat colony was maintained in the animal care facilities of the U.B.C. Acute Care Center. Breeding pairs, consisting of lean heterozygote females (Fa/fa) and obese homozygous males (fa/fa), were purchased from Charles River Laboratories. The offspring of these breeding pairs were lean (Fa/?) and obese (fa/fa) pups (approximate 1:1 ratio). The animals were housed in polyethylene cages with bedding in a room similar to that used for the Wistar rats, with laboratory chow and tap water available ad libitum. Breeding pairs and lactating females plus their litters were housed in separate cages. After weaning (21 days), 3-6 rats of either phenotype but the same sex were housed together. Animals were brought to the Physiology Department animal care facilities at least one day prior to use to minimize the effects of stress associated with transport. In these studies, Zucker rats of either sex were used between 12-20 weeks of age, by which time rats with the recessive genotype were overtly obese and therefore easily identified. 51 B. In Vitro Preparations 1. Isolated Islets Islets were isolated by collagenase digestion of the pancreas followed by dextran gradient purification modified from the technique of Van der Vliet et al (1988). a) . Materials Plastic or siliconized glassware was used throughout. Chemicals were of reagent grade unless otherwise stated; commercial sources and abbreviations are listed in the Appendix. Hanks' Balanced Salt Solution (HBSS) was prepared first as a 5X stock solution from concentrated powder (Gibco Laboratories, Grand Island, N Y ) and supplemented with 18.5 g/1 sodium bicarbonate (NaHC03) and 11.9 g/1 N'-2-hydroxyethylpiperazine-N'-ethanesulfonic acid (HEPES). The solution was filter-sterilized (0.22 iim filters; Falcon Labware, Becton-Dickinson, Lincoln Park, NJ) into autoclaved bottles and stored at 4 ° C until use. When needed, HBSS was prepared by diluting the stock solution 1:5 in sterile H 2 0 , adding sterile BSA (0.1 %; Fraction V) and adjusting to p H 7.2-7.4. b) . Islet Isolation Usually, animals were not starved prior to islet isolation. Zucker or Wistar rats were first anesthetized with 60 mg/kg (1 ml/kg) sodium pentobarbital (Somnotol R ; M . T . C . Pharmaceuticals, Cambridge, Ont.) administered intraperitoneally. The abdominal contents were then exposed by a midline laparotomy from sternum to pubis. The animal was sacrificed by cutting the diaphragm to induce respiratory arrest. The bowel was pulled back to allow access to the pancreas and the bile duct was cannulated anterogradely with polyethylene tubing (PE50; Clay-Adams, Parsippany, NJ) near the hilus of the liver. The cannula was secured with a single tie (3-0 silk; Ethicon Inc., Peterborough, Ont.) and the duct clamped at the distal (duodenal) terminus. Ice-cold HBSS (20 ml) containing 0.25 mg/ml collagenase Type XI was slowly 52 injected into the duct via the cannula, distending the pancreas to several times its normal size. The distended pancreas was then carefully resected. The gland was washed in a petri dish with HBSS, trimmed of fat and other tissue and placed in a 150 ml glass bottle on ice. After 30 min, the bottle, normally containing 2 glands, was transferred to a 37 ° C water bath and incubated for 22-25 min. The digestion was then terminated by the addition of ice-cold HBSS to a total volume of 100 ml and the pancreatic tissue was dispersed by aspirating the glands several times through a 10 ml plastic pipet. The digested tissue was washed twice by centrifugation (3 min, 100 X g) in 50 ml ice-cold HBSS. Ductal, vascular and undigested exocrine tissue was then removed by filtration through a wire mesh (approximately 800 Aim) and the filtrate was washed two more times in 50 ml HBSS. Islets were purified from the digested pancreatic tissue by centrifugation in a discontinuous dextran gradient. The pellet was suspended in 12 ml of 29 % (w/v) dextran (Industrial grade; m.w. 6 X 104 - 9 X 104) in HBSS. Four milliliters of 29 % dextran was layered beneath this suspension and 6 ml of 11 % dextran was carefully layered on top. The gradient was then centrifuged for 4 min at 40 X g followed by 10 min at 500 X g. Exocrine tissue formed a pellet beneath the 29 % dextran layer while the majority of islets remained suspended in this layer or rose to the interface of the 29 % and 11 % layers. Islets were harvested by diluting the supernatant to 50 ml with HBSS and centrifuging at 500 X g for 5 min. The pellet was then resuspended in 10-20 ml HBSS and stored on ice until islets from 4-8 rats had been pooled (1000-5000 islets). The pooled islets were viewed under a dissecting microscope (magnification 25X; Zeiss, West Germany) in a petri dish painted with black enamel to provide contrast. A Pasteur pipet (length 5.75"; Fisher Scientific Co., Fair Lawn, NJ) was used to remove non-islet tissue and islet fragments. The pure islets were then washed 2X by centrifugation (5 min, 500 X g) in 50 ml sterile HBSS and resuspended in culture medium. The medium used for short term culture of rat islets was CMRL-1066 (Gibco) plus 10 % heat-inactivated Calf Supreme Serum (Gibco), 2 m M 5 3 L-glutamine and an antibiotic-antimycotic mixture (Gibco) consisting of 100 U / m l penicillin, 1000 Mg/ml streptomycin, 50 Mg/ml gentamycin and 0.25 Mg/ml amphotericin B. It was usually made fresh on the day of the experiment or stored at 4 ° C for no more than 1 week. c). Hormone Secretion from Isolated Islets The islets were handpicked under the dissecting microscope using gentle mouth suction through rubber tubing and a 0.22 fim filter (Millipore, Bedford, M A ) attached to a 200 /xi glass micropipette. Frayed islets and islets with non-islet tissue attached were discarded. Twenty randomly selected islets of variable size were transferred to individual wells on a 48 well plate (Costar, Cambridge, M A ) . Each well had been previously coated with collagen for 45 min, obtained by dissolving rat tail collagen fibers in 0.02 M acetic acid (approximately 3.5 g/1) for 48 h at 4 °C, centrifuging (20 min, 1000 X g) and storing the supernatant at 4 ° C until needed. Five hundred microliters of islet culture medium were added to each well containing islets and the islets were cultured in an incubator (95 % air / 5 % C O 2 ) at 37 ° C for 2 days prior to release experiments. The incubation medium used in release experiments was Dulbecco's Modified Eagle's Medium ( D M E ) prepared from concentrated powder (Gibco) and supplemented with 3.7 g/1 N a H C 0 3 , 2.38 g/1 H E P E S , 0.1 % B S A (RIA Grade), 200 K.I .U. /ml aprotinin (Trasylol R ), 110 mg/1 sodium pyruvate, 290 mg/1 l-glutamine, 15 mg/1 phenol red and the desired concentration of D-glucose. Glucose was stored as a 308 m M concentrated solution at 4 °C. The medium was adjusted to p H 7.35 and heated to 37 °C immediately prior to use. The cultured islets were first examined under the dissecting microscope. Wells with contamination or where a majority of islets had not adhered to the dish were not used. Culture medium was aspirated from each well using a Pasteur pipet; care was taken to exclude the islets. The islets were first pre-incubated (95 % air / 5 % C O 2 ; 37 °C) for 45 54 min in release medium (0.5 ml) with 4.4 m M glucose. This medium was then discarded and replaced with release medium containing the desired glucose concentration plus 12 /d aliquots of 50X concentrated solutions of the drug or peptide to be tested (total volume: 0.6 ml). After 1 h incubation, the release medium was transferred with a Pasteur pipet to 1.5 ml polypropylene tubes (Eppendorf R ; Brinkmann Instruments Co., Westbury, NJ) on ice containing an additional 500 K.I.U. (50 /il) aprotinin. The medium was centrifuged for 5 min (500 X g, 4 °C) and the supernatant stored frozen in 250 /xl aliquots at -70 ° C for assay of glucagon-like immunoreactivity (GLI) and somatostatin-like immunoreactivity (SLI). The remaining supernatant was stored at -20 ° C for immunoreactive insulin (IRI) assay. The islets from each well were extracted in 200 /ul acetic acid (2 M), boiled for 10 min and centrifuged (10 min, 500 X g). Aliquots (50 ul) of the supernatant were lyophilized with 10 /ul B S A (5 %; RIA grade) in siliconized test tubes and stored at -70 ° C prior to the measurement of islet SLI and G L I content; the remaining supernatant was stored at -20 ° C prior to the measurement of islet IRI content. Extracts for assay of G L I and SLI were lyophilized to remove acetic acid and therefore avoid the problem of the low p H of these acid extracts interfering with the RIA; this was not a problem with the IRI assay because extracts were typically diluted 1:1000 in assay buffer before addition to the assay. Recovery of G L I and SLI in standards processed in the same way as islet extracts was 102 % and 88 % (n=3), respectively. 2. FACS-Purified Islet Cells Fractions of /3-cells and non-/3-cells were obtained from isolated islets using a fluorescence-activated cell-sorter (FACS) via the technique of Pipeleers et al (1985a) with slight modifications. a). Materials HEPES-buffered Earle's medium (EH) was prepared in this laboratory and was of 55 the following composition: 124 m M NaCl, 5.4 m M KC1, 1.8 m M C a C l 2 , 0.8 m M M g S 0 4 , 1 m M NaH 2 P04 , 14.3 m M N a H C 0 3 and 10 m M H E P E S . Ca-free E H medium and Ca-free E H plus 1 m M [ethylenebis-(oxyethylenenitrilo)]tetraacetic acid ( E G T A ) were also prepared. Before use these buffers were supplemented with 0.1 % B S A (Fraction V) and 2.75 m M glucose, adjusted to p H 7.30 and filter-sterilized. A l l glassware was siliconized, b). Islet Cell Preparation Islets (1000-5000), obtained as described and cultured overnight in islet culture medium prior to use, were first dissociated to single cells by enzymatic digestion and mechanical dispersion. The islets were washed 2X by centrifugation (3 min, 500 X g) in 50 ml HBSS and then 2X in 50 ml Ca-free E H . The pellet was resuspended in 2 ml Ca-free E H plus E G T A at room temperature and periodically aspirated through a Pasteur pipet which had been drawn to a fine tip (200-400 /xm diameter) over a flame. After 10 min, trypsin (final concentration 200 Mg/ml) and DNase (final concentration 2 /xg/ml) in 2 ml Ca-free E H plus E G T A were added, and the cells were warmed to 30 °C and continually aspirated through the pipet for a further 8-15 min. The suspension was examined every 2-3 min under an inverted microscope (magnification 100X; Zeiss) to determine the degree of enzymatic digestion. When at least 50 % of the cells were dissociated as single cells, the digestion was stopped by filtering the suspension through a 100 /xm nylon screen (Nitex R , Switzerland) into a 50 ml tube containing ice-cold E H plus calcium. The filter removed undigested material and cell clumps. After centrifugation (5 min, 500 X g), the cells were resuspended in 1 ml Ca-free E H and maintained at 17 ° C for 15 min prior to cell sorting. The yield of islet cells using this procedure was 5 X 105 - 1 X 106 cells from 1000-5000 islets obtained within 90 min. These cells were generally 60-90 % viable as determined by exclusion of 0.4 % trypan blue. Immunocytochemical examination of the dissociated islet cells after culture (Section I.B.2.e) showed that the majority of the cells were /3-cells (50-75 %) and a-cells (5-20 %), with the remainder being S-, PP- and unidentified cells. 56 c) . Purification of B- and Non-0 Islet Cells The suspension of dissociated cells was sorted into /3-cell and non-/3 islet cell fractions using a F A C S - I V (Becton-Dickinson, Sunny Vale, C A ) in the laboratory of Dr. N. Kastrukoff (Acute Care Center, U.B.C.). Cells were selected on the basis of two parameters: light scatter activity and fluorescence. Compared to other islet cells, /3-cells have been shown to have larger light scatter activity (Nielsen et al, 1982) and to exhibit higher autofluorescence (emission measured between 510-550 nm) in 2.8 m M glucose at 17 °C due to increased levels of flavin adenine dinucleotide ( F A D ; Van de Winkel et al, 1982). Fluorescence/scatter dot plots of islet cells under these conditions consistently showed two subpopulations: one of higher fluorescence and scatter and one of lower fluorescence and intermediate scatter (Figure 2). Selection of appropriate windows around these populations allowed the collection of two fractions (0-cells and non-0 islet cells). The /3-cell fraction consisted of more than 98 % insulin-containing cells when examined immunocytochemically. More than 95 % of these cells excluded trypan blue. Non-/3 cells were not always found in this fraction but when observed the major contaminating cell types were a- and PP-cells; 5-cells were rarely seen. The yield of /3-cells varied widely, from 5 X 104 to 5 X 105 /3-cells/8 Wistar rat pancreata. The non-/3-cell fraction of lower scatter and autofluorescence varied in cell proportion, consisting of 40-70 % a-cells, 20-50 % /3-cells and a smaller percentage of S-, PP-, and unidentified cells. Typically, 80-90 % of the cells in this fraction excluded trypan blue. The yield of non-/3-cells was usually similar to that of the /3-cells. d) . Culture of 0-Cells and Non-/3-Cells Cell fractions from the F A C S were collected in plastic test tubes containing 1 ml islet culture medium. These tubes were immediately transferred to ice upon completion of the sorting procedure and transported to the laboratory. The cells were washed 2X by centrifugation (5 min, 500 X g) in 15 ml HBSS under sterile conditions and resuspended in 57 o o c <u o o B Light Scatter c c 01 E Z 01 O .-Tvi-u. fluorescence Light Scatter F I G U R E 2: Analysis of Dispersed Rat Islet Cells by Autofluorescence and Light Scatter on a Fluorescence Activated Cell-Sorter. Cells from approximately 2500 islets were dispersed to single cells as described in Methods and applied to the F A C S . (A) Representative dot-plot of cells sorted according to autofluorescence and light scatter. Gated windows show collected fractions: (a) non-/3-celI fraction, (b) /3-cell fraction. (B) Representative histogram of same cells showing cell number at increasing intensity of autofluorescence and light scatter. 58 islet culture medium. B-Cells were seeded 3000 cells/well in 200 ii\ islet culture medium in uncoated, flat-bottom 96-well plates (Costar) for insulin release experiments. In some experiments, B-cells and non-/3-cells were remixed in equal proportions to increase /3-cell sensitivity to secretagogues (Pipeleers et al, 1985b). This mixture was then plated at a density of 6000 cells/well in 200 *il on 96-well plates. For immunocytochemical characterization of the cell fractions, 1 X 104 - 2 X 104 cells in 500 fx\ culture medium were seeded on sterile, laminin-coated 18 mm coverslips (Fisher) in a 10 X 35 mm petri dish (Costar). The cells were allowed to adhere for 20-30 min, then an additional 2 ml culture medium was added to the dish. The non-/3-cell fraction culture medium was supplemented with 25 /xM N 6 ,2'-0-dibutyryladenosine 3':5'-cyclic monophosphate (dbcAMP) and 25 nM 3-isobutyl-l-methylxanthine (IBMX), which have been shown to enhance the survival of a-cells when cultured in the absence of /3-cells (Pipeleers et al, 1985c). A l l experiments were performed after 3 days culture at 37 °C in 95 % air / 5 % C 0 2 . e). Immunocytochemical Characterization of Cell Fractions The dilutions and sources of the various antisera used in these experiments are listed in the Appendix. Cultured cell fractions were washed 2X in 2-3 ml HBSS and then fixed for 5 min in Bouin's solution (75 % saturated picric acid, 25 % formaldehyde). The cells were washed in 70 % ethanol and then 2-3X in phosphate-buffered saline (8.0 g/1 NaCl, 0.2 g/1 K H 2 P 0 4 , 1.15 g/1 N a 2 H P 0 4 ) plus 0.1 % sodium azide (NaN 3 ; PBS-azide). First layer antisera to insulin, glucagon, somatostatin or pancreatic polypeptide (either alone or in combination with one other) at the appropriate dilution (in PBS-azide) was then applied. After 1-3 days incubation at 4 °C, the cells were washed 2-3X in PBS-azide, the second layer antisera was applied and incubated for a further 1 h at room temperature. The cells were then washed 2-3X in PBS-azide and the coverslips mounted on glass slides and viewed on an Axiophot microscope (magnification 200X; Zeiss). A representative 59 example of a /3-cell fraction stained for the presence of insulin and glucagon-containing cells is shown in Figure 3. The percentage of different cell types was approximated by counting stained cells in randomly selected fields, f). Insulin Secretion Experiments Release experiments on pure /3-cells and mixtures of B- and non-/3-cells were performed in the same manner. Viability of these cells after 3 days culture was routinely monitored in test wells by trypan blue exclusion; experiments were only done when viability of the adhered cells exceeded 90 %. The viability of cultured cells was unaffected by 2 h incubation in release medium. Experiments were performed on adhered cells, since the majority of unattached cells were not viable. Culture medium and unattached cells were removed by inversion of the plate and gentle tapping onto tissue paper. The cells were first pre-incubated (37 °C; 95 % air / 5 % C 0 2 ) for 45 min in 200 /xl D M E release medium (as for islets) with 4.4 m M glucose. Pre-incubation medium was then discarded and release medium with the desired glucose concentration (usually 17.8 mM) was added. Aliquots (10 /xl) of the desired drug or peptide (prepared as 25X concentrate in release medium) were added for a total volume of 250 /xl/well. After 1 h incubation, the release medium was transferred to chilled 1.5 ml Eppendorf R tubes, centrifuged (5 min, 500 X g, 4 °C) and the supernatant stored at -20 °C for IRI assay. Each well was then extracted in 200 /xl of 2 M acetic acid, boiled for 10 min, centrifuged and the supernatant stored for the measurement of cell IRI content. 3. /3TC3 Tumor Cell Line Insulin secretion and GIP receptor binding experiments were also performed on a /3-cell line derived from mouse insulinoma tissue. The cell line originally arose in a lineage of transgenic mice expressing an insulin-promoted, SV40 T antigen hybrid oncogene in pancreatic )3-cells (D'Ambra et al, 1990). 60 FIGURE 3: Immunocytochemical Staining of FACS-Purified 0-Cells. Representative B-cell fraction obtained by autofluorescence activated cell-sorting and stained for immunoreactive-insulin (fluorescein) and glucagon-like immunoreactivity (rhodamine). No glucagon-containing cells were observed. Magnification, 200X. 61 a) . Culture of 0TC3 Cells The cells were maintained in long-term culture in the M R C Regulatory Peptide Group by Dr. Z . Huang. The culture medium was D M E with 5.5 m M glucose, 3.7 g/1 N a H C 0 3 , 25 m M H E P E S , 110 mg/1 sodium pyruvate, 290 mg/1 l-glutamine and antibiotics (purchased from Terry Fox Laboratories, Vancouver, B.C.) and supplemented with 15 % heat inactivated horse serum (Gibco) and 2.5 % fetal bovine serum (Gibco). Cells were grown in 50 and 250 ml tissue culture flasks (Falcon) in a 37 ° C incubator (95 % air / 5 % C O 2 ) ; medium was changed 2X per week. The cells were subcultured when needed or when approximately 50 % of the cells were confluent, by harvesting with trypsin-EDTA (trypsin 1:250; E D T A 0.2 g/1). b) . Insulin Secretion Experiments Freshly harvested cells were seeded in the above culture medium (0.5 ml) onto uncoated 24 well plates (Falcon) at a density of 1 X 106 cells/well. The cells were cultured for a further 2 days to allow the cells to adhere to the plate. Viability of the attached cells was > 98 % as assessed by trypan blue exclusion; this was unaffected by 2 h culture in release medium. Release experiments were performed in identical medium and in the same manner as for pure /J-cells with the following exceptions. First, each well was washed once in 0.5 ml release medium prior to pre-incubation to remove unattached cells. Incubation medium was removed by gentle suction with a Pasteur pipet. Pre-incubations and release incubations were done in a total volume of 0.5 ml and each secretagogue was added as a 10 ul aliquot of a 50X concentrated solution in release medium. Finally, since IRI content did not vary significantly from well to well within a single plate, only 4 representative wells from each plate were extracted in 0.5 ml acetic acid (2 M) for the measurement of total IRI content. Adhered cells were scraped from these wells using a rubber spatula. 62 4. Perfused Pancreas The isolated, vascularly perfused rat pancreas was used to assess the biologic activity of two GIP receptor probes, biotinylated GIP l -30-NH 2 (B-GIP) and HPLC-purified i27/i25j_Qjp (Section II.A). The inhibitory effects of the pancreatic neuropeptide galanin on the insulin response to GIP and other insulin secretagogues was also examined using this preparation. a). Apparatus Perfusate was delivered to the pancreas via a peristaltic pump (No. 7553-30; Cole-Parmer Instrument Co., Chicago, IL) at a flow rate of 3.0 rnl/min. The perfusate was stirred in flasks that were gassed with a H20-saturated mixture of 95 % 0 2 / 5 % C 0 2 to maintain approximately p H 7.4. Before delivery to the organ, the perfusate was heated to 37.5 ° C by passage through tubing (PE160) coiled around a thermostatically-controlled heating block and then passed through a bubble trap and fine mesh filter. System pressure was monitored continuously using a transducer. The gland was maintained at a constant temperature of 37 ° C by a heating source underneath and a piece of plastic wrap protected the organ from circulating air. A n automated fraction collector with an ice-filled reservoir was used to collect the venous effluent into chilled test tubes. A n infusion pump (Model 940; Harvard Apparatus Co. Inc., Millis, M A ) was used to deliver most test drugs and peptides from 10 ml syringes, via PE90 tubing into a rubber bulb situated just before the perfusate entered the gland. Alternatively, secretagogues (e.g. glucose, arginine) were added directly to the perfusate in auxiliary flasks. The perfusate in these flasks was recirculated through tubing set up in parallel with that entering the gland. When desired, the stimulus was delivered to the pancreas by rapidly switching the input to the organ from one flask to the other, proximal to the bubble trap/filter. In addition, some agents were administered as gradients of increasing concentration. This was achieved by coupling two perfusate flasks of equal cross-sectional area in series and adding 2X the 63 desired final concentration of the test substance to perfusate in the distal flask, b). Surgical Procedure The technique for surgical isolation and vascular perfusion of the pancreas was adapted from Penhos et al (1969). Male Wistar rats (250-350 g) were fasted the evening prior to surgery (12-18 h) to aid the surgical dissection and diminish variation in basal hormone secretion. They were first anesthetized by intraperitoneal injection of sodium pentobarbital (60 mg/kg) and placed on a heated pad. The abdomen was exposed by a midline opening from sternum to pubis. The abdominal aorta and inferior vena cava were located, cleared of associated connective tissue and the artery was isolated with two untied ligatures (3-0 silk) in preparation for subsequent cannulation. A third loose ligature was placed around the aorta immediately below the diaphragm. The vasculature supplying the left and right kidneys and adrenal glands were then tied and the sigmoid colon was sectioned between double ligatures. Next, the connective tissue around the spleen was dissected and the organ was removed. The vasculature between the pancreas and the stomach along the greater curvature was singly ligated. The pylorus and esophagus, including the left gastric artery and vagus nerve trunks, were doubly ligated and sectioned between ties, allowing the removal of the stomach. A duodenal drainage tube was then secured into the small bowel distal to the level of the pancreas and the majority of the remaining gut, from this drainage tube to just caudal to the cecum, was isolated and removed. In preparation for portal vein cannulation, a single loose tie was placed around the vein where it exits the pancreas, while a second loose tie, closer to the liver, encompassed the portal vein, bile duct and associated vasculature. A third untied ligature was placed around the inferior vena cava near the right adrenal gland. A cannula (PE160 tubing) was then inserted into the abdominal aorta and secured at a level just below the trunk of the superior mesenteric artery. The loose aortic ligature near the diaphragm was tied, directing the flow of perfusate to the pancreas via the 64 superior mesenteric and celiac arteries, and 2 ml heparin in saline (0.9 % NaCl) was administered to the pancreatic vasculature via this route. The animal was then sectioned at the diaphragm. The ligature around the inferior vena cava was tied and a portal vein cannula (PEI60 tubing) secured for venous collection. The preparation was then placed upon the heating block, the arterial cannula was connected to the perfusion apparatus and the flow of perfusate started. Any leak in the preparation was immediately halted using ligatures or clamps. c) . Solutions (i) . Perfusate The perfusate was a Krebs' solution containing 3.0 % dextran (Clinical grade), 0.2 % B S A (RIA grade), and the desired concentration of glucose. It was made fresh on the day of perfusion by dilution of a concentrated Krebs' stock solution with saline followed by the addition of NaHCC>3, dextran, B S A and glucose. The final ionic composition of the perfusate was 120 m M NaCl, 4.4 m M KC1, 2.5 m M C a C l ^ 1.2 m M M g S 0 4 . 7 H 2 0 , 1.5 m M KH2PO4 and 25 m M N a H C C ^ . The final perfusate glucose concentration was verified to within 10 % of the desired concentration using a glucose analyzer (Beckman Instruments Inc., Fullerton, C A ) . Osmolarity of the perfusate was approximately 285 mOs. (ii) . Drugs and Peptides A l l peptides were dissolved in siliconized test tubes in a small volume of 0.01 M acetic acid plus 0.2 % B S A (RIA grade) and maintained as a stock solution at 4 °C for use that day. Acetylcholine chloride (ACh) was stored as 1 M solution in distilled H 2 0 at 4 °C. When delivered via the sidearm infusion pump, an aliquot of stock solution was diluted in perfusate in a 10 ml syringe to achieve the desired concentration. Glucose and arginine were added directly to perfusate flasks during the equilibration period. d) . Perfusion Procedure The preparation was perfused for a period of 20-30 min prior to the collection of 65 samples to equilibrate basal hormone secretion following the surgical preparation. During this time, the pump and heating apparatus were adjusted to maintain a steady flow rate (3.0 ml/min) and perfusate temperature (37.5 °C) . Different experimental protocols were used in GIP gradient and galanin infusion experiments. In GIP gradient studies, fractions of venous effluent were collected for 1 min into ice-cold test tubes and stored at -20 ° C for subsequent IRI assay. In galanin infusion experiments, fractions were collected every 5 min in tubes containing 5000 K.I.U. aprotinin to prevent degradation of somatostatin. Aliquots (0.5 ml) from these samples were taken for IRI and SLI assay and stored at -20 °C; an additional 500 K.I.U. of aprotinin was added to samples for assay of SLI before freezing. II. GIP R E C E P T O R STUDIES Attempts to demonstrate the presence of specific receptors for GIP on rat pancreatic islets and /3TC3 cells were made using both a histological approach and radioreceptor binding assays. A . GIP Receptor Probes Two types of probes were developed to investigate the putative GIP receptor. 1 2 5 I -GIP (purified by H P L C ) was used in radioreceptor binding assays and B-GIP was prepared as a probe for localizing 0-cell GIP receptors using avidin-biotin immunocytochemistry. 1. 1 2 5 I -GIP a). Preparation of 1 2 5 I -GIP 1 2 5 I -GIP used in receptor binding studies (and RIAs) was prepared by chloramine T iodination (Kuzio et al, 1974) followed by purification of the iodinated peptide by gel filtration and H P L C . A gel filtration column was prepared prior to the iodination by pouring a sufficient volume of pre-swollen Sephadex R G-15 beads (Pharmacia, Uppsala, Sweden) in 0.2 M acetic acid into a 0.5 X 10 cm plastic pipet with a glass wool plug, allowing to settle and equilibrating the column by running overnight (approximately 250 66 jul/min) in column buffer (0.2 M acetic acid plus 5.0 % BSA; R I A Grade). On the day of the iodination, the column buffer was supplemented with 2 % aprotinin and equilibrated for at least 1 h. The iodination reaction was performed in a siliconized test tube containing 10-15 jug porcine GIP 1-42 ( E G III fraction; Brown et al, 1970; Brown, 1971) dissolved in 100 ul phosphate buffer (0.4 M), p H 7.5. The phosphate buffer was prepared by titrating 0.4 M N a 2 H P 0 4 with 0.4 M N a H 2 P 0 4 to p H 7.5. Five microliters of N a 1 2 5 I (0.5 mCi) and 10 ul chloramine T (4 mg/ml in 0.4 M phosphate buffer) were added to the reaction vessel, followed 15 s later by 20 ul sodium metabisulfite (14.8 mg/ml in 0.4 M phosphate buffer). The tube was gently mixed by tapping and the reaction mixture applied to the G-15 column. The column was eluted with the described column buffer at a flow rate of approximately 250 jul/min and fractions of 400 /ul were collected (1.5 min/fraction). Radioactivity in 10 /ul aliquots of the eluent fractions was counted for 0.10 min in a Searle gamma counter (NCS Instruments, Inc.) with approximately 69 % counting efficiency; a typical column profile is shown in Figure 4. Free N a 1 2 5 I eluted after the 1 2 5 I -GIP. Peak fractions from the column were tested for their ability to adsorb to charcoal. Ten microliter aliquots of each peak fraction was diluted in GIP assay buffer (Section III.AAa) to approximately 5000 cpm/100 /ul. Then, 100 /ul aliquots of these diluted fractions were added (in triplicate) to tubes containing 900 /xl assay buffer. Dextran-coated charcoal was prepared by dissolving 2.5 mg/ml dextran (T70) in 0.04 M phosphate buffer (pH 6.5), adding 2.5 mg/ml activated charcoal (Norit R Decolorizing Carbon (Neutral)), 0.75 % aprotinin and 5 % charcoal-extracted plasma (CEP; Section III.A.l.a) and stirring for 2 h at 4 °C. Two hundred microliters of dextran-coated charcoal were added to each tube and the mixture vortexed and centrifuged (30 min, 1000 X g). Both the supernatant and the charcoal pellet were counted and the percent of total radioactivity 67 F I G U R E 4: Profile of Gastric Inhibitory Polypeptide Iodination Mixture Eluted on Sephadex R G-15. Iodination mixture was applied to Sephadex R G-15 column immediately following iodination as described in Methods. Samples were collected every 1.5 min and 1 2 5 I radioactivity measured in 10 n\ aliquots. Free N a 1 2 5 I eluted after 1 2 5 I -GIP. 'Denotes samples pooled for use in RIA and binding studies. 68 bound (% B) to charcoal (i.e. 1 2 5 I-GIP) calculated as: % B = 100 X C p e i i e t / ( C p e i i e t + ^supernatant) where C = cpm. Peak fractions with the greatest adsorption to charcoal (usually >95 %) were pooled; generally, the fractions used included those with the highest radioactivity in the 1 2 5 I -GIP peak and 1-3 fractions on the descending portion of this peak. These fractions were previously shown to have the greatest immunoreactivity in the RIA, possibly due to decreased substitutions of iodine into the GIP molecule (Kuzio et al, 1974). Usually, the pooled fractions were diluted to 2.5 X 106 cpm/100 /d in a 1:1 mixture of column buffer and acid ethanol (750 ml/1 95 % ethanol plus 15 ml/1 concentrated H Q ) and stored at -20 ° C until use in the RIA or H P L C purification. However, the concentrated tracer in column eluent was sometimes applied directly to the H P L C to maximize the yield of purified 1 2 5 I -GIP. b). H P L C Purification of 1 2 5 I -GIP Purification of various 1 2 5 I -GIP products of the iodination mixture was performed by reverse-phase H P L C on a /nBondapak C18 column (Waters Associates Inc., Milford, M A ) using two Beckman model HOB Solvent Delivery Module pumps (Beckman Instruments Inc, San Ramon, CA) controlled by a programmable Beckman Model 421A Controller. This system was used exclusively for purifying radioiodinated peptides. The solvents used were acetonitrile ( C H 3 C N ) plus 0.1 % (v/v) trifluoroacetic acid (TFA) and H 2 0 plus 0.1 % T F A approximate p H 2.5. Water was obtained from a MilliQ H 2 0 filtration system (Waters) and both solvents were degassed and filtered through either 0.22 /xM ( H 2 0 ) or 0.45 /xm ( C H 3 C N ) filters (Waters) before use. The procedure was performed at room temperature. 1 2 5 I -GIP was injected using a 25 /Ltl needle syringe (Hamilton Co., Reno, N V ) into a Beckman Model 210A injector port with 100 jul capacity. Multiple injections of the iodination mixture were usually required to obtain a sufficient yield of purified tracer for use in receptor binding studies. Separation of the various iodinated forms of GIP was 69 accomplished by increasing the concentration of C H 3 C N in a linear gradient over 10 min at a flow rate of 1 ml/min; the starting and finishing concentrations of C H 3 C N were altered until optimal separation was achieved. Radioactivity of the column eluent was measured using a Model 170 Radioisotope Detector (Beckman) and charted on a Recordall Series 5000 recorder (Fisher). Eluent fractions were collected every 0.5 min into 13 X 100 mm glass test tubes containing 10 /xl B S A (5 %; R IA Grade) and 10 /xl (100 K.I.U.) aprotinin. Aliquots (10 /xl) of these fractions were counted and the desired peak fractions were pooled and subjected to a stream of 100 % N 2 gas for 30 min to evaporate the C H 3 C N . Aliquots containing a known amount of radioactivity (usually 1 X 106 cpm) were frozen (-70 °C) in siliconized test tubes, lyophilized and stored at -20 ° C until use. In some experiments, an excess of GIP (approximately 10 /xg) was added to a small amount of 1 2 5 I -GIP (5 X 105 cpm) before application to the H P L C system to determine if uniodinated GIP eluted separately from 1 2 5 I -GIP. GIP immunoreactivity was then determined by R I A in non-peak fractions. c). Specific Activity of 1 2 5 I -GIP Specific activity of the 1 2 5 I -GIP label was estimated by two methods. First, the total radioactivity (cpm) eluted from the G-15 column was determined as the sum of the radioactivity in each fraction collected. The percent of the total radioactivity in the 1 2 5 I -GIP peak was then calculated. Assuming 100 % recovery of added radioactivity (0.5 mCi), the specific activity (S.A.) was calculated as: S.A.(mCi/mg) = 0.5 mCi X % radioactivity in GIP peak/amount of GIP used (mg). Using this method, specific activity of the unpurified 1 2 5 I -GIP was usually found to be 10 - 50 mCi/mg. Specific activity of both unpurified 1 2 5 I -GIP and fractions of HPLC-purified 1 2 5 I -GIP was also estimated from self-displacement curves, using increasing concentrations of 1 2 5I-labeled peptide in a normal GIP RIA. The ratio of bound to free tracer (B/F) was calculated for each concentration of 1 2 5 I -GIP and the results plotted as B / F vs. 1 2 5 I -GIP (B 70 + F; cpm). On the same axes, B / F vs. GIP standard concentration (pg/100 /xl) was also plotted (Figure 5A). Assuming that the antiserum used bound 1 2 5 I -GIP and unlabeled GIP equally well, both the GIP standard concentration and 1 2 5 I -GIP radioactivity were determined for several values of B / F in the mid-range of the curve and a plot of GIP mass (pg) vs. radioactivity (cpm) was produced (Figure 5B). After determining the slope of this plot, specific activity (S.A.) of the label was calculated as S.A. = l/(slope X counting efficiency). d). Analysis of HPLC-Purified 1 2 5 I -GIP Peak fractions of HPLC-purified 1 2 5 I -GIP were analysed to determine the degree of incorporation of iodine into the two tyrosine residues (positions 1 and 10) on the GIP molecule. This was accomplished by pronase digestion of each peak of iodinated peptide collected from the H P L C into individual amino acids, separation of the mono- and diiodinated tyrosine residues by gel filtration and identification of these residues by comparison of their elution times to commercial standards (Jorgensen and Larsen, 1980). The column used for separation of monoiodinated tyrosine (MIT) and diiodinated tyrosine (DIT) was a glass 1.0 X 50 cm G-10 (Sephadex R ) column in tris-maleate buffer (pH 5.9). The buffer was prepared by titrating 0.05 M tris base with 0.05 M maleic acid to p H 5.9 and adding 0.1 % N a N 3 . The G-10 beads were swollen in this buffer prior to addition to the column and the column was allowed to equilibrate overnight at a flow rate of approximately 200 /xl/min before use. Lyophilized fractions (5 x 105 - 1 X 106 cpm) of HPLC-purified 1 2 5 I -GIP were reconstituted in 500 /xl distilled H 2 0 . Five hundred microliters of 10 mg/ml protease TypeXIV (Pronase E) was added and the mixture was incubated overnight (16-18 h) at 37 °C. A n aliquot containing 1 X 105 cpm (100-200 /xl) was then transferred to a siliconized tube containing 10 mg/ml MIT (3-iodo-l-tyrosine) and 10 mg/ml DIT (3,5-diiodo-l-tyrosine) standards in 1 ml ice-cold column buffer. This mixture was applied to the column 71 1.51 <D " D C 3 O m 1.0-0.5-o.o-125 10 100 1 0 0 0 10000 Mass GIP Standard <pg) (-o) or I Radioactivity (cpm X 10" 2 ) (-•-) B 300-1 ra "S 2 0 0 ra •a c ra 55 0. CO (/> ra 2 100 I Radioactivity (cpm X 10 ) F I G U R E 5: Self-Displacement of GIP by 1 2 : > I-GIP. Graphs were used to determine the specific activity of HPLC-purified 1 2 5 I-G1P (peak 2) as described in Methods. (A) Plots of bound/free vs. mass of GIP standard (open circles) and 1 2 5 I -G I P radioactivity (closed circles). (B) Plot of mass of GIP standard vs. 1 2 5 I -GIP radioactivity as determined from (A). The inverse of the slope of graph (B) was used to calculate the specific activity of peak 2 (297 mCi/mg). 72 and eluted at a flow rate of 200 /xl/min overnight; fractions (4 ml) were collected every 20 min. The elution times of the 1 2 5 I-Tyr residues liberated from digestion of 1 2 5 I -GIP were determined by counting 100 n\ aliquots of each fraction. These were compared to the elution times of the MIT and DIT standards determined by measuring absorbance at 297 nm (MIT) and 312 nm (DIT) on an ultraviolet spectrophotometer (Pye Unicam, Cambridge, England). In separate experiments, the elution times of free N a 1 2 5 I and undigested GIP fractions were also determined. e). 1 2 7 / 1 2 5 I - G I P (i). Preparation of 1 2 7 / 1 2 5 I - G I P The biological activity of the HPLC-purified, iodinated GIP used in receptor binding studies was assessed by comparison of the insulinotropic activity of iodinated GIP to the unmodified peptide using the isolated perfused rat pancreas. To avoid perfusing the rat with high levels of radioactivity and having 1 2 5 I -GIP in venous effluent interfere with the insulin RIA, 1 2 7 / 1 2 5 I - G I P was prepared using the stable isotope N a 1 2 7 I with trace amounts of N a 1 2 5 I . The procedure for preparation of 1 2 7 / 1 2 5 I - G I P was similar to that of 1 2 5 I -GIP. A larger amount of GIP (approximately 0.5 mg) was used to ensure adequate yield for use in the perfusion study and the amounts of other reagents used in the iodination were increased accordingly. The peptide ( E G III; 400-600 Mg) was first dissolved in 100 /xl 0.4 M phosphate buffer, p H 7.5, in a siliconized test tube. Sodium iodide (Na 1 2 7 I) was dissolved in phosphate buffer in an equimolar concentration with GIP (approximately 1 mM). N a 1 2 5 I was then diluted in the N a 1 2 7 I solution to 1 juCi/100 jul; the molar ratio of N a 1 2 7 I : N a 1 2 5 I used in the iodination was approximately 4 X 104:1. One hundred microliters of this N a 1 2 7 I / N a 1 2 5 I mixture was then added to the reaction vessel. The procedure was identical to the iodination and purification of GIP using N a 1 2 5 I , except that the volumes of sodium metabisulfite and chloramine T were increased tenfold. Following 73 gel filtration of the iodination mixture on G-15, radioactivity due to the trace levels of N a 125j was measured in each fraction. Fractions in the first peak corresponding to those normally utilized in a N a 1 2 5 I GIP iodination (i.e. the peak fraction plus the two immediately following) were pooled and stored in column buffer at 4 ° C until application to the H P L C (within 48 h). (ii). H P L C Purification of ^V^h-GlP H P L C purification of 1 2 7 / 1 2 5 i -GIP was performed on a separate H P L C system equipped for measuring absorbance to allow detection of 1 2 7 / 1 2 5 I - G I P and to avoid contamination with 1 2 5 I . A Waters H P L C system was used, consisting of two Model 510 pumps, a Model 712 WISP R (Waters Intelligent Sample Processor) controller and a Model 441 absorbance detector. The system was integrated by a microcomputer with a chromatography software program (Maxima R 820, Waters). The solvents and column used were identical to that used for the purification of 1 2 5 I -GIP so that the profiles obtained would be comparable. Flow rate was 1 ml/min and 1 2 7 / 1 2 5 I -GIP was eluted by increasing the concentration of C H 3 C N in a linear gradient over 10 min. A n automated fraction collector was used to collect eluent fractions (0.5 min/fraction) and the radioactivity in 100 /xl aliquots measured, allowing comparison of the H P L C profiles obtained with 1 2 5 I -GIP and 1 2 7 / 1 2 5 I - G I P . Acetonitrile was evaporated from peak fractions with N 2 and 100 /xl aliquots of these fractions were then lyophilized and stored at -20 °C. 2. Biotinylated GIP (B-GIP) a). Preparation of B-GIP A n analog of GIP (synthetic porcine GIP l-30-NH 2) was biotinylated by a technique described for the coupling of biotin to amine functional groups (Hazum, 1986). The analog was chosen because it was shown to have similar insulinotropic activity to GIP 1-42 (Pederson et al, 1990), while having 3 less lysine residues than the native hormone, thus minimizing the production of multi-biotinylated GIP derivatives. 74 Approximately 0.5 mg of the peptide was dissolved in a small volume of distilled H 2 0 in a siliconized tube and then diluted to a concentration of 1 m M with H E P E S (100 m M , p H 8.5). Ten millimolar biotin-N-hydroxysuccinimide ester (BNHS) in dimethylsulfoxide (DMSO) was added to the tube in a molar ratio of 4:1 (BNHS:peptide). The solution was mixed by gentle tapping and D M S O was added dropwise until the mixture clarified. The biotinylation reaction was then allowed to proceed at room temperature overnight (16-18 h). A C18 SepPak R cartridge (Waters) was used to purify peptide products from the reaction mixture. The SepPak was first primed with 10 ml C H 3 C N (plus 0.1 % T F A ) , followed by 10 ml H 2 0 (plus 0.1 % T F A ) and finally, 10 ml air. The reaction mixture was applied to the cartridge and washed with 10 ml H 2 0 plus T F A followed by 10 ml of 10 % C H 3 C N . It was then eluted by sequential application of 2 ml of increasing concentrations of C H 3 C N in 10 % increments. Absorbance (280 nm) of each eluent fraction was measured and those with detectable absorbance pooled. The product was then lyophilized and stored at -20 ° C until use. b). H P L C Purification of B-GIP Since biotinylation of GIP 1-30, with 3 amine groups, would theoretically yield multiple biotinylated forms of the peptide with varying biological activity (in addition to unaltered GIP 1-30), H P L C purification of B-GIP was attempted. A small amount of B-GIP was weighed (approximately 10 /ng) and dissolved in 100 /il H 2 0 plus 0.1 % T F A . The solution was then applied to the Waters H P L C system used for purification of 1 2 7 / 1 2 5 I - G I P . Peak fractions were lyophilized and stored at -20 °C. B. GIP Receptor Binding Studies The presence of a specific receptor for GIP on /3TC3 cells and on isolated rat islets was investigated using a radioreceptor binding assay employing HPLC-purified 1 2 5 I -GIP as tracer. 75 1. Assay Buffer The receptor assay buffer used in both the BTC3 cell and isolated islet binding studies was modified from one previously used in a GIP receptor binding assay in a hamster pancreatic /3-cell line (Amiranoff et al, 1984). It was composed of 50 m M NaCl, 5 m M KC1, 10 m M M g C l ^ 50 m M H E P E S , 10 m M glucose, 1 % B S A (Fraction V), 0.5 mg/ml bacitracin and 2 % aprotinin. It was usually prepared on the day of experiment or stored under sterile conditions at 4 °C for no more than 1 week. Immediately prior to use, the buffer was adjusted to p H 7.4 with 5 M HC1. In some islet binding experiments (indicated in text) 3 % B S A was used, since preliminary studies suggested that the increased B S A concentration decreased nonspecific binding. 2. Peptides Lyophilized fractions of HPLC-purified 1 2 5 I -GIP were reconstituted in assay buffer at 4 ° C to a concentration of approximately 1 X 106 cpm/ml immediately before use. Porcine GIP 1-42 ( E G III), synthetic GIP (l-30-NH 2), rat growth-hormone releasing factor l -29-NH 2 (GRF), porcine secretin, porcine vasoactive intestinal polypeptide (VIP), human glucagon-like peptide I (7-36-NH2) and cyclic somatostatin-14 were all weighed out fresh on the day of experiment; porcine glucagon, rat galanin and the C-terminal octapeptide of cholecystokinin, sulfated on the tyrosine residue (CCK-8) were stored as concentrated, lyophilized aliquots at -20 °C. A l l peptides were dissolved in assay buffer at 4 ° C immediately prior to use at a concentration 10X the final desired concentration in the assay. 3. Binding Assays For both /3TC3 cells and isolated islets, binding studies were performed by measuring displacement of 1 2 5 I -GIP (approximately 435 pM) by increasing concentrations of GIP 1-42. Specific binding of 1 2 5 I -GIP was determined by performing incubations of the radiolabeled peptide with and without an excess of unlabeled GIP (1 /xM for /3TC3 cells; 1 76 or 10 /xM for islets). Specificity of the 1 2 5 I -GIP binding was determined by examining the ability of an excess amount (1 /xM) of various other peptides, including most of the known peptides in the glucagon-secretin family that are structurally related to GIP, to displace 1 2 5 I -GIP. A l l incubations were performed for 1 h at 15 °C, since these conditions were previously shown to be optimal for 1 2 5 I -GIP binding to membrane preparations of a hamster /3-cell tumor line (Maletti et al, 1984). a) . /3TC3 Cells /3TC3 cells were cultured and harvested as described in Section I.B.3.a. The cells were plated at a density of 5 X 106 cells/well on 24-well plates (Falcon) and cultured for 2 days before use. On the day of experiment, non-adhered cells were aspirated and discarded. The adhered cells were first washed in 0.5 ml assay buffer at 4 °C, then 0.4 ml assay buffer was added to each well followed by 50 /xl of 1 2 5 I -GIP (approximately 5 X 104 cpm). Next, 50 /xl of 10X concentrated solution of unlabeled GIP or other peptide or 50 /xl assay buffer was added for a total volume of 0.5 ml. Each condition was tested in duplicate. The plate was then covered and incubated at 15 °C. After 1 h, the buffer in each well was gently aspirated using a Pasteur pipet and washed once with 0.5 ml ice-cold assay buffer. The cells were then solubilised in 0.1 M N a O H (0.5 ml) and transferred to test tubes for the measurement of cell-associated radioactivity on a gamma counter, using a different, siliconized Pasteur pipet for each well. Blank tubes were included for the measurement of background radioactivity. After counting for 5 min, the cell extracts of 4 random wells from each plate were diluted to 1 ml by the addition of 0.1 M N a O H , centrifuged (5 min, 500 X g) in 1.5 ml Eppendorf R tubes and the supernatant stored at -20 ° C for the measurement of total protein content (Section III.C). b) . Isolated Islets Islets from 4-8 male Wistar rats were isolated as described and cultured for 2 days in islet culture medium in a petri dish. On the day of experiment, the islets were washed 2X 77 by centrifugation in 15 ml ice-cold HBSS (plus 0.1 % BSA) and then resuspended in HBSS in a petri dish under a dissecting microscope. Islets were then handpicked into 400 /xl polyethylene microcentrifuge tubes (Brinkmann) and stored on ice in 200-400 /xl HBSS until use. In initial experiments, 50 islets per tube were used; however, in some experiments where indicated in the text, 100 islets per tube were used. The assay incubation medium was prepared in a separate set of microcentrifuge tubes at 4 °C. 1 2 5 I -GIP was first added to each tube as a 20 /xl aliquot containing 2 X 104 cpm, followed by 20 /xl unlabeled GIP, other peptide or assay buffer (in duplicate). Assay buffer (60 /xl) was then added for a total volume of 100 /xl and each tube was gently tapped to ensure adequate mixing. The islets were washed by centrifugation (15 s, 500 X g) in a microcentrifuge (Model 152 Microfuge, Beckman) in 200 /xl ice-cold assay buffer. The pellet was resuspended in 100 /xl assay buffer and the suspension of islets in each tube was transferred, using a siliconized Pasteur pipet, to the incubation tubes containing 1 2 5 I -GIP for a total volume in the islet binding assay of 200 /xl. Each tube was gently agitated, sealed and transferred to a rack for incubation at 15 °C. After 1 h, the tubes were transferred to ice, centrifuged (15 s, 500 X g) on a refrigerated table and the supernatant discarded. The islet pellet was resuspended in 200 /xl ice-cold assay buffer; a different, siliconized Pasteur pipet was used for each tube. Following recentrifugation, the tips of each tube containing the islet pellet were cut, transferred to test tubes and counted for 5 min on a gamma counter. After counting, the islet pellet in each tube was solubilised in 0.1 M N a O H (1 ml), transferred to 1.5 ml Eppendorf R tubes, centrifuged (5 min, 500 X g) and the supernatant stored at -20 ° C for protein assay. 4. Calculations Background counts in blank tubes were first subtracted from the counts in each 78 tube. Specific binding in each well was then calculated as the difference between total binding and nonspecific binding (in the presence of excess porcine GIP 1-42 ( E G III)). The total protein content of each islet extract and of extracts of 4 random wells per plate of 0TC3 cells was determined. Since total protein in extracts of /3TC3 cells did not vary appreciably from well to well within a single plate, the mean of 4 random wells was taken as the total protein content in each calculation. 1 2 5 I -GIP binding was then calculated as the specific binding (cpm) per microgram of protein; in all cases the mean of duplicate determinations was used. III. P E P T I D E Q U A N T I F I C A T I O N A . Radioimmunoassay The radioimmunoassay (RIA), a competitive binding assay capable of measuring picomolar quantities of most peptides with high specificity and reproducibility, was used to determine the concentration of immunoreactive islet hormones in incubation media, cell extracts and perfusion samples. A n RI A for GIP was also used to estimate the specific activity of the GIP tracer employed in receptor binding studies. Common units of measurement were used for expression of hormone concentration; the conversion factors for calculation of hormone concentration in Systeme International (SI) units are given in the Appendix. 1. Insulin Immunoreactive insulin (IRI) concentrations in perfusion samples, cell incubation samples and cell extracts were determined using an R I A modified from Albano et al (1972). a) Assay Buffer The diluent buffer used in the insulin R IA consisted of 0.04 M phosphate buffer (pH 7.5) with 5 % charcoal-extracted plasma (CEP). Charcoal-extracted plasma was prepared 79 from outdated human plasma (obtained from the Red Cross blood bank; Vancouver, B.C.) as follows. After removal of erythrocytes by centrifugation (30 min, 1000 X g), the plasma was filtered through sharkskin filter paper (15 cm; Schleicher and Schuell, Inc., Keene, NH) and stirred with 1 % charcoal for 1 h at 4 °C. The charcoal was removed from the slurry by centrifugation (30 min, 1000 X g) and subsequent filtration of the supernatant through sharkskin. Ten milliliter aliquots of the plasma were stored at -20 ° C until use. b) Antiserum A n antiserum (GP01) raised in a guinea pig against rat insulin was used in all insulin RIAs. This antiserum was found not to cross-react in the R I A with GIP, glucagon, somatostatin or galanin at the concentrations used in these experiments. The antiserum was stored at -20 ° C in lyophilized 100 jul aliquots at a dilution of 1:10. When required, these aliquots were reconstituted in insulin RIA buffer to a dilution of 1:5000 and stored until the day of assay as 1 ml aliquots at -20 °C. The final dilution used in the R I A was 1:106 to achieve a "zero" binding (in the absence of unlabeled insulin) to porcine 1 2 5 I -insulin of approximately 50 %. At this dilution, IRI concentrations as low as 10-20 /uU/ml could reliably be detected in samples. c) 1 2 5I-Insulin A l l reagents were prepared immediately prior to the iodination procedure and siliconized glassware was used throughout. The procedure was carried out at room temperature. First, 10-50 /ug of porcine insulin was dissolved in 10 jul H O (0.01 M) and diluted with 0.2 M phosphate buffer (pH 7.4) to a final concentration of 5 /ug/10 /ul. Ten microliters of this solution (5 /ug insulin) was then added to the reaction vessel and gently mixed with 1 mCi N a 1 2 5 I (10 ul). The oxidation reaction was initiated by the addition of 25 jul chloramine T (4 mg/ml in 0.2 M phosphate buffer). After 10 s, 100 /ul sodium metabisulfite (2.4 mg/ml in 0.2 M phosphate buffer) was added and the mixture was gently 80 mixed to terminate the reaction. After a further 45 s, excess iodine (50 /xl potassium iodide; 10 mg/ml in 0.2 M phosphate buffer) was added followed by 1.8 ml of 0.04 M phosphate buffer. Purification of 1 2 5I-insulin from free 1 2 5 I was achieved by adsorption to silica as follows. The reaction mixture (approximately 2 ml) was added to a test tube containing 10 mg microfine silica (QUSO G-32). After thorough vortexing, the solution was centrifuged (30 s, 100 X g) and the supernatant containing free 1 2 5 I was discarded. The silica pellet, to which the 1 2 5I-insulin was adsorbed, was washed twice by centrifugation in 3.0 ml distilled H 2 0 to remove additional unincorporated 1 2 5 I . Finally, the 1 2 5I-insulin was eluted from the silica by the addition of 3 ml acid ethanol. After centrifugation, the supernatant was diluted by the further addition of 2 ml acid ethanol and 1.5 ml distilled H 2 0 . This solution was then stored at -20 °C and could be used for up to 6 weeks. Percent incorporation of 1 2 5 I into the peptide was determined after each iodination. The silica pellet and ten microliter aliquots of the dilute 1 2 5I-insulin in acid ethanol ( C a c i d ethanol) a n ^ the silica/reaction mixture prior to centrifugation (Qotal) were counted for 0.1 min, corrected for volume and percent incorporation calculated by the following formula: % Incorporation = ( C a c i d ethanol + Qil ica pelletVQotal where C = cpm. Percent incorporation was generally 40-70 %. For use in the RIA, an aliquot of the 1 2 5I-insulin stock was usually diluted in an appropriate volume of assay buffer to achieve approximately 1 X 104 cpm/100 /xl. The dilution was later increased to 2000 cpm/100 /xl. This did not affect the sensitivity or accuracy of the RIA, as confirmed by a comparison of the IRI concentration of control samples (Section III.Al.e) determined using either concentration of 1 2 5I-insulin. d). Standards One hundred micrograms of pure, lyophilized rat insulin (activity approximately 25 U/mg) was dissolved in a small volume of distilled H 2 0 and diluted to 4.26 U / m l 81 (approximately 200 ng/ml) in 0.04 M phosphate buffer (pH 7.4) containing 6 % B S A (Pentex R ), 0.24 g/1 sodium merthiolate, and 6.0 g/1 NaCl. This stock solution was stored as 1 ml aliquots at -20 °C. When required, these aliquots were further diluted 1:26.6 in assay buffer to achieve a concentration of 160 /xU/ml and frozen as 2 ml aliquots until the day of assay. For use in the RIA, an aliquot was thawed and serially diluted in assay buffer to obtain standard curve insulin concentrations of 160, 80,40, 20,10 and 5 /xU/ml. e) . Controls Inter- and intra-assay variation were monitored using appropriate control tubes in every assay. Controls were obtained by pooling samples collected from perfused rat pancreas preparations stimulated with 17.8 m M glucose and 10 m M 1-arginine for 30 min. Immunoreactive insulin content of the pooled effluent was determined and it was diluted in assay buffer to achieve a concentration of 60 /zU/ml. Two milliliter aliquots of the control samples were stored at -20 °C until needed. Deviation of control values of greater than 10 MU/ml from 60 juU/ml invalidated the assay. Inter- and intra-assay variation of these control samples were 9 ± 2 % and 7 ± 3 % , respectively. f) . Separation Separation of bound and free 1 2 5I-insulin was achieved using dextran-coated charcoal. The charcoal was prepared by dissolving 5.0 g/1 dextran (T-70) in 0.04 M phosphate buffer (pH 7.4) and adding charcoal (50.0 g/1). The slurry was stirred overnight at 4 ° C for at least 1 h prior to use in the assay. g) . Procedure Assays were set up in 12 X 75 mm borosilicate glass test tubes. A l l procedures were carried out on a refrigerated table at approximately 4 °C. Each R I A included total count, standard, zero and control tubes (assayed in triplicate) and sample and non-specific binding (NSB) tubes (assayed in duplicate). Separate NSB tubes were included for standards, controls and samples. When necessary, samples were diluted in assay buffer so that the IRI 82 concentration fell within the range of the steepest slope of the standard curve. Binding of antisera and peptide was allowed to reach equilibrium prior to the addition of iodinated insulin (disequilibrium assay). One hundred microliters of either sample, control, standard or buffer were first added to tubes containing 700 pi assay buffer. After vortexing, 100 jul antiserum were added to each tube except the NSB tubes, and the mixtures were vortexed again and incubated for 24 h at 4 °C. 1 2 5I-insulin (100 iii) was then mixed into each tube for a total assay volume of 1 ml. The assays were incubated for a further 24-48 h prior to separation. Two hundred microliters of dextran-coated charcoal were added to each tube (except total count tubes) and the mixture was vortexed, centrifuged for 30 min (1000 X g, 10 °C) , and the supernatant (containing antibody-bound peptide) decanted. After drying (at least 4 h), the charcoal pellets (containing unbound peptide) were counted for either 2 min (1 X 104 cpm/tube) or 3 min (2000 cpm tube) on a gamma counter. The insulin RIA was also routinely performed at half-volume (0.5 ml) when the volume of sample collected was insufficient for a full-volume assay. h). Calculations The concentration of IRI in each sample was determined using an R I A software program (RIA Analysis vl.0) on a microcomputer. This program transformed standard curve data into a logit-log plot. The program was then used to calculate the concentration of IRI in controls and unknown samples from the mean of duplicate counts. 2. Glucagon Glucagon-like immunoreactivity (GLI) was determined in the islet incubation media and cell extracts by an RIA for pancreatic glucagon modified from Gregor and Riecker (1985). a). Assay Buffer The buffer used in the G L I RIA was a 0.06 M phosphate buffer (pH 7.4) containing 0.01 M ethylenediaminetetraacetic acid ( E D T A ) , 0.2 M NaCl, 3 % B S A (Pentex R ) , 0.5 g/1 83 sodium azide and 100 K.I .U./ml aprotinin. b) . Antibody A monoclonal antibody (23.6.B4) raised to the N-terminal of pancreatic glucagon was used in the assay. This antibody was previously shown to detect pancreatic glucagon standards as low as 2-3 fmol/tube (approximately 10 pg/tube) with less than 0.01 % cross-reactivity with structurally-related GIP. The antibody was stored (as supplied) at 4 ° C in hybridoma culture medium (Gregor and Riecker, 1985). It was diluted in assay buffer 1:1 X 104 for a final dilution of 1:1 X 105 in the assay. In some assays, the antibody was used at a final dilution of 1:2 X 105 to increase sensitivity. At this dilution, zero binding was approximately 35 %. Standard curves were usually sensitive to concentrations of glucagon in the range of 50 to 1000 pg/ml. c) . 1 2 5I-Glucagon Monoiodinated 1 2 5I-glucagon with specific activity of 2000 Ci/mmol was purchased in lyophilized form (Amersham Canada Ltd., Oakville, Ont) . The iodinated peptide was reconstituted in 1 ml distilled H 2 0 and further diluted in assay buffer to obtain approximately 8 X 105 cpm/ml. Aliquots (1 ml) were then dispensed into 1.5 ml Eppendorf R tubes and could be stored for up to 8 weeks at -20 °C. When required, an aliquot was thawed and diluted in 40 ml assay buffer to 2000 cpm/100 /xl. d) . Standards Highly purified, porcine pancreatic glucagon, which has an identical amino acid sequence to rat pancreatic glucagon, was dissolved to a concentration of 2 mg/ml in a solution consisting of 0.1 M formic acid, 0.14 M lactose, 11 m M citric acid, 6 m M 1-cysteine, 1000 K.I .U/ml aprotinin and 0.4 m M B S A (Pentex R ). It was lyophilized in 10 /xg (5 /xl) aliquots and stored at -20 °C. For use in the assay, an aliquot was reconstituted in 1 ml assay buffer and further diluted 1:1562.5 (256 /xl:4 ml) in assay buffer to achieve a concentration of 6400 pg/ml. This was then serially diluted to obtain a range of standards 84 including 6400, 3200, 1600, 800, 400, 200,100,50 and 25 pg/ml. e) . Controls Perfusate samples and incubation media containing G L I could not be reliably stored for extended periods of time without significant loss of measurable GLI, precluding the use of these samples as controls. Therefore, inter-assay variation was determined by the inclusion of normal glucagon standards (200 or 400 pg/ml) in the assay, while intra-assay variation was calculated by comparing duplicate determinations of G L I in random samples inserted at the beginning and end of most RIAs. The inter- and intra-assay variation determined by this method were 1 6 ± 5 % and 11±4%, respectively. f) . Separation Dextran-coated charcoal was used to separate bound and free 1 2 5I-glucagon. The charcoal was prepared by dissolving 2.5 g/1 dextran (T70) in 0.04 M phosphate buffer (pH 6.5) followed by the addition of 12.5 g/1 charcoal. The slurry was left stirring at 4 ° C overnight prior to addition to the assay. On the day of use, 7 % C E P was added and the mixture stirred for at least 1 h before use. g) . Procedure A disequilibrium assay protocol was followed for the glucagon RIA, identical to that used for insulin with the following exceptions. First, after addition of antibody and samples the assay was incubated for 48 h before the addition of 1 2 5I-glucagon, after which the assay was incubated for a further 24 h prior to the addition of charcoal. Second, 500 n\ of dextran-coated charcoal was added to each tube for separation. Finally, the assay was counted for 3 min. h) . Calculations The percent bound (% B ) 1 2 5I-glucagon for each sample and standard was calculated as follows: % B = 100 X ( C N S B " Qample)/Qotal> w ^ E R E ^ = C P M A N ( * ^ S B = non-specific binding. A standard curve of % B (ordinate) vs. glucagon standard 85 concentration (abscissa) was then drawn on semilogarithmic graph paper. The concentration of G L I in samples was determined from the calculated % B by reading from the standard curve. 3. Somatostatin The RIA used for measurement of somatostatin-like immunoreactivity (SLI) in islet incubation medium, cell extracts and perfusion samples has been previously described (Mcintosh et al, 1978; 1987). a) . Assay Buffer Stock buffer for the SLI RIA was prepared by dissolving sodium barbital (4.90 g), sodium acetate (0.32 g), NaCl (2.55 g) and ethylmercurithiosylicyclic acid sodium salt (merthiolate; 0.10 g) in 700 ml distilled H 2 0 , adjusting to p H 7.4 with HC1 and diluting to a final volume of 11 with distilled H 2 0 . This was stored at 4 ° C until required. The buffer used in the assay was prepared by adding 5.0 g/1 B S A (Pentex R ) and 100 K.I .U./ml (1 %) aprotinin to this stock buffer. b) . Antibody A monoclonal antibody ( S O M A 03) to somatostatin (Buchan et al, 1985) was used in the RIA. Previous characterization of this antibody for use in the R I A showed that it was effective at a final dilution of 1: 4 X 106, did not cross-react with GIP and detected somatostatin-14 and somatostatin-28 with equal sensitivity (Mcintosh et al, 1987). Stock solution of the antibody from crude mouse ascites was diluted 1:1 X 104 in assay buffer to obtain a titer of 1:1 X 106 (1:4 X 106 final dilution in the assay). c) . 125I-Somatostatin 1 2 5I-Tyr1-somatostatin was prepared by chloramine T iodination of synthetic Tyr 1 -somatostatin followed by purification of the iodinated peptide by adsorption to silica. It was stored as lyophilized aliquots of 1 X 106 cpm. O n the day of assay, aliquots were reconstituted in 2 m M ammonium acetate buffer (pH 4.6) and purified on a CM-cellulose 86 (Sephadex CM-52) column (0.9 X 10 cm). After equilibration with 2 m M ammonium acetate, the label was applied to the column and then washed by pumping 20 ml of 2 m M buffer through the column (flow rate: 1 ml/min). ^I-Tyr^somatostatin was then eluted with 0.2 M ammonium acetate (pH 4.6). The radioactivity in each fraction (2 min/fraction) was counted, peak fractions pooled, neutralized with 2 M N a O H and diluted in assay buffer to 3500 cpm/100 /xl for use in the assay. d) . Standards Standards were prepared by dissolving synthetic cyclic somatostatin-14 in acetic acid (0.1 M) and B S A (Pentex ; 0.5 g/1) to obtain a concentration of 100 Mg/ml. Aliquots of 50 /xl (5 /itg) were then lyophilized and stored at -20 ° C until use. On the day of assay, an aliquot was reconstituted in 100 /xl distilled H 2 0 followed by 400 /xl assay buffer to obtain a concentration of 10 /xg/ml. This was further diluted in assay buffer (1:2 X 104) to a concentration of 500 pg/ml. Serial dilution of this solution in assay buffer generated the standards used in the assay: 500, 250,125, 62.5, 31.25,15.6, 7.8 and 3.9 pg/ml. e) . Controls As with glucagon, samples containing SLI could not be stored for several months without loss of immunoreactivity. Inter- and intra-assay variation were therefore determined as for glucagon and were found to be 1 2 ± 3 % and 5 ± 1 % , respectively. f) . Separation Dextran-coated charcoal was prepared by dissolving 2.5 g/1 dextran (T70) in 0.05 M phosphate buffer (pH 7.5), adding activated charcoal (12.5 g/1) and stirring overnight at 4 °C prior to use. Before addition to the assay, 0.1 % C E P was added to the charcoal mixture and it was stirred for at least 1 h. g) . Procedure The protocol was similar to the insulin and glucagon R I A with the following differences. A l l assay constituents were added to the tubes the same day in the following 87 order: 100 standard or sample, 100-300 iA assay buffer, 100 /d antibody and 100 fj.1 label for a total volume of 400 ^I/tube. In assays for SLI in samples of perfused pancreas effluent or islet incubation medium, 100 ul perfusate or release medium, respectively, was added to standard and zero tubes in place of assay buffer. Release medium was used to dilute samples from islet incubation studies, while dilution of lyophilized islet extracts was done in assay buffer. After addition of ^I-Tyr^somatostatin, the assays were vortexed and incubated for approximately 72 h at 4 °C. One milliliter dextran-coated charcoal (plus CEP) was then added to each tube (except total counts) for separation of bound and unbound peptide. After centrifugation and decantation, the charcoal pellets were counted for 3 min on a gamma counter, h). Calculations The concentration of SLI in samples was determined by calculation of the % B and reading of the SLI concentration from a hand-drawn standard curve (as for glucagon). 4. Gastric Inhibitory Polypeptide A n RIA for GIP was used, in addition to an enzyme-linked immunosorbent assay (ELISA), to estimate the concentration of GIP in samples of B-GIP, 1 2 7 / 1 2 5 I - G I P and HPLC-purified GIP used in perfusions and GIP receptor studies. Standard curves for GIP were also generated using different peaks of HPLC-purified 1 2 5 I -GIP to estimate the specific activity of these products of 1 2 5 I-GB? purification. The procedure for this R I A was modified from the technique of Kuzio et al (1974). a) . Assay Buffer The buffer was 0.04 M phosphate buffer (pH 6.5) containing 0.5 % C E P and aprotinin (200 K.I.U/ml). b) . Antiserum The antiserum used was RK343F which has been shown to have negligible cross-reactivity with insulin and glucagon (Morgan et al, 1978). It was stored as 5 jul aliquots 88 lyophilized neat and stored at -20 °C. When required, aliquots were reconstituted in 15 ml assay buffer (1:3000 dilution); the final dilution in the assay was 1:15000. At this titer, the antiserum was shown to detect immunoreactive GIP (IR-GIP) with a sensitivity of 11 pg per tube (Morgan et al, 1978). c) . 1 2 5 I -GIP 1 2 5 I -GIP was prepared as described in section II.A.l.a. For use in the RIA, the 1 2 5 I -GIP in acid ethanol was diluted in assay buffer to a concentration of 5000 cpm/100 ul. d) . Standards Porcine GIP ( E G III; approximately 100 ixg) was dissolved in 0.2 M acetic acid plus 0.5 % B S A (Pentex R ) and aprotinin (200 K.I.U./ml) to a concentration of 100 Mg/ml. Aliquots of 1 Mg (10 M0 were then lyophilized and stored until use at -20 °C. When needed, an aliquot was dissolved in 100 y\ acetic acid (0.1 M) ; a further 100 ixl of assay buffer was added to achieve a concentration of 5 Mg/ml. This was further diluted in assay buffer 40 M1:10 ml to a concentration of 20 ng/ml. Serial dilution of this solution in assay buffer yielded a range of standard concentrations including 2000, 1000, 500, 250, 125, 62.5, 31.2, 15.6 and 7.8 pg/100/xl. For determination of the concentration of B-GIP in some samples, a standard curve was constructed using synthetic GIP l-30-NH 2 . A small amount of GIP l -30-NH 2 (5-10 Mg) was weighed out and immediately diluted in assay buffer to a concentration of 20 ng/ml; this standard was then serially diluted to produce a range of standards as for GIP 1-42. e) . Separation Bound and unbound 1 2 5 I -GIP were separated by adsorption of bound peptide to polyethylene glycol ( P E G 8000; Carbowax R ). The P E G solution was prepared by dissolving 250 g/1 P E G in distilled H 2 0 and allowing to stir at 4 ° C for 1 h. f) . Procedure The protocol for this RIA was similar to that described for the previous RLAs with 89 minor differences. The total volume of the assay was 500 /xl- Lyophilized samples of HPLC-purified B-GIP, 1 2 7 / 1 2 5 I - G I P and GIP were reconstituted and diluted in assay buffer. Standards or samples (100 /xl) and antiserum (100 /xl) were added to tubes containing 200-400 /xl assay buffer to a volume of 400 /xl; the tubes were then vortexed and incubated at 4 ° C for 24 h. One hundred microliters 1 2 5 I -GIP (5000 cpm) was added to each tube and the assays were incubated for a further 24 h. Bound and unbound peptide were separated by the addition of P E G (500 /xl) to each tube. The assays were vortexed well and centrifuged at 1000 X g for 45 min. The supernatant was gently decanted and after drying, the P E G pellets were counted on a gamma counter for 2 min. B. Enzyme-Linked Immunosorbent Assay (ELISA) The E L I S A was developed as a rapid method for screening antibodies. However, the E L I S A has recently been shown to be a fast, sensitive and reproducible method of measuring the concentration of peptide hormones (e.g. insulin) without requiring the use of radioactive isotopes as in the R I A (MacDonald and Gapinski, 1989). In this study, the E L I S A was used to estimate the amount of immunoreactive GIP (IR-GIP) in lyophilized aliquots of HPLC-purified fractions of 1 2 7 / 1 2 5 I - G I P . This was necessary because the trace amounts of radioactive 1 2 5 I -GIP present would interfere with the RIA. The E L I S A was also used, in parallel with the RIA, to estimate the amount of IR-GIP in samples of B-GIP. 1. Gastric Inhibitory Polypeptide Although an E L I S A for measuring GIP concentrations has not been previously described, the technique has been routinely used in this laboratory to screen monoclonal antibody-producing clones for the production of antibodies to GIP. The procedure involves the coating of multi-well plates with GIP followed by the addition of hybridoma incubation media possibly containing GIP antibodies. Antimouse immunoglobulin, conjugated to alkaline phosphatase, is then added to each well. The presence of bound antibody is detected by addition of phosphatase substrate, which changes absorbance after 90 reacting with the enzyme in wells with bound antibody. This method was modified as follows to allow the estimation of IR-GIP levels in samples. a) . Assay Buffers The carbonate-bicarbonate buffer used to coat the plates with antigen was a solution of of N a 2 C 0 3 (1.59 g/1), NaHCC«3 (2.93 g/1) and N a N 3 (0.20 g/1). It was adjusted to p H 9.6 with concentrated N a O H and stored at 4 ° C for no longer than 3 weeks. PBS-Tween buffer (PBS plus 5 ml/1 Tween-20 R ) was made as a 10X concentrated stock solution, stored at room temperature and diluted in distilled H 2 0 when needed. Diethanolamine buffer (10 %) was prepared in a container sealed from light by combining 97 ml/1 diethanolamine, 0.2 g/1 N a N 3 and 100 mg/1 M g C l 2 . Concentrated HC1 was added dropwise until p H 9.8 was achieved and the solution was stored in the dark at 4 ° C until needed. b) . Antisera Preliminary standard curves generated with different GIP antisera at various titers suggested that a monoclonal antibody (575), raised to porcine GIP in this laboratory, at a final concentration of 1 Mg/ml was most effective at detecting porcine GIP 1-42 and GIP 1-30 in the ELISA. It was prepared immediately prior to use in the E L I S A by dilution of a 100 Mg/ml stock solution stored at -20 ° C in PBS-Tween. c) . Standards Standards of porcine GIP 1-42 ( E G III) and synthetic GIP (l-30-NH 2) were prepared by weighing a small amount (5-10 Mg) of either peptide and dissolving in PBS-Tween to a concentration of 1 Mg/ml. This concentration of peptide was then serially diluted in PBS-Tween in siliconized glass test tubes to generate a range of standard concentrations from 1 Mg/ml to 1 ng/ml. d) . Procedure Lyophilized samples of B-GIP and 1 2 7 / 1 2 5 I - G I P were reconstituted and diluted in 91 carbonate-bicarbonate buffer. Standards and samples were then added (in duplicate) as 100 jul aliquots to a 96 well plate (Falcon 3912) and the plate covered in foil. Following overnight incubation at 4 ° C to allow the peptide to adhere to the plate, the buffer was discarded and each well washed 3X with PBS-Tween. One hundred microliters of fetal calf serum (1 % in PBS-Tween) was added to each well to block non-specific binding sites on the plate. After 1.5 h incubation at room temperature, each well was again washed 3X in PBS-Tween and 100 jzl of the GIP antisera added. Following a further 1 h at room temperature, the plate was washed as before and 100 ul dilute (1:1000) alkaline phosphatase conjugated antimouse immunoglobulin (Tago, Burlington, C A ) in PBS-Tween was added to each well. The plate was then incubated for 1 h at room temperature in the dark, and unbound conjugated antisera was discarded by washing in PBS-Tween (3X). The alkaline phosphatase reaction was then initiated by the addition of 100 /xl/well of a solution of phosphatase substrate (1 mg/ml) in diethanolamine buffer. This solution was prepared in the dark immediately prior to use by dissolving 1 tablet (5 mg/tablet) of disodium p-nitrophenyl phosphate (104-Phosphatase Substrate, Sigma) in 5 ml diethanolamine buffer. The reaction was allowed to proceed in the dark at room temperature until the phosphatase reaction had sufficiently developed, then the absorbance (490 nm) in each well was determined on a Microelisa Autoreader R MR580 (Dynatech Laboratories, Inc., Alexander, V A ) . A standard curve of absorbance (ordinate) vs. GIP concentration (abscissa) was drawn and IR-GIP levels in samples determined from this curve, e). Controls The accuracy of this E L I S A was validated by cross-determination of the amount of IR-GIP in lyophilized samples of B-GIP using the GIP RIA, and by inclusion in the E L I S A of a lyophilized GIP RIA standard reconstituted in carbonate-bicarbonate buffer to a concentration of 100 ng/ml. The IR-GIP concentration of this standard as determined by E L I S A was 86±13 % (n=3). 92 C. Protein Assay Total protein content of wells containing islets or /0TC3 cells in GIP receptor binding assays was measured using a commercial protein assay kit (BioRad Laboratories, Richmond, C A ) . The microassay protocol, capable of detecting protein levels as low as 5 Mg, was used. Lyophilized albumin standard (BioRad) was reconstituted in distilled H2O to a concentration of 1.4 mg/ml and stored as 0.5 ml aliquots at -20 °C. When needed, an aliquot was diluted to produce a range of standard concentrations including 5, 10, 15, 20 and 25 /xg/ml. Since islet and /3TC3 cell extracts were stored in 0.1 M N a O H , standards were diluted with this solution. Sample or standard (800 JLXI) plus 200 /xl concentrated protein dye reagent (BioRad) were mixed well in test tubes and allowed to stand for at least 5 min. Absorbance (595 nm) was then measured for each standard and sample and a standard curve of absorbance (ordinate) vs. mass of protein (/xg; abscissa) drawn. The amount of protein in each sample was then determined from the curve. IV. A N A L Y S I S O F D A T A A . Isolated Islets, FACS-Purified /3-Cells and /3TC3 Cell Secretion In isolated islets, FACS-purified /3-cells and /3TC3 cell secretion experiments, hormone release was calculated as a percent of the total cell or islet content of that peptide for each well. The data were expressed as the mean ± S.E.M. of at least 3 experiments (as indicated) with each condition tested in triplicate. Statistical significance was determined using one way analysis of variance with either Dunnett's test (when multiple comparisons were being made to the same control), or Scheffe's test (for multiple comparisons between different groups). B. Perfused Pancreas 1. Galanin Experiments The data were expressed as the mean (± S.E.M.) release of IRI (juU/min) or SLI 93 (pg/min) during 5 min collection periods for these experiments. The percentage change in hormone release during infusion of galanin or vehicle was then calculated as follows: [(release during 10 min galanin or vehicle infusion - release during 10 min prior to infusion)/release during 10 min prior to infusion] X 100. For statistical comparison, the percentage change during vehicle infusion was compared to the percentage change during rat or porcine galanin infusion using the Mann-Whitney U test. 2. GIP Gradient Experiments Total IRI output (mU/45 min) over the entire perfusion was calculated for each pancreas and was expressed as the mean (± S.E.M.) for each condition. Mean IRI output during perfusion with GIP 1-42, 1 2 7 / 1 2 5 I-GIP, GIP 1-30 B-GLP, or glucose alone were then compared using using the Mann-Whitney U test. In all cases, p < 0.05 was considered statistically significant. 94 RESULTS I. E F F E C T S O F GIP O N H O R M O N E S E C R E T I O N F R O M I S O L A T E D R A T P A N C R E A T I C ISLETS A . Insulin Secretion 1. Effect of Glucose Concentration on GIP-Stimulated Insulin Secretion After two days culture, isolated rat islets were incubated in medium containing glucose concentrations ranging from 2.75 m M to 17.8 m M in the absence or presence of GIP (1 or 10 nM). At 2.75 m M glucose and in the absence of GIP, 1.11±0.36 % of the total islet immunoreactive insulin (IRI) content was released in 1 h (Figure 6). Increasing the glucose concentration of the medium significantly enhanced IRI secretion only at the highest glucose concentration tested (17.8 mM). When GIP (1 n M or 10 nM) was added to the incubation medium, no change in IRI secretion compared to glucose alone was observed at low glucose concentrations (2.75 or 4.4 mM). Both concentrations of GIP produced slight but statistically insignificant increases in IRI release from the islets in the presence of 6.6 m M glucose. However, a marked enhancement of IRI secretion in response to GIP was seen in the presence of either 8.9 or 17.8 m M glucose. Both 1 and 10 n M GIP induced similar increases (approximately 3X) in IRI output compared to control (Figure 6). 2. Concentration-Dependency of GIP-Stimulated Insulin Secretion in Low (2.75 mM) and High (17.8 mM) Glucose Isolated islets incubated in 2.75 m M glucose with increasing concentrations of GIP (0.1-100 nM) were significantly stimulated by only the highest concentration of GIP tested (100 nM). The IRI response to 2.75 m M glucose plus 100 n M GIP 95 (2 .19±0.80 %) was slightly greater than the response to 2.75 m M glucose alone (1.10±0.47 %) (Figure 7). In the presence of 17.8 m M glucose, IRI release was enhanced 2X compared to 2.75 m M glucose (2 .03±0.40 % vs. 1,11±0.36 %). At the higher glucose levels, a concentration-dependent IRI response to GIP was observed. The stimulation of IRI release by GIP in 17.8 m M glucose was statistically significant at GIP concentrations of 1 n M or greater and appeared to be maximal at 10 n M GIP, which produced a 3X increase in IRI secretion compared to 17.8 m M glucose alone. 3. Modulation of GIP-Stimulated Insulin Secretion by Acetylcholine and Arginine In the presence of 2.75 m M glucose, arginine (10 mM) and GIP (10 nM) each produced slight, statistically insignificant increases in IRI secretion from isolated islets (Figure 8A). When islets were incubated with these concentrations of arginine and GIP together, a 2.5X increase in IRI release compared to 2.75 m M glucose alone was observed. However, the response to 10 m M arginine plus 10 n M GIP (2.75±0.78 %) was slightly less than the sum of the individual responses to arginine and GIP alone (3.48±0.72 %). Acetylcholine (ACh; 1 nM) and GIP (10 nM) each significantly enhanced IRI release from isolated islets in the presence of 17.8 m M glucose (Figure 8B). When islets were subjected to stimulation by both GIP (10 nM) and A C h (1 /xM) in 17.8 m M glucose, IRI secretion was increased 3.5 X compared to 17.8 m M glucose alone. The response to stimulation by both A C h and GIP together (7 .08±1.20 %) was again slightly less than would be expected if the responses to these two secretagogues were additive (10.30±1.97 %). 96 B. Glucagon Secretion 1. Effect of Glucose Concentration on GIP-Stimulated Glucagon Secretion The secretion of glucagon-like immunoreactivity (GLI) from isolated islets in response to GIP (1 or 10 nM) was measured in the presence of different glucose concentrations (2.75-17.8 mM). In the absence of GIP, G L I release was similar (1-1.5 % of total islet G L I content in 1 h) at all glucose concentrations tested except 4.4 m M glucose, where G L I secretion was elevated approximately 2X (2.51±0.41 %). Both concentrations of GIP tested (1 and 10 nM) significantly enhanced G L I release only in the presence of 2.75 m M glucose (Figure 9). At higher concentrations of glucose (4.4-17.8 mM), no effect of GIP on G L I secretion was consistently observed. At intermediate glucagon concentrations (6.6 and 8.9 mM), GIP had no significant effect on G L I release, although in the presence of the higher concentration of GIP (10 nM) G L I secretion tended to increase. 2. Concentration-Dependency of GIP-Stimulated Glucagon Secretion in 2.75 m M Glucose The stimulatory effect of GIP on G L I secretion in 2.75 m M glucose was more closely examined using a range of GIP concentrations from 0.1 to 100 n M . A modest stimulation of G L I secretion was apparent at GIP concentrations of 1 n M or greater (Figure 10). The maximally effective concentration of GIP appeared to be 10 n M (1.95±0.34 % vs 1.2±0.14 %). 3. Modulation of GIP-Stimulated Glucagon Secretion by Arginine The addition of 10 m M arginine to islet incubation medium containing 2.75 m M glucose had no significant effect on G L I secretion (Figure 11), although G L I release tended to be slightly higher when arginine was present. When islets were exposed to both GIP (10 nM) and arginine (10 mM) in 2.75 m M glucose, the G L I response (2 .34±0.38 %) was not significantly enhanced compared to the response to GIP (1 .79±0.40 %) or arginine (1.75±0.33 %) alone. 97 C. Somatostatin Secretion 1. Effect of Glucose Concentration on GIP-Stimulated Somatostatin Secretion In the absence of GIP, alteration of the glucose concentration of the incubation medium had no significant effect on the secretion of somatostatin-like immunoreactivity (SLI). Further, neither concentration of GIP tested (1 or 10 nM) affected SLI secretion in the presence of 4.4 or 8.9 m M glucose (Figure 12). However, in 2.75 m M glucose, both 1 and 10 n M GIP produced modest increases in SLI secretion compared to 2.75 m M glucose alone. At higher glucose concentrations (8.9 and 17.8 mM), slight (approximately 1.5X) increases in SLI release in response to GIP were also observed. In 8.9 m M glucose, the stimulatory effect of GIP on SLI secretion was statistically significant at a GIP concentration of 10 n M but not 1 n M , whereas in 17.8 m M glucose, the SLI response was greater at the lower GIP concentration (1 nM) (Figure 12). 2. Concentration-Dependency of GIP-Stimulated Somatostatin Secretion in 17.8 m M Glucose In the presence of 17.8 m M glucose, addition of GIP (0.1-100 nM) augmented SLI release over basal levels. The lowest concentration of GIP to produce a significant increase in SLI release was 1 n M (Figure 13). The stimulatory effect of GIP appeared to be concentration-dependent, although the SLI response to 1 n M GIP (1.15±0.24 %) was similar to that observed in the presence of 10 n M GIP (1 .09±0.19 %). The greatest stimulation of SLI secretion was seen at the maximum concentration of GIP tested (100 nM). 3. Modulation of GIP-Stimulated Somatostatin Secretion by Acetylcholine SLI secretion in the presence of 17.8 m M glucose (0 .62±0.08 %) was unaffected by the addition of 1 /xM A C h (0.89±0.10 %). However, stimulation of 98 SLI release from cultured islets by 17.8 m M glucose plus 10 n M GIP (1.20±0.19 %) was suppressed by 1 /LIM A C h to levels similar to that seen in 17.8 m M glucose alone (0.87±0.14 %) (Figure 14). 99 F I G U R E 6: Effect of Glucose Concentration on GIP-Stimulated Insulin Secretion from Isolated Rat Islets. Islets (20) were incubated for 1 h in glucose concentrations ranging from 2.75 to 17.8 mM in the absence of GIP (- • -) or in the presence of 1 n M (- O -) or 10 nM (- • -) GIP. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total islet IRI content. *p <0.05 compared to glucose alone, n =4 for all conditions. 100 c 3 c o o w <o o CD rr 8 7 6 5 4 3 2 1 0-.1 1 [GIP] (nM) 1 0 100 F I G U R E 7: Concentration-Dependency of GIP-Stimulated Insulin Secretion from Isolated Rat Islets. Islets (20) were incubated for 1 h in GIP concentrations ranging from 0.1 to 100 n M in either 2.75 (- O -) or 17.8 (- • -) m M glucose. Data is expressed as the m e a n ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total islet IRI content. *p<0.05 compared to 2.75 or 17.8 m M glucose alone. n = 6 for all conditions. 101 B 2.75 mM + GIP + Arg + Arg Glucose (10 nM) (10 mM) +GIP 1 0T CO c o u <D re rr. 8-6-4-2-T T T 0 - M L ^ _ l U ^ l U ^ J 17.8 mM +GIP +ACh +ACh Glucose (10 nM) (1 U.M) +GIP F I G U R E 8: Modulation of GIP-Stimulated Insulin Secretion from Isolated Rat Islets by Acetylcholine and Arginine. Islets (20) were incubated for 1 h in (A) 2.75 m M glucose alone or plus either 10 n M GIP, 10 m M arginine (Arg) or 10 n M GIP plus 10 m M Arg as indicated; or (B) 17.8 m M glucose alone or plus either 10 n M GIP, 1 M M acetylcholine (ACh) or 10 n M GIP plus 1 / i M A C h as indicated. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total islet IRI content. *p<0.05 compared to (A) 2.75 or (B) 17.8 m M glucose alone; § p < 0 . 0 5 compared to 17.8 m M glucose plus 1 nM A C h . n=6 for all conditions. 102 4 i c CD ••—> C o o 3 -^ 2-0 CO cd o CD 1 1 1 1 — 2.75 4.4 6.6 8.9 17.8 [Glucose] (mM) F I G U R E 9: Effect of Glucose Concentration on GIP-Stimulated Glucagon Secretion from Isolated Rat Islets. Islets (20) were incubated for 1 h in glucose concentrations ranging from 2.75 to 17.8 m M in the absence of GIP (-•-) or in the presence of 1 n M (-o-) or 10 n M (-•-) GIP. Data is expressed as the mean ( ± S E M ) glucagon-like immunoreactivity (GLI) released as a percent of total islet G L I content. *p<0.05 compared to glucose alone. n=4 for all conditions. 103 F I G U R E 10: Concentration-Dependency of GIP-Stimulated Glucagon Secretion from Isolated Rat Islets. Islets (20) were incubated for 1 h in GIP concentrations ranging from 0.1 to 100 nM in 2.75 m M glucose. Data is expressed as the mean ( ± S E M ) glucagon-like immunoreactivity (GLI) released as a percent of total islet G L I content. *p<0.05 compared to 2.75 m M glucose alone. n=5 for all conditions. 104 C CD *-» C o o CD CO CO © CD rr 2.75 mM + GIP + Arg + Arg Glucose (10 nM) (10 mM) + GIP F I G U R E 11: Modulation of GIP-Stimulated Glucagon Secretion from Isolated Rat Islets by Arginine. Islets (20) were incubated for 1 h in 2.75 m M glucose alone or plus either 10 n M GIP, 10 m M arginine (Arg) or 10 n M GIP plus 10 m M Arg as indicated. Data is expressed as the mean ( ± S E M ) glucagon-like immunoreactivity (GLI) released as a percent of total islet IRI content. *p<0.05 compared to 2.75 m M glucose alone. n=5 for all conditions. 105 5n DC • 1 -CO 0 H 1 1 1 1 1 2.75 4.4 6.6 8.9 17.8 [Glucose] (mM) F I G U R E 12: Effect of Glucose Concentration on GIP-Stimulated Somatostatin Secretion from Isolated Rat Islets. Islets (20) were incubated for 1 h in glucose concentrations ranging from 2.75 to 17.8 m M in the absence of GIP or in the presence of 1 n M (-O) or 10 n M (-•-) GIP. Data is expressed as the mean ( ± S E M ) somatostatin-like immunoreactivity (SLI) released as a percent of total islet SLI content. *p< 0.05 compared to glucose alone, n=4 for all conditions. 106 C CD • C o o CD GO cd CD CD cc CO 2.5-1 2.0 1.5-1.0 0.5-0.0 H .1 1 [GIP] (nM) •m, 10 100 F I G U R E 13: Concentration-Dependency of GIP-Stimulated Somatostatin Secretion from Isolated Rat Islets. Islets (20) were incubated for 1 h in GIP concentrations ranging from 0.1 to 100 nM in the presence of 17.8 m M glucose. Data is expressed as the mean ( ± S E M ) somatostatin-like immunoreactivity (SLI) released as a percent of total islet SLI content. "p<0.05 compared to 17.8 m M glucose alone, n=6 for all conditions. 107 2.0' c <D *-* c o o CD CO CO _Q) CD rr CO 1.5 1.0-0.5-0.0 17.8 mM +GIP +GIP Glucose (10 nM) + ACh (1 jiM) F I G U R E 14: Modulation of GIP-Stimulated Somatostatin Secretion from Isolated Rat Islets by Acetylcholine. Islets (20) were incubated for 1 h in 17.8 m M glucose alone or plus either 10 n M GIP, 1 tiM acetylcholine (ACh) or 10 n M GIP plus 1 /xM A C h as indicated. Data is expressed as the mean ( ± S E M ) somatostatin-like immunoreactivity (SLI) released as a percent of total islet SLI content. *p<0.05 compared to 17.8 m M glucose alone; § p < 0 . 0 5 compared to 17.8 m M glucose plus 10 n M GIP. n=5 for all conditions. 108 II. E F F E C T S O F GIP A N D O T H E R INSULIN S E C R E T A G O G U E S O N INSULIN S E C R E T I O N F R O M R A T P A N C R E A T I C /3-CELLS PURIFIED B Y F L U O R E S C E N C E - A C T I V A T E D C E L L - S O R T I N G A . Effect of Glucose on Insulin Secretion from FACS-Purified /3-Cells FACS-purified /3-cells incubated in 17.8 m M glucose secreted 2.07±0.32 % of their total IRI content in 1 h; this was not significantly greater than the IRI release observed in the presence of 4.4 m M glucose (1.94±0.35 %). B. Effect of GIP on Insulin Secretion from FACS-Purified /3-Cells In the presence of 17.8 m M glucose, GIP produced a concentration-dependent stimulation of IRI release (Figure 15). However, the stimulatory effect was not apparent until concentrations of at least 10-100 n M GIP were employed. Although the greatest IRI response to GIP (approximately 2X) was seen in the presence of the highest concentration of GIP tested (1000 nM), maximal stimulation of IRI secretion may not have been achieved by this concentration since the concentration-response curve appeared to still be increasing at this level of GIP. C. Effect of Acetylcholine, Cholecystokinin-8, Glucagon and Glucagon-Like Peptide-1 (7-36-NH2) on Insulin Secretion from FACS-Purified /3-Cells A l l secretagogues were studied in incubation medium containing 17.8 m M glucose. The effect of glucagon on IRI secretion from FACS-purified /3-cells was examined using porcine glucagon concentrations ranging from 0.1 to 100 n M (Figure 16A). Glucagon caused a modest stimulation of IRI release from pure /3-cells at 0.1 n M (1.64±0.17 % vs 1.24±0.18 %). Increasing the glucagon concentration of the medium produced further increases in IRI secretion. At 100 n M , IRI release was 109 approximately 2X that observed in the presence of 17.8 m M glucose alone (Figure 16A). GLP-1 (7-36-NH-2) was also a potent stimulus for IRI release from these cells (Figure 16B). The maximal IRI response to GLP-1 (7-36-NH2), about 2X basal IRI secretion, was observed at a concentration of 1 n M . Higher concentrations of the peptide did not further enhance IRI secretion; in fact, the IRI response to 10 n M GLP-1 (7-36-NH2) appeared to be slightly less than that produced by 1 n M GLP-1 (7-36-NH 2). The effect of A C h on pure /3-cell IRI secretion was studied at A C h concentrations of (0.1-10 /uM) (Figure 17A). Only 10 nM A C h generated a significant increase in IRI secretion compared to 17.8 m M glucose alone (3.03±0.34 % vs 2.15±0.39 %). CCK-8 also produced an increase in IRI secretion only at the highest concentration tested (100 nM) (Figure 17B); the octapeptide was ineffective at a concentration of 10 n M (2.39±0.57 % vs 2.04±0.37 %). D . Modulation of GIP-Stimulated Insulin Secretion from FACS-Purified B-Cells by Glucagon The effect of GIP on IRI secretion from FACS-purified /3-cells was examined in the presence of 17.8 m M glucose plus 10 n M glucagon (Figure 18A). The lowest effective concentration of GIP under these conditions was 10 n M , which produced an increase in IRI release of more than 2X (4.33±0.99 % vs 1.99±0.33 %). E . Effect of GIP on Insulin Secretion from Mixtures of FACS-Purified /3- and Non-/3-Cells Fractions of B- and non-/3-cells obtained from the F A C S were mixed in equal proportions prior to culture and subjected to stimulation by GIP in 17.8 m M glucose (Figure 18B). A significant increase in IRI release was observed at 1 n M GIP (2.30±0.39 % vs 1.60±0.22 %), which is 100X lower than the minimum 110 concentration of GIP required to stimulate IRI secretion from FACS-purified 8-cells in the absence of glucagon. GIP-stimulated IRI secretion from B- and non-B-cells was concentration-dependent, with 100 n M GIP causing an increase in IRI release of approximately 2X (3.25±0.48 % vs 1.60±0.27 %). I l l F I G U R E 15: Effect of GIP on Insulin Secretion from FACS-Purified Rat /3-Cells. FACS-purifed /3-cells (3000) were incubated for 1 h in GIP concentrations ranging from 0.1 to 1000 nM in the presence of 17.8 m M glucose. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total cell IRI content. *p<0.05 compared to 17.8 m M glucose alone. n=5 for all conditions. 1 1 2 A 4n F I G U R E 16: Effect of Glucagon and Glucagon-Like Peptide-1 (7-36-NH2) on Insulin Secretion from FACS-Purified Rat /3-Cells. FACS-purifed /3-cells (3000) were incubated for 1 h either in 17.8 m M glucose alone or plus (A) 0.1-100 nM glucagon (n=4) or (B) 0.1-100 n M glucagon-like peptide-1 (7-36-NH2) (n=5). Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total cell IRI content. *p<0.05 compared to 17.8 m M glucose alone. 113 C CD » C o o CD CO CO _a> CD rr B c <D c o o CD CO cd J D CD rr 3-5 4 3H _ 1-o-« .1 1 (ACh] (M-M) 10 '—it — • 0 10 100 [CCK] (nM) F I G U R E 17: Effect of Acetylcholine and Cholecystokinin-8 on Insulin Secretion from FACS-Purified Rat /3-Cells. FACS-purifed /3-cells (3000) were incubated for 1 h either in 17.8 m M glucose alone or plus (A) 0.1-10 nM acetylcholine (ACh) (n=6) or (B) 10 and 100 n M cholecystokinin-8 (CCK-8) (n=5). Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total cell IRI content. *p < 0.05 compared to 17.8 m M glucose alone. 114 A 7n 6-0 H <—// i — * — — i 0 .1 1 1 0 1 0 0 [GIP] (nM) F I G U R E 18: Effect of GIP on Insulin Secretion from FACS-Purified Rat 0-Cells in the Presence of Glucagon or FACS-Purified Non-0-Cells. (A)FACS-purifed 0-cells (3000) were incubated for 1 h in GIP concentrations ranging from 0.1 to 100 n M in the presence of 17.8 m M glucose and 10 n M glucagon; n=4 for all conditions. (B) 3000 FACS-purified 0-cells were cultured with 3000 FACS-purified non-j8-cells and incubated for 1 h in GUP concentrations ranging from 0.1 to 100 n M in the presence of 17.8 m M glucose; n=5 for all conditions. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total cell IRI content. *p<0.05 compared to 17.8 m M glucose alone. 115 III. E F F E C T S O F GIP O N INSULIN S E C R E T I O N F R O M BTC3 C E L L S A . Hormone Content of /3TC3 Cells R I A of cell extracts of 0TC3 cells revealed that the IRI content of these cells was 7 9 6 ± 5 8 M U / 1 0 5 cells (n=3). G L I was also measurable in extracts of /3TC3 cells (720 pg/10 6 cells; n=2). The molar ratio of I R L G L I in these cells was approximately 66:1. However, the concentration of G L I in the incubation medium of j8TC3 cells (105 cells in 0.5 ml) after 1 h was below the limit of detectability of the glucagon RIA used (approximately 50 pg/ml). B. Effect of GIP on Insulin Secretion in the Presence of 4.4 and 17.8 m M Glucose Basal (4.4 m M glucose) IRI secretion from BTC3 cells, expressed as a percent of total cell IRI content (10.0±0.85 %), was elevated 3-5X compared to normal rat islets and FACS-purified rat 0-cells. In 4.4 m M glucose, GIP (10"10-10"6 M) had no effect on IRI release from J0TC3 cells (Figure 19A). However, GIP did produce a weak, concentration-dependent stimulation of IRI release when the glucose concentration was 17.8 m M (Figure 19B). The maximum stimulation of IRI release was observed at 10"7 M GIP, which was also the lowest concentration of GIP to produce a statistically significant stimulation of IRI secretion. C. Effect of GIP on Insulin Secretion in the Presence of IB M X and Forskolin Addition of the phosphodiesterase inhibitor I B M X (1 mM) to incubation medium containing 17.8 m M glucose produced a significant (approximately 1.5X) enhancement of IRI secretion over 17.8 m M glucose alone (Figure 20A). GIP (10 nM) failed to further stimulate IRI release above IBMX-stimulated levels. 116 Similarly, the adenylate cyclase activator forskolin (1 /iM) enhanced IRI secretion in 17.8 m M glucose 1.5 fold and this effect was unaltered by the addition of 10 n M GIP (Figure 20B). D . Effect of GLP-1 (7-36-NH2) on Insulin Secretion In the presence of 17.8 m M glucose, 10 n M GLP-1 (7-36-NH2) caused a small increase in IRI release compared to glucose alone (12.58±1.31 % vs 9.93±1.14 117 c CO c o o CD t o CO CD CO rr E 15n 10-5-X X B c C o o CD t o 0J CD 15n la-s' X T X -10 -9 - 8 log[GIP] (M) -7 -6 F I G U R E 19: Effect of GIP on Insulin Secretion from 0TC3 Cells in 4.4 and 17.8 m M Glucose. 0TC3 cells (105) were incubated for 1 h in GIP concentrations ranging from 10"1U M to 10"° M in the presence of (A) 4.4 mM glucose (n=3) or (B) 17.8 m M glucose (n=5). Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total cell IRI content. *p<0.05 compared to 17.8 m M glucose alone. 118 A c CO c o o CD CO CO CD 20 15-10 o 5-I Glucose (mM) 17.8 IBMX (mM) GIP (nM) 17.8 1.0 17.8 1.0 10 B 20-15 c tg c o o °^ 10 CD t o a> CD cc 5-Glucose (mM) Forskolin ((iM) GIP (nM) 17.8 17.8 1.0 17.8 1.0 10 F I G U R E 20: Effect of GIP on Insulin Secretion from /3TC3 Cells in the Presence of I B M X and Forskolin. BTC3 cells (105) were incubated for 1 h in 17.8 m M glucose, plus either (A) 1 m M 3-isobutyl-l-methylxanthine (IBMX) or (B) 1 /xM forskolin in the absence or presence of 10 n M GIP as indicated. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total cell IRI content. *p<0.05 compared to 17.8 m M glucose alone. n=4 for all conditions. 119 IV. E F F E C T S O F GIP O N H O R M O N E S E C R E T I O N F R O M I S O L A T E D P A N C R E A T I C ISLETS O F L E A N A N D O B E S E Z U C K E R R A T S A . Hormone Content of Zucker Rat Islets Before and After Culture Islet content of insulin, glucagon and somatostatin was measured in freshly isolated and cultured islets from lean and obese Zucker rats (Table 3). Both before and after culture, the IRI content of obese Zucker rat islets was approximately 2X that of lean rats. During two day culture, IRI content of islets from lean and obese animals decreased by similar amounts (25-30 %). Conversely, the G L I and SLI content of lean rat islets was slightly higher than that of obese animals. The decrease in islet content of G L I and SLI in both lean and obese rat islets during culture (45-60 %) was somewhat greater than that observed for IRI. After culture, islet content of G L I and SLI remained higher in lean rat islets than in obese rat islets (Table 3). T A B L E HI: Hormone Content of Zucker Rat Islets Before and After Culture Lean Obese IRI (mU/islet) Before Culture After Culture 832±92 601144 b 1942±242 a 1327±283 a,b G L I (pg/islet) Before Culture After Culture 378±49 251±41 2 7 0 ± 7 0 130±33 a,b SLI (pg/islet) Before Culture After Culture 36.6±6.7 19.5±6.7 6 29.9±5.2 12.4±1.7 6 a: p < 0.05 vs. lean for same condition. b: p<0.05 vs. before culture for same phenotype. IRI: immunoreactive insulin; GLI : glucagon-like immunoreactivity; SLI: somatostatin-like immunoreactivity 120 B. Insulin Secretion from Lean and Obese Zucker Rat Islets 1. Effect of GIP on Insulin Secretion from Zucker Rat Islets During a 1 h incubation of 20 cultured lean or obese rat islets in medium containing 4.4 m M glucose, basal secretion of IRI per islet tended to be greater in obese than lean rats (10.09±1.95 vs 5.54±1.27 /iU/islet*h; p<0.08). This difference became more marked and statistically significant under conditions stimulatory for insulin secretion (0.2 or 2.0 n M GIP and/or 8.9 m M glucose) (Figure 21). In the presence of 8.9 m M glucose and 2.0 n M GIP, IRI secretion per islet was 3X greater from obese rat islets. To account for differences in islet IRI content between lean and obese Zucker rats, IRI secretion in response to glucose and GIP was expressed as a percent of total islet IRI content in each experimental well. Cultured islets from lean Zucker rats secreted 1.11±0.14 % of their islet IRI content under basal conditions (4.4 m M glucose). The addition of GIP (0.2 or 2.0 nM) to incubation medium containing 4.4 m M glucose had no effect on IRI secretion from lean rat islets (Figure 22A). Similarly, increasing the glucose concentration to 8.9 m M was without effect on IRI release. In the presence of 8.9 m M glucose, GIP (0.2 or 2.0 nM) appeared to produce a moderate elevation of IRI secretion although there was considerable variability in this response and the effect was not consistently seen. Cultured islets from obese Zucker rats secreted a significantly lower percent of the total islet IRI content than lean rat islets (0.54±0.11 % vs 1.11±0.14 %) under basal conditions (4.4 m M glucose). Like lean rats, increasing the glucose concentration to 8.9 m M did not affect IRI secretion. However, unlike the lean animals, islets from obese Zucker rats were consistently responsive to GIP even in the presence of 4.4 m M glucose (Figure 22B). In obese rat islets, GIP (2.0 but not 0.2 nM) produced a significant stimulation of IRI secretion compared to glucose alone at both 4.4 and 8.9 m M glucose. 121 2. Effect of Somatostatin on GIP-Stimulated Insulin Secretion from Zucker Rat Islets Somatostatin-14 (SS-14; 1.0 nM) suppressed 2.0 n M GIP-stimulated IRI secretion from obese Zucker rat islets in 8.9 m M glucose to levels similar to that seen with 8.9 m M glucose alone (Figure 23). In lean rat islets, IRI release was inhibited by SS-14 only to about 70 % of GIP-stimulated levels. As with the IRI response to GIP, there was substantial variation in IRI release from lean rat islets in the presence of GIP and SS-14 and therefore an inhibitory effect of SS-14 could not be statistically detected in lean islets in these experiments. The percent inhibition of GIP-stimulated IRI release by SS-14 was approximately 2X greater in islets from obese rats (58.6±7.3 % vs 31.2±12.6 %) although this difference was not statistically significant (0.05 <p<0.10). C. Glucagon Secretion from Lean and Obese Zucker Rat Islets 1. Effect of Glucose on Glucagon Secretion from Zucker Rat Islets Expressed as a percent of total islet G L I content, G L I secretion from lean vs. obese Zucker rat islets was not significantly different in the presence of 4.4 m M glucose (Figure 24). When the incubation medium glucose concentration was increased to 8.9 m M , no alteration in G L I release from lean rat islets was observed (Figure 24A), while G L I secretion from obese rat islets was suppressed to nearly 50 % of basal levels (2.12±0.36 % vs 1.21±0.17 %; Figure 24B). 2. Effect of GIP on Glucagon Secretion from Zucker Rat Islets Under all conditions tested, 0.2 n M GIP was without effect on G L I release (Figure 24). However, 2.0 n M GIP elicited modest (1.5-2.0X) increases in G L I secretion from both lean and obese rat islets in the presence of 4.4 m M glucose. In 8.9 m M glucose, GIP (2.0 nM) produced a significant stimulation of G L I secretion only from islets of obese Zucker rats (Figure 24B). Lean rat islets were unaffected by this concentration of GIP at the higher glucose level (Figure 24A). 122 D . Somatostatin Secretion from Lean and Obese Zucker Rat Islets 1. Effect of GIP and Glucose on Somatostatin Secretion from Zucker Rat Islets SLI release from lean and obese rat islets was measured in the above experiments in which significant alterations in IRI and G L I secretion were observed. However, no pattern in SLI release in response to glucose (4.4 and 8.9 mM) and GIP (0.2 or 2.0 nM) emerged (Figure 25). Although basal secretion of SLI appeared to be slightly elevated in islets from lean rats compared to the obese animals (3.46±0.73 % vs 2.32±0.87 %), statistical evaluation of the data did not reveal any significant differences in SLI release under any of the conditions tested. 123 50-T 40-1 1 30-3 20-CD W CO CD a io-0" Glucose (mM) GIP (nM) T i , T . * * * i * i T . * t * 1 1 4.4 4.4 4.4 8.9 8.9 8.9 0.2 2.0 - 0.2 2.0 F I G U R E 21: Insulin Secretion (per Islet) from Lean and Obese Zucker Isolated Rat Islets in the Presence of Glucose and GIP. Twenty islets from lean (•) or obese (•) Zucker rats were incubated for 1 h in 4.4 or 8.9 m M glucose either in the absence of GIP or with 0.2 or 2.0 n M GIP as indicated. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released (/uU/islet.h). *p<0.05 compared to lean rat islets for the same condition. n=6 for all conditions. 124 C CO •4—' c o o CD CO <d CD rr B c CO c o o CD CO cd CD rr 3.51 3.0 2.5 2.0 1.5 1.0 0.5 0.0 3.5 3.0 2.5 2.01 1.5 1.0 0.5 0.0 LEAN X OBESE Glucose (mM) 4.4 4.4 4.4 GIP(nM) - 0.2 2.0 8.9 T 8.9 0.2 8.9 2.0 F I G U R E 22: Effect of Glucose and GIP on Insulin Secretion from Isolated Islets of Lean and Obese Zucker Rats. Twenty islets from (A) lean or (B) obese Zucker rats were incubated for 1 h in 4.4 or 8.9 m M glucose either in the absence of GIP or with 0.2 or 2.0 n M GIP as indicated. Data was derived from experiments shown in Figure 20 and is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total islet IRI content.. *p<0.05 compared to 4.4 or 8.9 m M glucose alone. n=6 for all conditions. 125 3.0 n 2.5-Glucose (2.0 nM) + SS-14 (1 nM) F I G U R E 23: Effect of Somatostatin-14 on GIP-Stimulated Insulin Release from Isolated Islets of Lean and Obese Zucker Rats. Twenty islets from lean (•) or obese (0) Zucker rats were incubated for 1 h either in 8.9 m M glucose alone, plus 2.0 n M GIP or plus 2.0 n M GIP and 1.0 n M somatostatin-14 (SS-14) as indicated. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) released as a percent of total islet IRI content. *p<0.05 compared to 8.9 m M glucose alone. § p < 0 . 0 5 compared to 8.9 m M glucose plus 2.0 n M GIP. n = 6 for all conditions. 126 C 3 c o o CD CO cd _CD CD rr CD 4.0 T 3.5-3.0-2.5" 2.0" 1.5" 1.0-0.5-I o.o-X X LEAN X II B c CD C o <_> CD CO cd CD rr 4.0 3.5-3.0-2.5-2.0-1.5-1.0 0.5 0.0 T X I OBESE X Glucose (mM) 4.4 4.4 4.4 GIP{rtM) - 0.2 2.0 8.9 8.9 0.2 8.9 2.0 F I G U R E 24: Effect of Glucose and GIP on Glucagon Secretion from Isolated Islets of Lean and Obese Zucker Rats. Twenty islets from (A) lean (n = 5 for all conditions) or (B) obese (n = 6 for all conditions) Zucker rats were incubated for 1 h in 4.4 or 8.9 m M glucose either in the absence of GIP or with 0.2 or 2.0 n M GIP as indicated. Data is expressed as the mean (+SEM) glucagon-like immunoreactivity (GLI) released as a percent of total islet G L I content. . *p<0.05 compared to 4.4 or 8.9 m M glucose alone. 127 8n c co c o u CD w CO © CD rr CO 4-2- 1 1 Glucose (mM) 4 - 4 GIP (nM) -I I I T 1 1 I 4.4 4.4 8.9 8.9 8.9 0.2 2.0 - 0.2 2.0 F I G U R E 25: Effect of Glucose and GIP on Somatostatin Secretion from Isolated Islets of Lean and Obese Zucker Rats. Twenty islets from lean (•) or obese (•) Zucker rats were incubated for 1 h in 4.4 or 8.9 m M glucose either in the absence of GIP or with 0.2 or 2.0 n M GIP as indicated. Data is expressed as the mean ( ± S E M ) somatostatin-like immunoreactivity (SLI) released as a percent of total islet SLI content. n=5 for all conditions. 128 V . E F F E C T S O F G A L A N I N O N INSULIN A N D S O M A T O S T A T I N S E C R E T I O N S T I M U L A T E D B Y GIP A N D O T H E R INSULIN S E C R E T A G O G U E S A . Effects of Galanin on Insulin Secretion 1. Effect of Porcine Galanin on GIP-Stimulated Insulin Secretion from the Perfused Rat Pancreas In the presence of a perfusate glucose concentration of 8.9 m M , which is above the threshold for the insulinotropic action of GIP in this preparation (Pederson and Brown, 1976), 0.2 n M GIP produced a prompt stimulation of IRI secretion from the perfused pancreas to > 10X basal secretion levels. Infusion of 50 n M porcine galanin 10 min following the introduction of GIP rapidly suppressed GIP-stimulated IRI release to approximately 50 % of stimulated levels (Figure 26). Upon cessation of galanin infusion, IRI secretion returned to pre-infusion levels. 2. Effect of Porcine and Rat Galanin on Insulin Secretion from the Perfused Rat Pancreas Stimulated by Acetylcholine, Cholecystokinin-8 and Arginine A C h (5 M M ) also caused a marked increase in IRI secretion over basal levels in the presence of 4.4 m M glucose. Porcine galanin, when infused for 10 min at a concentration (50 nM) which markedly inhibited GIP-stimulated IRI secretion, had no effect on IRI release during stimulation with A C h (Figure 27). Rat galanin was also without effect under these conditions. The effects of rat and porcine galanin were also examined on IRI secretion during stimulation of the perfused rat pancreas by C C K and arginine. The results, expressed as the percent change in IRI secretion during galanin infusion from pre-infusion levels, are summarized in Figure 28. Like A C h , CCK-stimulated IRI secretion (in the presence of 8.9 m M glucose) was unaffected by either porcine or 129 rat galanin. However, arginine-stimulated IRI release was significantly suppressed during infusion of the neuropeptide. The inhibitory effects of rat and porcine galanin on arginine-stimulated IRI release were of approximately the same magnitude at the concentration of neuropeptide used (50 nM). 3. Comparison of the Inhibitory Effect of Rat and Porcine Galanin on GIP-Stimulated Insulin Secretion from Rat /3-Cells The inhibitory potencies of rat and porcine galanin on GIP-stimulated IRI release in vitro were compared over a range of galanin concentrations. Cultured mixtures of FACS-purified B- and non-/3-cells were used for this purpose since this preparation was shown to be more sensitive to the insulinotropic action of GIP than FACS-purified /3-cells alone. In the presence of 17.8 m M glucose, 10 n M GIP produced a 1.7 fold increase in IRI secretion from these cells (4 .1±0.4 vs. 2 .4±0 .3%; p<0.05 compared to 17.8 m M glucose alone). Both rat and porcine galanin inhibited the response to GIP in a concentration-dependent manner (Figure 29). The inhibitory effects of rat and porcine galanin on GIP-stimulated IRI release appeared also to be similar in magnitude, becoming statistically significant at a concentration of 1 n M for both the rodent and the porcine versions of the peptide. B. Effects of Galanin on Somatostatin Secretion 1. Effect of Porcine Galanin on Somatostatin Secretion from the Perfused Rat Pancreas in the Presence of GIP SLI secretion was not significantly altered during stimulation of the perfused pancreas with 8.9 m M glucose plus 0.2 n M GIP, although a slight decrease in pancreatic SLI release was apparent during exposure to these insulin secretagogues. Infusion of porcine galanin, at a concentration (50 nM) known to inhibit gastric SLI release from the perfused rat stomach (Kwok et al, 1988), had no effect on pancreatic SLI secretion under these conditions (Figure 30). 130 FIGURE 26: Effect of Porcine Galanin on GIP-Stimulated Insulin Secretion from the Perfused Rat Pancreas. The perfusate glucose concentration was 4.4 mM during periods 1 and 11-12 and 8.9 mM during periods 2-10. GIP (0.2 nM) and either 50 nM porcine galanin (Gal; n=7) or a control vehicle ( o ; n=6) were infused as indicated. Data is expressed as the mean (±SEM) immunoreactive insulin (IRI) release (MU/min) per 5 min time period. Percent change in IRI release during galanin infusion was significantly lower than control (p < 0.05). 131 rr 300 n c <zZ 200-CO cd CD rr 100-4 6 8 5 min Periods 1 o F I G U R E 27: Effect of Porcine and Rat Galanin on Acetylcholine-Stimulated Insulin Secretion from the Perfused Rat Pancreas. The perfusate glucose concentration was 4.4 m M throughout. Acetylcholine (ACh; 5 fj.M) and either 50 n M porcine galanin (-OS n=6), 50 n M rat galanin (-•-; n=7) or a control vehicle (-•-; n=6) were infused as indicated. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) release (/iU/min) per 5 min time period. 132 13 .2 a> DC 120-| -5- IOOH o> c cd .e o 0\ 80-60' 40-20-ACh X -20J B a> at 601 40-1 "• a> ID CC -20-T T T CCK <p o> c cd x: u 80-60-40-20-© 0-CC nr -20H Arginine -40-Control Porcine Rat Galanin Galanin FIGURE 28: Summary of the Effects of Porcine and Rat Galanin on Acetylcholine-, Cholecystokinin-8-, and Arginine-Stimulated Insulin Secretion from the Perfused Pancreas. Rat or porcine galanin (10 nM) or a control vehicle was infused for 10 min during the second phase of IRI secretion stimulated by either (A) 5 M M ACh, (B) 0.9 nM CCK-8 (plus 8.9 mM glucose) or (C) 20 mM arginine. Data is expressed as mean (±SEM) percent change of IRI secretion during 10 min infusion of rat or porcine galanin or control vehicle. n=6 for all experiments except n=5 for rat galanin infusion during CCK-8 or Arg stimulation and n=7 for rat galanin infusion during ACh stimulation. *p<0.05 compared to control. 133 F I G U R E 29: Comparison of the Inhibitory Effect of Porcine and Rat Galanin on GIP-Stimulated Insulin Secretion from Mixtures of FACS-Purified Rat (3- and non-j8-Cells. FACS-purifed 0-cells (3000) were cultured with 3000 FACS-purified non-0-cells and incubated for 1 h in the presence of 17.8 m M glucose plus 10 n M GIP and either rat (-o) or porcine (-•-) galanin in concentrations ranging from 10"10 to 10"7 M ; n=4 for all conditions. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (LRI) released as a percent of total cell IRI content. *p<0.05 compared to 17.8 m M plus 10 nM GIP alone glucose alone. 134 FIGURE 30: Effect of Porcine Galanin on Somatostatin Secretion from the Perfused Rat Pancreas in the Presence of GIP. The perfusate glucose concentration was 4.4 mM during periods 1 and 11-12 and 8.9 mM during periods 2-10. GIP (0.2 nM) and either 50 nM porcine galanin (Gal; -•-; n= 7) or a control vehicle (-o; n=6) were infused as indicated. Data is expressed as the mean (±SEM) somatostatin-like immunoreactivity (SLI) release (pg/min) per 5 min time period. 135 VI. I N V E S T I G A T I O N S INTO T H E E X I S T E N C E O F A GIP R E C E P T O R O N P A N C R E A T I C 0 -CELLS A . Development of GIP Receptor Probes 1. Iodinated ( 1 2 5I)-GIP a). H P L C Purification of 1 2 5 I -GIP H P L C analysis of the products of chloramine T iodination of porcine GIP 1-42 ( E G m), on an acetonitrile gradient of 32-38 % over 10 min (in 0.1 % T F A ) , revealed a heterogeneous population of iodinated peptides consisting of at least 6 peaks (numbered 1-6, Figure 31). Free N a 1 2 5 I always eluted immediately following injection of the iodination mixture. Peaks 1-4 were consistently discernible and were characterized by the larger magnitude of peaks 2 and 4. The later eluting peaks (5 and 6) were not always observed. Non-iodinated porcine GIP, when combined in excess with a trace amount of the iodinated mixture and measured by R I A and ELISA, eluted prior to the radioactive peaks (Figure 31). The elution time of peak 2 on this gradient was 19.5 min from the time of injection, while porcine GIP 1-42 and 3-42 eluted at 13.5 min and 11 min, respectively. This gradient was used for the purification of 1 2 5 I -GIP for use in all binding and biological activity studies. Since baseline resolution of these peaks was not achieved using this gradient, the purity of collected fractions corresponding to peak 2 was tested by reapplying the collected peak to the H P L C system. On an acetonitrile gradient of 32-38 % over 10 min, peak 2 eluted as a single narrow peak with little free N a 1 2 5 I (Figure 32A), suggesting that the collected fraction was of high purity despite the only partial resolution of this peak during the original purification. This was confirmed by application of the peak to a narrower acetonitrile gradient (33-35 % over 10 min; Figure 32B). A slightly broader, but still monocomponent peak was observed. 136 When stored lyophilized in 2 % aprotinin and 0.1 % BSA, this degree of purity was consistently observed 5-6 weeks following the original purification of peak 2, suggesting that no significant degradation of the purified fraction occurred. Improved resolution of the different peaks seen in the chromatogram in Figure 31 was attempted by isocratic elution of 1 2 5 I -GIP in 32 % acetonitrile for 20 min, followed by a more gradual increase in the acetonitrile concentration (32-38 % over 40 min). This protocol also yielded 6 peaks (numbered 1-6, Figure 33). Although peaks 1 and 3 were broad and less clearly defined, peaks 2 and 4-6 were plainly distinguishable. Each peak was separated by a return to baseline 1 2 5 I radioactivity. Because the GIP preparation used in the iodination procedure (porcine GIP 1-42, fraction E G III; Brown et al, 1969; 1970) contains small amounts of GIP 3-42 and C C K , iodination of these contaminating peptides may contribute to the heterogeneity of the H P L C elution profile of 1 2 5 I -GIP 1-42. Therefore, synthetic porcine GIP 1-30 was iodinated by an identical procedure as GIP 1-42 and the iodination mixture was eluted on the acetonitrile gradient used to produce the chromatogram in Figure 31. Elution of 1 2 5 I -GIP 1-30 by this protocol yielded at least 4 peaks (Figure 34). The peaks numbered 1-4 resembled peaks 1-4 of 1 2 5 I -GIP 1-42, in that peaks 2 and 4 were greater in magnitude. However, peak 2 of 1 2 5 I -GIP 1-30 eluted later than peak 2 of 1 2 5 I -GIP 1-42, (21.5 vs. 19 min). Further, unlike 1 2 5 I -GIP 1-42, peak 1 of 1 2 5 I -GIP 1-30 was partially resolved into 2 individual components. Some 1 2 5 I radioactivity eluted after peak 4, although individual peaks could not be distinguished. b). Iodination State of Different Peaks of HPLC-Purified 1 2 5 I -GIP In all subsequent discussion, HPLC-purified 1 2 5 I -GIP refers to iodinated porcine GIP 1-42 ( E G III) purified according to the protocol and elution profile shown in Figure 31. To analyse the iodination state of the tyrosine residues in the 137 different peaks obtained from H P L C purification, each peak was digested to single amino acid residues with pronase and was subjected to gel filtration on Sephadex R G-10. The elution time of the 125I-radioactivity was then compared to the elution times of mono- and diiodinated (MIT and DIT) standards. Figure 35 illustrates a representative profile obtained by gel filtration of peak 2 following pronase digestion. The fraction contained no detectable radioactivity coeluting with free N a 1 2 5 I or DIT, while a large single peak of radioactivity eluted simultaneously with the M I T standard. The % of recovered radioactivity (excluding free Na 1 2 5 I ) due to M I T and DTT was calculated for peaks 1-4 and the data summarized in Table IV. Peaks 1-4 all contained mostly MIT radioactivity. Of these, peak 2 was most pure, consisting of almost 100 % MIT. The later eluting peaks (3 and 4) had increasing amounts of DIT radioactivity (9.6 and 15.5 %, respectively). Poor resolution and yield of peaks 5 and 6 precluded them from this analysis. However, as shown in Table IV, the high specific activity of peaks 5 and 6 indicates a high degree of incorporation of 1 2 5 I into the GIP molecule relative to peaks 1-4. c). Binding of Different Peaks of HPLC-Purified 1 2 5 I -GIP to 0TC3 Cells Peaks 1-4 were tested for their ability to bind to BTC3 cells using a binding assay protocol adapted from Amiranoff et al, 1984. Although all 4 peaks showed binding displaceable by 10"6 M GIP (Figure 36), peak 2 demonstrated the greatest specific (total - non-specific) binding. Binding by peaks 1 and 4 tended to be slightly less than peak 2, and peak 3 exhibited the lowest specific binding to /3TC3 cells. Peak 2 was therefore chosen as the radioligand for further binding studies due to its relatively high yield and purity from H P L C , high degree of specific binding to /3TC3 cells, and the high ratio of mono- to diiodinated tyrosine residues. 138 d). Biological Activity of HPLC-Purified I-GIP (i) H P L C Purification of 1 2 7 / 1 2 5 I - G I P Using the perfused pancreas, the biological activity of peak 2 of I-GIP obtained from H P L C purification was tested using 1 2 7 I -GIP plus a trace amount of 1 2 5 I -GIP to avoid perfusing the organ with high levels of radioactivity. 1 2 7 / 1 2 5 I - G I P was purified by H P L C using a protocol similar to that used for 1 2 5 I -GIP in Figure 31. A separate H P L C system (Waters) was used to enable the measurement of absorbance and the different 1 2 7 / 1 2 5 I - G I P peaks were identified by measurement of 1 2 5 I radioactivity in collected fractions. Since GIP 1-42 eluted in the void volume when applied to this H P L C system in 32 % acetonitrile, a gradient of 30-36 % acetonitrile was used. This gradient produced retention times for GIP and I-GIP similar to that obtained when a gradient of 32-38 % was employed in the H P L C purification of 1 2 5 I -GIP using the Beckman H P L C (Figure 31). The elution profiles of 1 2 7 / 1 2 5 I - G I P as obtained by absorbance and by radioactivity measurement were similar with the major peaks nearly superimposable (Figure 37). Although only 3 major peaks could be distinguished by measurement of radioactivity, the first of these peaks appeared to have more than one component when the absorbance profile was examined. Because this major peak had the highest level of radioactivity and its retention time (20.5 min) was similar to that obtained for peak 2 in Figure 31 (19.5 min), this peak was identified as "peak 2". However, "peak 1" may also be a constituent of this fraction since absorbance measurement detected at least one other component in this peak. Non-iodinated GIP, measured by R I A and ELISA, eluted prior to the iodinated peaks. (ii) Effect of Peak 2 of HPLC-Purified 1 2 7 / 1 2 5 I - G I P on Insulin Secretion from the Perfused Rat Pancreas The fractions corresponding to "peak 2" from the elution profile in Figure 37 were pooled, lyophilized, reconstituted in perfusate and presented to the perfused 139 rat pancreas as a gradient from 0-1.0 n M in the presence of 17.8 m M glucose (Figure 38). The stimulatory effect of GIP 1-42 ( E G III) on IRI secretion was clearly seen only 7 min after the commencement of the GIP gradient, corresponding to a GIP concentration in the perfusate of approximately 150 p M . 1 2 7 / 1 2 5 I - G I P also produced a marked stimulation of IRI secretion although it did not become apparent until a concentration of approximately 375 p M was achieved. Both GIP (226±25 mU/45 min) and 1 2 7 / 1 2 5 I - G I P (136±34 mU/45 min) significantly enhanced the total IRI output over the 45 min perfusion period compared to glucose alone (41±11 mU/45 min), although the iodinated peptide was only about 60 % as insulinotropic as the native hormone. d) . Displacement of HPLC-Purified 1 2 5 I -GIP from ;0TC3 Cells by GIP Increasing concentrations of natural GIP decreased the binding of 1 2 5 I -GIP (HPLC-purified peak 2) to /3TC3 cells in a concentration-dependent manner (Figure 39). Specific and non-specific binding was 0.67±0.11 and 0.28±0.05 % per 100 Mg cell protein, respectively. Displacement by GIP was significant at GIP concentrations as low as 500 p M and appeared to be maximal at 100 n M . Numerous other peptides were tested for their ability to displace 1 2 5 I -GIP, including synthetic GIP 1-30-NH2, rat G R F , porcine secretin, porcine VIP, human GLP-1 (7-36-NH 2), porcine glucagon, rat galanin, somatostatin-14 and CCK-8. At a concentration of 1 / i M , GIP 1-30 produced a displacement of 1 2 5 I -GIP comparable to GIP 1-42. Glucagon (1 /xM) produced a slight (approximately 20 %) inhibition of 1 2 5 I -GIP binding but was without effect at 100 n M . No other peptide tested significantly altered 1 2 5 I -GIP binding to /3TC3 cells. e) . Displacement of HPLC-Purified 1 2 5 I -GIP from Isolated Islets by GIP The binding assay used to detect GIP binding sites on /8TC3 cells was used to investigate the presence of GIP receptors on isolated rat islets after short-term culture. Specific binding of 1 2 5 I -GIP to islets was 0.77±0.15 % per 100 Mg islet 140 protein, while non-specific binding was 8.04±1.53 % per 100 ng islet protein. Increasing concentrations of GIP 1-42, incubated with 1 2 5 I -GIP (peak 2; approximately 50 pM) in tubes containing 100 rat islets, displaced bound 1 2 5 I -GIP in a concentration-dependent fashion (Figure 40). The lowest concentration of GIP to significantly inhibit 1 2 5 I -GIP binding was 1 n M . Maximal inhibition of binding appeared to occur at 1 nM GIP. GIP 1-30 (1 /xM) displaced 1 2 5 I -GIP to a similar extent as GIP 1-42 (approximately 60 % of control). No other peptide, including glucagon, had any effect on 1 2 5 I -GIP binding at concentrations up to 1 / i M . B. Biotinylated GIP 1-30 1. H P L C Characterization of B-GIP It was anticipated that biotinylation of GIP l -30-NH 2 by the technique described in Methods would produce a heterogeneous mixture of multi-biotinylated forms of B-GIP of varying biological activities; therefore, attempts were made to characterize the product(s) of GIP 1-30 biotinylation by reverse-phase H P L C . B-GIP was applied to a gradient of 30-36 % acetonitrile (10 min), on which GIP 1-30 had a retention time of approximately 19 min. However, no peaks were oberved when up to 25 jug of B-GIP was applied to this gradient (Figure 41A). This indicated that no detectable amount of unaltered GIP 1-30 remained in the biotinylated product. To determine if B-GIP forms would elute at higher concentrations of acetonitrile, the mixture was applied to an acetonitrile gradient of 40-80 % (10 min). At least two peaks were discernible (Figure 41B). As measured by R I A and ELISA, both of these peaks contained immunoreactive GIP, while fractions collected outside of the peak areas had no measurable IR-GIP. This confirmed that altered, presumably biotinylated forms of GIP 1-30 were present in the product, although the nature of these components could not be identified. 141 2. Biological Activity of B-GIP Since no GIP 1-30 was detectable in the B-GIP product and H P L C purification of different forms of B-GIP resulted in low yield, unpurified B-GIP was used to compare the biological activity of this potential GIP receptor probe to GIP 1-30. The perfused rat pancreas responded to 17.8 m M glucose plus a linear gradient of GIP 1-30 (0-1.0 nM) with a strong increase in IRI secretion that became apparent at a GIP 1-30 concentration of approximately 225 p M . The response to a gradient of 0-1.0 n M B-GIP was similar (Figure 42). Total IRI output for the 45 min perfusion period was significantly greater during stimulation by either GIP 1-30 (189±24 mU/45 min) or B-GIP (154±29 mU/45 min) when compared to the response to 17.8 m M glucose alone (41±11 mU/45 min). 142 FIGURE 31: HPLC Elution Profile of ^I-GIP 1-42 in a Gradient of 32-38 % Acetonitrile. ^I-GIP 1-42 (400 /il; approximately 107 cpm) was eluted on an acetonitrile gradient of 32-38 % over 10 min (as indicated) on a /xBondapak C18 column (Waters). The early (0-5 min) eluting peaks are free N a 1 2 5 ! and are due to multiple injections of small volumes (100 ix\) of 125I-GIP 1-42. The elution times of noniodinated GIP 3-42 and 1-42 are indicated by the arrows. Major peaks are numbered 1-6 for reference in the text. 143 80' r25 70' -20 CD C o o < 60' 50' 40' 30' ' -15 >» .2 o 10 •§ £S CO •5 ~ 20" B — i — 10 — i 1 1 1— 15 20 25 30 35 Time (min) O CO-CO u o CO cr Time (min) FIGURE 32: HPLC Purity Analysis of Collected Fractions of Peak 2 of HPLC-Purified ^I-GIP. Fractions corresponding to peak 2 in the HPLC purification of ^ I-GIP (Figure 31) were collected, reapplied to the HPLC (approximately 105 cpm in 100 /ul) and eluted on acetonitrile gradients of (A) 32-38 %, 10 min and (B) 33-35 %, 10 min (as indicated) on a ^ Bondapak C18 column (Waters). 144 FIGURE 33: HPLC Elution Profile of ^I-GIP 1-42: Isocratic Elution in 32 % Acetonitrile. ^I-GIP 1-42 (100 /d; approximately 5 X 105 cpm) was eluted isocratically for 20 min in 32 % acetonitrile followed by an acetonitrile gradient of 32-38 % over 40 min (as indicated) on a /xBondapak C18 column (Waters). The early (5 min) eluting peak is free Na 1 2 5 ! . Major peaks are numbered 1-6 for reference in the text. 145 FIGURE 34: HPLC Elution Profile of ^I-GIP 1-30 in a Gradient of 32-38 % Acetonitrile. ^I-GIP 1-30 (Synthetic) (200 approximately 106 cpm) was eluted on an acetonitrile gradient of 32-38 % over 10 min (as indicated) on a /xBondapak C18 column (Waters). The two early (0-5 min) eluting peaks are free N a 1 2 5 ! and are due to two injections (100 /xl each) of ^I-GIP 1-30. Major peaks are numbered 1-4 for reference in the text. 146 F I G U R E 35: Analysis of Iodination State of HPLC-Purified 1 2 5 I - G I P (Peak 2). Peak 2 obtained from HPLC-purification of ^ I - G I P (Figure 31) was subjected to pronase digestion and filtration on Sephadex R G-10. 1 2 5I-Radioactivity (cpm) was determined in 100 /xl aliquots from 20 min fractions (flow rate: 200 /xl/min). The elution times of free N a 1 2 5 I and mono- and diiodinated tyrosine (MIT and DIT, respectively) are indicated. T A B L E IV: Iodination State of Different Peaks of HPLC-Purified I-GIP % M I T % D I T S.A. Iodine Atoms/ (mCi/mg) Molecule GIP Unpurified N / D N / D 82.7 0.21 Peak 1 99.6 0.4 293.2 0.74 Peak 2 99.9 0.1 296.7 0.75 Peak 3 90.4 9.6 74.8 0.19 Peak 4 84.5 15.5 33.1 0.083 Peak 5 N / D N / D 512.4 1.29 Peak 6 N / D N / D 439.4 1.11 % M I T and % DIT: Percent of total incorporated radioactivity due to mono- or diiodinated tyrosine residues, respectively. S.A.: Specific activity as determined from self-displacement curves. N / D : Not determined. 148 3 o C L CD O C O =1. E C L O C D c m o CD C L 00 4n 3 -T 2-1 -peak 1 peak 2 peak 3 Peak Number peak 4 F I G U R E 36: Specific Binding of Different Peaks of HPLC-Purified 1 2 5 I -GIP to /3TC3 Cells. Peaks 1-4 were collected from H P L C purification of 1 2 5 I -GIP as shown in Figure 31. Specific binding was calculated as the total binding in the presence of 5 X 104 cpm (approximately 50 p M ^ I - G I P ) minus the non-specific binding (in the presence of 1 /xM GIP 1-42). n=3 for each peak. 149 FIGURE 37: HPLC Elution Profile of W^-GJTP 1-42 in a Gradient of 30-36 % Acetonitrile. ^ / ^ I - G I P 1-42 (100 pi; approximately 100 /xg or 105 cpm) was eluted on a /iBondapak C18 column (Waters) on an acetonitrile gradient as follows: 30 % acetonitrile, 0-17 min; 30-36 % acetonitrile, 17-27 min; 36 % acetonitrile 27-30 min; 36-60 %, 30-35 min. The elution times of noniodinated GIP 3-42, GIP 1-42 and chloramine T (CT) are indicated. Absorbance is expressed in arbitrary units (volts) as determined by the Waters HPLC computer program (Maxima R 820). 150 F I G U R E 38: Effect of HPLC-Purified U1/125I-GIP (Peak 2) on Insulin Secretion from the Perfused Rat Pancreas. Perfusate glucose concentration was 17.8 m M throughout. HPLC-Purified 1 2 7 / 1 2 5 I - G I P (-•-; n=5) or GIP 1-42 (o- ; n=4) was introduced as a linear gradient (represented by the straight line) of 0-1.0 nM. Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) release (liU/min). The total IRI output during gradient infusion of GIP 1-42 or ^ / ^ I - G I P was significantly greater than the IRI output during perfusion with 17.8 m M glucose alone (-o; n=4). 151 [GIP] (M) F I G U R E 39: Displacement of HPLC-Purified 1 2 5 I-GIP from /3TC3 Cells by GIP. BTC3 cel lsXSXIO 5 cells/well) were incubated with 5 X 10 cpm (approximately 50 pM) H P L C -purified 1 2 5 I - G I P (peak 2) plus increasing concentrations of GIP 1-42 in 0.5 ml assay buffer for 1 h. ^ I - G I P binding is expressed as a percent of control (in the absence of GIP). 152 F I G U R E 40: Displacement of HPLC-Purified ^ I - G I P from Cultured Rat Islets by GIP. Isolated rat islets (100 islets/well) after 2-day culture were incubated with 2 X 104 cpm (approximately 50 pM) HPLC-Purified 1 2 5 I -GIP (peak 2) plus increasing concentrations of GIP 1-42 in 200 /xl assay buffer for 1 h. *p<0.05 compared to control (in absence of GIP). 153 FIGURE 41: HPLC Analysis of Biotinylated GIP 1-30. Biotinylated GIP 1-30 (B-GIP) was eluted on an acetonitrile gradient of (A) 30-36 % over 10 min (as indicated) or (B) 40-80 % over 10 min (as indicated) on a MBondapak C18 column (Waters). Mass of B-GLP injected was approximately (A) 25 Mg and (B) 10 /xg. The elution time of GIP 1-30 is indicated in each figure. 154 12000n Time (min) F I G U R E 42: Effect of Biotinylated GIP 1-30 on Insulin Secretion from the Perfused Rat Pancreas. Perfusate glucose concentration was 17.8 m M throughout. Biotinylated GIP 1-30 (B-GIP; - • - ; n=5) or GIP 1-30 ( -O - ; n=5) was introduced as a linear gradient (represented by the straight line) of 0-1.0 n M . Data is expressed as the mean ( ± S E M ) immunoreactive insulin (IRI) release (/iU/min). The total IRI output during gradient infusion of GIP 1-30 or B-GIP was significantly greater than the IRI output during perfusion with 17.8 m M glucose alone (-o; n=4). 155 DISCUSSION The studies presented in this thesis were undertaken to assess the influence of GIP on islet hormone secretion using in vitro approaches. The insulinotropic action of GIP has been documented in numerous studies previous to this work and the hormone has gained wide acceptance as an important component of the enteroinsular axis (Brown, 1988; Creutzfeldt and Ebert, 1988). Further, GIP has been implicated in certain disorders of insulin secretion and carbohydrate metabolism (Creutzfeldt and Ebert, 1988; Krarup, 1988; Brown et al, 1989). Yet with few exceptions, past investigations have utilized in vivo or in situ approaches to the study of GIP and islet hormone secretion due to the poor responsiveness of in vitro preparations to GIP. While these studies provided clear proof that GIP was a physiologic incretin, they yielded little information regarding the interaction of GIP at the level of the islet cells, its mechanism of action, the existence of a GIP receptor and the effects of GIP on islet glucagon and somatostatin secretion. A better understanding of these aspects of the action of GIP, which could most effectively be gained through in vitro studies, is critical to understanding the possible role of GIP in pathophysiological states. The impetus for the present investigations arose from the demonstration that following culture, isolated islets regained their sensitivity to GIP (Siegel and Creutzfeldt, 1985). In this thesis, cultured preparations of isolated islets and FACS-purified 0-cells, the 0TC3 cell line as well as the perfused pancreas were all used to investigate the role of GIP in the regulation of islet hormone secretion. The glucose-dependency of the insulinotropic action of GIP was clearly demonstrated in the experiments with isolated, cultured, Wistar rat islets. When the glucose concentration of the incubation medium was below or near fasting glucose levels in the rat (2.75 or 4.4 mM, respectively; Tasaka et al, 1975), physiological 156 concentrations of GIP were without effect on insulin secretion. However, a glucose concentration which approximated post-prandial plasma levels in the rat (8.9 m M ; Tasaka et al, 1975) was sufficient to potentiate the stimulatory action of GIP on the /3-cell. The maximal potentiation of insulin secretion by GIP was seen at the highest glucose concentration tested (17.8 mM). This glucose-dependency of GIP has been observed consistently in vivo and in the perfused pancreas preparation and has been described as a protective mechanism against the stimulation of insulin secretion during hypoglycemia, as well as a mechanism for increasing the release of insulin during post-prandial increases in blood glucose (Pederson and Brown, 1976). The present demonstration that this mechanism is intact in vitro in isolated islets suggests that it involves an interaction of glucose and GIP at the /3-cell. This idea was further supported by the demonstration of glucose-dependency in /3TC3 cells, although the effect in this cell line was somewhat less pronounced and was only tested at two glucose concentrations (4.4 and 17.8 mM). The possible involvement of other islet hormones in this mechanism was not supported by these data, since glucagon, which stimulates insulin secretion, was lower in 17.8 m M glucose plus GIP than in 2.75 m M glucose plus GIP. Further, somatostatin, which inhibits insulin secretion, was released in higher concentrations in the presence of 17.8 m M glucose plus GIP. The exact mechanism of interaction remains undefined but has been suggested to involve synergism between /3-cell intracellular messengers produced by glucose metabolism and GIP receptor binding (Dahl, 1983; Zawalich and Rasmussen, 1990). However, the possibility that glucose influences GIP binding to its receptor on the /3-cell cannot be ruled out since only one glucose concentration was used in the GIP receptor binding experiments. The binding assay described here should now allow this possibility to be tested. Glucose-dependency has also been shown for other insulin secretagogues, including C C K (Verspohl and Ammon, 1987), A C h (Garcia et al, 1988) and GLP-I (7-36-NH2) (Weir et al, 1989). Since most insulinotropic 157 agents appear to require the presence of glucose, a shared mechanism seems likely, possibly at a common site in the intracellular signal transduction pathways of these secretagogues. The concentration of GIP (1 nM) that was used to produce a stimulation of insulin secretion from cultured islets in the presence of 8.9 m M glucose, was close to the post-prandial levels measured in rat plasma by R IA (Chan et al, 1984). In 17.8 m M glucose, islets appeared to be sensitive to GIP concentrations as low as 0.1 n M , with a clear 2.5 fold stimulation seen at 1 n M . These results confirmed the experiments of Siegel and Creutzfeldt (1985), who suggested that cultured islets are sensitive to concentrations of GIP in the physiological range. In addition, this demonstration substantiates the claim that GIP is a physiologic incretin and further, that the insulinotropic effect of GIP is mediated by direct interaction with the islet. Certain insulin secretagogues (e.g. A C h , C C K ) have been shown to produce an insulin response immediately after islet isolation by collagenase digestion (Verspohl and Ammon, 1987; Garcia et al, 1988), while other insulinotropic agents (e.g. GIP, glucagon) were found to stimulate insulin secretion more effectively after short periods of culture (Rabinovitch et al, 1978). It has been suggested that certain receptors, such as the GIP receptor, are more susceptible to damage by collagenase. Presumably, receptors destroyed during enzymatic and mechanical digestion of the pancreas are replaced during culture. Islet responsiveness to GIP has been shown to be markedly increased after only 4 h culture following collagenase isolation (Siegel and Creutzfeldt, 1988). This may explain why previous attempts to demonstrate the presence of GIP receptors were successful in 0-cell lines in continuous culture but not collagenase-isolated islets, while in the present study, displaceable binding of 1 2 5 I -GIP to cultured islets was shown. It should also be considered that signal transduction may be altered in cultured islets compared to freshly isolated islets. Glucose-stimulated insulin release is much weaker from 158 cultured islets, as was observed in these experiments, in which no significant insulin response was seen in the absence of GIP until 17.8 m M glucose was used as the stimulus. It has been suggested that the diminished insulin response to glucose in cultured islets was related to decreased levels of islet pyridine nucleotides and diminished calcium uptake (Siegel et al, 1983; Verspohl et al, 1988). In this regard, it is interesting that freshly isolated islets seem to be more sensitive to secretagogues which are thought to mediate their insulinotropic action via the stimulation of phosphoinositol metabolism ( C C K and ACh), whereas only cultured islets are sensitive to secretagogues which may act via the stimulation of adenylate cyclase (GIP and glucagon). The stimulatory effect of GIP on insulin secretion appeared to be concentration-dependent in these experiments at both low and high glucose concentrations. In 17.8 m M glucose, which had the maximum potentiating effect on GIP-stimulated insulin secretion both in this preparation and in the perfused rat pancreas (Pederson and Brown, 1976), the islet was most sensitive to changes in the concentration of GIP between 0.1 and 1 n M . Insulin secretion was twofold greater in the presence of 1 n M GIP compared to 0.1 n M GIP. This sensitivity to changes in GIP levels within the physiological range of concentrations for the peptide suggests that GIP is an important regulator of insulin release during hyperglycemia. The maximal effect of GIP on insulin release from the islets was seen at 10 n M , which may have approached the maximal stimulatory capacity of the jS-cell, since the addition of 1 nM A C h to medium containing 17.8 m M glucose and 10 n M GIP produced no additional increase in the release of insulin. In 2.75 m M glucose, a 100X greater concentration of GIP (100 nM) was required to produce a significant stimulation of insulin secretion. Such a GIP level would not be achieved under normal physiological conditions and therefore, under fasting conditions, GIP probably exerts no influence on insulin secretion. 159 Although A C h (1 /iM) was by itself a potent stimulator of insulin secretion from the islets, it did not further enhance GIP-stimulated insulin secretion in this preparation. In the perfused rat pancreas, GIP and A C h have been shown to have a potentiating interaction on insulin secretion in 8.9 m M glucose (Verchere et al, 1991), while their stimulatory effects in 17.8 m M glucose were less than additive (Muller et al, 1986). Thus, the high glucose concentration employed in these experiments (17.8 mM) may have masked the synergistic interaction of GIP and A C h that was found in other studies (Zawalich et al, 1989; McCullough et al, 1985). Several explanations for the present observation are possible. First, as described, the maximal stimulatory capacity of the /3-cell may have been attained in the presence of these stimuli. The concentration of GIP used was tenfold greater than that used in the perfused pancreas studies. Second, somatostatin levels in the incubation medium may have increased and inhibited insulin secretion. The addition of somatostatin antiserum to static incubations of isolated islets has been shown to enhance insulin secretion (Schatz and Kullek, 1980). Using a flow-through perifusion system, in which somatostatin secreted from the islets would be washed away, Zawalich et al (1989) showed that A C h potentiated the insulinotropic effect of C C K and GIP. However, since A C h suppressed GIP-stimulated somatostatin secretion in the present study, less of an inhibitory influence of endogenous somatostatin on the /3-cell would be expected in the presence of A C h . In 2.75 m M glucose, neither arginine (10 mM) nor GIP (10 nM) was able to significantly stimulate insulin secretion from isolated islets. However, when incubated together, the secretagogues produced a 2.5 fold stimulation of insulin release over 2.75 m M glucose alone. A similar effect has been observed in the perfused rat pancreas, where GIP weakly stimulated insulin secretion in low glucose in the presence of arginine. These data suggest that under fasting conditions, /3-cell sensitivity to GIP can be induced by the presence of non-glucose secretagogues. 160 Thus, there may be certain physiologic conditions, such as following a high protein, low carbohydrate meal, when GIP may be able to influence insulin release. Since GIP release can be induced by basic amino acids such as arginine (Thomas et al, 1978), it follows that GIP may have an indirect role in the regulation of protein metabolism via insulin (Unger and Dobbs, 1978). Since arginine is a "fuel" stimulus for the /3-cell (Prentki and Matschinsky, 1987), the ability of GIP to stimulate insulin secretion may depend more on the metabolic state of the /3-cell than on the glucose concentration per se. Indeed, the presence of intermediates of glucose metabolism such as glyceraldehyde have been shown to be sufficient to allow the insulinotropic action of GIP to occur (Dahl, 1983). Pederson and Brown (1978) showed that in the presence of low glucose concentrations, GIP has a stimulatory effect on glucagon secretion from the perfused rat pancreas. The present results provide the first demonstration that near physiological concentrations of GIP (1 nM) can stimulate glucagon secretion from isolated rat islets. As in the perfused pancreas, the effect of GIP on a-cell secretion was dependent upon the presence of low glucose levels, reaching statistical significance only when the glucose concentration of the incubation medium was 2.75 m M . The data supports the idea that GIP is a physiological regulator of glucagon secretion. When considered with its observed effects on /3-cell secretion, these studies suggest that GIP is capable of influencing carbohydrate metabolism in both fasting and fed states, by stimulating the release of either glucagon or insulin according to the prevailing level of glycemia. A glucagonotropic action of GIP has also been demonstrated in two previous in vitro studies. In neonatal rat pancreatic monolayer cultures, 0.2 n M GIP was shown to stimulate glucagon secretion in both low (1.67 mM) and high (16.5 mM) glucose (Fujimoto et al, 1978). The apparent lack of influence of glucose on GIP-stimulated glucagon release in that study was not seen in the present investigation, 161 although at intermediate glucose concentrations (6.6 and 8.9 mM) a tendency towards stimulation by the highest GIP concentration (10 nM) was observed. The disparity may be related to differences in the experimental preparations (isolated islets vs pancreatic monolayer cultures) or age of the animals used (adult vs neonatal). Considerably higher concentrations of GIP (100 nM) were required to stimulate the secretion of glucagon from freshly isolated mouse islets (Bailey et al, 1990). As was observed in these experiments, 10 n M GIP had no effect on glucagon secretion from isolated mouse islets in 5.6 or 16.7 m M glucose. The lack of a glucagon-releasing effect of GIP at high glucose concentrations in isolated islets and in the perfused pancreas may be partly due to a local inhibitory effect of insulin, secreted in response to increased glucose and GIP levels. In the perfused pancreas, especially, where insulin may be delivered by the islet vasculature directly to the a-cells in very high concentrations (Samols et al, 1988), any direct stimulatory effect of GIP on the a-cell could be suppressed. In the present study, insulin concentrations in the islet incubation medium in the presence of 17.8 m M glucose plus 10 n M GIP attained levels of 0.1-1.0 m U / m l . Although this concentration is lower than the concentration of exogenous insulin usually required to inhibit glucagon secretion in vitro (approximately 25 mU/ml) (Samols et al, 1986), the actual concentration of insulin that the a-cell was exposed to following B-cell secretion and diffusion of insulin through the a-cell mantle was probably several fold greater than the measured value. In 6.6 and 8.9 m M glucose, where a weak stimulation of glucagon secretion by 10 n M GIP was apparent, GIP-stimulated insulin release was not as pronounced and therefore the suppressive effect of insulin on GIP-induced a-cell secretion may not have been as great as in 17.8 m M glucose. The possible local inhibitory effects of insulin may also not have been as prominent in the rat pancreatic monolayer study of Fujimoto et al (1978), where the a-cells do not surround a core of /3-cells and therefore might not be exposed to as high a 162 concentration of endogenous insulin as in intact islets. As with insulin secretion, glucagon release from the cultured islet preparation was found to be insensitive to changes in the glucose concentration of the incubation medium. In the absence of GIP, glucagon secretion was approximately the same at all glucose concentrations tested except 4.4 m M , where it was observed to be elevated over other glucose concentrations, including 2.75 m M . Although glucose has been shown to suppress glucagon secretion from the perfused pancreas (Pederson and Brown, 1978) and from isolated islets in certain studies (Verspohl and Ammon, 1978), a lack of glucose inhibition of glucagon secretion has been recently documented in isolated islets (Bolaffi et al, 1990; Vara and Tamarit-Rodriguez, 1988). The reason for the lack of effect of glucose on glucagon secretion in some situations is unclear and was not a focus of this particular study. If the control of glucagon release is modulated to a large degree by endogenous insulin as discussed, then the diminished insulin response to high glucose seen in this study might have contributed to the lack of glucose-mediated suppression of a-cell secretion. The maximum glucagon response observed from the islets, in 2.75 m M glucose plus 100 n M GIP, was less than a 2X increase in secretion over basal. The stimulatory effect of GIP on glucagon release in 2.75 m M glucose was also found to be modest in the perfused rat pancreas (Pederson and Brown, 1978). However, the a-cell response to GIP in that preparation was substantially augmented by the addition of 10 m M arginine. Since the arginine concentration of the islet incubation medium used in this study (DME) was only 0.4 mM, experiments were performed on cultured islets using 2.75 m M glucose D M E supplemented with an additional 10 m M arginine. In contrast to the situation in the perfused pancreas, only a weak glucagon response to arginine alone was observed in the islets and the stimulatory effect of GIP on the a-cell was not further enhanced by the presence of arginine. 163 The lack of a potent effect of arginine on the a-cell may be related to alterations in the signal transduction pathway for arginine-stimulated glucagon secretion during culture. The mechanism of arginine-induced glucagon secretion is unknown but has been suggested to involve increases in a-cell intracellular C a 2 + (Wang and McDaniel, 1990). As discussed, this pathway has been shown to become attenuated in the B-cell during two day culture of isolated islets (Verspohl et al, 1988). In addition, it is possible that the maximum a-cell response to GIP was already achieved at very low concentrations of GIP (0.1-1.0 nM), since 0.1 n M generated a non-significant yet discernible increase over basal levels, while GIP concentrations greater than 1 n M did not produce obvious further increases in glucagon secretion. In this study, basal glucagon levels were not always detectable and therefore the results of several experiments were excluded from analysis. Further, the inter-assay variation of the glucagon R I A was relatively large (16 %), which contributed to the high degree of variance in these results. It would therefore be of interest to more closely test the effects of GIP concentrations between 0 and 1 n M on glucagon secretion in this preparation. If a more sensitive RIA is not available, the glucagon concentration in samples of incubation medium could be augmented by increasing the number of islets per well, or by concentrating the samples by lyophilization followed by reconstitution of the sample in a small volume of assay buffer. GIP has been shown to be a potent stimulator of gastric somatostatin secretion in the perfused rat stomach (Mcintosh et al, 1981), suggesting a pathway whereby it exerts its inhibitory influence on gastric acid secretion. Yet studies on a possible role for GIP in the regulation of islet somatostatin secretion have yielded equivocal results. In the few experiments conducted to date, pharmacological concentrations of GIP (>10 nM) have generally been required to produce a stimulation of somatostatin secretion from the perfused pancreas or isolated islets. The experiments presented here show that GIP, at a concentration (1 nM) near the 164 physiological range for the peptide, is capable of stimulating pancreatic somatostatin release. The effect appeared to be both concentration-dependent and subject to modulation by glucose. In the presence of glucose concentrations near fasting and post-prandial levels in the rat (4.4 and 6.6 m M , respectively), neither 1 nor 10 n M GIP had a noticeable effect on somatostatin secretion from isolated islets. However, a stimulatory effect of GIP became apparent in the presence of glucose concentrations mimicking hypo- or hyperglycemic conditions. In 2.75 m M glucose, a modest stimulation was produced by both 1 and 10 n M GIP. The strongest stimulation by GIP was seen in the presence of 17.8 m M glucose. A 2-fold increase was observed when 1 n M GIP was added to medium containing 17.8 m M glucose, and higher concentrations of the hormone (up to 100 nM) appeared to further enhance 5-cell secretion at this glucose level. Although a clear glucose-dependency was not observed, the results indicate that glucose modulates GIP-stimulated somatostatin from the islet. A C h (1 pM) potently suppressed the stimulatory action of 10 n M GIP in 17.8 m M glucose. The islet 6-cell and its sensitivity to the presence of A C h thus showed striking similarity to the gastric S-cell. In the perfused rat stomach, GIP-stimulated somatostatin release was abolished by cholinergic or vagal stimulation (Mcintosh et al, 1981). While these experiments have clearly shown that GIP is capable of inducing islet somatostatin secretion, the conditions under which this stimulation occurs in vivo must still be determined. In the in situ perfused rat pancreas, in which the in vivo islet vascular flow is more accurately represented, glucose plus GIP (0.2 nM) was found (Figure 30) to be without effect on pancreatic somatostatin release. Other investigators have demonstrated stimulation of somatostatin secretion from the perfused pancreas of rat (Schmid et al, 1990) or dog (Ipp et al, 1977) when higher concentrations of GIP (> 10 nM) were employed. In one study in the 165 perfused human pancreas, 1.0 n M GIP was sufficient to enhance somatostatin secretion (Brunicardi et al, 1986b). It must be considered that in the perfused pancreas, insulin secreted in response to GIP may suppress somatostatin release via the islet microvasculature. Infusion of insulin antiserum has been shown to augment somatostatin release from the perfused rat pancreas (Samols et al, 1988). In addition, infusion of exogenous insulin was found to suppress GIP-stimulated somatostatin secretion in the perfused human pancreas (Brunicardi et al, 1986b). It follows that at low glucose concentrations, when the insulinotropic action of GIP is greatly diminished, a more noticeable GIP stimulation of islet somatostatin secretion might be unmasked. A modest stimulation of S-cell secretion was observed in 2.75 m M glucose in this study, and this effect might become even more apparent in vivo, where insulin is delivered directly to the 5-cell via the vasculature. It is difficult to ascribe a physiological function to the increase of islet somatostatin secretion produced by GIP, since the function of pancreatic somatostatin remains unclear. If islet somatostatin acts primarily as a local inhibitor of insulin secretion, as has been suggested (Orci and Unger, 1975), the increase in somatostatin release produced by high glucose and GIP concentrations may serve to prevent the hypersecretion of insulin when these stimuli are present. In low glucose, islet somatostatin may help diminish the inappropriate stimulation of insulin secretion by GIP, and could even contribute to the mechanism of the glucose-dependency of GIP-stimulated insulin release. However, the question of whether a paracrine action of somatostatin exists in the islets is still unanswered and awaits more advanced experimental techniques. Islet somatostatin may also influence exocrine pancreatic secretion, as acinar tissue is thought to be exposed to large concentrations of islet hormones via the islet-exocrine portal blood flow (Williams and Goldfine, 1985). Somatostatin has been shown to exert a potent inhibitory effect on pancreatic amylase secretion 166 (Garry et al, 1989). It follows that GIP may indirectly alter acinar secretion via its stimulation of islet somatostatin release. GIP was previously found to augment ACh-induced amylase secretion from the perfused rat pancreas (Muller et al, 1986); it was suggested that this effect was mediated indirectly through the stimulation of insulin release by GIP, which in turn enhanced pancreatic exocrine secretion. In that study the stimulatory action of GIP on the exocrine pancreas was only seen when A C h was infused concomitantly with GIP. That finding bears particular consideration in view of the observation in the present study that A C h potently suppressed GIP-stimulated somatostatin release. It is tempting to speculate that in the presence of A C h , such as would be seen during increased vagal activity associated with meal ingestion, GIP-stimulated somatostatin release is diminished and therefore the predominant effect of GIP on exocrine secretion would be stimulatory (via insulin). Conversely, in the absence of vagal activity, such as in the interdigestive period (when GIP levels are known to remain elevated; Salera et al, 1983), GIP-stimulated somatostatin release would suppress exocrine pancreatic secretion. The availability of an in vitro islet preparation sensitive to GIP should facilitate future investigation into pathophysiological conditions in which possible islet defects in the action of GIP have been implicated. In one such situation, the Zucker fatty rat, enhanced /3-cell sensitivity to GIP has been suggested to be a contributor to the fasting hyperinsulinemia observed in these animals. Chan et al (1984) found that the insulinotropic action of GIP was glucose-dependent in the perfused pancreas of lean but not obese rats. It was proposed that a /3-cell defect existed in fat animals, either at the level of signal transduction, the putative /3-cell GIP receptor, or /3-cell interactions with other islet hormones, which altered islet sensitivity to GIP. However, the perfused pancreas is an inappropriate model for elucidating the nature of such a cellular or molecular defect, since alterations in 167 pancreatic /0-cell mass, or neural or vascular influences could account for the obtained results. Therefore, the proposed defect was further investigated using islets isolated from Zucker rats. Cultured islets from obese rats secreted more insulin per islet than those from lean rats. This difference was apparent under basal conditions (4.4 m M glucose) but became considerably more pronounced when insulin secretion was stimulated by glucose and GIP. The observation may partly be explained by the clear differences in islet size and % B-ceYL content between the two phenotypes that have been previously demonstrated immunocytochemically (Larsson et al, 1977; Chan et al, 1984). In this study, islets isolated from fat rats contained approximately double the insulin content of those from lean rats. This ratio remained unchanged after two days culture, although total islet insulin content became diminished by 25-30 % during culture of both lean and obese rat islets. The higher level of insulin secretion from isolated obese Zucker rat islets has been well documented (Hayek and Woodside, 1979; Hayek, 1980; Trimble and Herberg, 1983; Tokuyama et al, 1991). Hayek (1980) showed that the elevated insulin release from obese Zucker rat islets could not be normalized by long-term tissue culture (up to 21 days). This suggested that the putative islet defect in obese rats was intrinsic and could not be eliminated by removing the islet from its metabolic environment. Further, that finding validated the use of cultured islets in the present study to examine the possible role of GIP, which was deemed necessary due to the unresponsiveness of non-cultured islets to GIP. Obese Zucker rat islets clearly secreted more insulin when stimulated by glucose and GIP compared to lean rat islets. However, to properly compare the relative sensitivity of lean and obese rat islets to these secretagogues, correction for the greater size and insulin content of obese islets must be made. Surprisingly, when insulin release was calculated as a fraction of the total islet insulin content, secretion 168 from lean rat islets was greater than obese rats in 4.4 m M or 8.9 m M glucose in the absence of GIP. This finding differed from previous work using freshly isolated islets which showed that when expressed as a percent of islet insulin content or total islet protein, insulin secretion from obese Zucker rat islets was the same or greater than that of lean rat islets under both basal and glucose-stimulated conditions (Hayek and Woodside, 1979; Verspohl et al, 1986b; Tokuyama et al, 1991). Since obese Zucker rat islets appear to undergo a greater loss of sensitivity to glucose than lean rat islets in identical culture conditions, the increased pancreatic hypersensitivity to glucose seen in obese animals (Kuffert et al, 1988) is more likely to be a compensatory response to some other metabolic alteration rather than a primary islet defect. GIP (2.0 nM) stimulated the release of insulin from obese rat islets in both 4.4. and 8.9 m M glucose. In contrast, GIP was without a significant effect on insulin secretion from lean rat islets at either glucose concentration. The results indicate that islets of obese Zucker rats are more sensitive to the insulinotropic action of GIP than those of lean rats. Thus, the increased insulin response to GIP previously seen in the perfused pancreas of obese Zucker rats (Chan et al, 1984) was not simply due to the increased pancreatic /3-cell mass in these animals but rather an enhanced islet sensitivity to the insulinotropic action of GIP. Further, the data agree with the finding of Chan et al (1984), that in obese rats, the insulinotropic action of GIP is not glucose-dependent. The present observation that GIP stimulated insulin release from obese rat islets in 4.4 m M glucose supports the hypothesis that a /3-cell defect exists in these animals which appears to lower the glucose threshold for the insulinotropic action of GIP. The lack of glucose-dependency of the action of GIP may contribute to the fasting hyperinsulinemia seen in these animals, since even in fasting conditions the /3-cell would be subject to continual stimulation by GIP. 169 The lack of responsiveness of lean rat islets to GIP in 8.9 m M glucose was somewhat surprising in view of past demonstrations that islets from lean Zucker rats resemble those of normal Wistar rats in their response characteristics (Verspohl et al, 1986b). Wistar rat islets were found in this study to respond to 8.9 m M glucose plus 1.0 n M GIP with a 3-fold increase in insulin secretion. Further, the perfused pancreas of lean rats has been shown to be responsive to physiological levels of GIP (Chan et al, 1984). Although a tendency towards stimulation in the presence of 8.9 m M glucose plus 2.0 n M GIP was observed with isolated lean Zucker rat islets, there was wide variation in the insulin response seen in these experiments. This variation may have been due to the unknown genotype of the lean rats used in this study. Lean rats of heterozygote genotype (Fa/fa) were recently found to have an insulin response to glucose in the perfused pancreas that was less than obese rats yet significantly greater than that of homozygous (Fa/Fa) lean rats (Blonz et al, 1985). This suggested that the presence of a fa gene in lean animals confers some of the physiological traits seen in the obese phenotype. In the present study, the genotype of the lean rats used was unknown, but presumably occurred in a ratio of approximately 2:1 (heterozygoushomozygous). Since in a typical experiment the islets from 2-4 rats were pooled, an unknown mixture of islets from both genotypes would have been placed in each experimental well, contributing to the lack of a consistent response pattern. The exact cause of the /3-cell hypersensitivity to GIP was not elucidated, although several hypotheses should be considered. First, the density and/or affinity of GIP receptors on the /3-cell might be increased in obese rats. Refinement of the GIP receptor binding assay developed in this work should allow the future examination of this possibility. Alterations in 0-cell signal transduction might also cause an increase in sensitivity to GIP. Black et al (1988) demonstrated that the insulin response to the adenylate cyclase activator forskolin was exaggerated in 170 isolated islets of the ob/ob mouse. Further, forskolin-stimulated insulin secretion could be inhibited by verapamil in islets of lean but not obese mice, suggesting that the increased sensitivity to forskolin was linked to a defect in the /3-cell voltage-dependent C a 2 + channel. Since GIP may act via the stimulation of adenylate cyclase (Siegel and Creutzfeldt, 1985), a similar defect may be present in obese Zucker rat islets. If this were the case, the defect may not be specific for GIP and increased secretion should also be observed in the presence of other secretagogues which act through the c A M P pathway, such as GLP-I (7-36-NH2). Beck and Max (1987) found that GIP-stimulated incorporation of fatty acids into adipose tissue was enhanced in obese Zucker rats, suggesting that other tissues also exhibit altered sensitivity to GIP in obesity. However, this effect was not due to an alteration in the c A M P pathway, since GIP was without effect on c A M P levels in adipose tissue in both lean and obese Zucker rats (Beck and Max, 1988). The possibility that intra-islet effects may be involved in the increased insulin response to GIP in obese islets was also examined by determining the glucagon and somatostatin responses to GIP. Despite the clearly increased size of obese rat islets, islet content of both glucagon and somatostatin was greater in the islets of lean rats both before and after culture. In agreement with this, circulating glucagon levels have been shown to be lower in obese Zucker rats (Eaton et al, 1976). In the presence of 4.4 m M glucose, GIP induced increases in glucagon release of similar magnitude in both lean and obese Zucker rat islets. However, in 8.9 m M glucose, GIP stimulated the release of glucagon only from obese rat islets. Thus, it appears that the mechanism underlying the glucose-dependency of the action of GIP is also uncoupled in the a-cells of obese rats. Glucagon secretion from obese but not lean rat islets was suppressed by increasing the glucose concentration from 4.4 to 8.9 m M in the absence of GIP. This contrasts with the results of Rohner-Jeanrenaud and Jeanrenaud (1988), who found that administration of glucose in vivo decreased 171 circulating glucagon levels in lean but not obese rats. However, the Zucker rats used by that group exhibit markedly different metabolic traits than the ones used in this study, including abnormal glucose tolerance. Somatostatin secretion from the perfused pancreas has been shown to be elevated in obese Zucker rats (Seino et al, 1981). It was suggested that the increase in somatostatin release was a compensatory attempt to suppress insulin hypersecretion via a paracrine mechanism. No differences in somatostatin secretion between lean and obese Zucker rat islets were observed in these experiments, although insulin secretion from obese rat islets appeared to be more sensitive to the inhibitory effects of somatostatin. The present study could not distinguish whether enhanced /?-cell sensitivity to GIP is a primary defect in obese rats or a compensatory response to peripheral insulin resistance or other metabolic changes. The observation that differences in islet sensitivity to GIP between lean and obese rats were still apparent after two day culture under identical conditions suggests that the altered sensitivity to GIP is not a secondary manifestation of other metabolic changes. However, Hayek (1982) showed that islets isolated from fatty Zucker rats just prior to the onset of obesity and cultured for 1 week had insulin secretory patterns identical to cultured lean rat islets. In addition, Chan (1984) found that the onset of increased /3-cell sensitivity to GIP in the perfused rat pancreas did not occur until 5 weeks of age, which is after the age that changes in adipose tissue (Boulange et al, 1979) and pancreatic sensitivity to glucose (Blonz et al, 1985) have been shown to appear. Regardless, increased sensitivity of both /3-cells and adipose tissue to the actions of GIP may contribute to the development and maintenance of the obese state. The isolated islet studies presented provide strong evidence that under certain conditions, GIP acts directly on islet cells to stimulate the secretion of insulin, glucagon and somatostatin. Yet because these experiments were performed 172 using heterogeneous cell preparations, the presence of intra-islet modulatory influences could not be excluded. Therefore, indisputable evidence of a direct effect of GIP on the /3-cell was not ascertained by the studies in islets. As this was a primary goal of this thesis, pure /3-cell preparations were also used to investigate the insulinotropic action of GIP. Two types of homogeneous /3-cell preparations were exploited for their unique advantages in studying /3-cell secretion: a neoplastic /3-cell line (the mouse /3TC3 cell), which is available in vast quantities for receptor binding studies, and FACS-purified /3-cells, which were obtained from normal rat pancreatic tissue and therefore may be more physiologically relevant than tumor cell lines. Using the F A C S , a cell preparation was obtained that consisted of >98 % B-cells. The cells were cultured for two days, since this was shown to increase responsiveness to GIP in isolated islets (Siegel and Creutzfeldt, 1985). In 17.8 m M glucose, GIP was found to stimulate insulin secretion from FACS-purified /3-cells in a concentration-dependent manner. The results support the idea that GIP exerts its insulinotropic action by direct interaction with pancreatic /3-cells. However, the effect was considerably weaker than that observed in cultured islets. The minimum effective concentration of GIP was 100 n M , which was 100X greater than the concentration of GIP required to stimulate insulin secretion from islets. There are several possible explanations for the diminished insulin response to GIP observed in FACS-purified /8-cells. Cell damage or death caused during the enzymatic digestion and sorting procedure must first be considered, although this is an unlikely cause of these observations for the following reasons. First, the F A C S selected cells partly on the basis of F A D fluorescence (Pipeleers et al, 1985a), a parameter which should have excluded most non-living or damaged cells. Second, viability tests before and after culture, as well as following incubations in release medium, showed that > 90 % of the cells were viable. Third, the cells behaved like living cells in that they formed multicellular aggregates and adhered to the culture 173 dish; non-attached cells were always discarded. Fourth, and most importantly, the cells responded to physiological concentrations of glucagon (0.1 nM) and GLP-I (7-36-NH 2) (1.0 nM). The insulin response to glucagon was similar to that observed by Pipeleers et al (1985b). Thus, the defect appears to be specific for the stimulatory action of GIP. Whether this defect exists at the receptor level or intracellularly could not be ascertained by these experiments, although some speculations are made in the following discussion. As discussed, the results obtained here and in other studies (Siegel and Creutzfeldt, 1985; 1988) support the idea that the GIP receptor is particularly susceptible to enzymatic damage yet can be restored in culture. In the procedure for purification of /3-cells, a second enzymatic step is added (trypsinization) following collagenase digestion as well as mechanical manipulations (cell dispersion and sorting). It follows that /3-cell GIP receptors may be exposed to further damage during this procedure. Moreover, it is unknown whether the recovery of /3-cell GIP receptors in culture is at all dependent upon 0-cell contacts with other islet cells. This is a possibility since mixing pure j0-cells with cells from the non-/3-cell (a-cell-enriched) fraction prior to culture greatly enhanced /3-cell sensitivity to GIP. The density of functional GIP receptors on FACS-purified /3-cells may therefore be lower, even after culture, than in other preparations. This could not be determined in the present study, since the yield of /3-cells obtained from the F A C S was insufficient for GIP receptor binding experiments, even though others have used FACS-purified /3-cells to study the /3-cell glucagon receptor (Van Schravendijk et al, 1985). Alterations in signal transduction in pure /3-cells may also be involved in the diminished insulin response to GIP. A consistent observation in this and other studies has been that FACS-purified /3-cells were relatively insensitive to the stimulatory effects of glucose, yet retained sensitivity to glucagon, forskolin and 174 theophylline (Pipeleers et al, 1985b; Wang et al, 1990). Pipeleers et al (1985b) suggested that in the absence of contacts with other islet cells, pure single /3-cells were unable to produce sufficient levels of c A M P to induce an insulin response to glucose. Insulin secretagogues which act via increasing /3-cell c A M P levels were able to increase /3-cell production of c A M P and thus potentiate glucose-induced insulin secretion. Wang et al (1990) recently showed that the sensitivity of purified /3-cells to glucose could also be greatly enhanced by cytochalasins B and D . It was suggested that c A M P and the cytochalasins modulate the state of the /3-cell cytoskeleton to sensitize the cell to increases in the intracellular calcium concentration ([Ca 2 +]i). Since GIP weakly stimulated the secretion of insulin from these cells, it is possible that the peptide acts only partly or not at all through the activation of adenylate cyclase. A lack of stimulation of adenylate cyclase activity was reported in membrane preparations of insulin-secreting hamster pancreatic tumors even though GIP receptor binding to this tissue was demonstrated (Maletti et al, 1984). In addition, GIP-stimulated incorporation of fatty acids into adipose tissue was found not to be mediated by c A M P (Beck and Max, 1988). However, GIP has been shown to produce a modest increase in c A M P levels in cultured islets (Siegel and Creutzfeldt, 1985) and in human insulinoma tissue (Maletti et al, 1987) in the presence of IBMX. Further, in the In III /3-cell line, half-maximal stimulation of c A M P production by GIP was seen at a concentration (30 nM) similar to that required to induce a half-maximal insulin response (10 nM) and near the (4 nM) for the high affinity GIP receptors demonstrated on this tissue (Amiranoff et al, 1984). Therefore, whether or not GIP activation of adenylate cyclase in the /3-cell is a step in the mechanism of action of GIP-induced insulin release remains unresolved and deserves further study. Unfortunately, unlike previous reports (Schuit and Pipeleers, 1985), the yield of pure /3-cells obtained by cell-sorting was 175 not adequate for the measurement of intracellular c A M P levels. The sensitivity of FACS-purified /3-cells to the insulinotropic action of GIP was markedly increased by the addition of 10 n M glucagon to the incubation medium or by culturing the /3-cells with an equal number of cells from the a-cell enriched fraction. In fact, the minimum concentration of GIP (1 nM) required to produce an insulin response from /3- plus non-/3-cell mixtures was similar to the lowest effective concentration of the peptide in cultured islets. Therefore, it appears that /8-cell contacts with non-/3-cells, most likely a-cells, are critical for GIP-induced insulin release to occur. It could not be established how contacts with a-cells confer /3-cell sensitivity to GIP. Since the presence of glucagon also enhanced GIP-stimulated insulin secretion, it must be considered as a possible mediator of this effect. Glucagon increases /3-cell c A M P (Schuit and Pipeleers, 1985) and so GIP may require /3-cell c A M P levels to be elevated for the cell to be sensitive to its stimulatory action. In support of this hypothesis, Brunicardi et al (1986a) found that theophylline, which inhibits c A M P breakdown, potentiated GIP-stimulated insulin release from the perfused rat pancreas but was without effect on glucagon-stimulated insulin secretion. However, the results of other studies have not supported this idea, since GIP-induced insulin release was shown to be potentiated by secretagogues which act via the phosphoinositol pathway (Sandberg et al, 1988; Verchere et al, 1991) but not those which act via c A M P (Sandberg et al, 1988; Fehmann et al, 1989). In addition, the microarchitecture of the islet in situ makes a local influence of glucagon on the /3-cell unlikely, since the direction of islet blood flow is known to be /3- to a-cell in the rat (Samols et al, 1988) and only a minority of /3-cells directly contact a-cells in the periphery of the islet. The possibility that the decreased sensitivity to GIP of FACS-purified /3-cells is related to the lack of direct communication between islet cells in this preparation also warrants consideration. Cells within a single islet have been shown to be 176 coupled intracellularly and electrophysiologically, such that activation of a single islet cell of lower stimulatory threshold could result in the recruitment of other cells within the same islet to secretory activity (Pipeleers, 1987). Thus, increased numbers of contacts of B-cells with other islet cells has been shown to increase insulin secretion (Pipeleers et al, 1982; Bosco et al, 1989). In the present study, the degree of reaggregation of pure, single /3-cells into multicellular groups was not closely regulated. Some single cells remained isolated while others appeared to make close contacts with other /3-cells. If the mechanism of GIP-stimulated insulin secretion is more dependent than other insulin secretagogues upon intra-islet communication, disruption of this process could diminish the insulin response to GIP. Glucose-stimulated insulin release appears to be more dependent upon homologous contacts among /3-cells than contacts with other islet cell types (Pipeleers et al, 1985b; Bosco et al, 1989). The results of this study suggest that heterologous contacts (with a-cells) are more important for GIP-stimulated insulin release from /3-cells. However, since the cell density of the mixed B- and non-/3-cell cultures was 2X that of the pure /3-cell cultures, the increased sensitivity of the mixed cell population may have been at least partly due to the greater degree of homologous contacts between /3-cells. Functional subpopulations of /3-cells with varying thresholds of sensitivity to glucose have recently been demonstrated both in vivo (Stefan et al, 1987) and in individual /3-cells in culture (Hiriart and Ramirez-Medeles, 1991). It is therefore also possible that only certain subpopulations of /3-cells are sensitive to the stimulatory actions of GIP. Upon stimulation by GIP, these cells would recruit neighboring islet cells to secretory activity. Using immunocytochemical techniques, Fujimoto et al (1978) localised GIP receptors in neonatal pancreatic monolayer cell cultures to a minority subpopulation of these cells, although the cell type was not identified. This could explain the apparent low density of GIP receptors in normal 177 islet tissue. Further, if such a subpopulation of cells exists, these cells may reside in the mantle of the islet near a-cells, since contacts with a-cells appear to increase B-cell sensitivity to GIP. The two biologically active GIP receptor probes described in this thesis may be useful tools to further investigate the localisation of GIP receptors in islets. GIP resembled CCK-8 more than glucagon or GLP-I (7-36-NH2) in insulin-releasing potency in FACS-purified /3-cells. Both CCK-8 and A C h , which act via increasing phosphoinositol metabolism and intracellular calcium (Prentki and Matschinsky, 1987), were considerably less effective than glucagon or GLP-I (7-36-N H 2 ) in this preparation. A n insulinotropic effect of these secretagogues, like GIP, has been observed at 100X lower concentrations in isolated islets (Verspohl and Ammon, 1987; Garcia et al, 1988). Thus, sensitivity to the phosphoinositol-calcium pathway may be diminished in pure /3-cells. Measurement of [Ca 2 + ]j in purified /3-cells (by fura-2AM microspectrofluorimetry) revealed that the minimum concentrations of CCK-8 and A C h required to produce an increase in [Ca 2 + ] ; varied markedly between individual cells (J. Wang, unpublished observations). If individual /3-cells have different thresholds of sensitivity to CCK-8 and A C h , as this suggests, then the diminished insulin response to these secretagogues may be related to the lack of intercellular communication between islet cells in this preparation. GIP, at concentrations as high as 10 /xM, failed to induce [Ca 2 + ]j changes in pure /3-cells, implying that the hormone does not act via the phosphoinositol-calcium pathway (J. Wang, unpublished observations). However, GIP (0.2 nM) has been shown to suppress carbachol-stimulated inositol phosphate production in neonatal rat pancreatic cell cultures (Mazzu et al, 1989). The mouse /3-cell line used in receptor binding studies (/3TC3) was similar to FACS-purified /3-cells in its poor responsiveness to GIP. This could not be attributed to damage to GIP receptors during handling of the cells, since the 178 receptor binding experiments clearly showed that nanomolar concentrations of GIP displaced 1 2 5 I -GIP from these cells. Therefore, a defect in the coupling of GIP binding to secretion may exist in /8TC3 cells. Whether this defect is similar or identical to the one in FACS-purified /3-cells is unknown. The results of studies with this and other tumor cell lines must be interpreted with caution, since the responsiveness of these cells often changes with repeated passages in culture. The /3TC3 cell line was chosen for use in the binding studies presented here because it was shown to closely resemble normal /3-cells in response characteristics (D'Ambra et al, 1990). Yet basal secretion of insulin from these cells was found to be 10X that claimed by D'Ambra and co-workers. This high basal secretory rate may have masked further stimulation by GIP. The reason for the high basal secretion of insulin was not ascertained, but may have been related to experimental differences, particularly in the culture and handling of the cells. To see if the poor insulin response to GIP in /3TC3 cells was linked to an inability to produce c A M P as was suggested for purified rat jS-cells (Pipeleers et al, 1985b), the effects of I B M X and forskolin on insulin secretion were examined. Both agents augmented insulin secretion in 17.8 m M glucose. Yet, neither inhibition of c A M P degradation (by IBMX) or direct stimulation of c A M P production (by forskolin) was able to increase the sensitivity of /3TC3 cells to GIP. Thus, the reason for the poor sensitivity of these cells to GIP was not elucidated, but could not be attributed to /3-cell insensitivity to cAMP. The finding that GIP stimulated insulin secretion from isolated islets, F A C S -purified /3-cells and /3TC3 cells suggested that GIP receptors must be present on B-cells to mediate its insulinotropic action. This idea was supported by previous demonstrations of GIP binding to the In III cell line (Amiranoff et al, 1984) and human insulinoma tissue (Maletti et al, 1987). Yet prior to the present study, GIP binding to normal islet tissue had not been shown. Brown et al (1989) suggested two 179 major reasons for this: the lack of a suitable tracer for GIP binding studies and the relatively low density of functional GIP receptors on isolated islets and islet cells. In the present study, attempts were made to overcome both of these potential problems. A suitable probe for use in GIP receptor studies was obtained by H P L C purification of iodinated GIP. H P L C analysis of the products of GIP iodination suggested the presence of numerous iodinated peptides. These likely included various multi-iodinated forms of GIP 1-42 and possibly contaminating peptides known to be present in the GIP fraction used ( E G III), namely GIP 3-42 and C C K . However, a similar H P L C elution profile was obtained with iodinated synthetic GIP 1-30, suggesting that none of the major peaks were due to iodinated forms of contaminants such as C C K . The H P L C profile somewhat resembled that obtained by Maletti et al (1984), who successfully used HPLC-purified 1 2 5 I -GIP to demonstrate receptor binding to membrane preparations of hamster B-ce\\ tumors, in that two peaks were particularly prominent. Non-iodinated GIP 1-42 and the biologically inactive fragment GIP 3-42 eluted well before the major radioactive peaks, eliminating the possibility that the native hormone could interfere in the binding assay by displacing 1 2 5 I -GIP. Although at least 6 peaks of radioactivity were apparent on most H P L C elution profiles of 1 2 5 I -GIP, 4 peaks were chosen for further analysis because of their high yield and reproducibility. It is desirable that radioligands used in radioreceptor binding assays be of high specific activity without significant loss of biological activity on the target tissue. Thus, although increased specific activity confers greater radioactivity per bound molecule in a binding assay, increased incorporation of 1 2 5 I into the peptide may interfere with binding of the radioligand to its receptor, particularly if the tyrosine residues to which 1 2 5 I is attached occur in the biologically active site of the molecule. Analysis of the iodination state of the different peaks obtained from 180 H P L C purification of 1 2 5 I -GIP revealed that the 4 major peaks all contained predominantly mono-iodinated tyrosine residues, suggesting that the iodination procedure employed minimized the production of di-iodinated forms. This is probably a result of the short exposure time to chloramine T (15 s) which is used in the iodination procedure to minimize oxidative damage to the GIP molecule. The earlier eluting peaks (1 and 2) appeared to contain exclusively mono-iodinated tyrosine residues, whereas the later eluting peaks (3 and 4) contained increasing amounts of di-iodinated tyrosine residues. Therefore, increasing the degree of incorporation of 1 2 5 I appears to increase the retention time of the GIP molecule. Since peaks 1 and 2 were determined to have approximately 0.75 1 2 5 I atoms per GIP molecule, as calculated from the specific activity, it is most likely that only one tyrosine residue in each of these peaks is mono-iodinated. Although the iodination state of the tyrosine residues of peaks 5 and 6 was not analysed due to insufficient yield, the higher specific activity of these peaks and thus the higher theoretical number of 1 2 5 I atoms per GIP molecule suggests the presence of more di-iodinated tyrosine residues in these fractions. As expected, peaks 1 and 2 were found to have a higher specific activity (approximately 3.5-fold) than that of unpurified 1 2 5 I -GIP. However, the specific activities of peaks 3 and 4, as determined by self-displacement curves, were lower than the unpurified tracer. In addition, the number of 1 2 5 I atoms per GIP molecule indicated that only 10-20 % of the GIP molecules in peaks 3 and 4 were iodinated. This result seems surprising in view of the demonstration that non-iodinated GIP clearly eluted before the radioactive peaks, suggesting that non-iodinated GIP should not have been present in the radioactive fractions. It is possible that incorporation of 1 2 5 I into the GIP molecule interfered with binding of the molecule to the antibody used in the RIA (RK343F; Dr. Linda Morgan, England), since determination of specific activity by self-displacement curves assumes that the 181 antibody has equal affinity for both the iodinated peptide and the native peptide. The binding site of antibody RK343F is not known, although it was found in the present study to detect GIP 1-30 and therefore is N-terminally directed. Since the tyrosine residues on GIP occur at positions 1 and 10, it follows that iodination of GIP at these sites may interfere with binding with this particular antibody, resulting in an underestimation of the true specific activity. Using a protocol adopted from the receptor binding assay developed by Maletti et al (1984), peaks 1-4 were tested for their ability to bind to /3TC3 cells. Specific binding (displaced by 1 /xM GIP) was demonstrated with all four peaks, although peak 2 consistently exhibited slightly higher specific binding than the other 3 peaks. Peak 2 was therefore tested for its ability to stimulate insulin release from the perfused rat pancreas. The peak was only slightly less effective than the native hormone in stimulating insulin secretion in the presence of 17.8 m M glucose. In addition, further H P L C analysis of peak 2 suggested that the fractions collected during elution of the peak were of high purity. Thus, peak 2 appeared to meet the criteria of an appropriate radioligand, in that it was of relatively high specific activity yet biologically active, and was displaceable by the native hormone. Binding of 1 2 5 I -GIP (peak 2) to BTC3 cells was inhibited by non-iodinated GIP 1-42 in a concentration-dependent manner. The lowest concentration of GIP 1-42 capable of producing a significant displacement of 1 2 5 I -GIP was 0.5 n M . This concentration is within the physiological range of concentrations of the peptide and similar to the lowest concentration of GIP (1 nM) shown to stimulate insulin secretion from cultured islets. Much higher (500 X) concentrations were required to produce a significant stimulation of insulin secretion from /3TC3 cells, although a slight insulin response was apparent at concentrations as low as 1 n M . As discussed earlier, the results suggest the presence of a defect in the mechanism coupling GIP receptor binding to insulin secretion in /3TC3 cells. Binding of 1 2 5 I -GIP was highly 182 specific in that no other peptide tested, including those structurally related to GIP could displace the radioligand. The results concur with several previous studies which demonstrated the presence of specific receptors for GIP on neoplastic /3-cell tissue (Maletti et al, 1984; 1987; Amiranoff et al, 1984). In addition, the data provide the first indication that the GIP receptor on the /3-cell is distinct from the receptor for GLP-I (7-36-NH2), which shows marked homology with the GIP molecule and comparable potencies in the stimulation of /3-cell secretion. GIP 1-30 (1 juM) displaced 1 2 5 I -GIP to a similar extent as GIP 1-42. The results support the recent demonstration that this fragment of the GIP molecule was as potent as GIP 1-42 in the perfused rat pancreas (Pederson et al, 1990). Although a full range of concentrations of GIP 1-30 should still be tested for their ability to inhibit 1 2 5 I -GIP binding, these preliminary results indicate that the C-terminal part of the GIP molecule is not essential for GIP receptor binding on the /3-cell. This binding assay should now allow a more rapid and complete assessment of the biologically active site of GIP, by testing the ability of various synthetic fragments of GIP to displace 1 2 5 I -GIP. In addition, since the /3TC3 cell line is a homogeneous preparation of B-cells, these data provide further support for the idea that GIP acts directly on the B-cell to exert its insulinotropic effect. Binding of 1 2 5 I -GIP to isolated, cultured rat islets was also inhibited by GIP 1-42 in a concentration-dependent manner, starting at a concentration of 1 n M . As with /3TC3 cells, the GIP receptor on rat islets appears to be very specific for GIP, since a number of peptides, including CCK-8 and GLP-I (7-36-NH2) were unable to displace 1 2 5 I -GIP even in micromolar concentrations. Previous to these experiments, GIP receptor binding had only been shown on /3-cells that had been transformed for long-term culture or were obtained from insulinoma tissue. The present demonstration of GIP receptors on isolated islets was probably made possible not only by the production of a suitable radioligand, but also by increasing 183 the number of functional GIP receptors available for binding through tissue culture and the use of a large number of islets (100 islets/experimental well). These results show that specific receptors for GIP exist on normal islet tissue. Further, when considered with the results of the binding studies performed on /8TC3 cells, the existence of GIP receptors on normal pancreatic /3-cells is strongly implicated. The demonstration of this receptor is further evidence for a physiological role for GIP as an incretin. Still, this receptor binding assay clearly needs further refinement. The considerable number of islets required to demonstrate displaceable binding of 1 2 5 I -GIP precludes the testing of a large number of different conditions. This is essential to properly evaluate the density of these receptors on islet tissue, their affinity for the GIP molecule, and the influences of glucose concentration, p H , time, temperature, and ionic strength on GIP binding to its receptor. In addition, these experiments have only demonstrated the existence of a GIP receptor on whole islets, which are a heterogeneous population of islet cells. Since GIP was shown in these studies to stimulate somatostatin and glucagon release from isolated islets, the presence of GIP receptors on a- and S-cells must also be considered likely. These could be of different affinity for GIP or may even bind different parts of the GIP molecule. The gastric 5-cell, unlike the pancreatic /3-cell, was recently shown to be less responsive to stimulation by GIP 1-30 than GIP 1-42, suggesting that the GIP receptor on the gastric 5-cell may differ from the /3-cell GIP receptor (Pederson et al, 1990). Hopefully, increases in the yield of FACS-purified islet cells will allow the future demonstration of GIP receptors on individual islet cell types. Further development of the GIP receptor binding assay should eventually permit the study of /3-cell GIP receptors in various pathophysiological conditions in which /3-cell sensitivity to GIP has been shown to be altered. For example, the present experiments confirmed the hypothesis of Chan et al (1984), that the 184 exaggerated insulin response of the perfused pancreas of obese Zucker rats to GIP was at least partly due to an increased /3-cell sensitivity to GIP. While an alteration in signal transduction in the /3-cells of obese rats cannot be excluded as a possible mechanism for increased sensitivity, changes in the number and/or affinity of /3-cell GIP receptors may be responsible for the enhanced insulin response. In other conditions, changes in circulating levels of GIP appear to cause alterations in the insulin response to the peptide. Following jejunoileal bypass (JIB) in the rat> circulating GIP levels were found to be elevated, while the insulin response to GIP in the perfused pancreas was diminished by 70 % (Pederson et al, 1982). Conversely, after 6 days total parenteral nutrition (TPN), plasma GIP concentrations remained near fasting levels in the rat, and insulin secretion from the perfused pancreas during stimulation by GIP was increased 30 % compared to control rats (Pederson et al, 1985). It was suggested that an elevation or reduction in plasma GIP levels resulted in a down- or up-regulation, respectively, of /3-cell GIP receptors. Biotinylated GIP 1-30 (B-GIP) was also produced and assessed for its suitability as a potential probe in GIP receptor studies. Biotinylated peptides are increasingly being used in the study of hormone receptors for several reasons. First, biotinylation is an easier procedure than iodination, and avoids problems associated with the handling of radioactivity and the decrease in usefulness of a radioligand due to decay. Also, the coupling of the biotin molecule to the peptide does not involve an oxidative procedure which could cause damage to certain amino acid residues as in iodination procedures (Lambert et al, 1982). Second, using different protocols, the biotin molecule can be coupled to desired residues outside of the biologically active site of the peptide so as not to interfere with receptor binding (Wilchek and Bayer, 1988). For example, biotin can easily be coupled to e-amine groups (lysine residues). Since the biologically active fragment GIP 1-30 has only 185 two lysine residues (positions 16 and 30) compared to five in the 42 amino acid native peptide, GIP 1-30 was used in the present study instead of GIP 1-42, thus reducing the potential number of multi-biotinylated forms of the peptide produced. Surprisingly, when the characterization of B-GIP was attempted by reverse-phase H P L C , no discernible peaks were observed in a gradient of 30-36 % acetonitrile, even though GIP 1-30 eluted as a single peak in this gradient. In fact, elution of B-GIP was not observed unless concentrations of acetonitrile of >40 % were used. Two peaks, each containing GIP immunoreactivity, were seen when B-GIP was eluted on a gradient of 40-80 % acetonitrile. The results suggest that the hydrophobicity of the GIP molecule was greatly altered by the coupling to biotin. This has not been previously reported for the characterization of numerous other biotinylated peptides. The yield of HPLC-purified forms of B-GIP was too low to be used in testing for biological activity and therefore the unpurified mixture was tested on the perfused rat pancreas. The insulinotropic activity of B-GIP was comparable to GIP 1-30, suggesting that biotinylated forms of the peptide retained biological activity. The result could not be attributed to the presence of unaltered GIP 1-30 in the reaction mixture, since no GIP 1-30 was detected when the biotinylated product was applied to the H P L C system. However, the possibility that non-biotinylated yet altered forms of GIP 1-30 were present in the reaction product cannot be excluded; these may have contributed to the observed insulin response. Further experiments testing the usefulness of B-GIP as a GIP receptor probe were not performed due to an insufficient amount of the probe being generated. However, these preliminary experiments suggest that further characterization of B-GIP or the continued development of other forms of biotinylated GIP is worthwhile. Such a probe may prove to be a versatile tool in future GIP receptor studies and may have certain advantages over iodinated GIP. Biotinylated parathyroid hormone (PTH) has 186 recently been used in a flow cytometry receptor binding assay requiring < 1 X 104 cells per experimental condition. Comparable sensitivity in a B-GIP binding assay might allow the detection of GIP receptor binding to FACS-purified /3-cells, which were obtained in yields too low for use in 1 2 5 I -GIP receptor binding assays. B-GIP might also be used to purify GIP receptors from /3-cell membrane preparations by affinity chromatography using avidin coupled to sepharose (Wilchek and Bayer, 1988). The perfused pancreas and FACS-purified /3-cells were used to investigate the interaction of GIP with the recently discovered neuropeptide galanin. Galanin has been shown in numerous studies to exert a potent inhibitory effect on insulin secretion under a variety of conditions (Ahren et al, 1988). Yet in marked contrast to previous studies which found that galanin suppressed stimulated insulin secretion regardless of the secretagogue, the present results indicate that the inhibitory action of galanin may be specific for certain B-ce\\ stimuli. Porcine galanin (10 nM) potently suppressed GIP-stimulated insulin secretion from the perfused rat pancreas. Although the same concentration of porcine galanin inhibited arginine-stimulated insulin secretion, it was without effect on CCK-8- or ACh-induced insulin release. The lack of effect of porcine galanin on the insulin response to CCK-8 and A C h was not due to a diminished biological activity of the porcine neuropeptide on rat j3-cells, since rat galanin was found to have the same effect as porcine galanin under every condition tested. These results imply that GIP acts on the /3-cell by a mechanism distinct from CCK-8 and A C h . Whether or not GIP- and arginine-stimulated insulin release act via a common intracellular pathway, as has been previously suggested (Elahi et al, 1982; Brunicardi et al, 1986a), could not be discounted by this study since galanin inhibited the insulin response to both secretagogues. The possibility that the inhibitory effect of galanin on insulin secretion is 187 more pronounced when GIP is the stimulus was recently supported by the work of Amiranoff et al (1988). Porcine galanin was found to potently inhibit forskolin- and GIP-stimulated insulin secretion and c A M P production in the Rinm5f /3-cell line, but was without effect on carbamoylcholine-stimulated insulin secretion. It was suggested that galanin exerted its inhibitory action on the /3-cell via the inhibition of c A M P production. Such a mechanism may also explain the inhibitory effect of galanin on argimne-stimulated insulin release since arginine has been shown to produce a modest increase in c A M P levels in islets (Garcia-Morales et al, 1984). It is unlikely that the neuropeptide influences increases in [Ca 2 + ] i induced by stimulation of phosphoinositol metabolism, since galanin failed to alter the insulinotropic action of either CCK-8 or A C h , which both act via that pathway (Prentki and Matschinsky, 1987). A stimulus-specific inhibitory effect of galanin in the rat was also observed by Yoshimura et al (1989), who found that porcine galanin inhibited glucose-, but not arginine- or potassium-stimulated insulin release in the perfused rat pancreas. Although the present study found galanin to inhibit arginine-stimulated insulin secretion, this effect was somewhat weaker than its action on the insulin response to GIP. In addition, the concentration of galanin used in this study (10 nM) was 10-fold the concentration employed by Yoshimura and colleagues. A weak inhibitory effect of galanin on arginine-stimulated insulin release has been demonstrated in other studies (Silvestre et al, 1987; Miralles et al, 1990). The physiological significance of the apparent stimulus-specificity of the inhibitory action of galanin remains to be determined. Galanin has been proposed to be a sympathetic neurotransmitter involved in the hyperglycemic response to stress, via its inhibition of insulin secretion and stimulation of glucagon secretion (Dunning and Taborsky, 1988). Since GIP levels remain high in the interdigestive period (Salera et al, 1983) and therefore could stimulate the release of insulin 188 during a stress-induced increase in blood glucose, it is possible that the presence of galanin prevents an inappropriate stimulation of insulin secretion under these conditions. In apparent conflict with the results of these experiments, Miralles et al (1988) found that CCK-8-stimulated insulin secretion was suppressed by porcine galanin in the perfused rat pancreas. However, in those experiments a priming dose of galanin was given prior to the introduction of CCK-8 followed by infusion of galanin throughout the stimulation, whereas in this study galanin was infused for only a 10 min period during the second phase of CCK-8-stimulated insulin secretion. It is possible that an inhibitory action of galanin on CCK-8 (or ACh)-stimulated secretion would only be evident when the presence of the neuropeptide precedes the stimulation. Indeed, it has been shown that galanin infusion inhibited the first phase of glucose-stimulated insulin release only when commenced 4 min prior to the glucose stimulation, suggesting that galanin acts on early events of the stimulus-secretion coupling in the /3-cell (Yoshimura et al, 1989). Porcine and rat galanin were found to have identical effects on insulin secretion from rat /3-cells under every condition tested and the two peptides were shown to be equipotent inhibitors of GIP-stimulated insulin release in vitro. These results confirm that homologous galanin inhibits insulin secretion in the rat (Miralles et al, 1990). The data also support the recent demonstration that the two peptides were equipotent in their inhibition of circulating insulin levels in response to glucose in the anesthetized rat (Schnuerer et al, 1990). The results are in accordance with recent suggestions that the C-terminal part of the galanin molecule is not necessary for its inhibitory action on the /3-cell, since the 3 amino acid differences between the porcine and rodent peptides reside near the C-terminus (Hermansen et al, 1989; Amiranoff et al, 1989). Porcine galanin was without effect on pancreatic somatostatin release in the 189 presence of GIP. The concentration of porcine galanin used (10 nM) was previously shown to inhibit gastric somatostatin release from the perfused rat stomach (Kwok et al, 1988), suggesting that gastric and pancreatic 6-cells differ in their sensitivity to galanin. Previous studies using the perfused rat pancreas also found no effect of porcine galanin on pancreatic somatostatin release under either basal or stimulated conditions (Silvestre et al, 1987; Miralles et al, 1988). However, more recent experiments have shown that rat galanin exerts a powerful inhibitory effect on somatostatin secretion from the perfused rat pancreas (Miralles et al, 1990). This suggested that the pancreatic 6-cell of the rat, unlike the /3-cell, is only sensitive to homologous galanin and supports the idea that the C-terminal part of the galanin molecule appears to be essential for its inhibitory action on the S-cell (Hermansen et al, 1989). Homologous galanin has also been shown to be a potent inhibitor of pancreatic somatostatin secretion in the pig (Messel et al, 1990) suggesting that galanin is an important neural regulator of islet somatostatin release. The effect of rat galanin on somatostatin secretion in the presence of GIP was not tested in the present experiments. In summary, these studies have fully examined the regulation of islet hormone secretion by GIP using a variety of in vitro experimental approaches. The experiments with cultured islets demonstrated that GIP exerts a powerful influence over the secretion of not only insulin, but also glucagon and somatostatin from the pancreatic islet. A n important observation was that these effects were observed at GIP levels within or near the physiological range of concentrations for the peptide, supporting the hypothesis that GIP has a role in the regulation of pancreatic endocrine secretion. This influence appears to occur via direct interaction of GIP with islet cells, since the effects were observed in vitro in islets and pure B-cells, and the existence of specific islet receptors for GIP was demonstrated. The action of GIP on islet hormone secretion was subject to considerable modulation by glucose, 190 as well as other fuel and neural inputs to the islets. In particular, GIP-stimulated insulin and glucagon secretion were clearly dependent upon the prevailing glucose concentration. GIP-stimulated somatostatin release appeared to be more influenced by the inhibitory effects of A C h . The diminished potency of GIP in FACS-purified /3-cells suggested that B-cell contacts with other islet cells, particularly a-cells, were important for GIP-stimulated insulin secretion to occur. This novel cell preparation promises to yield important new information about B-cell function. Some implications were also made about the mechanism of GIP-stimulated insulin secretion from these studies, even though direct measurements of /3-cell intracellular messengers were not made. The studies in pure /3-cells did not fully support the idea that GIP acts through the stimulation of adenylate cyclase, like other members of the secretin-glucagon peptide family. However, the specificity of the inhibitory effects of galanin for GIP-stimulated insulin release did suggest that GIP acts via a mechanism distinct from CCK-8 and A C h . The possibility that GIP activates a novel or lesser known intracellular pathway in the B-cell should be considered in future studies. In addition, the presence of a defect in the islets of obese Zucker rats was confirmed, which may contribute to hyperinsulinemia in these animals. This defect was not elucidated but appears to involve enhanced islet sensitivity to the insulinotropic action of GIP, mediated probably by either an increase in the density and/or affinity of /3-cell GIP receptors in these animals or by an alteration in /3-cell signal transduction. Finally, a receptor binding assay for GIP was developed and was used to show, for the first time, that GIP exerts its influence on the pancreatic islet by interaction with specific receptors. Future studies should examine the nature of these receptors and their possible role in the pathophysiology of certain metabolic conditions. 191 REFERENCES Adrian T E , Bloom SR, Hermansen K, Iversen J (1978) Pancreatic polypeptide, glucagon and insulin secretion from the isolated perfused canine pancreas. Diabetologia 14:413-417. Ahren B, Hakanson R, Lundquist I, Sjolund K, Sundler F (1981) GIP-like immunoreactivity in glucagon cells. Interactions between GIP and glucagon on insulin release. Acta Physiologica Scandinavica 112:233-242. Ahren B, Paquette T L , Taborsky G J Jr (1986a) Effect and mechanism of vagal nerve stimulation on somatostatin secretion in the dog. American Journal of Physiology 250:E212-E217. Ahren B, Rorsman P, Berggren P-O (1988) Galanin and the endocrine pancreas. F E B S Letters 2:233-237. Ahren B, Taborsky G J Jr (1986) The mechanism of vagal nerve stimulation of glucagon and insulin secretion in the dog. Endocrinology 118:1551-1557. Ahren B, Taborsky G J Jr, Porte D Jr (1986b) Neuropeptidergic versus cholinergic and adrenergic regulation of islet hormone secretion. Diabetologia 291:827-836. Ahren B, Veith R C , Taborsky G J Jr (1987) Sympathetic nerve stimulation versus pancreatic norepinephrine infusion in the dog: 1) Effects on basal release of insulin and glucagon. Endocrinology 121:323-331. Albano J D M , Ekins RP, Maritz G , Turner R C (1972) A sensitive, precise radioimmunoassay of serum insulin relying on charcoal separation of bound and free hormone moieties. Acta Endocrinologica 70:487-509. Alumets JR, Hakanson R, O'Dorisio T, Sjolund K, Sundler F (1978) Is GIP a glucagon cell constituent? Histochemistry 58:253-256. Amenta F, Cavallotti C, de Rossi M , Tonelli F, Vatrella F (1983) The cholinergic innervation of human pancreatic islets. Acta Histochemica 73:273-278. Amiranoff B, Couvineau A , Vauclin-Jacques N, Laburthe M (1986) Gastric inhibitory polypeptide receptor in hamster pancreatic beta cells. Direct cross-linking, solubilization and characterization as a glycoprotein. European Journal of Biochemistry 159:353-358. Amiranoff B^ Lorinet A - M , Lagny-Pourmir I, Laburthe M (1988) Mechanism of galanin-inhibited insulin release: occurrence of a pertussis-toxin-sensitive inhibition of adenylate cyclase. European Journal of Biochemistry 177:147-152. Amiranoff B, Lorinet A - M , Yanaihara N, Laburthe M (1989) Structural requirements for galanin action in the pancreatic /3-cell line Rinm5F. European Journal of Pharmacology 163:205-207. 192 Amiranoff B, Servin A L , Rouyer-Fessard C, Couvineau A , Tatemoto K, Laburthe M (1987) Galanin receptors in a hamster pancreatic /?-cell tumour: identification and molecular characterization. Endocrinology 121:284-289. Amiranoff B, Vauclin-Jacques N, Laburthe M (1984) Functional GIP receptors in a hamster pancreatic beta cell line, In III: specific binding and biological effects. Biochemical and Biophysical Research Communications 123:671-676. Amland PF, Jorde R, Burhol P G , Giercksky K - E (1984) Similar plasma GIP responses in obese and lean subjects after an oral test meal and after intraduodenal stimulation with fat and glucose. International Journal of Obesity 8:649-653 Amland PF, Jorde R, Aanderund P, Burhol P G , Giercksky K - E (1985a) Effects of intravenously infused porcine GIP on serum insulin, plasma C-peptide, and pancreatic polypeptide in non-insulin-dependent diabetes in the fasting state. Scandinavian Journal of Gastroenterology 20:315-320. Amland PF, Jorde R, Burhol P G , Giercksky K - E (1985b) Effect of intravenously infused porcine GIP on serum insulin in obese and lean subjects studies with the hyperglycemic clamp technique. Scandinavian Journal of Gastroenterology 20:309-314. Andersen D K , Elahi D, Brown JC, Tobin JD, Andres R (1978) Oral glucose augmentation of insulin secretion: interactions of gastric inhibitory polypeptide with ambient glucose and insulin levels. Journal of Clinical Investigation 62:152-161. Argiles J M (1989) The obese Zucker rat: a choice for fat metabolism. 1968-1988: Twenty years of research on the insights of the Zucker mutation. Progress in Lipid Research 28:53-66. Bagdade JD, Bierman E L , Porte D Jr (1967) The significance of basal insulin levels in the evaluation of the insulin response to glucose in diabetic and nondiabetic subjects. Journal of Clinical Investigation 46:1549-1557. Baetens D , Malaisse-Lagae F, Perrelet A Orci L (1979) Endocrine pancreas: three-dimensional reconstruction shows two types of islets of Langerhans. Science 206:1323-1324. Bailey CJ, Wilkes L C , Conlon J M , Armstrong P H , Buchanan K D (1990) Effects of gastric inhibitory polypeptide, vasoactive intestinal polypeptide and peptide histidine isoleucine on the secretion of hormones by isolated mouse pancreatic islets. Journal of Endocrinology 125:375-379. Balasse E , Ooms H (1973) Role of plasma free fatty acids in the control of insulin secretion in man. Diabetologia 9:145-151. Banting F G , Best C H (1922) The internal secretion of the pancreas. Journal of Laboratory and Clinical Medicine 7:256-271. 193 Banting F G , Best C H , Collip JB, Campbell WR, Fletcher A A (1922) Pancreatic extracts in the treatment of diabetes mellitus. Canadian Medical Association Journal 12:141. Bayliss W M , Starling E H (1902) The mechanism of pancreatic secretion. Journal of Physiology (London) 28:235-334. Beck (1989) Gastric inhibitory polypeptide: a gut hormone with anabolic functions. Journal of Molecular Endocrinology 2:169-174. Beck B, Max JP (1983) Gastric inhibitory polypeptide enhancement of the insulin effect on fatty acid incorporation into adipose tissue in the rat. Regulatory peptides 7:3-8. Beck B, Max JP (1987) Hypersensitivity of adipose tissue to gastric inhibitory polypeptide action in the obese Zucker rat. Cellular and Molecular Biology 33:555-562. Beck B, Max JP (1988) The effects of gastric inhibitory polypeptide and insulin on fatty acid uptake in adipose tissue are not mediated through cyclic A M P in lean and obese Zucker rats. Hormone and Metabolic Research 20:24-27. Bell GI (1986) The glucagon superfamily: precursor structure and gene organization. Peptides 7:27-36. Bell GI, Santerre RF , Mullenbach G T (1983) Hamster preproglucagon contains the sequence of glucagon and two related peptides. Nature 302:716-718. Bergman R N , Miller R E (1973) Direct enhancement of insulin secretion by vagal stimulation of the isolated pancreas. American Journal of Physiology 225:481-486. Bernard C (1849) Chiens rendus diabetiques. Comptes Rendus des Seances de la Societe de Biologie et de ses Filiales (Paris) 1:60. Berson SA, Yalow RS (1959) Immunoassay of endogenous plasma insulin in human subjects by immunologic methods. Nature 184:1648-1649. Berthoud H R (1984) The relative contribution of the nervous system, hormones, and metabolites to the total insulin response during a meal in the rat. Metabolism 33:18-25. Black M A , Fournier L A , Heick H M , Begin-Heick N (1990) Different insulin-secretory responses to calcium-channel blockers in islets of lean and obese (ob/ob) mice. Biochemical Journal 249:401-407. Blonz E R , Stern JS, Curry D L (1985) Dynamics of pancreatic insulin release in young Zucker rats: a heterozygote effect. American Journal of Physiology 248:E188-E193. Bloom SR (1974) Hormones of the gastrointestinal tract. British Medical Bulletin 30:62-67. 194 Bloom SR* Bryant M G , Cochrane UPS (1975) Normal distribution and post prandial release of gut hormones. Clinical Science and Molecular Medicine 49:3. Bloom SR, Edwards A V (1975) The release of pancreatic glucagon and inhibition of insulin in response to stimulation of the sympathetic innervation. Journal of Physiology (London) 253:157-173. Bloom SR, Edwards A V (1981) Pancreatic endocrine responses to stimulation of the peripheral ends of the vagus nerves in conscious calves. Journal of Physiology (London) 315:31-41. Bolaffi JL, Heldt A Lewis L D , Grodsky G M (1986) The third phase of in vitro insulin secretion: evidence for glucose sensitivity. Diabetes 35:370-373. Bolaffi JL, Rodd G , M a Y, Grodsky G M (1990) Effect of glucagon or somatostatin on desensitized somatostatin secretion. Endocrinology 126:1750-1755. Bonner-Weir S, Orci L (1982) New perspectives on the microvasculature of the islets of Langerhans in the rat. Diabetes 37:616-621. Bosco D, Orci L, Meda P (1989) Homologous but not heterologous contact increases insulin secretion of individual pancreatic B-cells. Experimental Cell Research 184:72-80. Boulange A , Planche E , DeGasquet P (1979) Onset of genetic obesity in the absence of hyperphagia during the first week of life in the Zucker rat (fa/fa). Journal of Lipid Research 20:857-864. Bray G A (1969) Studies on the composition of adipose tissue from the genetically obese rats. Proceedings of the Society for Experimental Medicine and Biology 131:1111-1114. Bray G A York D A (1972) Studies on food intake of genetically obese rats. American Journal of Physiology 223:176-179. Brazeau P, Vale W, Burgus R, Ling N, Butcher M , Rivier J, Guillemin R (1973) Hypothalamic peptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science 179:77-79. Britton SW (1925) Studies on the conditions of activities in endocrine glands (XVIII). The nervous control of insulin secretion. American Journal of Physiology 74:291-307. Brown JC (1971) A gastric inhibitory peptide I. The amino acid composition and tryptic peptides. Canadian Journal of Biochemistry 49:255-261. Brown JC (1982) Gastric inhibitory polypeptide. In: Monographs on Endocrinology. Springer-Verlag, Berlin, Germany. Brown JC (1974) Gastric inhibitory polypeptide (GIP). In: Endocrinology 1973. S. Taylor, ed., Heinemann, London, pp. 276-284. 195 Brown JC (1988) Enteroinsular axis. Baillieres Clinical Endocrinology and Metabolism 2(2):359-373. Brown JC, Buchan A M J , Mcintosh CHS, Pederson R A (1989) Gastric inhibitory polypeptide. In: Handbook of Physiology, Section 6, The Gastrointestinal System. Eds. S. G . Shultz, G . M . Makhlouf, B. B. Rauner. American Physiology Society, Bethesda, M D , pp. 403-430. Brown JC, Dryburgh JR (1971) A gastric inhibitory polypeptide. II. The complete amino acid sequence. Canadian Journal of Biochemistry 49:867-872. Brown JC, Koop H , Mcintosh CHS, Otte SC, Pederson R A (1980) Physiology of gastric inhibitory polypeptide. Frontiers in Hormone Research 7:132-144. Brown JC, Magee D F (1967) Inhibitory action of cholecystokinin on acid secretion from Heidenhain pouches induced by endogenous gastrin. Gut 8:29-31. Brown JC, Mutt V, Pederson R A (1970) Further purification of a polypeptide demonstrating enterogastrone activity. Journal of Physiology (London) 209:57-64. Brown JC, Pederson R A (1970) A multiparameter study on the action of preparations containing cholecystokinin-pancreozymin. Scandinavian Journal of Gastroenterology 5:537-541. Brown JC, Pederson R A (1976) GI Hormones and insulin secretion. In: Endocrinology Proceedings of the Fifth International Congress of Endocrinology. Excerpta Medica, Amsterdam, 2:568-570. Brown JC, Pederson R A , Jorpes JE, Mutt V (1969) Preparation of a highly active enterogastrone. Canadian Journal of Physiology and Pharmacology 47:113-114. Brunicardi F C , Druck P, Seymour N E , Sun YS, Elahi D , Andersen D K (1990) Selective neurohormonal influences in islet cell secretion in the isolated perfused human pancreas. Journal of Surgical Research 48:273-278. Brunicardi F C , Elahi D, Andersen D K (1986a) Gastric inhibitory polypeptide activation of beta-cell calcium transport: further evidence of a common mechanism with organs. Canadian Journal of Physiology and Pharmacology (Symposium on Gastrointestinal Hormones) 64:A227. Brunicardi F C , Sun YS, Druck P, Elahi D, Yamada T, Andersen D K (1986b) Regulation of somatostatin release in the isolated perfused human pancreas. Canadian Journal of Physiology and Pharmacology (Symposium on Gastrointestinal Hormones) 64:A166. Bryce G F , Johnson PR, Sullivan A C , Stern JS (1970) Insulin and glucagon: plasma levels and pancreatic release in the genetically Zucker rat. Hormone and Metabolic Research 9:366-370. 196 Buchan A M J , Ingman-Baker J, Levy J, Brown JC (1982) A comparison of the ability of serum and monoclonal antibodies to gastric inhibitory polypeptide to detect immunoreactive cells in the gastroenteropancreatic system of mammals and reptiles. Histochemistry 76:341-349. Buchan A M J , Polak J, Capella C, Solcia E , Pearse A G E (1978) Electron immunocytochemical evidence for the K cell localisation of gastric inhibitory polypeptide (GIP) in man. Histochemistry 56:37-44. Buchan A M J , Sikora L K J , Levy JG, Mcintosh CHS, Dyck I, Brown JC (1985) A n immunocytochemical investigation with monoclonal antibodies to somatostatin. Histochemistry 83:175-180. Burhol P G , Jorde R, Waldun H L (1980) Radio-immunoassay of plasma gastric inhibitory polypeptide (GIP), release of GIP after a test meal and duodenal infusion of bile and immunoreactive plasma GIP components in man. Digestion 20:336-345. Cantor P ^1986) Evaluation of a radioimmunoassay for cholecystokinin in human plasma. Scandinavian Journal of Clinical Laboratory Investigation 46:213-221. Carlquist M , Maletti M , Jornvall H , Mutt V (1984) A novel form of gastric inhibitory polypeptide (GIP) isolated from bovine intestine using a radioreceptor assay. Fragmentation with staphylococcal protease results in GIP 1-3 and GIP 4-42, fragmentation with enterokinase in GIP 1-16 and GIP 17-42. European Journal of Biochemistry 145:573-577. Cataland S, Crockett SE, Brown JC, Mazzaferri E L (1974) Gastric inhibitory polypeptide stimulation by oral glucose in man. Journal of Clinical Endocrinology and Metabolism 39:223-228. Chan C B (1985) On the role of the enteroinsular axis in obesity. Ph.D. Thesis, The University of British, Columbia, Vancouver, Canada. Chan CB, Pederson R A , Buchan A M J , Tubesing K B , Brown JC (1984) Gastric inhibitory polypeptide (GIP) and insulin release in the obese Zucker rat. Diabetes 33:536-542. Chan CP, Stern JS (1982) Adipose lipoprotein lipase in insulin-treated diabetic lean and obese Zucker rats. American Journal of Physiology 242:E445-E450. Chey WY, Lee Y H , Hendricks J G , Rhodes R A , Tai H - H (1978) Plasma secretin concentrations in fasting and postprandial state in man. Digestive Diseases 23:981-988. Childs G V , Naor Z , Hazum E , Tibolt R, Westlund K N , Hancock M B (1983) Localization of biotinylated gonadotropin releasing hormone on pituitary monolayer cells with avidin-biotin-peroxidase complexes. Journal of Histochemistry and Cytochemistry 31:1422-1425. 197 Chow B-C, Morrow G W , H o M , Pederson R A , Mcintosh CHS, Brown JC, MacGillivray R T A (1990) Expression of recombinant human glucose-dependent insulinotropic polypeptide in Eschirichia coli by sequence-specific proteolysis of a protein A fusion protein. Peptides 11:1069-1074. Cleator IG, Gourlay R H (1975) Release of immunoreactive gastric inhibitory polypeptide (IR-GIP) by oral ingestion of food substances. American Journal of Surgery 130:128-135. Cook D L , Hales C N (1984) Intracellular A T P directly blocks K + channels in pancreatic B-cells. Nature 311:271-273. Cooper GJS, Willis A C , Clark A Turner R C , Sim RB, Reid K B M (1987) Purification and characterization of a peptide from amyloid-rich pancreases of type 2 diabetic patients. Proceedings of the National Academy of Science (U.S.A.) 84:8628-8632. Crespin SR, Greenough W, Steinberg D (1969) Stimulation of insulin secretion by infusion of free fatty acids. Journal of Clinical Investigation 48:1934-1943. Creutzfeldt W (1979) The incretin concept today. Diabetologia 16:75-85. Creutzfeldt W, Ebert R (1977) Release of gastric inhibitory polypeptide (GIP) under normal and pathological conditions in man. In: Bajaj JS (ed) Diabetes. Excerpta Medica, Amsterdam, pp. 63-75. Creutzfeldt W, Ebert R (1988) The incretin concept. Advances in Metabolic Disorders 11:333-367. Creutzfeldt W, Ebert R, Willms B, Frerichs H , Brown JC (1978) Gastric inhibitory polypeptide (GIP) and insulin in obesity: increased response to stimulation and defective feedback control of serum levels. Diabetologia 14:15-24. Creutzfeldt W, Feurle G , Ketterer H (1970) Effect of gastrointestinal hormones on insulin and glucagon secretion. New England Journal of Medicine 282:1139-1141. Curry D L (1986) Insulin content and insulinogenesis by the perfused rat pancreas: effect of long term glucose stimulation. Endocrinology 118:170-175. Curry D L , Bennett L L , Grodsky G M (1968) Dynamics of insulin secretion by the perfused rat pancreas. Endocrinology 83:572-584. Curry D L , Stern JS (1985) Dynamics of insulin hypersecretion by obese Zucker rats. Metabolism 34:791-796. Cushman SW, Zarnowski MJ, Franzusoff AJ , Salans L B (1978) Alterations in glucose metabolism and its stimulation by insulin in isolated adipose cells during the development of genetic obesity in the Zucker fatty rat. Metabolism 27:1930-1940. 198 Dahl M A (1983) O n the mechanism of action of glucose-dependent-insulinotropic polypeptide. Ph.D. Thesis, The University of British, Columbia, Vancouver, Canada. D'Ambra R, Surana M , Efrat S, Starr R G , Fleischer N (1990) Regulation of insulin secretion from 0-cell lines derived from transgenic mice insulinomas resembles that of normal 0-cells. Endocrinology 126:2815-2822. Daniel P M , Henderson JR (1967) The effect of vagal stimulation on plasma insulin and glucose levels in the baboon. Journal of Physiology (London) 192:317-326. Dunning B E , Ahren B, Bottcher G , Sundler F, Taborsky G J Jr (1987) The presence and actions of N P Y in the canine endocrine pancreas. Regulatory Peptides 18:253-265. Dunning B E , Ahren B, Veith R C , Bottcher G , Sundler F, Taborsky G J Jr (1986) Galanin: a novel pancreatic neuropeptide. American Journal of Physiology 251:E127-E133. Dunning B E , Ahren B, Veith R C , Taborsky G J Jr (1988) Nonadrenergic sympathetic neural influences on basal pancreatic hormone secretion. American Journal of Physiology 255:E1785-E792. Dunning B E , Taborsky G J Jr (1988) Galanin - sympathetic neurotransmitter in the endocrine pancreas? Diabetes 37:1157-1162. Dunning B E , Taborsky G J Jr (1989) Galanin release during pancreatic nerve stimulation is sufficient to influence islet function. American Journal of Physiology 256:E191-E198. Dupre J (1964) A n intestinal hormone affecting glucose disposal in man. Lancet ii:672-673. Dupre J, Beck JC (1966) Stimulation of release of insulin by an extract of intestinal mucosa. Diabetes 15:555-559. Dupre J, Curtis JD, linger R H , Waddell RW, Beck JC (1969) Effects of secretin, pancreozymin or gastrin on the response of the endocrine pancreas to administration of glucose or arginine in man. Journal of Clinical Investigation 48:748-757. Dupre J, Greenidge N, McDonald TJ, Ross S A Rubinstein D (1976) Inhibition of action of glucagon in adipocytes by gastric inhibitory polypeptide. Metabolism 25:1197-1199. Dupre J, Ross SA, Watson D , Brown JC (1973) Stimulation of insulin secretion by gastric inhibitory polypeptide in man. Journal of Clinical Endocrinology and Metabolism 37:826-828. Eaton RP, Conway M , Schade DS (1976) Endogenous glucagon regulation in genetically hyperlipemic obese rats. American Journal of Physiology 230:1336-1341. Ebert R, Creutzfeldt W (1982) Influence of gastric inhibitory polypeptide antiserum on glucose-induced insulin secretionin rats. Endocrinology 111:1601-1606. 199 Ebert R, Illmer K, Creutzfeldt W (1979) Release of gastric inhibitory polypeptide (GIP) by intraduodenal acidification in rats and humans and abolishment of the incretin effect of acid by GIP infusion in rats. Gastroenterology 76:515-523. Eckel R H , Fujimoto WY, Brunzell J D (1979) Gastric inhibitory polypeptide enhances lipoprotein lipase activity in cultured preadipocytes. Diabetes 28:1141-1142. Edwards A V (1984) Neural control of the endocrine pancreas. In: 10 Recent Advances in Physiology. P.F. Baker, Ed., Pitman Press, Bath, England, pp.277-315. Elahi D , Andersen D K , Brown JC, Debas H T , Hershcopf RJ, Raizes GS, Tobin JD, Andres R (1979) Pancreatic a- and /3-cell responses to GIP infusion in normal man. American Journal of Physiology E185-E191. Elahi D , Andersen D K , Muller D C , Tobin JD, Brown JC, Andres R (1984) The enteric enhancement of glucose-stimulated insulin release. The role of GIP in aging, obesity, and non-insulin-dependent diabetes mellitus. Diabetes 33:950-957. Elahi D, Raizes GS, Andres R, Hershcopf RJ, Muller D C , Tobin JD, Andersen D K (1982) Interaction of arginine and gastric inhibitory polypeptide on insulin release in man. American Journal of Physiology 242:E343-E351. El-Munshid H A , Hakanson R, Lindberg G , Sundler F (1980) Effects of various gastrointestinal peptides on parietal cells and endocrine cells in the oxyntic mucosa of rat stomach. Journal of Physiology (London) 305:249-265. Elrick M L , Stimmler L, Hlad CJ, Arai Y (1964) Plasma insulin response to oral and intravenous glucose administration. Journal of Clinical Endocrinology 24:1076-1082. Elwin C E (1974) Gastric acid responses to antral application of some amino acids, peptides, and isolated fractions of a protein hydrosylate. Scandinavian Journal of Gastroenterology 9:239-247. Fahrenkrug J, Schaffalitzky de Muckadell OB, Kiihl C (1978) Effect of secretin on basal-and glucose-stimulated insulin release in man. Diabetologia 14:229-234. Falko J M , Crockett SE, Cataland S, Mazzaferri E L (1975) Gastric inhibitory polypeptide (GIP) stimulated by fat ingestion in man. Journal of Clinical Endocrinology and Metabolism 41:260-265. Fehmann H C , Goke B, Goke R, Trautmann M E , Arnold R (1989) Synergisitc stimulatory effect of glucagon-like peptide-1 (7-36)amide and glucose-dependent insulin-releasing polypeptide on the endocrine rat pancreas. FEBS Letters 252:109-112. Feldman M , Walsh JH, Wong H C , Richardson C T (1978) Role of gastrin heptadecapeptide in the acid secretory response to amino acids in man. Journal of Clinical Investigation 61:308-313. 200 Floyd J C Jr, Fajans SS, Conn JW, Knopf RF , Rull J (1966) Stimulation of insulin secretion by amino acids. Journal of Clinical Investigation 45:1487-1502. Floyd JC Jr, Fajans SS, Knopf RF, Pek S, Thiffault C A , Conn JW (1970) Synergistic effect of essential ammo acids and glucose upon insulin secretion in man. Diabetes 19:109-115. Frohman L A , Goldman JK, Schnatz JD, Bernardis L L (1967) Effects of vagotomy and vagal stimulation on insulin secretion. Diabetes 16:443-448. Fujimoto W Y (1981) Effect of gut peptides on glucose-stimulated insulin release by monolayer cultures of neonatal rat islet cells. Hormone and Metabolic Research 13:135-138. Fujimoto WY, Ensinck JW, Merchant FW, Williams R H , Smith P H , Johnson D G (1978) Stimulation by gastric inhibitory polypeptide of insulin and glucagon secretion by rat islet cultures. Proceedings of the Society for Experimental Biology and Medicine 157:89-93. Garcia M C , Hermans MP, Henquin JC (1988) Glucose-, calcium- and concentration-dependence of acetylcholine stimulation of insulin release and ionic fluxes in mouse islets. Biochemical Journal 254:211-218. Garcia-Morales P, Dufrane SP, Sener A , Valverde I, Malaisse WJ (1984) Inhibitory effect of clonidine upon adenylate cyclase activity, cyclic A M P production, and insulin release in rat pancreatic islets. Bioscience Reports 4:511-521. Garry DJ, Garry M G , Williams J A Mahoney WC, Sorenson R L (1989) Effects of islet hormones on amylase secretion and localization of somatostatin binding sites. American Journal of Physiology 256:G897-G904. Gerich JE, Charles M A Grodsky G M (1974) Characterization of the effects of arginine and glucose on glucagon and insulin release from the in vitro perfused rat pancreas. Journal of Clinical Investigation 54:833-841. Gerich JE, Charles M A , Grodsky G M (1976) Regulation of pancreatic insulin and glucagon secretion. Annual Review of Physiology 38:353-388. Girardier L, Seydouyx J, Campfield L A (1976) Control of a and B cells in vivo by sympathetic nervous input and selective hyper- and hypo-glycaemia. Journal of Physiology (Paris) 72:801-804. Gillespie IE, Grossman MI (1964) Inhibitory effect of secretin and cholecystokinin on Heidenhain pouch responses to gastrin extract and histamine. Gut 5:342-345. Godfrey D A , Matschinsky F M (1975) Enzymes of the cholinergic system in islets of Langerhans. Journal of Histochemistry and Cytochemistry 23:645-651. 201 Gregor M , Riecken E - O (1985) Production and characterization of N-terminally and C-terminally directed monoclonal antibodies against pancreatic glucagon. Gastroenterology 89:571-580. Grill V , Ostenson C - G (1983) Muscarinic receptors in pancreatic islets of the rat. Demonstration and dependence on long-term glucose environment. Biochimica et Biophysica Acta 756:159-162. Grodsky G M (1972) Threshold distribution hypothesis for packet storage of insulin and its mathematical modelling. Journal of Clinical Investigation 51:2047-2059. Grodsky G M (1989) A new phase of insulin secretion. How will it contribute to our understanding of /3-cell function? Diabetes 38:673-678. Grodsky G M , Batts A A , Bennett L L , Vcella C, McWilliams NB, Smith D F (1963) Effects of carbohydrates on secretion of insulin from isolated rat pancreas. American Journal of Physiology 205:638-644. Haist R E (1965) Effects of changes in stimulation on the structure and function of islet cells. In: O n the Nature and Treatment of Diabetes. Eds. B. S. Liebel, G . A . Wrenshall. Excerpta Medica, Amsterdam, pp. 12-30. Havel PJ, Taborsky G J Jr (1989) The contribution of the autonomic nervous system to changes of glucagon and insulin secretion during hypoglycemic stress. Endocrine Reviews 10:322-350. Hayek A (1980) Insulin release in long-term culture from isolated islets of obese and lean Zucker rats. Hormone and Metabolic Research 12:85-86. Hayek A , Woodside W (1979) Correlation between morphology and function in isolated islets of the Zucker rat. Diabetes 28:565-569. Hazum E (1986) Preparation and use of biotinylated neuroendocrine peptides. Methods in Enzymology 124:47-57. Hermansen K (1980) Secretion of somatostatin from the normal and diabetic pancreas. Studies in vitro. Diabetologia 19:388-392. Hermansen K (1984) Effects of cholecystokinin (CCK)-4, non-sulfated CCK-8, and sulfated CCK-8 on pancreatic somatostatin, insulin and glucagon secretion in the dog: studies in vitro. Endocrinology 114:1770-1775. Hermansen K, Yanaihara N, Ahren B (1989) On the nature of the galanin action on the endocrine pancreas: studies with six galanin fragments in the perfused dog pancreas. Acta Endocrinologica (Copenhagen) 121:545-550. 202 Hildebrand P, Ensinck JW, Ketterer S, Delco F, Mossi S, Bangerter, Beglinger C (1991) Effect of a cholecystokinin antagonist on meal-stimulated insulin and pancreatic polypeptide release in humans. Journal of Clinical Endocrinology and Metabolism 72:1123-1129. Hiriart M , Ramirez-Medeles M C (1991) Functional subpopulations of indiviual pancreatic B-cells in culture. Endocrinology 128:3193-3198. Hoenig L C , MacGregor L C , Matschinsky P M (1986) In vitro exhaustion of pancreatic B cells. American Journal of Physiology 250:E502-E511. Hoist JJ, Gronholt R, Schaffalitzky de Muckadell OB, Fahrenkrug J (1981a) Nervous control of pancreatic endocrine secretion in pigs. I. Insulin and glucagon responses to electrical stimulation of the vagus nerves. Acta Physiologica Scandinavia 111:1-7. Hoist JJ, Gronholt R, Schaffalitzky de Muckadell OB, Fahrenkrug J (1981b) Nervous control of pancreatic endocrine secretion in pigs. II. The effect of pharmacological blocking agents on the response to vagal stimulation. Acta Physiologica Scandinavia 111:9-14. Hoist JJ, Gronholt R, Schaffalitzky de Muckadell OB, Fahrenkrug J (1981c) Nervous control of pancreatic endocrine secretion in pigs. V . Influence of the sympathetic nervous system on the pancreatic secretion of insulin and glucagon, and on the insulin and glucagon response to vagal stimulation. Acta Physiologica Scandinavia 113:278-283. Hoist JJ, Lindkaer Jensen S, Schaffalitzky de Muckadell OB, Fahrenkrug J (1980) Secretin and vasoactive intestinal polypeptide in the control of the endocrine pancreas. Frontiers in Hormone Research 7:119-131. Hoist JJ, Lindkaer Jensen S, Knuhtsen S, Nielsen O V (1983) Autonomic nervous control of pancreatic somatostatin secretion. American Journal of Physiology 245:E542-E548. Hoist JJ, Orskov C, Vagn Nielsen O, Schwartz T (1987) Truncated glucagon-like peptide I, an insulin-releasing hormone from the distal gut. FEBS Letters 211:169-174. Hoist JJ, Schaffalitzky de Muckadell OB, Fahrenkrug J, Lindkaer S, Nielsen O V , Schwartz T W (198 Id) Nervous control of pancreatic endocrine secretion in pigs. III. The effect of acetylcholine on the pancreatic secretion of insulin and glucagon. Acta Physiologica Scandinavia 111: 15-22. Inagaki N, Seino Y, Takeda J, Yano H , Yamada Y, Bell GI, Eddy R L , Fukushima Y, Byers M G , Shows TB, Imura H (1989) Gastric inhibitory polypeptide: structure and chromosomal localization of the human gene. Molecular Endocrinology 3:1014-1021. Ipp E , Dobbs R E , Arimura A Vale W, Harris V, Unger R H (1977) Release of immunoreactive somatostatin from the pancreas in response to glucose, amino-acids, pancreozymin-cholecystokinin, and tolbutamide. Journal of Clinical Investigation 60:760-765. 203 Ipp E , Dobbs R E , Arimura A , Vale W, Unger R H (1977) The effects of gastrin, gastric inhibitory polypeptide, secretin, and the octapeptide of cholecystokinin upon immunoreactive somatostatin release by the perfused canine pancreas. Journal of Clinical Investigation 60:1216-1219. Iversen J (1973) Effect of acetylcholine on the secretion of glucagon and insulin from the isolated, perfused canine pancreas. Diabetes 22:381-387. Ivy A C , Oldberg E (1927) Contraction and evacuation of the gall-bladder caused by highly purified "secretin" preparation. Proceedings of the Society for Experimental Biology and Medicine 25:113-115. Jansen J B M , Lamers C B H W (1983) Radioimmunoassay of cholecystokinin: production and evaluation of antibodies. Journal of Clinical Chemistry and Clinical Biochemistry 21:387-394. Johnson PR, Stern JS, Greenwood M R C Hirsch J (1978) Adipose tissue hyperplasia and hyperinsulinemia in Zucker obese female rats: a developmental study. Metabolism 27:1941-1954. Jones IR, Owens D R , Luzio S, Williams S, Hayes T M (1989) The glucose-dependent insulinotropic polypeptide response to oral glucose and mixed meals is increased in patients with Type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia 32:668^677. Jones IR, Owens D R , Moody AJ , Luzio SD, Morris T, Hayes T M (1987) The effects of glucose-dependent insulinotropic polypeptide infused at physiological concentrations in normal subjects and type 2 (non-insuhn-dependent) diabetic patients on glucose tolerance and B-cell secretion. Diabetologia 30:707-712. Jorde R, Amland PF, Burhol P G , Giercksy K - E , Ebert R (1983a) GIP and insulin responses to a test meal in healthy and obese subjects. Scandinavian Journal of Gastroenterology 18:1115-1119. Jorde R, Burhol P G , Schulz T B (1983b) Fasting and postprandial plasma GIP values in man measured with seven different antisera. Regulatory Peptides 7:87-94. Jorde R, Burhol P G , Waldum H L , Schulz TB, Lygren I, Florholmen J (1980) Diurnal variation of plasma gastric inhibitory polypeptide in man. Scandinavian Journal of Gastroenterology 15:617-619. Jorgensen K H , Larsen U D (1980) Homogeneous mono-1 2 5I-insulins. Preparation and characterization of mono- 1 2 5I-(Tyr A19)-insulin. Diabetologia 19:546-554. Jornvall H , Carlquist M , Kwauk S, Otte SC, Mcintosh CHS, Brown JC, Mutt V (1981) Amino acid sequence and heterogeneity of gastric inhibitory polypeptide (GIP). F E B S Letters 123:205-210. 2 0 4 Kaneto A , Kosaka K (1974) Stimulation of insulin and glucagon secretion by acetyl choline infused intrapancreatically. Endocrinology 95:676-681. Kaneto A , Kosaka K, Nakao K (1967) Effects of stimulation of the vagus nerves on insulin secretion. Endocrinology 80:530-536. Kaneto A , Miki E , Kosaka K (1974) Effects of vagal stimulation on glucagon and insulin secretion. Endocrinology 95:1005-1010. Karam J H , Grodsky G M , Forsham P H (1963) Excessive insulin response to glucose in obese subjects as measured by immunochemical assay. Diabetes 12:197-204. Kofod H , Hansen B, Lernmark A, Hedeskov CJ (1986) Secretin and its C-terminal hexapeptide potentiates insulin release in mouse islets. American Journal of Physiology 250:E107-E113. Komatsu R, Matsuyama T, Namba M , Watanabe N, Itoh H , Kono N, Tarui S (1989) Glucagonostatic and insulinotropic action of glucagonlike peptide l-(7-36)-amide. Diabetes 38:902-905. Kosaka T, Lim R K S (1930) Demonstration of the humoral agent in fat inhibition of gastric secretion. Proceedings of the Society of Experimental Biology and Medicine 27:890-891. Krarup T (1988) Immunoreactive gastric inhibitory polypeptide. Endocrine Reviews 9:122-134. Krarup T, Madsbad S, Moody A J (1985) Immunoreactive gastric inhibitory polypeptide response to a meal during the first eighteen months after diagnosis of type 1 (insulin dependent) diabetes mellitus. Journal of Clinical Endocrinology and Metabolism 60:120-125. Krarup T, Saurbrey N, Moody AJ , Kiihl C, Madsbad S (1987) Effect of porcine gastric inhibitory polypeptide on /S-cell function in type I and type II diabetes mellitus. Metabolism 36:677-682. Kreymann B, Ghatei M A , Williams G , Bloom SR (1987) Glucagon-like peptide-1 7-36: a physiological incretin in man. Lancet 2:1300-1303. Kuffert A , Stern JS, Curry D L (1988) Pancreatic hypersensitivity to glucose by young obese Zucker rats (fa/fa). Metabolism 37:952-957. Kurose T, Seino Y, Nishi S, Tsuji K, Taminato T, Tsuda K, Imura H (1990) Mechanism of sympathetic neural regulation of insulin, somatostatin, and glucagon secretion. American Journal of Physiology 258:E220-E227. Kuzio M , Dryburgh JR, Malloy K M , Brown JC (1974) Radioimmunoassay for gastric inhibitory polypeptide. Gastroenterology 66:357-364. 2 0 5 Kwok Y N , Verchere CB, Mcintosh CHS, Brown JC (1988) Effect of galanin on endocrine secretions from the isolated perfused rat stomach and pancreas. European Journal of Pharmacology 145:49-54. Labarre J (1927) Sur 1'augmentation de la teneur en insuline du sang veineux pancreatique apres excitation du nerf vague. Comptes Rendus des Seances de la Societ de Biologie et de ses Filiales (Paris) 96:193-196. Labarre J (1932) Sur les possibilities d'un traitement du diabete par l'incretine. Bulletin Academie Royale Society Medicine Belgium 12:620-634. Lacey RJ, Berrow NS, London J M , Lake SP, James RF , Scarpello J H B , Morgan N G (1990) Differential effects of B-adrenergic agonists on insulin secretion from pancreatic islets isolated from rat and man. Journal of Molecular Endocrinology 5:49-54. Lambert D T , Stachelek C, Varga J M , Lemer A M (1982) Iodination of melanotropin. Journal of Biological Chemistry 82111-82115. Langerhans P (1869) Contributions to the microscopic anatomy of the pancreas. M D Thesis, Berlin. (Translated by H . Morrison; The Johns Hopkins Hospital Press, Baltimore, M D , 1937) Lardinois C K , Richeson R B III, Starich G H , Mazzu D, Mazzaferri E L (1990) Gastric inhibitory polypeptide mechanisms of augmenting insulin secretion. Life Sciences 47:1015-1022. Larsson LI, Boder G B , Shaw W N (1977) Changes in the islets of Langerhans in the obese Zucker rat. Laboratory Investigation 36:593-598. Lauritsen K B , Moody A J (1978) The response of gastric inhibitory polypeptide (GIP) and insulin to glucose in duodenal ulcer patients. Diabetologia 14:149-153. Leclerq-Meyer V, Brissun G R , Malaisse WJ (1971) Effect of adrenaline and glucose on release of glucagon and insulin in vitro. Nature (New Biol) 231:248-249. Lerner R L , Porte D Jr (1970) Uniphasic insulin responses to secretin stimulation in man. Journal of Clinical Investigation 49:2276-2280. Liddle R A , Goldfine ID, Rosen MS, Taplitz R A , Williams J A (1985) Cholecystokinin bioactivity in human plasma. Molecular forms, responses to feeding, and relationship to gallbladder contraction. Journal of Clinical Investigation 75:1144-1152. Lindkaer Jensen S, Fahrenkrug J, Hoist JJ, Kiihl C, Vagn Nielsen O, Schaffalitzky de Muckadell O B (1978) Secretory effects of secretin in isolated perfused porcine pancreas. American Journal of Physiology 235:E381-E386. Lindskog S and Ahren B (1987) Galanin: effects on basal and stimulated insulin and glucagon secretion in the mouse. Acta Physiologica Scandinavia 129:305-309. 206 Lindskog S and Ahren B (1988) Galanin and pancreastatin inhibit stimulated insulin secretion in the mouse: comparison of effects. Hormone Research 29:237-240. Loew E R , Gray JS, Ivy A C (1940) Is a duodenal hormone involved in carbohydrate metabolism? American Journal of Physiology 129:659-663. Lopez L C , Frazier M L , Su CJ, Humar A , Sanders G F (1983) Mammalian pancreatic preproglucagon contains three glucagon-related peptides. Proceedings of the National Academy of Sciences (U.S.A.) 80:5485-5489. Loubatieres-Mariani M M , Chapal J, Alric R, Loubatieres A (1973) Studies of the cholinergic receptors involved in the secretion of insulin using isolated perfused rat pancreas. Diabetologia 9:439-446. Lukinius A , Wilander E , Westermark G T , Engstrom U , Westermark P (1989) Colocalization of islet amyloid polypeptide (LAPP) and insulin in the j3-cell secretory granules of human pancreatic islets. Diabetologia 32:240-244. Luzi L, DeFronzo R A (1989) Effect of loss of first phase insulin secretion on hepatic glucose production and tissue glucose disposal in man. American Journal of Physiology 275:E241-E246. MacDonald M J (1990) Elusive proximal signals for insulin secretion. Diabetes 39:1461-1466. MacDonald MJ, Gapinski JP (1989) A rapid E L I S A for measuring insulin in a large number of research samples. Metabolism 38:450-452. Madison L L , Seyffert W A Jr, Unger R H , Baker B (1968) Effect of plasma free fatty acids on plasma glucagon and serum insulin concentrations. Metabolism 17:301-304. Magee D F , Nakamura M (1966) Action of pancreozymin preparations on gastric secretion. Nature 212:1487-1488. Malaisse WJ, Herchuelz A , Levy J, Sener A , Pipeleers D G , Devis G , Somers G , Van Obberghen E (1976) The stimulus-secretion coupling of glucose-induced insulin release. XIX. The insulinotropic effect of glyceraldehyde. Molecular and Cellular Endocrinology 4:1-12. Malaisse WJ, Malaisse-Lagae F, Wright P H , Asnore J (1967) Effects of adrenergic and cholinergic agents upon insulin secretion in vitro. Endocrinology 80:975-978. Malaisse WJ, Sener A , Koser M , Ravazzalo M , Malaisse-Lagae F (1977) The stimulus-secretion coupling of glucose-induced insulin release. Insulin release due to glycogenolysis in glucose-deprived islets. Biochemical Journal 104:447-454. Malaisse WJ, Sener A , Herchuelz A , Jutton JC (1979) Insulin release: the fuel hypothesis. Metabolism 28:373-386. 207 Maletti M , Altaian J-J, Hui Bon Hoa D, Carlquist M , Rosselin G (1987) Evidence of functional gastric inhibitory polypeptide (GIP) receptors in human insulinoma. Binding of synthetic human GIP 1-31 and activation of adenylate cyclase. Diabetes 36:1336-1340. Maletti M , Carlquist M , Portha B, Kergoat M , Mutt V , Rosselin G (1986) Structural requirements for gastric inhibitory polypeptide (GIP) receptor binding and stimulation of insulin release. Peptides 7(suppl l):75-78. Maletti M , Portha B, Carlquist M , Kergoat M , Laburthe M , Marie JC, Rosselin G (1984) Evidence for and characterization of specific high affinity sites for gastric inhibitory polypeptide in pancreatic /3-cells. Endocrinology 115:1324-1331. Manaka H , Taniguchi H , Katshuhiko W, Takahashi H , Takahashi K, Katagiri T, Sasaki H (1987) Glucagon-like peptide-I in the rat pancreas. Biomedical Research 8:1-8. Martin RJ, Harris RBS, Jones D D (1986) Evidence for a central mechanism of obesity in the Zucker fatty rat (fa/fa). Proceedings of the Society for Experimental Biology and Medicine 183:1-10. Maxwell V, Shulkes A , Brown JC, Solomon T E , Walsh JH, Grossman MI (1980) Effect of gastric inhibitory polypeptide on pentagastrin-stimulated acid secretion in man. Digestive Diseases and Sciences 25:113-116. Mazzaferri E L , Starich G H , Lardinois C K , Bowen G D (1985) Gastric inhibitory polypeptide responses to nutrients in Caucasians and American indians with obesity and noninsulin-dependent diabetes mellitus. Journal of Clinical Endocrinology and Metabolism 61:313-321. Mazzu D, Starich G H , Lardinois C K (1989) Intracellular phosphate quantitation in rat primary beta cell suspensions. The Endocrine Society (71st Annual Meeting) :A145. McCullough A J , Marshall JB, Bingham CP, Rice BL, Manning L D , Kalhan SC (1985) Carbachol modulates GIP-stimulated insulin release from rat pancreatic lobules in vitro. American Journal of Physiology 248:E299-E303. Mcintosh C H S (1985) Gastrointestinal somatostatin: distribution, secretion and physiological significance. Life Sciences 37:2043-2058. Mcintosh CHS, Arnold R, Bothe E , Becker M , Kobberling J, Creutzfeldt W (1978) Gastrointestinal somatostatin: extraction and radioimunoassay in different species. Gut 19:655-633. Mcintosh CHS, Kwok Y N , Tang C, Brown JC (1987) The use of monoclonal antibodies in radioimmunoassays for gastrointestinal hormones. Journal of Clinical Immunoassay 10:79-84. 208 Mcintosh CHS, Pederson R A , Koop H , Brown JC (1981) Gastric inhibitory polypeptide stimulated secretion of somatostatin like immunoreactivity from the stomach: inhibition by acetylcholine or vagal stimulation. Canadian Journal of Physiology and Pharmacology 59:468-472. Mclntyre N, Holdsworth C D , Turner DS (1964) New interpretation of oral glucose tolerance. Lancet ii:20-21. Meda P, Perrelet A , Orci L (1984) Gap junctions and cell-to-cell coupling in endocrine glands. Modern Cell Biology 3:131-196. Messell T, Harling H , Bottcher G , Johnsen A H , Hoist JJ (1990) Galanin in the porcine pancreas. Regulatory Peptides 28:161-176. Miller R E (1975) Neural inhibition of insulin release from the isolated canine pancreas. American Journal of Physiology 228:114-119. Miller R E (1981) Pancreatic neuroendocrinology: peripheral neural mechanisms in the regulation of the islets of Langerhans. Endocrine Reviews 2:471-494. Miralles P, Peiro E , Silvestre R A , Villanueva M L , Marco J (1988) Effects of galanin on islet cell secretory responses to VIP, GIP, 8-CCK, and glucagon by the perfused rat pancreas. Metabolism 37:766-770. Miralles P, Peiro E , Degano P, Silvestre R A , Marco J (1990) Inhibition of insulin and somatostatin secretion and stimulation of glucagon release by homologous galanin in perfused rat pancreas. Diabetes 39:996-1001. Moltz JH, McDonald JK (1985) Neuropeptide Y : Direct and indirect action on insulin secretion in the rat. Peptides 6:1155-1159. Moody A J , Thim L, Valverde I (1984) The isolation and sequencing of human gastric inhibitory polypeptide (GIP). F E B S Letters 172:142-148. Moore B, Edie ES, Abram J M (1906) On the treatment of diabetes mellitus by acid extract of duodenal mucous membrane. Biochemical Journal 1:28-38. Morgan (1979) Immunoassay of gastric inhibitory polypeptide: investigations into its role in carbohydrate metabolism. Annals of Clinical Biochemistry 16:6-14. Morgan L M , Morris BA, Marks V (1978) Radioimmunoassay of gastric inhibitory polypeptide. Annals of Clinical Biochemistry 15:172-177. Mosjov S, Heinrich G , Wilson IB, Ravazzola M , Orci L, Habener JF (1986) Preproglucagon gene expression in pancreas and intestine diversifies at the level of post -translational processing. Journal of Biological Chemistry 261:11880-11889. 209 Mosjov S, Weir G C , Habener JF (1987) Insulinotropin: glucagon-like peptide (7-37) co-encoded in the glucagon gene is a potent stimulator of insulin release in the perfused rat pancreas. Journal of Climcal Investigation 79:616-619. Muller M K , Pederson R A , Brown JC (1982) Interaction of gastric inhibitory polypeptide and arginine in the perfused pancreas of rats treated with triiodothyronine. Canadian Journal of Physiology and Pharmacology 60:297-301. Muller M K , Scheck T, Goebell H (1986) Interaction of acetylcholine and gastric inhibitory polypeptide on endocrine and exocrine rat pancreatic secretion: augmentation of acetylcholine-induced amylase and volume secretion by the insulinotropic action of gastric inhibitory polypeptide. Digestion 33:45-52. Nakaki T, Nakadate T, Ishii K, Kato R (1980) Postsynaptic alpha-2 adrenergic receptors in isolated rat islets of Langerhans: inhibition of insulin release and cyclic 3':5'-adenosine monophosphate accumulation. Journal of Pharmacology and Experimental Therapeutics 216:607-612. Nauck M , Stockmann F, Ebert R, Creutzfeldt W (1986) Reduced incretin effect in Type 2 (non-insulin-dependent) diabetes. Diabetologia 29:46-52. Nielsen D A , Lernmark A , Berelowitz M , Bloom G D , Steiner D F (1982) Sorting of pancreatic islet cell subpopulations by light scattering using a fluorescence-activated cell sorter. Diabetes 31:299-306. Niijima A (1989) Nervous regulation of metabolism. Progress in Neurobiology 33:135-147. Nishi S, Seino Y, Ishida H , Seno M , Taminato T, Sakurai H , Imura H (1987) Vagal regulation of insulin, glucagon, and somatostatin secretion in vitro in the rat. Journal of Clinical Investigation 79:1191-1196. O'Connor M D L , Landahl H , Grodsky G M (1980) Comparison of storage- and signal-limited models of pancreatic insulin secretion. American Journal of Physiology 238:R378-389. Ohneda A , Aguilar-Pavada E , Eisentraut A Unger R H (1969) Control of pancreatic glucagon secretion by glucose. Diabetes 18:1-10. Okabayashi Y, Otsuki M , Baba S (1989) Cholecystokinin in the entero-insular axis. Diabetes Research and Clinical Practise 7:S79-S85. Orci (1982) Macro- and micro-domains in the endocrine pancreas. Diabetes 31:538-565. Orci L, Unger R H (1975) Functional subdivision of islets of Langerhans and possible role of D-cells. Lancet 2:1243. 210 Orskov C, Hoist JJ, Nielsen O V (1988) Effect of truncated glucagon-like peptide-1 [proglucagon-(78-107) amide] on endocrine secretion from pig pancreas, antrum, and nonantral stomach. Endocrinology 123:2009-2013. Ostenson C - G , Efendic S, Grill V (1990) Abnormal regulation by glucose of somatostatin secretion in the perfused pancreas of N I D D M rats. Pancreas 5:347-353. Otsuki M , Sakamoto C, Yuu H , Maeda M , Morita S, Ohki A , Kobayashi N, Terashi K, Okano K, Baba S (1979) Discrepancies between the doses of cholecystokinin- or caerulin-stimulating exocrine and endocrine responses in isolated perfused rat pancreas. Journal of Clinical Investigation 63:478-484. Palmer JP, Werner PL, Hollander P, Ensinck JW (1979) Evaluation of the control of glucagon secretion by the parasympathetic nervous system in man. Metabolism 28:549-552. Patel Y (1984) Role of the sympathetic nervous sytem in glucagon response to insulin hypoglycemia in normal and diabetic rats. Diabetes 33:1154-1159. Patel Y C , Wheatley T, Ning C (1981) Multiple forms of immunoreactive somatostatin: comparison of distribution in neural and non-neural tissues and portal plasma of the rat. Endocrinology 109:1943-1949. Patton GS, Ipp E , Dobbs R E , Vale W, Orci L, Unger R H (1976) Response of pancreatic immunoreactive somatostatin to arginine. Life Sciences 19:1957-1960. Pederson R A , Brown JC (1976) The insulinotropic action of gastric inhibitory polypeptide in the perfused isolated rat pancreas. Endocrinology 99:780-785. Pederson R A , Brown JC (1978) Interactions of gastric inhibitory polypeptide, glucose and arginine on insulin and glucagon secretion from the perfused rat pancreas. Endocrinology 102:610-615. Pederson R A , Brown JC (1979) Effect of cholecystokinin, secretin and GIP on insulin release from the isolated perfused rat pancreas. Canadian Journal of Physiology and Pharmacology 57:1233-1237. Pederson R A , Buchan A M J , Zahedi-Asl S, Chan CB, Brown JC (1982) Effect of jejunoileal bypass in the rat on the enteroinsular axis. Regulatory Peptides 5:53-63. Pederson R A Innis SM, Buchan A M J , Chan CB, Brown JC (1985) The effect of total parenteral nutrition (TPN) on the enteroinsular axis in the rat. Regulatory Peptides 10:199-206. Pederson R A , Mochizuki T, Yanaihara C, Yanaihara N, Brown JC (1990) Reduced somatostatinotropic effect of a GIP fragment (pGIP[l-30]-NH2) with insulinotropic activity. Digestion 46(suppl 1):A238. 211 Pederson R A , Schubert H E , Brown JC (1975) Gastric inhibitory polypeptide. Its physiologic release and insulinotropic action in the dog. Diabetes 24:1050-1056. Penhos JC, Wu C - H , Basabe JC, Lopez N, Wolff F W (1969) A rat pancreas-small gut preparation for the study of intestinal factors and insulin release. Diabetes 18:733-738. Perley MJ , Kipnis D M (1967) Plasma insulin responses to oral and intravenous glucose: studies in normal and diabetic subjects. Journal of Clinical Investigation 46:1954-1962. Pettersson M , Ahren B, Lundquist I, Bottcher G , Sundler F (1987) Neuropeptide Y : intrapancreatic neuronal localization and inhibitory effects on glucose-stimulated insulin secretion in the mouse. Cell and Tissue Research 248:43-48. Pipeleers D G (1984) Islet cell interactions with pancreatic 0-cells. Experientia 40:1114-1126. Pipeleers D G (1987) The biosociology of pancreatic B cells. Diabetologia 30:277-291. Pipeleers D G , in't Veld P A Maes E, Van de Winkel M (1982) Glucose-induced insulin release depends on functional cooperation between islet cells. Proceedings of the National Academy of Science (U.S.A.) 79:7322-7325. Pipeleers D G , in't Veld P A Van de Winkel M , Maes E , Schuit F C , Gepts W (1985a) A new in vitro model for the study of pancreatic A and B cells. Endocrinology 117:806-816. Pipeleers D G , Schuit F C , in't Veld P A Maes E , Hooghe-Peters E L , Van de Winkel M , Gepts W (1985b) Interplay of nutrients and hormones in the regulation of insulin release. Endocrinology 117:824-833. Pipeleers D G , Schuit F C , Van Schravendijk C F H , Van de Winkel M (1985c) Interplay of nutrients and hormones in the regulation of glucagon release. Endocrinology 117:817-823. Polak J M , Bloom SR (1986) Somatostatin localization in tissues. Scandinavian Journal of Gastroenterology 21(suppl. 119):11-21. Polak J M , Bloom SR, Kuzio M , Brown JC,' Pearse A G E (1973) Cellular localisation of gastric inhibitory polypeptide in the duodenum and jejunem. Gut 14:284-288. Porte D Jr (1967) Beta adrenergic stimulation of insulin release in man. Diabetes 16:150-155. Porte D Jr (1991) j8-cells in type II diabetes mellitus. Diabetes 40:166-180. Porte D Jr, Girardier L, Seydoux J, Wollheim CB, Kanazawa Y, Posternak J (1973) Neural regulation of insulin secretion in the dog. Journal of Clinical Investigation 52:210-214. Porte D Jr, Pupo A A (1969) Insulin responses to glucose: evidence for a two pool system in man. Journal of Clinical Investigation 48:2309-2318. 212 Porte D Jr, Williams R H (1966) Inhibition of insulin release by norepinephrine in man. Science 152:1248-1250. Prentki M , Matschinsky F M (1987) C a 2 + , cAMP, and phospholipid-derived messengers in coupling mechanisms of insulin secretion. Physiological Reviews 4:1185-1248. Rabinovitch A , Cuendet GS, Sharp G W G , Renold A E , Mintz D H (1978) Relation of insulin release to cyclic A M P content in rat pancreatic islets maintained in tissue culture. Diabetes 27:766-773. Rabinovitch A , Dupre J (1972) Insulinotropic and glucagonotropic activities of crude preparation of cholecystokinin-pancreozymin. Clinical Research 20:945 (Abstract). Rasmussen H , Zawalich K C , Ganesan S, Calle R, Zawalich WS (1990) Physiology and pathophysiology of insulin secretion. Diabetes Care 13:655-666. Rehfeld JF, Stadil F (1973) The effect of gastrin on basal- and glucose-stimulated insulin secretion in man. Journal of Clinical Investigation 52:1415-1426. Rehfeld JF, Stadil F (1975) The glucose-induced gastrointestinal stimulation of insulin secretion in man: relation to age and to gastrin release. European Journal of Clinical Investigation 5:273-283. Reichlin S (1983) Somatostatin. New England Journal of Medicine 309:1495-1501. Reimers J, Nauck M , Creutzfeldt W, Strietzel J, Ebert R, Cantor P, Hoffmann G (1988) Lack of insulinotropic effect of endogenous and exogenous cholecystokinin in man. Diabetologia 31:271-280. Rizza R A , Oyer PE , Gerich J E (1979) Role of glucagon, catecholamines, and growth hormone in human glucose counterregulation: effects of somatostatin and combined a- and /3-adrenergic blockade on plasma glucose recovery and glucose flux rates after insulin-induced hypoglycemia. Journal of Clinical Investigation 64:62-71. Robertson RP, Porte D Jr (1973) Adrenergic modulation of basal insulin secretion in man. Diabetes 22:1-8. Rocha D , Faloona G R , Unger R H (1972) Glucagon-stimulating activity of twenty amino acids in dogs. Journal of Clinical Investigation 51:2346-2351. Rohner-Jeanrenaud F, Hochstrasser A - C , Jeanrenaud B (1983) Hyperinsulinemia of preobese and obese fa/fa rats is partly vagus nerve mediated. American Journal of Physiology 244:E317-E322. Rohner-Jeanrenaud F, Jeanrenaud B (1988) Abnormal regulation of pancreatic glucagon secretion in obese fa/fa rats. Diabetologia 31:235-240. 213 Ross SA, Dupre J (1978) Effects of ingestion of triglyceride or galactose on secretion of gastric inhibitory polypeptide and on responses to intravenous glucose in normal and diabetic subjects. Diabetes 27:327-333. Ross SA, Dupre J, Brown JC (1977) Hypersecretion of gastric inhibitory polypeptide following oral glucose in diabetes mellitus. Diabetes 26:525-529. Roy M W , Lee K C , Jones S, Miller R E (1984) Neural control of pancreatic insulin and somatostatin secretion. Endocrinology 115:770-775. Rushkakoff RJ, Goldfine ID, Carter JD, Liddle R A (1987) Physiological concentrations of cholecystokinin stimulate amino-acid induced insulin release in humans. Journal of Clinical Endocrinology and Metabolism 65:395-401. Sakamoto C, Goldfine ID, Roach E , Williams J A (1985) Localization of saturable C C K binding sites in rat pancreatic islets by light and electron microscope autoradiography. Diabetes 35:38-43. Salera M , Giacomoni P, Pironi L, Cornia G , Capelli M , Marini A , Benfenati F, Miglioli M , Barbara L (1982) Gastric inhibitory polypeptide (GIP) release after oral glucose: relationship to glucose intolerance, diabetes mellitus, and obesity. Journal of Clinical Endocrinology and Metabolism 55:329-336. Salera M , Giacomoni P, Pironi L, Ustra C, Capelli M , Giorgi A , Miglioli M , Barbara L (1983) Orcadian rhythm of gastric inhibitory polypeptide (GIP) in man. Metabolism 32:21-24. Samols E , Bonner-Weir S, Weir G C (1986) Intra-islet insulin-glucagon-somatostatin relationships. Clinics in Endocrinology and Metabolism 15:33-58. Samols E , Stagner JI (1990) Islet somatostatin - microvascular, paracrine, and pulsatile regulation. Metabolism 9(suppl 2):55-60.7 Samols E , Stagner JI (1991) Islet perfusion is B-+A-+D in the human pancreas. Diabetes 40(suppl 1):73A. Samols E , Stagner JI, Ewart R B L , Marks V (1988) The order of islet microvascular cellular perfusion is B A D in the perfused rat pancreas. Journal of Clinical Investigation 82:350-353. Samols E , Stagner JI, Weir G C (1981) Autonomic function and control of pancreatic somatostatin. Diabetologia 20:388-392. Samols E , Weir G C (1979) Adrenergic modulation of pancreatic A , B, and D cells, a-adrenergic suppression and ^-adrenergic stimulation of somatostatin secretion, a-adrenergic stimulation of glucagon secretion in the perfused dog pancreas. Journal of Clinical Investigation 63:230-238. 214 Sandberg E , Ahren B, Tendler D , Efendic S (1988) Cholecystokinin-33 potentiates and vasoactive intestinal polypeptide inhibits gastric inhibitory polypeptide-induced insulin secretion in the perfused rat pancreas. Acta Endocrinologica (Copenhagen) 117:545-551. Sarson D L , Bryant M G , Bloom SR (1980) A radioimmunoassay of gastric inhibitory polypeptide in human plasma. Journal of Endocrinology 85:487-496. Sarson D L , Wood SM, Kansal PC, Bloom SR (1984) Glucose-dependent insulinotropic polypeptide augmentation of insulin. Physiology or pharmacology? Diabetes 33:389-393. Schafer R, Schatz H (1979) Stimulation of (pro-)insulin biosynthesis and release by gastric inhibitory polypeptide in isolated islets of rat pancreas. Acta Endocrinologica (Copenhagen) 91:493-500. Schatz H , Kullek U (1980) Studies on the local (paracrine) actions of glucagon, somatostatin and insulin in isolated islets of rat pancreas. F E B S Letters 122:207-210. Schauder P, Brown JC, Frerichs H , Creutzfeldt W (1975) Gastric inhibitory polypeptide: effect on glucose-induced insulin release from pancreatic islets in vitro. Diabetologia 11:483-484. Schauder P, Mcintosh C, Arends J, Arnold R, Frerichs H , Creutzfeldt W (1976) Somatostatin and insulin release from isolated rat pancreatic islets stimulated by glucose. F E B S Letters 68:225-227. Schmid R, Schusdziarra V, Aulehner R, Weigert N, Classen M (1990) Comparison of GLP-I (7-36 amide) and GIP on release of somatostatin-like immunoreactivity and insulin from the isolated rat pancreas. Zeitschrift fur Gastroenterologie 28:280-284. Schnuerer E M , Rokaeus A , Carlquist M , Bergman T, Dupre J, McDonald T J (1990) Rat and porcine galanin are equipotent in inhibiting insulin responses to glucose in the anesthetized rat. Pancreas 5:70-74. Schuit F C , Pipeleers D G (1985) Regulation of adenosine 3',5'-monophosphate levels in the pancreatic B cell. Endocrinology 117:834-840. Schuit F C , Pipeleers D G (1986) Differences in adrenergic recognition by pancreatic A and B cells. Science 232:875-877. Seino S, Seino Y, Takemura J (1981) Somatostatin, insulin and glucagon secretion in obese rats. American Journal of Physiology 241:E146-E150. Seltzer HS, Allen E W , Herron A L Jr, Brennan M T (1967) Insulin secretion in response to glycemic stimulus: relation of delayed initial release to carbohydrate intolerance in mild diabetes mellitus. Journal of Clinical Investigation 46:323-335. 215 Service FJ, Rizza R A , Westland R E , Hall L D , Gerich JE, Go V L W (1984) Gastric inhibitory polypeptide in obesity and diabetes mellitus. Journal of Clinical Endocrinology and Metabolism 58:1133-1140. Sharp R, Gilbert S, Cook J, Jennings A , Burr IM (1974) Cholinergic modification of glucose-induced biphasic insulin release in vitro. Journal of Clinical Investigation 53:710-716. Shiekh SP, Hoist JJ, Skak-Nielsen T, Knigge U , Warberg J, Theodorsson-Norheim E , Hokfelt T, Lundberg J M , Schwartz T W (1988) Release of N P Y in pig pancreas: dual parasympathetic and sympathetic regulation. American Journal of Physiology 255:G46-G54. Shimizu T, Parker JC, Najafi H , Matschinsky F M (1988) Control of glucose metabolism in pancreatic /8-cells by glucokinase, hexokinase, and phosphofructokinase. Model study with cell lines derived from 0-cells. Diabetes 37:1524-1530. Shino A , Matsuo T, Iwatsuka H , Suzuoki F (1973) Structural changes of pancreatic islets in genetically obese rats. Diabetologia 9:413-421. Siegel E G , Creutzfeldt W (1985) Stimulation of insulin release in isolated islets by GIP in physiological concentrations and its relation to islet cyclic A M P content. Diabetologia 28:857-861. Siegel E G , Creutzfeldt W (1988) Hormone responsiveness of collagenase-isolated rat islets is markedly increased by short-term maintenance in tissue culture. Hormone and Metabolic Research 20:184-185. Siegel E G , Wollheim CB, Janjic D, Ribes G , Sharp G W G (1983) Involvement of C a + + in the impaired glucose-induced insulin release from islets cultured at low glucose. Diabetes 32:993-1000. Silvestre R A , Miralles P, Monge L, Moreno P, Villanueva L, Marco J (1987) Effects of galanin on hormone secretion from the in situ perfused rat pancreas and on glucose production in rat hepatocytes in vitro. Endocrinology 121:378-383. Soon-Shiong P, Debas H T , Brown JC (1979) The evaluation of gastric inhibitory polypeptide (GIP) as the enterogastrone. Journal of Surgical Research 26:681-686. Soon-Shiong P, Debas H T , Brown JC (1984) Bethanecol prevents inhibition of gastric acid secretion by gastric inhibitory polypeptide. American Journal of Physiology 247:G171-Sorenson R L , Lindell D V , Elde RP (1980) Glucose stimulation of somatostatin and insulin release from the isolated, perfused rat pancreas. Diabetes 29:747-751. Stefan Y, Meda P, Neufeld M , Orci L (1987) Stimulation of insulin secretion reveals heterogeneity of pancreatic B cells in vivo. Journal of Clinical Investigation 80:175-183. 216 Stolz DJ , Martin R J (1982) Role of insulin in food intake, weight gain and lipid deposition in the Zucker obese rat. Journal of Nutrition 112:997-1002. Suzuki S, Kawai K, Ohashi S, Mukai H , Yamashita K (1989) Comparison of the effects of various C-terminal and N-terminal fragment peptides of glucagon-like peptide-1 on insulin and glucagon release from the isolated perfused rat pancreas. Endocrinology 125:3109^ 3114. Suzuki S, Kawai K, Ohashi S, Mukai H , Murayama Y, Yamashita K (1990) Reduced insulinotropic effects of glucagonlike peptide l-(7-36)-amide and gastric inhibitory polypeptide in isolated perfused diabetic rat pancreas. Diabetes 39:1320-1325. Szecowka J, Grill V , Sandberg E , Efendic S (1982a) Effect of GIP on the secretion of insulin and somatostatin on the accumulation of cyclic A M P in vitro in the rat. Acta Endocrinologica (Copenhagen) 99:416-421. Szcowka J, Lins PE, Efendic S (1982b) Effects of cholecystokinin, gastric inhibitory polypeptide, and secretin on insulin and glucagon secretion in rats. Endocrinology 110:1268-1272. Taborsky G J Jr, Ensinck JW (1984) Contribution of the pancreas to circulating somatostatin-like immunoreactivity in the normal dog. Journal of Clinical Investigation 73:216-223. Takeda J, Seino Y, Tanaka K-I, Fukumoto H , Kayano T, Takahashi H , Mitani T, Kurono M , Suzuki T, Tobe T, Imura H (1987) Sequence of an intestinal c D N A encoding human gastric inhibitory polypeptide precursor. Proceedings of the National Academy of Sciences 84:7005-7008. Tasaka Y, Sekine M , Wakatsuki T (1975) Levels of pancreatic glucagon, insulin and glucose during twenty-four hours of the day in normal subjects. Hormone and Metabolic Research 7:205-206. Tatemoto K, Rokaeus A , Jornvall, McDonald TJ, Mutt V (1983) Galanin - a novel biologically active peptide from porcine intestine. F E B S Letters 164:124-128. Thomas FB, Sinar D, Mazzaferri E L , Cataland S, Mekhjian S, Caldwell JH, Fromkes JJ (1978) Selective release of gastric inhibitory polypeptide by intraduodenal amino acid infusion in man. Gastroenterology 74:1261-1265. Tokuyama Y, Kanatsuka A Ohsawa H , Yamaguchi T, Makino H , Yoshida S, Nagase H , Inoue S (1991) Hypersecretion of islet amyloid polypeptide from pancreatic islets of ventromedial hypothalamic-lesioned rats and obese Zucker rats. Endocrinology 128:2739-2744. Trimble E R , Herberg L (1983) Leucine-induced insulin and somatostatin secretion from pancreatic islets of Zucker rats: effect of pair feeding. In: Lessons from Animal Diabetes. Eds. E Shafrir, A E Renold. John Libbey & Co., London, England, pp. 25-27. 217 Unger R H (1983) Insulin-glucagon relationships in the defense against hypoglycemia. Diabetes 32:575-583. Unger R H , Dobbs (1978) Insulin, glucagon, and somatostatin secretion in the regulation of metabolism. Annual Review of Physiology 40:307-343. Unger R H , Eisentraut A M (1969) Entero-insular axis. Archives of Internal Medicine 123:261-266. Unger R H , Eisentraut A M , McCall MS, Madison L L (1962) Measurements of endogenous glucagon in plasma and the influence of blood glucose concentration upon its secretion. Journal of Clinical Investigation 41:682-689. Unger R H , Ohneda A, Aguilar-Parada E , Eisentraut A M (1969) The role of aminogenic glucagon secretion in blood glucose homeostasis. Journal of Clinical Investigation 48:810-822. Unger R H , Orci L (1981) Glucagon and the A cell. New England Journal of Medicine 304:1518-1524; 1575-1580. Uvnas-Wallenstein K, Nilson G (1978) A quantitative study of the insulin release induced by vagal stimulation in anaesthetized cats. Acta Physiologica Scandinavia 102:137-142. Van der Vliet JA, Meloche R M , Field M J , Chen D, Kaufman D B , Sutherland D E R (1988) Pancreatic islet isolation in rats with ductal coliagenase distension, stationary digestion, and dextran separation. Transplantation 45:493-495. Van de Winkel M , Pipeleers D G (1982) Autofluorescence-activated cell sorting of pancreatic islet cells: purification of insulin-containing B-cells according to glucose-induced changes in cellular redox state. Biochemical and Biophysical Research Communications 114:835-842. Van Schravendijk C F H , Foriers A , Hooghe-Peters E L , Rogiers V, De Meyts P, Sodoyez JC, Pipeleers D G (1985) Pancreatic hormone receptors on islet cells. Endocrinology 117:841-848. Vara E , Tamarit-Rodriguez J (1988) Islet secretion of immunoreactive thyrotropin-releasing hormone and the "paracrine-like" effects of its exogenous administration. Acta Endocrinologica (Copenhagen) 118:429-436. Verchere CB, Kwok Y N , Brown JC (1991) Modulation of acetylcholine-stimulated insulin release by glucose and gastric inhibitory polypeptide. Pharmacology 42:273-282. Verspohl EJ, Ammon H P T (1987) Cholecystokinin (CCKg) regulates glucagon, insulin, and somatostatin secretion from isolated islets: interaction with glucose. Pfugers Archives 410:284-287. 218 Verspohl EJ, Ammon HPT, Williams JA, Goldfine ID (1986a) Evidence that cholecystokinin interacts with specific receptors and regulates insulin release in isolated rat islets of Langerhans. Diabetes 34:390-394. Verspohl EJ, Kaimer P, Wahl M , Ammon H P T (1988) Decreased insulin secretory responses of pancreatic islets during culture in the presence of low glucose is associated with diminished 4 5 C a 2 + net uptake, N A D P H / N A D P + and G S H / G S S G ratios. Life Sciences 43:209-219. Verspohl EJ, Melien M C , Ammon H P T (1986b) Decreased insulin binding to pancreatic islets of genetically obese if a/fa) rats. Diabetes Research and Clinical Practice 2:29-67. Verspohl EJ , Tacke R, Mutschler E , Lambrecht G (1990) Muscarinic receptor subtypes in rat pancreatic islets: binding and functional studies. European Journal of Pharmacology 178:303-311. Wang C C , Grossman MI (1951) Physiological determination of release of secretin and pancreozymin from intestine of dog with transplanted pancreas. American Journal of Physiology 164:527-545. Wang JL, Easom R A , Hughes JH, McDaniel M L (1990) Evidence for a role of microfilaments in insulin release from purified /3-cells. Biochemical and Biophysical Research Communications 171:424-430. Wang JL, McDaniel M L (1990) Secretagogue-induced oscillations of cytoplasmic C a 2 + in single B and a-cells obtained from pancreatic islets by fluorescence-activated cell sorting. Biochemical and Biophysical Research Communications 166:813-818. Weir G C , Bonner-Weir S (1990) Islets of Langerhans: the puzzle of intraislet interactions and their relevance to diabetes. Journal of Clinical Investigation 85:983-987. Weir G C , Mosjov S, Hendrick G K , Habener JF (1989) Glucagonlike peptide 1 (7-37) actions on endocrine pancreas. Diabetes 38:338-342. Westermark P, Wernstedt C, Wilander E , Hayden D W , O'Brien T D , Johnson K H (1987) Amyloid fibrils in human insulinoma and islets of Langerhans of the diabetic cat are derived from a neuropeptide-like protein also present in normal islets. Proceedings of the National Academy of Sciences (U.S.A.) 84:3881-3885. Wilchek M , Bayer E A (1988) The avidin-biotin complex in bioanalytical applications. Analytical Biochemistry 171:1-32. Williams JA, Goldfine ID (1985) The insulin-pancreatic acinar axis. Diabetes 34:980-986. Woods SC, Porte D Jr (1974) Neural control of the endocrine pancreas. Physiological Reviews 54:596-619. 219 Yajima H , Fujii N, Akaji K, Sakurai M , Nomizu M , Misuta K, Aona M , Moriga M , Inoue K, Hisotani R, Tobe T (1985) Synthesis of a 42-residue peptide corresponfing to the entire amino acid sequence of human GIP. Chemical and Pharmaceutical Bulletin (Tokyo) 33:3578-3580. Yoshimura T, Ishizuka J, Greeley G H Jr, Thompson JC (1989) Effect of galanin on glucose-, arginine-, or potassium-stimulated insulin release. American Journal of Physiology 256:E619-E623. Zawalich WS, Rasmussen H (1990) Control of insulin secretion: a model involving C a 2 + , c A M P and diacylglycerol. Molecular and Cellular Endocrinology 70:119-137. Zawalich WS, Rognstad R, Pagliara AS, Matschinsky F M (1977) A comparison of the utilization rates and hormone-releasing actions of glucose, mannose, and fructose in isolated pancreatic islets. Journal of Biological Chemistry 252:8519-8523. Zawalich WS, Takuwa N, Takuwa Y, Diaz V A Rasmussen H (1987) Interactions of cholecystokinin and glucose in rat pancreatic islets. Diabetes 36:426-433. Zawalich WS, Zawalich K C , Rasmussen H (1989) Interactions between cholinergic agonists and enteric factors in the regulation of insulin secretion from isolated perifused rat islets. Acta Endocrinologica 120:702-707. Zucker L M , Zucker T F (1961) Fatty, a new mutation in the rat. Journal of Heredity 52:275-278. Zucker T F , Zucker L M (1962) Hereditary obesity in the rat associated with high serum fat and cholesterol. Proceedings of the Society of Experimental Biology and Medicine 110:165-171. Zunz E , Labarre,J (1929) Contributions a l'etude des variations physiologiques de la secretion interne du pancreas: Relations entre les secretions externe et interne du pancreas. Archives of International Physiology 31:20-44. 220 APPENDIX I: CHEMICAL SOURCES C H E M I C A L S O U R C E Acetic acid (glacial) B D H Acetonitrile B D H Acetylcholine chloride Sigma Alkaline phosphatase conjugated antimouse immunoglobin T A G O Ammonium acetate Baker Amphoterocin B Gibco Aprotinin (Trasylol R ) Miles Biotin-N-hydroxysuccinimide ester Vector Bovine serum albumin (Fraction V) Sigma Bovine serum albumin (RIA Grade) Sigma, Miles Calcium chloride Fisher Calf Supreme Serum Gibco Carbon decolorizing neutral (activated charcoal) Fisher Chloramine T Eastman Kodak Cholecystokinin-8 Sigma Citric acid B D H CMRL-1066 medium Gibco Collagenase (Type XI) Sigma 1-Cysteine Sigma Deoxyribonuclease Sigma Dextran (Clinical, Industrial Grade) Sigma Dextran T70 Pharmacia ISr^'-O-dibutyryladenosine 3':5'-cyclic monophosphate Sigma Diethanolamine B D H 3,5-Diiodo-l-tyrosine Sigma Dimethylsulfoxide Sigma Disodium p-nitrophenyl phosphate Sigma Dulbecco's Modified Eagle's Medium (powder) Gibco, Sigma Dulbecco's Modified Eagle's Medium (sterile solution) Terry Fox Labs Ethanol Commercial Alcohol [Ethylene-bis-(oxyethylenenitrilo)]-tetraacetic acid Fisher Formaldehyde (histological grade) Fisher Formic acid B D H Forskolin Sigma Galanin (porcine, rat) Peninsula Gastric inhibitory polypeptide (1-42; E G III) M R C Gastric inhibitory polypeptide (1-30) Shizuoka, Japan Gentamycin sulphate Sigma Glucagon Novo 1 2 5I-Glucagon Amersham Glucagon-like peptide-I (7-36-NH2) Peninsula Glucose (50 % commercial solution) Abbott 221 C H E M I C A L S O U R C E Glutamine Sigma Growth hormone releasing factor Sigma Hank's Balanced Salt Solution (powder) Gibco Heparin Fisher Hydrochloric acid Fisher N'-2-hydroxyethylpiperazine-N'-ethanesulfonic acid B D H Insulin (porcine, rat) Novo 3-iodo-l-tyrosine Sigma 3-isobutyl-l-methylxanthine Sigma Lactose Sigma Laminin Sigma Magnesium chloride Fisher Magnesium sulphate Fisher Maleic acid Eastman Kodak Phenol red Sigma Picric acid B D H Polyethylene glycol 8000 Fisher Potassium chloride Fisher Potassium iodide Fisher Potassium phosphate (monobasic) Fisher Protease Type X I V (Pronase E) Sigma Protein assay reagents Biorad Q U S O (microfine silica, G-32) Philadelphia Quartz Sephadex (CM-52, G-10, G-10) Pharmacia Sodium acetate Fisher Sodium azide Baker Sodium barbital Fisher Sodium bicarbonate Fisher Sodium chloride Fisher Sodium carbonate Fisher Sodium hydroxide Fisher Sodium 1 5iodide Amersham Sodium merthiolate Eastman Kodak Sodium metabisulphite Fisher Sodium pentobarbital M T C Sodium phosphate (monobasic and dibasic) Fisher Sodium pyruvate Gibco Somatostatin (synthetic cyclic and Tyr 1) Peninsula Trifluoracetic acid Pierce Tris Base Sigma Trypsin Worthington Trypsin-EDTA Sigma Tween 20 Calbiochem Vasoactive intestinal polypeptide Peninsula 222 APPENDIX II: LIST OF ABBREVIATIONS A C h B-GIP B N H S B S A C C K C E P DNase dbcAMP DIT D M E D M S O E D T A E G T A E H E L I S A F A C S GIP G L I G L P G R F HBSS H E P E S H P L C 125j I B M X IRI K I U MIT NSB PBS P E G PP R I A S E M SLI T F A VIP Acetylcholine chloride Biotinylated gastric inhibitory polypeptide Biotin-N-hydroxysuccinimide ester Bovine serum albumin Cholecystokinin Charcoal extracted plasma Deoxyribonuclease N ,2'-0-dibutyryladenosine 3':5'-cyclic monophosphate 3,5-Diiodo-l-tyrosine Dulbecco's Modified Eagle's Medium Dimethylsulfoxide disodium ethylenediaminetetraacetate [Ethylene-bis-(oxyethylenenitrilo)]-tetraacetic acid HEPES-buffered Earle's medium Enzyme-linked immunosorbent assay Fluorescence-activated cell-sorter Gastric inhibitory polypeptide Glucose-dependent insulinotropic polypeptide Glucagon-like immunoreactivity Glucagon-like peptide Growth hormone releasing factor Hank's Balanced Salt Solution N'-2-Hydroxyethylpiperazine-N'-ethanesulfonic acid High pressure liquid chromatography 1 2 5Iodine 3-Isobutyl- 1-methylxanthine Immunoreactive insulin Kallikrein inactivating units 3-Iodo-l-tyrosine Non-specific binding Phosphate-buffered saline Polyethylene glycol Pancreatic polypeptide Radioimmunoassay Standard error of the mean Somatostatin-like immunoreactivity Trifluoracetic acid Vasoactive intestinal polypeptide 223 APPENDIX i n : ANTIBODY SOURCES A . FIRST L A Y E R Antigen Antibody/ Antiserum Source Dilution Glucagon (porcine) Insulin (rat) Somatostatin 23.6.B4 GP01 Pancreatic Polypeptide n/a (avian) S O M A 08 Gregor, Berlin 1:100 monoclonal, a U.B.C. (guinea 1:1000 pig antiserum) Buchanan, Belfast 1:2000 (rabbit antiserum) U.B .C. monoclonal, b 1:1000 a: tissue culture medium b: ascites B. S E C O N D L A Y E R Second Antibody Source Dilution Conjugate Rabbit anti-guinea pig Donkey anti-rabbit Goat anti-mouse Miles Jackson Jackson 1:300 1:2000 1:2000 Fluorescein Fluorescein Rhodamine 224 APPENDIX IV: SYSTEME INTERNATIONAL (SI) UNITS Measurement SI Unit Common Unit Conversion factors Common-^SI SI-^common Glucose GIP Glucagon Insulin Somatostatin m M p M p M p M p M mg/dl pg/ml pg/ml MU/ml pg/ml 0.056 0.201 0.287 6.00 0.611 18 4.98 3.48 1.67 1.64 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0101044/manifest

Comment

Related Items