Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Release and metabolism of gastric inhibitory polypeptide Kieffer, Timothy James 1994

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

Item Metadata

Download

Media
831-ubc_1994-894199.pdf [ 4.54MB ]
Metadata
JSON: 831-1.0088076.json
JSON-LD: 831-1.0088076-ld.json
RDF/XML (Pretty): 831-1.0088076-rdf.xml
RDF/JSON: 831-1.0088076-rdf.json
Turtle: 831-1.0088076-turtle.txt
N-Triples: 831-1.0088076-rdf-ntriples.txt
Original Record: 831-1.0088076-source.json
Full Text
831-1.0088076-fulltext.txt
Citation
831-1.0088076.ris

Full Text

RELEASE AND METABOLISM OF GASTRIC INHIBITORY POLYPEPTIDE by TIMOTHY JAMES KIEFFER  B.Sc., The University of British Columbia, 1989  A THESIS SUBMrrIED iN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Physiology  We accept this thesis as conforming to the required standard  THE UNWERS1TY OF BRITISH COLUMBIA April 1994 © Timothy James Kieffer, 1994  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 In presenting this thesis  publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of The University of British Columbia Vancouver, Canada Date  DE-6 (2/88)  t’t  11  ABSTRACT This thesis reports methodology that was developed to isolate and enrich intestinal endocrine cell preparations by centrifugal elutriation and short-term culture to enable the study of the regulation of IRGIP secretion at the cellular level. Adherent canine epithelial cell cultures contained —10% IRGIP cells. The release of IRGIP from these cells was significantly increased by incubation with depolarizing concentrations of K, glucose, somatostatin immunoneutralizing antibody, or GRP, in addition to pharmacological elevations of intracellular cAMP or Ca . It is concluded that this method provides a 2 means of further investigating the cellular mechanisms controlling GIP release. Enteroendocrine tumor cell lines were also investigated as an alternate source of GIP cells. A cell line derived from intestinal tumors of transgenic mice (STC- 1) was subcloned to produce a cell line with —30% IRGIP (STC 6-14). HPLC of STC 6-14 extracts indicated that the tumor cell derived IRGIP eluted with synthetic porcine GTP 1-42. Release of IRGIP from STC 6-14 cells; increased in a concentration dependent fashion in response to glucose, was augmented by the addition of somatostatin neutralizing antibody, and attenuated by exogenous somatostatin. Immunoreactive somatostatin (IRSS) release was significantly increased by adding GIP to the incubation medium. It is concluded that this cell line represents a means of rapidly obtaining large numbers of GIP cells, and thus should be useful to investigate stimulus-secretion coupling in the GIP cell. GIP secreted by STC 6-14 cells was metabolized by a serum constituent to biologically inactive GIP 3-42. ‘ 1-GIP 2 5 was purified by HPLC and used as a highly sensitive means to further investigate the degradation of GIP by serum. The removal of the N-terminal dipeptide by serum could be blocked by diprotin A, a competitive inhibitor of dipeptidyl peptidase IV (DPP IV). No GIP 3-42 was produced by incubation of GIP with serum from rats specifically lacking DPP IV. Infusion of 1-GIP 25 into rats, followed ‘ by HPLC analysis, indicated that 50% was metabolized to 1-GIP 25 3-42 by —1.5 mm. It is ‘ concluded that DPP IV is a primary degradative and inactivating enzyme of GIP.  111  TABLE OF CONTENTS  ABSTRACT  ii  TABLE OF CONTENTS LIST OF FIGURES  viii  ACKNOWLEDGEMENTS  xiii CHAPTER 1 iNTRODUCTION  1.1  OVERVIEW  1.2  DISCOVERY OF GIP AS AN ENTEROGASTRONE  1.3  1.4  1.5  1  1 .2. 1  Enterogastrone concept  2  1 .2.2  Evidence for existence of OW  2  1.2.3  IsolationofGlP  3  1 .2.4  Glucagon superfamily  4  ENTEROGASTRONE ACTIONS OF GIP 1 .3. 1  Effect of GIP on gastric secretion  5  1 .3.2  Mechanism of gastric action of GIP  5  1 .3 .3  Interaction with other enterogastrones  7  DISCOVERY OF GIP AS AN INCRETIN 1 .4. 1  Incretin concept  8  1 .4.2  Enteroinsular axis  9  1 .4,3  Evidence for GIP as an incretin  10  INSULINOTROPIC ACTIONS OF GIP 1 .5. 1  Effect of OP on insulin secretion  11  1 .5 .2  Other potential incretins  12  1 5 3  Interactions between insulin secretagogues  13  1 .5.4  Relative contribution of GIP vs tGLP-I  14  .  .  iv 1 .5.5  Participation of GIP and tGLP-I in an enteroendocrine loop  15  1 .5.6  Mechanism of action of GIP on the B-cell  16  1 .5 .7  Nature of glucose dependency  18  1.6  OTHERACTIONSOFG1P  1.7  GJPGENE  19  1 .7. 1  Gene structure and post-translational processing  20  1 .7.2  Gene regulation  22  1.8  GIP CELL DISTRIBUTION AND CHARACThRIZATION  1.9  MEASUREMENT OF GIP  22  1 .9. 1  Development of GIP radioimmunoassay  24  1 .9.2  Physiological levels of GIP  24  1 .9.3  Immunoreactive forms of GIP  25  1.10 GIP SECRETION 1. 10. 1 Response to a mixed meal  27  1 10.2 Response to carbohydrate  28  1.10.3 Response tofat  32  1 10.4 Response to protein  34  1. 10.5 ANS control of GIP release  36  1 10.6 Other modulators of GIP release  38  1. 10.7 Feedback inhibition of GIP release  39  .  .  .  1.11  GIP RECEPTORS 1. 11 1 Biologically active site(s)  42  1 11 .2 OW binding studies  44  .  .  1.12 OP METABOLISM  47  GIP PATHOPHYSIOLOGY  49  1.14 THESIS INVESTIGATION  52  1.13  V  CHAPTER 2 GASTRIC INHIBITORY POLYPEPTIDE RELEASE FROM ISOLATED CANINE AND PORCINE ENDOCRINE CELLS 2.1  INTRODUCTION  2.2  MATERIALS and METHODS 2.2. 1  Animal tissues  54  2.2.2  Isolation of mucosal cells  55  2.2.3  Enrichment of mucosal endocrine cells  55  2.2.4  Endocrine enriched cell culture  56  2.2.5  Immuocytochemistry  57  2.2.6  Release experiments  58  2.2.7  Secretagogues  59  2.2.8  Radioimmunoassay for GIP  60  2.2.9  Radioimmunoassay for SS  63  2.2. 10 Expression of results 2.3  2.4  54  64  RESULTS 2.3. 1  Mucosa digestion  65  2.3.2  Endocrine cell enrichment and culture  65  2. 3 .3  ICC of intact mucosa and cultured cells  67  2.3 .4  IRGIP secretion in response to K  67  2. 3 .5  IRGIP secretion in response to glucose  68  2.3 .6  IRGIP secretion in response to A23 187  68  2.3 .7  IRGIP secretion in response to GRP  69  2.3.8  IRGIP secretion in response to forskolin  69  DISCUSSION  83  vi CHAPTER 3 GASTRIC INHIBITORY POLYPEPTIDE RELEASE FROM A TUMORDERIVED CELL LINE 3.1  INTRODUCTION  3.2  MATERIALS and METHODS  3.3  3.2. 1  Culture and production of STC 6-14 cell line  90  3 .2.2  Release experiments  91  3.2.3  Secretagogues  92  3.2.4  Peptide quantification  93  3.2.5  Immunocytochemistry  93  3.2.6  HPLC  93  3 .2.7  Expression of results  94  RESULTS 3.3. 1  Characterization of STC 6-14 cells  94  3 .3 .2  IRGIP secretion in response to glucose  95  3 3 .3  IRSS secretion in response to glucose and GIP  96  .  3.4  89  DISCUSSION  103  CHAPTER 4 GASTRIC INHIBITORY POLYPEPTIDE METABOLISM BY DIPEPTIDYL PEPTifiASE IV 4.1  INTRODUCTION  4.2  MATERIALS and METHODS  107  4.2. 1  Incubation of GIP with serum  108  4.2.2  Dipeptidyl peptidase IV-negative rats  109  4.2.3  Dipeptidyl peptidase IV assay  109  4.2.4  HPLC purification of 1-GIP 25 ‘  110  vii  4.3  4.4  4.2.5  Incubation of ‘I-GW with serum  .111  4.2.6  In vivo experiments with 125 1-GIP  112  4.2.7  Dipeptidyl peptidase IV in Zucker rats and NIDDM subjects  113  RESULTS 4.3. 1  Incubation of GIP with serum  113  4.3 .2  Dipeptidyl peptidase IV assay  114  4.3 .3  Label specific activity  114  4.3.4  Incubation of purified ‘I-GIP with Wistar rat serum  115  4.3.5  Incubation of purified ‘I-GIP with DPP IV-negative rat serum 116  4.3.6  Incubation of purified ‘I-GIP with CEP  117  4.3 .7  The analysis of I-GIP 25 degradation in vivo ‘  117  DISCUSSION  144  CHAPTER 5 SUMMARY SUMMARY AND FUTURE DIRECTIONS  151  REFERENCES  160  vi” LIST OF FIGURES  1  Amino acid sequences of porcine, human, bovine, and rodent gasthc inhibitory polypeptide  4  2  Sequence of a cDNA encoding the human GIP precursor  21  3  A) IRGIP and IRSS content from the canine single cell suspension prior to elutriation and the two fractions resulting from elutriation. B) IRGIP content from isolated porcine cell fractions  71  4  Canine jejunum immunostained for GIP using the peroxidase method  72  5  Isolated canine epithelial cells after 40 h culture, immunostained for GIP.  6  Isolated canine epithelial cells after 40 h culture, immunostained for somatostatin  74  Canine jejunum stained for neutral and acid mucins using a combined alcian blue-periodic-acid Schiff technique  75  7  8  9  10  11  12  Isolated canine epithelial cells after 40 h culture, stained for neutral and acid mucins using a combined alcian blue-periodic-acid Schiff technique.  .  .  .  .  .  .  73  76  Effect of K concentrations of 10 to 55 mM on basal IRGIP release (5 mM K+) from isolated canine and porcine endocrine cells in the presence of 5 mM glucose  77  Concentration-response relationship between glucose (5 to 20 mM) and LRGIP secretion from isolated canine epithelial cells  78  Concentration-response relationship between glucose (5 to 20 mM) and IRGIP secretion from isolated porcine epithelial cells  79  Effect of 1, 5, and 10 IIM A23 187 on basal release of IRGIP in the presence of 5 mM glucose  80  ix 13  Effect of graded concentrations of GRP (0 to 100 nM) on IRGIP release from cultured canine epithelial cells  81  Effect of 0.1 to 100 .tM forskolin in the presence of 5 mM glucose on IRGIP release from isolated canine epithelial cells  82  15  Cultured STC 6-14 cells immunostained for GIP  97  16  HPLC elution profile of natural porcine GIP and synthetic porcine GIP from a C 18 column with an acetonitrile gradient of 28-33% over 10 mm  98  14  17  IRGIP content of HPLC fractions: A) Overlay of elution profiles for natural porcine GIP and synthetic porcine GW. B) Overlay of elution profiles for STC 6-14 cell extract and culture medium  18  19  20  21  Effect of 250 j.tg/ml anti-somatostatin antibody (SOMA-lO) or 10 nM somatostatin on the dose-response relationship between glucose concentrations (5 to 20 mM) and IRGIP secretion from STC 6-14 cells  99  .  .  .  .  100  Concentration-response relationship between glucose concentrations (5 to 20 mM) and IRSS secretion from STC 6-14 cells  101  Effect of graded concentrations of porcine GTP (0-100 nM) on IRSS release from STC 6-14 cells in the presence of 5 mM glucose  102  Determination of relative DPP IV levels in Fischer DPP TV-positive) and DPP TV-negative rats from rate of production of p-nitroaniline. A) Standard curve of absorbance for increasing concentrations of p nitroaniline (mM) B) Plot of p-nitroaniline concentration vs. time (mm) deterimined from absorbance values converted to mM p-nitroaniline from curve A  119  x 22  23  24  25  26  Determination of the relative levels of DPP IV in NIDDM and normal subjects from the production of p-nitroaniline after 45 mm  120  Profile of synthetic porcine GIP iodination mixture eluted on Sephadex G 15  121  A) % bound for a GIP standard curve using increasing GIP concentrations, and a self displacement curve using increasing amounts of ‘I-GIP. B) Plot of mass of GIP standard vs. I1 -GIP radioactivity as deterimined from A) for a number of points  122  A) HPLC elution profile of ‘I-GIP at 32% acetonitrile for 10 mm, followed by a gradient to 38% from 10 to 20 mm. B) HPLC elution profile of peak 2 after 25 h incubation at 37°C  123  Both figures are HPLC elution profiles of 1-GIP 25 incubated with Wistar ‘ rat serum (10%) at 37°C. A) duration = 10 mm. B) duration = lh  124  27  Both figures are HPLC elution profiles of 1-GJP 25 incubated with Wistar rat ‘ serum (10%) at 37°C. A) duration = ih. B) duration = 125  28  Both figures are HPLC elution profiles of 1-GIP 25 incubated with Wistar rat ‘ serum (10%) at 37°C. A) duration = Iih. B) duration = 24Ji 126  29  Both figures are HPLC elution profiles of 125 1-GIP incubated with Wistar rat serum (10%) at 4°C. A) duration = 4..h. B) duration = 127  30  Both figures are HPLC elution profiles of 125 1-GIP incubated with Wistar rat serum (10%) at 4°C. A) duration = .19J . B) duration = 24Ji 1 128  31  Both figures are HPLC elution profiles of 1-GIP 25 incubated with Wistar ‘ rat serum (10%) and diprotin A (0.1 mM) at 37°C. A) duration = B) duration  =  6h  129  xi 32  Both figures are HPLC elution profiles of 1-GJP 25 incubated with Wistar ‘ rat serum (10%) and diprotin A (0.1 mM) at 37°C. A) duration = jh. B) duration  33  34  35  130  =  Both figures are HPLC elution profiles of 1-GIP 25 incubated with Wistar ‘ rat serum (10%) and aprotinin (2%) at 37°C. A) duration = i.h. B) duration =  131  Both figures are HPLC elution profiles of 1-GIP 25 incubated with Wistar ‘ rat serum (10%) and aprotinin (2%) at 37°C. A) duration = jjj . 1 B) duration 2j1 2  132  Both figures are HPLC elution profiles of 125 1-GIP incubated with Wistar rat serum (10%) and bacitracin (50 U/mi) at 37°C. A) duration = 2.h. B) duration  =  19.5 h  133  36  Both figures are HPLC elution profiles of 1-GIP 25 incubated with Wistar ‘ rat serum (10%) containing diprotin A (0.1 mM), aprotinin (2%) and bacitracin (50 U/mi), at 37°C. A) duration = 3.25 h. B) duration = 2h.... 134  37  Both figures are HPLC elution profiles of 125 1-GIP incubated with DPP IVnegative rat serum (10%) at 37°C. A) duration = 2h. B) duration = lb.... 135  38  Both figures are HPLC elution profiles of 125 1-GJP incubated with DPP IVnegative rat serum (10%) and aprotinin (2%) at 37°C. A) duration = 4...h. B)duration=.j2  39  40  136  Both figures are HPLC elution profiles of 125 1-GIP incubated with DPP IVnegative rat serum (10%) and aprotinin (2%) at 37°C. A) duration = lib. B)duration=27h  137  Both figures are HPLC elution profiles of 1-GIP 25 incubated with charcoal ‘ extracted human plasma (10%) used in the GIP MA, at 37°C. A) duration = j. B) duration =  138  xii 41  42  43  Both figures are HPLC elution profiles of 125 1-GIP incubated for 1 h with Wistar rat serum (neat) containing diprotin A (0.1 mM), aprotinin (2%), purified by SepPak, lyophilized, and reconstiuted. A) incubation on ice. B) incubation at 37°C  139  Both figures are HPLC elution profiles of 125 1-GIP infused into a Wistar rat, collected in 2 ml serum with diprotin A (0.1 mM) and aprotinin (2%), purified by SepPak, lyophilized and reconstituted. A) duration = 2 mm. B) duration =5 mm  140  Both figures are HPLC elution profiles of 101P 25 infused into a Wistar ‘ rat, collected in 2 ml serum with diprotin A (0.1 mM) and aprotinin (2%), purified by SepPak, lyophilized and reconstituted. A) duration = 10 mm. B) duration =20 mm  44  Both figures are HPLC elution profiles of 1-GTP 25 infused into a DPP IV‘ negative rat, collected in 2 ml serum with diprotin A (0.1 mM) and aprotinin (2%), purified by SepPak, lyophilized and reconstituted. A) duration = 2 mm. B) duration = 10 mm  45  141  142  Graphs summarizing the conversion rate of 1-GIP 25 1-42 to 1-GIP ‘ 25 3-42 ‘ both in vitro and in vivo calculated from peak areas of Figures 5 to 22, plotted as % GIP 3-42 (% of GIP 1-42 + 3-42) vs. time. A) Relative % 1-GIP 25 3-42 production in Wistar rat serum at 37°C ‘ (normal serum), + diprotin A (0.1 mM), + bacitracin (50 U/mI), or @ 4°C. B) Rate of formation of 125 1-GIP 3-42 from 101P 25 1-42 infused in a ‘ Wistar rat  143  xli’  ACKNOWLEDGEMENTS  I was introduced to the U.B.C. Physiology Department in 1988 by Dr. John Brown, and was fortunate to obtain a summer studentship position working for him. I thank Dr. Brown for cultivating my interest in research and aiding in my development as a scientist. Other members of the then MRC Regulatory Peptide Group were also equally important in my training. Dr. Kenny Kwok emphasized the need for ‘controls’ and his pharmacology background was helpful. Dr. Alison Buchan guided me through immunocytochemical techniques and the isolation of gut endocrine cells. Dr. Chris McIntosh was an endless source of answers, and directed my work with the HPLC system. Finally, my supervisor, Dr. Ray Pederson, continually guided my research and supported me throughout my degree. He also provided annual ‘retreats’ to Mayne Island, which I will always remember. Throughout my degree, I received a great deal of technical support, particularly from Haidee Barker, Dr. Zhongxian Huang, Narinder Dhatt, Leslie Checknita, Marie Langton, John Sanker and Joe Tay. Fellow graduate students were also a large part of my experience, especially Glenn Morrow who provided constant feedback on experiments, and the odd ‘one-on-one’ challenge. Finally, I would like to thank my family and Stephanie for continual support and encouragement.  1 CHAPTER 1 INTRODUCTION  1.1 OVERVIEW In 1902, Bayliss and Starling noted that water and bicarbonate secretion from the pancreas increased when weak acid was introduced into the duodenum, but not when administered intravenously. Furthermore, the intravenous injection of an acid extract of upper intestinal mucosa resulted in copious secretion from the pancreas. Thus, the presence of “secretin” was proposed. Bayliss and Starling also searched for a general name that would convey the meaning of “chemical messenger”. It was W. B. Hardy that proposed the name “hormone” derived from the Greek word for “I arouse to activity”, and although it did not suggest the property of messenger, it was adopted. The word first appeared in print in Starling’s Croonian Lecture of 1905. The original concept  -  that a chemical substance called a hormone, liberated by  one kind of cell, and carried by the blood stream to act on a distant target cell  -  represented a major advance in physiological thinking. Hormones are principal players in the maintenance of physiological homeostasis, as perturbation of this equilibrium leads to their appropriate corrective secretion. These responses are monitored and feedback loops exist to inhibit further stimulation of hormone release in a regulatory fashion. To be effective, this feedback must be accompanied by the clearance of previously secreted hormone from the circulation through enzymatic inactivation and/or removal of the substance by the liver or kidneys. Investigation of a hormone first requires isolation and identification of the active principle in order that the biological response be reproduced by administration of exogenous hormone. The demonstration of an endocrine-mediated biological response then allows examination of other components of the system. Important physiological questions are left outstanding without knowledge of stimulus-secretion coupling,  2 circulating hormone levels, and duration of action and metabolism in plasma. The goal of these investigations was to further understand the secretion and metabolism of the hormone GIP (Gastric Inhibitory Polypeptide).  1.2 DISCOVERY OF GIP AS AN ENTEROGASTRONE 12.1 ENTEROGASTRONE CONCEPT It was first reported in 1886 by Ewald and Boas that the addition of olive oil to a test meal produced an inhibition of gastric secretion in human subjects. In 1910, Pavlov demonstrated that fat added to a meal fed to dogs inhibited the secretion of acid and pepsin. In 1926, Farrell and Ivy made the observation that the acid response to a meal could be inhibited by injection of extracts of the duodenal mucosa and suggested that a hormonal mechanism must be involved. Kosaka and Lim (1930) proposed the term “enterogastrone” to describe a putative hormone which was secreted in response to fat or its digestive products in the intestinal lumen and inhibited gastric acid secretion.  12.2 EVIDENCE FOR EXISTENCE OF GIP Two candidate intestinal hormones apparently satisfying the requirements for enterogastrone were secretin (Bayliss and Starling, 1902) which was isolated for its ability to stimulate pancreatic secretion, and cholecystokinin (CCK; Ivy and Oldberg, 1927), a potent stimulator of gall bladder contraction. Impure preparations of both were found to have acid inhibitory activity in the vagally denervated (Heidenhain) canine gastric pouch (Gillespie and Grossman, 1964). However, in apparent conflict with the described acid inhibitory effects of CCK, Magee and Nakamura (1966) observed that CCK preparations were capable of stimulating acid secretion under fasting conditions when administered intravenously. Brown and Pederson (1970a) tested the hypothesis that the gastric effects resulted from the actions of factors other than CCK present in the impure preparations. The acid secretory effects of two different preparations of CCK  3 designated 10% and 40% pure on the basis of gallbladder stimulating potency were compared in dogs prepared with Heidenhain pouches.  The 40% pure preparation  produced a greater stimulatory effect on acid secretion than the 10% pure preparation at doses that yielded comparable gallbladder activity. It was proposed that either a gastric stimulant had been concentrated or an inhibitor of acid secretion had been removed during the purification procedure. Support for the latter was provided in 1971 (Pederson) when it was demonstrated that the 10% pure preparation of CCK was a more potent inhibitor of pentagastrin stimulated acid secretion than the 40% pure preparation.  12.3 ISOLATION OF GIP The canine Heidenhain pouch was used as a bioassay model for the isolation of the putative inhibitor of acid secretion. Brown et al. (1969, 1970) purified the active substance from extracts of hog duodeno-jejunal mucosa to a degree of homogeneity suitable for amino acid analysis to be performed, and named it gastric inhibitory polypeptide, or GIP (Brown, 1971; Brown and Dryburgh, 1971). Pure GIP was shown to be a potent inhibitor of gastric acid and pepsin secretion in the dog (Pederson and Brown, 1972). The amino acid sequence of the porcine peptide, as determined by Brown and Dryburgh (1971) and later corrected by Jörnvall eta!. (1981) is shown in Figure 1, along with that of human, bovine, and rat GIP. The high degree of conservation of the peptide among species suggests that GIP has an important physiological role.  4  PORCINE: HUMAN:  1 2 4 7 3 5 6 8 9 10 11 12 13 14 Tyr—Aia—Giu—Gly—Thr—Phe—Ile—Ser—Asp—Tyr—Ser—Ile—Ala-Met  BOVINE: RODENT: PORCINE: HUMAN: BOVINE: RODENT: PORCINE: HUMAN: BOVINE:  15 16 17 18 19 20 21 22 23 24 25 26 27 28 Asp-Lys-Iie-Arg-Gln-Gln-Asp-Phe-Val-Asn-Trp-Leu-Leu-Aia -His-  29 32 34 30 31 33 35 36 37 40 41 38 39 42 Gin-Lys-Giy-Lys-Lys-Ser-Asp-Trp-Lys-His-Asn-Ile-Thr-Gin -Asn  RODENT:  -lie-Asn-  -Leu  FIGURE 1: Amino acid sequences of porcine, human, bovine, and rodent gastric inhibitory polypeptide (GIP). From Jömvall et a!., 1981 (porcine); Moody et at., 1984 (human); Carlquist et at., 1984 (bovine) and Sharma et at., 1992 (rat).  12.4 GLUCAGON SUPERFAMILY The sequence of GIP, as well as the structure of its precursor peptide (Takeda et at., 1987) and the gene encoding it (Inagaki et at., 1989) indicate that GIP belongs to the glucagon superfamily of peptides, which share sequence homologies.  This family  includes glucagon, glicentin, the glucagon-like peptides (GLP-I and GLP-II), secretin, vasoactive intestinal peptide (VIP), peptide histidine-methionine (PHM), peptide histidine-isoleucine (PHI), peptide histidine-valine (PHV), growth hormone releasing factor (GRF), pituitary adenylate cyclase activating peptides (PACAP-38 and PACAP 27), helodermin, helospectins (I and II), extendins (3 and 4), and somatoliberin (Arimura, 1992; Eng et at., 1990; Parker et a!., 1984; Raufman et at., 1991; Rai et at., 1993; Robberecht et a!., 1991). It has been suggested that these hormones probably arose from a common ancestral gene (Bell, 1986; Campbell and Scanes, 1992).  5 1.3 ENTEROGASTRONE ACTIONS OF GIP 13.1 EFFECT OF GIP ON GASTRIC SECRETION Dogs surgically prepared with Heidenhain pouches of the body of the stomach were used to determine the acid inhibitory action of GJP and served as a bioassay for its isolation. Using this vagally and sympathetically denervated gastric pouch preparation, GIP dose-dependently inhibited pentagastrin-stimulated acid secretion, with the highest dose able to produce —80% inhibition (Pederson and Brown, 1972). However, in the vagally innervated gastric remnant, the same authors observed only —40% inhibition of acid secretion induced by insulin hypoglycemia. Subsequent studies in man (Maxwell et al., 1980) and rat (El-Munshid et al., 1980) also found GIP to be a weak inhibitor of pentagastrin-stimulated acid secretion when innervation was intact, and it was speculated that GIP may not be physiologically important as an enterogastrone. In a comparative study in the dog stomach, Soon-Shiong et al. (1979) reproduced the profound inhibition of pentagastrin-stimulated acid secretion observed earlier in the Heidenhain pouch but also found Gil’ had only weak effects on the innervated gastric remnant. Yamagishi and Debas (1980) noted that if oleic acid was introduced into the duodenum, a GIP infusion was able to completely inhibit acid secretion in the innervated dog stomach in response to a meal.  The possible involvement of a cholinergic  mechanism antagonistic to the action of GIP on the stomach was suggested by the observation that the acid inhibitory effect of this hormone in the denervated gasthc pouch of the dog could be blocked by the intravenous infusion of urecholine (Soon-Shiong et at., 1979). It therefore seemed possible that GIP exerted its inhibitory effect on the parietal cell indirectly, via the release of an inhibitor under cholinergic control.  1.3.2 MECHANISM OF GASTRICACTIONOF GIP In 1981, McIntosh et al. suggested that gastric somatostatin might mediate the acid inhibitory action of GJP. In testing this hypothesis, they found a strong stimulation  6 of immunoreactive somatostatin (IRSS) release from the perfused stomach by GIP. In addition, both acetyicholine and vagal stimulation potently inhibited the somatostatinotropic activity of GIP. These data accounted for the observation of weak acid inhibitory activity of GIP in the innervated stomach.  McIntosh et al. (1981)  proposed a model whereby GIP released by the presence of fat in the duodenum stimulated somatostatin secretion upon reaching the stomach.  However, the  cholinergically or neurally mediated inhibition of GIP-stimulated IRSS secretion in the rat was only partially blocked by atropine (McIntosh et al., 1979). Naloxone partially reversed the vagally induced inhibition of GIP-stimulated IRSS release, suggesting that opioid peptides were released and also inhibited IRSS release during the period of vagal stimulation (McIntosh et a!., 1983).  Subsequent studies have demonstrated that  noncholinergic transmitters, including the opioid peptides (enkephalins, dynorphins) and the tachykinins (substance P, neurokinin A), also inhibit somatostatin secretion and reverse the actions of GIP on the stomach (Kwok et a!., 1988a; McIntosh et al., 1989; 1990). In order for GIP to act as a physiological inhibitor of acid secretion, a mechanism capable of antagonizing the parasympathetic influence would be required. McIntosh et a!. (1981) suggested that in addition to stimulating GIP secretion, fat in the duodenum may activate sympathetic fibres which provide such control. Evidence for this hypothesis was supplied by the observation that catecholamines stimulate gastric IRSS release in the perfused stomach (McIntosh etal., 1981; Goto etal., 1981; Koop etal., 1981). McIntosh et a!. (1981) also observed a marked increase in IRSS with stimulation of preganglionic sympathetic nerves together with simultaneous infusion of atropine. Alternatively, Brown et a!. (1989) suggested that the inhibitory effect of GIP may be most profound when the parasympathetic activity of the stomach is minimal, such as the interdigestive period, when OP levels have been shown to remain elevated (Jorde et a!., 1980; Salera et  7 a!., 1983). While the mechanism of action is still not completely understood, GIP appears to fulfill the requirements of an enterogastrone.  13.3 INTERACTION WITH OTHER ENTEROGASTRONES Sequencing of the preproglucagon gene by two separate groups (Bell et al., 1983; Lopez et at., 1983) revealed that the precursor peptide contained, in addition to glucagon and glicentin, the sequence of two glucagon-like peptides (GLP-I and GLP-II). Mojsov et at. (1986) showed that post-translational processing of the preprohormone yielded a C-terminal 30 residue form of GLP-I, with amino acids 7-36-NH 2 (tGLP-I). As a member of the glucagon superfamily (section 1.2.4), tGLP-I has a high degree of homology with GIP, sharing 11 amino acids. Like GIP, tGLP-I has also been shown to inhibit pentagastrin and meal stimulated acid secretion in man (Schjoldager et a!., 1989; O’Halloran et a!., 1990; Wettergren et al., 1993).  In a seeming paradox, tGLP-I  stimulated H production in an enriched preparation of isolated rat parietal cells (Schmidler et al., 1991). It was suggested that in vivo, direct stimulation by tGLP-I of the parietal cells might be counterbalanced by indirect inhibitory mechanisms that are excluded in the in vitro cell system. Like GIP, tGLP-I may act indirectly via somatostatin release, as tGLP-I has been reported to stimulate somatostatin release in the isolated perfused rat stomach (Eissele et at., 1990). It is quite likely, therefore, that GIP fulfills the enterogastrone role by acting in concert with other peptides, such as CCK, secretin, and tGLP-I, in addition to nervous inhibitory mechanisms that may be initiated during the digestive and absorptive process in the upper small bowel. Nauck et al. (1992) recently questioned whether GIP and tGLP-I were in fact physiological enterogastrones in humans. These investigators examined the effects of intravenous human GIP and/or tGLP-I on pentagasthn-stimulated gastric volume and acid output. Pentagastrin significantly stimulated acid output, but neither GIP nor tGLP I, either alone or in combination, reduced pentagastrin-stimulated gasthc acid secretion.  8 It was therefore concluded that these hormones were not likely enterogastrones in man. However, these authors did not acknowledge the data from the previous decade which clearly indicated that the enterogastrone effects of these hormones are likely mediated by gastric somatostatin, which is also under neural control.  Thus, in the absence of  intraduodenal stimuli (i.e. not a physiologically relevant design), there may be no sympathetic activity to antagonize the parasympathetic inhibition of somatostatin release, and thus no reduction in gastric acid secretion (see section 1.3.2). Thus, to date, the role of GIP as an enterogastrone in man is unresolved.  1.4 DISCOVERY OF GIP AS AN INCRETIN 1.4.1 INCRETIN CONCEPT In 1902, Bayliss and Starling speculated that signals arising in the gut after ingestion of nutrients may elicit endocrine responses and affect the disposal of carbohydrates, as well as stimulating pancreatic bicarbonate secretion. This came as a result of their discovery of secretin as a regulator of pancreatic secretion. Moore et al. (1906) postulated that the duodenum produced a “chemical excitant” for pancreatic secretion, the absence of which caused diabetes. Their attempted treatment of diabetes mellitus by injection of gut extracts was unsuccessful. Zunz and Labarre (1929) and Labarre and Still (1930) prepared an intestinal extract free of secretin activity which was able to produce hypoglycemia in dogs.  Thus Labarre (1932) introduced the term  “incretin” to describe humoral activity of the gut that might enhance the endocrine secretion of the pancreas. Interest in the search for the active principle in the duodenum waned with the isolation of insulin, allowing the effective control of blood glucose in diabetes mellitus (Banting et a!., 1922). Also, studies in 1940 by Loew et a!. questioned the existence of gastrointestinal factors with insulin-releasing potency. It was not until the development of a reliable radioimmunoassay for insulin in the 1960’s by Berson and Yalow, allowing  9 measurement of circulating levels of this hormone, that interest was renewed. It became apparent that the direct effect of nutrients and their metabolites on the islets could not adequately account for the amount of insulin released. Absorption of nutrients from the gut seemed to be accompanied by the release of additional factors. The insulin response to intravenous glucose was much smaller than to either oral glucose (Elrick et al., 1964) or intrajejunal infusion (Mcintyre et a!., 1964), even though there was a greater increase in blood glucose levels with intravenous administration. It was concluded from these observations that a humoral substance was released from the jejunum during glucose absorption which acted to stimulate insulin release. Perly and Kipnis (1967) estimated that as much as half of the insulin secreted following an oral glucose load was a result of gastrointestinal factors.  1.4.2 ENTEROINSULAR AXIS In 1969, Unger and Eisentraut named the connection between the gut and the pancreatic islets described above the “enteroinsular axis”. While Labarre had noted an incretin or hormonal effect from the gut on insulin release (1932), he had provided earlier evidence that insulin could be released by vagal stimulation (Labarre, 1927). It became apparent that the gastrointestinal neural system might participate in the enteroinsular axis not only by direct stimulation of the islets, but also by regulating the release of an incretin into the blood. As it was clear that the pancreas played a major role in maintaining circulating glucose levels, it also seemed logical that nutrients, like glucose, were able to act directly on islets to control hormone release. Creutzfeldt (1979) therefore suggested that the enteroinsular axis encompass nutrient, neural and hormonal signals from the gut to the islet cells secreting insulin, glucagon, somatostatin, or pancreatic polypeptide. Furthermore, Creutzfeldt (1979) defined the criteria for fulfillment of the hormonal or incretin part of the enteroinsular axis as: 1) it must be released by nutrients, particularly  10 carbohydrates, and 2) at physiological levels, must stimulate insulin secretion in the presence of elevated blood glucose levels.  1.4.3 EVIDENCE FOR GIP AS AN INCRETIN In 1966, Dupré and Beck demonstrated that a crude preparation of CCK possessed insulinotropic activity. In 1972, Rabinovitch and Dupré found that this insulinotropic action could be removed by further purification of the CCK. This resembled the loss of the acid inhibitory activity previously reported by Brown and Pederson (1970a) in the purification of GIP from CCK, and led Dupré to the hypothesis that Gil’ may have insulin-releasing capabilities.  By 1973, Dupré et al. demonstrated that a purified  preparation of GIP infused intravenously in humans in concert with glucose, resulted in the stimulation of insulin release and an improvement in glucose tolerance. The insulin response was sustained for the duration of the GIP infusion and was not observed in the euglycemic state. While no specific GIP antagonists have been developed to further establish the physiological contribution of GIP to the postprandial insulin response, antisera to GIP have been produced which may suppress its biological activity. One of the first attempts at GIP immunoneutralization study was by Lauritsen et al. (1981). Insulin release was measured in rats given an oral glucose tolerance test (OGTT) and compared to an OGTT plus infusion of GLP antisera via the jugular vein. They found that the insulin release relative to the glycemic stimulus (the insulinogenic index) was significantly depressed by anti-Gil’ serum and was comparable to that observed in rats which received intravenous glucose. The GIP antiserum completely suppressed the incretin effect of the oral glucose, suggesting GIP to be the only incretin. However, the GIP antiserum used in these studies was prepared using a 40% pure GIP preparation which was derived from porcine CCK extracts. It is therefore probable that the antisera preparation also bound other potential incretin factors contained in the impure gut extract.  11 Ebert and coworkers (1979a) tested the effect of intravenous GIP antisera raised to a much purer form of Gil? on intestinal stimuli for IRGIP release in rats. During intraduodenal acid administration along with intravenous glucose infusion, GIP antibodies strongly inhibited the initial increase in insulin levels, although after 20-30 mm, circulating insulin levels were the same as in control rats. Similarly, in 1982, Ebert and Creutzfeldt found the incretin effect of GIP was strongest immediately after an intraduodenal glucose load, and the GIP antisera did not completely block the incretin effect. It was therefore concluded from these studies that GIP was not the exclusive incretin and that additional gut factors with insulinotropic activity existed. Since the nervous system also influences glucose-induced insulin secretion following an oral or intraduodenal glucose load (Porte et al., 1973), the preserved incretin effect after GIP antiserum injection could be related to the neural part of the enteroinsular axis. To circumvent potential neural effects, Ebert et a!. (1983) examined the insulinotropic potency of differently prepared gut extracts. They found that intravenous rat gut extracts exerted insulinotropic activity even after removal of GIP by immunoadsorption. Approximately 50% of the incretin activity of the gut extracts remained suggesting the existence of other hormonal gut factors with insulinotropic activity.  1.5 INSULINOTROPIC ACTIONS OF GIP 15.1 EFFECT OF GIP ON INSULIN SECRETION The observation by Dupré et a!. (1973) that GIP was not insulinotropic under euglycemic conditions suggested glucose played an important role in regulating this action of GIP. It was also established that the insulinotropic action of GIP was glucoseconcentration dependent in vivo in dog (Pederson et a!., 1975b), man (Elahi et a!., 1979) and in the perfused rat pancreas (Pederson and Brown, 1976) where a glucose threshold of approximately 5.5 mM for the insulinotropic action of GIP was observed. This prompted Brown and Pederson (1976a) to suggest that GIP be given the alternate  12 designation Glucose-dependent Insulinotropic Polypeptide. Thus while ingestion of fat has been shown to be a potent stimulus for IRGIP release (see section 1.10.3), no increase in insulin was observed unless intravenous glucose was administered as well (Brown, 1974; Cleator and Gourlay, 1975; Ross and Dupré, 1978). There have even been reports that GIP may attenuate the stimulation of insulin secretion under euglycemic conditions (Opara and Go, 1993). These observations demonstrate that the glucose-dependency of GIP-stimulated insulin secretion provides an important safeguard against hypoglycemia by preventing the inappropriate stimulation of insulin release during a high fat, low carbohydrate meal. Not only was it shown that the insulinotropic action of GIP depended on the presence of a threshold glucose concentration, but it also appeared that glucose could potentiate the action of GIP on the B-cell. At a fixed GIP concentration, increased glucose concentrations stimulated insulin secretion in more than an additive manner (Pederson and Brown, 1976). In the perfused rat pancreas, the maximum potentiating action of glucose on GIP-stimulated insulin release was observed at approximately 16 mM (Brown, 1982). In the presence of 17.8 mM glucose, when GIP was delivered to the rat pancreas in a linear gradient from 0 to 200 pM, the insulinotropic effect of GIP was initiated at concentrations as low as 70 pM (Pederson et a!., 1982), concentrations within the physiological range (see section 1.9.2). It was also noted that the effect of GIP on insulin release from the perfused rat pancreas was dose-dependent (Pederson and Brown, 1976).  15.2 OTHER POTENTIAL INCRETINS  Several other gastrointestinal hormones have been investigated as potential incretins involved in the enteroinsular axis, including gastrin, secretin, CCK, and tGLP-I (Creutzfeldt and Ebert, 1988; Brown, 1988). Kreymann et a!. (1987) showed that tGLP-I was present in the small intestine of man, and observed a sustained increase in plasma  13 levels following an oral glucose load or ingestion of a mixed meal. At postprandial concentrations, Mojsov et al. (1987) showed potent insulinotropic actions of tGLP-I in the perfused rat pancreas. In humans, tGLP-I infusion at physiological concentrations in the presence of an intravenous glucose load significantly enhanced insulin release and significantly reduced peak plasma glucose concentrations. Furthermore, like GIP, the effectiveness of tGLP-I has been shown to be glucose-dependent (Weir et al., 1989). It therefore appears that tGLP-I also fulfills the criteria of an incretin.  15.3 INTERACTIONS BETWEEN INSULIN SECRETAGOGUES Rasmussen and his colleagues have suggested that interactions between various insulin secretagogues occur at the level of signal transduction with the B-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 B-cell, while those that act via the same intracellular pathway tend to have additive stimulatory effects. Evidence for this hypothesis is the observation of potentiating interactions between GIP and acetyicholine (Verchere et a!., 1991) as well as CCK (Sandberg et a!., 1988) in the perfused rat pancreas. GIP appears to act by stimulating cAMP production in the B-cell (see section 1.5.6), while acetyicholine and CCK act via the phosphoinositide pathway and mobilization of intracellular calcium (Prentki and Matschinsky, 1987). On the other hand, activators of cAMP, such as tGLP-I tend to exert additive effects with GIP on insulin release (Fehmann et al., 1989). In addition to interacting at the B-cell with other hormones and neurotransmitters, GIP stimulated insulin release is subject to modulation by nutrients such as arginine. In 197 8, Pederson and Brown demonstrated that 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 OW. At higher concentrations (20 mM) arginine attenuated the insulinotropic effect of GIP, suggesting that GIP and  14 arginine acted on the B-cell via a similar mechanism (Pederson and Brown, 1978). Support for this hypothesis has come from the observation of similar interactions of arginine and GIP on insulin release in vivo in man (Elahi et al., 1982).  15.4 RELATIVE CONTRIBUTION OF GIP vs tGLP-I There have been numerous studies aimed at determining the relative importance of GIP and tGLP-I in the enteroinsular axis. While some studies indicated both peptides were equally effective (Schmid et al., 1990; Kieffer et a!., 1993), other dose-response analyses of the insulinotropic activity of GIP and tGLP-I indicated that lower plasma concentrations of tGLP-I are necessary to augment insulin secretion, particularly at elevated glucose concentrations (D’Alessio et al., 1989; Holst et aL, 1987; Mojsov et a!., 1987; Shima et a!., 1988; Weir et a!., 1989; Nauck et a!., 1989; Krarup et al., 1987c; Kreyman et a!., 1987). On the other hand, while tGLP-I appears more potent than GIP, the rise in IRGIP following oral glucose is greater in magnitude than the increment in plasma immunoreactive tGLP-I after a similar load (Nauck et a!., 1986; Salera et al., 1983; Kreyman et aL, 1987; ørskov and Hoist, 1987; Ørskov et al., 1991; Nauck et a!., 1993). Nauck et al. (1993) examined the incretin effect of Gil’ and/or tGLP-I infused at doses chosen to produce plasma levels roughly comparable to values measured after oral glucose in human subjects (--450 pM and —50-60 pM respectively). It was observed that the doses used were sufficient to augment the B-cell response to intravenous glucose, to values not significantly different from those after oral glucose. It was also noted that GIP made a major contribution to the incretin effect after oral glucose, while tGLP-I appeared to mediate a smaller proportion. Nauck et al. (1993) concluded that GIP and tGLP-I together were sufficient to explain the full incretin effect after oral glucose in normal individuals. Fehmann et a!. (1989) and Suzuki et a!. (1992) noted synergistic stimulatory incretin effects of tGLP-I and GIP on the rat pancreas and concluded that the hormones act “in concert” to guarantee adequate insulin responses to a meal. Kreymann et a!.  15 (1987), found that when human volunteers consumed a test breakfast, plasma IRGIP levels rose more rapidly than tGLP-I, implying that GIP mediated the early insulin response to a meal while tGLP-I was more important in the late response.  15.5 PARTICIPATION OF GJP AND tGLP-I IN AN ENTEROENDOCRINE LOOP The observed differences in IRGIP and tGLP-I secretion rates have been interpreted differently by different investigators. The two insulinotropic peptides GIP and tGLP-I have distinct patterns of distribution in the intestine. GIP cells are located mainly in the upper small intestinal mucosa, found in highest concentrations in the duodenum and, to a much lesser extent, in the ileum and colon (see section 1.8). tGLP-I producing L-cells, on the other hand, are located predominantly in the ileum and colon, with 10- to 40-fold lower concentrations of proglucagon-derived peptides in the duodenum (Eissele et al., 1992). As nutrients rarely reach the ileum before postprandial insulin responses are observed, the role of tGLP-I as a physiological incretin acting from this location has been questioned. In 1991, Roberge and Brubaker made the interesting observation that intestinal proglucagon-derived peptide secretion was stimulated equally by fat in the duodenum as compared to fat in the ileum. It was concluded that duodenal fat either stimulates the enteric nervous system or the secretion of a factor that, in turn, stimulates the release of intestinal proglucagon-derived peptides, including tGLP-I. Brubaker (1991) also reported on the regulation of intestinal proglucagon-derived peptide secretion by intestinal regulatory peptides using a fetal rat intestinal culture model. It was observed that GIP stimulated intestinal proglucagon-derived peptide secretion at physiological concentrations and in a dose-dependent fashion. This prompted the group to hypothesize an ‘enteroendocrine loop’, whereby GIP might stimulate the release tGLP-I before arrival of nutrients in the ileum, allowing tGLP-I to act as an incretin. Recently, Roberge and Brubaker (1993) investigated this hypothesis. They found, as did Kreymann et al. (1987), that the rise in plasma IRGIP levels in response to  16 duodenal nutrients occurred slightly before intestinal proglucagon-derived peptide levels, suggesting a relationship between the two peptides. Intravenous infusion of GIP at a dose sufficient to yield concentrations similar to those observed after duodenal fat administration induced a 2-fold increase in plasma levels of intestinal proglucagon derived peptides that were independent of glycemic levels. No increment in intestinal proglucagon-derived peptides was found in response to infusion of CCK.  It was  therefore concluded that this enteroendocrine loop between the duodenal peptide GIP and the ileal proglucagon-derived peptides may account for some of the early rises in secretion of tGLP-I observed in response to nutrient ingestion. Nauck et al. (1993) however questioned this hypothesis when intravenous GIP infusions failed to reproduce an increase in tGLP-I levels mimicking levels attained after oral glucose in human subjects.  1.5.6 MECHANISM OF ACTION OF GIP ON THEJ3-CELL The mechanism by which GD? stimulates insulin release from the B-cell is not fully understood. This has largely been due to the lack of homogenous preparations of Bcells, which have only recently become available. Using a hamster pancreatic B-cell line (In III), Amiranoff et al. (1984) showed that GIP produced a concentration-dependent increase in cAMP content in the cells which paralleled insulin release. The idea that GIP acted via stimulation of B-cell G-protein coupled adenylate cyclase was later substantiated by studies with cultured rat islets (Siegle and Creutzfeldt, 1985), the rat insulinoma cell line RINm5F (Gallwitz et a!., 1993), and human insulinoma tissue in vitro (Maletti et a!., 1987). The closely related peptide tGLP-I was also shown to increase cAMP levels in a rat islet cell line (Drucker et al., 1987; Göke and Conlon, 1988; Gallwitz et a!., 1993). An alternative pathway for the augmentation of insulin secretion by GD? was suggested by Lardinois et a!. (1990). Using neonatal rat islet cell cultures, they observed that GIP-stimulated insulin release could be suppressed by  17 inhibitors of the membrane associated enzyme phospholipase A2 and intracellular lipoxygenase and cyclooxygenase. It was concluded that GIP exerted its influence in part by modulating membrane associated phospholipase A2 activity, and that the formation of intracellular lipoxygenase products appeared to be a pivotal step in the insulinotropic action of GIP. More recently, Wahi et a!. (1992) investigated the potential for a role of ionic fluxes in pancreatic beta cells as a target for the action of GIP. Using mouse pancreatic islets, it was observed that the amplification of insulin release by GIP did not occur in the absence of Ca 2 in the extracellular medium, and did not involve changes in the cellular content of inositol triphosphate. It was therefore concluded that the amplifying effect of GIP on insulin release was due to an effect on Ca 2 uptake, and subsequent increase in electrical activity. It was also pointed out that an additional contribution of the other islet cell types (c and 6) should also be taken into account, although the mouse islet contains more than 80% B-cells. A potential role for glucagon is suggested by the fact that GIP has been shown to increase glucagon release (Szecowka et al., 1982b), and the observation by Pipeleers et al. (1985) that glucose-induced insulin release from isolated B-cells depended on the cAMP levels, which were markedly increased after addition of cAMP, glucagon, or pancreatic tx-cells. 2 (Bu) Using a hamster B-cell line (HIT T15) Lu et al. (1993) found that both tGLP-I and GIP-stimulated increases in cAMP resulted in increased extracellular Ca 2 influx through voltage-dependent Ca 2 channels. Neither peptide altered phosphoinositide metabolism, further underlining the fact that the mobilization of intracellular Ca 2 from endoplasmic reticulum is not involved in the GIP and tGLP-I signal transduction pathways. Yada et a!. (1993) also observed that tGLP-I potentiated rises in cAMP and resultant increases in the cytosolic free Ca 2 concentration in size selected rat pancreatic B-cells. It was determined that the Ca 2 influx resulted from an enhanced activity of L type Ca 2 channels in the B-cell plasma membrane.  18 15.7 NATURE OF GLUCOSE-DEPENDENCY The glucose-dependency of the insulinotropic action of GIP has been welldocumented in rat, dog, man (Dupré et a!., 1973; Pederson et a!., 1975b; Elahi et al., 1979; Pederson and Brown, 1976; Zawalich et a!., 1993, Siegel et al., 1992), and in 13-cell lines (Kieffer et a!., 1993; Lu et a!., 1993). Other insulin secretagogues such as CCK (Verspohi and Ammon, 1987; Zawalich et a!., 1987), acetylcholine (Garcia et a!., 1988; Verchere et a!., 1991) and tGLP-I (Weir et a!., 1989; Komatsu et a!., 1989) have also exhibited glucose-dependency. The mechanism by which GIP and other insulinotropic agents are glucose-dependent is not fully understood. D-glyceraldehyde, an intermediate of the glycolytic pathway, was able to potentiate the insulinotropic action of GIP in the absence of glucose from the perfused rat pancreas (Brown, eta!. 1981). Mannoheptulose, which blocks glycolysis, abolished GIP-stimulated insulin secretion in the same preparation (MUller et a!., 1982). These studies strongly suggest that glucose metabolism is a prerequisite for GIP-stimulated insulin release to occur. Rasmussen et at. (1990) proposed that GIP-induced increases in cAMP content leading to the conversion of latent to operative Ca 2 channels are ineffective in producing sufficient Ca 2 influx for insulin release unless the B-cell is partially depolarized by the actions of glucose. Postprandial changes in plasma glucose concentrations are thus suggested to primarily serve as conditional modifiers of insulin secretion. Calmodulin, activated by glucose-induced rises in intracellular Ca , has been shown to stimulate 2 adenylate cyclase activity in rat islets (Sharp et at., 1980) and was proposed by Holz and Habener (1992) to be required for tGLP-I stimulated insulin release. Although this may also explain the glucose-dependency of GIP, 2 -Ca calmodulin does not always appear to activate adenylate cyclase activity in islets (Thams et a!., 1982). Furthermore, transgenic mice with elevated levels of B-cell calmodulin develop severe diabetes, even though their pancreatic B-cells contain reserve levels of insulin (Epstein et a!., 1992).  19 Another possible mechanism by which glucose may sensitize the B-cell to further stimulation by insulinotropic agents is through the expression of functional hormone receptors. The transport of glucose, and/or its metabolism resulting in the production of various intracellular messengers could be involved in the expression or allosteric activation of receptors for insulinotropic hormones such as GIP and tGLP-L Thus in the absence of glucose, a reduction in the expression of functional hormone receptors would be expected and it would therefore be predicted that GIP or tGLP-I would not alter cAMP levels; an observation that was recently made using a hamster B-cell line (Lu et al., 1993). If in fact glucose transport is necessary for ‘correct GIP receptor expression, then in cases where glucose transport is altered, a corresponding alteration in the insulinotropic action of GIP would be expected.  Reduced expression of the B-cell high Km glucose  transporters (GLUT2) has been demonstrated in the non-insulin-dependent diabetes mellitus model, the obese Zucker rat (Johnson et al., 1990; Orci et al., 1990). This observation has also been observed in transformed B-cells, combined with the presence of a glucose transporter isoform with a lower Km for glucose (GLUT1; Brant et al., 1992) and alterations in the expression of various hexokinases (Visher et a!., 1987). In both the obese Zucker rat and the transformed B-cell line, BTC3, a significant reduction in the glucose threshold for GIP-stimulated insulin release has been observed (Chan et al., 1984; Kieffer et a!., 1993). Examination of the hypothesis that altered GIP receptor expression may contribute to these observations may soon be possible with the recent cloning of the GIP receptor (Usdin et al., 1993).  1.6 OTHER ACTIONS OF GIP In addition to enterogastrone and insulinotropic actions, GIP has been suggested to have other functions, although whether they are physiological or not is uncertain. This list includes actions on gastrointestinal motility, mesenteric blood flow, intestinal secretion, and release of anterior pituitary hormones (Brown et a!., 1989). The recent  20 discovery of the GIP receptor in tissues outside the gut (brain; see section 1.11.2) suggests that there may be as yet undiscovered actions of this hormone. Potential anabolic actions of GJP are better documented. In the liver, GIP has been demonstrated to inhibit glucagon-induced lipolysis and diminish glucagon-stimulated hepatic glucose production (Hartmann et at., 1986). GIP may also increase glucose uptake in adipocytes and other peripheral tissues such as muscle (Hauner et a!., 1988). These insulin-like actions are compatible with the insulinotropic effect of GIP in promoting glucose utilization. In adipose tissue, GIP has been shown to increase lipoprotein lipase activity, (Eckel et at., 1979),  inhibit glucagon-induced lipolysis (Dupré et at., 1976), and  potentiate the insulin-stimulated incorporation of fatty acids into triglycerides (Beck and Max, 1983). GIP has also recently been shown to stimulate fatty acid synthesis in adipose tissue (Oben et al., 1991). A role for GIP in the control of fat metabolism is not surprising considering the potent stimulatory effect of fat on IRGIP release, which in the absence of glucose, is not insulinotropic (Beck, 1989).  1.7 GIP GENE 1.7.1 GENE STR UCTURE AND POST-TRANSLATIONAL PROCESSING The sequence of the human GIP gene was reported by Takeda et at. in 1987. The cDNA sequence coding the human GIP precursor (Figure 2) indicates that human preproGlP is a protein of 153 amino acids, with a predicted molecular weight of 17,107 daltons. The sequence indicates that proteolytic processing at single arginine residues at either end of the GIP sequence would yield GIP 1-42 (Figure 2).  The structural  organization of human preproGlP was reported by this group to be: putative signal peptide (21 amino acids), NH -terminal peptide (30 amino acids), GIP (42 amino acids) 2 and COOH-terminal peptide, (60 amino acids). Subsequently, Inagaki et al. (1989) revealed that the human GIP gene consists of six exons separated by five introns, with exons 3 and 4 encoding mature human GIP. Recently, rat preproGiP was reported to be  21 144 amino acids in length and comprised of the GIP peptide itself, N- and C-terminal flanking peptides of 22 and 59 amino acids respectively and a typical hydrophobic signal peptide (Sharma et a!., 1992).  5’  -  —70 —90 —80 —60 —50 AGGCTCAGAAGGTCCAGAAATCAGGGGAAGGAGACCCCTATCTGTCCTTCTTCTGGAA  —20 —40 —30 —10 1 10 20 GAGCTGGAAAGGAAGTCTGCTCAGGAAATAACCTTGGAAGATGGTGGCCACGAAGACCTT MetVa lAlaThrLysThrPh 40 30 50 60 70 80 TGCTCTGCTGCTGCTGTCCCTGTTCCTGGCAGTGGGACTAGGAGAGAAGAAAGAGGGTCA eA1aLeuLeuLeuLeuSerLeuPhLuA1aVa1G1yLeuG1yG1uLysLysG1uG1yHi 100 110 120 90 130 140 CTTCAGCGCTCTCCCCTCCCTGCCTGTTGGATCTCATGCTAAGGTGAGCAGCCCTCAACC sPheSerAlaLeuProSerLeuProValGlySerHisAlaLysValSerSerProGlnPr 150 160 170 180 190 200 TCGAGGCCCCAGGACGCGGAAGGGACTTTCATCAGTGACTACAGTATTGCCATGGACAA oArgGlyP roArgTyrAlaGluGlyThrPhelleSerAspTyrSerfleAlaNetAspLy 210 220 230 240 250 260  GATTCACCAACAAGACTTTGTGAACTGGCTGCTGGCCCAAAAGGGGAAGAAGATGACTG $ I1eHisG1nG1nAspPheVa1AsnTrpLeuIeuA1aG1nLysG1yLysLysAsnAspTr 270 280 290 310 300 320 GAAACACAACATCACCCAGAGGGAGGCTCGGGCGCTGGAGCTGGCCAGTCAAGCTAATAG PlaysHisAsnhleThrGlnArgGluAlaArgAlaLeuGluLeuAlaSe rGlnAlaAsnAr 340 330 350 360 370 380 GAAGGAGGAGGAGGCAGTGGAGCCACAGAGCTCCCCAGCCAAGAACCCCAGCGATGAAGA gLysGiuGluGluAlaValGiuP roGlriSerSerP roAlaLysAsnProSerAspGluAs 400 390 410 420 430 440 TTTGCTGCGGGACTTGCTGATTCAAGAGCTGTTGGCCTGCTTGCTGGATCAGACAAACCT pLeuLeuArgAspLeuLeul leGlnGluLeuLeuAlaCysLeuLeuAspGlnThrAsnLe 450 460 470 480 490 500 CTGCAGGCTCAGGTCTCGGTGACTCTGACCACACCCAGCTCAGGACTCGATTCTGCCCTT uCysArgLeuArgSe rArg 510 520 530 540 550 560 CACTTAGCACCTGCCTCAGCCCCACTCCAGAATAGCCAAGAGCCCCCTGT 570 580 590 600 610 TTATGCTAAGTCGAGCCCATTGTGAAATTTATTAAJATGACTACTGAGCACT  -  3’  FIGURE 2: Sequence of a cDNA encoding the human GIP precursor. The numbering of nucleotides starts from the first adenine of the initiating methionine to the last nucleotide just before the poly(A) tract. Residues in the 5’ untranslated region have negative numbers. The deduced amino acid sequence of preproGlP is indicated. The regions of the putative signal sequence and mature GIP are underlined and in bold, respectively. (From Takeda et a!., 1987)  22 1.7.2 GENE REGULATION The mechanism of GIP gene regulation is not completely understood.  The  promoter region of the human GIP gene contains potential binding sites for multiple transcription factors, including Spi, AP-1 and AP-2, but the roles of these sites are unknown (Inagaki et at., 1989). More recently, this same group investigated whether GIP gene expression is regulated by cAMP and glucose. Two cAMP response element (CRE) binding protein (CRE-BP) binding sites were identified in the promoter region (Someya et at., 1993).  CRE is an inducible enhancer of genes which can be transcribed in  response to increased cAMP levels (Comb et at., 1986). CRE-BP1 can dimerize with the c-jun protooncogene product (c-Jun) (Macgregor et at., 1990), and interestingly, c-Jun represses the cAMP-induced activity of the insulin promotor (Inagaki et at., 1992). A possible mechanism of this c-Jun action is by its ability to inhibit the binding of some CRE-BP’s, such as CRE-BP1, to the CREs of the insulin gene by formation of a heteroclimer. These results, and another observation that the level of c-Jun is dramatically increased by glucose deprivation in hamster insulinoma (HIT T15) cells (Inagaki et al., 1992), suggest that glucose may regulate expression of the human insulin gene through CREs and c-Jun. Mutation analysis showed that the two CREs are required for basal promoter activity. Interestingly, the GIP promoter activity was repressed by c-Jun like the human insulin promoter, possibly through the CREs.  1.8 Gil’ CELL DISTRIBUTION AND CHARACTERIZATION Proof of the existence of endocrine cells in the gastrointestinal mucosa came with the use of electron microscopy by Solcia et a!. (1967). Several morphologically distinct cell types were soon identified and correlated with the production of members of the expanding family of gut endocrine peptides. Immunoreactive GIP (IRGIP) cells have been located in the upper small intestine of ruminants (Bunnett and Harrison, 1986), man, pig, dog (Buffa et al., 1975) and rats (Buchan et at., 1982). In the gastrointestinal tract of  23 dog and man, IRGIP is present in cells predominantly in the mid-zone of the glands in the duodenum, and to a lesser extent in the jejunum (Polak et al., 1973). Other studies have found a few IRGIP cells as far as the terminal ileum in rat (Buchan et al., 1982) and man (Fern et aL, 1983). IRGIP cells are part of the amine precursor uptake decarboxylation (APUD) series (Pearse, 1968). Ultrastructural studies of human IRGIP cells indicated a characteristic appearance of the K-cell; intracellular secretory granules having a small electron-dense core surrounded by a concentric electron-lucent halo (Buchan, 1978). In the dog, however, IRGIP cells identified in the duodenum contained uniformly electrondense secretory granules consistent with the cell type recognized as the I-cell of the endocrine cell classification (Usellini et al., 1984). Thus some species specificity exists regarding the ultrastructure of the GIP cell. Some immunocytochemical studies indicated that GIP was co-localized with glucagon in the pancreatic a-cells in mammals (Smith et at., 1977; Alumets et at., 1978; Ahrén et a!., 1981). Other studies indicated, however, that the staining of the pancreatic a-cells could be blocked by preincubation of the antisera with glicentin (Larson and Moody, 1980) or glucagon (Buchan et al., 1978), suggesting that these results were due to cross-reactivity of the GIP antisera used with glucagon or the glucagon-like peptides and their precursors. Furthermore, studies using RNA blot analysis have only detected the presence of human preproGlP mRNA in human intestine, and not in the pancreas (Takeda et at., 1987; Inagaki et at., 1989; Sharma et a!., 1992). It is interesting to note that IRGIP has been localized to a distinct cell type in the islet organ of elasmobranchs, although its function there is not known (Hazeiwood, 1989). It appears that GIP cells are not found phylogenetically in the gut mucosa until a distinct islet organ evolves upon which the peptide may act, i.e. the hagfish, suggesting the possibility of a functional enteropancreatic axis at this evolutionary level (Falkmer et a!., 1980).  24 1.9 MEASUREMENT OF GIP 1.9.1 DEVELOPMENT OF GIP RADJOIMMUNOASSAY With the purification of GIP came the possibility of development of techniques for measuring endogenous hormone secretion.  This was  a prerequisite for the  establishment of GIP as a hormone, in order to demonstrate the polypeptide in blood and tissues and release from tissues by physiological mechanisms. The development of a radioimmunoassay (RIA) for GIP was first achieved in 1974 by Kuzio et al. Antisera were produced in guinea pigs by subcutaneous immunization with GIP emulsified with Freund’s adjuvant, followed by boosting with GIP conjugated to bovine serum albumin using the carbodiimide method. Antisera were shown to have no measurable cross reactivity with natural secretin, synthetic glucagon, synthetic human gastrin, pure porcine CCK, pure porcine motilin, and pure porcine vasoactive intestinal peptide. GIP was labeled with  125J  by a minor modification of the chloramine-T technique, and purified on  a Sephadex G-15 column. The RIA developed had a sensitivity range of 25 to 250 pg, and Kuzio et al. (1974) reported a mean fasting serum concentration of 237 ± 14 pg/mi and >1200 pg/mi after a meal.  1.9.2 PHYSIOLOGICAL LEVELS OF GIP Subsequent to the development of the first GIP RIA, other assays followed (Moody and Lauritsen, 1977; Morgan et al., 1978; McLoughlin et al., 1979; Ebert et al., 1979b; Sarson et al., 1980; Burhol et al., 1980a; Wolfe and McGuigan, 1982; Jorde et al., 1983; Sheu et al., 1987; Wishart et al., 1992; Moody et a!., 1992). There appeared to be a general agreement on the pattern of IRGIP release after ingestion of nutrients, whereas the absolute values measured differed widely, ranging from 9 to 80 pmol/l in the fasting state, and from 35 to 700 pmol/l after meals. In 1983, Jorde et al. attempted to resolve this disparity by measuring fasting and postprandial IRGIP values in man with seven different antisera, including two of their own. Even under the same assay conditions,  25 mean fasting IRGIP levels ranged from 12 to 92 pmol/l, and the mean postprandial IRGIP values ranged from 35 to 235 pmol/l. It was concluded that this variability was a result of different cross reactivities of the antisera with human GIP. Porcine GIP (two amino acids different) of varying purities had been used as standards and to raise antisera for RIAs used to measure IRGIP in man. Amland et a!. (1984) measured fasting and postprandial IRGIP values in pigs, rats, dogs and man with five different antisera and porcine GIP standards and tracers. The mean IRGIP values in rats, dogs, and man varied considerably, depending on the antiserum used, whereas all the antisera recorded fairly similar IRGIP values in pigs. These findings again demonstrated the species differences in immunological properties of GIP, and thus the importance of using GIP from the appropriate species as standards. The sequencing of human GIP and its successful production by both chemical synthesis (Yajima et a!., 1985; Fujii et al., 1986) and recombinant DNA techniques (Chow et a!., 1990) should have allowed the proper evaluation of IRGJP concentrations in man. Oddly enough, only two groups at present appear to have used human GIP in assays employed for determining plasma IRGIP levels in man. Kreymann et al. (1987) used synthetic human GIP as standards for the measurement of IRGIP in human subjects in response to oral glucose and a test meal. Basal values were 47± 16 (mean ± SEM) and peaked at 174 ± 24, 30 mm after receiving oral glucose. The material used for tracer production was not stated.  Using synthetic human GIP for tracer and standards, Nauck et a!. (1992)  measured basal IRGIP levels in human subjects as 69 ± 18 pM, and peak IRGIP concentrations after oral glucose of 340 ± 39 pM. It is not surprising that these peak levels are higher than others previously reported in response to oral glucose.  1.9.3 IMMUNOREACTIVE FORMS OF GIP In 1975, Brown et a!. reported that the antisera used in the RIA developed by Kuzio et a!. (1974) identified more than one molecular form of IRGIP in plasma and gut  26 tissue. In fact, all seven of the antisera tested by Jorde et al. (1983) recognized three molecular forms of IRGIP to variable degrees, likely contributing to the marked discrepancies in reported IRGIP levels. When subjected to gel filtration, one high molecular weight component eluted in the void volume of the column, while the two other components eluted at positions corresponding to molecular weights of approximately 8000 and 5000 daltons (Brown et a!., 1975). The 5000 IRGIP form is the major component of GIP in both blood and tissues and has been used for the establishment of RIAs (see Figure 1 for sequence). After gel filtration of extracts of porcine and human small intestine, a fourth IRGTP form was found eluting between 8000 IRGIP and void volume (Krarup and Holst, 1984). This fourth form has never been detected in plasma. Krarup and Hoist (1984) concluded that since all five antisera tested reacted with this IRGIP form it may represent a prohormone to 5000 IRGIP. -IRGIP) has been detectable after gel filtration of 0 No void volume IRGIP (V extracts of gut mucosa (Bacarese-Hamilton et al., 1984; Krarup and Hoist, 1984). This component could be removed from plasma by alcohol extraction, indicating its association with high molecular weight constituents of plasma (Krarup, 1988). This suggestion had been made by Dryburgh (1977), when it was noted that pre-treatment of serum by boiling or by addition of 6 M urea diminished V -IRGJP and increased 5000 0 IRGIP. In addition, V -IRGIP does not appear to contribute to the increase in IRGIP in 0 response to glucose, fat, or a meal (Krarup et a!., 1985; Krarup et a!., 1987b). Krarup (1988) concluded that V -IRGIP may represent large molecular weight proteins 0 interfering in the assays. While 5000 IRGIP was shown to increase in plasma 60 mm after intraduodenal glucose or fat had been given in man, 8000 IRGIP showed small and inconsistent changes after both stimuli (Krarup et al., 1985; Krarup et a!., 1987b). It was concluded by Krarup (1988) that 8000 IRGIP contributed little to the increase in IRGIP after glucose and fat, and only modestly to that observed after a meal. A partially purified 8000 IRGIP from  27 porcine gut did not stimulate insulin release even when used at pharmacological doses in an isolated porcine pancreas system (Krarup et al., 1987a). Otte et al. (1984) found that the entire sequence of GIP was not contained in an 8000 IRGIP form extracted from pig gut and concluded that it was not a proform of GIP. Since 8000 IRGIP cross reacts with most GIP antisera (Jorde et a!., 1983; Krarup, 1988) there must be a degree of homology between this form and 5000 IRGIP. It is possible, therefore, that 8000 IRGIP is derived from a GIP precursor.  1.10 GIP SECRETION 1.10.1 RESPONSE TO A MIXED MEAL It is generally agreed that IRGIP levels increase approximately six times above basal in response to ingestion of a mixed meal, although absolute values vary significantly (Kuzio et al., 1974; Morgan et a!., 1978; Jorde et a!., 1980; Jorde et al., 1983; Amland et a!., 1984). Plasma levels of IRGIP remained elevated for up to six hours after a meal, and thus during a 24 h period, with three meals ingested, IRGIP was significantly high all day (Jorde et a!., 1980; Salera et al., 1983; Jones et al., 1985). Beck et a!. (1984) observed that ingestion of a meal with a higher caloric level yielded a significantly greater IRGIP response, even though blood glucose levels were not significantly different. It was concluded that this was an adaptation to the greater amount of food ingested, and that the GIP producing cells have a rapid mechanism of adapting to the ingested caloric load.  Service et a!. (1983) also observed a highly significant  association between meal size and post-prandial plasma IRGIP responses. Oektedalen et al. (1983), compared the IRGIP response to a meal in previously fasted and fed individuals. It was observed that the IRGIP response to oral glucose of a test meal was augmented after food deprivation. It was suggested that the increased catecholamine levels associated with fasting may have caused the increased IRGIP release, as the adrenergic nervous system has been shown to modify the IRGIP release  28 (see section 1.10.5). Hampton et a!. (1983) found the IRGIP response to a meal in rats fed a high fat diet for 4 days prior was greater than controls. Deschodt-Lanckman et al. (1971) showed that pancreatic lipase activity in the rat could be stimulated by a high fat diet. This increased lipase activity might result in an increase in the rate of absorption of fat, and thus a greater IRGIP response (section 1.10.3). Morgan et a!. (1988), however, only noted a small but significant increase in IRGIP release by human subjects previously on a high fat diet for 35 days, suggesting that species differences are likely to exist.  1.10.2 RESPONSE TO CARBOHYDRATES Evidence supporting the role of GIP as an incretin came with the first observation of its release in response to oral glucose, followed almost immediately by an increase in plasma insulin levels (Cataland et al., 1974). A direct relationship between the glucose load ingested and the release of IRGIP was observed in dog (Pederson et al., 1975b) and man (Falko et al., 1980). The fact that luminal stimulation by glucose was necessary for IRGIP release was proven when no changes in circulating levels of this hormone were observed when glucose was administered intravenously (Cataland et a!., 1974; Pederson et a!., 1975b; Cleator and Gourlay, 1975; Andersen et a!., 1978). The digestion of complex carbohydrates occurs in the gut through the action of pancreatic enzymes and brush border hydrolases such as sucrase and maltase. Final products of digestion, D-glucose, D-galactose, and D-fructose, can then be absorbed by enterocytes lining the upper third of the intestinal villi. The absorptive process occurs in the duodenum and jejunum and is generally complete before the chyme reaches the ileum. The pyranoses are absorbed by a two-stage process. Glucose and galactose enter epithelial cells by an energy requiring Na-dependent transporter, SGLT1 (Wright, 1993). In contrast, fructose is absorbed down its concentration gradient by the facilitative glucose transporter, GLUT5 (Burant et a!., 1992). The second step is the downhill transport of sugar out of the enterocyte across the basolateral membrane into the blood,  29 by the facilitated sugar transporter GLUT2 (Thorens, 1993). During some point in this process, endocrine cells release GIP; the actual step that triggers this secretion is however not understood. In an attempt to determine the carbohydrate specificity for GIP release, Morgan (1979) studied the IRGIP responses to the sugars glucose, galactose, sucrose, and fructose given as an oral carbohydrate tolerance test in human subjects, or perfused through the rat small intestine. In both models, IRGIP was released by glucose, galactose and sucrose, but not fructose. It was also noted that the response to sucrose was significantly delayed relative to the rise in response to glucose and galactose. The results indicated that IRGIP secretion was dependent on the active transport of monosaccharides. The delay in IRGIP response to sucrose was explained by the fact that sucrose must be hydrolysed into fructose and glucose by a brush border hydrolase prior to absorption. Similar results were reported by Sirinek et al. (1979; 1983) from studies of intraduodenal infusion of carbohydrates in dogs. Sirinek et a!. (1983) suggested, on the basis of their data, that the structural integrity of the glucose molecule from the C-i to C-4 carbon atom, a free aldehyde group on the C-i carbon atom and a cyclic structure are all necessary for both the active transport of glucose and the release of GIP. Further evidence for the necessity of active sugar transport of monosaccharides for IRGIP release has been supplied by studies in which transport is blocked. Morgan (1979) observed that reduction of the post-prandial rise in blood glucose by guar (a gum produced from the seeds of the Indian Cluster bean, Cyonopsis tetraagonoloba) resulted in a significant reduction in IRGIP release.  Sykes et al. (1980) blocked the Na  dependent carrier protein with phloridzin, a 8-glucoside that binds but is not transported (Hopfer, 1987) and found that glucose-stimulated IRGIP release in rats was abolished. This group also provided further information on the minimal structural requirements for the release of IRGIP in response to carbohydrates. Of the numerous monosaccharides tested, only those with the structural requirements for active transport by the Na  30 dependent hexose pathway (Hopfer, 1987) were effective as stimulants of IRGIP release. Furthermore, their observations that x-methylglucoside and 3-O-methylglucose stimulated significant IRGIP release indicated that transport out of the cell, or metabolism, was not necessary for IRGIP release, respectively. While similar results were obtained by Flatt et al. (1989) with mice, others have observed no IRGIP response to 3-O-methylglucose in rats (Fushiki et a!., 1992). While the majority of data seem to suggest that the Na-dependent glucose transporter was involved in the mechanism by which GIP cells ‘sense’ glucose, the evidence was not definitive. Unlike Morgan (1979) and Sykes et al. (1980), Flatt et al. (1989) observed significant IRGIP release by fructose. In addition, Flatt et a!. (1989) demonstrated a delayed but large IRGIP response by 2-deoxyglucose, which is not transported by the glucose or fructose luminal transport system (Hopfer, 1987). Sykes et a!. (1980), however,  found no significant response to luminal perfusion of 2-  deoxyglucose in the rat. Whether these observations indicate species differences or a mechanism other than the Na-dependent glucose transporter for GIP release is not clear. In hopes of better characterizing the mechanism by which GIP cells ‘sense’ glucose, Fushiki et a!. (1992) studied IRGIP release into the portal vein of rats in response to duodenal infusion of glucose in the presence of inhibitors of some putative glucose sensors and carriers in the intestinal lumen. Gymnemic acid, the active principle of Gymnema sylvestre leaves, is a glucuronide of triterpene which can inhibit Na dependent active glucose transport in the small intestine (Yoshioka, 1986). Fushiki et a!. (1992) found that like phloridzin, gymnemic acid markedly suppressed the increase in portal IRGIP in response to intraduodenal glucose infusion.  In the presence of  cytochalasin B, a competitive inhibitor of facilitative glucose transporters, there was no significant change in the portal IRGIP concentration after glucose administration, suggesting that a facilitative glucose transporter was not directly involved in the IRGIP release by the small intestinal endocrine cells (Fushiki et a!., 1992).  31 Fushilci et at. (1992) investigated the possibility that IRGIP release by glucose occurred by a mechanism similar to that by which glucose acts to release insulin from the pancreatic B-cell. Glucokinase has been demonstrated to be a key enzyme for glucose recognition by B-cells as well as liver cells (Matschinsky, 1990).  Infusion of  mannoheptulose, an inhibitor of glucokinase (Matschinsky, 1990), did not affect the portal vein plasma IRGIP concentration (Fushiki et a!., 1992). Glybenclamide, a potent insulin secretagogue acting via closure of the ATP-sensitive K channels (Trube et a!., 1986), also had no effect on portal vein plasma IRGIP concentration. These findings suggested that the mechanism underlying GIP release in response to luminal glucose in the small intestine was different from that underlying insulin release in response to plasma glucose in the pancreas. A final putative mechanism for glucose ‘sensing’ by GIP cells was mediation by a neural pathway (Fushiki et at., 1992). In 1978, Mei found a glucoreceptor existed in the duodenum and proximal jejunum, which transmitted via vagal neurons. During recording from nodose ganglia by means of extracellular glass microelectrodes, Mei (1978) noted that glucose and galactose preferentially stimulated the sensor, but fructose did not. Furthermore, this duodenal glucose sensor could be blocked by 1% procaine infusion. Fushiki et a!. found that infusion of 10 g/l procaine and 10 g/l lidocaine in saline did not affect the glucose-induced portal vein plasma IRGIP concentration, suggesting that the procaine-sensitive vagal glucoreceptor was not involved in IRGIP release. Overall, the present data suggest that glucose-mediated GIP release does not occur as a result of the direct actions of the duodenal procaine-sensitive vagal glucose sensor, a cytochalasin B sensitive facilitative glucose transporter, a mannoheptulose sensitive glucokinase, or the ATP-sensitive K channels.  It appears that neither glucose  metabolism, nor glucose transport out of the cell are necessary for GIP release, although this requires further verification. It seems glucose transport via a phloridzin sensitive Na-dependent transporter like SGLT1 is involved in glucose recognition leading to GIP  32 release but some data indicate a specificity different from that known for SGLT1. Furthermore, it is not known whether the GIP cells directly ‘sense’ glucose, or whether messages are transmitted by neighbouring glucose-transporting mucosal cells. Sykes et a!. (1980) postulated that tight junctions between the two types of cells allowing changes in the flux of sodium ions generated in the mucosal cells by interaction between the stimulatory monosaccharides and their transport proteins might offer a possible explanation for their ability to stimulate release of IRGIP.  1.10.3 RESPONSE TO FATS Triglycerides form the primary lipids of a normal diet. Their digestion occurs primarily in the duodenum and upper jejunum, where pancreatic lipases act on fat emulsified by bile. In the intestinal lumen, triglycerides are converted to free fatty acids and monoglycerides, form micelles, and then diffuse among the microvilli that form the brush border. The high lipid solubility of the fatty acids and monoglycerides promotes their diffusion into the enterocytes. The subsequent fate of the fatty acids depends on their size. Fatty acids containing less than 10-12 carbon atoms pass from the mucosal cells directly into the portal blood, where they are transported as free (unesterified) fatty acids. Larger fatty acids are reesterified with monoglycerides to re-form triglycerides, which accumulate in chylomicra and are ejected by exocytosis Out of the cell, into the lymph. In order for GIP to fulfill the role of an enterogastrone, it is necessary to establish its release by some stage of this fat digestion process. The first report of IRGIP release in response to fat was made in 1974 by Brown shortly after the development of a GJP radioimmunoassay (Kuzio et al., 1974). Brown observed a peak in serum IRGIP levels in human volunteers approximately 2 h after ingestion of 100 ml corn oil suspension (Lipomul). A similar response was noted by Failco et at. (1975), and Cleator and Gourlay (1975), and these investigators also reported no increase in serum glucose or insulin concentrations. These studies showed quite  33 conclusively that endogenously released IRGIP was not insulinotropic in the absence of hyperglycemia. Pederson et a!. (1975b) studied the IRGIP response to graded oral doses of fat and glucose in dogs, and found triglycerides produced a greater and more prolonged IRGIP response (5 h) than the glucose loads (2.5 h). It was suggested that the rate of gastric emptying might be a contributing factor to these observations. Pederson et a!. (1975b) pointed out that the release of IRGIP by fat and the nature of this response fit well with an enterogastrone-like action, whereas the rapid response to oral glucose was more relevant to the potentiation of a rapidly rising insulin response. In 1981, Ross and Shaffer found that the process of hydrolysis was necessary for GIP release. This came from the observation that there was no IRGIP release from patients with cystic fibrosis and defective fat lypolysis following ingestion of corn oil, but that it returned to levels similar to controls with coingestion of pancreatic enzymes. A similar conclusion was made by Ohneda et al. (1983) when it was observed that an IRGIP response to fat in pancreatectomized dogs occurred only during coingestion of pancreatic enzymes. Imamura et at. (1988) found a reduced IRGIP response to ingestion of a fat-enriched meal in patients with disturbed fat metabolism resulting from external biliary drainage. Interestingly duodenal infusion of bile alone has been demonstrated to increase IRGIP release (Burhol et a!., 1980b). Thus emptying of the gallbladder in response to the presence of fat in the duodenum may enhance the release of IRGIP in response to fat hydrolysis. The suggestion that fatty acids were the product of triglyceride hydrolysis responsible for IRGIP release came from observations that monoglycerides had no effect on IRGIP levels (Ross and Shaffer, 1981; Ohneda eta!., 1984; Williams eta!., 1981). In 1976, O’Dorisio et a!. reported only a modest IRGIP response to an intraduodenal infusion of medium chain triglycerides in dogs, when compared to the IRGIP release previously obtained from long chain fatty acids (Brown, 1974; Falko et a!., 1975). Other studies on the effects of fatty acid chain length on IRGIP release have since confirmed  34 that long chain fatty acids exert strong stimulation, while medium or short chain fatty acids do not (Ross and Shaffer, 1981; Ohneda et a!., 1984 Kwasowski et al., 1985). In addition to fatty acid chain length, it also appears that the degree of saturation might affect their ability to stimulate IRGIP release. Lardinois et a!. (1988) compared the IRGIP response to saturated, monosaturated and polyunsaturated fats in man, and found the latter yielded the smallest IRGIP secretion. As short and medium chain fatty acids are absorbed faster than long chain, GIP release is not related only to the rate of cellular uptake of fatty acids. Kwasowski et a!. (1985) suggested that the stimulus for fat-induced GIP release might be generated during the intracellular handling and metabolism of fatty acids. Short- and medium-chain fatty acids are transferred across the intestinal epithelium without esterification. However, long-chain fatty acids are conveyed to the smooth endoplasmic reticulum for esterfication prior to incorportation into chylomicrons and eventual exocytosis into the intercellular compartment. The GIP-releasing action of fatty acids may be coupled therefore to the extent of esterification, an energy-consuming metabolic process confined to long-chain fatty acids. Previous studies on the mechanism of carbohydrate-stimulated GIP release have shown that only actively transported sugars stimulate IRGIP secretion (see section 1.10.2). Thus there may be a common link between metabolic and secretory events responsible for nutrient-regulation of GIl? release from the intestinal GIP cells.  1.10.4 GIP RESPONSE TO PROTEINS The majority of protein digestion occurs in the duodenum and jejunum, where pancreatic proteases such as trypsin, chymotrypsin, and carboxypeptidase act to convert dietary protein to small peptides. The brush border of the upper small intestine contains a number of peptidases, including aminopeptidases, dipeptidases and dipeptidyl aminopeptidases, which further reduce the peptides to oligopeptides and amino acids. The small peptides and amino acids are then transported across the brush border plasma  35 membrane into the epithelial cells. Transport of dipeptides and tripeptides across the brush border plasma membrane is a secondary active transport process powered by the electrochemical potential difference of Na across the membrane. Normally amino acids are transported into the enterocyte by way of certain specific amino acid transport systems, some dependent on Na+ and others not. Once in the enterocyte, transport systems of the basolateral membrane then transport amino acids out of the cell, again with some dependent on Na, and others not. Brown (1974) reported that ingestion of a meat extract did not result in a significant release of IRGIP. A similar result was observed by Cleator and Gourlay (1975) with either beef steak or a meat extract, while a small but significant response to leucine was noted. Thomas et al. (1976) were able to demonstrate that intraduodenal administration of a mixture of amino acids elevated serum IRGJP concentrations (peaking at 30 mm) in association with an increased insulin response. More specifically, they showed that a mixture containing arginine, histidine, isoleucine, lysine and threonine caused a marked rise in integrated IRGIP and insulin secretion, while a combination of methionine, phenylalanine, tryptophan and valine had only a minimal effect on release of these two hormones.  It was suggested by Thomas et a!. (1978) that the higher  concentration of amino acids as compared to the previous studies was responsible for the observed IRGIP response, as concentrations of 10 or 20 mM in the small intestine demonstrated no effect.  O’Dorisio et at. (1976) demonstrated a significant IRGIP  response with a mixture of amino acids totaling —38 mM. In 1982, Schulz et al. examined the IRGIP response to amino acids in rats treated with a corticosteroid or alloxan. These compounds are thought to increase Na K ATPase activity, and thus the Na-dependent active absorption process (Schulz et a!., 1982). A mixture of amino acids, including those shown by Thomas et at. (1976) to stimulate IRGIP release, were also found to be stimulatory by Schulz et a!. (1982), and, furthermore, both groups of treated rats showed significantly higher serum IRGIP levels.  36 These results were attributed to an increased absorption of the amino acids owing to an increased Na K ATPase activity in the intestinal mucosa of corticosteroid- and alloxan treated rats. These data therefore suggested that as with glucose or fat, the release of GIP after amino acid stimulation seemed to depend more on an active absorption than on their presence within the intestinal lumen. In 1982, Wolfe and McGuigan measured the IRGIP response following a peptone meal in the dog. The increase in IRGIP release in response to the peptone meal in this study was maximal in peripheral venous serum at 15 mm. This is similar to the response to glucose feeding and contrasts with the more prolonged response to fat ingestion. IRGIP release has not been demonstrated previously after feeding or infusion of a protein of peptone meal that is free of carbohydrate or fat. It was suggested that these differences might result from the sensitivity of the GIP radioimmunoassay used in this study which measured basal and peak IRGIP levels at 34 ± 10 and 323 ± 95 pg/ml, respectively; values lower than those previously reported. No significant increase in insulin was noted, consistent with the glucose-dependent action of GIP. Wolfe and McGuigan (1982) suggested that their demonstration of GIP release after a peptone meal accompanying peptone-stimulated gastrin release and gastric acid secretion was supporting of the enterogastrone role of GIP.  1.10.5 A UTONOMIC NERVOUS SYSTEM CONTROL OF GIP RELEASE Conflicting data, both in human and animal studies, exist regarding the involvement of cholinergic and adrenergic factors in GIP secretion. Vagotomy has been demonstrated either to enhance (Thomford et al., 1974; Imamura et al., 1984; Yoshiya et al., 1985), reduce (Lauritsen et al., 1982; Imamura et al., 1984), or to have no effect (Gayle and Ludewig, 1978; Yoshiya et al., 1985) on the IRGIP response to nutrient ingestion. Sham feeding, which produced cephalic-vagal stimulation, had no effect on IRGIP release (Taylor and Feldman, 1982), as did direct vagal stimulation (Berthoud et  37 al., 1982). Studies with atropine have demonstrated either reduced (Baumert et al., 1978; Larrimer et al., 1978) or no effect (Sirinek eta!., 1981; Nelson eta!., 1986) on the IRGIP response to nutrient ingestion. Finally, acetylcholine decreased the IRGIP response to intragastric glucose administration (Williams and Beisbroeck, 1980). As with parasympathetic stimulation, controversy also exists on the effect of the sympathetic nervous system on GIP release. Epinephrine has been shown to inhibit (Williams and Biesbroeck, 1980) or have no effect (Sirinek et al., 1977b) on glucosestimulated IRGIP release. Frier et a!. (1984) examined the influence of adrenergic denervation on the IRGIP response to feeding in man.  There was no significant  difference in meal-stimulated IRGIP release in sympathectomized subjects (with a complete transection of the spinal cord above the sympathetic outflow) compared to controls. Blockade of cz-adrenorecetors has been shown to augment (Sirenek et a!., 1977b) or have no effect on nutrient stimulated IRGIP release (Flaten, 1981), while 13adrenoreceptor antagonists seem to inhibit nutrient stimulated IRGIP release (Sirinek et al., 1977b; Flaten, 1981). Stimulation of x-adrenoreceptors has been shown to reduce the IRGIP response to oral glucose (Salera et al., 1982a), while B-adrenoreceptor stimulation has been shown to increase basal and glucose stimulated IRGIP release (Flaten et a!., 1982; Kogire et a!., 1990), or have no effect (Salera et a!., 1982a). The multiple effects of the autonomic nervous system on the gastrointestinal tract probably conthbute to the difficulty of establishing the role of its components on IRGIP release. Alterations in gastric emptying, intestinal secretion and absorption of fluids, motility, and blood flow by the autonomic nervous system are all likely to influence IRGTP release. It is therefore difficult to make inferences on the direct action of the  sympathetic or parasympathetic nervous system on GIP cells.  38 1.10.6 OTHER MODULATORS OF GIP RELEASE Somatostatin appears to be an effective inhibitor of GIP release. Pederson et al. (1975a) demonstrated that a rapid intravenous injection of somatostatin in dogs delayed the increase in serum concentrations of insulin, JRGIP and glucose following oral glucose. During a sustained somatostatin infusion, the IRGIP response to oral glucose was suppressed until cessation of the infusion, even though serum glucose levels were elevated to the same levels as controls. This suggested that inhibition of glucosestimulated IRGIP release was not a result of the inhibition of glucose absorption. This hypothesis was confirmed by Creutzfeldt and Ebert (1977), who demonstrated that IRGIP and insulin responses to a test meal were completely suppressed during somatostatin infusion, whereas blood glucose levels were not different from controls. Kraenzlin et a!. (1985b) demonstrated that a long acting somatostatin-analogue, SMS 201 995, also effectively suppressed the postprandial release of IRGIP. Intravenous gastrin has been demonstrated to increase the IRGIP response to intraduodenal glucose in dogs (Sirinek et at., 1977a). This augmentation could not have been caused by the stimulation of gastric H secretion, since acidification of the duodenal glucose did not augment the IRGIP release. Jorde et a!. (1981) further examined this putative effect of gastrin by examining the IRGIP response to a mixed liquid test meal in a group of achiorhydric patients with high serum gastrin levels compared to a normal group of human subjects. A significantly higher IRGIP response was observed in the achlorhydric group than in the control group. Studies on patients after total antero duodeno-pancreatectomy that have undetectable levels of serum gastrin show a significant IRGIP response to a test meal, indicating that gastrin is not a prerequisite for GIP release (Creutzfeldt et al., 1976). However, these patients have a very rapid gastric emptying, which may explain their elevated serum IRGIP levels. Evidence for a relationship between GIP and gastrin has also been supplied by Morgan et a!. (1985).  These investigators observed a highly significant negative  39 correlation between circulating gastrin and IRGIP levels after control and guar gum meals. Addition of guar gum to an oral glucose load reduced the postprandial secretion of both IRGIP and insulin. Studies in dogs by Wolfe et al. (1983) using anti-GIP antibodies have shown that IRGIP can function as a physiological inhibitor of gastric acid secretion through its effect on gastrin release. Morgan et al. (1985) therefore suggested that the guar gum attenuated IRGIP response to the protein meal, leading to an attenuated somatostatin response and as a result unrestrained gastrin secretion. Exogenous GIP infusions have been shown to inhibit gastrin release (Villar et al., 1976; Arnold et al., 1978), the effect seemingly mediated via an increase in gastric somatostatin secretion (McIntosh et al., 1979). In addition to gastrin, other compounds have also been demonstrated to alter IRGIP release. Increases in IRGIP levels have been noted in response to intravenous administration of the neuropeptide gastrmn-releasing peptide (McDonald et al., 1981; Greely et al., 1986a; 1986b), while the neuropeptide calcitonin gene-related peptide inhibited basal levels of IRGIP in man (Kraenzlin et a!., 1985a). A reduction in postprandial IRGIP release has also been noted in response to intravenous morphine (Champion et a!., 1982) and oral administration of the prostaglandins enprostil or rioprostil (Nicholl et a!., 1986; Demol and Wingender, 1989; Schwartz and Saito, 1989). It is not clear whether these actions result from direct interactions with GIP cells, or indirectly by slowing gastric emptying, reduced intestinal nutrient absorption, or by the action of another mediating peptide or neurotransmitter.  1.10.7 FEEDBACK INHIBITION OF GIP RELEASE In 1975, Brown et a!. proposed a feedback inhibitory control mechanism of GIP release involving insulin.  This hypothesis was based on their observation that an  intravenous bolus injection of insulin during ingestion of fat significantly reduced the triglyceride-induced IRGIP increment.  Support came from studies demonstrating  40 attenuated IRGIP responses to oral fat ingestion during simultaneous infusion of glucose, and, thereby, stimulation of endogenous insulin release (Cleator and Gourlay, 1975; Crockett et al., 1976; Ross and Dupr, 1978; Ebert et at., 1979a). In 1978 Sirinek et at. suggested that insulin was also capable of attenuating the GIP response to oral glucose. However, Andersen et at. (1978) noted that during a euglycemic clamp with a continuous insulin infusion resulting in hyperinsulinemia, a significant IRGIP response to oral glucose was noted. Andersen et at. concluded that glucose-stimulated GIP release was not inhibited in the presence of marked hyperinsulinemia. This finding was supported by the observations that hyperinsulinemia induced by exogenous insulin (Service et at., 1978), or endogenous insulin (Collier et at., 1984) had no effect on carbohydratestimulated IRGIP secretion in normal subjects. Furthermore, in juvenile diabetics without insulin reserve, fat-, but not glucose-induced IRGIP secretion was reduced by exogenous insulin infusion (Creutzfeldt et at., 1980). In a recent investigation with rats, however, Bryer-Ash et at. (1994) observed that the IRGIP response to oral glucose was suppressed by hyperinsulinemia, and this suppression was attenuated when hyperinsulinemia was accompanied by hyperglycemia. It is therefore possible that species differences may exist in the insulin feedback inhibition of GIP release by oral glucose. Verdonk et a!. (1980), using the glucose insulin clamp technique, questioned the effectiveness of insulin as an inhibitor of fat-stimulated IRGIP secretion. These authors used hypoglycemic, euglycemic, and hyperglycemic clamp conditions and found that insulin, under euglycemic conditions, did not exert an inhibitory effect on fat-stimulated IRGIP levels.  Since in their study, the hypoglycemic clamp increased and the  hyperglycemic clamp decreased the IRGIP response to oral fat, they suggested an effect of glycemia itself on IRGIP secretion in the presence of hyperinsulinemia. However, Stöckmann et a!. (1984) postulated that the 2 h hyperinsulinemic period preceding ingestion of fat in the study by Verdonk et a!. may have prevented the demonstration of  41 the insulin effect on fat-induced IRGIP release. Stöckmann et a!. (1984) found that the response of IRGIP to oral fat was inhibited by 63% if insulin infusion was started at the time of fat ingestion, whereas no inhibition was seen if a 2 h hyperinsulinemic period proceeded the fat load. It was therefore concluded that insulin does inhibit fat-induced IRGIP secretion in normal man, but prior hyperinsulinemia masks this insulin effect, probably by decreasing the sensitivity of the GIP cells to insulin. Recently, Takahashi et al. (1991) compared the effect of insulin and glucose on fat-induced GIP and tGLP-I release in humans. The response of both hormones to fat ingestion was measured during Continuous glucose infusion and during a hyperinsulinemic euglycemic glucose clamp.  The release of GIP and tGLP-I was  suppressed in the hyperglycemic, hyperinsulinemic state. However, while GIP was also suppressed in the normoglycemic hyperinsulinemic state, the release of tGLP-I was not. Takahashi et a!. (1991) thus concluded, as previously reported, that insulin inhibited fatinduced GIP, but the secretion of tGLP-I was more likely inhibited by a direct action of glucose. Dryburgh et al. (1980a; 1980b) investigated the possibility that C-peptide might inhibit the release of IRGIP. At a dose three-times that found in the fed rat, C-peptide totally abolished the IRGIP response of the perfused rat intestine to fat stimulation. It was also demonstrated that C-peptide released after stimulation of the pancreas by glucose and tolbutamide administration given intravenously in combination with insulin antiserum significantly inhibited fat-stimulated IRGIP release. However, this mechanism could not account for the reduced IRGIP response to fat during concomitant insulin infusion observed by Stöckmann et a!. (1984) under their experimental conditions. C peptide levels were not elevated, and were in fact decreased by about 50% by insulin infusion, consistent with the suggested feedback inhibition of insulin secretion by insulin (Service et a!., 1978; Elahi et a!., 1982).  42  1.11 GIP RECEPTORS 1.11.1 BIOLOGICALLY ACTIVE SITE(S) It is possible, that the enterogastric and insulinotropic actions of GIP are mediated via two different receptor types, which interact with different regions of the polypeptide. Evidence for this hypothesis emerges from studies testing the biological activity of various regions of the GIP molecule. The cleavage of GIP by trypsin was first used during amino acid sequence determination of the polypeptide (Brown et a!., 1970). An unpurified mixture of tryptic fragments, including the major products 1-16, 19-30, and 34-42, had no enterogastrone activity, as tested in the denervated pouch of the fundus of the stomach in dogs. Brown and Pederson (1970b) reported that a purified cyanogen bromide produced C-terminal fragment with amino acids 15-42 was sufficient to inhibit acid secretion in the same model. Moroder et a!. (1978) tested synthetic porcine GIP 1’ inhibitor of gastric acid secretion. However, it was later t 38 and found it to be a “poor found that there was an extra glutamine residue in this sequence at position 30, which perhaps conthbuted to that observation. Pederson et al. (1990) observed only a weak somatostatinotropic effect in a combined isolated perfused pancreas and stomach preparation of the rat with synthetic porcine GIP 1-30. Furthermore, Rossowski et a!. (1992) found that unlike the parent peptide, porcine GIP 1-30 was not able to inhibit pentagastrin-stimulated gastric acid secretion in rats.  These observations, when  combined with that made by Brown and Pederson (1970b) indicate that the C-terminal portion of the polypeptide is likely the site of the acid inhibitory action of GIP. The first GIP fragment tested for insulinotropic activity was the purified cyanogen bromide fragment 15-42 produced by Brown and Pederson (1970b). In 1976, these investigators tested this C-terminal peptide in the perfused rat pancreas and found that it was approximately 40% as potent as an insulinotropic agent as intact porcine GIP (1976b). Using the same bioassay, Maletti et al. (1986) established that an HPLC purified fragment of bovine GIP resulting from enterokinase digestion (amino acids 17-  43 42) was approximately 32% as insulinotropic as native bovine GIP. The N-terminal synthetic fragments bovine GIP 1-39 (Sandberg et at., 1986), porcine GIP 1-38 (Moroder et al., 1978), human GIP 1-31 (Carquist, 1987) and porcine GIP 1-30 (Pederson et al., 1990; Gallwitz et a!., 1993) were shown to be equipotent with native GIP. In 1986, Schmidt et a!. found that HPLC purified porcine GIP 19-42 derived by proteolytic cleavage of the natural peptide was not significantly insulinotropic in isolated rat pancreatic islets. This result, when considered with those by Maletti et a!. (1986), and Pederson et al. (1990), suggests that the GIP region of 19-30 does not contain all the necessary amino acids for its insulinotropic activity, and 17-30 would be predicted to retain approximately 32% insulinotropic action. Blundell et al. (1976) pointed out that hydrophobic amino acid regions of the glucagon superfamily are probably important for receptor binding. The amino acid sequences 6-14 and 19-27 constitute the hydrophobic regions in GIP. It is also evident that the region 19-30 has so far been shown to be completely conserved in all species (Figure 1). Interestingly, in 1981, Brown et at. reported that the purest preparations of GIP existing at that time, contained a minor peptide component contributing approximately 5% to the total peptide content as shown by isotachophoresis and HPLC. Concomitantly, Jörnvall et at. confirmed the heterogeneity in the porcine OP preparation, and sequence analysis of the minor component suggested that it corresponded to GIP 3-42 (lacking residues 1 and 2). Jörnvall et at. (1981) reported 20% contamination by this fragment in GIP preparations, and suggested that it was formed by secondary processing or degradation, through susceptibility to attack by aminopeptidase, elastase, dipeptidyl aminopeptidase or related enzymes in the intestine.  Again in 1981, Brown et al.  compared the insulinotropic and somatostatinotropic actions of these two components in isolated perfused rat pancreas and stomach preparations. GIP 3-42 was found to lack significant actions on both IRSS and insulin release when compared with intact GIP 1-42. The lack of insulinotropic activity of OP 3-42 was confirmed by Moody et a!. (1981)  44 and Schmidt et a!. (1986). In 1987, Schmidt et a!. reported that this shorter form of OW comprised 32% of available natural GIP, and also found 4% CCK to be present. Furthermore, they noted that GIP 3-42 did not exhibit antagonistic activity to GIP 1-42, even at a 10-fold molar excess. Cariquist et a!. (1984) used Staphylococcus aureus V8 to cleave the N-terminal tripeptide from bovine GIP, resulting in GIP 4-42, and tested its insulinotropic activity in the perfused rat pancreas (Maletti et a!., 1986). In 3 Out of 6 rats tested, a “small” stimulation of insulin release as compared to GIP 1-42 was noted (—10%). It is not clear why the biological activity of inactive GIP 3-42 appears to be regained with the further removal of N-terminal amino acids.  1.11.2 GIP BINDING STUDIES An important step in confirming the actions of a peptide hormone on target tissues is the demonstration of specific binding sites. Attempts at studying the receptors that mediate the action of GIP on normal tissues have proven very difficult. Brown et a!. (1989) suggested two reasons to explain the unsuccessful attempts. Firstly, it was noted that iodination of GIP resulted in a heterogenous population of iodinated peptides, potentially altering the receptor binding and biological activity. Secondly, methods used to isolate B-cells or islets appeared to critically influence the GIP receptor. Indirect evidence for this was the observation that isolated pancreatic islets or islet cells only responded to pharmacological concentrations of GIP, unlike responses to other stimuli (Schafer and Schatz, 1979; Schauder et a!., 1975; Schauder et at., 1976). This implied that the GIP receptor was particularly susceptible to enzymatic damage during cell isolation. Verchere (1991) was able to demonstrate binding sites for GIP on cultured rat islets using HPLC purified, biologically active ‘ I-GIP. 2 5 Significant displacement of the radioligand by GIP was observed at concentrations as low as 1 nM. Tumor cells have been recognized as a potential source for large numbers of OP receptors. Malleti et a!. (1984) and Couvineau et a!. (1984) were able to demonstrate  45 high affinity binding sites in membrane preparations from hamster B-cell tumors. Maletti et al. (1984) ensured that the label was homogenous by HPLC purification, and demonstrated biological activity in the perfused rat pancreas. Amiranoff et al. (1984; 1985) provided evidence for specific GIP binding sites in the pancreatic tumor cell line In III. Furthermore, GIP binding was correlated with an increase in cAMP levels and insulin release. Binding of 1-GIP 25 in these studies was found to be saturable and could ‘ not be displaced by peptides structurally related to GIP. Both high affinity (Amiranoff et al.: KD  =  7 nM; 3000 binding sites/cell, Maletti et a!.: KD  =  2 nM; 219 binding sites/cell)  and low affinity (Amiranoff et al. :KD = 800 nM; 150,000 binding sites/cell, Maletti et a!.: KD  =  39; 1250 binding sites/cell) binding sites for GIP were identified. By using a cross-  linker to prevent dissociation of 125 1GJP bound to In III membranes, Amiranoff et al. (1986) identified a 59-kDa membrane protein that specifically bound GIP. Functional GIP receptors have also been demonstrated in human insulinoma plasma membranes (Maletti et al., 1987), again with both high affinity (KD  =  0.2 nM) and low affinity (KD  =  8.4 nM) binding sites, and in the mouse derived B-cell line, BTC3 (Kieffer et al., 1993). These receptors belong to a distinct family of G-protein coupled receptors referred to as the secretin-VIP receptor family (Christophe eta!., 1986; Rosselin, 1986). Binding sites for GIP in vivo were reported by Whitcomb et al. (1984) using  125J..  GIP in rats. Specific and displaceable binding of GIP to its receptors was demonstrated in the pancreas, glandular portion of the stomach, throughout the intestine and various muscle groups. No OP receptors could be demonstrated in the liver, adrenal gland, spleen, kidney, submandibular gland, testis, epididymis, prostate or seminal vesicles. Recently, using a molecular approach, Usdin et a!. (1993) cloned a novel receptor, which, when expressed in a cell line demonstrated activation only by GIP. Northern blots, reverse-transcription PCR and in situ hybridization demonstrated the receptor in tissues known to respond to OP (pancreas, gut, adipose tissue) and in novel locations such as the heart, brain, and inner layers of the adrenal cortex, where physiological effects of OP  46 have not been described. No receptor mRNA was found in kidney, spleen or liver. As neither GIP nor its effects have been described in the central nervous system, except at high concentrations (Ottlecz et al., 1985) these findings might suggest the presence of a novel, homologous peptide in the brain. Interestingly, the related peptide tGLP-I has been observed and its release demonstrated in the rat brain (Shimazu et al., 1987; Kreymann et a!., 1989) where it might act as a neurotransmitter. As the results of biological assays with GIP fragments do not demonstrate interactions with the receptor as either an agonist or antagonist, the availability of specific receptor binding assays is important. Some fragments of GIP shown to be biologically active (GIP 1-31; 1-30; 17-42) were shown to competitively inhibit the binding of  125j..  GIP (Maletti et a!., 1987; Kieffer et al., 1993; Maletti et a!., 1986; Gallwitz et a!., 1993). On the other hand, fragments not expected to yield insulinotropic activity, such as GIP 19-30 (see section 1.11.1), did not displace binding of the tracer to insulinoma membranes (Maletti et al., 1986). Likewise, the GIP fragment 1-27 was unable to displace OW binding to RINm5F cells (Gallwitz et al., 1993). Interestingly, the GIP fragment 4-42, which demonstrated only partial insulinotropic action, was capable of competing equally with GIP 1-42 for tracer binding (Maletti et a!., 1986). In this regard, GIP 4-42 may act as a partial antagonist. In 1991, Fehmann and Habener showed that the tGLP-I receptor in the glucoseresponsive B-cell line HIT-T15 could rapidly and reversibly desensitize in response to supraphysiological concentrations of tGLP-I. Preperifusion with GIP had no effect on the tGLP-I response indicating the tGLP-I receptor on the HIT-T 15 cells was distinct from that of OW. Gallwitz et a!. (1993) arrived at the same conclusion with the RINm5F cell line when it was observed that I-GIP 25 binding was not displaced by tGLP-I, and ‘  125I  tGLP-I binding was not altered by GIP. Fehman and Habener (1991a) also suggested that the OW receptor could undergo homologous desensitization from their observation that prior exposure of HIT-T15 cells to GIP (100 nM) also reduced the insulin secretion  47 during stimulation with 10 nM GIP.  This hypothesis was also supported by the  demonstration that high concentrations of GJP inhibit both insulin secretion from the perfused pancreas (Szecowka et al., 1982a) and cAMP generation in human (Malleti et al., 1987) and rat (Gallwitz et al., 1993) insulinoma cells, possibly by homologous desensitization. Furthermore, in rats subjected to total parenteral nutrition for a period of 6 days, there was a 30% increase in insulin release from the perfused pancreas in response to GIP (Pederson et al., 1985). It was concluded that subjecting the pancreas to chronic low concentrations of GIP could induce an increase in the sensitivity or number of GIP receptors.  1.12 GIP METABOLISM In 1975, Brown et al. determined the half-life of porcine IRGIP in the plasma of normal subjects to be approximately 21 mm. Identical values were obtained using porcine GJP in man by Elahi et a!. (20.4 ± 2.37 mm; 1979) and Sarson et al. (20.3 ± 1.2 mm; 1982), and the metabolic clearance rate was calculated as 2.6 ± 0.1 ml/kg•min (Sarson et a!., 1982). While even higher values have been obtained for the half-life of tGLP-I in rat (39.5 ± 15.5 mm; Oshima et a!., 1988) and man (45.9 ± 8.8 mm; Oshima et al., 1991), other related peptides are cleared from the plasma much quicker, with halflives of 5 mm or less (secretin: 2.5 mm, Häcki et a!., 1977; vasoactive intestinal peptide: 1 mm, Modlin et a!., 1978; glucagon: 5.5 mm, Jaspan and Rubenstein, 1977). It therefore appears that exogenously administered GIP is cleared relatively slowly from human plasma. Wolfe and McGuigan (1982) measured the half-time of disappearance for porcine IRGIP in the dog at 7.6 ± 1.5 mm, thus indicating that some species differences might exist. In 1979, Elahi et al. reported that the prevailing state of glycemia could influence the metabolic half-life of GIP. Using porcine GIP in human subjects, Elahi et a!. (1979) found that under euglycemic conditions, the half-life was 26.3 ± 4.57 mm, during mild  48 hyperglycemia (8 mM), 23.2 ± 3.16 mm, and under moderate hyperglycemia (13 mM) dropped to 13.1 ± 2.59 mm. As the moderate hyperglycemia was also associated with hyperinsulinemia, Elahi and associates (1979) were unable to conclude which factor may have been responsible for the shortened IRGIP half-life. In order to differentiate between these two factors, Andersen et at. (1980) examined the half-life of exogenous porcine GIP in fasting dogs during hyperinsulinemia and hyperglycemia, alone and in combination. The half-life was significantly reduced by hyperglycemia alone or in combination, but there was no change with hyperinsulinemia alone, suggesting that hyperglycemia, but not hyperinsulinemia could significantly enhance the metabolic clearance rate of IRGIP. Recently, Sheu et at. (1987) also noted a greater metabolic clearance rate of porcine GIP in fed versus fasted rats. However, not all investigators have reported similar observations. Nauck et at. (1989) observed no difference in the half-life (18 mm) or metabolic clearance rate (‘—6 mlIkgmin) of human GIP infused in normal subjects during basal (5 mM) or hyperglycemic (8 mM) clamps. All pancreatic and gastrointestinal hormones must traverse the liver prior to reaching the general circulation. The liver is major clearance site for some hormones, such as insulin, of which approximately 50% of that secreted is removed with each transhepatic circulation (Ishida et at., 1983; Stoll et al., 1970). On the other hand, there appears to be no hepatic extraction of endogenous or exogenous Gil’ (Hanks et at., 1984; Chap et al., 1987). The kidney also inactivates or clears several polypeptide hormones, such as insulin, glucagon and gastrin from the circulation (Chamberlain and Stimmier, 1967; Sherwin et al., 1976; Davidson et al., 1973). O’Dorisio et at. (1977) found a renal arterial-venous IRGIP difference of 39% in dogs during intraduodenal perfusion of glucose, and found that uremic patients had higher than normal fasting and stimulated IRGIP concentrations. It therefore appears that the kidney is the major site of GIP clearance.  49 An important factor that must be considered when examining the metabolism of a hormone is the possibility that hormone measurements associated with immunoreactivity may not necessarily correspond to the biological activity of that hormone. There are numerous proteases in plasma that could potentially render a hormone biologically inactive, yet still immunoreactive in an RIA. Thus, while measurements of the half-life of IRGIP levels in plasma are useful, they must also be combined with studies examining the duration of biological activity of circulating GIP. The importance of this issue has recently been underscored by the observation that dipeptidyl peptidase IV hydrolyzes GIP in human serum to the biologically inactive fragment GIP 3-42 and thus may play an important role in the metabolism of GIP (Mentlein et a!., 1993). Present GIP antisera may be unable to distinguish this truncated N-terminal form, suggesting that GIP 3-42 likely contributes to IRGIP measurements in the circulation. Further discussion of this issue appears in Chapter 4 of this thesis.  1.13 GIP PATHOPHYSIOLOGY The possibility exists that GIP may be involved in the etiology of pathophysiological states. In particular, in light of the fact that GIP is a major incretin candidate of the enteroinsular axis, the role of GIP in diseases associated with insulin deficiency and excess have been extensively investigated. The results of these studies, however, leave no clear answer to the function of GIP in these states. For instance, in non-insulin dependent diabetes mellitus (NIDDM), circulating IRGIP levels have been reported to be increased (Brown et al., 1975; Mazzaferri et a!., 1985; Elahi et at., 1984; Cox et al., 1981; Lardinois et at., 1985; Jones et a!., 1989b; Ross et at., 1977), normal (Service et at., 1984; Levitt et at., 1980; Alam et al., 1992), or decreased (Alam and Buchanan, 1980; Service et a!., 1984; Groop, 1989; Nauck et al., 1993) following the oral administration of nutrients. Similarly, in obese subjects, some studies have demonstrated elevated fasting levels of IRGIP (Elahi et a!., 1984; Mazzaferri et al., 1985; Salera et al.,  50 1982a), and others exaggerated (Creutzfeldt et a!., 1978; Elahi et a!., 1984; Mazzaferri et a!., 1985; Ebert et al., 1979a; Jones et al., 1989a; Fukase et a!., 1993), normal (Lauritsen et al., 1980; Jorde et al., 1983; Amland et al., 1984; Ebert and Creutzfeldt, 1989) or blunted (Service et al., 1984; Groop, 1989) levels after oral nutrient ingestion. Among other defects in insulin secretion in NIDDM patients, there is a reduced or absent incretin effect: insulin release is no longer stimulated more by oral as compared to “isoglycemic” intravenous glucose (Perley and Kipnis, 1967; Nauck et a!., 1986; Tronier et a!., 1985). A reduced incretin effect could be caused by impaired secretion of relevant incretin hormones or by B-cell insensitivity to their insulinotropic action. While the data on GIP secretion are not conclusive, reports of the insulinotropic action of GIP in NIDDM patients seem to be in agreement.  In streptozotocin diabetic rats, the  insulinotropic effectiveness of GIP was significantly reduced (Suzuki et a!., 1990). In human diabetic subjects, there appears to be a significantly lower insulin response to intravenous GIP when compared to normal subjects (Jones et al., 1987; Krarup et a!., 1987c; Nauck et a!., 1993; Meneilly et a!., 1993). Meneilly et a!. (1993) found that the sulfonylurea glyburide significantly increased the B-cell response to GIP infusion in diabetic subjects, but the observed insulin release was still approximately 10-fold less than that of controls. Interestingly, Nauck et al. (1993) found no significant difference in the insulin response to tGLP-I in diabetic vs. normal subjects. Holz et a!. (1993) postulated that since tGLP-I appeared able to compensate for a defect in the glucose signaling pathway that regulates insulin secretion from B-cells, making them “glucose competent”, tGLP-I might be useful for the treatment of NIDDM. The presence of a glucose threshold for the insulinotropic action of GIP is teleologically necessary to prevent inappropriate insulin release and hypoglycemia. Since untreated patients with NIDDM have plasma glucose levels above this threshold even in the fasting state, Jones et a!. (1989) investigated whether GIP was insulinotropic during fasting in these individuals. It was found that the glucose threshold for the insulinotropic  51 action of GIP was not altered in the NIDDM patients, and thus there was a significant insulin response to intravenous GIP even during fasting. The resulting secretion of insulin was, however, insufficient to decrease plasma glucose levels. In normal subjects, persistently elevated GIP levels only have a short lived effect on insulin release, as plasma glucose levels are only above the necessary threshold for a relatively short time following meals. Jones et al. (1989b) reasoned that in untreated NIDDM subjects exhibiting an insulinotropic effect to GIP in the hyperglycemic fasting state, persistently elevated GIP would be likely to exert an effect on insulin secretion throughout the day. Obesity is often associated with glucose intolerance and hyperinsulinemia in both the fed and fasted state (Perley and Kipnis, 1967; Creutzfeldt et a!., 1978; Chan et al., 1984). Unlike diabetics, however, B-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., 1985). Elahi et a!. (1984) calculated B-cell sensitivity to GIP from the insulin and IRGIP responses to oral glucose and found it normal in obese individuals. The GIP sensitivity of B-cells from the hyperinsulinemic obese Zucker rat has also been measured. Chan et a!. (1984) found that the insulin response to GIP was enhanced in the pancreas of obese animals and additionally, that the glucose threshold for the insulinotropic action of GIP was well below fasting levels in these rats. Thus, as with NIDDM, there may be inappropriate GIP insulinotropic activity occurring associated with obesity. How might the insulinotropic activity of GIP become uncontrolled? A defect in the feedback control of GIP secretion by insulin was proposed as the explanation for the exaggerated [RGIP release after food stimulation in obesity (Brown and Otte, 1978) and in NLDDM (Crockett et a!., 1976; Ross et al., 1977). Hampton et a!. (1983) studied the effects of pre-treatment with a high fat diet on the IRGIP and insulin responses to oral fat and glucose in rats. The pre-treatment resulted in increased IRGIP secretion in response to oral fat and abolition of the feedback inhibition of exogenous insulin on fat-stimulated  52 IRGIP release. It was also noted that some degree of insulin resistance was established.  Willms et al. (1978), found that 5 days of caloric restriction abolished the exaggerated IRGIP response to fat or a mixed meal in obese subjects, suggesting that a previously high calorie intake might have been responsible for the excessive IRGIP production. Creutzfeldt et al. (1978) also reported an exaggerated IRGIP response in obese subjects given a high calorie mixed meal. This was not, however, observed in subjects after 5 days of dietary restriction. Hampton et al. (1983) suggested that this might be due to a decreased responsiveness of the GIP secreting cells to insulin due to a reduction in insulin receptor numbers. Investigation of insulin receptors on human monocytes has shown that insulin receptor number can be altered by dietary changes (Pedersen et al., 1980) and that these changes can occur rapidly within 24 h (Schluter et a!., 1980). -  1.14 THESIS INVESTIGATION In summary, the hormone GLP, has been extensively studied as a potential enterogastrone and an incretin involved in the enteroinsular axis. In addition, evidence has been presented that suggests GIP has other roles, such as direct actions on fat metabolism. These activities have been substantiated by the location of specific GIP receptors on the respective target tissues. Studies have revealed specific regions of the molecule to be important for receptor interactions, and activation of intracellular messenger systems have been associated with GIP binding. The endocrine cells that release GIP have been identified, and the gene coding for the prepropeptide sequenced. Numerous GIP assays have been established allowing for the measurement of GIP responses to the ingestion of nutrients or the intravenous administration of various secretagogues, and have enabled an estimation of the half-life of GIP immunoreactivity in plasma. Observations from circulating IRGIP profiles do not allow direct assessment of the actions of GIP secretagogues and inhibitors because of the potential interactions  53 among paracrine, endocrine, neural and luminal influences at the level of the GIP cell. These confounding factors can only be overcome by studying GIP release at the cellular level. Such methods allow the investigation of many components potentially influencing GIP release, individually and in combination. Studies at the cellular level also allow investigation of the intracellular mechanisms by which agents act on the GIP cell. One objective of these thesis investigations was therefore to develop methods that would enable the study of GIP secretion at the cellular level in a controlled environment. Observations of circulating IRGIP levels have also been used to evaluate the immunoreactive half-life of the hormone. It is not known, however, how accurately these measurements reflect the time of biological activity of circulating GIP. There is evidence to suggest that peptidases may convert GIP to a biologically inactive fragment, that is immunologically indistinguishable from the native form. Reports of circulating IRGIP levels may therefore have little meaning in terms of the actual physiological potency of plasma GJP. A second aim of the thesis investigations was consequently to carefully examine the metabolic processing of GIP and to determine a physiological time frame during which this process occurs.  A better understanding of GIP secretion and  metabolism is critically important to resolving the potential role this hormone may have in both normal and pathophysiological situations.  54 CHAPTER 2 GIP RELEASE FROM ISOLATED CANINE AND PORCINE ENDOCRINE CELLS  2.1 INTRODUCTION A major factor impeding study of isolated GIP cells is the diffuse nature of the distribution of endocrine cells throughout the small intestine. Separation techniques to produce enriched preparations of gastric endocrine cells were developed by Soil and colleagues in 1984. These techniques were used to study the release of somatostatin and have been modified to permit the study of the local regulation of peptide secretion from purified gastrin (Giraud et a!., 1987; Campos et a!., 1990), cholecystokinin (Barber et al., 1986b; Koop and Buchan, 1992), neurotensin (Barber et a!., 1986a), peptide YY (Aponte et a!., 1988), motilin (Poitras et al., 1993) and enteroglucagon cells (Buchan et al., 1987). The aim of these studies was to modify these techniques to allow for the isolation, enrichment, and culture of intestinal endocrine cells in order to study the local regulation of GIP secretion.  2.2 MATERIALS AND METHODS 22.1 ANIMAL TISSUES Initial studies were performed with adult mongrel dogs (n  =  10), but due to  unavailability of dogs, the model was later adapted for use with pigs (n  =  10). Fasted  male and female dogs or pigs were sedated with phentanol triperidol (0.1 mi/kg) administered with atropine (0.05 mg/kg), anesthetized with intravenous sodium pentobarbital (30 mg/kg) and prepared for abdominal surgery. The animals were then bled from the abdominal vena cava, and the upper small bowel was removed. The intestine was immediately cut open and washed in ice-cold Hanks’ balanced-salt solution (HBSS; Gibco Laboratories, Burlington, Ont.) —pH 7.4 containing 0.1% bovine serum albumin (BSA fraction V; Sigma Chemical, St. Louis, MO).  55 22.2 ISOLATION OF MUCOSAL CELLS The canine or porcine mucosa was blunt dissected from the intestine, weighed and chopped finely with scalpels. Aliquots (10 g) of canine mucosa were initially digested in 50 ml basal medium Eagle (BME; Gibco) —pH 7.4, with a mixture of enzymes consisting of 75 U/mi type I coliagenase (Sigma), 75 U/mi type XI collagenase (Sigma), 0.9 U/mi type IX protease (Sigma) and 1 U/mi trypsin (Worthington Biochemical Corp, Freehold NJ) for 1 h in a shaking water bath at 37°C. The total volume was then doubled with HBSS-BSA and allowed to settle —10 mm. The supemate containing detached cells was discarded, and the remaining tissue was further digested in the enzyme mixture for two 45 mm  periods, with each step followed by the addition of 300 1.11 of 0.5 M  ethylenediaminetettraacetic acid (EDTA; BDH, Toronto, Ont.) for 15 mm. A similar procedure was followed for the porcine mucosa, but as it was much thinner, the digestions were performed only with 75 U/mi type I collagenase (Sigma). The cell suspension resulting from digest 3 or 4 of the canine and porcine mucosa was filtered through Nitex mesh (200 Lm, B & SH Thompson, Scarborough, Ont.) and washed and centrifuged at 200 x g (Baxter Biofuge) twice with HBSS-BSA supplemented with 0.01% dithiothreitol (DTT; Sigma) and 0.001% deoxyribonuclease (DNAse; Sigma). The cells were then filtered a second time through fine Nitex mesh (62 jim, B & SH Thompson), counted, and diluted in HBSS-BSA-DTT-DNAse to 6 x 106 cells/mi for elutriation.  22.3 ENRiCHMENT OF MUCOSAL ENDOCRINE CELLS Counter-flow centrifugal elutriation of cells was first reported by Lindahl in 1948. This procedure separates cells on the basis of their sedimentation coefficients. The cell suspension is pumped into a rotating chamber, and cells are held where their sedimentation rate is balanced by the flow of fluid through the separation chamber. Different fractions of homogeneous cells are then ‘eluted’ by either increasing the flow rate through the chamber or decreasing the centrifugal speed. The appropriate flow rates  56 and centrifugation speeds were determined empirically. Batches consisting of 1.5 x 108 canine or porcine dispersed cells were introduced into the Beckman elutriator (model J221M/E; Beckman, Los Angeles, CA) via a pump (Cole Parmer) connected to a sterile source of HBSS-BSA. The enzyme dispersed canine cells were loaded into the elutriator chamber at a rotor speed of 2500 rpm with a flow rate of 25 mi/mm and washed for 2 mm A 100 ml fraction (Fl) was collected after changing the flow rate to 30 mI/mm. A second 100 ml fraction (F2) was obtained at a rotor speed of 2100 rpm and a flow rate of 55 ml/min. The elutriation procedure for the dispersed porcine mucosal cells was slightly different. Cells were loaded at a rotor speed of 3000 rpm and a flow rate of 25 mi/mm. The Fl and F2 fractions were eluted at 2800 rpm, 30 mI/mm and 2200 rpm, 55 mi/mm respectively. Samples of cells (5 x 106, in triplicate) from the pre-eluted cells, Fl and F2 were extracted in 1 ml of 2 M acetic acid in 1.5 ml Eppendorf microcentrifuge tubes. The tubes were boiled for 10 mm, centrifuged for 5 mm at 7000 x g (Baxter Biofuge) at 4°C and the supernate stored at -20°C for subsequent radioimmunoassay (RIA). Prior to being assayed, these extracts were neutralized with 10 M NaOH. Results indicated that the majority of canine and porcine IRGIP cells eluted in F2, and consequently these cells were cultured for immunocytochemical and hormone release studies.  22.4 ENDOCRINE ENRICHED CELL CULTURE Canine and porcine cells from F2 were concentrated by centrifuging at 200 x g for 10 mm, and then resuspended in sterile culture medium. The canine culture medium consisted of 47.5% Dulbecco’s modified Eagle medium (DMEM, Terry Fox Laboratory, Vancouver, BC) and 47.5% Ham’s F-12K (Terry Fox Lab), containing 5.5 mM glucose, supplemented with 5% fetal calf serum (FCS, Gibco), 2 ng/ml nerve growth factor (Collaborative Research, Bedford, MA), 8 pg/ml insulin (Sigma), 1 rig/mI hydrocortisone (Sigma), 50 pg/ml gentamycin sulfate (Sigma), 0.25 jig/mI amphotericin B (Fungizone,  57 Gibco), 50 U/mi penicillin (Gibco), 50 ag/ml streptomycin (Gibco), and 20 iM cytosine B-D-arabinofuranoside (Gibco). The porcine culture medium was identical with the exception of 10% heat inactivated porcine serum substituted for the 5% FCS. Plates for culturing the cells (12-well Costar, Cambridge, MA) were coated with collagen to promote cell adherence. To prepare the collagen coated plates, rat tail collagen was dissolved in 0.0 17 M acetic acid (3.5 mg collagenlml) and then added to each well and left for 45 mm. The solution was then aspirated, and the plates allowed to air dry prior to .10 m11.we111 in the collagen coated storage at 4°C. Cells from F2 were seeded at 5 x 6 culture plates and maintained at 37°C in a humidified atmosphere of 5% CO 2 in air for 40 h.  22.5 IMMUNOCYTOCHEMISTRY Both cultured cells and intact intestine were prepared for immunocytochemical analysis. Sections of intestine (1 cm ) were fixed in Bouin’s solution (75 ml saturated 2 picric acid, 25 ml 40% formaldehyde, 5 ml acetic acid) for 1 h, and then washed and stored in 70% ethanol prior to processing. The tissue was dehydrated through a graded series of alcohol and xylene using a histomatic tissue processor (model 166; Fisher). The samples were then embedded in paraffin wax using a histo-center IT-N (Thermolyne). Sections 5 I.Im thick were cut using a microtome (1 130/Biocut; Reichert-Jung) and dried on gelatin coated slides. Prior to staining, wax was removed by soaking the sections for 10 mm in xylene followed by clearing in petroleum ether for 2 mm. Cultured cells were washed 2 x in phosphate-buffered saline (8.0 g/l NaCI, 0.2 g/l , 4 P 2 KH O 1.15 g/l 4 H 2 Na ) PO plus 0.1% sodium azide (NaN ; PBS-azide). The cells were then fixed in 3 Bouin’s solution for 10 mm and then washed 3 x and stored in PBS-azide. Tissue sections and cultured cells were immunostained with primary antibodies for all hormones listed in Table 1 using the avidin-biotin peroxidase method. To prevent nonspecific staining caused by the presence of endogenous peroxidase activity, samples  58 were immersed in PBS-azide containing 0.3% H 0 (BDH Inc. Toronto, Ont.) for 30 mm 2 at room temperature and then washed in PBS-azide. The fixed cells and 5 jim sections of jejunum were incubated with the antibodies diluted in PBS-azide containing 10% swine serum as listed in Table 1 for a period of 48 h at 4°C. After washing 3 x in PBS-azide, the biotinylated secondary antibodies (Table 1) were applied and incubated for 1 h at room temperature. The secondary antibodies were then localized with a premixed (10 mm) avidin-biotin peroxidase solution (Vectastain ABC Kit; Vector Laboratories, Burlingame CA) for 1 h. The resulting complex was identified with pH 7.6 tris buffer solution of diaminobenzidine (BDH; 125 p.g/100 ml) containing 40 mg 4 NH C 1, 200 mg D-glucose, and 0.3 mg glucose oxidase (Sigma). Staining was monitored with a Zeiss Axiophot microscope, and the reaction was terminated after the appearance of positive cells (15-45 mm) by washing in PBS-azide. In order to quantify staining, a minimum of 10 groups of approximately 100 cells were observed for positive cells, which were then expressed as a % of the total number counted. The cultures and intact mucosa were also stained for mucin containing cells using a combined alcian blue-periodic-acid Schiff technique. Fixed cells were treated with 1% alcian blue (Fisher, Fair Lawn, NJ) in 3% acetic acid for 5 mm  and then washed in  distilled water. The cells were treated with 1% aqueous periodic acid solution for 5 mm and again washed in distilled water. Finally, they were washed in Schiff’s reagent (0.5% basic fuchsin, Fisher; 0.5% sodium metabisulphite, Fisher; 0.5% hydrochloric acid, and 0.5% decolorizing charcoal, Fisher) for 15 mm followed by 10 mm of running water. Neutral mucins stained red and acid mucins blue (pink and purple respectively in figures).  2.2.6 RELEASE EXPERIMENTS On the experimental day, the culture medium was removed by pipet and the porcine or canine cells washed with 1 ml release medium, consisting of DMEM (Sigma) supplemented with 25 mM NaHCO , 25 mM HEPES, 5 mM glucose, 1% FCS, and 2% 3  59 aprotinin (Trasylol, 500 kallikrein-inhibiting U/mi, Miles Pharmaceuticals, Rexdale, Ont.). This was followed by addition of 1 ml release medium containing the appropriate glucose concentration to each well. Stock solutions of other secretagogues were added as 25 .tl of a 40 x concentration to 975 j.tl of release medium in 5 mM glucose, in triplicate. Cells were incubated for 2 h, after which the release medium was removed and stored in 1.5 ml Eppendorf microcentrifuge tubes on ice. The release medium was centrifuged for 5 mm  at 7000 x g (Baxter Biofuge) at 4°C and the supernate stored at -20°C for  subsequent GIP and somatostatin RIA. Adherent cells in control wells (2 wells per plate) were extracted in 2 M acetic acid for determination of total cell peptide content per well. After centrifugation to remove particulate matter, the supernate was stored at -20°C. Prior to being assayed, these extracts were neutralized with 10 M NaOH. Peptide secretory values from each well were expressed as percentage of total cell peptide content (%TCC) as measured from extracts.  2.2.7 SECRETAGOGUES The somatostatin antibody (SOMA 10; Regulatory Peptide Group, Vancouver, BC, Canada) was diluted in release medium to 10 jig/mI/well. Porcine gastrin releasing peptide (GRP, Penninsula Laboratories mc, Belmont CA) was diluted to a concentration of 4 x 10-6 M with release medium and kept on ice. This solution was serially diluted to form other stock solutions of 4 x iOfl, 4 x 10-8, and 4 x 10 M. Forskolin (Sigma) was prepared as a stock solution of 4 x iO M in ethanol and serially diluted in DMEM release medium to concentrations of 4 x 10, 4 x iO, and 4 x 10-6 M. A23 187 (Sigma) was prepared as a stock solution of 4 x 10 M in dimethylsulfoxide (DMSO) and serially diluted in DMEM release medium to concentrations of 2 x i0 and 4 x lO M. The effect of the solvents ethanol and DMSO on IRGIP or IRSS release was tested by adding identically diluted solvents to medium in corresponding control wells. Peptide values  60 resulting in these controls were subtracted from the appropriate values obtained with forskolin or A23 187.  22.8 RADJOIMMUNOASSAY FOR GIP The RIA used to measure immunoreactive GIP (IRGIP) was originally described by Kuzio et a!. (1974) and modified by Morgan et a!. (1978). The ‘ 1-GIP 2 5 used for the RIA was prepared by a slight modification of the chloramine-T method as was performed by Greenwood et al. (1963). The column for label purification was prepared prior to the iodination with Sephadex G-15 (Pharmacia, Uppsala, Sweden). The G-15 was first swollen in 0.2 M acetic acid with the ‘fines’ repeatedly decanted, and then degassed under water vacuum for —2 h. A 1:1 ‘slurry’ of gel and acetic acid was then poured into a 0.5 x 20 cm plastic pipet with a glass wool plug and allowed to settle at an eluent (0.2 M acetic acid plus 5.0% BSA; RIA Grade) flow rate of -250 I.tl/min for 4-5 h. The column buffer was supplemented with 2% aprotinin for -1 h before the iodination. The iodination reaction was performed in a siliconized test tube containing 5 p.g of porcine GIP (Regulatory Peptide Group) dissolved in 100 p.1 40 mM pH 7.5 phosphate buffer (10 x stock prepared by titrating 0.4 M Na 4 H 2 PO with 0.4 M NaH 4 P 2 O to pH 7.5). In a fume hood behind lead, 0.5 mCi Na 1 (NEN Research Products, Wilmington, DE) and 10 p.1 125 chioramine-T (Fisher, Fair Lawn, NJ; 4 mg/ml phosphate buffer) were added while the reaction vessel was agitated. The reaction was stopped after 15 sec by the addition of 20 p.1 sodium metabisulfite (Fisher; 12.6 mg/mi phosphate buffer). Separation of the free iodine from the 125 1-labeled hormone was achieved using gel filtration on the Sephadex G-15 column previously poured. Fractions of 400 p.1 were collected at a flow rate of 250 p.1/mm and 10 p.1 aliquots were counted (LKB Wallace 1277 Gammamaster). A typical column profile obtained with these counts is shown in Figure 23. Fractions comprising the ‘ 1-GIP 2 5 peak were assayed for ‘damage’ or the ability of -  the label to bind to charcoal. The 10 jii fractions were diluted to —5000 cpm/100 p.1 of assay  61 buffer (pH 6.5, 40 mM phosphate buffer containing 5% CEP and 2% aprotinin). Then, 100 p.1 aliquots of the diluted fractions were added in triplicate to tubes containing 900 p.1 assay buffer. Dextran-coated charcoal was prepared by dissolving 2.5 mg/ml dextran (T-70, Pharmacia) in 40 mM phosphate buffer (pH 6.5), adding 2.5 mg/mi activated charcoal (Norit Decolorizing Carbon), 0.75% aprotinin, plus 5% CEP, and stirring for 2h at 4°C. Each tube of diluted label received 200 p.1 of the dextran-coated charcoal and was vortexed and centrifuged (30 mm, 1000 x g). Both the supematant and the charcoal pellet were counted and the percent of total radioactivity bound (%B) to charcoal (i.e. 125 1-GIP) was calculated. The fraction from the 1-GIP 25 peak with the highest radioactivity and the first 1-2 fractions ‘ on the descending portion of the peak usually had the greatest adsorption to charcoal (usually >95%). These fractions were also 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). Therefore, these fractions were pooled and diluted to 2.5 x 106 cpmIlOO p.1 in a 1:1 mixture of column buffer and acid ethanol (750 mill 95% ethanol plus 15 mi/l concentrated HC1), and stored at -20°C until used in the RIA. The rabbit anti-porcine GIP serum used for the RIA (RK343F also referred to as LMR34) was generously supplied by Dr LM Morgan (University of Surrey, UK). Negligible cross-reactivity of this antibody was demonstrated with cholecystokinin, insulin, pancreatic polypeptide, glucagon, secretin, and vasoactive intestinal peptide (Morgan et al., 1978). This antisera was stored as 100 p.1 aliquots of a 1:40 dilution in assay buffer at -20°C, and diluted to 1:6000 for use in the RIA (final dilution in RIA of 1:3 x 10). Stock standards for the RIA were prepared from 100 p.g porcine GIP (Regulatory Peptide Group) dissolved in 0.2 M acetic acid plus 0.5% BSA (Pentex) and 2% aprotinin to a concentration of 100 p.g/ml. Aliquots of 1 p.g (10 p.1) were then lyophilized and stored at -20°C. When needed, an aliquot was dissolved in 100 p.1 acetic acid (100 mM) and 100 p.1 assay buffer to achieve a concentration of 5 p.g/ml. This was then further diluted to 20 ng/ml  62 by diluting 40 p1 up to 10 ml in assay buffer, and then serially diluted in assay buffer to provide a range of standards down to 78 pg/ml. On the first day of the assay, 200 p1 assay buffer was added to all sample tubes followed by either 100 p1 of porcine GIP standard (7.8  -  2000 pgIlOO p1) or 100 p1 of  unknown sample, and 100 p1 of antibody (LMR34). For the non-specific binding (NSB) and zero-binding (Bo) tubes, 100 p1 of buffer was used to replace the antibody or standard, respectively. All tubes were vortexed and incubated 24 h at 4°C. On day two, 100 p1 of porcine GIP tracer (—5000 cpmIlOO pi) was added to each tube and vortexed. On day three, 500 pi of polyethylene glycol (250 g/l H O PEG 8000; Fisher) was added to all but the total 2 counts and the tubes were vortexed and centrifuged at 1000 x g for 45 mm, after which the supematant was decanted and the pellets were allowed to dry overnight. The specific radioactivity in the bound fraction (B) was then determined by counting the pellets for 2 mm on a gamma counter (LKB Wallace 1277 Gammamaster) and subtracting the NSB. Sample values were computed from a spline smooth plot of B expressed as a percentage of the Bo -  against concentration using an IBM 386 microprocessor with RiaCalc (Pharmacia). Controls for the GIP RIA were prepared by dissolving 1 pg porcine GIP (Regulatory Peptide Group) in 500 ml of 40 mM phosphate buffer (pH 6.5) containing 0.5% BSA (Pentex) and 2% aprotinin (200 pg/100 jil). Aliquots of 1 ml were stored in Eppendorf tubes at -20°C until use in the RIA. Inter- and intra-assay variation calculated from these controls was 10 ± 2% and 5 ± 1%, respectively. Control studies were also performed to examine the effect of incubation of GIl’ at 37°C in release medium, with and without isolated cells. No significant difference was noted in standard curves prepared from porcine GIP standards under normal assay conditions or after 2 h incubation with cells.  63 22.9 RADIOIMMUNOASSAY FOR SOMATOSTATIN The RIA used to measure immunoreactive somatostatin (IRSS) was as described by McIntosh et a!. (1978; 1987). Synthetic Tyr’-somatostatin was iodinated by the chloramine-T method, and the labeled peptide was separated from free  125J  and damaged’  material by adding 1 ml hormone free plasma followed by 20 mg microfine silica (QUSO G32; Philadelphia Quartz Co., USA). After centrifugation the pellet was washed twice with 1 ml distilled water, and the label eluted with 1 ml acetic acid/acetone/water (0.1/3.9/4). The label was then diluted in 0.1 M acetic acid containing 0.5% BSA (Pentex, Miles Labs) to 1 x 106 cpm/10 p.! and lyophilized and stored at -20°C. On the day of the RIA, an aliquot of label was dissolved in 1 ml 2 mM ammonium acetate buffer (pH 4.6) and applied to a column (0.9 x 10 cm) of CM-cellulose (CM 52; Whatman) equilibrated in the same buffer. The column was then washed with —50 ml buffer pumped at a flow rate of 1 mI/mm and the label was eluted with 200 mM ammonium acetate (pH 4.6). The peak fraction (2 mm/fraction) was then neutralized with 2 M NaOH, and diluted in assay buffer (50 mM barbital buffer pH 7.4 containing 0.01% merthiolate, 1% aprotinin, and 0.5% BSA) to -P3500 cpm/100 1.tl. A somatostatin monoclonal antibody (SOMA 03; Regulatory Peptide Group) was used in the assay, prepared from a stock of 1:100 (in 0.9% NaC1 containing 0.1% NaN 3 and 0.5% BSA) stored at 7°C. Antibody was diluted to 1:1 x 106 in assay buffer the day of the RIA for a final concentration of 1:4 x 106. This antibody was shown not to crossreact with GIP, gastrin or motilin and is directed towards the central region of the somatostatin- 14 molecule, recognizing somatostatin- 14 and -28 equally (McIntosh et aL, 1987). Assay standards were prepared by dissolving synthetic cyclic somatostatin-14 (Penninsula) in 100 mM acetic acid containing 0.5% BSA (Pentex) to obtain a concentration of 100 p.g/ml. Aliquots of 5 p.g were then lyophilized and stored at -20°C. On the day of the assay, an aliquot was reconstituted in 100 p.! cold distilled water  64 followed by 400 j.tl assay buffer to obtain a concentration of 10 p.g/ml. This was then serially diluted (1:10, 1:10, 1:10, 1:20) to a concentration of 500 pglml, and then sequentially 1:2 to 3.9 pg/mi for a total of 8 standards for the RIA. The same general procedure was used for the RIA as in the GIP assay, with the following differences. All the assay constituents were added to the tubes the first day in the following order, 100 p.! standard or sample, 100-300 p.1 assay buffer, 100 .tl antibody and 100 p.! label for a total volume of 400 p.1/tube. The assays were then vortexed and incubated for —72 h at 4°C. The bound peptide was then separated from the unbound by adding 1 ml dextran-coated charcoal (plus CEP) to each tube (except total counts), allowing the assays to sit —15 mm, centrifuging at 1000 x g for 30 mm, and then decanting the supematant. After drying, the pellet was counted for 3 mm on a gamma counter, and sample values were computed from a spline  -  smooth plot of standard B  expressed as a percentage of the Bo against concentration. Samples containing known amounts of somatostatin for controls could not be kept for long periods of time because of loss of immunoreactivity. Interassay variation was therefore computed from standards of 10 standard curves at 11 ± 2%.  Intraas say  variation as computed from standards (20 duplicates) placed randomly throughout the assay was 5 ± 1%.  2.2.10 EXPRESSION OF RESULTS Release data were calculated as mean ± SEM for a percentage of the total cell content (%TCC) to reduce variation between animals.  Statistical significance was  determined using the StudenCs t test and was set at the 5% level. Support for using this method as opposed to an analysis of variance test followed by one of the a posteriori tests was provided by Sinclair (1988).  65 2.3 RESULTS 2.3.1 MUCOSA DIGESTION The digestion protocol described for both canine and porcine mucosal cells was empirically determined to give the greatest yield of IRGIP cells.  Other methods  attempted, such as tissue homogenization, were not successful. The duration of time between bleeding the animal and retrieving the tissue into cold Hank’s (period of tissue anoxia) was determined to be a critical factor for the cell survivability rate. Porcine bowel supplied by a slaughter house which was anoxic for 5-10 mm usually did not yield significant numbers of viable endocrine cells. In addition, these animals were usually not fasted prior to slaughter, and the presence of pancreatic and intestinal enzymes in the intestinal lumen undoubtediy contributed to tissue degradation. While the enzymatic digestion used in dispersing the canine mucosal cells undoubtedly resulted in cell damage, collagenase on its own was not sufficient. With just type I collagenase, 8-10 digests were required to disperse the mucosa and this repetitive process led to excessive cell death. In contrast, the porcine mucosa was much thinner and digestion with type I collagenase was found sufficient to disperse the cells after 3-4 digests. The enzyme mixture used for canine mucosa completely digested the porcine tissue in 1-2 digests, but the cells did not recover from this process. In both the canine and porcine experiments, extracts of cells were kept from each digestion for assay of IRGIP content. It was usually noted that the later digests (3 and 4) had the greatest concentration of IRGIP. Early digests were observed to contain greater numbers of red blood cells, cellular debris, and non-viable cells. Viability of cells from later digests was —98%, as assessed by trypan blue exclusion.  2.3.2 ENDOCRINE CELL ENRICHMENT AND CULTURE The elutriation procedure yielded two main fractions, designated Fl and F2. The Fl fraction contained most of the bacteria, red blood cells, and debris that was present in  66 the digested cells. The F2 fractions for both porcine and canine cells were collected under similar elutriation conditions, and made up a population of cells of similar size. The IRGIP and IRSS content of 5 x 106 canine cells from the eluted fractions was compared to that of the same number of cells from a pre-elutriation (PE) sample (Figure 3A). While there was no significant difference between the PE and Fl fractions, the F2 fractions contained approximately 3 times the IRGIP and IRSS concentrations as the PE fractions  (p<O.005).  The canine F2 fraction (5 x 106 cells) contained 38 ± 6 ng IRGTP  and 7.0 ± 0.6 ng IRSS. A similar enrichment of IRGIP was observed in the elutriated porcine cells (Figure 3B), with the F2 fraction (15.4 ± 2.5 ng) containing —3 times the PE fraction (5.0 ± 0.8 ng). Although IRSS was not consistently measured in the porcine extracts, preliminary results suggest that while there was enrichment, considerably less IRSS was present when compared to the same number of eluted canine cells (Fl: 52 pg, F2:321 pg;n=2). The culture conditions were found to be an important factor in the success of the peptide secretion experiments. Collagen coated plates provided a suitable substrate for endocrine cells to attach to. These cells did not adhere to non-coated plates, and other substrates, such as poly-l-lysine, laminin, and fibronectin produced various degrees of success. As only 20  -  25% of the cells in each well adhered, the maximum number of  cells deemed possible to culture (5 x 106) were added to each well. This was important in order to attain sufficient numbers of IRGIP cells to produce measurable levels of IRGIP. Attempts at culturing the dispersed and enriched cells in large flasks and then transferring to Eppendorfs for release experiments were also not successful. After 40 h culture, 98% of adherent cells were viable as assessed by trypan blue exclusion.  The viability  decreased rapidly as cells were cultured longer. Extracts of canine cells contained 11.5 ± 2.5 ng IRGIP/well and 1.4 ± 0.2 ng IRSS /well, while porcine cells contained 6.6 ± 0.5 ng IRGIP/well.  67 23.3 IMMUNOCYTOCHEMISTRY OF INTACT MUCOSA AND CULTURED CELLS Figure 4 shows canine jejunum immunostained for IRGIP using the peroxidase method. The IRGIP cells, shown by arrows are sparsely distributed, and account for  <  0.1% of the cells of the intact mucosa. A similar observation was made for IRSS distribution (data not shown). Figures 5 and 6 show the isolated canine mucosal cells after enrichment and 40 h culture, immunostained using the peroxidase method for GIP and SS, respectively. At the time of plating, the F2 fraction contained mainly single, spherical cells, but after culture, endocrine cells appeared to flatten and adopt a more ovoid or triangular cell appearance (see Figure 5 & 6). During the culture period, adhered clusters of up to 200 cells formed, including the endocrine cells, and these cells were observed to form connections with neighbouring cells. IRGIP cells accounted for approximately 10% of the adherent canine cell population, while IRSS cells accounted for 5% on average. A similar proportion of cultured porcine cells stained positive for IRGJP (8%), while IRSS cells added up to —1% of the adherent population (data not shown). A large portion of cells in the intact mucosa were mucin containing (Figure 7), and accounted for —80% of the adherent canine (Figure 8) and porcine cell population. Less than 1% of the canine or porcine cultured cells stained for motilin, secretin, gastrin/CCK, glicentin, or neurotensin (data not shown).  2.3.4 IRGIP SECRETION IN RESPONSE TO K The effect of depolarization of the isolated canine and porcine endocrine cells was investigated using K. Figure 9 shows the IRGIP secretory response to increasing concentrations of K from 10 to 55 mM over a 2 h period from canine and porcine cells. Basal release of IRGIP in 5 mM glucose, 5 mM K was 2.7 ± 0.4 %TCC (310 ± 45 pg/ml) for canine cells and 4.6 ± 0.4 %TCC (304 ± 23 pg/ml) for porcine cells. Potassium concentrations ranging from 20 to 55 mM significantly stimulated IRGIP release from canine cells in a concentration-dependent fashion when compared to basal  68 (p<O.O5). A similar response was observed in the porcine cells, however, a significant increase in IRGIP release was also observed at 10 m’vI K (p<O.O5). Stimulation at 55 mM K was 18.8 ± 2.1 %TCC (2.2 ± 0.2 ng/weli), or approximately 7 times basal release  ) for canine cells and approximately 4 times basal release (p<O.OOl) for porcine 1 (p<O.OO cells (19.9 ± 2.0 %TCC or 1.3 ± 0.1 ng/well). Basal release of IRSS from canine cells was 2.0 ± 0.2 %TCC (28 ±3 pg/mi) and increased significantly by 55 mM K to 4.1 ± 0.3 %TCC (57 ±4 pg/well; p<O.O5, n =3, data not shown).  23.5 IRGIP SECRETION INRESPONSE TO GLUCOSE Figures 10 and 11 show the effects of graded glucose concentrations from 5 to 20 mM on IRGIP secretion from canine and porcine cells, respectively. In the canine cells, the release of IRGIP was enhanced significantly (p<O.O5) over basal by glucose levels of 15 and 20 mM, while the addition of 10 mM glucose had no significant effect. In contrast, IRGIP release from porcine cells was significant at glucose concentrations  10  mM (p.cz0.O5). With the addition of 10 p.g/well somatostatin antibody SOMA 10 to the canine cells, IRGIP release in the presence of 5 mM glucose significantly increased to 6.1 ± 0.7 %TCC (p<O.O5), comparable to that induced by 20 mM glucose alone (Figure 10). Glucose had a minimal effect on IRSS release from canine cells, yielding a significant increase only at 20 mM glucose (from 2.0 ± 0.2 %TCC to 2.8 ± 0.3 %TCC; p<O.05, n  =  4, data not shown).  2.3.6 IRGIP SECRETION IN RESPONSE TO A23187 The involvement of Ca 2 in the release of IRGIP was investigated using the Ca 2 ionophore A23 187. Figure 12 shows the effect of A23 187 at concentrations of 1, 5, and 10 .tM on IRGIP release from canine epithelial cells in the presence of 5 mM glucose. While the addition of 5 p.M A23 187 produced only a small but significant (p<O.05) increase over basal IRGIP release, the addition of 10 p.M A23187 resulted in an IRGIP  69 output approximately 4.1 times basal (to 11.2 ± 1.8 %TCC;  p<O.Ol).  IRSS release was  increased significantly by 10 pM A23 187, from 2.0 ± 0.2 %TCC to 3.9 ± 0.6 %TCC (p<O.O5, n =4, data not shown).  2.3 .7 JR GIP SECRETION INRESPONSE TO GRP To investigate the potential for peptidergic control of IRGIP secretion, the effect of graded concentrations of porcine GRP on IRGIP release from canine epithelial cells in the presence of 5 mM glucose was examined (Figure 13). GRP significantly (p<O.O5) stimulated IRGIP release in a concentration-dependent fashion with significant release occurring at a concentrations of 1 nM and greater. At 100 nM GRP, IRGIP release was 7.2 ± 1.0 %TCC, approximately 2.7 times basal (p<O.Ol). This concentration of GRP also significantly increased IRSS release from 2.0 ± 0.2 %TCC to 2.9 ± 0.4 %TCC (p<O.O5, n =4, data not shown).  2.3.8 IRGIP SECRETION IN RESPONSE TO FORSKOLIN A possible role for adenylate cyclase in IRGIP secretion was investigated using forskolin. Figure 14 shows IRGIP release from canine epithelial cells in response to forskolin concentrations ranging from 0.1 to 100 .tM. Forskolin at a concentration of 1 JiM significantly (p<O.O5) increased IRGIP release over basal. The maximum effect observed was achieved with the highest forskolin concentration tested (100 JiM) where IRGIP release was approximately 4.9 times basal (to 13.1 ± 1.8 %TCC; p<O.Ol). Concentrations of forskolin from 1-100 JiM also significantly increased IRSS release, with 100 JiM yielding 3.3 ± 0.5 %TCC compared to 2.0 ± 0.2 %TCC in the absence of forskolin (p<O.05, n =4, data not shown).  70  Peptide  Dilution  Source  Species  Gastrin/CCK Glucagon  1:100  Walsh  1:100  Gregor  Mouse Mouse  Secretin  1:500  Polak  Rabbit  Somatostatin Insulin  1:1000 1:1000  RPG  Mouse  RPG  Guinea Pig  Glicentin  1:1000  Moody  Rabbit  Neurotensin  1:2000  RPG  PP  1:2000 1:5000  Buchanan  Mouse Rabbit  RPG  Mouse  1:5000  RPG  Rabbit IgG Mouse IgG  1:300 1:300  Vector  Mouse Goat  Vector  Goat  Guinea Pig IgG  1:300  Vector  Goat  GIP Motilin  TABLE 1: Antisera used in this thesis. CCK cholecystokinin; GIP gastric inhibitory polypeptide; PP pancreatic polypeptide RPG Regulatory Peptide Group; Dr. 3. Walsh Cure, UCLA; Dr. J. Polak Hammersmith Hospital, RPMS; Dr. A. Moody Novo Nordisk, Cophenhagen; Dr. K. Buchanan University of Belfast; Vector Laboratories Inc. Burlingame CA. -  -  -  -  -  -  -  -  -  71 A  50-  *  IRGIP  Q  i  IRSS  .  20-  T  T T  1o. 0.  I  I  PE  Fl  I  F2  FRACTION B 20.  In  10-  c5.  0. PE  I  I F2  FRACTION FIGURE 3: A) IRGIP and IRSS content (ng/5 x 106 cells) from the canine single cell suspension prior to elutriation (PE) and the two fractions resulting from eluthation (Fl & F2). For IRGIP n = 9, for IRSS n = 4. B) IRGIP content from isolated porcine cell fractions. n = 10. * represents significance to a level of p<O.05 when compared to PE.  72  c-,/ ‘3 1  I,.  -  i,tfç  .  —  ,  :v.  \  p. L’;:  i\ .  )  ‘:  FIGURE 4: Canine jejunum immunostained for GIP using the peroxidase method. Magnification is 50 x. CM = circular muscle, MM = muscularis mucosae, SM = submucosa, V = villus.  73  f  /  ,  FIGURE 5: Isolated canine epithelial cells after 40 h culture, immunostained for GIP. Magnification is 400 x. Arrow indicates a positive cell.  74  ,_  r  FIGURE 6: Isolated canine epithelial cells after 40 h culture, immunostained for somatostatin. Magnification is 200 x. Arrow indicates a positive cell.  75  FIGURE 7: Canine jej unum stained for neutral (red) and acid (blue) mucins using  a combined alcian blue-periodic-acid Schiff technique. Magnification is 50 x.  76  •  4% •LF  :/  LL  V 41 liir•\. -  •  •.  ••  FIGURE 8: Isolated canine epithelial cells after 40 h culture, stained for neutral (red) and acid (blue) mucins using a combined alcian blue-periodic-acid Schiff technique. Magnification is 400 x.  77  25.  CANINE  * *  PORCINE  20.  I * *  15. 10.  0 5  I  I  35  55  K (mM)  FIGURE 9: Effect of K concentrations of 10 to 55 mM on basal IRGIP release (5 mM K+) from isolated canine and porcine endocrine cells in the presence of 5 mM glucose. In this and subsequent figures, levels are expressed as mean ± SEM %TCC. responses were compared to basal release (5 mM glucose, 5 ni?vI K+). *  p<0.05, canine n =5, porcine n  =  8.  IRGIP  78  8-  0  —r 5  —  —r —r—r5+cS 10 15 GLUCOSE (mM) —  —  —  —r 20  -  FIGURE 10: Concentration-response relationship between glucose (5 to 20 mM) and IRGIP secretion from isolated canine epithelial cells. 5+czS indicates 5 mM glucose with 10 jig/ml anti-somatostatin monoclonal antibody (n compared to basal. *p<005, n 6.  =  4).  IRGIP responses were  79  15 *  *  *  0 10  15  20  GLUCOSE (mM)  FIGURE 11: Concentration-response relationship between glucose (5 to 20 mM) and IRGIP secretion from isolated porcine epithelial cells. IRGIP responses were compared to basal. *p<O.05, n =7.  80  15. *  *  5.  —  A23187 (jiM)  FIGURE 12: Effect of 1, 5, and 10 jiM A23187 on basal release of IRGIP in the presence of 5 mM glucose. IRGIP responses were compared to basal. *p<0.05, n = 10.  81  10. *  8  J  c.)  *  *  -I  4. 20-  —  0  0.1  1  1  10  —  100  GRP (nM)  FIGURE 13: Effect of graded concentrations of GRP (0 to 100 nM) on IRGIP release from cultured canine epithelial cells. IRGIP responses were compared to basal. *p<O.05, n =4.  82  15. 1  *  _I__  * _:r_  0  —  0  0.1  1  —r 10  —  100  FORSKOLIN (j.tM)  FIGURE 14: Effect of 0.1 to 100 pM forskolin in the presence of 5 mM glucose on IRGIP release from isolated canine epithelial cells. IRGIP responses were compared to basal. *p<O.05, n =4.  83 2.4 DISCUSSION Although it is known that IRGIP is released into the circulation in response to a meal, the mechanisms by which the components of a meal in the lumen of the small intestine act to cause GIP secretion are unknown. A greater understanding of these mechanisms will only become available with in vitro experiments to explore the direct effect of various substances on the GIP cell. Studying the release of IRGIP at the cellular level requires enrichment of the otherwise diffusely located endocrine cells or the use of tumor-derived cell lines which express IRGIP. The objective of these experiments was to develop a model for the isolation, enrichment and culture of canine and porcine intestinal endocrine cells. Dispersion of canine and porcine duodenal and jejunal mucosa by sequential enzyme digestion progressively increased the IRGIP content of the resulting cell suspension. This cell suspension was further enriched for endocrine cells by elutriation followed by culture. The culture technique has often been found useful after counterfiow elutriation to facilitate further enrichment in the studied cell content by providing selective attachment of the endocrine cells to the culture dishes (Aponte et al., 1988; Barber et a!., 1986; Buchan et a!., 1987; Giraud et a!., 1987; Koop and Buchan, 1992; Soil et a!., 1984). The overall enrichment of cultured endocrine cells was as much as 100-fold, with IRGIP and IRSS cells accounting for 10% and 5% respectively of the adherent canine cells, plus 8% and 1% respectively of the attached porcine cell population. The majority of the remaining cells were mucin and goblet cells, staining positive for neutral and/or acidic mucins. The fact that 1 to 2 days was required for cells to adhere and adopt a more normal appearance indicates that these cells likely sustained damage as a result of the chemical and mechanical forces of the digestion procedure. Using similar methods, Barber et al. (1986) documented the beneficial effects of short term culture of isolated canine ileal mucosal cells. Freshly isolated neurotensin containing cells incubated at 4°C for 60 mm  84 released —12% of their neurotensin content, while after 48 h culture, basal release was —1%.  The culture period was thus important for the isolated cells to functionally  stabilize. In the present investigations, basal release of IRGIP was 2.7 ± 0.4 %TCC (canine) and 4.6 ± 0.35 %TCC (porcine), values similar to those previously reported from other comparably studied isolated cells. Membrane potential difference plays an important role in the control of many cellular processes, including endocrine secretion (Barber et a!., 1986; Matthews and Sakamoto, 1975; Taraskevich and Douglas, 1984). To assess the ability of the cells to release peptide in response to a depolarizing stimulus, cells were exposed to a series of K concentrations. Basal IRGIP release in the presence of 5 mM K was significantly increased by K concentrations of 20 mM and greater for canine cells and from 10 mM for porcine cells. Release in response to 55 mM K was approximately 7 times and 4 times that of the basal canine and porcine cell release, respectively.  This same  concentration of K doubled IRSS release compared to basal in the canine cells. K likely produces these effects by depolarizing the cells, thus opening Ca 2 channels and increasing the intracellular Ca 2 concentration, leading to peptide release. IRGP release by K indicated these cells were electhcally active and that depolarization resulted in exocytosis. Glucose is thought to be one of the most important physiological stimuli for GIP secretion in vivo. The isolated porcine and canine endocrine cells responded to graded glucose by releasing IRGIP in a concentration-dependent fashion. There was, however, no significant increase in IRGIP release from the canine endocrine cells until a concentration of 15 mM glucose was reached in the release medium. physiological diets result in luminal concentrations of glucose in dogs of  Normal <  20 mM  (Ferraris et al., 1990). Therefore, one would expect canine GIP cells to be sensitive to lower concentrations of glucose then was observed in this study. A probable explanation for the apparent lack of sensitivity of the isolated canine endocrine cells to glucose is that  85 the GIP cells may have been tonically inhibited. Unlike the intact intestinal mucosa, the isolated epithelial cells in this model are in a non-vascular system. Metabolic products and other secreted peptides that would normally be removed by the circulation remain in the medium. Immunocytochemistry of the cultured canine intestinal mucosal cells revealed that cultures contained 5% IRSS cells, and IRSS levels during basal conditions were 28 ± 3 pg/mi. IRSS release was increased modestly by 20 mM glucose to 39 ± 1 pg/mi. Although IRSS cells may not appear adjacent to IRGIP cells in the intact gut (Buchan et at., 1982), somatostatin has been previously demonstrated to inhibit IRGIP release in man (Salera et al., 1982b), dog (Pederson et a!., 1975a) and rat (Ho et at., 1987). It is possible, therefore, that endogenous somatostatin present in the medium inhibited IRGIP release.  Support for this is the observation that the addition of  somatostatin immunoneutralizing antibody (Seal et al., 1987) to basal medium significantly increased IRGIP release. The significant IRGIP release from porcine endocrine cells at a lower glucose concentration (10 mM) might reflect the fewer IRSS cells present, and resultant lower levels of endogenous IRS S. Considering that glucose transport is a requirement for GIP release (see section 1.10.2), the sensitivity of the isolated cells to glucose may also be a reflection of the activity of luminal glucose transporters. It is possible that these transporters were damaged by the digestive process, and were not functioning normally after 40 h culture. Alternatively, these transporters which exist in a glucose-free environment (during the interdigestive period), may be down regulated by 40 h culture in 5.5 mM glucose. It is also possible that the GIP cells may not possess luminal glucose transporters and thus require messages from neighbouring ‘glucose-sensitive’ cells in order to release GIP in response to luminal glucose. The degree of IRGIP response to glucose may therefore be related to the degree of reaggregation of the cells upon the seeding on collagen coated plates in order that the appropriate connections are made. Thus other factors could  86 potentially contribute to the weak IRGIP response to glucose observed from the isolated cells. The Ca 2 ionophore A23187 increases cytosolic Ca 2 concentrations. Concentrations of 5 .tM and greater were able to significantly increase IRGIP release from the cultured canine epithelial cells  .  At 10 .tM A23187, release of IRGIP was  approximately 4.1 times that of basal. These data suggest that signal transduction mechanisms in the IRGIP cell involve an increase in intracellular Ca 2 concentration. IRSS levels were also increased to twice that of basal by 10 tM A23 187. Somatostatin secretion from isolated canine fundic mucosal cells (Chan and Soil, 1988) or human antral cells (Buchan et a!., 1990) has been reported to be stimulated by treatment with A23 187. In contrast, cultured fetal rat intestinal cells did not respond to modulation of 2 fluxes (Brubaker et at., 1990). It is not known if the action of A23 187 on IRSS Ca secretion observed in the present study was direct, or whether IRGIP release in response to A23 187 stimulated IRSS release. GIP has been shown previously to stimulate gastric IRSS release (McIntosh eta!., 1981). Gasthn releasing peptide (GRP) has been located in neurons and nerve fibres of canine duodenum and jejunum (Vigna et at., 1987). In rats and dogs, GRP infusion has been demonstrated to cause IRGIP release (Greely et at., 1986a, 1986b; McDonald et at., 1981) although it was not known whether this was a direct effect on Gil? cells. GRP is thought to act through 1P 3 to release Ca 2 from intracellular stores (Gallacher et at., 1990). It was therefore of interest to determine the effects of GRP on IRGIP secretion in the isolated epithelial cell model. Concentrations of GRP from 1 to 100 nM produced a significant increase in IRGIP release over basal in a concentration-dependent fashion. This provides evidence that GRP may be involved in vivo in the control of GIP release, possibly by an IP 2 / 3 Ca mediated pathway. It is, however, notable that other isolated endocrine cell preparations have yielded significant responses to GRP at concentrations as low as 0.01 fM (Campos et a!., 1990). It is possible, therefore, that IRGIP release in  87 response to GRP observed in the present study was mediated by a different subtype of the GRP receptor, or via a related receptor, such as that for neuromedin B. The addition of 100 nM GRP also resulted in a modest increase in IRSS release (to 2.9 ± 0.4 %TCC). GRP infusion in dogs has previously been demonstrated to increase plasma IRSS (Schusdziarra et a!., 1980). To determine if activation of cAMP-dependent intracellular pathways would result in IRGIP secretion from the cultured canine epithelial cells, forskolin, an activator of adenylate cyciase was tested. Forskolin at concentrations of 1 to 100 iM produced significant increases in IRGIP release when compared to basal. The effect of 100 p.M forskolin was disproportionally higher than at lower concentrations tested, and may be reflective of the lack of specificity of forskolin for adenylate cyclase at this dose. High concentrations of forskolin may also affect a variety of ion channel functions, independent of cAMP systems (Leidenheimer et at., 1990; Ticku and Mehta, 1990; Watanabe et a!., 1987). With these limitations in mind, it appears that canine GIP cells are responsive to activation of adenylate cyclase. Whether the effects on stimulation of 2 and cAMP systems yield secretory responses that are additive or potentiating is a Ca subject for further investigation.  Forskolin concentrations of 1 to 100 p.M also  significantly increased the release of IRSS (p<O.O5). Somatostatin release in response to dibutyryl cAMP or forskolin has been reported in isolated canine fundic somatostatin cells (Soil et a!., 1984) in human antral somatostatin cell cultures (Buchan et at., 1990), and in fetal rat intestinal cells in culture (Brubaker et al., 1990). In conclusion, a method has been established to enrich duodenal/jejunal endocrine cells, yielding large adhered cell clusters in culture of approximately 10% IRGIP in the canine and 8% IRGIP in the porcine preparation. Both depolarization by K and nutrient stimulation with glucose yielded an increase in IRGIP release from the cultured canine and porcine cells.  In addition, pharmacological investigation of potential signal  transduction systems involved in stimulus-secretion coupling in canine GIP cells  88 suggested the participation of both cAMP and Ca 2 dependent pathways. Evidence was -  also provided for a direct role of GRP on receptor-dependent stimulation of IRGIP release. The use of a somatostatin immunoneutralizing antibody indicated that basal IRGIP release was tonically inhibited by exogenous IRS S. The fewer IRSS cells in the isolated porcine cells may, therefore, have contributed to the greater basal release and sensitivity to lower glucose or K concentrations. IRSS secretion was significantly stimulated by high levels of all the above agents used to examine IRGIP release, but it is unclear if these are direct actions or a result of endogenously released peptides such as GIP. This model may prove useful for further elucidating the cellular mechanisms controlling the release of the intestinal endocrine hormones OW and somatostatin. Furthermore, it may be possible to adapt these methods to isolate epithelial cells from the rat, where models for disease states potentially characterized by abnormal GIP release, such as obesity, exist. Such studies at the cellular level may be the only way to resolve possible aberrant OW release, and the mechanisms responsible.  89 CHAPTER 3 GIP RELEASE FROM A TUMOR-DERIVED CELL LINE (STC 6-14)  3.1 INTRODUCTION The rationale for studying GIP release at the cellular level was presented in chapter 2. By using cell preparations, the culture environment can be controlled and progressive changes in intracellular and intercellular events can be directly monitored (Hassall et a!., 1989). A major factor impeding such studies is the diffuse disthbution of the entero-endocrine cells (including IRGIP cells) throughout the small intestine. One strategy employed has been the development of methods for the isolation enrichment and culture of gut endocrine cells, and the adaptation of these methods to specifically examine GIP release (chapter 2). While this model has proven useful for studying the mechanisms controlling IRGIP release, endocrine cell yield remains a critical problem. Recently, hormone secreting cell lines have been obtained from neuroendocrine tumors of transgenic mice expressing oncogenes in neuroendocrine cells. The general strategy involves the construction of a hybrid gene composed of the regulatory region from a hormone gene linked to the coding region of an oncogene. This hybrid gene is then transferred into a mouse germ line via microinjection of fertilized eggs and the resulting transgenic mice are able to express the oncogene in the cell types that normally express the hormone regulatory sequences (Hanahan, 1988). Recently, two lineages of transgenic mice were crossed to produce double transgenics. One of these lines carried a hybrid gene construct linking the rat insulin promoter, which drives expression in pancreatic B-cells, to the SV4O early region encoding the potent oncogene large Tag (RIP lTag2; Hanahan, 1985). The other line carried a polyoma small T antigen gene linked to the rat insulin promoter (RIP2PyST1; Seth et a!., 1991). Progeny of these mice followed two distinct tumorigenesis pathways, either resulting in the development of encapsulated B-cell tumors of the pancreas, or unexpected neuroendocrine cell tumors of  90 the small intestine (Grant et a!., 1991). These tumors were excised, minced and injected as small clumps into mice to yield cell lines (Grant et a!., 1991). The tumor cells did not stain with periodic acid-Schiff/alcian blue, indicating a lack of mucus secretion, but retained strong positivity for the Grimelius’ silver method and chromogranin A, revealing their endocrine nature (Rindi et al., 1990).  It appeared that a well differentiated  endocrine cell producing secretin was the original target of transgene expression in the gut of RIP1Tag2IRIP2PyST1 mice, as T-antigen-positive mucosal neuroendocrine cells only expressed secretin (and not the other tested hormones). A high proportion of cells stained for secretin, and thus the name ecretin producing lumor Lells (or -Iype Lell; abbreviated STC) was given to the cell line. Other cells stained for the different areas of the proglucagon molecule (GLPI, GLPII, glicentin and glucagon) and scattered rare cells reacted to neurotenisn and pancreatic polypeptide antisera, while no immunoreactivity was detected for insulin in the STC-1 cells (Rindi et al., 1990). Experiments reported in this chapter explored the use of this tumor cell line to study GIP release at the cellular level.  3.2 MATERIALS AND METHODS 32.1 CULTURE AND PRODUCTION OF STC 6-14 CELL LINE The STC-1 cell line was kindly supplied by Dr. D. J. Drucker (Banting and Best Research Group, Department of Medicine, Toronto Ont).  Cells were cultured in  Dulbecco’s modified Eagle’s medium (DMEM; Terry Fox Laboratory, Vancouver, BC) supplemented with 25 mM glucose, 2.5% fetal calf serum (FCS; Sigma Chemical, St. Louis, MO), 12.5% horse serum (HS; Gibco Laboratories, Burlington, Ont.), 1-glutamine (Sigma, 0.15 mg/mi), penicillin (Gibco, 50 units/mi), and streptomycin (Gibco, 50 p.g/ml) at 37°C in 5% CO 2 in air in 250 ml Falcon tissue culture flasks (Becton Dickinson Canada Inc., Toronto, Ont.).  The cells were subcultured as required or when  approximately 80% confluency was reached, by harvesting with trypsin-EDTA (Sigma).  91 Some cells were cultured on collagen coated 12-well plates until —80% confluency was achieved and fixed in Bouins for immunocytochemical analysis, as described in section 2.2.5. STC-1 cells were also routinely frozen in media containing DMSO and stored in liquid nitrogen. Data from immunocytochemical studies indicated that a small number of STC- 1 cells were immunoreactive for GIP. In an attempt to increase the proportion of cells expressing GIP, the STC- 1 cells were sub-cloned. Cells were plated at an average of 1 cell/well in 96 well Falcon culture plates (10 plates; Becton Dickinson Canada Inc.) and cultured as above to produce clones. Supematant from each well containing clones was stored at -20°C for later IRGIP determination using the GIP RIA described in section 2.2.8. Cells from RIA-positive wells were transferred to progressively larger wells to increase cell numbers, and also immunostained to verify the presence of IRGIP containing cells (see section 2.2.5 for methods). Sub-clones were produced with varying proportions of IRGIP expressing cells. However, those clones with a majority of IRGIP cells appeared slow growing and difficult to culture. A stable clone, STC 6-14, was therefore chosen as the cell line to use for IRGIP release studies. The STC 6-14 cells were transferred to culture flasks and maintained as above to establish sufficient cell numbers for experiments to be conducted.  Cells were routinely frozen in media  containing DMSO and stored in liquid nitrogen. The cells were subcultured as required or when approximately 80% were confluent, by harvesting with trypsin-EDTA. Cells from passages 26-39 were used in this study.  32.2 RELEASE EXPERIMENTS For IRGIP secretory studies, 5 x iO cells in 1 ml of culture medium were plated in each well of collagen coated Costar 12-well plates and cultured for 4 days. On the experimental day, the medium was removed by pipet, and changed to 5 mM glucose DMEM (prepared as described above) and incubated 4-5 h. The culture medium was  92 then removed by pipet and the cells washed with 1 ml release medium consisting of DMEM supplemented with 5 mM glucose, 1% FCS, and 2% aprotinin (Trasylol, 200 kallikrein-inhibiting U/ml, Miles Pharmaceuticals, Rexdale, Ont.). This was followed by addition of 1 ml of release medium containing the appropriate stimulus to each well. Each stimulus was tested in triplicate, for n  =  1. The cells were incubated for 2 h, after  which the release medium was removed and stored in 1.5 ml Eppendorf microcentrifuge tubes on ice. Samples were centrifuged for 5 mm at 7000 x g (Baxter Biofuge) at 4°C and the supematant stored at 200C for subsequent peptide determination. Adherent cells in control wells (2 wells per plate) were extracted in 2 N acetic acid for determination of total cell peptide content per well. After boiling 10 mm and centrifugation to remove particulate matter, the supernatant was stored at -20°C for subsequent IRGIP or immunoreactive SS (IRSS) assay. IRGIP and IRSS secretory values from each well were expressed as percentage of total cell content (%TCC) of IRGIP or IRSS as measured from extracts. Control plates that were not used for the secretion experiments were fixed with Bouins for 10 mm and then stored in phosphate buffered saline (PBS) to be subsequently examined for peptides by immunocytochemistry.  32.3 SECRETAGOGUES Porcine GIP (pGIP; Regulatory Peptide Group, Vancouver, BC) was diluted to a concentration of 4 .tM with glucose-free DMEM release medium and kept on ice. This solution was serially diluted to 400, 40, and 4 nM. Final concentrations of pGJP per well were obtained by adding 25 jil of the different pGIP solutions to 1 ml medium in each well. An appropriate volume of a solution of 400 nM SS (SS-14; Peninsula, Belmont CA) was added to release media to give a final concentration of 10 nWwell before the media were added to the appropriate wells. The SS antibody (SOMA-lO; Regulatory Peptide Group) was diluted in release medium to 250 p.glmllwell. Synthetic porcine GIP used for HPLC standard was obtained from Peninsula.  93 32.4 PEPTIDE QUANTIFICATION Peptides (IRGIP & IRSS) were quantified by RIA as described in sections 2.2.8 and 2.2.9. Acetic acid extracts were adjusted to pH —7 with 10 M NaOH prior to assaying.  32.5 IMMUNOCYTOCHEMISTRY The STC 6-14 cells were immunostained with primary antibodies for all hormones listed in Table 1 using the avidin-biotin peroxidase method as previously described in section 2.3.3.  3.2.6 HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY (HPLC) Reverse-phase HPLC was used to characterize the IRGIP extracted from STC 6-14 cells, and in the culture medium. A pellet of 1.5 x iO cells was extracted with 5 ml 2 N acetic acid, followed by boiling 10 mm  and centrifugation (7000 x g, 10 mm).  The  supernatant was then dried using a SpeedVac (Savant Instruments Inc., Farmingdale, NY) and stored at -20°C. Medium that had cultured STC 6-14 cells for 4 days was also processed for HPLC analysis. Samples of 20 ml were adjusted to pH —2 with acetic acid. SepPak cartridges (C ; Waters) were prepared by priming with 5 ml acetonitrile plus 0.1% 18 thfluoroacetic acid 3 (CH C N  +  0.1% TFA), followed by 10 ml H 0 (plus 0.1% TFA). The 2  media samples were then applied to the cartridge and washed with 50 ml H 0 plus TFA, 2 followed by 20 ml 10% 3 CH C N plus TFA. It was then eluted with 2 ml 60% 3 CH C N plus TFA, and the products were dried by Speed Vac and stored at -20°C. An aliquot of purified culture medium or the cell extract (10 .tg/jil) was dissolved in distilled filtered water and 75 .tl of this solution was loaded onto a tBondpak 1 C 8 column (4.6 x 250 mm; Waters). The HPLC equipment (Waters) consisted of a model 512 WISP autoinjector, two model 510 pumps, a model 441 absorbance detector, and a NEC computer and printer for signal processing using the Waters Maxima 820 program. The loaded material was eluted using a  94 gradient that had previously been empirically determined to resolve the components of natural porcine GIP. Samples were loaded at 28% CH CN plus 0.1% TFA, and after 1 mm, 3 were eluted by a CH CN gradient (28 3  -  33% in water) containing 0.1 % TFA, run over 10  mm at a flow rate of 1 mI/mm. The column was then maintained a further 3 mm at 33% CH C 3 N followed by a 4 mm gradient up to 70% CH CN. After 2 mm at 70% CH 3 CN, it was 3  then re-equilibrated to 28% 3 CH C N over 2 mm. The run was complete after a 10 mm period of equilibration, and was followed by 100 p.1 H 0 injected and subjected to the same gradient 2 to ensure no residual GIP remained on the column and was collected in the next sample run. Fractions were collected every 0.5 mm and were dried using a SpeedVac prior to being stored at -20°C. Samples were then dissolved in 500 p.1 assay buffer prior to being assayed for IRGIP.  32.7 EXPRESSION OF RESULTS Release data are presented as mean ± SEM for a percentage of the total cell content (%TCC). Statistical significance was determined using the Student’s t test and was set at the 5% level.  3.3 RESULTS 3.3.1 CHARACTERIZATION OF STC 6-14 CELLS Sub-cloning of the STC- 1 cell line resulted in a stable cell line, designated STC 614, that consisted of approximately 30% IRGIP cells (Figure 15). Roughly the same proportion of cells exhibited SS immunoreactivity, and <2% immunostained for either glucagon, gastrin, secretin, glicentin, pancreatic polypeptide, or neurotensin.  No  immunoreactivity for insulin was observed. In order to examine IRGIP expressed in STC 6-14 cells, HPLC analysis of cell extracts was employed. The narrow gradient was chosen to resolve the components of natural porcine GIP in a fashion similar to that which had been previously reported (see section 1.1 1.1). Three major components of  95 natural porcine GIP were resolved, with retention times of 11.37, 12.47 and 13.12 mm respectively, while synthetic porcine GIP eluted as a mono-component at 13.01 mm (Figure 16). For following runs, fractions were collected and assayed for IRGIP, as samples were below the absorbance detection limits of the HPLC system. Analysis of a natural porcine Gil? standard in this fashion yielded 3 IRGIP peaks (a, b, c), the second of which (b) eluted in the same fraction as synthetic porcine GIP (Figure 17A). On a separate elution, RIA analysis of collected fractions revealed 1 major IRGIP peak for the STC 6-14 cell extract, but 3 major IRGIP peaks (i,  iii) for the STC 6-14 culture  medium (Figure 17B). While the IRGIP in the cell extract was present in the same fractions as synthetic porcine GIP 1-42, the 3 IRGIP peaks from cell medium were in different fractions. The largest IRGIP fraction of the cell medium (peak i) eluted at the same fraction as peak a of natural porcine G]P.  33.2 IRGIP SECRETION INRESPONSE TO GLUCOSE The effects of SS and SS monoclonal antibody (SOMA-lO) on glucose stimulated IRGIP release are shown in Figure 18. After 4 days culture, each well of cells on average contained 33.3 ± 1.4 ng of IRGIP, and basal release in the presence of 5 mM glucose was 733 ± 58 pgml-’2h’ (2.20 ± 0.17 %TCC).  Incubation of the cells at glucose  concentrations of 10 mM and higher resulted in significantly greater IRGIP release than at 5 mM glucose, to a maximum as tested of 4.20 ± 0.42 %TCC at 20 mM. In the presence of 10 nM SS however, glucose had no significant effect on IRGIP release. The addition of SOMA-lO to the graded glucose concentrations resulted in significantly greater IRGIP responses at all but 20 mM glucose. At 5 mM glucose, IRGIP release was increased by the addition of SOMA-lO to a level comparable to that obtained for 15 mM glucose alone.  96 3.3.3 IRSS SECRETION IN RESPONSE TO GLUCOSE AND GIP The release of IRSS from STC 6-14 cells in response to 4 concentrations of glucose is shown if Figure 19. The average content of IRSS in each well after 4 days culture was 18.4 ± 1.5 ng. Basal release of IRSS was 2.05 ± 0.19 %TCC, and increased significantly in response to 15 and 20 mM glucose. The addition of natural porcine GIP at concentrations of 1 nM and greater to release medium containing 5 mM glucose also significantly increased IRSS release in a concentration dependent fashion (Figure 20). At the highest pGIP concentration tested (100 nM) IRSS release was doubled (4.10 ± 0.44 %TCC).  97  • —  I  L •,.  -  “  F.  a  •4 —  1  b•.  ,..  —  . •  .  -  Ii) c(-  .IJ  •  FIGURE 15: Cultured STC 6-14 cells immunostained for GIP (black), x 400.  98  -5.6  -5.7 LI)  .nGIP sGIp -5.9  20 TIME (mm)  FIGURE 16: HPLC elution profile of natural porcine GIP (5 .tg) and synthetic porcine GIP (3 tg) from a C 18 column with an acetonitrile gradient of 28-33% over 10 mm, at a flow rate of 1 mi/mm.  99 A b  40  NATURAL pGIP  3  —0---—  SYNTHETIC pGIP  2 a C  1  0 10  30 FRACTION  40  50  B 30 STC 6-14 MEDIA I  —0——  STC 6-14 EXTRACT  120  11  1O 111  0 0  10  20 30 FRACTION  40  50  FIGURE 17: Immunoreactive GIP (IRGIP) content of HPLC fractions eluted from a C 18 column with an acetonitrile gradient of 28-33% over 10 mm. A) Overlay of elution profiles for natural porcine GIP and synthetic porcine GIP. B) Overlay of elution profiles for STC 6-14 cell extract and culture medium.  100  7 6 c.)  5 4 3 2 1 0 GLUCOSE (mM)  FIGURE 18: Effect of 250 pgfm1 anti-somatostatin antibody (SOMA-lO) or 10 nM somatostatin (SS) on the dose-response relationship between glucose concentrations (5 to 20 mM) and IRGIP secretion from STC 6-14 cells. For glucose alone, n = 10; for glucose anti-SS, n  7; for glucose + 10 nM SS, n = 4. In this and subsequent figures, IRGIP levels are expressed as mean ± SEM percent total cell content (TCC). IRGIP responses were compared to 5 mM glucose alone. * p< 0.05. +  =  101  4  *  *  3  Ci) Ci)  1-  0.  —  5  I  —  I  —  10 15 GLUCOSE (mM)  I  -  20  FIGURE 19: Concentration-response relationship between glucose concentrations (5 to 20 mM) and immunoreactive somatostatin (IRSS) secretion from STC 6-14 cells. n = 9. IRSS responses were compared to 5 mM glucose. * p< 0.05.  102  S  * *  4 *  c-)  3  Ci)  2 1 0 0  0.1  1  10  100  GIP (nM)  FIGURE 20: Effect of graded concentrations of porcine GIP (0-100 nM) on immunoreactive somatostatin (IRSS) release from STC 6-14 cells in the presence of 5 mM glucose. n =7. IRSS responses were compared to 5 mM glucose in the absence of GIP.  *  p< 0.05.  103 3.4 DISCUSSION Cell lines producing gut peptides are able to proliferate in culture and can provide large numbers of viable tumor cells free of contamination by other cell types. They are therefore valuable models for studying the synthesis, storage and release of gut peptides in definable environments. The STC-1 cell line was derived from an intestinal endocrine tumor that developed in mice carrying transgenes consisting of the rat insulin promoter linked to the potent viral oncogene SV4O T antigen and to the polyoma virus small T antigen (Rindi et a!., 1990). This cell line has been used previously to study the release of cholecystokinin (CCK) and secretin (Chang et al., 1992; Mangel et al., 1993). In these studies, it was concluded that the STC-1 cells may serve as a useful model system to investigate the secretory mechanisms for these peptides. In the present study, these primarily ecretin-secreting tumor ells were sub-cloned to enrich the population of IRGIP secreting cells, resulting in the STC 6-14 clone. Immunohistochemical analysis of the STC 6-14 cells revealed that IRGIP and IRSS were the main endocrine cell types comprising the cell line, and that small numbers of glucagon, gastrin, secretin, glicentin, pancreatic polypeptide, and neurotensin were also present. The fact that the STC cell lines secrete a number of peptides must be considered.  Heterogeneity of hormone expression might be related to the known  plasticity of endocrine cells when transformed (Tischler, 1983). Rindi et a!. (1990) noted that the spectrum of hormones expressed apparently increased with tumor progression, suggesting that the proliferating intestinal neuroendocrine cell is able to switch to multiple alternative differentiated states. This quality has been ascribed to stem cells of the intestinal epithelium (Cheng and Leblond, 1974a; 1974b; Pictet et a!., 1976). Rindi et al. (1990) also noted it was surprising that in RIP1Tag2/RIP2PyST1 the regulatory sequences of rat insulin II gene induced transformation in the secretin cell, an epithelial endocrine cell anatomically and functionally quite distinct from the insulin producing Bcell. This might be explained by the fact that the pancreas and upper gut have a common  104 embryologic origin (Pictet and Rutter, 1972; Fallcmer et at., 1984). Whether more than one peptide is secreted from one cell was not determined, but is the subject of further investigation. CCK and secretin released from the STC-1 cell line have been previously shown to have the same retention times on HPLC as the naturally occurring peptides (Chang et a!., 1992). HPLC was also employed in the current investigations to characterize GIP expressed by the STC 6-14 cells. The gradient derived for the elution of GIP by HPLC was capable of resolving three main components of natural porcine GIP, similar to that which had been previously reported (Brown et al., 1981; Jömvall eta!., 1981; Schmidt et a!., 1987). The main component had a retention time comparable to that of synthetic porcine GIP, and was therefore assigned GIP 1-42. The other two peaks were identified by comparisons to previous HPLC profiles of natural porcine OW (Brown et a!., 1981; Jörnvall et al., 1981; Schmidt et a!., 1987). These reports identified the two preceding peaks as CCK and GIP 3-42, respectively. As the content of GIP in the STC 6-14 culture medium and extracts loaded on the HPLC (using the same gradient conditions) was below the absorbance detection limit, eluted fractions of subsequent runs were collected, examined for GIP immunoreactivity, and compared to identically treated GIP standards run the same day. Three main IRGIP components were resolved from natural porcine GIP. The largest component (b) was found in the same fraction as synthetic porcine GIP and was assigned GIP 1-42. Peak a of the natural GIP preparation eluted prior to peak b and was hypothesized to be GIP 3-42 based on the previous reported elution profiles of this OW preparation. The difference in elution profiles between the absorbance (Figure 16) and the GIP immunoreactivity (Figure 17A) may have occurred as a result of replacing the C 18 column between these runs. IRGIP in the eluted STC 6-14 extract was found in the same fractions as synthetic porcine GIP and was assigned GIP 1-42. However, IRGIP in eluted STC 6-14 culture medium was found in three fractions, the largest of which (peak i) eluted in the same fractions as peak a of the natural GIP,  105 believed to be GIP 3-42. The other IRGIP peaks of the STC 6-14 culture medium could represent putative forms of GIP that were previously suggested to exist (see section 1.9.3), or could be other 1RGIP fragments. Thus while 1RGIP contained in the STC 6-14 cells does appear to have identical HPLC characteristics to synthetic GIP, IRGIP purified from the cell culture medium appears to have been degraded, possibly largely to GIP 342. This phenomenon is further investigated in chapter 4 of this thesis. Glucose is thought to be one of the main physiological regulators of GIP secretion in vivo. Concentrations of glucose from 10 to 20 mM significantly increased IRGIP release from STC 6-14 cells, however, there was only an approximately 2-fold increase in IRGIP release at 20 mM when compared to basal. This increase in IRGIP secretion is significantly smaller than that noted in the isolated canine and porcine endocrine cells (Figures 10 & 11).  It is possible that these tumor cells have glucose regulatory  mechanisms that differ from those of normal endocrine cells. Such a case has previously been observed with other tumor-derived endocrine cell lines (Brant et at., 1992; Nagamatsu and Steiner, 1992; Visher et al., 1987). In view of the high proportion of somatostatin-secreting cells present in the STC 6-14 cell cultures, it is likely that endogenous somatostatin contributes to the blunted IRGIP response observed following addition of glucose. Studies of glucose-stimulated IRGIP release in the presence of the somatostatin monoclonal antibody tend to support this hypothesis.  This antibody  significantly increased glucose-stimulated IRGIP release at glucose concentrations less than 20 mM, presumably by immunoneutralizing IRSS. Furthermore, the addition of exogenous SS was able to completely suppress glucose-stimulated IRGIP release, at all glucose concentrations tested. SS has been previously demonstrated to inhibit IRGIP release in man (Salera et a!., 1982), dog (Pederson et a!., 1975) and rat (Ho et al., 1987). The cellular mechanisms responsible for the inhibitory effect of SS on IRGIP secretion are presently not understood, probably because of the prior lack of suitable in vitro models.  106 Concomitant determination of IRSS revealed that glucose concentrations of 15 and 20 mM significantly increased IRSS release when compared to basal. Previous in vivo studies have also demonstrated the release of gut IRSS in response to luminal glucose (Schusdziarra et at., 1978). It is not clear in the present case whether this secretion is a direct result of glucose, or whether glucose-stimulated IRGIP is acting as a secretagogue of IRSS. In order to examine if IRGIP could stimulate release of IRSS from the STC 6-14 cells, pGIP was added to the basal release medium. At concentrations of 1 nM and greater, pGIP significantly increased IRSS release.  Verification that  endogenously released IRGIP stimulates IRSS release requires the development of antibodies capable of neutralizing the action of IRGIP.  GIP has previously been  demonstrated to stimulate secretion of IRSS from the perfused rat stomach (McIntosh et al., 1981). In conclusion, the tumor-derived STC 6-14 cell line represents a readily available source of large numbers of IRGIP and IRSS cells for intensive study of cellular mechanisms of hormone release. IRGIP expressed by these cells has the same HPLC retention time as porcine GIP 1-42. Glucose-stimulated IRGIP release is augmented by the addition of a somatostatin antibody, possibly by immunoneutralization of endogenous IRSS. The addition of exogenous SS suppressed glucose-stimulated IRGIP release. IRSS release from STC 6-14 cells is enhanced by the addition of glucose, but it is unclear if this increased release is secondary to glucose-stimulated IRGIP release. The addition of exogenous pGIP stimulated IRSS release in a concentration dependent fashion. The STC 6-14 cell line may thus be a useful model to further study the interactions and release of GIP and SS.  107 CHAPTER 4 METABOLISM OF GIP  4.1 INTRODUCTION Studies in chapter 3 indicated that while IRGIP extracted from STC 6-14 cells eluted with the same retention time on HPLC as native porcine GIP, GIP secreted from these cells had a significantly different HPLC profile. Furthermore, it appeared that the major constituent of IRGIP in the STC 6-14 culture medium eluted identically with GIP 3-42 that was a component of preparations of porcine GIP produced by the method of Brown et at. (1970). This observation suggested that the conversion of GIP 1-42 to GIP 3-42 might be occurring in the medium. As this culture medium contained a total of 15% serum, it was also hypothesized that the N-terminal Tyr-Ala of GIP could be the substrate for an enzyme present in serum. The physiological implication of this process was evident from the observation that GIP 3-42 separated from the natural preparations by HPLC was biologically inactive (see section 1.11.1). Jömvall et a!. (1981) suggested that the GIP fragment which constituted >20% of purified GIP was formed by secondary processing or degradation, through susceptibility to attack by aminopeptidase, elastase, dipeptidyl aminopeptidase or related enzymes in the intestine. As GIP belongs to a superfamily of peptides sharing sequence homology (see section 1.2.4), it could be predicted that such an enzyme would act on other hormones with similar N-terminal amino acid sequences.  In 1986, Frohman et al.  observed that growth hormone-releasing hormone (GRF 1-44) was cleaved to 3-44 by an enzyme in plasma. Furthermore, this enzyme product, like the GIP fragment, was biologically inactive. As a member of the glucagon superfamily, N-terminal amino acids (1 and 2) of GRF are identical to GIP (Tyr-Ala). In 1989, Frohman et at. identified the plasma enzyme responsible as dipeptidyl peptidase IV (DPP IV).  They clearly  108 demonstrated that this enzyme was present in human plasma and quickly inactivated GRF in vivo. The objective of studies in this chapter was to determine if DPP IV was also responsible for the putative production of GIP 3-42 found in the STC 6-14 growth medium, and more importantly, whether this enzyme might play an important role in GIP metabolism in vivo. Furthermore, in light of a potential role for OP in some disease states (see section 1.13), it was also of interest to determine if levels of DPP IV were altered in the obese Zucker rat, and NTDDM subjects. It was hypothesized that elevations of plasma DPP IV leading to rapid incretin degradation, might manifest as hyperglycemia due to a reduction in the biological potency of the enteroinsular axis.  4.2 MATERIALS AND METHODS 4.2.1 INCUBATION OF GIP WITH SERUM In order to test the hypothesis that the serum component of the STC 6-14 culture medium was responsible for the degradation of GIP, synthetic porcine OW (5 j.tglml; Peninsula) was incubated at 37°C in 40 mM phosphate buffer (pH 7.0), with and without serum (12.5% HS, 2.5% FCS). At time intervals of 1, 10, 30 mm; and 1, 3, 6, and 24 h, 100 jil aliquots were transferred to Eppendorf tubes, immediately flash frozen by submersion in ethanol and dry ice and then stored at -70°C. Sample volumes of 50 p1 were then analyzed by reverse-phase HPLC as previously described in section 3.2.6. These studies indicated a time-dependent breakdown of GIP 1-42 to the product believed to be GIP 3-42, in the presence of serum. Similar experiments were performed with and without the addition of 0.1 mM diprotin A (Sigma). Diprotin A is a bacterial tripeptide (Ile-Pro-ile) that has been demonstrated to competitively inhibit DPP IV (Umezawa et al., 1984) and was shown previously by Frohman et al. (1989) to greatly reduce the conversion of GRF 1-44 to GRF 3-44.  109 42.2 DIPEPTIDYL PEPTIDASE IV-NEGAT1VE RATS Recently, studies have demonstrated that Fischer-344 rats from the Japanese Charles River Inc. (DPP TV-negative) specifically lack DPP IV, whereas Fischer-344 rats from sources in the United States (DPP TV-positive) possess normal DPP IV activity (Tiruppathi et a!., 1990; Watanabe et al., 1987). It was felt that these animals would serve as ideal models with the DPP TV-negatives serving as controls, to investigate a possible role for DPP IV in the metabolism of GIP in vivo. Two breeding pairs of these rats were generously supplied by Dr. F. H. Liebach (Medical College of Georgia, Augusta, Georgia) and a colony was established at the department of Physiology at UBC.  42.3 DIPEPTIDYL PEPTIDASE IV ASSAY In order to verify the lack of DPP TV in the DPP TV-negative rats, a DPP TV assay was established, using a slightly modified method from that which was previously described (Matumura, 1985). Blood was collected from the tall of five DPP TV-negative and five DPP TV-positive rats, and the serum separated. The substrate for the enzyme was a solution of 1.4 mM Gly-Pro-p-nitroanilide (Sigma) in 114 mM Tris buffer pH 8.0 (700 mg Tris HC1  +  844 mg Tris Base to 100 ml distilled H 0). Test tubes containing 1 2  ml of the Tris buffer were incubated at 37°C, followed by the addition of 100 il test serum. The absorbance at 410 nm was then recorded immediately and at 5 mm intervals, using a U.V. spectrophotometer (Canlab Pye Unicam SP8-100) to monitor the appearance of the yellow product of enzyme activity, p-nitroaniline. A standard curve was prepared by reading the absorbance of p-nitroaniline (Sigma) solutions of 0.005 to 1 mM in Tris buffer (Figure 21A). Absorbance of serum samples with substrate (time  =  0) was  subtracted from all further timed readings. The resulting absorbance values were then determined from the standard curve, and the resulting conversion rate to p-nitroaniline was plotted.  110 25 ‘ 4.2.4 HPLC PURIFICATION OF 1-GIP In order to perform in vivo studies to confirm the action of DPP IV on GIP, it was desirable to develop a highly sensitive assay system to monitor the conversion of GIP 142 to GJP 3-42. Since GIP 1-42 can be separated from GIP 3-42 by HPLC (see section 3.3.1), it was hypothesized that HPLC could also resolve 125 1-GIP 1-42 from 1-GIP 25 3‘ 42. The use of radioactive tracer would enable the use of physiological levels of GIP in vivo, and thus allow for a highly sensitive, accurate estimation of the conversion of GIP 1-42 to inactive GIP 3-42. Tracer was therefore prepared, using the methods outlined in section 2.2.8, with the exception that synthetic porcine GIP (Peninsula) was used instead of the natural purified form (known to contain a high proportion of GIP 3-42; see section 1.11.1). Immediately after iodination, the peak fraction of 125 1-GIP was purified on an HPLC system that was used exclusively for purifying radioiodinated peptides. Water (distilled and treated through a Waters MilliQ H 0 filtration system) and 2 acetonitrile (CH CN) were prepared by adding 0.1% thfluoroacetic acid (TFA) to each, 3 and degassing and filtering through either 0.22 I.Lm (H 0) or 0.45 jim (CH 2 CN) filters 3 (Waters). These solvents were delivered to a j.tBondapak C18 column (Waters) using two 1 lOB Solvent Delivery Module pumps (Beckman Instruments Inc, San Ramon, CA) controlled by a programmable 421A Controller (Beckman). The 1-GIP 25 was injected ‘ using a 100 jil needle syringe (Hamilton Co., Reno, NV) into a 210A Injector Port (Beckman) with 100 jil capacity. The tracer was loaded at a CH CN concentration of 3 32%, and eluted by increasing the CH CN content to 38% over 10 mm, and maintained 3 for a further 5 mm.  The column was then washed by increasing the CH CN 3  concentration to 60% over 5 mm, and re-equilibrated at 32% for 10 mm prior to injection of another sample.  This program had previously been empirically determined by  Verchere (1991) to provide separation of the tracer components. Radioactivity of the column eluant was measured with a 170 Radioisotope Detector (Beckman) and charted on a Recordall Series 5000 recorder (Fisher). Eluant fractions were collected every 0.5  111 mm into 13 x 10 mm siliconized glass test tubes containing 10 p.1 BSA (5%; RIA Grade). Aliquots (10 p.1) of these fractions were counted to ensure recovery and separation, and peak fractions were pooled, frozen at -70°C, lyophiized, and stored at -20°C. The specific activity of the 1-GIP 25 was estimated by two methods. First, the ‘ total radioactivity (cpm) eluted from the G- 15 column for the first peak (‘ 1-GIP) was 25 determined as the sum of the radioactivity in each fraction of the peak  (= x p.Ci).  Assuming that all 5 p.g GIP was iodinated, the specific activity (S.A.) was then estimated as: S.A. (p.CiIp.g)  = x p.Ci/5 jig. Specific activity of the purified tracer was estimated  from self-displacement curves, using increasing concentrations of 1-GIP 25 in a GIP RIA. ‘ A displacement curve was formed beginning with --5000 cpm, and then doubling amounts up to 160,000 cpm. A standard curve was also prepared similar to the normal GIP RIA (see section 2.2.8), with the exception that the label was added at the same time as the standards. In this fashion, both displacement curves allowed equal time for labeled and unlabeled peptide to compete for binding to the antibody. The ratio of bound to total tracer (%B) was then calculated for each concentration of ‘I-GIP and the results plotted as %B vs. total counts added, on the same axes as the standard curve. Assuming that the antiserum used bound 1-GIP 25 and unlabeled GIP equally well, both the GIP standard ‘ concentration and 1-GIT’ 25 radioactivity were determined for several values of %B in the ‘ mid-range of the curve and a plot of GIP mass (pg) vs radioactivity (cpm) was produced. The S.A. of the label was then estimated as the inverse slope of this plot.  4.2.5 INCUBATION OF ‘ 1-GIP 2 5 WITH SERUM The hypothesis that 1-GIP 25 could be metabolized by serum and the components ‘ resolved by HPLC was tested. Purified label (peak 2) was diluted to 500,000 cpm/50 p.1 in 40 mM phosphate buffer (pH 7.0) and 50 p.1 aliquots were added to siliconized test tubes containing 400 p.1 of the same buffer. After gentle mixing, 50 p.1 of rat serum was added to each tube, and incubated at 37°C. Some samples were also incubated at 4°C to  112 examine the effect of temperature. Similar serum/label mixtures were prepared with buffer containing either 0.1 mM diprotin A, 2% aprotinin, 50 U/mi bacitracin (Upjohn), or all three combined as a ‘cocktail’. After incubation times ranging from 10 mm to 48 h, 100 j.il aliquots were removed and analyzed by HPLC for breakdown of 1 I-GP 1-42.  42.6 IN VIVO EXPERIMENTS WITH 125 1-GIP  Experiments were next carried out to measure the degradation of 1-GTP 25 1-42 in ‘ vivo. Age matched Fischer 344 rats and DPP TV-negative rats (—350 g) were fasted overnight and anesthetized by intraperitoneal injection of sodium pentobarbital (60 mg/kg). The right jugular vein was cannulated with PE 90 tubing filled with heparinized saline and attached to a 3-way stopcock. The abdomen was then exposed by a midline incision and the abdominal aorta cannulated with heparinized saline-filled PE16O tubing attached to a 3-way stopcock. At time 0, —5 x 106 cpm purified GIP label (peak 2; —22.5 ng GIP) in a volume of 100 I.Ll 40 mM phosphate buffer (pH 7.0) was injected into the jugular vein, followed by 200-300 j.ti of heparinized saline. At intervals of 2, 5, 10, and 20 mm, 2 ml blood samples were collected from the dorsal aorta and transferred immediately to Eppendorf tubes on ice containing diprotin A and aprotinin (final concentration 0.1 mM and 2%, respectively). Samples were then centrifuged at 7,000 x g for 5 mm at 4°C and the serum removed and stored on ice. AC 18 SepPak cartridge (Waters) was used to purify the 1-GIP 25 from the serum. ‘ The SepPak was first primed with 5 ml CH CN (plus 0.1% TFA), followed by 10 ml H 3 0 2 (plus 0.1% TFA). The serum sample was then applied to the cartridge, and washed with 10 ml H 0 plus TFA followed by 10 ml of 20% CH 2 CN plus TFA. The label was then 3 eluted into a siliconized test tube containing 10 j.tl BSA (5%; RIA Grade) with 2 ml 40% CN plus TFA, lyophilized and stored at -20°C. Each step of the procedure was 3 CH monitored with a portable gamma detector to ensure maximal recovery of the tracer.  113 0 and analyzed by 2 Lyophilized samples were then dissolved in 100 jil of distilled H HPLC as described in section 4.2.4. Control experiments were performed in order to ensure that during the sample collection procedure and sample treatment steps, no further degradation of the label occurred. Approximately 6 x iO cpm was divided into two aliquots of 1 ml serum containing diprotin A (0.1 mM) and aprotinin (2%). One aliquot was maintained on ice for 1 h, while the other was incubated at 37°C for 1 h. Both samples were then purified using the SepPak, and lyophilized as the rest of the samples. These controls were then analyzed by HPLC for any degradation of the ‘I-GIP.  4.2.7 DIPEPTIDYL PEPTIDASE IV IN ZUCKER RATS AND NIDDM PATIENTS Serum samples were collected from fed Wistar rats (4) plus fed lean and fat Zucker rats (4 of each) and stored at -20°C. In collaboration with Dr. G. S. Meneilly (Department of Medicine, UBC) samples were also collected from 12 fasted non-obese NTDDM patients receiving treatment with oral hypoglycemic agents, and 12 fasted age matched controls with normal glucose tolerance, and stored at -20°C. All samples were then assayed for DPP IV by the method described in section 4.2.3.  4.3 RESULTS 4.3.1 INCUBATION OF GIP WITH SERUM Incubation of GIP with serum yielded a time-dependent degradation, resulting in products that eluted by HPLC identically to that observed for STC 6-14 cell culture medium (Figure 17 chapter 3). This process did not occur in the absence of serum, and was considerably reduced by the addition of 0.1 mM diprotin A (data not shown).  114 4.3.2 DIPEPTIDYL PEPTIDASEIVASSAY The assay used to measure serum levels of DPP IV relied on the ability of this enzyme to convert Gly-Pro-p-nitroanilide to the colored product, p-nitroaniline. The development of the product was monitored with a spectrophotometer at 410 nm and was converted to mM p-nitroaniline produced using a standard curve (Figure 21A). Under the assay conditions used, serum from Fischer rats was able to produce p-nitroaniline at a rate of —1.66 I.LM/min, while production with serum from DPP TV-negative rats occurred at a much slower rate of —0.17 p.M/mm (Figure 21B). There was no significant difference in DPP IV activity in the same serum volume from Fischer, Wistar, lean Zucker, or fat Zucker rats (data not shown). Similarly, there was no significant difference in the DPP IV activity in serum from NIDDM patients tested when compared to controls. Absorbance values at 45 mm were corrected for the absorbance of the serum  +  substrate  at time 0, and then converted to mM p-nitroaniline produced using the standard curve. The average concentration of p-nitroaniline after 45 mm for control subjects was 0.134 ± 0.004 mM (—2.97 p.M/mm), and for diabetic subjects was 0.128 ± 0.004 mM (—2.84 .tM/min).  4.3.3 LABEL SPECIFIC ACTIVITY The specific activity of the 125 1-GIP eluted from the G-15 column during iodination (Figure 23) was estimated at —35 p.Ci/p.g. This material was then purified by reverse-phase HPLC, which revealed a heterogeneous population of iodinated peptides, consisting of 4 major components, the largest being peak 2 (Figure 25A). The specific activity of this fraction was estimated from self displacement curves to be —100 p.Ci/p.g (Figure 24), representing a significant enrichment compared to the unpurified tracer.  115 I-GIP WITH WISTAR RAT SERUM 4.3.4 INCUBATION OF PURIFIED 125 The stability of HPLC purified ‘I-GW 1-42 was investigated by incubation at 37°C for up to 24 h in the absence of serum. At intervals of --6 h, labeled GIP was removed and analyzed by HPLC. Even after 24 h, a single component was observed 1-GJP 1-42 eluted slightly later (—21-22 mm) then when collected (Figure 25B). This 125 as peak 2 (—19 mm; Figure 25A), perhaps attributable to the use of a new C 18 column. A single component was also observed after 10 mm incubation with 10% Wistar rat serum 125 had resolved into two distinct major components. (Figure 26A), but after 1 h, 1-GIP One component eluted in the original position (—21-22 mm) and thus consisted of 1251 GIP 1-42 (peak 1), while the new component eluted earlier (19-20 mm). This separation is similar to that observed in the identification of GIP 3-42 in natural GJP preparations (see Figure 16) and therefore this peak 2 was assigned 1-GIP 25 3-42. With time, as ‘ 25 and serum were analyzed by HPLC, a progression from ‘ aliquots of incubated 1-GIP 1-GIP 1-42 to putative 1-GLP 125 25 3-42 was observed (Figures 27 & 28). Furthermore, ‘ other ‘ 1-GIP 2 5 fragments with different elution times appeared at a significantly slower rate, the majority eluting at —3 mm (peak 3; Figures 27 & 28). The effect of temperature on the metabolism of 1-GIP 25 1-42 was investigated by ‘ performing the same experiments at 4°C. Both peaks 1 & 2 could be observed, indicating that the conversion of 125 1-GIP 1-42 to putative ‘ I-GIP 3-42 was still occurring (Figures 2 29 & 30). However, the rate of this reaction (as determined by the changes of the areas of the two peaks with time) was considerably slower. Interestingly, at this temperature, an absence of other peaks including peak 3 was noted, indicating no further degradation of the tracer had occurred (Figures 29 & 30) The effect of enzyme inhibitors on the degradation of ‘ 1-GIP 2 5 was also investigated. Diprotin A, a competitive inhibitor of DPP IV, was able to significantly r s IGIP 3-42. This was evident from reduce the degradation of 1-GIP 25 1-42 to putative 2 ‘ the significantly reduced rate of appearance of peak 2 and reduction in peak 1 when 0.1  116 mM diprotin A was added to the serum and ‘ 1-GTP 2 5 (Figures 31 & 32). However, this inhibitor did not alter the degradation of 125 1-Gll’ to other fragments, with considerable radioactivity eluting in peak 3 as incubation time progressed (Figures 31 & 32). In contrast, the protease inhibitor aprotinin had no significant effect on the conversion of IGIp 1-42 to putative 1-GIP 125 25 3-42, as evident from the reduction of peak 1 and ‘ production of peak 2 with time by Wistar rat serum in the presence of 2% aprotinin (Figures 33 & 34). However, this inhibitor did reduce the rate of production of peak 3 (Figures 33 & 34).  The antibiotic bacitracin (50 U/mi) significantly reduced the  conversion of 1-GJP 25 1-42 to putative ‘ ‘ 1-GJP 2 5 3-42 by serum (peak 1 to peak 2; Figure 35), but not to the same degree as diprotin A. Bacitracin, however, greatly reduced the further fragmentation of ‘ 1-GIP 2 5 as indicated by the near absence of other peaks, including peak 3 (Figure 35). When used in combination, these inhibitors effectively reduced the degradation of ‘ 1-GIP 2 5 (Figure 36). After 20 h incubation of ‘ 1-GIP 2 5 with Wistar rat serum and the inhibitor ‘cocktail’ at 37°C, —35% existed as putative 12 I-GIP 342 (peak 2) and the remainder as ‘I-GIP 1-42 (peak 1).  4.3.5 INCUBATION OF PURIFIED 125 1-GIP WiTH DPP [V-NEGATIVE RAT SERUM  The effect of DPP TV-negative serum on the degradation of GW 25 as analyzed by ‘ HPLC is shown in Figure 37. For a period of up to 5 h, no peak 2 could be resolved, suggesting that no ‘ 1-GIP 2 5 3-42 was produced. However, as with Wistar rat serum, other peaks representing other I1 -GIP fragments were observed, including peak 3 (Figure 37). The addition of 2% aprotinin significantly reduced the rate of production of other  l2Sj  GIP fragments by serum from DPP P1-negative rats, as evident from the reduction in the size of other peaks, including peak 3 (Figures 38 & 39).  117 4.3.6 INCUBATION OF PURIFIED 125 I-GIP WiTH CEP The buffer used for the GIP RIA contains 5% charcoal extracted human plasma (CEP; see section 2.2.8). In order to determine if this plasma contains enzymes capable of degrading GIP, CEP was diluted and incubated at 37°C with ‘ 1-GIP 2 5 and was analyzed by HPLC. After 3 h incubation, —48% ‘ 1-GIP 2 5 was found in peak 2, the putative ‘ 1-GIP 2 5 3-42 (Figure 40). Furthermore, by 7 h, further significant degradation to other ‘ 1-GIP 2 5 fragments had occurred, including a significant proportion in peak 3 (Figure 40). These observations would suggest that the addition of this plasma to the GIP assay buffer would promote the degradation of the GIP tracer, and possibly the standards and samples.  4.3 .7 THE ANALYSIS OF 1-GIP 25 DEGRADATION IN VIVO ‘ Methods were developed to investigate the degradation of GIP in vivo. It was expected that 125 1-GIP 1-42 would be rapidly degraded to ‘ 1-GIP 2 5 3-42 in the circulation, and rapid delivery of the tracer and expeditious blood sampling were deemed necessary. Label was therefore delivered by injection through a jugular vein cannula and plasma samples were collected directly from the aorta. It was also imperative that any further degradation subsequent to sample collection be prevented. Based on in vitro results it was determined that in order to keep degradation to a minimum, samples should be collected into tubes in ice, containing diprotin A and aprotinin (final concentrations 0.1 mM and 2%, respectively). The collected samples were then purified by SepPak as described earlier (section 4.2.6).  Recovery of tracer from plasma collection to  lyophilization was approximately 80%. In order to verify that there was minimal label degradation through all steps of the procedure, 125 1-GIP was added to serum containing the inhibitors, and was treated identically to the experimental samples. However, prior to being purified by SepPak, one aliquot was stored on ice for 1 h, and another was incubated at 37°C for 1 h. The results of the HPLC analysis are shown in Figure 41.  118 Neither condition resulted in measurable degradation of the label, indicating that experimental samples were treated appropriately. The degradation of label occurred very quickly in vivo. By 10 mm, only --9% of the ‘ 1-GIP 2 5 was represented by peak 1 1-GIP 125 1-42; Figure 42 & 43). The remainder ( of tracer appeared to be ‘ 1-GIP 2 5 3-42 (peak 2) with no other degradation products discernible. In contrast, there was no degradation of 125 1G1P infused into DPP IVnegative rats by 10 mm (Figure 44). In order to summarize some of the data on 125 1-GIP breakdown in serum, the relative proportions of putative ‘ 1-GIP 2 5 3-42 and 125 1-GIP 1-42 were computed by measuring the area of peak 1 and peak 2. By repeating this for various time intervals, plots of the rate of conversion to putative 125 1-GIP 3-42 were produced (Figure 45). Figure 45A indicates that the time required to yield 50% putative ‘ 1-GIP 2 5 3-42 was —2.2 h in 10% normal Wistar rat serum. This time was significantly reduced by the addition of bacitracin to --5.6 h, and to 16.1 h by storage at 4°C. The addition of diprotin A lengthened the time required to acquire 50% putative ‘ 1-GIP 2 5 3-42 to —29 h. Similar computations were done with the in vivo data from one Wistar rat. Only —1.5 mm was required for 50% of the ‘ 1-GTP 2 5 to be converted to putative ‘ 1-GIP 2 5 3-42, a rate approximately 88-times faster than was observed with the 10% rat serum in vitro. No 1-GIP 1-42 to putative ‘ degradation of 125 1-GIP 2 5 3-42 was observed in DPP TV-negative rats, in vitro or in vivo.  119 A 1.0 0.8  z  0.6  C  0.4  (I)  0.2 0.0 5 p-NITRO ANILINE (mM) B 0.8 ----0----  0.6  z  0.5  z  0.3  FISCHER DPP IV NEG. CONTROL  C  z  0.2 0.0 20 30 TIME (mm)  FIGURE 21: Determination of relative DPP IV levels in Fischer (DPP IV-positive) and DPP TV-negative rats from rate of production of p-nitroaniline. A) Standard curve of absorbance for increasing concentrations of p-nitroaniline (mM); n = 10. B) Plot of p nitroaniline concentration vs. time (mm) determined from absorbance values converted to mM p-nitroaniline from curve A. Control = substrate in the absence of serum; n = 8.  120  4-  z -  z C -  0-—  I  CONTROLS  —  DIABETICS  FIGURE 22: Determination of the relative levels of DPP IV in NDDM (diabetic) and normal (control) subjects from the production of p-nitroaniline after 45 mm. Average concentrations of p-nitroaniline were 0.134 ± 0.004 mM (--2.97 p.M/mm) for controls (n 12), and 0.128 ± 0.004 mM (—2.84 p.M/mm) for diabetics (n not significantly different. =  =  12). These values are  121  S 4 3 1 125 Na  2 1 0 10  20 FRACTION  30  FIGURE 23: Profile of synthetic porcine GIP iodination mixture eluted on Sephadex G 15. The iodination mixture was applied to the column immediately following iodination, and fractions were collected every 1.5 mm. 1-radioactivi 25 ‘ ty was measured from 10 i.tl aliquots, to produce the profile. Free Na 1 eluted after 1-GIP. 125 25 *denotes samples ‘ used.  122  A 50-  40-  30-  o  20-  10. 0.(  0.1  I  I  I  1  10  100  Mass GIP Standard (pg x 102) (....) 125i Radioactivity (cpm x 1O) (-.-)  B 0.50.4oZ 0.30.20.10-  0  I  0  I  I  I  2 4 6 8 10 125! Radioactivity (cpm x 10)  FIGURE 24: A) % bound (%B) for a GIP standard curve using increasing GIP concentrations (empty circles), and a self displacement curve using increasing amounts of 25 (filled circles). B) Plot of mass of GIP standard vs. 1-GIP ‘ 1G1P 25 radioactivity as ‘ determined from A for a number of points. The inverse of the slope of this line was used to calculate the specific activity of HPLC purified ‘I-G1P (peak 2; —100 .tCi/tg).  123 A 35 30 4  25  E  15  V  1  10 5 0 0  5  30  35  TIME (mm) B 543E  2-  V  10 0  I  I  I  I  I  I  10  I  5  15  20  25  30  35  TIME (mm)  FIGURE 25: A) HPLC elution profile of 1-GIP 25 (—10 cpm) at 32% acetonitrile for 10 ‘ mm, followed by a 10 mm gradient up to 38%. After a further 5 mm at 38%, the column was washed by a 5 mm gradient up to 60%, and then at 30 mm was returned to 32% over 5 mm. Major peaks numbered 1-4 are discussed in the text. B) HPLC elution profile of peak 2 (‘ 1-GIP in subsequent figures) after 24 h incubation at 37°C (—5 x i0 cpm). 25  124 A 101  8-  0O1I52IO53IO3I5  TIME (mm) B 108-  -  1  1  0I:III  TIME (mm)  FIGURE 26: Both figures are HPLC elution profiles of ‘ 1-GIP 2 5 (—10 cpm) incubated with Wistar rat serum (10%) at 37°C. A) duration = 10 mm. The single peak corresponds to 1-GIP 25 1-42. B) duration = j. Peak 1 is 1-GIP ‘ 25 1-42, peak 2 ‘ corresponds to 1-GIP 25 3-42. In this and subsequent figures, the acetonitrile gradient is ‘ identical to that described for Figure 2.  125  A 108-  2..  __1 0— )  5  I  I  I  I  I  10  15  20  25  30  35  TIME (mm) B 10. 8-  E .  3  0JII  TIME (mm)  FIGURE 27: Both figures are HPLC elution profiles of 1-GIP 25 (—1O cpm) incubated ‘ with Wistar rat serum (10%) at 37°C. A) duration = 111. B) duration = 6h. In both figures, peak 1 is I-GIP 1-42, peak 2 corresponds to 1-GIP 25 3-42, and peak 3 consists ‘ of unidentified 1-GIP 25 fragments. ‘  126 A 1086S  .?01’52’O2’53’O3’5 TIME (mm) B 108-  -  5  4..  Ca  1’01’52’02’53’03’5 TIME (mm)  FIGURE 28: Both figures are HPLC elution profiles of 125 1-GIP (—10 cpm) incubated with Wistar rat serum (10%) at 37°C. A) duration = .11h. B) duration = 24h. In both figures, all peaks consist of unidentified 125 1-GIP fragments.  127  A 108-  61  °O1’52’O2’53’O3’5  TIME (mm) B 108-  -  TIME (mm)  FIGURE 29: Both figures are HPLC elution profiles of 1-GIP 25 (—10 cpm) incubated ‘ with Wistar rat serum (10%) at 4°C. A) duration = 4h. B) duration = J1. In both figures, peak 1 is 1-Gll? 25 1-42, and peak 2 corresponds to 1-GIP ‘ 25 3-42. ‘  128 A 108-  -  4 E  0111111  TIME (mm) B 108-  0I;III  TIME (mm)  FIGURE 30: Both figures are HPLC elution profiles of 1-GIP 25 (—‘1O cpm) incubated ‘ with Wistar rat serum (10%) at 4°C. A) duration = 19h. B) duration = 24h. In both figures, peak 1 is 1-GJP 25 1-42, and peak 2 corresponds to ‘I-GW 3-42. ‘  129 A 10. 86-  I TIME (mm) B 1086-  I TIME (mm)  FIGURE 31: Both figures are HPLC elution profiles of 125 1-GIP (-10 cpm) incubated with Wistar rat serum (10%) and diprotin A (0.1 mM) at 37°C. A) duration = 2Ji. B) duration = j 25 1-42, peak 2 corresponds to 1-GIP ‘ . In both figures, peak 1 is 1-GIP 1 25 ‘ 3-42, and peak 3 consists of unidentified I-GW 25 fragments. ‘  130 A 1086.  E1’01’52’02’53’03’5 TIME (mm) B 103  86-  2-  2 vJ  00  5  I  I  I  I  I  10  15  20  25  30  35  TIME (mm)  FIGURE 32: Both figures are HPLC elution profiles of 1-GIP 25 (—10 cpm) incubated ‘ with Wistar rat serum (10%) and diprotin A (0.1 mM) at 37°C. A) duration = i2Ji. B) duration  =  25 1-42, peak 2 corresponds to 1-GIP ‘ ..h. In both figures, peak 1 is 1-GIP 25 ‘  3-42, and peak 3 and others consist of unidentified ‘ 1-GIP 2 5 fragments.  131  A 108-  6-  o.I’III  TIME (mm) B 1086-  I  0E1I0  115  2’O  5  310  315  TIME (mm)  FIGURE 33: Both figures are HPLC elution profiles of ‘ 1-GIP 2 5 (‘—10 cpm) incubated with Wistar rat serum (10%) and aprotinin (2%) at 37°C. A) duration = iji. B) duration 1-GIP 1-42, peak 2 corresponds to 1-GIP = i.h. In both figures, peak 1 is 125 25 3-42, and ‘ peak 3 consists of unidentified 125 1-GIP fragments.  132  A  10-  8-  1111110 TIME  I  (mm)  B  10-  8-  : TIME (mm)  FIGURE 34: Both figures are HPLC elution profiles of 125 1-GIP (—1O cpm) incubated with Wistar rat serum (10%) and aprotiniri (2%) at 37°C. A) duration = iLh. B) duration 5j In both figures, peak 1 is ‘I-GW 1-42, peak 2 corresponds to 115 . 1 1-GIP 3-42, and peak 3 and others Consist of unidentified 125 1-GIP fragments. =  133 A 1086-  E  .  ohol’52’02’53’o3’5  TIME (mm) B 10. 864.  :  3 H I 5 2 I 5 I O O 3 : lI  TIME (mm)  FIGURE 35: Both figures are HPLC elution profiles of 1-GIP 25 (—1O cpm) incubated ‘ with Wistar rat serum (10%) and bacitracin (50 U/mi) at 37°C. A) duration = 2h. B) duration = 19.5 h. In both figures, peak 1 is 1-GIP 25 1-42, peak 2 corresponds to 1251.. ‘ GIP 3-42, and peak 3 consists of unidentified ‘I-GW fragments.  134 A 1081 6-  TIME (mm) B 1086-  I TIME (mm)  FIGURE 36: Both figures are HPLC elution profiles of 125 1-GIP (-10 cpm) incubated with Wistar rat serum (10%) containing diprotin A (0.1 mM), aprotinin (2%) and bacitracin (50 U/mi), at 37°C. A) duration = 3.25 h. B) duration = 2Ji. In both figures, peak 1 is 1-GIP 25 1-42, and peak 2 corresponds to 1-GIP ‘ 25 3-42. ‘  135 A 108-  4. 23 0. 0  5  10  15  20  25  30  35  310  I  2’S  TIME (mm) B 1086-  E ,.  110  1’S  210  TIME (mm)  FIGURE 37: Both flgures are HPLC elution profiles of 1-GIP 25 (--10 cpm) incubated ‘ with DPP IV-negative rat serum (10%) at 37°C. A) duration 2..h. B) duration ih. In both figures, peak 1 is 125 1-GIP 1-42, and peak 3 consists of unidentified 125 1-GIP fragments.  136 A 108a  6-  E ..  200  5  I  I  10  15  1  20  25  I  30  35  TIME (mm) B 1086-  I TIME (mm)  FIGURE 38: Both figures are HPLC elution profiles of 1-GIP 25 (—1O cpm) incubated ‘ with DPP TV-negative rat serum (10%) and aprotinin (2%) at 37°C. A) duration = 4..h. B) duration &h. In both figures, peak 1 is 1-GIP 25 1-42, and peak 3 consists of ‘ unidentified 1-GIP 25 fragments. ‘  137 A 1086-  01I0  115  210  215  315  3’O  TIME (mm) B 10. 8. 6. E  3  -  Ca  1101’52’02’53’03’5 TIME (mm)  FIGURE 39: Both figures are HPLC elution profiles of 1-GTP 25 (—10 cpm) incubated ‘ with DPP TV-negative rat serum (10%) and aprotinin (2%) at 37°C. A) duration = iih. B) duration = 27 h. In both figures, peak 1 is 125 1-GIP 1-42, and peak 3 and others consist of unidentified ‘ 1-GIP 2 5 fragments.  138 A 1086E  .  ‘‘‘‘‘I’  TIME (mm) B 108-  -  4,. V  0EiI0  1’S  2’:  25  310  I  TIME (mm)  FIGURE 40: Both figures are HPLC elution profiles of 1-GIP 25 (--10 cpm) incubated ‘ with charcoal extracted human plasma (10%) used in the GIP RIA, at 37°C. A) duration iii. B) duration = Lh. In both figures, peak 1 is 1-GIP 25 1-42, peak 2 corresponds to ‘ 25 3-42, and peak 3 consists of unidentified 1-GJP ‘ 1-GIP 25 fragments. ‘  =  139 A 7.5.  ;- 5.0.  1  E 2.5.  0.0 C  I  I  I  I  I  I  I  5  10  15  20  25  30  35  25  30  35  TIME (mm) B 7.5-  5.01 E V  2.5-  0.00  I  5  10  15  20  TIME (mm)  FIGURE 41: Both figures are HPLC elution profiles of 1-GIP 25 (-40 cpm) incubated ‘ for 1 h with Wistar rat serum (neat) containing diprotin A (0.1 mM), aprotinin (2%), purified by SepPak, lyophilized, and reconstituted as described in section 4.2.6. A) incubation on ice. B) incubation at 37°C. In both figures, peak 1 is 125 1-GIP 1-42.  140 A 1086-  .O1’52:2’53’O3’5 TIME (mm) B 7.5.  ;-‘  5.0-  115  I  I  2’S  TIME (mm)  FIGURE 42: Both figures are HPLC elution profiles of 1-GIP 25 infused into a Wistar ‘ rat, collected in 2 ml serum with diprotin A (0.1 mM) and aprotinin (2%), purified by SepPak, lyophilized and reconstituted, as described in section 4.2.6. A) duration = 2 mm. B) duration = 5 mm. In both figures, —8 x iO cpm was loaded, peak 1 is 125 1-GIP 1-42 and peak 2 corresponds to 1-GIP 25 3-42. ‘  141 A 2.5-  2.0-  TIME (mm) B 2.5.  2.0.  I TIME (mm)  FIGURE 43: Both figures are HPLC elution profiles of 1-GIP 25 infused into a Wistar rat, ‘ collected in 2 ml serum with diprotin A (0.1 mM) and aprotinin (2%), purified by SepPak, lyophilized and reconstituted, as described in section 4.2.6. A) duration = 10 mm. B) duration = 20 mm. In both figures, ‘—4 x 10 cpm was loaded, peak 1 is “ 1-GIP 1-42, peak 5 2 corresponds to 1-GJP 25 3-42, and peak 3 consists of unidentified 125 ‘ 1-GIP fragments.  142 A 10861 E  42-  0-  0  5  I  I  I  I  I  I  10  15  20  25  30  35  25  30  35  TIME (mm) B 2.5  1 2.0  1.5 E 1.0  0.5 0  5  10  15  20  TIME (mm)  FIGURE 44: Both figures are HPLC elution profiles of 1-GIP 25 infused into a DPP IV‘ negative rat, collected in 2 ml serum with diprotin A (0.1 mM) and aprotinin (2%), purified by SepPak, lyophilized and reconstituted, as described in section 4.2.6. A) duration = 2 mm, —10 cpm loaded. B) duration = 10 mm, --.5 x iO cpm loaded. In both figures, peak 1 I.G1P 1-42 and peak 3 consists of unidentified ‘I-GIP fragments. is 125  143 A 100-  Normal Serum  0  +  Diprotin A  ----0----  +  Bacitracin  --A  @ 4°C  -0  75  -  1 50__0  25  —  /  yt —.4  /  I  ,___  _  I,.  0  5  10 TIME (h)  15  20  B 100.  75  50.  25-  0.  o  I  2  I  4 6 TIME (mm)  I  I  8  10  FIGURE 45: Graphs summarizing the conversion rate of 1-GIP 25 1-42 to 1-GIP ‘ 25 3-42 ‘ both in vitro and in vivo calculated from peak areas of Figures 5 to 22, plotted as % GIP 3-42 (% of GIP 1-42 + 3-42) vs. time. A) Relative % 1-GIP 25 3-42 production in Wistar ‘ rat serum at 37°C (normal serum), + diprotin A (0.1 mM), + bacitracin (50 U/ml), and 4°C. B) Rate of formation of 1-GIP 25 3-42 from 125 ‘ 1-GIP 1-42 infused in a Wistar rat.  @  144 4.4 DISCUSSION The enzyme DPP IV preferentially cleaves peptides and proteins having either X Pro, X-Hyp or X-Ala at their N-termini (Walter et a!., 1980), and to a lesser extent X-Ser, X-Thr, and X-Val (Martin et a!., 1993). DPP IV was first identified in rat liver (Hopsu Havu and Glenner, 1966), and has since been found ubiquitously distributed throughout mammalian tissues, with highest activity in the kidney and the intestinal brush-border membrane, as well as in insects, bacteria, and yeast (see Yaron and Naider, 1993 for review). The primary structure of rat liver DPP IV was deduced from its cDNA, with a calculated mass for the 767-residue polypeptide of 88,107 daltons (Ogata eta!., 1989). In keeping with the widespread occurrence of DPP IV, evidence suggests that the enzyme is involved in a number of biochemical processes. This list includes; renal transport and intestinal digestion of proline-containing peptides, immunological activation of immunocompetent cells, and fibronectin-mediated adhesion (Yaron and Naider, 1993). DPP IV also appears to have a role in the inactivation of biologically relevant peptides (Ahmad et a!., 1992; Frohman et a!., 1989; Mentlein et a!., 1993), and it is the involvement of DPP IV in the metabolism of GIP that was the focus of research discussed in this chapter. The observation made in chapter 3 that IRGIP in STC 6-14 culture medium eluted differently than IRGIP extracted from the cells, suggested the possibility of GIP degradation. These observations were reproduced by incubating GIP in serum, yielding a time-dependent breakdown of GIP 1-42. The retention time of the major metabolic product, eluting just prior to GIP 1-42 standards, suggested it was GIP 3-42. GIP 3-42 has been sequenced, and shown to be a significant component of natural GIP preparations, eluting just prior to GIP 1-42 on reverse-phase HPLC (JUrnval eta!., 1981, Schmidt et a!., 1987). Further evidence that this component of GIP degradation by serum was GIP 3-42 was supplied by the observation that the inclusion of diprotin A, a competitive inhibitor of DPP IV, significantly reduced the production of the peak  145 believed to be GIP 3-42. The enzyme DPP IV, had previously been shown to be responsible for removal of the same two N-terminal residues from growth hormone releasing factor (GRF) in human serum (Frohman et a!., 1986; 1989). Subsequent to these studies, a report confirming the degradation of GIP by DPP 1V in serum was published (Mentlein et a!., 1993). These authors found that DPP IV purified from human placenta hydrolyzed the N-terminal dipeptide from GIP, GRF, tGLP-I, and PHM. Km values of 4- 34 jiM and Vmax values of 0.6 3.8 j.tmolmin-’mg 4 -  protein were determined for the enzymatic degradation of the 3 peptides by the purified peptidase. The same fragments were identified when GIP or tGLP-I were incubated with human serum, and analyzed by HPLC. Furthermore, as was observed in the present study, this reaction could be inhibited by 0.1 mM diprotin A. Mentlein et al. (1993) concluded that DPP IV initiates the metabolism of GIP in serum. Confirmation of a role for DPP IV in the metabolism of GIP required in vivo evidence. Frohman et a!. (1986) proposed a role for DPP IV in GRF metabolism, by demonstrating degradation of GRF 1-44 to GRF 3-44 in vivo. These observations were accomplished by injecting subjects with GRF and analyzing serum by HPLC. There were, however, several disadvantages to this system, one being the requirement of infusing supraphysiological doses of hormone. Subjects were injected with GRF at 1 jig/kg, and even at this dose, plasma concentrations were well beneath the concentrations required for UV detection after HPLC, thus necessitating the collection of eluted fractions and analysis by rad.ioimmunoassay. Furthermore, the interpretation of these experiments were compounded by the presence of endogenous hormone. A primary objective of the present investigations was therefore to develop a highly sensitive assay to detect the conversion of GIP 1-42 to GIP 3-42 in vivo, to support the in vitro results. The HPLC profile obtained with the ‘ 1-GIP 2 5 in the present study, is very similar to that previously reported by others (Maletti et a!., 1986; Verchere, 1991). Maletti et al. (1986) found that tracer from the two largest peaks was able to stimulate insulin release  146 from the isolated perfused rat pancreas similarly to natural GIP. These authors also had preliminary evidence from enzyme cleavage that both of these iodinated fractions were iodinated only at the tyrosine in position 10, and not at the tyrosine in position 1. Verchere (1991) investigated the iodination state of the four major peaks of 125 1-GIP obtained after HPLC. Peaks 1-4 contained predominantly monoiodinated tyrosines (MIT), with peak 2 being the most pure, consisting of almost 100% MIT. Furthermore, in a binding assay using tumor derived B-cells (BTC3), peak 2 had the greatest specific binding (Verchere, 1991). In combination, these investigations suggested that peak 2 consisted of GIP exclusively monoiodinated at the tyrosine in position 10. This peak was therefore deemed ideal for the present studies, as any iodination at tyrosine in position 1 could potentially alter the ability of DPP IV to cleave GIP between residues 2 and 3. Furthermore, this peak appeared to retain biological activity (Maletti et al., 1986; Verchere, 1991), and made up the greatest proportion of the iodinated product, allowing for maximum yields. All subsequent experiments were therefore performed with peak 2 of the HPLC separated ‘I-GIP. In order to confirm that the purified 1251..Gu) would still serve as a substrate for DPP IV, and that the products of enzyme degradation could be readily identified and quantified, experiments involving the incubation of GJP with serum followed by HPLC were repeated with the tracer. These in vitro experiments clearly demonstrated the time dependent metabolism of GIP into products that could be resolved.  As had been  observed with the metabolism of natural GIP in the present experiment and by Mentlein et a!. (1993), there appeared to be one primary product which eluted just prior to ‘ 1-GIP 2 5 1-42, that was considered ‘I-GIP 3-42. The time required for conversion of 50% of the ‘I-GP to 1-GJP 25 3-42 under the assay conditions was estimated at 2.2 h. Other peaks ‘ appeared indicating further degradation of the tracer, but these reactions occurred at a much slower rate. Further evidence that the primary product was 125 1-GIP 3-42 was provided by the observation that the inclusion of diprotin A in the reaction significantly  147 1-GIP 3-42 reduced the rate of production of this peak (50% 125  @ —29 h). Furthermore,  rats specifically lacking DPP IV, served as suitable controls, and serum from these 25 1-42 to 1-GIP ‘ animals was unable to metabolize 1-GIP 25 3-42. ‘ This highly sensitive model also allowed for further characterization of the metabolism of 125 1-GIP 1-42. As expected, the inhibitor aprotinin had no effect on the production of 1-GLP 25 3-42. Aprotinin did however significantly reduce subsequent ‘ metabolism to other unidentified products, likely by inhibiting the action of other enzymes such as trypsin. Buckley and Lundquist (1992) reported that bacitracin was an effective inhibitor of enzymatic cleavage of tGLP-I, and therefore this inhibitor was also investigated. Bacitracin significantly reduced the degradation of 125 1-GIP 1-42 to  1251..  GIP 3-42, but not to the degree that was observed with diprotin A (50% 125 1-GIP 3-42  @  —5.6 h). However, while diprotin A did not appear to significantly affect subsequent label degradation, bacitracin was highly effective. Even after 19.5 h incubation at 37°C, very small amounts of other products were identifiable (other than 1-GIP 25 3-42). In ‘ combination, diprotin A, aprotinin, and bacitracin provide highly effective protection against 1-GJP 25 degradation. ‘ Interestingly, incubation with serum at 4°C resulted in significant degradation of 25 to 1-GIP ‘ 1.GW 25 3-42, although the rate was much slower than the same reaction at ‘ 25 3-42 ‘ 37°C (50% 1-GIP  @ —16.1 h). Other enzymes capable of metabolizing GIP  appeared to be completely ineffective at this temperature, however, as indicated by the lack of other significant peaks (products) even after 24 h incubation. Thus in addition to including the enzyme inhibitor ‘cocktail’, these results would suggest that storage at 4°C is an effective method to prevent degradation of GIP. While this study clearly supports the current practice of performing GIP assays at 4°C, the use of charcoal extracted plasma (CEP), a major assay constituent, requires further consideration. CEP was clearly able to degrade the tracer into multiple products, as indicated by the multiple peaks obtained by HPLC.  This degradation of ‘ 1-GIP, 2 5 could very likely also extend to sample  148 degradation, which could affect the binding of the products by the antiserum. These studies, therefore, clearly do not support the use of plasma in the GIP radioimmunoassay. This model thus appeared to fulfill the requirements of a suitable assay for the metabolism of GJP, and therefore in vivo experiments were performed. Approximately 5 x 106 cpm 1-GIP 25 with a specific activity of —100 tCi/p.g (equivalent to 22.5 ng/GIP), ‘ was injected in each rat. Thus for a plasma volume of —10 ml, the GIP concentration would be —2 ng/ml, a value within the physiological range. This method thus allows the use of concentrations at least 10-fold lower then those used by Frohman et a!. (1986) in the investigation of in vivo metabolism of GRF. As expected, the rate for ‘ 1-GIP 2 5 degradation was much more rapid in vivo than was observed in vitro with diluted serum. Only —1.5 mm was required for half of the ‘ 1-GIP 2 5 1-42 to be converted to ‘ 1-GIP 2 5 342. This represents an 88-fold reduction compared to the time observed in vitro for equivalent degradation. This result is comparable to that obtained by Frohman et al. (1986) who found that 1 mm after injection of 1 pjg/kg GRF 1-44, 43 ±7 % of the total immunoreactive GRH was GRH 3-44. This short half-life of DPP IV substrates is likely a reflection of the fact that DPP IV is found in high concentrations as an ectoenzyme of the plasma membrane of numerous cell types, in addition to plasma (Yaron and Naider, 1993). There are many implications of these results that should be considered. The data presented suggest that DPP IV is a primary enzyme involved in the degradation of GIP in vivo. The product of this reaction is GIP 3-42, a polypeptide that has been previously demonstrated to lack somatostatinotropic and insulinotropic activity (Brown eta!., 1981; Moody et a!., 1981; Schmidt et al., 1986; 1987). There have been no assays reported to date using antisera that can distinguish between GIP 1-42 and GIP 3-42. Given this information, the significance of sustained elevation of circulating GIP levels must be re evaluated. For example, reports of GIP half-lives of —20 mm (see section 1.12) and circulating levels elevated for hours (see section 1.10.1) must be considered in light of the  149 time course of conversion of GIP 1-42 to GIP 3-42. Thus the insulinotropic activity appears regulated in such a fashion that secreted GIP only acts on the pancreas for one circulation and is then inactivated to prevent overproduction of insulin and hypoglycemia. While GIl’ 3-42 is not an antagonist of the insulinotropic activity of GIl’ 1-42, other smaller fragments such as GIP 4-42 could act as antagonists, or may even be insulinotropic (see section 1.11.2).  It remains possible, therefore, that the further  enzymatic degradation of GIP that was observed to occur, albeit at a comparatively slow rate, might yield GIP fragments smaller than 3-42 with biological activity, or the ability to act as antagonists. It is also possible that the addition of  125J  to the tyrosine at position  10 of the GIP polypeptide to produce the tracer used for metabolism studies hindered enzymatic cleavage near this residue. These queries can be addressed by the separation, identification, and determination of biological activities of further serum degradation products of GIP. Over activity of DPP IV could result in increased metabolism of the insulinotropic hormones Gil’ and tGLP-I, and thus a reduced insulin response to the secretion of these hormones. This might then result in insufficient insulin secretion in response to a glucose load, yielding diabetic-type symptoms. Interestingly, urinary concentrations of DPP IV have been shown to be elevated in NIDDM patients (Nagata et a!., 1988; Takasawa et a!., 1990; Nukada et a!., 1992). As this elevation preceded the onset of microalbuminuria that is associated with nephropathy in diabetics, it was suggested that the alteration of the proximal tubules occurs prior to pathological changes of the glomeruli and brings about enzyme leakage from the brush border. If this enzyme dissociation is a response to the elevated glucose levels, then it is possible that maintained high glucose could also result in DPP IV dissociation from other tissues, such as endothelium, and lead to elevated levels in the plasma. However, neither in the obese Zucker rat, nor in the 12 NIDDM subjects tested, were DPP IV activity rates significantly different from controls. The  150 possibility of altered DPP IV activity in subjects with diabetes requires further investigation. The mechanisms that regulate serum levels of DPP IV are not currently understood. In light of the metabolic role of this enzyme for degradation of incretins, it is possible that DPP IV, like incretins, is influenced by nutrient absorption. Suzuki et al. (1993) recently reported that rat intestinal DPP IV levels could be modified by diet. After 7 days on a high proline (gelatin) diet, DPP IV activity and mRNA levels in brush-border membranes were three- to six-fold greater. No analysis of serum concentrations were made in this study. Mentlein et al. (1993) measured serum DPP IV levels in subjects pre and postprandial and found no significant difference with an n  =  3. Clearly, regulation of  serum DPP IV levels requires further investigation. In conclusion, separation of the cleavage products of GIP in vitro and in vivo, combined with the influence of enzyme inhibitors and serum from DPP IV-negative rats indicates that dipeptidyl peptidase IV is a principal enzyme resulting in the degradation and biological inactivation of GIP. Preliminary data indicate no significant difference in DPP IV activity for either obese Zucker rats or NIDDM subjects when compared to controls. The ubiquitous nature of this enzyme, including epithelial cells of the intestine, might explain the isolation of considerable quantities of GIP 3-42 from natural GIP preparations. This and smaller enzyme fragments of OW undoubtedly contribute to the overall immunoreactivity measured by GIP radioimmunoassays currently in use. Furthermore, the inclusion of plasma in assay buffer, results in significant degradation of as well as perhaps GIP in standards and unknowns. This can be prevented by including diprotin A, aprotinin, and bacitracin in the assay buffer, in addition to maintaining all samples at 4°C.  151  CHAPTER 5 SUMMARY AND FUTURE DIRECTIONS  The role of gastric inhibitory polypeptide as a hormone has been well established. The polypeptide has been isolated, purified, and chemically identified, and its cells of origin located. Secretagogues for GIP release in vivo have been identified. The gene coding GIP has been isolated, sequenced, and the regulation of its transcription investigated. Induction of endogenous hormone release is accompanied by insulin release, and GIP injection in amounts that result in physiologically attainable levels mimic the biological actions. A GIP receptor has been cloned, sequenced, and located on target tissues, and the signal transduction system has been studied. Finally, putative roles for GIP in pathophysiological conditions, such as diabetes mellitus and obesity have been suggested, although data are equivocal. What is least well understood about Gil’, are the mechanisms controlling its release, and its duration of biological activity i.e. metabolism. The objectives of investigations reported in this thesis were, therefore, firstly to develop methods to investigate GIP release at the cellular level, and secondly, to develop a sensitive assay to monitor the degradation of GIP in plasma. A major difficulty in studying the release of GIP at the cellular level is the diffuse distribution of the endocrine cells of origin in the upper small intestine. Methodology was therefore developed involving preferential enrichment of intestinal endocrine cell preparations by centrifugal elutriation and short-term tissue culture in order to enable the study of the local regulation of IRGIP secretion from isolated endocrine cells. Canine intestinal duodenal and jejunal epithelial cell preparations enriched for endocrine cells released IRGIP in response to depolarization, as well as the principal GIP secretagogue identified in studies in vivo, glucose. It was demonstrated indirectly that glucose stimulated IRGIP release was inhibited by somatostatin, which is also likely a mediator of GIP release in vivo.  Increasing intracellular levels of cAMP or 2 Ca by  152 pharmacological activators resulted in significant secretion of IRGIP, indicating that these intracellular compounds are likely involved as messengers in the signal transduction system regulating GIP release. Finally, receptor-dependent activation of intracellular + messenger systems by the neuropeptide GRP yielded concentration-dependent 2 Ca increases in IRGLP output. This provides one of possibly many examples of neuropeptide control mechanisms for GIP release. While most of these observations were made with isolated canine epithelial cells, preliminary data suggests that porcine tissue is suitable for the continuation of these studies. These studies represent the first data on GIP release at the cellular level. Other in vitro models, such as mucosal sections or isolated bowel preparations, have for the most part proven unsuccessful for studying GIP release, due to the problems associated with ischemia in intestinal tissues, and the diffuse distribution of the GIP cells. With the ability to culture the isolated cells, comes the ability to readily control their environment  -  a key factor in the investigation of stimulus-secretion coupling. Understanding the mechanism by which a compound acts on the GIP cell, whether it be luminal, endocrine, paracrine, or neuronal, is very difficult to establish from in vivo studies. The methods used in these investigations resulting in cultured GIP cells, allowed for observations of the effects of luminal (glucose), endocrine/paracrine (somatostatin) and neuronal (GRP) stimuli on GIP release. There are many unanswered questions which remain concerning GIP release, which may be addressed using this model. For example, it might now be possible to investigate the influence of insulin on GIP release to resolve the potential role of insulin as a feedback-inhibitor of GIP secretion. The mechanisms by which nutrients, such as glucose and fat, act to cause GIP release may also be investigated. It is not known what intracellular processes in the metabolism of these nuthents result in GIP expression. Techniques are available in this laboratory to directly examine Ca 2 flux in individual cells in response to secretagogues. This could facilitate answering these remaining  153 questions, and corroborate data presented in this thesis suggesting that changes in intracellular Ca 2 are important in the regulation of GIP release. It also remains to be determined whether nuthents or their digestion products act directly on the GIP cell, or indirectly on neighbouring cells to stimulate GIP release. If the spatial organization of cells is important in allowing signal transduction from nutrient sensitive cells to GIP cells, then it can be speculated that the reaggregation of dispersed cells during the culture process might influence the GIP response to certain stimuli. With the identification of the glucose transporter responsible for the activation of OW release, and its localization on specific cell types, it will also be important to examine its distribution in the cultured cells. Specifically, it is of interest whether the cells attach and aggregate on the collagen coated wells in similar patterns that occur in vivo. To this end, a small chamber was constructed to enable culture of isolated cells on a microscope stage from which time-lapse photographs may be taken to observe the process of cell aggregation, growth of cellular projections, formation of cell to cell connections etc. which might facilitate these studies. One negative aspect of the static release studies described in this thesis is that metabolites, and other secretory products of cells in the cultures may influence the release of GJP. Evidence for this was supplied by the ability of somatostatin immunoneutralizing antibody to increase IRGIP release. It would therefore be desirable to examine the release of GIP from cells in a perifusion apparatus, where fresh medium is continually flowing over the cells and thus removing any secreted compounds. Such a system would also allow investigation of the dynamics of the GIP response to various secretagogues. For example, the IRGIP response may be biphasic, or exhibit desensitization to prolonged stimuli. Finally, a molecular approach would also be beneficial, allowing determination of the action of secretagogues on GIP mRNA levels. It would then be possible to address how much of the observed release of GIP is from GIP stores, and how much is from newly synthesized hormone.  154 Tumor cell lines were also investigated as an alternate source of GIP cells for study. A cell line derived from intestinal tumors of transgenic mice (STC- 1) was subcloned to produce a stable cell line with approximately 30% IRGIP and 30% IRSS cells. This new clone (STC 6-14) was used to study the release of IRGIP in response to glucose and interactions with somatostatin. HPLC of extracts of STC 6-14 cells indicated that the tumor cell derived IRGIP had the same retention time as natural porcine GIP 1-42. Release of IRGIP from STC 6-14 cells; increased in a concentration dependent fashion by glucose, was attenuated by the addition of somatostatin, and was augmented by addition of a somatostatin neutralizing antibody, presumably by immunoneutralizing endogenously released somatostatin. The addition of exogenous porcine GIP in the  presence of 5 mM glucose produced a concentration dependent increase in IRSS release. These represent the first reported data on GIP release from a tumor cell line. The STC 614 cells may be useful to further investigate the cellular mechanisms controlling the release of GIP and somatostatin, in addition to their interactions. This cell line provides a readily available, inexpensive supply of GIP cells. Studies on the cellular regulation of GIP secretion should be done concomitantly with studies on primary cells. Caution must be used when interpreting data from transformed cells, which might be regulated in a significantly different fashion than their natural counterpart. For example, in most malignant epithelial cells, there is an impairment of the metabolism of glucose with high rates of glucose consumption, aerobic glycolysis and an unusual accumulation of glycogen as compared with the normal tissue (see section 3.4 and Zweibaum et a!., 1991 for review). Such cellular alterations might therefore result in aberrant hormone regulation and release mechanisms. A multidisciplinary approach including parallel studies with primary cells is therefore desirable. Most intestinal receptors were initially identified in isolated intestinal epithelial cell preparations from various species (Laburthe and Amiranoff, 1989). Cell lines offer several advantages over other tissue preparations for studying neurohormonal receptors  155 (Zweibaum et a!., 1991). Studies of the GIP receptor on pancreatic tumor derived 13-cells are reviewed in section 1.11.2. In light of the fact that the GIP receptor on somatostatin cells may be different then that on 13-cells (see section 1.11.1), the somatostatin secreting STC 6-14 cell line may be a useful model to further investigate this hypothesis, particularly in view of the recent cloning of the GIP receptor (Usdin et a!., 1993). Previous studies have used somatostatin-secreting tumor cell lines to study the binding characteristics of tGLP-I (Fehmann and Habener, 1991b; Gros et a!., 1993), but no such reports on GIP binding have been forthcoming to date. All of the studies suggested for the primary GIP cells should also be performed on the STC 6-14 cells. However, more importantly, further experiments should first be aimed at investigating characteristics of the tumor cells. In particular, it is important to determine if the intracellular organization of these cells, as studied by electron microscopy, is characteristic of normal endocrine cells, e.g. secretory granule size and arrangement. Specifically, as this population of cells secretes a multitude of peptides (see section 3.3.1), it is of interest to determine if there are cells expressing more than one hormone. These studies are currently in progress at the Department of Physiology, UBC. In addition, the STC 6-14 cell line provides a source of murine GIP for sequencing and assessment of the biological activity. Studies described in chapter 4 implicate DPP IV as the main degradation and inactivating enzyme for GIP in the circulation.  This hypothesis is based on the  observations of the degradation of GIP 1-42 to putatitive GIP 3-42. Evidence that the DPP IV product is GIP 3-42 is supplied by its elution position just prior to GIP 1-42, and the finding that this inactive peptide was absent or reduced by including diprotin A, a specific inhibitor of DPP IV or by incubation with serum from rats DPP IV dificient. Furthermore, a recent report by Mentlein et a!. (1993) confirmed that GIP 1-42 is degraded to GIP 3-42 in vitro by purified DPP IV or serum, by sequencing the reaction products.  In order to confirm that GIP 3-42 is the main product of experiments  156 performed in this thesis, sequencing analysis should also be performed. In order to’ avoid 27 could be prepared and purified similar to that ‘ sequencing radioactive peptide, 1-GIP 1-GIP. After incubation with serum, the products could be which was performed for 125 separated using the same HPLC protocols, collected and then sequenced. Studies performed in vivo suggest that 50% of a bolus injection of GIP can be inactivated to GIP 3-42 in as little as 1.5 mm. Rapid hormone inactivation in conjunction with the glucose-dependency of GIP, may act to ensure the prevention of hyperinsulinemia and subsequent hypoglycemia. Since GIP 3-42 is biologically inactive, yet still immunoreactive, reports of circulating IRGIP levels in health and disease must be re-considered. Information regarding the circulating levels of biologically inactive GIP fragments can be misleading. Therefore, it is desirable to obtain N-terminal directed antibodies to establish an RIA, that can be used to distinguish between GIP 1-42 and GJP 3-42. Such an assay would provide a clearer picture of circulating levels of biologically active GIP secreted. As discussed in section 4.4, there is very little currently known about the mechanisms that control the levels of the enzyme DPP IV. If diet can alter intestinal brush border levels of this enzyme, this might lead to modulation of serum concentrations. Alterations in the levels of DPP IV could potentially have profound effects on the biological actions of GIP and its other substrates eg. tGLP-I. Elevated DPP IV levels would be expected to reduce the half-life of biologically active GIP, and thus reduce the incretin effect, potentially leading to hyperglycemia.  Initial studies  investigating the DPP IV activity in serum from obese Zucker rats and NIDDM subjects revealed no abnormalities. However, it does remain possible that other factors, such as diet or medical treatment, might alter DPP IV activity. Furthermore, serum samples were collected from NTDDM patients in the fasting state, during which GIP is normally neither secreted nor insulinotropic. In order to test the hypothesis that DPP IV serum activity is altered in NJDDM subjects, it is important, therefore, to also examine the enzyme activity  157 during the post-absorptive state, when GIP is insulinotropic. In addition, due to the ubiquitous distribution of DPP IV, with many potential sites for the metabolism of GIP other than in serum, in vivo GIP degradation studies should also be performed with NIDDM subjects.  As diabetes is a multifactorial disorder, with many potential  contributing abnormalities, further investigation into a potential role for DPP IV is warranted. The DPP TV-negative rats provide a model for the investigation of the consequences of a deficiency of this enzyme. The half-life for biological activity of GIP in these animals is considerably longer then controls, limited only by degradation by other enzymes and extraction by organs such as the kidney. It might be conceived that the greater duration of insulinotropic activity of GIP (and tGLP-I) could potentially result in over production of insulin and hypoglycemia. However, as glucose is the primary source of cell metabolic energy, it might be expected that these animals adapt physiologically to ensure an appropriate insulin response to ingestion of a meal. Preliminary investigations indicate that these animals do in fact have a normal insulin response to an oral glucose tolerance test, resulting in the appropriate maintenance of blood glucose levels. One protective measure that could explain this observation is the glucose-dependent nature of GIP. As soon as elevated plasma glucose levels (following an oral glucose challenge) return to basal (below threshold for GIP action), any nonmetabolized GIP 1-42 will not be insulinotropic, thus preventing hypoglycemia. Alternatively, or additionally, desensitization of B-cells to GIP could result from elevated levels of GIP 1-42. In light of the fact that elevated GIP or tGLP-I levels have been demonstrated to cause B-cell desensitization to these hormones (see section 1.11.2), it is of interest to determine if similar phenomena are also occurring in the DPP TV-negative rats.  Preliminary studies suggest that the pancreas from these animals is in fact  desensitized to GIP, as evident from a reduced insulin response from the perfused  158 pancreas to GIP in DPP TV-negative rats when compared to controls (Pederson and Kieffer, unpublished observations). Desensitization of the DPP TV-negative rat B-cells to the incretins GIP and tGLP-I may occur as a result of down-regulation of the receptors to these incretins. This hypothesis could be tested by examining the binding of these peptides to B-cells from DPP IV-negative and control rats. Methods for rat B-cell isolation and examination of GIP binding have been previously described (Verchere et a!., 1991). It would also be interesting to examine the effect of total parenteral nutrition (TPN) on the enteroinsular axis in these rats. In previous studies with normal rats, Pederson et al. (1985) showed that while the IRGIP response to oral glucose was normal following TPN, the isolated perfused pancreas showed a 30% increase in the insulin release in response to GIP after TPN. It was hypothesized that the increase in B-cell sensitivity to GIP may be causally connected to the exposure of the pancreas to chronically low levels of GIP during TPN. Perhaps this procedure could reverse the B-cell desensitization to GIP observed in the DPP IV-negative rats. Evidence provided in this thesis that implicates DPP IV as a principal inactivating enzyme of GIP suggests a strategy for the development of GIP analogs with a longer halflife in the circulation. DPP IV has an absolute requirement for the L configuration of the amino acid residue, both in the penultimate and the N-terminal position (Yaron and Naider, 1993). Thus substitution of D-amino acids for the L-amino acids His-Ala at the N-terminus of tGLP-I has been reported to result in resistance to DPP IV cleavage (Buckley and Lundquist, 1992). It is, therefore, likely that substitution of the GIP Nterminal amino acids Tyr-Ala with corresponding D configurations might make the resulting analogue resistant to degradation by DPP IV, with an attendant increase in insulinotropic potency. Such a compound could potentially have therapeutic actions in subjects with NIDDM who have insufficient insulin release to regulate glucose metabolism. This concept is supported by recent observations that tGLP-I (which may be  159 more insulinotropic then native GIP; see section 1.5.4) might have antidiabetogenic effects in subjects with NIDDM (Gutniak et al., 1992; Holz et al., 1993; Nathan et al., 1992; Nauk et a!., 1993). In conclusion, GIP is a hormone with important anabolic functions, of which the most extensively studied is undoubtedly its insulinotropic action. The role of GIP as an incretin in the enteroinsular axis has been well established, and potential links with obesity and NTDDM have been made. The experiments outlined in this thesis have been directed towards gaining a better understanding of the cellular mechanisms controlling GIP release and biological inactivation of this hormone in vivo. Data have been presented which implicate both the cAMP and Ca 2 intracellular messenger systems in the release of GIP, and the enzyme dipeptidyl peptidase IV in the rapid inactivation of GIP in vivo. The models developed to examine the release and metabolism of GIP that are presented in this thesis should enable further future investigation of these aspects of GIP physiology in both normal and pathophysiological states.  160 REFERENCES Ahmad S, Wang L, Ward PE (1992) Dipeptidyl(amino)peptidase TV and aminopeptidase M metabolize circulating substance P in vivo. J. Pharmacol. Exp. Therap. 260:12571261. Ahrén B, Häkanson R, Lundquist I, Sjolund K, Sundler F (1981) GIP-like immunoreactivity in glucagon cells. Interactions between GIP and glucagon on insulin release. Acta Physiol. Scand. 112:233-242. Alam MJ, Buchanan KD (1980) Gastric inhibitory polypeptide (GIP) responses in primary diabetes. [Suppi. 1] Regul. Pept. p 52 (Abstract). Alam MJ, Buchanan KD (1993) Conflicting gastric inhibitory polypeptide data: possible causes. Diabetes Res. Clin. Prac. 19:93-101. Alam MJ, Kerr JI, Cormican K, Buchanan KD (1992) Gastric inhibitory polypeptide (GIP) response in diabetes using a highly specific antiserum. Diabetic. Med. 9: 542545. Aluments JR, Häkanson R, O’Dorisio T, Sjolund K, Sundler F (1978) Is GIP a glucagon cell constituent? Histochem. 58:253-256. Amaland PF, Jorde R, Revhaug A, Myhre ESP, Burhol P0, Giercksky KE (1984) Fasting and postprandial GIP values in pigs, rats, dogs, and man measured with five different GIP antisera. Scand. J. Gastroenterol. 19:1095-1098. 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. Eur. J. Biochem. 159:353-358. Amiranoff B, Vauclin-Jacques N, Laburthe M (1984) Functional GIP receptors in a hamster pancreatic beta-cell line ml 11: specific binding and biological effects. Biochem. Biophys. Res. Commun. 123:671-676. Amiranoff B, Vauclin-Jacques N, Laburthe M (1985) Interaction of gastric inhibitory polypeptide (GIP) with the insulin-secreting pancreatic beta-cell line In 111: characteristics of GIP binding sites. Life Sci. 36:807-8 13. Andersen DK, Elahi D, Brown JC, Tobin JD, Andrs R (1978) Oral glucose augmentation of insulin secretion: interactions of gastric inhibitory polypeptide with ambient glucose and insulin levels. J. Clin. Invest. 62:152-161. Andersen DK, Putnam WS, Hanks JB, Wise JE, Lebovitz HE, Jones RS (1980) Enhancement of the metabolic clearance of gastric inhibitory polypeptide (GIP) by hyperglycemia. Regul. Pept. [Suppl.J 1:S3. Aponte GW, Taylor IL, Soil AH (1988) Primary culture of PYY cells from canine colon. Am. J. Physiol. 254:G829-G836. Arimura A (1992) Receptors for pituitary adenylate cyclase-activating polypeptide: comparisons with vasoactive intestinal peptide receptors. Trends Endocrinol. Metab. 3:288-294.  161 Arnold R, Ebert R, Creutzfeldt W, Becker HD, Borger H (1978) Inhibition of gastric acid secretion by gastric inhibitory polypeptide (GIP) in man. Scand. 3. Gastroenterol. [Suppi. 49] 13:11. Bacarese-Hamilton AJ, Adrian TE, Bloom SR (1984) Human and porcine immunoreactive gastric inhibitory polypeptides (IR-GIP) are not identical. FEBS Lett. 168:125-128. Banting FG, Best CH, Collip JB, Campbell WR, Fletcher AA (1922) Pancreatic extracts in the treatment of diabetes mellitus. Can. Med. Assoc. J. 12:141. Barber DL, Buchan AMJ, Walsh JH, Soil AH (1986a) Isolated canine ileal mucosal cells in short-term culture: a model for the study of neurotensin release. Am. J. Physiol. 250:G374-G384. Barber DL, Walsh IH, Soll AH (1986b) Release and characterization of cholecystokinin from isolated canine jejunal cells. Gastroenterol. 9 1:627-636. Baumert Jr JE, Cataland S, Tetirick Jr CE, Pace G, Mazzaferri EL (1978) Effect of atropine on meal-stimulated gastrin and gastric inhibitory polypeptide (GIP) release. 3. Clin. Endocrinol. Metab. 46:473-476. Bayliss WM, Starling EH (1902) The mechanism of pancreatic secretion. 3. Physiol. (London) 28:235-334. Beck B (1989) Gastric inhibitory polypeptide: a gut hormone with anabolic functions. 3. Mol. Endocrinol. 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. Regul. Pept. 7:3-8. Beck B, Villaume C, Chayviaile JA, Gariot P. Ulmer M, Desalme A, Debry G (1984) Influence of caloric intake on gastric inhibitory polypeptide, VIP and gastrin release in man. Peptides 5:403-406. Bell GI (1986) The glucagon superfamily: precursor structure and gene organization. Peptides 7:27-36. Bell GI, Santerre RF, Mullenbach GT (1983) Hamster preproglucagon contains the sequence of glucagon and two related peptides. Nature 302:716-718. Berson SA, Yallow RS (1959) Immunoassay of endogenous plasma insulin in human subjects by immunologic methods. Nature 184:1648-1649. Berthoud HR, Trimble ER, Moody AJ (1982) Lack of gastric inhibitory polypeptide (GIP) response to vagal stimulation in the rat. Peptides 3:907-912. Blundell TL, Dockerill 5, Sasaki K, Tickle U, Wood SP (1976) The relation of structure to storage and receptor binding of glucagon. Metab. [Suppl. 1] 25:1331-1336. Bodanszky M, Bodanszky A (1986) Conformation of peptides of the secretin-VIP glucagon family in solution. Peptides [Suppl. 1] 7:43-48.  162 Brant AM, MeCoid S, Thomas HM, Baldwin SA, Davies A, Parker JC, Gibbs EM, Gould GW (1992) Analysis of the glucose transporter content of islet cell lines: implications for glucose-stimulated insulin release. Cellular. Sig. 4:641-650. Brown JC (1971) A gastric inhibitory polypeptide. I. The amino acid composition and the tryptic peptides. Can. J. Biochem. 49:255-261. Brown JC (1974) Gastric inhibitory polypeptide (GIP). In: Endocrinology 1973. S. Taylor, ed., Heinemann, London, pp. 276-284. Brown JC (1982) Gastric inhibitory polypeptide. In: Monographs on Endocrinology. Springer-Verlag, Berlin, Germany. Brown JC (1988) Enteroinsular axis. In: Baillières Clinical Endocrinology and Metabolism 2(2) :359-373. Brown JC, Buchan AMJ, McIntosh CHS, Pederson RA (1989) Gastric inhibitory polypeptide. In: Handbook of physiology, the gastrointestinal system. Eds. S.G. Schultz, G.M. Makhlouf, B.B. Rauner. Am. Phsyiol. Soc. Bethesda MD, pp. 403-430. Brown JC, Dahl M, Kwauk 5, McIntosh CHS, Muller M, Otte SC, Pederson RA (1981) Actions of GTP. Peptides 2 (suppi. 2) 24 1-245. Brown JC, Dryburgh JR (1971) A gasthc inhibitory polypeptide II. The complete amino acid sequence. Can. J. Biochem. 49:867-872. Brown JC, Dryburgh JR, Ross SA, Dupré J (1975) Identification and actions of gastric inhibitory polypeptide. Recent Prog. Horm. Res. 31:487-532. Brown JC, Mutt V. Pederson RA (1970) Further purification of a polypeptide demonstrating enterogastrone activity. J. Physiol. 209:57-64. Brown JC, Otte SC (1978) Gastrointestinal hormones and control of insulin secretion. Diabetes 27:782-789. Brown JC, Pederson, RA (1970a) A multiparameter study on the action of preparations containing choleocystokinin-pancreozymin. Scand. J. Gastroenterol. 5:537-541. Brown JC, Pederson, RA (1970b) Cleavage of a gastric inhibitory polypeptide with cyanogen bromide and the physiological action of the C-terminal fragment. Proc. Physiol. Soc. June 52P (Abstract). Brown JC, Pederson RA (1976a) GI hormones and insulin secretion. Endocrinology. Proceedings of the Vth International Congress of Endocrinology 2:568-570. Brown JC, Pederson RA (1976b) The insulinotropic action of gastric inhibitory polypeptide in the perfused isolated rat pancreas. Endocrinology 99:780-785. Brown JC, Pederson, RA, Jorpes E, Mutt V (1969) Preparation of highly active enterogastrone. Can. 3. Physiol. Pharmacol. 47:133-114. Brubaker PL (1991) Regulation of intestinal proglucagon-derived peptide secretion by intestinal regulatory peptides. Endocrinology 128:3175-3182.  163 Brubaker PL, Drucker DJ, Greenberg GR (1990) Synthesis and secretion of somatostatin 28 and -14 by fetal rat intestinal cells in culture. Am. J. Physiol. 258:G974-G981. Bryer-Ash M, Cheung A, Pederson RA (1994) Feedback regulation of glucose-depedent insulinotropic polypeptide (GIP) secretion by insulin in conscious rats. Regul. Pept. (in press). Buchan AMJ, Ingman-Baker J, Levy 3, 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. Histochem. 76:341-349. Buchan AMJ, Barber DL, Gregor M, Soll AH (1987) Morphologic and physiologic studies of canine ileal enteroglucagon-containing cells in short-term culture. Gastroenterol. 93:791-800. Buchan AMJ, Curtis SB, Meloche RM (1990) Release of somatostatin immunoreactivity from human antral D cells in culture. Gastroenterol. 99:690-696. Buchan AMJ, Polak JM, Capella C, Solcia E, Pearse AGE (1978) Electron immunocytochemical evidence for the K cell localisation of gastric inhibitory polypeptide (GIP) in man. Histochem. 56:37-44. Buckley DI, Lundquist P (1992) Analysis of the degradation of insulinotropin [GLP-1(737)j in human plasma and production of degradation resistant analogs. Regul. Pept. 40:117 (Abstract). Buffa B, Polak TM, Pearse AGE, Solcia E, Grimelius L, Capella C (1975) Identification of the intestinal cell storing gastric inhibitory polypeptide. Histochem. 43:249-255. Bunnett NW, Harrison FA (1986) Immunocytochemical localization of gastric inhibitory polypeptide and glucagon in the alimentary tract of ruminants. Quart. 3. Exp. Physiol. 7 1:433-441. Burant CF, Takeda 3, Brot-Laroche E, Bell GI, Davidson NO (1992) Fructose transporter in human spermatozoa and small intestine is GLUT5. J. Biol. Chem. 267:1452314526. Burhol PG, Jorde R, Waldun HL (1980a) 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. Digest. 20:336-345. Burhol PG, Lygren I, Waldum HL, Jorde R (1980b) The effect of duodenal infusion of bile on plasma VIP, GIP, and secretin and on duodenal bicarbonate secretion. Scand. J. Gastroenterol. 15:1007-1011. Campbell RM, Scanes CG (1992) Evolution of the growth hormone-releasing factor (GRF) family of peptides. Growth Reg. 2:175-191. Campos RV, Buchan AMJ, Meloche RM, Pederson RA, Kwok YN, Coy DH (1990) Gastrin secretion from human antral G cells in culture. Gastroenterol. 99:36-44.  164 Cariquist M (1987) Solid phase synthesis of a 31-residue fragment of human glucosedependent insulinotropic polypeptide (GIP) by the continuous flow polyamide method. Acta Chem. Scand. 41:494-498. Cariquist M, Maletti M, Jörnvall 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 GJP 17-42. Eur. J. Biochem. 145:573-577. Cataland 5, Crockett SE, Brown JC, Mazzaferri EL (1974) Gastric inhibitory polypeptide stimulation by oral glucose in man. I. Clin. Endocrinol. Metab. 39:223-228. Chamberlain MJ, Stimmler L (1967) The renal handling of insulin. J. Clin. Invest. 49:791-797. Champion MC. Sullivan SN, Bloom SR, Adrian TE, Christofides ND (1982) The effects of naloxone and morphine on postprandial gastrointestinal hormone secretion. Am. 3. Gastroenterol. 77:617-620. Chan CB, Pederson RA, Buchan AMJ, Tubesing KB, Brown JC (1984) Gastric inhibitory polypeptide (GIP) and insulin release in the obese zucker rat. Diabetes 33:536-542. Chan CB, Soil AH (1988) Pertussis toxin-sensitive cholinergic inhibition of somatostatin release form canine D-cells. Am. J. Physiol. 255:G424-G428. Chang CL, Chang TM, Chey WY (1992) Characterization of the release of cholecystokinin and secretin from a murine intestinal cell line. Regul. Peptides 40:122 (Abstract). Chap Z, O’Dorisio TM, Cataland 5, Field JB (1987) Absence of hepatic extraction of gastric inhibitory polypeptide in conscious dogs. Dig. Dis. Sci. 32:280-284. Cheng H, Leblond CP (1974a) Origin, differentiation and renewal of the four epithelial cell types in the mouse small intestine: III enteroendocrine-cells. Am. J. Anat. 141:503-520. Cheng H, Leblond CP (1974b) Origin, differentiation and renewal of the four epithelial cell types in the mouse small intestine: V unitarian theory of the origin of the four epithelial cell types. Am. 3. Anat. 141:537-562. Chow BC, Morrow GW, Ho M, Pederson RA, McIntosh CHS, Brown JC, MacGillivray RTA (1990) Expression of recombinant human glucose-dependent insulinotropic polypeptide in Esherichia coli by sequence-specific proteolysis of a protein A fusion protein. Peptides 11:1069-1074. Christophe 3, Svoboda M, Lambert M, Waelbroeck M, Winand 3, Dehaye JP, Vandermeers-Piret MC, Vandermeers A, Robberecht P (1986) Effector mechanisms of peptides of the VIP family. Pept. [Suppl. 1] 7:101-107. Cleator IG, Gourlay RH (1975) Release of immunoreactive gastric inhibitory polypeptide (IR-GIP) by oral ingestion of food substances. Am. J. Surg. 130: 128-135.  165 Collier G, McLean A, ODea K (1984) The effect of co-ingestion of fat on the metabolic response to slowly and rapidly absorbed carbohydrates. Diabetologia 26:50-54. Comb M, Birnberg NC, Seasholtz HE, Goodman HM (1986) A cyclic AMP- and phorbol ester-inducible DNA element. Nature 323:353-356. Couvineau AB, Amiranoff N, Vauclin-Jacques, Laburthe M (1984) The GIP receptor on pancreatic beta cell tumor: molecular identification by covalent cross-linking. Biochem. Biophys. Res. Commun. 122:283-288. Cox JS, ODorisio TM, Cataland S, Crockett SE (1981) Gastric inhibitory polypeptide hypersecretion in diabetes mellitus: effect of sulfonylurea treatment. J. Clin. Endocrinol. Metab. 52: 1002-1005. Creutzfeldt W (1979) The incretin concept today. Diabetologia 16:75-85. Creutzfeldt W, Ebert R (1977) Release of gastric inhibitory polypeptide (GIl?) to a test meal under normal and pathophysiological conditions in man. In: Bajaj JS (ed) Diabetes. Excerpta Medica, Amsterdam, pp 63-75. Creutzfeldt W, Ebert R (1988) The incretin concept. In: Advances in Metabolic Disorders Academic Press 11:333-367. Creutzfeldt W, Ebert R, Arnold R, Frerichs H, Brown JC (1976) Gastric inhibitory polypeptide (GIP), gastrin and insulin: response to a test meal in coeliac disease and after duodeno-pancreatectomy. Diabetologia 12:279-286. 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 serum levels. Diabetologia 14:15-24. Creutzfeldt W, Talaulicar M, Ebert R, Willms B (1980) Inhibition of gastric inhibitory polypeptide (GIP) release by insulin and glucose in juvenille diabetes. Diabetes 29:140-145. Crockett SE, Cataland S, Falko JM, Mazzaferri EL (1976) The insulinotropic effect of endogenous gastric inhibitory polypeptide in normal subjects. J. Clin. Endoncrinol. Metab. 42:1098-1103. D’Alessio DA, Fujimoto WY, Ensinck JW (1989) Effects of glucagon-like peptide 1 (736) on release of insulin, glucagon, and somatostatin by rat pancreatic islet cell monolayer cultures. Diabetes 38:1534-1538. Davidson WD, Springberg PD, Falkinburg NR (1973) Renal extraction and excretion of endogenous gastrin in the dog. Gastroenterol. 64:955-961. Demol P, Wingender W (1989) Effects of rioprostil on the postprandial glucose, GIP, insulin and gastrin levels in volunteers. Scand. J. Gastroenterol. [Suppl. 164] 24:112117. Deschodt-Lanckman M, Robberecht P. Camus J, Christophe J (1971) Short-term adaptation of pancreatic hydrolases to nutritional and physiological stimuli in adult rats. Biochimie 53:789-796.  166 Drucker DJ, Philippe J, Mojsov S, Chick WL, Habener iF (1987) Glucagon-like peptide I stimulates insulin gene expression and increases cyclic AMP levels in a rat islet cell line. Proc. Natl. Acad. Sci. USA, 84: 3434-3438. Dryburgh JR (1977) Immunological techniques in the investigation of the physiological functions of gastric inhibitory polypeptide and motiin. Vancouver, Canada: Univ. of B.C. PhD dissertation. Dryburgh JR. Hampton SM, Marks V (1980a) Endocrine pancreatic control of the release of gastric inhibitory polypeptide: a possible physiological role for C-peptide. Diabetologia 19:397-401. Dryburgh IR, Hampton SM, Morgan LM, Marks V (1980b) C-peptide-induced inhibition of GIP release. Regul. Pept. [Suppi.] 1:28. Dupré J, Greenidge N, McDonald TJ, Ross SA, Rubinstein D (1976) Inhibition of action of glucagon in adipocytes by gastric inhibitory polypeptide. Metabol. 25:1197-1199. Dupré J, Ross SA, Watson D, Brown JC (1973) Stimulation of insulin secretion by gastric inhibitory polypeptide in man. 3. Clin. Endocrinol. Metab. 37:826-828. Dupré J, Beck JC (1966) Stimulation of release of insulin by an extract of intestinal mucosa. Diabetes 15:555-559. Ebert R, Creutzfeldt W (1982) Influence of gastric inhibitory polypeptide antiserum on glucose-induced insulin secretion in rats. Endocrinology 111:1601-1606. Ebert R, Frerichs H, Creutzfeldt W (1979a) Impaired feedback control of fat induced gastric inhibitory polypeptide secretion by insulin in obesity and glucose intolerance. Eur. 3. Clin. Invest. 9:129-135. Ebert R, Illmer K, Creutzfeldt W (1979b) Release of gastric inhibitory polypeptide (GIP) by intraduodenal acidification in rats and humans and abolishment of the incretin effect by OW-antiserum in rats. Gastroenterol. 76:515-523. Ebert R, Unger H, Creutzfeldt W (1983) Preservation of incretin activity after removal of gastric inhibitory polypeptide (GIP) from rat gut extracts by immunoadsorption. Diabetologia 24:449-454. Eckel RH, Fujimoto WY, Brunzell IL) (1979) Gasthc inhibitory polypeptide enhances lipoprotein lipase activity in cultured preadipocytes. Diabetes 28:1141-1142. Elahi D, Andersen DK, Brown JC, Debas HT, Hershcopf RJ, Raizes GS, Tobin JD, Andres R (1979) Pancreatic a- and B-cell responses to GIP infusion in normal man. Am. 3. Physiol. 237:E185-E191. Elahi D, Andersen DK, Muller DC, Tobin JD, Brown C JC, Andres R (1984) The enteric enhancement of glucose-stimulated insulin release. The role of GIP in aging, obesity and noninsulin-dependent diabetes mellitus. Diabetes 33:950-957. Elahi D, Raizes OS, Andres R, Hershcopf RJ, Muller DC, Tobin JO, Andersen DK (1982) Interaction of arginine and gastric inhibitory polypeptide on insulin release in man. Am 3. of Physiol. 242:E343-E35 1.  167 El-Munshid HA, Häkanson R, Lindberg G, Sundler F (1980) Effects of various gastrointestinal peptides on parietal cells and endocrine cells in the oxyntic mucosa of rat stomach. J. Physiol. (London) 305:249-265 Elrick ML, Stimmler L, Hlad CJ, Arai Y (1964) Plasma insulin response to oral and intravenous glucose administration. J. Clin. Endocrinol. 24: 1076-1082. Eissele R, Göke R, Willemer S, Harthus H-P, Vermeer H, Arnold R, Göke B (1992) Glucagon-like peptide- 1 cells in the gastrointestinal tract and pancreas of rat, pig and man. Eur. J. Clin. Invest. 22:283-29 1. Eissele R, Koop H, Arnold R (1990) Effect of glucagon-like peptide-1 on gastric somatostatin and gastrin secretion in the rat. Scand. J. Gastroenterol. 25 :449-454. Eng J, Andrews PC, Kleinman WA, Singh L, Raufman JP (1990) Purification and structure of extendin-3, a new pancreatic secretagogue isolated from Heloderma horridum venom. J. Biol. Chem. 265:20259-20262. Epstein PN, Ribar TJ, Decker GL, Yaney G, Means AR (1992) Elevated B-cell calmodulin produces a unique insulin secretory defect in transgenic mice. Endocrinology 130:1387-1393. Ewald CA, Boas J (1886) Beitrage zur Physiologie und Pathologie der Verdauung II. Virchows Arch [Path Anat] 104:271-305. Falkmer 5, Ebert R, Creutzfeldt W (1980) The entero-insular axis In: Frontiers of Hormone Research, vol 7. Creutzfeldt W, Ed. Basel: S Karger, pp 1-6. Fallcmer S, El-Salhy M, Titlbach M (1984) Evolution of the neuroendocrine system in vertebrates. A review with particular reference to the phylogeny and postnatal maturation of the islet parenchyma. In: Evolution and Tumor Pathology of the Neuroendocrine System, Falkmer S, Hakanson R, Sundler F, eds. Amsterdam: Elsevier Science Publication, pp 59-87. Fallcmer S, Ostberg Y (1985) Comparative morphology of pancreatic islets in animals. In: The diabetic pancreas, BW Vollc and ER Arquilla, chapt. 2. New york: Elsevier/North Holland. Failco JM, Crockett SE, Cataland 5, Mazzaferri EL (1975) Gastric inhibitory polypeptide (GIP) stimulated by fat ingestion in man. J. Clin. Endocrinol. Metab. 41:260-265. Falko 3M, Crockett SE, Cataland 5, ODorisio TM, Kramer W, Mazzaferri EL (1980) The effect of increasing doses of ingested glucose on insulin and gastric inhibitory polypeptide (GIP) concentrations in man. Clin. Endocrinol. 13:587-593. Farrell JI, Ivy AC (1926) Studies on the motility of the transplanted gastric pouch. Am. J. of Physiol. 46:227-228. Fehmann HC, Göke B, Göke R, Truatmann ME, Arnold R (1989) Synergistic stimulatory effect of glucagon-like peptide- 1 (7-36) amide and glucose-dependent insulinreleasing polypeptide on the endocrine rat pancreas. Fed. Eur. Biochem. Soc. 252:109-112.  168 Fehmann HC, Habener JF (1991a) Homologous desensitization of the insulinotropic glucagon-like peptide-I(7-37) receptor on insulinoma (HIT-T15) cells. Endocrinology 128:2880-2888. Fehmann HC, Habener IF (1991b) Functional receptors for the insulinotropic hormone glucagon-like peptide-I(7-37) on a somatostatin secreting cell line. Fed. Eur. Biochem. Soc. 279:335-340. Ferraris RP, Yasharpour S, Lloyd KCK, Mirzayan R, Diamond 3M (1990) Luminal glucose concentrations in the gut under normal conditions. Am. 3. Physiol. 22:G822G837. Fern GL, Adrian TE, Ghatei MA, O’Shaugnessy DJ, Probert L, Lee YC, Buchan AMJ, Polak JM, Bloom SR (1983) Tissue localization and relative distribution of regulatory peptides in separated layers from the human bowel. Gastroenterol. 84:777-786. Flaten 0 (1981) The effect of adrenoceptor blockade on the release of gastric inhibitory polypeptide after intraduodenal glucose in humans. Scand. 3. Gastroenterol. 16:641645. Flaten 0, Sand T, Myren 3 (1982) Beta-adrenergic stimulation and blockade of the release of gastric inhibitory polypeptide and insulin in man. Scand. 3. Gastroenterol. 17:283-288. Flatt PR, Kwasowski P, Bailey CJ (1989) Stimulation of gastric inhibitory polypeptide release in ob/ob mice by oral administration of sugars and their analogues. 3. Nutr. 119:1300-1303. Frier BM, Corrall 3M, Adrian TE, Bloom SR, Strachan RK, Heading RC (1984) The influence of adrenergic denervation on the response to feeding of the gastropancreatic system in man. Clin. Endocrinol. 21:639-647. Frohman LA, Downs TR, Heimer EP, Felix AM (1989) Dipeptidylpeptidase IV and trypsin-like enzymatic degradation of human growth hormone-releasing hormone in plasma. 3. Clin. Invest. 83:1533-1540. Frohman LA, Downs TR, Williams TC, Heimer EP, Pan YCE, Felix AM (1986) Rapid enzymatic degradation of growth hormone-releasing hormone by plasma in vitro and in vivo to a biologically inactive product cleaved at the NH 2 terminus. 3. Clin. Invest. 78:906-913. Fujii N, Sakurai M, Akaji K, Nomizu M, Yajima H, Mizuta K, Aono M, Moriga M, Inoue K, Hosotani R, Tobe T (1986) Studies on peptides. CXXXIX. Solution synthesis of a 42-residue peptide corresponding to the entire amino acid sequence of human glucose-dependent insulinotropic polypeptide (GIP). Chem. Pharm. Bull. 34:2397-2410. Fukase N, Igarashi M, Takahashi H, Manaka H, Yamatani K, Daimon M, Tominaga M, Sasaki H (1993) Hypersecretion of truncated glucagon-like peptide-l and gastric inhibitory polypeptide in obese patients. Diabet. Med. 10:44-49. Fushiki T, Kojima A, Imoto T, Inoue K, Sugimoto E (1992) An extract of Gymnema sylvestre leaves and purified gymnemic acid inhibits glucose-stimulated gastric inhibitory peptide secretion in rats. 3. Nutr. 122:2367-2373.  169 Gallacher DV, Hanley MR, Petersen OH, Roberts ML, Squire-Pollard LG, Yule DI (1990) Substance P and bombesin elevate cytosolic Ca 2 by different molecular mechanisms in a rat pancreatic acinar cell line. J. Physiol. 426:193-207 Gallwitz B, Witt M, Fölsch UR, Creutzfeldt W, Schmidt WE (1993) Binding specificity and signal transduction of receptors for glucagon-like peptide-1(7-36)amide and gastric inhibitory polypeptide on RINm5F insulinoma cells. J. Mol. Endocrinol. 10:259-268. Garcia MC, Hermans MP, Henquin IC (1988) Glucose-, calcium- and concentrationdependence of acetyicholine stimulation of insulin release and ionic fluxes in mouse islets. Biochem. J. 254:2 1 1-2 18. Gayle RG, Ludewig RM (1978) Effect of truncal, selective and highly selective vagotomy on fat-induced gastric inhibitory polypeptide release. Surg. Forum 29:397402. Gillespie IE, Grossman MI (1964) Inhibitory effect of secretion and cholecystokinin on Heindenhain pouch responses to gastrin extract and histamine. Gut 5:342-345. Giraud AS, Soll AH, Cuttita F, Walsh JH (1987) Bombesin stimulation of gastrin release from canine gastrin cells in primary culture. Am. J. Physiol. 252:G413-G420. Göke R, Conlon JM (1988) Receptor for glucagon-like peptide-1(7-36)amide on rat insulinoma-derived cells. 3. Endocrinol. 166:357-362. Goto Y, Berelowitz M, Frohman LA (1981) Effect of catecholamines on somatostatin secretion by isolated perfused rat stomach. Am. J. Physiol. 240:E274-E278. Grant SGN, Seidman I, Hanahan D, Bautch V (1991) Early invasiveness characterizes metastatic carcinoid tumors in transgenic mice. Cancer Res. 5 1:4917-4923. Greely Jr GH, Spannagel A, Hill FLC, Thompson JC (1986a) Comparison of the actions of bombesin, gastrin-releasing peptide-27, neuromedin B, and gastrin-releasing peptide-lO in causing release of gastrin and gastric inhibitory peptide in rats. Proc. Soc. Exp. Biol. Med. 183:136-139. Greely Jr GH, Spannagel A, Trowbridge I, Thompson JC (1986b) Effect of bombesin and gastrin-releasing peptide on the release of gastric inhibitory polypeptide and insulin in rats. Proc. Soc. Exp. Biol. Med. 182:540-542. Greenwood FC, Hunter WM, Glover IS (1963) The preparation of j13 1 -labelled human growth hormone of high specific radioactivity. Biochem. 1. 89:114-123. Groop PH (1989) The influence of bodyweight, age and glucose tolerance on the relationship between GIP secretion and beta-cell function in man. Scand. J. Clin. Lab. Invest. 49:367-379. Gros L, Thorens B, Bataille D, Kervran A (1993) Glucagon-lilce peptide-1-(7-36)amide, oxyntomodulin, and glucagon interact with a common receptor in a somatostatin secreting cell line. Endocrinology 133:631-638.  170 Gutniak M, Ørskov C, Hoist JJ, Ahrén B, Efendic S (1992) Antidiabetogenic effect of glucagon-like peptide-1 (7-36)amide in normal subjects and patients with diabetes mellitus. N. Engi. J. Med. 326:1316-1322. Hicki WH, Bloom SR, Mitznegg P, Domschke W, Domschke S, Belohiavek D, Demling L, WUnsch E (1977) Plasma secretin and pancreatic bicarbonate response to exogenous secretin in man. Gut 18:191-195. Hanks JB, Andersen DK, Wise JE, Putnam WS, Meyers WC, Jones RS (1984) The hepatic extraction of gastric inhibitory polypeptide and insulin. Endocrinology 115:1011-1018. Hampton SM, Kwasowski P, Tan K, Morgan LM, Marks V (1983) Effect of pretreatment with a high fat diet on the gastric inhibitory polypeptide and insulin responses to oral triolein and glucose in rats. Diabetologia 24:278-28 1. Hanahan D (1984) Heritable formation of pancreatic B cell tumors in transgenic mice expressing recombinant insulin/simian virus 40 oncogenes. Nature (Lond.) 315:115122. Hanahan D (1988) Dissecting multistep tumorigenesis in transgenic mice. Annu. Rev. Genet. 22:479-519. Hanks JB, Andersen DK, Wise JE, Putnam WS, Meyers WC, Jones RS (1984) The hepatic extraction of gastric inhibitory polypeptide and insulin. Endocrinology 115:1011-1018. Hartmann E, Ebert R, Creutzfeldt W (1986) Insulin-dependent inhibition of hepatic glycogenolysis by gastric inhibitory polypeptide (GIP) in perfused rat liver. Diabetologia 29:112-114. Hassall CJS, Allen TGJ, Pittam BS, Burnstock G (1989) The use of cell and tissue culture techniques in the study of regulatory peptides. EXS 56:113-136. Hauner H, Glatting G, Kaminska D, Pfeiffer EF (1988) Effects of gastric inhibitory polypepetide on glucose and lipid metabolism of isolated rat adipocytes. Ann. Nutr. Metab. 32:282-288. Hazelwood RL (1989) Evolutionary, phylogenetic, and comparative considerations. In: The endocrine pancreas. Prentice Hall. chpt. 10. Ho LT, Pu HF, Sheu WJ, Wang WC, Wang PSG (1987) Inhibition of somatostatin of glucose-induced release of gastric inhibitory polypeptide in rats. Chin. J. Physiol. 30:45-53. Holst JJ, Ørskov C, Vagn Nielson 0, Schwartz T (1987) Truncated glucagon-like peptide 1, an insulin-releasing honuone from the distal gut. FEBS Lett. 211:169-174. Holz GG, Habener IF (1992) Signal transduction crosstalk in the endocrine system: pancreatic B-cells and the glucose competence concept. TIBS 17:388-393. Holz GG, Kühtreiber WM, Habener iF (1993) Pancreatic beta-cells are rendered glucosecompetent by the insulinotropic hormone glucagon-like peptide- 1(7-37). Nature 361:362-365.  171 Hopfer U (1987) Membrane transport mechanisms for hexoses and amino acids in the small intestine. In: Physiology of the gastrointestinal tract. 2nd ed. LR Johnson (ed) Raven Press, New York, Chpt 55. Hopsu-Havu VK, Glenner GG (1966) A new dipeptide naphthylamidase hydrolyzing glycyl-prolyl-13-naphthylamide. Histochemie 7:197-201. Imamura M, Kameyama J, Naito H, Sato T, Ohneda A (1984) Influence of vagotomy upon GIP release in patients with peptic ulcer. Tohuku J. Exp. Med. 143:335-344. Imamura M, Takahashi M, Sasaki I, Yamauchi H, Sato T (1988) Effects of the pathway of bile flow on the digestion of fat and release of gastrointestinal hormones. Am. J. Gastroenterol. 83:386-392. Inagaki N, Maekawa T, Sudo T, Ishii 5, Seino Y, Imura H (1992) C-Jun represses the human insulin promoter activity that depends on multiple cAMP response elements. Proc. Nati. Acad. Sci. USA 89: 1045-1049. Inagaki N, Seino Y, Takeda J, Yano H, Yamada Y, Bell GI, Eddy RL, Fukushima Y, Byers MG, Shows TB, Imura H (1989) Gastric inhibitory polypeptide: structure and chromsomal location of the human gene. Mol. Endocrinol. 3:1014-1021. Ishida T, Chap Z, Chou J, Lewis R, Hartley C, Entman M, Field J (1983) Differential effects of oral, peripheral intravenous and intraportal glucose on hepatic glucose uptake and insulin and glucagon extraction in conscious dogs. J. Clin. Invest. 72:590601. Ivy AC, Oldberg E (1927) Contraction and evacuation of the gall-bladder caused by highly purified “secretin” preparation. Proc. Soc. Exp. Biol. Med. 25:113-115. Jaspan JB, Rubenstein AH (1977) Circulating glucagon plasma profiles and metabolism in health and disease. Diabetes 26:887-904, Johnson JH, Ogawa A, Chen L, Orci L, Newgard CB, Alam T, Unger RH (1990) Under expression of B cell high Km glucose transporters in noninsulin-dependent diabetes. Science 250:546-549. Jones JR. Owens DR, Luzio SD, Hayes TM (1989a) Obesity is associated with increased postprandial GIP levels which are not reduced by dietary restriction and weight loss. Diabet. Metab. 15:11-22. Jones IR, Owens DR, Luzio SD, Hayes TM (1989b) Glucose dependent insulinotropic polypeptide (GIP) infused intravenously is insulinotropic in the fasting state in type 2 (non-insulin dependent) diabetes mellitus. Horm. Metabol. Res. 21:23-26. Jones IR, Owens DR, Moody AJ, Luzio SD, Morris T, Hayes TM (1987) The effects of glucose-dependent insulinotropic polypeptide infused at physiological concentrations in normal subjects and type 2 (non-insulin-dependent) diabetic patients on glucose tolerance and B-cell secretion. Diabetologia 30:707-711. Jones IR, Owens DR, Vora J, Luzio SD, Hayes TM (1989c) A supplementary infusion of glucose-dependent inuslinotropic polypeptide (GIP) with a meal does not significantly improve the B cell response or glucose tolerance in type 2 diabetes mellitus. Diabetes Res. Clin. Prac. 7: 263-9.  172 Jones IR, Owens DR, Sarsons DL, Bloom SR (1985) Day profiles of glucose dependent insulinotropic polypeptide (GIP) in normal subjects. Horm. Metabol. Res. 17:660662. Jorde R, Burhol PG, Schulz TB (1983) Fasting and postprandial plasma GIP values in man measured with seven different antisera. Reg. Pept. 7:87-94. Jorde R, Burhol PG, Waldum HL (1981) The effect of a test meal on plasma gastric inhibitory polypeptide and serum gastrin in patients with achiorhydria and hypergastrinemia, and in patients with normal gastric H secretion. Digest. 21:304309. Jorde R, Burhol PG, Waldum HL, Schulz TB, Lygren I, Florholmen J (1980) Diurnal variation of plasma gastric inhibitory polypeptide in man. Scand. J. Gastroenterol. 15:617-619. Jörnvall H, Carlquist M, Kwauk S, Otte SC, McIntosh CHS, Brown JC, Mutt V (1981) Amino acid sequence and heterogeneity of gastric inhibitory polypeptide (GIP). FEBS Lett. 123:205-210. Kieffer TJ, Verchere CB, Fell CD, Huang Z, Brown JC, Pederson RA (1993) Glucosedependent insulinotropic polypeptide (GJP)-stimulated insulin release from a tumorderived 13-cell line (13TC3). Can. J. Physiol. Pharmacol. 7 1:917-922. Kogire M, Izukura M, Gomez G, Uchida T, Greely Jr G, Thompson JC (1990) Terbutaline, a 13 -adrenoreceptor agonist, inhibits gastric acid secretion and stimulates 2 release of peptide YY and gastric inhibitory polypeptide in dogs. Dig. Dis. Sci. 35:453-457. Koop H, Behrens I, McIntosh CHS, Pederson RA, Arnold R, Creutzfeldt W (1981) Adrenergic modulation of gastric somatostatin release in rats. FEBS Lett. 118:248250. Koop I, Buchan AMJ (1992) Cholecystokinin release from isolated canine epithelial cells in short-term culture. Gastroenterol. 102:28-34. Komatsu R, Matsuyama T, Namba M, Watanabe N, Itoh H, Kono N, Tarui 5 (1989) Glucagonostatic and insulinotropic action of glucagonlike peptide 1 -(7-36)-amide. Diabetes 38:902-905. Kosaka T, Lim RKS (1930) Demonstration of the humoral agent in fat inhibition of gastric secretion. Proc. Soc. Exp. Biol. Med. 27:890-891. Kraenzlin ME, Ch’ng JLC, Mulderry PK, Ghatei MA, Bloom SR (1985a) Infusion of a novel peptide, calcitonin gene-related peptide (CGRP) in man. Pharmacokinetics and effects on gastric acid secretion and on gastrointestinal hormones. Regul. Pept. 10:189-197. Kraenziin ME, Wood SM, Neufeld M, Adrian TE, Bloom SR (1985b) Effect of long acting somatostatin-analogue, SMS 201995, on gut hormone secretion in normal subjects. Experimentia (Birkhäuser Verlag) 41:738740.  173 Krarup T (1988) Immunoreactive gastric inhibitory polypeptide. Endocrine Rev. 9:122134. Krarup T, Hoist JJ (1984) The heterogeneity of gastric inhibitory polypeptide in porcine and human gastrointestinal mucosa evaluated with five different antisera. Reg. Pept. 9:35-46. Krarup T, Hoist JJ, Knuhtsen S, Moody AJ, Nielsen OV (1987a) Effect of porcine 8000 dalton immunoreactive gastric inhibitory polypeptide on the isolated perfused pig pancreas. Diabetes 36:7A (Abstract). Krarup T, Hoist JJ, Larsen KL (1985) Responses and molecular heterogeneity of IR-GIP after intraduodenal glucose and fat. Am. 3. Physiol. 249:E195-E200. Krarup T, Hoist JJ, Madsbad S (1987b) Heterogeneity of immunoreactive gastric inhibitory polypeptide in the plasma of newly diagnosed type 1 (insulin-dependent) diabetics. Acta Endocrinol. (Copenh.) 114:74-78. Krarup T, Saurbrey N, Moody AJ, KUhl C, Madsbad S (1987c) Effects of porcine gastric inhibitory polypeptide on beta-cell function in type I and II diabetes meilitus. Metabolism 36:677-682. Kreymann B, Ghatei MA, Burnet P, Williams G, Kanse S, Diani AR, Bloom SR (1989) Characterization of glucagon-like peptide-1-(7-36)amide in the hypothalamus. Brain Res. 502:325-331. Kreymann B, Ghatei MA, Williams G, Bloom SR (1987) Glucagon-like peptide 1 7-36: a physiological incretin in man. Lancet 2:1300-1304. Kuzio M, Dryburgh JR, Malloy KM, Brown JC (1974) Radioimmunoassay for gastric inhibitory polypeptide. Gastroenterol. 66:357-364. Kwasowski P, Flatt PR, Bailey CT, Marks V (1985) Effects of fatty acid chain length and saturation on gastric inhibitory polypeptide release in obese hyperglycaemic (ob/ob) mice. Biosc. Rep. 5:701-705. Kwok YN, McIntosh CHS, Sy H, Brown JC (1988a) Inhibitory action of tachykinins and neurokinins on somatostatin release from the isolated perfused stomach. J. Pharmacol. Exp. Therap. 246:726-731. Kwok YN, Verchere CB, McIntosh CHS, Brown JC (1988b) Effect of galanin on endocrine secretions from the isolated perfused rat stomach and pancreas. Eur. J. Pharmacol. 145:49-54. Labarre J (1927) Sur l’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) Stir les possibilities d’un traitement du diabete par Pincrétine. Bull. Acad. Roy. Soc. Med. Belg. 12:620-634. Labarre J, Still EV (1930) Studies on the physiology of secretin. Am. J. Physioi. 91:649653.  174 Lardinois CK, Mazzaferri EL, Stanch GH, Hollenbeck CB, Reaven GM (1985) Increased gastric inhibitory polypeptide is not reduced in patients with noninsulin-dependent diabetes mellitus treated with intense insulin therapy. J. Clin. Endocrinol. Metab. 61: 1089-1092. Lardinois CK, Richeson III RB, Stanch GH, Mazzu D, Mazzaferri EL (1990) Gastric inhibitory polypeptide mechanisms of augmenting insulin secretion. Life Sci. 47:1015-1022. Lardinois CK, Stanch GH, Mazzaferri EL, DeLett A (1988) Polyunsaturated fatty acids augment insulin secretion. J. Am. College Nutr. 6:507-5 15. Larrimer JN, Mazzaferri EL, Cataland S, Mekhjian HS (1978) Effect of atropine on glucose-stimulated gastric inhibitory polypeptide. Diabetes 27:638-642. Larrson LI, Moody AJ (1980) Glicentin and gastric inhibitory polypeptide immunoreactivity in endocrine cells of the gut and pancreas. J. Histochem. Cytochem. 28:925-933. Lauritsen KB, Christensen KC, Stokhoim KH (1980) Gastric inhibitory polypeptide (GIP) release and incretin effect after oral glucose in obesity and after jejunoileal bypass. Scand. J. Gastroenterol. 15:489-492. Lauritsen K, Hoist 33, Moody AJ (1981) Depression of insulin release by anti-GIP serum after oral glucose in rats. Scand. J. Gastroenterol. 16:417. Lauritsen KB, Moesgaard F, Nielson ML, Jensen SL (1982) Release of gastric inhibitory polypeptide and insulin in response to intrajejunal glucose in duodenal ulcer patients before and after truncal vagotomy. Scand. J. Gastroenterol. 17:507-511. Leidenheimer NJ, Browning MD, Dunwiddie TV, Hahner LD, Harris RA (1990) Phosphorylation-independent effects of second messenger system modulators on y aminobutyric acid receptor complex function. Mol. Pharmacol. 38:823-828. Levitt NS, Vinik AT, Child PT (1980) Glucose-dependent insulin-releasing peptide in noninsulin-dependent maturity-onset diabetes: effects of autonomic neuropathy. 3. Clin. Endocrinol. Metab. 51:254-258. Lindahl PE (1948) Principle of a counter-streaming centrifuge for the separation of particles of different sizes. Nature 161:648-649. Loew ER, Gray JS, Ivy AC (1940) Is a duodenal hormone involved in carbohydrate metabolism? Am. J. Physiol. 129:659-663. Lopez LC, Frazier ML, Su CJ, Human A, Sanders GF (1983) Mammalian pancreatic preproglucagon contains three glucagon-related peptides. Proc. Nati. Acad. Sc. (U.S.A.) 80:5485-5489. Lu M, Wheeler MB, Leng XH, Boyd Ill AE (1993) The role of free cytosolic calcium level in B-cell signal transduction by gastric inhibitory polypeptide and glucagon-like peptide 1(7-37). Endocrinology 132:94-100. Macgregor PF, Abate C, Curran T (1990) Direct cloning of leucine zipper proteins: Jun binds cooperatively to the CRE with CRE-BP1. Oncogene 5:45 1-458.  175 Magee DF, Nakamura M (1966) Action of pancreozymin preparations on gastric secretion. Nature 212: 1487-1488. Maletti M, Altman 33. Hoa DHB, 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: 13361340. Maletti M, Cariquist 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 [Suppi 1] 7:75-78. Maletti M, Portha B, Cariquist M, Kergoat M, Laburthe M, Marie JC, Rosselin G (1984) Evidence for and characterization of specific high affinity binding sites for gastric inhibitory polypeptide in pancreatic B-cells. Endocrinology 115:1324-1331. Mangel AW, Snow ND, Misukonis MA, Basavappa S, Middleton JP, Fitz JO, Liddle RA (1993) Calcium-dependent regulation of cholecystokinin secretion and potassium currents in STC-1 cells. Am. 3. Physiol. 264:G1031-G1036. Martin RA, Cleary DL, Guido DM, Zurcher-Neely HA, Kubiak TM (1993) Dipeptidyl peptidase IV (DPP-IV) from pig kidney cleaves analogs of bovine growth hormonereleasing factor (bGRF) modified at position 2 with Ser, Thr, or Val. Extended DPP IV substrate specificity? Biochim. Biophys. Acta 1164:252-260. Matschinsky FM (1990) Glucokinase as glucose sensor and metabolic signal generator in pancreatic B-cells and hepatocytes. Diabetes 39:647-652. Matthews EK, Sakamoto Y (1975) Electrical characteristics of pancreatic islet cells. J. Physiol. Lond. 246:421-437. Matumura C (1985) Dipeptidyl aminopeptidase IV (DAP IV) in urine. 3. Jpn. Pediatr. Soc. 33:287-291. Maxwell V, Shulkes A, Brown JC, Solomon TB, Walsh JH, Grossman MI (1980) Effects of gastric inhibitory polypeptide on pentagastrin-stimulated acid secretion in man. Dig. Dis. Sci. 25:113-116. Mazzaferri EL, Starich GH, Lardinois CK, Bowen GD (1985) Gastric inhibitory polypeptide responses to nutrients in caucasians and american indians with obesity and noninsulin-dependent diabetes mellitus. J. Clin. Endocrinol. Metab. 61:313-321. McDonald TJ, Ghatei MA, Bloom SR. Track NS, Radziuk 3, Dupré 3, Mutt V (1981) A qualitative comparison of canine plasma gastroenteropancreatic hormone responses to bombesin and the porcine gastrin-releasing peptide (GRP). Regul. Pept. 2:293-304. McIntosh C, Arnold R, Boothe E, Becker H, Köbberling 3, Creutzfeldt W (1978) Gastrointestinal somatostatin: extraction and radioimmunoassay in different species. Gut 19:655-663. McIntosh CHS, Bakich V, Brown JC, Kwok YN, Nishimura E, Pederson RA (1989) Somatostatin and gastrin secretion from the isolated perfused rat stomach: neural and hormonal regulation and the effect of the ulcerogen cysteamine. In: Szabo S (ed)  176 Ulcer disease: new aspects of pathogenesis and pharmacology. CRC Press, Boca Raton pp. 217-228. McIntosh CHS, ha X, Kwok YN (1990) Characterization of the opioid receptor type mediating inhibition of rat gastric somatostatin release. Am. J. Physiol. 259:0922G927. McIntosh CHS, Kwok YN, Mordhorst T, Nishimura E, Pederson RA, Brown JC (1983) Enkephalinergic control of somatostatin secretion from the perfused rat stomach. Can. J. Physiol. Pharmacol. 61:657-663. McIntosh CHS, Kwok YN, Tang C, Brown JC (1987) The use of monoclonal antibodies in RIA for gastrointestinal hormones. I. Clin. Immunoassay 10:79-84. McIntosh CHS, Pederson RA, Koop H, Brown JC (1979) Inhibition of GIP stimulated somatostatin-like immunoreactivity (SLI) by acetyicholine and vagal stimulation. In: Miyoshi A, ed. Gut Peptides: secretion, function and clinical aspects. Tokyo: Kodansha 100-104. McIntosh CHS, Pederson RA, Koop H, Brown JC (1981) Gasthc inhibitory polypeptide stimulated secretion of somatostatin-like immunoreactivity from the stomach: inhibition by acetyicholine or vagal stimulation. Can. J. Physiol. Pharmacol. 59:468472. McIntyre N, Holdsworth CD, Turner DS (1964) New interpretation of oral glucose tolerance. Lancet 11:20-2 1. McLoughlin JC, Buchanan KD, Alam MJ (1979) A glycoside-hydrolase inhibitor in treatment of dumping syndrome. Lancet 2:603-605. Mei N (1978) Vagal glucoreceptors in the small intestine of the cat. 3. Physiol. 282:485506. Meneilly GS, Bryer-Ash M, Elahi D (1993) The effect of glyburide on B-cell sensitivity to glucose-dependent insulinotropic polypeptide. Diabetes Care 16:110-114. Mentlein R, Gallwitz B, Schmidt WE (1993) Dipeptidyl-peptidase IV hydrolyses gastric inhibitory polypeptide, glucagon-like peptide- 1(7 -36)amide, peptide histidine methionine and is responsible for their degradation in human serum. Eur. J. Biochem. 214:829-835. Modlin IM, Mitchell SJ, Bloom SR (1978) The systemic release and pharmacokinetics of VIP. In: Bloom SR, ed. Gut hormones. Edinburgh: Churchill Livingstone, 470-474. Mojsov 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. J. Biol. Chem. 261:11880-11889. Mojsov S, Weir GC, Habener iF (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. I. Clin. Invest. 79:616-619. Moody AJ, Damm JØrgensen K, Thim L (1981) Stucture-function relationships in porcine GIP. Diabetologia 21:306 (Abstract).  177  Moody AJ, Krarup T, Larsen UD (1992) The reactivity of anti-porcine gastric inhibitory polypeptide (GIP) rabbit serum R65 with synthetic human and porcine GIP. Scand. J. Clin. Invest. 52:103-106. Moody AJ, Lauritsen KB (1977) A sensitive radio-immunoassay for plasma GIP. In: Speranza V, Basso N, Lezoche E (eds) Gastrointestinal Hormones and Pathology of the Digestive System Arti Grafiche Tris. Rome, p 148 (Abstract). Moody AJ, Thim L, Valverde I (1984) The isolation and sequencing of human gastric inhibitory polypeptide (GIP) FEBS Lett. 172:142-148. Moore B, Edie ES, Abram JM (1906) On the treatment of diabetes mellitus by acid extract of duodenal mucous membrane. Biochem. 1. 1:28-38. Morgan LM (1979) Immunoassayable gastric inhibitory polypeptide: investigations into its role in carbohydrate metabolism. Annals Clin. Biochem. 16:6-14. Morgan LM, Hampton SM, Tredger JA, Cramb R, Marks V (1988) Modification of gastric inhibitory polypeptide (GIP) secretion in man by a high-fat diet. Br. J. Nutr. 59:373-380. Morgan LM, Morris BA, Marks V (1978) Radioimmunoassay of gastric inhibitory polypeptide. Ann. Clin. Biochem. 15:172-177. Morgan LM, Tredger IA, Madden A, Kwasowski P, Marks V (1985) The effect of guar gum on carbohydrate, fat and protein stimulated gut hormone secretion. Br. J. Nutr. 53:467-475. Moroder L, Hallet A, Thamm P, Wilschowitz L, Brown JC, Wünsch E (1978) Studies on gastric inhibitory polypeptide: synthesis of the actatriacontapeptide GIP1-38 with full insulinotropic activity. Scand. I. Gastroent. [Suppi. 49] 13:129 (Abstract). Muller MK, Pederson RA, Brown JC (1982) Interaction of gastric inhibitory polypeptide and arginine in the perfused pancreas of rats treated with triiodothyronine. Can. 3. Physiol. Pharmacol. 60:297-301. Nagamatsu 5, Steiner D (1992) Altered glucose regulation of insulin biosynthesis in insulinoma cells: mouse BTC3 cells secrete insulin-related peptides predominantly via a constitutive pathway. Endocrinology 130:748-754. Nagata T, Umeda Y, Urakami M, Inada M (1988) Urinary dipeptidyl aminopeptidase IV (DAP IV) activity in patients with impaired glucose tolerance and diabetes mellitus. J. Jpn. Diabetic Soc. 31:169-171. Nathan DM, Schreiber E, Fogel H, Mojsov S, Habener JF (1992) Insulinotropic action of glucagonlike peptide-I-(7-37) in diabetic and nondiabetic subjects. Diabetes Care 15:270-276. Nauck MA, Bartels E, ørskov C, Ebert R, Creutzfeldt W (1992) Lack of effect of synthetic human gastric inhibitory polypeptide and glucagon-like peptide 1 [7-36 amide] infused at near-physiological concentrations on pentagastrmn-stimulated gastric acid secretion in normal human subjects. Digestion 52:214-221.  a  178 Nauck MA, Heimesaat MM, Ørskov C, Hoist JJ, Ebert R, Creutzfeldt W (1993) Preserved incretin activity of glucagon-like peptide 1 [7-36 amide] but not of synthetic human gastric inhibitory polypeptide in patients with type 2 diabetes mellitus. J. Clin. Invest. 91:301-307. Nauck MA, Homberger E, Siegel EG, Allen RC, Eaton RP, Ebert R, Creutzfeldt W (1986) Incretin effects of increasing glucose loads in man calculated from venous insulin and C-peptide responses. J. Clin. Endocrinol. Metab. 63:492-498. Nauck MA, Schmidt WE, Ebert R, Strietzel J, Cantor P, Hoffmann G, Creutzfeldt W (1989) Insulinotropic properties of synthetic human gastric inhibitory polypeptide in man: interactions with glucose, phenylalanine, and cholecystokinin-8. J. Clin. Endocrinol. Metab. 69:654-662. Nelson RL, Go VLW, McCullough AJ, Ilstrup DM, Servic FJ (1986) Lack of a direct effect of the autonomic nervous system on glucose-stimulated gastric inhibitory polypeptide (GIP) secretion in man. Dig. Dis. Sci. 3 1:929-935. Nicholl C, Sevelius H, Bloom SR (1986) Enprostil markedly reduces the postprandial rises in insulin and glucose-dependent insulinotropic peptide. Diabet. Med. 3:358a (Abstract). Nukada 0, Kobayashi M, Moriwake T, Kanzaki 5, Himei H, Yoda T, Seino Y (1992) Urinary glycylprolyl dipeptidyl aminopeptidase (GP-DAP) in insulin-dependent diabetic patients. Acta Paediatr. 81:907-911. Oben J, Morgan L, Fletcher J, Marks V (1991) Effect of the enter-pancreatic hormones, gastric inhibitory polypeptide and glucagon-like polypeptide-1(7-36) amide, on fatty acid synthesis in explants of rat adipose tissue. J. Endocrinol. 130:267-272. O’Dorisio TM, Cataland 5, Stevenson M, Mazzaferri EL (1976) Gastric inhibitory polypeptide (GIP). Intestinal distribution and stimulation by amino acids and medium-chain triglycerides. Am. J. Dig. Dis. 21:761-765. ODoriso TM, Sirinek KR, Mazzaferri EL, Cataland S (1977) Renal effects on serum gastric inhibitory polypeptide (GIP). Metab. 26:651-656. Oektedalen 0, Opstad PK, Jorde R (1983) Increased plasma response of gastric inhibitory polypeptide to oral glucose and a liquid meal after prolonged starvation in healthy man. Digest. 26:114-123. Ogata 5, Misumi Y, Ikehara Y (1989) Primary structure of rat liver dipeptidyl peptidase IV deduced from its cDNA and identification of the NH -terminal signal sequence as 2 the membrane-anchoring domain. J. Biol. Chem. 264:3596-3601. OHalloran DJ, Nilcou GC, Kreymann B, Ghatei MA, Bloom SR (1990) Glucagon-like peptide-1 (7-36)-NH : a physiological inhibitor of gastric acid secretion in man. J. 2 Endocrinol. 126:169-173. Ohneda A, Kobayashi T, Nihei J (1983) Response of gastric inhibitory polypeptide (GIP) to oral administration of nutrients in normal and pancreatectomized dogs. Reg. Pept. 5:263-272.  179 Ohneda A, Kobayashi T, Nihei J (1984) Response of gastric inhibitory polypeptide to fat ingestion in normal dogs. Reg. Pept. 8:123-130. Opara EC, Go VLW (1993) Dual effects of gastric inhibitory polypeptide on insulin secretion. Pancreas 8:39-43. Orci L, Ravazzola M, Baetens D, Inman L, Amherdt M, Peterson RG, Newgard CB, Johnson IH, Unger RH (1990) Evidence that down-regulation of B-cell glucose transporters in non-insulin-dependent diabetes may be the cause of diabetic hyperglycemia. Proc. Nail. Acad. Sci. USA 87:9953-9957. ørskov C, Holst JJ (1987) Radio-immunoassays for glucagon-like peptides 1 and 2 (GLP-1 and GLP-2). Scand. J. Clin. Lab. Invest. 47:165-174. ørskov C, Jeppesen JU, Madsbad S, Hoist JJ (1991) Progiucagon products in plasma of noninsulin-dependent diabetics and nondiabetic controls in the fasting state and after oral glucose and intravenous arginine. J. Clin. Invest. 87:415-423. Oshima I, Hirota M, Ohboshi C, Shima K (1988) Comparison of half-disappearance times, distribution volumes and metabolic clearance rates of exogenous glucagon-like peptide 1 and glucagon in rats. Regul. Pept. 21:85-93. Oshima I, Yamamoto C, Shima K (1991) Half-disappearance time of endogenous GLP-1 immunoreactivity in man. Horm. Metab. Res. 23:240-242. Otte SC, Mutt V. McIntosh CHS, Brown JC (1984) Purification and amino acid compostion of an 8,000 dalton immunoreactive form of GIP. Dig. Dis. Sci. [Suppl.] 29:63S (Abstract). Ottlecz A, Samson WK, McCann SM (1985) The effects of gastric inhibitory polypeptide (GIP) on the release of anterior pituitary hormones. Peptides 6:115-119. Parker DS, Raufman JP, ODonohue TL, Bledsoe M, Yoshida H, Pisano JJ (1984) Amino acid sequences of helospectins, new members of the glucagon superfamily, found in gila monster venom. J. Biol. Chem. 259:11751-11755. Pavlov IV (1910) The work of the digestive glands. Translated by Thompson WH. Griffin, London. Pearse AGE (1968) Common cytochemical and ultrastructural characteristics of cells producing polypeptide hormones (the APUD series) and their relevance to thyroid and ukimobranchial C cells and calcitonin. Proc. R. Soc. Lond. Biol. Sci. 170:71-80. Pedersen 0, Hjollund E, Lindskov HO, Schwartz Sorensen N (1980) Increased insulin receptors on monocytes from insulin dependent diabetics after high starch high fibre diets. Diabetologia 19:306 (Abstract). Pederson RA (1971) The isolation and physiological actions of gastric inhibitory polypeptide. Vancouver, Canada: Univ. of B.C. PhD dissertation. Pederson RA, Brown JC (1972) Inhibition of histamine-, pentagastrin-, and insulinstimulated canine gastric secretion by pure ‘gastric inhibitory polypeptide’. Gastroenterol. 62:393-400.  180 Pederson RA, Brown IC (1976) The insulinotropic action of gastric inhibitory polypeptide in the perfused rat pancreas. Endocrinology 99:780-785. Pederson RA, Brown IC (1978) Interactions of gastric inhibitory polypeptide, glucose and arginine on insulin and glucagon secretion from the perfused rat pancreas. Endocrinology 102:610-615. Pederson RA, Buchan AMJ, Zahedi-Asi S, Chan CB, Brown IC (1982) Effect of jejuno ileal bypass in the rat on the entero-insular axis. Regul. Pept. 5:53-63. Pederson RA, Dryburgh JR, Brown JC (1975a) The effect of somatostatin on release and insulinotropic action of gastric inhibitory polypeptide. Can. J. Physiol. Pharmacol. 53:1200-1205. Pederson RA, Innis SM, Buchan AMJ, Chan CB, Brown IC (1985) The effect of total parenteral nutrition (TPN) on the enteroinsular axis in the rat. Regul. Pept. 10:199206. Pederson RA, Mochizuki T, Yanaihara C, Yanaihara N, Brown JC (1990) Reduced somatostatinotropic effect of a GIP fragment (pGIP[l.. ) with insulinotropic 2 ]-NH 30 activity. Digestion 46 [Suppi. 1] 86 (Abstract) Pederson RA, Schubert HE, Brown IC (1975b) Gastric inhibitory polypeptide. Its physiologic release and insulinotropic action in the dog. Diabetes 24:1050-1056. Perley MI, Kipnis DM (1967) Plasma insulin responses to oral and intravenous glucose: studies in normal and diabetic subjects. I. Clin. Invest. 46:1954-1962. Pictet RL, Rall LB, Phelps P, Rutter WJ (1976) The neural crest and the origin of the insulin-producing and other gastro-intestinal hormone-producing cells. Science 191:191-192. Pictet RL, Rutter WI (1972) Development of the embryonic endocrine pancreas. In: Steiner DF, Frenkel M, eds. Handbook of Physiology, Sec. 7, Am. Physiol. Soc., Washington pp 25-66. Pipeleers DG, Schuit FC, Veld PA, Maes E, Hooghe-Peters EL, Van de Winkel M, Gepts W (1985) Interplay of nutrients and hormones in the regulation of insulin release. Endocrinology 117:824-833. Poitras P, Dumont A, Cuber IC, Trudel L (1993) Cholinergic regulation of motilin release from isolated canine intestinal cells. Pept. 14:207-2 13. Polak 3M, Bloom SR, Kuzio M, Brown IC, Pearse AGE (1973) Cellular localization of gastric inhibitory polypeptide in the duodenum and jejunum. Gut 14:284-288. Porte D Ir, Girardier L, Seydoux J, Woliheim CB, Kanazawa Y, Postemak 3 (1973) Neural regulation of insulin secretion in the dog. 3. Clin. Invest. 52:210-214. Prentki M, Matschinsky FM (1987) Ca , cAMP, and phospholipid-derived messengers 2 in coupling mechanisms of insulin secretion. Physiol. Rev. 4:1185-1248.  181 Rabinovitch A, Dupré J (1972) Insulinotropic and glucagonotropic activities of crude preparation of cholecystokinin-pancreozymin. Clin. Res. 20:945 (Abstract). Rai A, Singh G, Raffaniello R, Eng J, Raufman JP (1993) Actions of Helodermatidae venom peptides and mammalian glucagon-like peptides on gastric chief cells. Am. J. Physiol. 265:G1 18-G 125. Raufman JP, Singh L, Eng J (1991) Extendin-3, a novel peptide from Heloderma horridum venom, interacts with vasoactive intestinal peptide receptors and a newly described receptor on dispersed acini from guinea pig pancreas. J. Biol. Chem. 266:2897-2902. Rasmussen H, Zawalich KC, Ganesan S, Calle R, Zawalich WS (1990) Physiology and pathophysiology of insulin secretion. Diabetes Care 13:655-666. Rindi G, Grant SGN, Yiangou Y, Ghatei MA, Bloom SR, Bautch VL, Solcia E, Polak JM (1990) Development of neuroendocrine tumors in the gastrointestinal tract of transgenic mice. Am. 3. Pathol. 136:1349-1363. Robberecht P, Gourlet P. Cauvin A, Buscail L, De Neef P. Arimura A, Christophe J (1991) PACAP and VIP receptors in rat liver membranes. Am. J. Physiol. 260:G97G102. Roberge JN, Brubaker PL (1991) Secretion of proglucagon-derived peptides in response to intestinal luminal nutrients. Endocrinology 128:3169-3174. Roberge JN, Brubaker PL (1993) Regulation of intestinal proglucagon-derived peptide secretion by glucose-dependent insulinotropic peptide in a novel enteroendocrine loop. Endocrinology 133:233-240. Ross SA, Brown JC, Dupré 3 (1977) Hypersecretion of gastric inhibitory polypeptide following oral glucose in diabetes mellitus. Diabetes 26:525-529. Ross SA, Dupré 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, Shaffer EA (1981) The importance of triglyceride hydrolysis for the release of gastric inhibitory polypeptide. Gastroenterol. 80:108-111. Rosselin G (1986) The receptors of the VIP family peptides (VIP, secretin, GRF, PHI, PHM, GIP, glucagon and oxyntomodulin). Specificities and identity. Pept. [Suppl. 1] 7:89-100. Rossowski WJ, Zacharia S. Mungan Z, Ozmen V, Ertan A, Baylor LM, Jiang NY, Coy DH (1992) Reduced gastric acid inhibitory effect of a pGIP(1-30)NH 2 fragment with potent pancreatic amylase inhibitory activity. Regul. Pept. 39:9-17. Salera M, Ebert R, Giacomoni P. Pironi L, Venturi S, Corinaldesi R, Miglioli M, Barbara L (1982a) Adrenergic modulation of gastric inhibitory polypeptide secretion in man. Dig. Dis. Sci. 27:794-800.  182 Salera M, Giacomoni P. Pironi L, Ustra C, Capelli M, Giorgi A, Miglioli M, Barbara L (1983) Circadian rhythm of gastric inhibitory polypeptide (GIP) in man. Metabolism 32:21-24. Salera M, Pironi L, Giacomoni P, Venturi S, Capelli M, Miglioli M, Barbara L (1982b) Effect of somatostatin on fasting and glucose-stimulated gastric inhibitory polypeptide release in man. Digest. 24:126-132. Sandberg E, Ahrén B, Tendler D, Carlquist M, Efendic S (1986) Potentiation of glucoseinduced insulin secretion in the perfused rat pancreas by porcine GIP (gastric inhibitory polypeptide), bovine GIP, and bovine GIP (1-39). Acta. Physiol. Scand. 127:323-326. Sandberg E, Ahrén B, Tendler D, Efendic S (1988) Cholecystokinin-33 potentiates and vasoactive intestinal peptide inhibits gastric inhibitory polypeptide-induced insulin secretion in the perfused rat pancreas. Acta Endocrinol. (Copenhagen) 117:545-551. Sarson DL, Bryant MG, Bloom SR (1980) A radioimmunoassay of gastric inhibitory polypeptide in human plasma. J. Endocrinol. 85:487-496. Sarson DL, Hayter RC, Bloom SR (1982) The pharmacokinetics of porcine glucosedependent insulinotropic polypeptide (GJP) in man. Eur. J. Clin. Invest. 12:457-461. Schafer R, Schatz H (1979) Stimulation of (pro-)insulin biosynthesis and release by gastric inhibitory poypeptide isolated islets of rat pancreas. Acta Endocrinol. (Copenhagen) 91:493-500. 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. FEBS Lett. 68:225-227. Schjoldager BTG, Mortensen PE, Cliristiansen J, ørskov C, Holst JJ (1989) GLP-1 (glucagon-like peptide 1) and truncated GLP-1, fragments of human proglucagon, inhibit gastric acid secretion in humans. Dig. Dis. Sci. 34:703-708. Schluter KJ, Petersen KG, Kerp L (1980) Rapid changes in insulin binding affinity of monocytes in male and female volunteers after breakfast. Diabetologia 19:3 13 (Abstract). Schmid R, Schusdziarra V, Aulehner R, Weigert N, Classen M (1990) Comparison of GLP- 1 (7-36amide) and GIP on release of somatostatin-like immunoreactivity and insulin from the isolated rat pancreas. Z. Gastroenterol. 28:280-284. Schmidt WE, Siegel EG, Gallwitz B, Kümmel H, Ebert R, Creutzfeldt W (1986) Characterization of the insulinotropic activity of fragments derived from gastric inhibitory polypeptide. Diabetologia 29:591A (Abstract). Schmidt WE, Siegel EG, Kümmel H, Gallwitz B, Creutzfeldt W (1987) Commercially available preparations of porcine glucose-dependent insulinotropic polypeptide (GIP)  183 contain a biologically inactive GIP-fragment and cholecystokinin-3 31-39. Endocrinology 120:835-837. Schmidtler J, Schepp W, Janczewska I, Weigert N, FUrlinger C, Schusdziarra V. Classen M (1991) GLP-1-(7-36) amide, -(1-37), and -(1-36) amide: potent cAMP-dependent stimuli of rat parietal cell function. Am 3. Physiol. 260:G940-G950. Schulz TB, Jorde R, Burhol PG, Christensen 0 (1982) Augmented release of gastric inhibitory polypeptide into the portal vein in response to intraduodenal glucose and amino acids in anesthetized rats treated with methyiprednisolone of alloxan. Scand. J. Gastroenterol. 17:357-362. Schusdziarra V, Harris V, Conlon JM, Arimura A, Unger RH (1978) Pancreatic and gastric somatostatin release in response to intragastric and intraduodenal nutrients and HCL in the dog. 3. Clin. Invest. 62:509-5 18. Schusdziarra V, Rouiller D, Harris V. Pfeiffer EF, Unger RH (1980) Effect of bombesin upon plasma somatostatin-lilce immunoreactivity, insulin and glucagon in normal and chemically sympathectomized dogs. Regul. Pept. 1:89-96. Schwartz KE, Saito T (1989) Suppresion of alimentary lipemia in man by a prostaglandin analogue (enprostil). Atherosci. 79:23 1-236. Seal AM, Meloche RM, Liu YQE, Buchan AMJ, Brown JC (1987) Effects of monoclonal antibodies to somatostatin on somatostatin-induced and intestinal fat-induced inhibition of gastric acid secretion in the rat. Gastroenterol. 92:1187-1192. Service Fl, Hall LD, Westland RE, O’Brien PC, Go VLW, Matmond MW, Rizza RA (1983) Effects of size, time of day and sequence of meal ingestion on carbohydrate tolerance in normal subjects. Diabetologia 25:316-321. Service FJ, Nelson RL, Rubenstein AH, Go VLW (1978) Direct effect of insulin on secretion on insulin, glucagon, gastric inhibitory polypeptide and gastrin during maintenance of normoglycemia. J. Clin. Endocrinol. Metab. 47:488-493. Service FJ, Rizza RA, Westland RE, Hall LD, Gerich JE, Go VLW (1984) Gastric inhibitory polypeptide in obesity and diabetes. J. Clin. Endocrinol. Metab. 58:11331140. Sharma SK, Austin C, Howard A, Lo G, Nicholl CG, Legon S (1992) Characterization of rat gastric inhibitory peptide eDNA. J. Mo!. Endocrinol. 9:265-272. Sharp GWG, Wiedenkeller DE, Kaelin D, Siegel EG, Wollheim CB (1980) Stimulation of adenylate cyclase by Ca 2 and calmodulin in rat islets of langerhans. Diabetes 29:74-77. Sherwin RS, Basti C, Finkeistein FO (1976) Influence of uremia and hemodialysis on the turnover and metabolic effects of glucagon. 3. Clin. Invest. 57 :722-731. Sheu WJ, Pu HF, Wang SW, Ho LT, Wang PSG (1987) Metabolic clearance rate and secretion rate of gastric inhibitory polypeptide in the rat. Chin. J. Physiol. 30:25-33.  184 Shima K, Hirots M, Ohboshi C (1988) Effect of glucagon-like peptide-1 on insulin secretion. Regul. Pept. 22:245-252. Shimazu I, Hirota M, Ohboshi C, Shima K (1987) Identification and localization of glucagon-like peptide-1 and its receptor in rat brain. Endocrinology 121:1076-1082. Siegel EG, Creutzfeldt W (1985) Stimulation of insulin release in isolated islets by GIP in physiological concentrations and its relation to islet cyclic AMP content. Diabetologia, 28: 857-861. Siegel EG, Schulze A, Schmidt WE, Creutzfeldt W (1992) Comparison of the effect of GIP and GLP- 1 (7-36 amide) on insulin release from rat pancreatic islets. Eur. J. Clin. Invest. 22:154-157. Sinclair JD (1988) Multiple t tests are appropriate in science. TIPS 9:12-13. Sirinek KR, Cataland S, ODorisio TM, Mazzaferri EL, Crockett S. Pace WG (1977a) Augmented gastric inhibitory polypeptide response to intraduodenal glucose by exogenous gastrin and cholecystokinin. Surg. 82:438-442. Sirinek KR, Cataland S, O’Dorisio TM, Mazzaferri EL, Crockett S, Pace WG (1977b) Sympathetic neuroendocrine regulation of gastric inhibitory polypeptide response to glucose. Surg. Forum 28:377-379. Sirinek KR, Levine BA, O’Dorisio TM, Cataland S (1981) Parasympathetic neuroendocrine regulation of GIP response to glucose. J. Surg. Res. 30:343-347. Sirinek KR, Levine BA, O’Dorisio TM, Cataland S (1983) Gastric inhibitory polypeptide (GIP) release by actively transported, structurally similar carbohydrates. Proc. Soc. Exp. Biol. Med. 173:379-385. Sirinek KR, O’Dorisio TM, Levine BA, Cataland S (1979) Selective release of gastric inhibitory polypeptide by some actively transported, structurally similar carbohydrates. Surg. Forum 30:358-360. Sirinek KR, Pace WG, Crocket SE, O’Dorisio TM, Mazzaferri EL (1978) Insulin-induced attenuation of glucose-stimulated gastric inhibitory polypeptide secretion. Am. J. Surg. 135:151-155. Smith PH, Merchant FW, Johnson DG, Fujimoto WY, Williams RH (1977) Immunocytochemical localization of gastric inhibitory polypeptide-like material within A-cells of the endocrine pancreas. Am. J. Anat. 149:585-590. Soll AH, Yamada T, Park J, Thomas LP (1984) Release of somatostatin-like immunoreactivity from canine fundic mucosal cells in primary culture. Am. 3. Physiol. 247:G558-G566. Someya Y, Inagaki N, Maekawa T, Seino Y, Ishii S (1993) Two 3’,5’-cyclic-adenosine monophosphate response elements in the promoter region of the human gastric inhibitory polypeptide gene. Fed. Eur. Biochem. Soc. 3 17:67-73. Soon-Shiong P, Debas HT, Brown JC (1979) The evaluation of gastric inhibitory polypeptide (GIP) as the enterogastrone. 3. Surg. Res. 26:681-686.  185 Starling EH (1905) The Croonian lectures on: the chemical correlation of the functions of the body. The Lancet 2:339-341. Stockmann F, Ebert R, Creutzfeldt W (1984) Preceding hyperinsulinemia prevents demonstration of insulin effect on fat-induced gastric inhibitory polypeptide (GIP). Diabetes 33:580-585. Stoll RW, Touber IL, Menahan LA, Williams RH (1970) Clearance of porcine insulin, proinsulin, and connecting peptide by the isolated rat liver. Proc. Soc. Exp. Biol. Med. 133:894-896. Suzuki S, Kawai K, Ohashi S, Watanabe Y, Yamashita K (1992) Interaction of glucagon like peptide-1(7-36)amide and gastric inhibitory polypeptide or cholecystokinin on insulin and glucagon secretion from the isolated perfused rat pancreas. Metabol. 41:359-363. Suzuki S, Kawai K, Ohashi S. Mukai H, Murayama Y, Yamashita K (1990) Reduced insulinotropic effects of glucagonlike peptide I-(7-36)-amide and gastric inhibitory polypeptide in isolated perfused diabetic rat pancreas. Diabetes 39:1320-1325. Suzuki Y, Erickson RH, Sedimayer A, Chang SK, Ikehara Y, Kim YS (1993) Dietary regulation of rat intestinal angiotensin-converting enzyme and dipeptidyl peptidase IV. Am. J. Physiol. 264:G1 153-G1 159. Sykes S, Morgan LM, English J, Marks V (1980) Evidence for preferential stimulation of gastric inhibitory polypeptide secretion in the rat by actively transported carbohydrates and their analogues. J. Endocrinol. 85:201-207. Szecowka 3, Grill V, Sandberg E, Efendic S (1982a) Effect of Gil’ on the secretion of insulin and somatostatin and the accumulation of cyclic AMP in vitro in the rat. Acta Endocrinol. (Copenh.) 99:416-421. Szechowka 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. Takahashi H, Manaka H, Suda K, Fukase N, Sekikawa A, Eguchi H, Tominaga M, Sasaki H (1991) Hyperglycaemia but not hyperinsulinaemia prevents the secretion of glucagon-like peptide-1 (7-36 amide) stimulated by fat ingestion. Scand. 3. Clin. Lab. Invest. 51:499-507. Takasawa M, Kasahara S, Igarashi K (1990) Urinary glycylprolyl dipeptidylaminopeptidase acitivity in diabetes mellitus. 3. Jpn. Diabetic Soc. 33:287291. Takeda 3, Seino Y,Tanaka KI, Fukumoto H, Kayano T, Takahashi H, Mitani T, Kurono M, Suzuki T, Tobe T (1987) Sequence of an intestinal cDNA encoding human gastric inhibitory polypeptide precursor. Proc. Nat!. Acad. Sci. USA 84:7005-7008. Taraskevich PS, Douglas WW (1984) Electrical activity in adenohypophyseal cells and effects of hypophysiotropic substances. Federation Proc. 43:2373-2378.  186 Taylor IL, Feldman M (1982) Effect of cephalic-vagal stimulation on insulin, gastric inhibitory polypeptide, and pancreatic polypeptide release in humans. J. Clin. Endocrinol. Metab. 55:1114-1117. Thams P. Capito K, Hedeskov CJ (1982) Different effects of Ca -calmodulin on 2 adenylate cyclase activity in mouse and rat pancreatic islets. Biochem. 3. 206:97-102. Thomas FB, Mazzaferri EL, Crockett SE, Mekhjian HS, Gruemer ND, Cataland S (1976) Stimulation of secretion of gastric inhibitory polypeptide and insulin by intraduodenal amino acid perfusion. Gastroenterol. 70:523-527. Thomas FB, Sinar D, Mazzaferri EL, Cataland 5, Mekhjian HS, Caidwell JH, Fromkes 33 (1978) Selective release of gastric inhibitory polypeptide by intraduodenal amino acid perfusion in man. Gastroenterol. 74:1261-1265. Thomford NR, Sirinek KR, Crockett SE, Mazzaferri EL, Crockett 5, Pace WG (1974) Gastric inhibitory polypeptide response to oral glucose after vagotomy and pyloroplasty. Arch. Surg. 109:177-182. Thorens B (1993) Facilitated glucose transporters in epithelial cells. Annu. Rev. Physiol. 55:591-608. Ticku MK, Mehta AK (1990) y-aminobutyric acid receptor desensitization in mice spinal cord cultured neurons: lack of involvement of protein kinases A and C. Molec. Pharm. 38:719-724. Tiruppathi C, Miyamoto Y, Ganapathy V, Roesel RA, Whitford G, Leibach FH (1990) Hydrolysis and transport of proline-containing peptides in renal brush-border membrane vesicles from dipeptidyl peptidase TV-positive and dipeptidyl peptidase TV-negative rat strains. 3. Biol. Chem. 265:1476-1483. Tischler AS (1983) Gut peptide production by tumor cell lines in culture. Scand. 3. Gastroenterol. [Suppl.] 18:33-43. Tronier B, Deigard A, Andersen T, Madsbad 5 (1985) Absence of incretin effect in obese type II and diminished effect in lean type IT and obese subjects. Diabetes Res. Clin. Pract. [Suppl. 1] 1:S568 (Abstract). Trube G, Rorsman P, Ohno-Shosaku T (1986) Opposite effects of tolbutamide and diazoxide on the ATP-dependent K channel in mouse pancreatic B-cells. Pfluegers Arch. 407:493-499. Umezawa H, Aoyagi T, Ogawa K, Naganawa H, Hamada M, Takeuchi T (1984) Diprotins A and B, inhibitors of dipeptidyl aminopeptidase IV, produced by bacteria. 3. Antibiotics XXXVII:422-425. Unger RH, Eisentraut AM (1969) Entero-insular axis Arch. Intern. Med. 123:261-266. Usdin TB, Mezey E, Button DC, Brownstein MJ, Bonner TI (1993) Gastric inhibitory polypeptide receptor, a member of the secretin-vasoactive intestinal peptide receptor family, is widely distributed in periperal organs and the brain. Endocrinology 133:2861-2870.  187 Usellini L, Capella C, Solcia E, Buchan AMJ, Brown IC (1984) Ultrastructural localization of gastric inhibitory polypeptide (Gil’) in a well characterized endocrine cell of canine duodenal mucosa. Histochem. 80:85-89. Verchere CB (1991) The role of gastric inhibitory polypeptide in the regulation of pancreatic endocrine secretion. Vancouver, Canada: Univ. of B.C. PhD dissertation. Verchere CB, Kwok YN, Brown JC (1991) Modulation of acetyicholine-stimulated insulin release by glucose and gastric inhibitory polypeptide. Pharmacol. 42:273-282. Verdonk CA, Rizza RA, Nelson LE, Go VLW, Gerich JE, Service FJ (1980) Interaction of fat-stimulated gastric inhibitory polypeptide on pancreatic alpha and beta cell function. J. Clin. Invest. 65:1119-1125. Verspohl EJ, Ammon HPT (1987) Cholecystokinin (CCK8) regulates glucagon, insulin and somatostatin secretion from isolated islets: interaction with glucose. Pfugers Archives 410:284-287. Vigna SR, Mantyh CR, Giraud AS, Soll AH, Walsh JH, Mantyh PW (1987) Localization of specific binding sites for bombesin in the canine gastrointestinal tract. Gastroenterol. 93:1287-95. Villar HV, Fender HR, Rayford PL, Bloom SR, Ramus NI, Thompson IC (1976). Suppression of gastrin release and gastric secretion by gastric inhibitory polypeptide (GIP) and vasoactive intestinal polypeptide (VIP). Annals Surg. 184:97-102. Visher U, Blondel B, Wollheim CB, Hoppner W, Seitz HI, lynedjian PB (1987) Hexokinase isoenzymes in RIN-m5F insulinoma cells. Biochem. 1. 241: 249-255. Wahi MA, Plehn RI, Landsbeck EA, Verspohi EJ, Ammon HPT (1992) Are ionic fluxes of pancreatic beta cells a target for gastric inhibitory polypeptide? Mol. Cell. Endocrinol. 90:117-123. Walter R, Simmons WH, Yoshimoto T (1980) Proline specific endo- and exopeptidases. Mol. Cell. Biochem. 30:111-126. Watanabe K, Gola M (1987) Forskolin interaction with voltage-dependent K channels in Helix is not mediated by cyclic nucleotides. Neurosci. Lett. 78:211-216. Watanabe Y, Kojima T, Fujimoto Y (1987) Deficiency of membrane-bound dipeptidylaminopeptidase IV in a certain rat strain. Experientia 43:400-401. Weir GC, Mosjov S. Hendrick GC, Habener IF (1989) Glucagonlike peptide 1 (7-37) actions on endocrine pancreas. Diabetes 38:338-342. Wettergren A, Schjoldager B, Mortensen PE, Myhre J, Christiansen I, Holst II (1993) Truncated GLP-1 (proglucagon 78-107-amide) inhibits gastric and pancreatic functions in man. Dig. Dis. Sci. 38:665-673. Whitcomb DC, O’Dorisio TM, Cataland S, Nishikawara MT (1984) In vivo binding of insulin to target tissues using a new in vivo radioreceptor assay. Dig. Dis. Sci. 29:95S (Abstract).  188 Williams RH, Biesbroek TB (1980) Gastric inhibitory polypeptide and insulin secretion after infusion of acetylcholine, catecholamines, and gut hormones. Proc. Soc. Exp. Biol. Med. 163:39-47. Williams RH, May 3M, Biesbroek JB (1981) Determinants of gastric inhibitory polypeptide and insulin secretion. Metab. 30:36-40. Wilims B, Ebert R, Creutzfeldt W (1978) Gastric inhibitory polypeptide (GIP) and insulin in obesity: II. Reversal of increased responses to stimulation by starvation or food restriction. Diabetologia 14:379-387. Wishart J, Morris HA, Horowitz M (1992) Radioimmunoassay of gastric inhibitory polypeptide in plasma. Clin. Chem. 38:2156-2157. Wolfe MM, Hocking MP, Maico DG, McGuigan TB (1983) Effects of antibodies to gastric inhibitory peptide oa gastric acid secretion and gastrin release in the dog. Gastroenterol. 84:941-948. Wolfe MM, McGuigan JR (1982) Release of gastric inhibitory peptide following a peptone meal in the dog. Gastroenterol. 83:864-872. Wright E (1993) The intestinal Na/glucose transporter. Annu. Rev. Physiol. 55:575-589. Yada T, Itoh K, Nakata M (1993) Glucagon-like peptide-1-(7-36)amide and a rise in cyclic adenosine 3’,5’-monophosphate increase cytosolic free Ca 2 in rat pancreatic Bcells by enhancing Ca 2 channel activity. Endocrinology 133:1685-1692. Yajima H, Fujii N, Akaji K, Sakurai M, Nomizu M, Misuta K, Aona M, Moriga M, Inoue K, Histotani R, Tobe T (1985) Synthesis of a 42-residue peptide corresponding to the entire amino acid sequence of human GIP. Chem. Pharm. Bull. (Tokyo) 33:35783580. Yamagishi T, Debas HT (1980) Gastric inhibitory polypeptide (GIP) is not the primary mediator of the enterogastrone action of fat in the dog. Gastroenterol. 78:931-936. Yaron A, Naider F (1993) Proline-dependent structural and biological properties of peptides and proteins. Crit. Rev. Biochem. Mo!. Biol. 28:31-81. Yoshioka S (1986) Inhibitory effects of gymnemic acid and an extract from the leaves of Zizyphus jujuba on glucose absorption in the rat small intestine. J. Yonago Med. Assoc. 37:142-154. Yoshiya K, Yamamura T, Ishikawa Y, Utsunomiya 3, Takemura 3, Takeda 3, Seinao Y, Imura H (1985) Effect of truncal vagotomy on GLP release induced by intraduodenal glucose or fat in dogs. Digest. 31:41-46. Zawalich WS, Rasmussen H (1990) Control of insulin secretion: a model involving Ca , 2 cAMP and diacyglycerol. Mol. Cell. Endocrinol. 70:119-137. Zawalich WS, Takuwa N, Takuwa Y, Diaz VA, Rasmussen H (1987) Interactions of cholecystokinin and glucose in rat pancreatic islets. Diabetes 36:426-433.  ççz-€z dd A1 wsA puiisiwons9 qj (ojois(qj jo jooqpui :uI uo!lDunJ put uoiirniuaijjip ii’ Iu!1s’u jo sipnis u suq painno Jo pluAno’] ‘ 1ssi9 ‘j qunqw mnqp (1661)  u°  a  ‘v  vv-OVTE OSAqJ iui qai s1JDud  np uiui uix suoraxs i, sj suowpi :s11Du1d np uJ1u! uouis auu ij p snbiojoisAqd SUO!11I1A sp pni j suoT1nquwo (6z6T) f uiqrT ‘ zun  i-pdd  E8-LLVV icIj jn’>j SSuATSuodsJ Jp g uo (66T) H USSflWSJ ‘D)I P!T’”Z ‘SM  muonj Jo unuui  68 T  

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088076/manifest

Comment

Related Items